Surface Interactions of Surface Washing Agentshomepages.uc.edu/~sorialga/Karen PhD Dissertation.pdf · Interfacial tension and contact angle were measured for PBC in the presence

Surface Interactions of Surface Washing Agents: An Examination of Detergency, Interfacial Tension and Contact Angle

A dissertation submitted to

Division of Research and Advanced Studies of the University of Cincinnati

in partial fulfillment of the requirements for the degree of

DOCTOR OF PHILOSPHY

in the Department of Civil and Environmental Engineering of the College of Engineering

2007

by

Karen M. Koran

B.S., Biological Science, Cedarville University, 1992 M.S., Environmental Science, University of Cincinnati 1999

Committee Members: Dr. George A. Sorial (Chair) Dr. Makram T. Suidan, Dr. Albert D. Venosa, Dr. Paul L. Bishop

i

ABSTRACT

A laboratory testing protocol to evaluate the effectiveness of surface washing agents (SWAs) to

remove crude oil from solid substrates was developed in this study. Variables were tested to

determine their effect on SWA performance using Prudhoe Bay Crude (PBC) oil as a

representative oil. The protocol was most sensitive to the SWA-to-oil ratio (SOR), rotational

speed of mixing, and oil weathering; it was not greatly affected by volume of oil applied, oil-

SWA contact time, mixing time, or SWA concentration (when total applied mass was constant).

The protocol was tested using wet and dry substrates. Application of oil to dry substrates is

preferred based on lower variability and greater differential between treatments and controls.

Interfacial tension and contact angle were measured for PBC in the presence of 5 SWAs at 3

concentrations. SWAs were ranked based on 1) efficiency under the developed protocol, 2)

ability to reduce interfacial tension and 3) ability to increase oil-substrate contact angle. Four of

six SWAs had the same relative rankings under each of these criteria. The other two ranked high

based on ability to lower interfacial tension, but low for an increase in contact angle. The

efficiency of these two products under the protocol was mid to low. The best predictor of SWA

performance at high SOR was the term representing the difference between SWA-substrate and

oil-substrate interfacial tensions (σSC - σSO). This term is a measure of the preferential wetting of

the substrate surface by either the oil or the SWA wash water. A high correlation was observed

between SWA efficiency at 10:1 SOR and the σSC - σSO difference at 5% SWA solution

concentration. At low SOR, CMC was a better predictor of SWA performance.

ii

iii

ACKNOWLEDGMENTS

I would like to express my deepest gratitude to my advisor, Dr. George Sorial, for his guidance,

instruction and support during the course of my study. I am grateful to the members of my

committee, Dr. Makram Suidan, Dr. Albert Venosa, and Dr. Paul Bishop, for their advice,

contributions and support. I also would like to thank the U.S. Environmental Protection Agency

for funding this work.

I would like to thank colleagues and friends who helped me with my research and in so many

other ways. Special thanks go to Christopher Luedeker, Sylian Rodriguez, Satish Vishnubhatla

and Saurabh Vyas for their contributions to this work. Thanks also to Christopher Rulison,

Christina Bennett-Stamper, Dennis King and Keith Dennison for their help with interfacial

tension and contact angle measurements, SEM imaging, and statistical support. Special thanks to

my friends, especially Jennifer and Dominic Boccelli, Margaret Kupferle, and Moustafa

Moteleb, for their encouragement and for providing much needed diversions along the way.

My very special thanks to my husband, Joe, for his love and so much more; to Ethan and Colin,

my boys, the inspiration for everything I do; and to my parents, for their love and encouragement

and for always believing in me. This dissertation is dedicated to them.

iv

TABLE OF CONTENTS

Abstract ............................................................................................................................................ i

Acknowledgments.......................................................................................................................... iii

Table of Contents ........................................................................................................................... iv

List of Figures ..................................................................................................................................x

List of Tables .............................................................................................................................. xvii

1.0 Introduction .............................................................................................................................1

1.1 Overview .............................................................................................................................1

1.2 Surface Washing Agents ......................................................................................................2

1.2.1 Considerations for Use .............................................................................................2

1.2.2 Classification ...........................................................................................................4

1.2.3 Surfactant Structure and Function ..........................................................................5

1.2.4 Interfacial Phenomena and Mechanism of Action ..................................................8

1.2.5 Studies of Oil/Water/Surfactant Systems ...............................................................10

1.3 SWA Effectiveness Testing ..............................................................................................14

1.3.1 Effectiveness of SWAs in Field and Laboratory Studies ......................................14

1.3.2 Previous Laboratory Testing Protocols .................................................................17

1.4 Research Objectives .........................................................................................................20

1.4.1 Protocol Development and Testing ........................................................................20

1.4.2 Surface Interactions and Detergency .....................................................................21

2.0 Material and Methods ..........................................................................................................23

v

2.1 Materials ............................................................................................................................24

2.1.1 Crude Oil ..............................................................................................................24

2.1.2 SWAs and Surfactants ..........................................................................................24

2.1.3 Synthetic Seawater .................................................................................................24

2.1.4 Extraction Solvent ..................................................................................................25

2.1.5 Substrates ..............................................................................................................25

2.1.6 Experimental Apparatus ........................................................................................26

2.2 Experimental Procedures ...................................................................................................28

2.2.1 Glass Plate Testing Procedure ...............................................................................28

2.2.2 Natural Substrate Testing Procedure .....................................................................29

2.2.3 Extraction Methods ................................................................................................31

2.2.3.1 Aqueous Phase Extraction ......................................................................31

2.2.3.2 Glass Plate Extraction .............................................................................32

2.2.3.3 Natural Substrate Extraction ...................................................................32

2.3 Analytical Methods ...........................................................................................................34

2.3.1 UV-Vis Spectrophotometry: .................................................................................34

2.3.1.1 Oil Stock Standard Preparation ...............................................................34

2.3.1.2 Preparation of Standards from the Stock Standard Solutions .................35

2.3.1.3 Instrument Conditions and Calibration ...................................................36

2.3.2 Gas Chromatography/Mass Spectrometry .............................................................37

2.3.3 Spectrofluorometry ................................................................................................37

2.3.4 Thin Layer Chromatography ..................................................................................37

2.3.5 Surface Tension for Determining Critical Micelle Concentration .........................38

vi

2.3.6 Interfacial Tension .................................................................................................39

2.3.7 Contact Angle Measurement..................................................................................39

2.4 Identification of Significant Experimental Parameters ......................................................40

3.0 Preliminary Testing ...............................................................................................................47

3.1 Glass Plate Protocol ...........................................................................................................48

3.2 Fritted Disc Protocol ..........................................................................................................51

3.3 Revised Glass Plate Protocol .............................................................................................53

3.3.1 Oil Application.......................................................................................................53

3.3.2 Weathering Time ...................................................................................................55

3.3.3 SWA Application ...................................................................................................56

3.3.4 Oil-SWA Contact Time .........................................................................................57

3.3.5 Washing .................................................................................................................57

3.3.6 Testing Procedure .................................................................................................58

3.3.7 Effect of Plate Roughness on SWA Performance ..................................................60

3.3.8 Effect of SWA Dilution and SWA-Oil Ratio ........................................................60

3.3.9 Reduced Efficiency and Reproducibility ...............................................................61

4.0 Development of the Natural Substrate Protocol .................................................................70

4.1 Materials ............................................................................................................................70

4.2 Experimental Design ..........................................................................................................72

4.3 Effect of Testing Variables on Protocol .............................................................................73

4.3.1 Substrate Type .......................................................................................................73

4.3.2 Substrate Hydration and Drainage Time Effects ...................................................75

4.3.3 Mode and Pattern of Oil Application .....................................................................77

vii

4.3.4 Oil Weathering Time .............................................................................................78

4.3.5 Mode and Volume of SWA Application ...............................................................79

4.3.6 Oil-SWA Contact Time .........................................................................................80

4.3.7 SWA Dilution Effects ............................................................................................80

4.3.8 Oil Volume and SWA-to-Oil Ratios ......................................................................84

4.3.9 Mixing Speed and Time .........................................................................................87

4.3.10 Oil Type ...............................................................................................................88

4.3.10.1 SLC and Bunker C ................................................................................88

4.3.10.2 IFO-180 .................................................................................................89

4.4 Other Considerations for Use: Dispersability and Toxicity ..............................................91

4.5 Conclusions ........................................................................................................................92

5.0 Analytical Methodologies: Comparisons and Limitations ..............................................113

5.1 Spectrophotometry ...........................................................................................................114

5.1.1 Theory ..................................................................................................................114

5.1.2 Analytical Biases and Interferences .....................................................................115

5.1.2.1 Absorbance Interference Due to SWAs ................................................115

5.1.2.2 Water in DCM Emulsions .....................................................................116

5.2 Gas Chromatography/Mass Spectrometry .......................................................................117

5.2.1 Theory ..................................................................................................................117

5.2.2 Bias ......................................................................................................................118

5.2.3 Results ..................................................................................................................118

5.3 Spectrofluorometry ..........................................................................................................119

5.3.1 Theory ..................................................................................................................119

viii

5.3.2 Bias ......................................................................................................................120

5.3.3 Results ..................................................................................................................120

5.4 Thin Layer Chromatography............................................................................................121

5.4.1 Theory ..................................................................................................................121

5.4.2 Bias ......................................................................................................................121

5.4.3 Results ..................................................................................................................122

5.5 Conclusions ......................................................................................................................123

6.0 Testing of Variables: The Fractional Factorial Experiment............................................131

6.1 Design and Methods ........................................................................................................132

6.2 Results ..............................................................................................................................133

6.2.1 Main Effects .........................................................................................................133

6.2.2 Interactions ...........................................................................................................135

6.2.2.1 SWA Solution Concentration Interactions............................................135

6.2.2.2 SOR Interactions ...................................................................................136

6.2.2.3 Oil-SWA Contact Time ........................................................................137

6.2.2.4 Mixing Speed by Mixing Time Interaction ..........................................138

6.3 Conclusions ......................................................................................................................138

7.0 Comparison of Glass Plate and Natural Substrate Protocols ..........................................152

7.1 Physical and Chemical Properties of Substrates ..............................................................152

7.2 Surface Roughness ...........................................................................................................153

7.2.1 Scanning Electron Microscopy ............................................................................154

7.2.2 Other Methods for Determining Surface Area .....................................................157

7.2.3 Sand Particle Size Distribution ............................................................................159

ix

7.2.4 Sand Sphericity and Shape Factor .......................................................................159

7.3 Comparison of Experimental Results ..............................................................................160

7.4 Conclusions ......................................................................................................................161

8.0 Interfacial Chemistry of Surface Washing Agents ...........................................................175

8.1 Critical Micelle Concentration .........................................................................................175

8.1.1 Method .................................................................................................................176

8.1.2 Results .................................................................................................................177

8.2 Interfacial Tension ...........................................................................................................178

8.2.1 Method .................................................................................................................179

8.2.2 Results ..................................................................................................................181

8.3 Contact Angle and Competitive Wetting .........................................................................182

8.3.1 Method .................................................................................................................183

8.3.2 Results .................................................................................................................184

8.4 SWA Performance Under Proposed Protocol .................................................................185

8.5 Conclusions ......................................................................................................................189

9.0 Conclusions ...........................................................................................................................205

9.1 Protocol Comparison .......................................................................................................205

9.2 Testing of the Natural Substrate Protocol ........................................................................207

9.3 Interfacial Activity of SWAs ...........................................................................................210

9.4 Recommendations for Future Work.................................................................................211

References ...................................................................................................................................214

x

LIST OF FIGURES

Figure 1.1: Surfactant Induced Increase in Contact Angle ............................................................22

Figure 2.1: Rotational Mixer with Stainless-Steel Shafts, Glass Plates and Baffled

Beakers Used in the Glass Plate Protocol ......................................................................................41

Figure 2.2: Oiled Glass Plates Before and After Treatment with SWA Solution..........................41

Figure 2.3: Orbital Shaker Table, Beakers and Baskets Used in the Natural Substrate

Protocol ..........................................................................................................................................42

Figure 2.4: Templates for 5-Drop and 12-Drop Oil Application Pattern on Glass Plates .............43

Figure 2.5: Template for Application of Oil and SWA to Natural Substrates in 2”x2”

Wire Mesh Baskets ........................................................................................................................43

Figure 3.1 – Pictures of Experimental Apparatus, Including a) Glass Plates, Rotational

Mixer and Wash Jars; b) XCS and P Glass Plates After Washing ................................................62

Figure 3.2 – Removal of Weathered PBC from Glass Plates in Controls and 10%

Petroclean 2:1 SOR Treatments .....................................................................................................63

Figure 3.3 – Removal of Weathered PBC from Glass Plates in Controls and 10%


Figure 3.4 – Fritted Disc After Application of PBC Oil ................................................................64

Figure 3.5 – Rotational Mixer with Custom Designed Shafts, Etched Glass Plates, and

Baffled Beakers ..............................................................................................................................64

Figure 3.6 – Templates for a) Five and b) Twelve Drop Oil Application Patterns .......................64

xi

Figure 3.7 – Removal of Weathered PBC from Coarse and Extra Coarse Plates by

Corexit: Comparison of 5- and 12-Drop Oil Application Patterns ................................................65

Figure 3.8 – Removal of PBC from Coarse Glass Plate at Varied Weathering Times ..................65

Figure 3.9 – Effect of Oil-SWA Contact Time on Removal of Weathered PBC Oil from

Finely Ground Glass Plates by Corexit at 3:1 SOR .......................................................................66


Coarse Glass Plates by Biosolve at 10:1 SOR ...............................................................................66


Extra Coarse Glass Plates by Biosolve at 10:1 SOR .....................................................................67

Figure 3.12 – Release of PBC from Smooth, Fine and Coarse Plates in Controls and 20%


Figure 3.13 – Release of PBC from Smooth, Fine and Coarse Plates in Controls and

Corexit 1:1 SOR Treatments ..........................................................................................................68

Figure 3.14 – Effect of Plate Etching on PBC Release from SWA Treated Glass Plates .............68

Figure 3.15 – SWA Efficiency as a Function of Dilution and SOR ..............................................69

Figure 4.1 – Oiled Sand Before and After Washing for Control and Treatment ...........................94

Figure 4.2 – Effect of Substrate Type on Aquaclean Effectiveness to Remove PBC from

Dry and Wet Substrates .................................................................................................................95

Figure 4.3 – Removal of PBC Oil from Dry and Wet Sand for Untreated Control and

Three SWAs after 15 min Weathering ...........................................................................................95

Figure 4.4 – Effect of Substrate Drain Time on PBC Removal from Wet Sand after 15

min Weathering ..............................................................................................................................96

Figure 4.5 – PBC Removal by SWAs Under Dry and Wet Sand Applications.............................96

xii

Figure 4.6 – Comparison of Effectiveness Data for Two Operators Using Undiluted

SWAs on Wet and Dry Sand at 2:1 SOR .......................................................................................97

Figure 4.7 – Comparison of Effectiveness Data for Two Operators Using 50% Aqueous

SWA Solutions on Wet and Dry Sand at 1:1 SOR ........................................................................97

Figure 4.8 – Effect of Oil Weathering Time on PBC Release from Wet Sand .............................98

Figure 4.9 - Effect of contact time on removal of PBC by Corexit and Petroclean ......................98

Figure 4.10 – Effect of SWA Dilution with Different Application Mass for Aquaclean ..............99

Figure 4.11 – Effect of SWA Dilution with Different Application Mass for Petroclean ..............99

Figure 4.12 – Effect of SWA Dilution with Different Application Mass for Petrotech ..............100

Figure 4.13 – Effect of SWA Dilution with Different Application Mass for Superall ................100

Figure 4.14 – Effect of SWA Dilution with Same Application Mass for Aquaclean ..................101

Figure 4.15 – Effect of SWA Dilution with Same Application Mass for Petroclean ..................101

Figure 4.16 – Effect of SWA Dilution with Same Application Mass for Petrotech ....................102

Figure 4.17 – Effect of SWA Dilution with Same Application Mass for Superall .....................102

Figure 4.18 – Effect of Oil Volume on SWA Effectiveness at Four SOR ..................................103

Figure 4.19 – Effect of SOR on Aquaclean Effectiveness at Four Oil Volumes.........................104

Figure 4.20 - Effect of Oil Volume on Aquaclean Effectiveness at Three SOR .........................104

Figure 4.21 – Effect of SOR on Petrotech Effectiveness at Four Oil Volumes ...........................105

Figure 4.22 - Effect of Oil Volume on Petrotech Effectiveness at Three SOR ...........................105

Figure 4.23 – Effect of SOR on Oil Removal from Dry Sand .....................................................106

Figure 4.24 – Effect of Mixing Speed and Mixing Time on Removal of PBC from Wet

Sand in Untreated Controls after 18 Hours Weathering ..............................................................107

xiii

Figure 4.25 – Effect of Mixing Speed and Mixing Time on Removal of PBC from Wet

Sand in Untreated Controls after 18 Hours Weathering ..............................................................107

Figure 4.26 – Effect of Drain time on Release of Bunker C and South Louisiana Crude

Oils from Wet Sand after 18 Hours Weathering ..........................................................................108

Figure 4.27 - Effectiveness of Undiluted Aquaclean to Remove PBC, IFO-180 and

Bunker C from Wet Sand after 18 Hours Weathering .................................................................108

Figure 4.28 – Comparison of Weathering Times for Untreated IFO-180 Controls.....................109

Figure 4.29 – Comparison of Weathering Times for IFO-180 Treated with Undiluted

Aquaclean at 2:1 SOR ..................................................................................................................109

Figure 4.30 – Effect of Contact Time on Release of IFO-180 from Wet Sand by

Aquaclean and Petroclean ............................................................................................................110

Figure 4.31 – Effect of Mixing Speed and Mixing Time on Removal of IFO-180 from

Wet Sand in Untreated Controls after 18 Hours Weathering ......................................................111

Figure 4.32 - Effect of Mixing Speed and Mixing Time on Removal of IFO-180 from

Wet Sand in Aquaclean Treatments after 18 Hours Weathering .................................................111

Figure 5.1 – Mass Alkane or PAH by GC/MS vs. Mass of Oil by Spectrophotometer. .............124

Figure 5.2 – Percent Recovery of Alkanes and PAHs by GC/MS vs. Percent Recovery of

Oil by Spectrophotometer ............................................................................................................124

Figure 5.3 – Percent PAH Recovery vs. Percent Alkane Recovery by GC/MS. .........................125

Figure 5.4 – Percent Recovery of Alkanes, PAHs, Resins, and Asphaltenes by Iatroscan

versus Percent Recovery of Total Oil by Spectrophotometer ......................................................125

Figure 5.5 – Percent Recovery of Alkanes by Iatroscan versus Percent Recovery of Total

Oil by Spectrophotometer. ...........................................................................................................126

xiv

Figure 5.6 – Percent Recovery of PAHs by Iatroscan versus Percent Recovery of Total


Figure 5.7 – Percent Recovery of Resins by Iatroscan versus Percent Recovery of Total


Figure 5.8 – Percent Recovery of Asphaltenes by Iatroscan versus Percent Recovery of

Total Oil by Spectrophotometer...................................................................................................127

Figure 5.9 – Percent Recovery of Alkanes by Iatroscan vs. Percent Recovery of PAHs by

Iatroscan. ......................................................................................................................................128

Figure 6.1 – Interaction of SWA Solution Concentration and SOR ............................................141

Figure 6.2 – Interaction of SWA Solution Concentration and Oil SWA Contact Time ..............141

Figure 6.3 – Interaction of SWA Solution Concentration and Mixing Speed .............................142

Figure 6.4 – Interaction of SWA Solution Concentration and Oil SWA Mixing Time ..............142

Figure 6.5 – Interaction of SOR and Oil SWA Contact Time .....................................................143

Figure 6.6 – Interaction of SOR and Mixing Speed ....................................................................143

Figure 6.7 – Interaction of SOR and Mixing Time ......................................................................144

Figure 6.8 – Interaction of Oil-SWA Contact Time and Mixing Speed ......................................144

Figure 6.9 – Interaction of Oil-SWA Contact Time and Mixing Time .......................................145

Figure 6.10 – Interaction of Mixing Speed and Mixing Time .....................................................145

Figure 7.1 – SEM Images of Glass Plates at 100x Magnification ...............................................162

Figure 7.2 – SEM Images of a) ASTM 20/30 Sand at 160x Magnification and b) FilPro

Gravel at 40x Magnification ........................................................................................................163

Figure 7.3 – Replicate SEM Images of Fine Glass Plates at 100x Resolution ............................164

Figure 7.4 – Replicate SEM Images of Medium Glass Plates at 100x Resolution ......................165

xv

Figure 7.5 – Replicate SEM Images of Coarse Glass Plates at 100x Resolution ........................166

Figure 7.6 – Digitized 2-D and 3-D Images and Cross-Sectional Profile of Fine Glass

Plate at 100x Resolution ..............................................................................................................167

Figure 7.7 – Digitized 2-D and 3-D Images and Cross-Sectional Profile of Medium Glass


Figure 7.8 – Digitized 2-D and 3-D Images and Cross-Sectional Profile of Coarse Glass


Figure 7.9 – Cross-Sectional Profile of Fine, Medium and Coarse Plates Scaled to

Relative Size ................................................................................................................................170

Figure 7.10 – SEM Images and Particle Characteristics for ASTM 20/30 Sand .........................171

Figure 7.11 – Particle Size Distribution for Ten Replicate Sand Subsamples .............................172

Figure 7.12 – Removal of PBC from Coarse Plate and Sand at 1:1 SOR ...................................173

Figure 7.13 – Removal of PBC from Coarse Plate and Sand at 10:1 SOR .................................173

Figure 8.1 – Schematic Drawing of a Micelle .............................................................................191

Figure 8.2 – Method for Determining CMC from Surface Tension Data with Incremental

Changes in Solution Concentrations ............................................................................................192

Figure 8.3 – Determination of CMC for Five SWAs...................................................................192

Figure 8.4 – Schematic Drawing of Pendant Drop Method for Determining Interfacial

Tension .........................................................................................................................................193

Figure 8.5 – Measurement of Contact Angle for Oil on Sandstone Surface ...............................193

Figure 8.6: Contact Angles for PBC Oil on Sandstone in 100% SWA ......................................194

Figure 8.7: Contact Angles for PBC Oil on Sandstone in 50% Aqueous Solution ....................195

Figure 8.8: Contact Angles for PBC Oil on Sandstone in 5% Aqueous Solution ......................196

xvi

Figure 8.9 – Performance and Relative Ranking of SWAs at Four SOR ....................................197

Figure 8.10 – SWA Efficiency as a Function of the Summed Difference Between

Substrate-SWA and Substrate-Oil Interfacial Tensions at Three SWA Concentrations .............198

Figure 8.11 – SWA Efficiency as a Function of the Averaged Difference Between

Substrate-SWA and Substrate-Oil Interfacial Tensions at Three SWA Concentrations .............198

Figure 8.12 – SWA Efficiency at 10:1 SOR as a Function of IFT Difference at Three

SWA Concentrations ...................................................................................................................199

Figure 8.13 - SWA Efficiency at 1:1 SOR as a Function of Critical Micelle Concentration ......200

xvii

LIST OF TABLES

Table 1.1 – Surface Washing Agent Test Results for Environment Canada Protocol ..................22

Table 2.1 – Chemical and Physical Properties of PBC Oil...........................................................44

Table 2.2 – Surface Washing Agents Listed on the NCP Product Schedule .................................45

Table 2.3 – Major Ion Composition of “Instant Ocean” Synthetic Sea Salt..................................46

Table 2.4 – Experimental Levels to be Tested Under the Natural Substrate Protocol ..................46

Table 4.1 – Results of Baffled Flask Test for Dispersability .......................................................112

Table 4.2 – Effectiveness, Dispersability and Toxicity for Five SWAs ......................................112

Table 5.1 – Interference in Spectrophotometric Analysis Caused by 500 µl SWAs ...................129

Table 5.2 – Recovery of 50 µl PBC from Seawater in the presence of 500 µl SWA ..................129

Table 5.3 – Absorbance Values and Measured Masses of Oil in a Typical Sample and

Control Before and After Filtering with Sodium Sulfate.............................................................129

Table 5.4 – Comparison of Analytical Results Measured by UV/Visible

Spectrophotometry and Spectrofluorometry ................................................................................130

Table 6.1 – Main Effects .............................................................................................................146

Table 6.2 – SWA Solution Concentration by SOR Interaction ..................................................146

Table 6.3 – SWA Solution Concentration by Oil–SWA Contact Time Interaction ...................147

Table 6.4 – SWA Solution Concentration by Mixing Speed Interaction ...................................147

Table 6.5 – SWA Solution Concentration by Mixing Time Interaction .....................................148

Table 6.6 – SOR by Oil–SWA Contact Time Interaction ..........................................................148

Table 6.7 – SOR by Mixing Speed Interaction ...........................................................................149

xviii

Table 6.8 – SOR by Mixing Time Interaction .............................................................................149

Table 6.9 – Oil–SWA Contact Time by Mixing Speed Interaction .............................................150

Table 6.10 – Oil–SWA Contact Time by Mixing Time Interaction ............................................150

Table 6.11 – Mixing Speed by Mixing Time Interaction ............................................................151

Table 7.1 – Surface Characteristics for Fine, Medium and Coarse Glass Plates .........................174

Table 8.1 – Critical Micelle Concentrations for Five Water Soluble SWAs ...............................201

Table 8.2 – Measured Densities for Prudhoe Bay Crude Oil and SWA Solutions ......................201

Table 8.3 – Oil–Water Interfacial Tension Values for Undiluted SWAs ....................................202

Table 8.4 – Oil–Water Interfacial Tension Values for 50% Aqueous SWA Solutions ...............202

Table 8.5 – Oil–Water Interfacial Tension Values for 5% Aqueous SWA Solutions .................202

Table 8.6 – Contact Angles for PBC Oil Submerged in Undiluted SWA ...................................203

Table 8.7 – Contact Angles for PBC Oil Submerged in 50% Aqueous SWA Solutions ............203

Table 8.8 – Contact Angles for PBC Oil Submerged in 5% Aqueous SWA Solutions ..............203

Table 8.9 – Difference Between Substrate–SWA and Substrate–PBC Interfacial Tensions ......204

Table 9.1 – Comparison of Testing Protocols Using Corexit 9580 and PBC Oil .......................213

1

Chapter 1

INTRODUCTION

1.1 OVERVIEW

Surface washing agents (SWAs), also known as shoreline cleaning agents, can be used following

an oil spill event to enhance the removal of stranded oil from shoreline surfaces. SWAs are

designed to facilitate the release of stranded oil from substrate surfaces and subsequently transfer

that oil to near-shore receiving waters. In biologically sensitive areas, cleaning agents should not

disperse the oil into the receiving waters or promote oil penetration into permeable shoreline

matrices. Once released, the oil should re-coalesce to form a slick that can be recovered through

physical containment (Clayton, 1993) and mechanical skimming operations.

The use of SWAs as a remediation tool is generally recommended when conventional methods

cannot be employed. Conventional methods for shoreline cleaning include mechanical removal

technologies, such as high-pressure water washing, hot-water washing, steam cleaning, and

physical removal of beach substrate. These techniques, however, are often invasive, impractical,

or detrimental to the natural species within the impacted environment. Bioremediation is also

frequently used to degrade oil stranded on shorelines, and is stimulated through the addition of

nutrients or non-indigenous organisms (Prince et al., 1999; Lee and de Mora, 1999; Lee and

Merlin, 1999; Le Floch et al., 1999; Weise et al., 1999). The use of dispersants is typically

limited to open sea situations, where dispersion of the oil into the water column will have less of

an impact on local biota (National Research Council, 1989).

2

The U.S. EPA’s National Oil and Hazardous Substances Pollution Contingency Plan, or National

Contingency Plan (NCP, U.S. EPA, 1994), governs the use of SWAs, as well as oil dispersants,

biological additives, and other chemical agents. Currently the NCP requires submittal of toxicity

data for all products listed on the NCP Product Schedule. Dispersants and bioremediation agents

must also undergo effectiveness testing in accordance with the published testing protocols

developed by the U.S. EPA: the Baffled Flask Test (BFT) for dispersants (Sorial, et al., 2004a,

2004b; Venosa et al., 2002) and the Bioremediation Agent Test for bioremediation products

(Haines, et al., 2003, 2005). There is currently no protocol for evaluating the effectiveness of

SWAs. Having such a protocol would aid on-scene coordinators in making informed decisions

and enhance the likelihood of a successful clean-up operation.

The goal of this work was to develop a standardized and reproducible testing protocol to evaluate

the cleaning efficiency of SWAs. Oil removal efficiencies for SWA treatments were compared

to the washing efficiency of water without SWA. Since a good SWA should not disperse oil in

water, the dispersability of each SWA was evaluated separately using the BFT (Sorial et al.,

2004a, 2004b). In addition to protocol development, a fundamental study of the surface

interactions involved in detergency was conducted. The roles of critical micelle concentration,

oil-water interfacial tension, and oil-substrate contact angle in predicting SWA effectiveness

were evaluated.

3

1.2 SURFACE WASHING AGENTS

1.2.1 Considerations for Use

Prior to using SWAs in the field, it is important to evaluate the type of oil, geology of the

shoreline, water flow and other physical considerations, and biological sensitivity of the

impacted area. SWAs work best with Type IV heavy crude oil or light and medium crude oils

that are highly weathered (U.S. EPA Region IV). Lighter oils tend to evaporate and degrade

quickly and are not deposited in large quantities on shorelines. Heavier oils will form thick

foams that cling to shoreline substrates. As these oils weather, they form tar balls and asphalt

that are difficult to remove from surfaces. The effectiveness of a SWA to treat the specific oil

and shoreline of interest should be determined on a small scale prior to full scale use. SWAs

should not be used for oils that will sink in water upon mobilization as this will affect benthic

communities (Michel and Galt, 1995).

Shoreline types best suited for use of SWAs include man-made structures, rip-rap, boulders,

cobble, bedrock, etc., that can be cleaned without trapping removed oil in inaccessible spaces.

Marine shorelines range in features, and can have 1) narrow beaches formed from rounded or

flattened cobbles and pebbles, 2) wide beaches covered with a layer of sand or broken shell

fragments, or 3) steep cliffs with no beach at all. Freshwater shorelines are composed of

sediments and may be lined with trees or heavy vegetation. Oil that sticks to suspended sediment

particles or exposed surfaces are amenable to degradation. However, oil that penetrates deep

into the shoreline matrix becomes difficult to remediate. SWAs should be used to treat exposed

surfaces and in areas where the mobilized oil can be contained (U.S. EPA).

4

Physical and environmental conditions at the impacted site will also affect the performance of

SWAs. The following was recommended by U.S. EPA Region IV under its Shoreline Cleaner

Test and Evaluation Protocol:

1. Water Velocity: Current at the impacted area should be less than 1 knot. This will help

ensure refloated oil does not escape containment and contaminate clean environments

down current.

2. Wave Action: The treated area should not be exposed to breaking waves. The cleaning

agents require a soaking time, and continual bombardment will reduce effectiveness of

the SWA.

3. Accessibility: Area must be accessible to observers, monitors, sample collectors, and

workers.

4. Precipitation: Application during heavy rain, sleet, or snow should be avoided. Heavy

precipitation will significantly diminish cleaner effectiveness by reducing the soaking

time.

5. Temperature: If ambient air temperature is below 50EF, special consideration of the

shoreline cleaner's viscosity should be reviewed when selecting it for use. Consult the

manufacturer's recommended application criteria when practical.

Special consideration should be given to environmentally sensitive areas where threatened or

endangered species may be exposed to spilled oil and treatment chemicals. The dangers posed to

animals and plants by the physical properties of the oil include the prevention of respiration,

photosynthesis, or feeding. These must be weighed against the impacts associated with SWA

use, including toxicity of SWAs and the threat associated with mobilization of the oil in the

5

environment. SWAs should not be used near water intakes.

1.2.2 Classification

SWAs are products designed to increase the ease or efficiency of oil removal from shorelines

(Walker, et al.,1993). There are two classifications of SWAs: non-surfactant based solvents and

surfactant based agents. Non-surfactant based solvents act to lower the viscosity of the oil,

thereby facilitating its removal from beach substrate. Following treatment and washing, the

released oil floats to the surface of the receiving waters, where it can be contained and recovered

by mechanical methods. The solvents used for early oil spill clean up contained a high fraction

of aromatic compounds that were toxic to native organisms, and thus are no longer used. Current

formulations contain lower concentrations of aromatic compounds (Walker et al. 1993).

The second class of SWA contains surfactants as well as additives and/or solvents. The principle

active ingredients are the surfactants, which act both to remove the oil from the substrate and to

prevent its redeposition. Additives promote dissolution into the oil, while solvents increase the

solubility of surfactants in the products (Clayton et al., 1995). Surface-active agents can be

further classified as either dispersants or non-dispersants. Dispersants promote the dispersion of

oil into the water following its release from the substrate surface. As stated, this is not desirable

in biologically sensitive areas where toxicity effects are a concern. Surface-active agents that do

not promote oil dispersion contain surface-active wetting agents designed to release oil from

substrates and refloat the oil as a slick on the water surface. The surfactants in these SWAs tend

to be more hydrophilic in nature and preferentially dissolve into the water phase. This allows the

6

oil droplets to re-coalesce and form slicks that can be recovered. This is the category of SWA

that was addressed in this work.

1.2.3 Surfactant Structure and Function

Surfactants are amphipathic in structure, i.e., they contain both a hydrophilic head group and a

hydrophobic (lipophilic) tail group in the molecular structure. The head group is typically ionic

or highly polar and has a strong attraction for the solvent, while the tail group is a long-chain,

non-polar hydrocarbon with little affinity for the solvent. Due to its unique structure, surfactants

tend to align at system interfaces, with the hydrophobic group oriented away from the water and

the hydrophilic group extending into the water. When dissolved in water, the hydrophobic group

causes a distortion of the water structure and increases the free energy of the system. As a result,

less energy is required to bring the surfactant molecule to the surface. Thus, the surfactant

decreases surface tension, which is the amount of work required to create a unit area of surface

(Rosen 1978).

Surfactants are classified according to the nature of the hydrophilic head group, which can be

anionic, cationic, non-ionic, or zwitterionic (Rosen, 1978). Anionic surfactants are the most

common surfactants found in detergents. They contain a negative charge on the surface-active

portion of the molecule and include carboxylic acid salts, sulfonic acid salts, sulfuric acid ester

salts, phosphoric and poly-phosphoric acid esters, and perfluorinated anionics. Cationic

surfactants are effective anti-bacterial agents and are employed as disinfectants and antiseptic

agents, but seldom as detergents. These surfactants include long-chain amines and their salts,

diamines and polyamines and their salts, quaternary ammonium salts, polyoxyethylenated long-

7

chain amines, quaternized polyoxyethylenated long-chain amines, and amine oxides. Non-ionic

surfactants are generally used together with anionic surfactants as active ingredients in

shampoos, hand dish washing liquids and washing powders. Non-ionic surfactants include

polyoxyethylenated alkylphenols, alkylphenol ethoxylates, polyoxyethylenated straight-chain

alcohols, alcohol ethoxylates, polyoxyethylenated polyoxypropylene glycols,

polyoxyethylenated mercaptans, long-chain carboxylic acid esters, alkanolamine condensates,

alkanolamides, and tertiary acetylenic glycols. Zwitterionic surfactants have both positive and

negative charges on the surface-active portion of the molecule, and thus are highly pH dependent

(Rosen, 1978).

Surfactant hydrophobic groups are generally long chain hydrocarbon residues. They may consist

of long straight- or branched-chain alkyl groups, long-chain alkylbenzene residues,

alkylnaphthalene residues, rosin derivatives, high-molecular weight propylene oxide polymers,

long-chain perfluoroalkyl groups, and polysiloxane groups (Rosen, 1978).

At sufficient concentration, surfactants will tend to self-associate and form micelles. The

concentration at which this occurs is known as the critical micelle concentration (CMC). As

with adsorption at interfaces, aggregation of surfactant molecules in a solvent will reduce the

free energy of the system. Micelle formation is critical to the solubilization mechanism of oil

removal, in which oil is solubilized into the surfactant micelles. However, it competes with the

interface adsorption mechanism, which is the predominant mechanism of action for SWAs

(Karsa, 2000).

8

1.2.4 Interfacial Phenomena and Mechanism of Action

When present at low concentrations in a system, a surface-active agent will adsorb onto the

surfaces or interfaces of the system, thereby altering the free energies of those surfaces or

interfaces (Rosen, 1978). The physical, chemical, and electrical properties associated with phase

boundaries often differ significantly from those properties in bulk. The contribution these

interfacial properties have on the overall system is typically small. However, in cases such as

dispersions, where the phase boundary area is large relative to the volume of the system, the

properties of the phase boundary can be significant. In such cases, surfactants will play a major

role in the system. Surfactants will also be significant in flotation and detergency situations,

where the interfacial interactions are radically different than the bulk phase interactions, and thus

have a profound effect on the overall system (Rosen, 1978). Detergency, dispersion, and

flotation are phenomena of particular interest when exploring the washing effect of SWA

surfactants at oil-substrate and oil-water interfaces.

In the case of oil stranded on a shoreline, SWAs function to release oil that is adhered as a film

on a solid surface, such as a sand particle, gravel, rock or rip-rap. The surfactant accomplishes

this by aligning itself at the oil/water interface, with the lipophilic portion of the molecule

oriented toward the oil film and the hydrophilic portion extending into the water phase. This

alignment of surfactant molecules at the interface acts to lower the oil/water interfacial tension,

γ, according to the Gibbs isotherm equation (Gibbs, 1876). The equation connects surface

tension of a solution with the concentration of the solute as follows:

9

where c is the surfactant concentration in the bulk solution, γ is the surface tension of the

solution, R is the gas constant, T the temperature, and Γ is the specific adsorption. The equation

assumes that the solution is ideal. Lowering the oil/water interfacial tension results in a

subsequent increase in the contact angle between the oil and the solid substrate to which it is

adhered (Figure 1.1).

As the contact angle is increased, the oil film rolls up into droplets and is released. The dynamic

contact angle, θ, is greatly dependent upon surface tension and is defined by the Young equation:

γLV cosθ0 = γSV - γSL

where γ is the interfacial tension between the liquid-vapor (LV), solid-vapor (SV), and solid-

liquid (SL) phases (Young, 1805). Surface wetting occurs when the contact angle is less than

90°; detachment occurs when the contact angle equals 180°. Oil is prevented from redepositing

on the substrate by the presence of surfactant molecules oriented at the oil/water interface. In the

case of oil dispersants, the surfactant has a strong affinity for the oil, and thus remains at the

droplet interface, stabilizing the oil as finely dispersed droplets. Non-dispersant surfactants are

somewhat more hydrophilic, and after releasing the oil, will preferentially dissolve into the

water, allowing the oil to coalesce and float to the water surface (Clayton et al. 1995).

( ) Γ−=⎟⎟⎠

⎞⎜⎜⎝

⎛RT

cd

d

TPln

γ

10

1.2.5 Studies of Oil/Water/Surfactant Systems

The role of surfactants in detergency applications has been widely studied. However,

understanding the surface chemistry associated with detergency can be difficult. Detergency is a

complex mechanism dependent upon many factors, including water hardness, temperature,

electrolyte composition, hydrodynamic conditions, surfactant properties and concentration,

substrate properties, and the nature of the contaminant (Thompson, 1994). In addition, there are

several mechanisms for oil removal that may act interdependently or simultaneously to yield an

overall observed efficiency. The primary oil removal mechanisms are the roll-up mechanism

and emulsification. In the roll-up mechanism, the oil-water-substrate surface energies are altered

by the addition of surfactant, thus lowering the adhesive forces holding the oil to the substrate.

In emulsification, gravitational and hydrodynamic forces exceed the cohesive surface forces

holding the droplet together, causing it to fragment (Thompson, 1994). These mechanisms

function interdependently, with roll-up generally occurring as a precursor mechanism to

emulsification. A third mechanism, solubilization of the oil in surfactant micelles, is less likely

in the current application.

Temperature is a critical variable in determining cleaning efficiency for non-ionic surfactants.

Oil/water interfacial tension exhibits a minimum at the phase inversion temperature (PIT) for oil-

in-water (o/w) to water-in-oil (w/o) emulsions (Aveyard, 1985). Thus, the temperature at which

phase inversion occurs for a particular oil and surfactant is associated with optimal detergency

for that system. For non-ionic surfactants, the effective size of the hydrophilic head group

relative to the lipophilic tail is linked to phase inversion. As temperature increases, the hydration

of the head group decreases, and thus the effective size of the head group decreases. If the

11

surfactant head group is larger than the solvated tail, an o/w emulsion will form; if the lipophilic

tail is larger than the head group, a w/o emulsion will form. The phase inversion condition,

which corresponds to an interfacial tension minimum and optimal detergency, occurs when the

effective size of the head and tail are equal. Under these conditions, the entropies of micelle and

surface formation are similar, and thus the system energy is minimized (Aveyard, 1985).

Anionic surfactants are similarly affected by electrolyte concentration. High concentrations of

counterions in solution will screen the electrostatic interactions of the headgroup and thus reduce

its effective size. An interfacial tension minimum occurs when the dissociation of the surfactant

into ion and counterion are equal in the micelle and at the interface (Aveyard, 1986). This

corresponds to a phase inversion condition as the surfactant aggregate moves from the aqueous

to the oil phase.

Thompson (1994) studied oil detachment mechanisms associated with optimal detergency

conditions for an anionic surfactant (sodium hexadecyl o-xylene sulphonate), a non-ionic

surfactant (n-dodecyl pentaethylene glycol C12E5), and a surfactant mixture (sodium dodecyl

sulfate/C12E3). In that study, surface activity was altered by varying electrolyte concentration for

the anionic surfactant, temperature for the non-ionic surfactant, and composition for the mixture.

In several cases, two detergency maxima were observed, one attributable to emulsification, the

other to the roll up mechanism. The detergency maxima associated with emulsification

corresponds to phase inversion and a minimum in interfacial tension. While two mechanisms

exist and can function simultaneously, either mechanism can fail to operate if the contact angle

and interfacial tension are not optimal.

12

Several researchers have explored the factors affecting interfacial tension and contact angle as

they pertain to wettability in oil/water systems (Buckley, et al., 1994; Liu, et al., 1999).

Wettability is the tendency of oil to spread across a surface and in this context is an opposing

mechanism to detergency. The properties of oil/water systems that promote wetting will tend to

inhibit detergency, and vice versa. Factors such as pH, salinity, crude oil acid/base composition,

crude oil polar groups, and temperature can influence oil/water interfacial tension and thus affect

detergency and surfactant efficiency (Buckley, et al., 1994; Drummond and Israelachvili, 2002).

It has been traditionally assumed that the asphaltenic fraction in a crude oil is directly

responsible for its wetting behavior (Craig, 1971; Buckley et al., 1997). Depending on

conditions, asphaltenes may adsorb on the surfaces as colloidal aggregates or as single

molecules. Yang (2002) found that deposition of polar asphaltenes from oil caused an increase

in oil wetting for a crude oil/brine/mica system. The degree of alteration was dependent on

hydrophobicity of the asphaltene and the amount of deposition. Factors having the greatest

affect on wettability were asphaltene content and base number/acid number. There are other

species in the oil and brine that can modify the wettability of the system by activating or

deactivating the adsorption of the asphaltenes and other surface-active species on the different

interfaces (Buckley et al. 1998). Kowalewski et al. (2002) found that adding hexadecylamine to

n-decane increased the wetting properties of n-decane on sandstone in brine. This was caused by

adsorption of the additive on the rock surface. Wetting properties were observed to be dependent

on the concentration of the additive and to increase with adsorption of the additive on the

surface.

13

Standal et al. (1999) examined the effect of pH, salinity, and ionic strength on partition

coefficients, surface tension and interfacial tension for polar organics in an oil/water system.

Polar components, such as carboxylic, phenolic and indolic acids, organic bases, and metal

complexes, are thought to be the surface-active components of crude oil. Results of the Standal

et al. study indicate a correlation between the polar component distribution and interfacial

tension. The partitioning of an individual component between the oil and water phases is largely

dependent upon the pKa of that compound. Acidic compounds preferentially partition into the

oil phase at pH values less than the pKa; basic compounds partition into oil at pH values greater

than the pKa. Because both acidic and basic components are active, the oil/water interfacial

tension is highest near neutral pH and lowest at low or high pH. In the Standal study, an increase

in salinity enhanced the partitioning of polar components into the oil phase and reduced the

interfacial tension for the strong acid, weak acid, and base. A decrease in interfacial tension with

increasing concentration was observed for the weak acid and the base. Buckley et al. (1989)

found that polar functional groups from both oil and mineral can behave like acids or bases when

exposed at the oil/brine and brine/mineral interfaces. The interaction is pH- and salinity-

dependent and may be attractive or repulsive depending on the charge distribution, electric

potential, and sign of the charged sites on both interfaces.

Similar studies by Hoeiland et al. (2001) and Standal et al. (1999) were conducted to examine

the effects of acidic and polar fractions, respectively, on interfacial tension and contact angle.

Results of the Hoeiland study indicate a decrease in interfacial tension as a function of pH and

acid number. A decrease in interfacial tension with increasing pH was observed in the alkaline

pH range for three acid fractions. This is due to the increase in surface activity caused by

14

dissociation. The relative activity of each fraction correlates to the acid number of the original

crude oil, with highest activity exhibited by the fraction with the highest acid number. The acid

structures and types were determined to be more important than the actual acid concentrations.

Alkyl acids, phenols and cyclic acids had a significant effect on contact angle, while high

molecular ring-structured acids with carboxylic compounds did not.

1.3 SWA EFFECTIVENESS TESTING

1.3.1 Effectiveness of SWAs in Field and Laboratory Studies

While a controlled laboratory effectiveness test will provide a useful general guideline, the

effectiveness of a SWA in the field is greatly contingent upon site- and situation-dependent

factors. Operational and environmental variables can greatly influence the effectiveness of a

SWA. These variables include the mode and success of SWA application, the penetration and

mixing of the SWA into the oil, the local substrate characteristics that influence oil adhesion, and

the chemical and physical properties of the oil itself (Clayton et al., 1995; Walker et al., 1999).

Such chemical properties vary not only among different oils but also over time. Physical

weathering of oil on impacted shorelines may significantly reduce SWA efficiency. Thus, it is

recommended that small-scale, on-site testing of a SWA be performed prior to large-scale

application.

Attempts have been made to evaluate the cleaning performance of SWAs in both field and

laboratory settings. Some of these studies have attempted to assess the beneficial effects of

SWAs on oil-impacted vegetation. EXXON funded research to evaluate the impact of Corexit

9580 on three species of plants: Rhizophora, Spartina, and Scirpus (DeLaune et al. 1994; Teas et

15

al. 1993). Cleaning effectiveness and toxicity were evaluated for treated vegetation relative to

oiled, untreated vegetation. DeLaune et al. found that oil adhering to test vegetation treated with

Corexit 9580 was significantly less than for untreated vegetation based on visual assessment. In

the study by Teas et al., treatment of oiled red mangrove trees with Corexit reduced mortality as

long as the SWA was applied within the first seven days. Pezeshki et al. (1998) also evaluated

the effect of Corexit 9580 on the removal of South Louisiana Crude (SLC) and Arabian Medium

Crude (AMC) oils from Gulf Coast macrophyte species, specifically bulltongue (Sagittaria

lancifolia L.), three cornered grass (Scirpus olneyi E. & G.), and broadleaf cattail (Typha latifolia

L.). In a controlled greenhouse environment, plant stomatal functioning, photosynthesis,

respiration, regeneration, growth, and biomass were monitored for species receiving no oil or

cleaner, oil only, and oil plus cleaner. Similar field studies were conducted in two marsh habitats

in the Louisiana coastal wetlands (Pezeshki et al., 2001). In these studies, Corexit 9580

alleviated the short-term impact of oil on gas exchange function; however, it had no observable

effects on above ground biomass production or regeneration in S. patens, and no beneficial effect

on carbon fixation or the number of live shoots in S. lancifolia. Gonzalez et al., (1999),

examined the effect of oil on fungal biomass in an estuarine mangrove. Plots populated with

Rhizophora mangle trees received treatments of oil, oil plus fertilizer, oil plus shoreline cleaner,

or no oil and no cleaner. While fertilizers aided in enhancing biodegradation, the shoreline

cleaner showed no observable difference over the oil only treatment.

Research has also been conducted to determine SWA efficiency based on detergency alone, the

physical removal of oil from solid, non-vegetative substrates. Tumeo and Cote (1998) examined

the effect of wash water temperature and composition on the effectiveness of SWAs.

16

Laboratory column experiments were performed using beach substrate from Prince William

Sound. The porous gravel substrate was cleaned to remove existing contamination, re-oiled with

a known quantity of oil, and loaded into stainless steel columns 30 cm in length with an interior

diameter of 10 cm. SWA was then applied, and the column was flushed with either fresh or salt

water in an up-flow direction. The seawater experiments were performed at two temperatures,

10°C and 20°C. Results indicated that SWA efficiency is reduced at colder temperatures;

however, the removal efficiencies reported for all treatments were extremely low, and thus the

significance of these findings is questionable. The largest removal efficiency observed was only

1.69% for the SWA Grancontrol ‘O’. The greatest relative loss in efficiency was reported for

PES-51, which experienced a decline in efficiency from 1.41% at 20°C to 0.47% at 10°C.

Similarly, a decline in efficiency was noted when seawater was used instead of fresh water.

However, the greatest relative loss in efficiency was observed for Corexit 9580, which had an

efficiency of 2.01% in tap water and 0.69% in seawater. One SWA, PES-51, showed the

opposite trend, with efficiencies of 0.79% in tap water and 1.41% in seawater. The very low

removal efficiencies reported for all treatments suggest that this experimental design may not

have been optimal for this investigation.

Field testing of SWAs has also been attempted, although interpretation of field data is often

difficult and subjective. Corexit 9580 was used in Prince William Sound following the 1989

EXXON VALDEZ spill (Alaska Department of Environmental Conservation, 1993).

Measurements of total hydrocarbon concentrations (THC) in sediment samples were taken

before and after treatment (Fiocco et al. 1991). Due to the high variability in THC

concentrations across the test sites, many samples had to be taken, and the performance data for

17

treated plots relative to untreated plots were statistically inconclusive. Two SWAs, PES-51 and

Corexit 9580, were applied to shorelines following the 1994 MORRIS J. BERMAN spill in San

Juan, Puerto Rico. SWA results were compared to the use of high-pressure, hot-water alone on

beach rock and rip-rap (Michel and Benggio 1995). The consensus based on a visual comparison

of test plots was that both SWAs performed better than water alone, but there was no observable

difference in performance between PES-51 and Corexit 9580. PES-51 was also used on seawalls

and rip-rap following the 1993 Tampa Bay spill; however, no consensus on the effectiveness of

PES-51 was reached by response team members (Owens et al. 1995).

1.3.2 Previous Laboratory Testing Protocols

The development of a testing protocol to evaluate the effectiveness of SWAs has been

undertaken by numerous research laboratories. Most of these protocols are concerned with

assessing the detergency, or cleaning power, of a SWA, as opposed to toxicity or biological

impact. The goal of these testing procedures is to quantify the physical removal of oil from non-

vegetative substrates as a result of SWA action. The fundamental methodologies for these

protocols are similar. They involve the application of oil to a solid substrate (or the use of pre-

oiled substrate), weathering of the oil on the substrate, application of SWA at the manufacturer’s

recommended dosage, observance of a contact period for SWA penetration, and washing of the

substrate with water. The fraction of oil removed in the wash water and the fraction remaining

on the substrate are quantified. Oiled controls without SWA application are also tested.

The testing procedure developed by Environment Canada employs a stainless steel or porcelain

trough as the substrate. Oil is applied to one end of a pre-weighed trough, and the trough is then

18

suspended vertically for ten minutes. The trough is reweighed to determine the mass of oil

applied. SWA is applied evenly along the trough and allowed to contact with the oil for ten

minutes. The trough is subsequently rinsed with two aliquots of water, blotted dry, and

reweighed. Following solvent extraction of the wash water and the substrate, the oil removal is

quantified. Fingas et al. (1990) evaluated the results of tests conducted using this protocol and

reported oil removal efficiencies ranging from 1% to 52% for 26 cleaning agents tested. The

toxicity and dispersant effectiveness were also tested for each product. A negative correlation

was observed between dispersant effectiveness and cleaning effectiveness. Fingas et al.

concluded that the mechanisms of dispersancy and detergency are quite different, and thus, a

good SWA will be a poor dispersant, and vice versa. The top ten cleaning agents from that study

are listed in Table 1.1.

Clayton (Science Applications International Corporation, SAIC) modified the Environment

Canada protocol, through contract agreement with the U.S. EPA, and published a comparison of

the two methods in an EPA report (Clayton et al., 1993). The SAIC modified protocol used

stainless steel and porcelain coupons instead of troughs. Following application of the oil and

SWA, the coupons were lowered into separate beakers containing seawater. Agitation was

applied using a shaker table, and the coupons were subsequently removed, dried, and extracted,

along with the wash water. Clayton et al. tested this procedure, along with the Environment

Canada protocol, using stainless steel and porcelain substrates, two test oils (Prudhoe Bay Crude

and Bunker C), and three cleaning agents (Corexit 9580, Citrikleen XPC, and Corexit 7664).

Comparison of results revealed greater variability due to oil type and cleaning agent than to the

testing method used. Cleaning performance was lowest for Corexit 7664, with efficiency values

19

ranging from 0.1% to 36.6% across treatments and an overall mean efficiency value of 11.2%.

Corexit 9580 and Citrikleen XPC performed comparably for both oils and all testing procedures.

Mean efficiency values ranged from 32% to 63% for both SWA; the overall mean efficiency was

47.0% for Corexit 9580 and 46.6% for Citrikleen XPC.

Clayton further modified the testing protocol under the direction of the Marine Spill Response

Corporation (MSRC) for on-site field testing of SWAs (Clayton et al., 1995). The new protocol

was applied to natural substrates and was readily adaptable to field situations. Three natural

substrates were tested using this procedure: gravel, fragments of rock rip-rap, and blades of the

eelgrass Zostera marina. Two oils (Bunker C and Bonny Light) and two SWAs (PES-51 and

Corexit 9580) were studied. Substrates were placed in a wire mesh container and submerged in

seawater. A known quantity of oil was applied as a slick to the surface of the seawater, and the

mesh container was pulled vertically through the slick. The oiled substrate was allowed to

weather at ambient temperatures for 18-22 hours, and gravel and rock substrates were

additionally weathered in a 60°C drying oven for 1-2 hours. A known quantity of SWA was

applied to the substrate using a spray applicator. Following a contact period, the substrate was

washed with seawater, and the wash water and substrates were extracted with solvent. For

Bunker C oil, the oil removed without SWA was less than 3.2% for all substrates. The cleaning

efficiency for PES-51 was 30.4% for gravel, 11.5% for rip-rap, and 8.3% for eelgrass. Corexit

9580 achieved significantly higher removals: 52.9% for gravel, 21.1% for rip-rap, and 23.6% for

eelgrass. For Bonny Light, gravel removal efficiencies were 1.4% without SWA, 18.0% with

PES-51, and 38.2% for Corexit 9580.

20

A similar testing protocol was developed by EXXON using commercial aquarium gravel soaked

with Alaska North Slope Crude oil. According to this method, oiled gravel was placed on a glass

support in a stainless steel wire mesh basket. SWA was applied through spray application, and

the gravel was subsequently rinsed with seawater. The wash water and gravel substrate were

extracted to determine the percent oil removed.

The Centre de Documentation de Recherche et d’Experimentations Sur les Pollutions

Accidentelles des Eaux (CEDRE, Merlin and Le Guerroue, 1994) proposed a very basic testing

procedure, involving the application of oil to a glass slide. The mass of oil applied was

determined based on weight measurements before and after oil application. A spray gun was

used to apply the SWA, and after 10 minutes, the slide was sprayed with seawater to remove the

released oil.

1.4 RESEARCH OBJECTIVES

1.4.1 Protocol Development and Testing

The goal of this work was to develop a standardized effectiveness protocol to test the ability of

SWAs to remove oil from surfaces without dispersing the oil into the water column. The

protocol had to be reproducible in the hands of multiple operators and provide information that

can be used to predict effectiveness in the field. Several preliminary test designs were proposed

and evaluated. The protocol that was ultimately adopted used natural substrates to better reflect

real world applications. Research was conducted to determine the effect of protocol variables on

SWA effectiveness in removing weathered Prudhoe Bay Crude (PBC) oil from sand and gravel

(Koran et al., 2005, 2006). The following variables were tested at multiple levels: substrate type,

21

hydration, drain time, volume and mode of oil application, oil weathering time, SWA solution

concentration, SWA-to-oil ratio, oil-SWA contact time, rotational mixing speed, and mixing

time. A fractional factorial experiment was designed and completed to determine which

variables were significant to surface washing, and to establish optimal levels to be included in

the final protocol.

1.4.2 Surface Interactions and Detergency

In addition to the practical aspect of developing a reproducible SWA testing protocol, a

fundamental study of the surface interactions involved in detergency was conducted. The

primary mechanism of action for SWAs is the alignment of surfactants at the system interfaces

causing a reduction in oil-water interfacial tension and an increase in contact angle between the

oil and the substrate. As the contact angle increases, the oil film rolls up into droplets and is

released into the water. This mechanism is optimized by a surplus of free surfactant molecules at

the system surfaces. In this work, the roles of critical micelle concentration, oil-water interfacial

tension, and oil-substrate contact angle in predicting SWA effectiveness were evaluated. SWAs

were ranked based on each of these criteria, and the data were correlated with SWA effectiveness

under the developed protocol.

22

Figure 1.1: Surfactant Induced Increase in Contact Angle.

Surfactant molecule:

lipophilic tail

hydrophilic head

oilso

lidoil

solid

Surfactant molecule:

lipophilic tail

hydrophilic head

oilso

lidoil

solid

Table 1.1: Surface Washing Agent Test Results for Environment Canada Protocol.

Agent Oil Removed

(%) Toxicity*

Dispersant Effectiveness by

Swirling Flask Test (%)

D-Limonene 52 35 0

Penmul R-740 44 24 9

Corexit 9580 42 >5,600 0

Formula 2067 39 11 0

Citrikleen XPC 36 34 2

Formula 861 32 24 0

Corexit 7664 27 850 2

BP 1100 WD 21 120 6

Re-Entry 17 8 0

Palmolive dish soap 16 13 9

* LC50 for rainbow trout over 4 days; larger values denote less toxicity.

23

Chapter 2

MATERIALS AND METHODS

The criterion for evaluating SWA effectiveness was based on the amount of oil that can be

removed from a substrate with a given amount of test cleaner. Several testing designs were

proposed during the development stages of the protocol. Generally speaking, the protocol

involved application of oil to a substrate, weathering of the oil to allow time for it to adhere to

the substrate, treatment of the oiled substrate with SWA, and washing with seawater to release

any oil that had been lifted from the substrate surface. The wash water and the substrate were

extracted separately with methylene chloride, and the concentration of oil in the extracts was

quantified by spectrophotometry. The effectiveness of each SWA was calculated based on the

mass of oil released into the wash water relative to the total mass of oil applied. Untreated

controls were included in the experimental design to determine the amount of oil released by

washing with water alone. Critical micelle concentrations, oil-water interfacial tensions and oil-

substrate contact angles were measured and used as predictive tools for ranking SWAs. These

rankings were correlated to the measured effectiveness under the developed protocol. The

materials, testing procedures and analytical methods used for this work are described below.

2.1 MATERIALS

2.1.1 Crude Oil

Prudhoe Bay Crude (PBC), a medium weight EPA/American Petroleum Institute (API) standard

reference oil, was used as a representative crude oil for protocol development work. PBC has

been thoroughly characterized in previous EPA and API studies. It has a density of 0.894 g/ml at

24

25°C. Other physical and chemical characteristics are listed in Table 2.1. Some preliminary

testing was also done using IFO-180, a heavier weight fuel oil (source: Minerals Management

Service, Herndon, VA). Fresh IFO-180 has a density of 0.957 g/ml at 25°C.

2.1.2 SWAs and Surfactants

Twenty-seven SWAs are currently listed on the NCP Product Schedule, and their properties are

listed in Table 2. These SWAs are composed of surface active agents, solvents and/or additives.

The specific formulations for these SWAs are typically marked confidential; thus the surfactant

properties are generally unknown. For this work, 6 SWAs were selected from those listed on the

Product Schedule: Aquaclean (Madison Chem. Co., Inc.), Biosolve (Westford Chemical

Corporation), Corexit 9580 (Nalco/EXXON), PetroClean (Alabaster Corporation), Petrotech 25

(Petrotech America Corp.) and Superall (Stutton North Corporation). Aquaclean, Biosolve,

PetroClean, Petrotech25 and Superall are water soluble products; Corexit 9580 is oil soluble.

Dawn dishsoap and a pure surfactant, sodium dodecyl sulfate (SDS), were also tested

2.1.3 Synthetic Seawater

The synthetic seawater “Instant Ocean,” manufactured by Aquarium Systems of Mentor, OH,

was used as the wash water for all experiments. The synthetic seawater was prepared by

dissolving 34 g of the salt mixture in 1 liter of Milli-Q water (i.e., a salinity of 34 ppt). Table 3

provides a list of the ion composition of the sea salt mixture. Following the preparation, the

saltwater solution was allowed to equilibrate to the ambient temperature of the laboratory

(23±3°C) and filtered to remove any undissolved particulates.

25

2.1.4 Extraction Solvent

Methylene chloride (dichloromethane, DCM, pesticide quality) was used for preparation of oil in

DCM stock standards and for extraction of aqueous samples, solid substrates, and oil standards

in water. Oil in DCM standards and samples were analyzed directly by UV/visible

spectrophotometery. DCM extracts were solvent exchanged into hexane prior to analysis by

GC/MS. The purpose of this solvent exchange was to precipitate the asphaltene fraction of the

oil. Asphaltenes are not resolvable by GC/MS and will shorten the life of the column.

2.1.5 Substrates

The first protocol that was developed and tested utilized glass plates as the substrate. Glass is

composed of silica sand (silicon dioxide, SiO2) and has physical and chemical properties similar

to silica sand found on shorelines. Rectangular glass plates were fabricated and their surfaces

were varied by grinding one side of the glass plate with grinding powders. Five levels of surface

roughness were tested: smooth (no etching), fine, medium, coarse and extra coarse. Fine,

medium, and coarse surface grindings were obtained using grinding powder grit sizes of 400,

220, and 80, respectively. An extra-coarse surface was created using grinding materials used for

sand blasting.

For the natural substrate protocol, ASTM 20/30 silica sand and FilPro Filter Gravel from U.S.

Silica Company were used. ASTM 20/30 sand (plant: Ottawa, Illinois) is 99.8% whole grain,

unground silicon dioxide (quartz) with other trace metal oxides comprising the remaining 0.2%.

The sieve sizes 20 and 30 correspond to particle sizes of 0.850 mm and 0.600 mm, respectively;

97% of the particles are retained within this range. Sand grains are round with a hardness of 7 on

26

the Mohs scale. The FilPro Filter Gravel (plant: Mauricetown, New Jersey) is also whole grain,

crystalline silica. Particle sizes range from 1/16” to 1/8” and have a hardness of 7 on the Mohs

scale. The properties of these substrates are described in further detail in Chapter 7. A third

substrate, 5/8 x 3/8 common pea gravel (source unknown) was also tested but was not included

in the fractional factorial experiment. Experiments conducted during the protocol development

phase used only silica sand as the substrate. The two gravel substrates were tested after the other

protocol variables had been largely optimized. All substrates were acid washed with a 1M nitric

acid solution, rinsed until the wash water had a neutral pH, and baked overnight at 500°C prior to

use.

2.1.6 Experimental Apparatus

For the glass plate protocol, the experimental apparatus consisted of a six-shaft, rotational paddle

mixer (Phipps and Bird, model #7790-400) with custom stainless steel shafts, custom rectangular

glass plates, and custom baffled beakers. The experimental system is shown in Figure 2.1. The

mixer had an adjustable mixing speed with digital rpm readout. The glass plates were made

from 4 mm thick borosilicate glass and were rectangular (60 mm x 40 mm), with a tab (25 mm x

45 mm) centered at the top of the plate for attachment to the shafts. Oil and SWA solutions were

applied drop-wise using an Eppendorf Repeater Pro positive displacement pipetter or an

Eppendorf Research Pro Pipette capable of dispensing 5 :l–100 :l. Figure 2.2 shows oiled

plates before and after treatment with SWA. Baffled beakers (1L) were used to enhance mixing.

Aqueous extractions were performed in 500-ml separatory funnels with ground glass stoppers

and Teflon stopcocks. DCM extracts of aqueous samples were collected in 50-ml graduated

cylinders. Plates were rinsed with clean DCM until all visible oil was removed and the rinse

27

solvent ran clear. The DCM rinse was collected in a 250-ml beaker and transferred to a 50-ml

graduated cylinder. All samples were stored at 4°C in air-tight glass vials with Teflon lined

caps.

For the natural substrate protocol, sand and gravel were contained in 2” x 2” x 2” baskets

constructed of 30-mesh stainless steel wire cloth and supported by a stainless steel frame

(Hillside Wire Cloth Co., Inc., Bloomfield, NJ). The experimental system is shown in Figure

2.3. Oil and SWA solutions were applied using an Eppendorf Repeater Pro positive

displacement pipetter or an Eppendorf Research Pro pipetter capable of dispensing 5 :l–100 :l.

Non-baffled Pyrex 600-ml beakers were used to hold the seawater and submerged baskets during

washing. The beakers were positioned on a LabLine orbital platform shaker (model 3520) to

provide washing agitation. The wash water was extracted in 250-ml separatory funnels with

ground glass stoppers and Teflon stopcocks. The substrate-filled baskets were extracted in clean

600-ml Pyrex beakers by shaking on the orbital platform shaker. DCM extracts of aqueous and

substrate samples were collected in 50- and 250-ml graduated cylinders, respectively, and stored

in air-tight glass vials with Teflon lined caps.

2.2 EXPERIMENTAL PROCEDURES

2.2.1 Glass Plate Testing Procedure

The following is the generalized testing procedure used for conducting experiments under the

glass plate protocol. Multiple levels were tested for oil volume and application pattern,

28

weathering time, SWA dilution, SOR, contact time, mixing speed, and mixing time, and the

protocol was modified accordingly. The glass plate protocol is discussed in detail in Chapter 3.

• Clean and DCM-rinse all glassware prior to use.

• Using an Eppendorf positive displacement repeat pipetter with 500 µl Brinkmann

disposable tip, dispense oil drop-wise onto each of six plates using the patterns shown in

Figure 2.4. Plates should be kept on a level surface.

• Weather the oil on the plate in a ventilated hood at ambient temperature.

• Using an Eppendorf repeater pipette, dispense SWA to the oiled area. Due to oil

spreading, a larger volume of SWA will be required to cover the oiled area. Typically a

10x-volume is sufficient. Thus, ten SWA drops should be applied to the area covered by

each oil drop.

• Let the SWA contact with the oil for the specified contact time.

• Meanwhile, attach stainless steel shafts to the 6-shaft rotational mixer and elevate the

shafts to the highest position.

• Dispense 400 ml artificial seawater into each of 6 baffled beakers.

• After the oil-SWA contact period, attach the glass plates to the stainless steel shafts.

During attachment, each glass plate should be held over a baffled beaker so that any oil or

SWA that runs off the plate will fall into the wash water.

• Lower the shafts so that the plates are submerged in the seawater. The top of the plate

should be ¼” to ½” below the water surface and the edges of the plate should not touch

the bottom or sides of the beaker.

29

• Turn the mixer on and allow the plates to rotate at the specified mixing speed for the

specified mixing time.

• Turn the mixer off and immediately elevate the plates above the surface of the water.

Allow the plates to drain in this position for 5 minutes.

• Remove the glass plates from the stainless steel shaft; place each in a separate 250 ml

beaker.

• Extract the wash water according to the method outlined in Section 2.2.3.1, Aqueous

Phase Extraction.

• Extract the glass plates according to the method outlined in Section 2.2.3.2, Glass Plate

Extraction.

2.2.2 Natural Substrate Testing Procedure

The following is the generalized testing procedure that was used for conducting initial

experiments. Multiple levels were tested for oil volume and application pattern, weathering

time, SWA dilution, SOR, contact time, mixing speed, and mixing time, and the protocol was

modified accordingly. Experimental variables and testing levels are listed in Table 2.4. Gravel

substrates followed the same testing procedure. The natural substrate protocol is discussed in

detail in Chapter 4.

• Clean and DCM-rinse all glassware.

• Using a 25-ml graduated cylinder, dispense 25-ml of acid washed sand into six stainless

steel wire mesh baskets. Label the baskets 1-6.

30

• For wet sand experiments, submerge each basket in seawater (or freshwater) for 5

minutes, remove and drain for 5 minutes prior to oil application.

• Using an Eppendorf positive displacement repeat pipetter with disposable tip, dispense the

appropriate volume of oil in a 1, 5 or 9 drop pattern on the surface of the sand. See Figure

2.5 for drop patterns.

• Allow the oil to weather on the sand for the specified weathering time in a ventilated hood

at ambient temperature.

• Using an Eppendorf repeater pipetter, dispense the appropriate volume of SWA to each oil

spot on the sand in four of the six baskets. SWA application volumes, dilutions, SOR and

drop patterns will vary depending on the conditions being tested. The two oiled sand

baskets that did not receive SWA are untreated controls.

• Allow the SWA to contact the oil for the specified contact time.

• Meanwhile, dispense 100 ml artificial seawater (or freshwater) into each of six 600-ml

beakers. Label the beakers 1-6 to correspond to the basket number and secure the beakers

on the orbital platform shaker.

• Following the oil-SWA contact period, place each basket in a beaker so that the sand is

completely submerged in seawater. The basket will rest on the bottom of the beaker, and

the top of the sand should be ¼” to ½” below the water surface.

• Turn the platform shaker on and allow the water to wash through the baskets for the

specified mixing time at the specified mixing speed.

• Turn off the mixer and immediately elevate the baskets above the surface of the water.

Allow the excess water to drain out of the sand.

31

• Place each basket in a separate 600-ml beaker. Label the beakers 1-6 to correspond to the

basket number.

• Extract the 100 ml wash water according to the method outlined in Section 2.2.3.1,

Aqueous Phase Extraction.

• Extract the sand-filled baskets according to the method outlined in Section 2.2.3.3,

Substrate Extraction.

2.2.3 Extraction Methods

2.2.3.1 Aqueous Phase Extraction: Aqueous samples were extracted with DCM using 500-ml

(glass plate protocol) and 250-ml (natural substrate protocol) separatory funnels with Teflon

stopcocks and ground glass stoppers. Oil in DCM standards were prepared in seawater and

extracted according to Sections 2.3.1.1 and 2.3.1.2 below. The following procedure was

followed for extraction of wash water:

• Transfer the entire wash water sample to a separatory funnel with Teflon stopcock.

• Rinse the walls and bottom of sample beaker with four 5-ml aliquots of DCM, transferring

each DCM rinse to the separatory funnel.

• Place a ground glass stopper in the top of the separatory funnel and shake for 15 seconds,

venting as needed.

• Allow the separatory funnel to remain stationary for at least 2 minutes to allow phase

separation.

• Remove the stopper, open the stopcock and collect the DCM extract in a 50 ml graduated

cylinder. For this first extraction, do not drain the last 1-2 ml of extract from the funnel as

it may contain an emulsion.

32

• Extract the wash water two more times with 15-ml aliquots of DCM and drain to the oil-

water interface, collecting each extract in the 50 ml graduated cylinder.

• Bring extract to a final volume of 50 ml with clean DCM. If a larger volume is used,

record the final volume and adjust the recovery calculations accordingly

2.2.3.2 Glass Plate Extraction: Glass plates were extracted with DCM according to the

following procedure:

• Place each glass plate in a separate 250-ml beaker.

• Rinse the plate with clean DCM until all visible oil is removed and the rinse solvent runs

clear. This will require approximately 15-20 ml DCM.

• Transfer the DCM extract to a 25 ml graduated cylinder.

• Rinse the inside of the beaker with 4-5 ml fresh DCM and transfer the rinse to the

graduated cylinder.

• Bring the extract to a final volume of 25 ml. If a larger volume is used, record the final

volume and adjust the recovery calculations accordingly.

2.2.3.3 Natural Substrate Extraction: Sand and gravel were extracted using the procedure

outlined below:

• Drain the excess water from the sand basket into the sample beaker.

• Place the basket in a clean 600-ml beaker.

• Add 50 ml DCM to the beaker so the level of DCM comes just to the surface of the

substrate.

• Place beakers on the orbital shaker and shake for 5 minutes at 150 rpm.

33

• Remove the beaker from the shaker and tilt the beaker to allow the DCM to wash all sides

of the basket above the DCM level.

• Drain the DCM extract into a 250-ml separatory funnel. A small amount of water will

also be transeferred.

• Repeat this extraction procedure two more times with 40-ml aliquots of DCM, transferring

each extract to the separatory funnel.

• Using a DCM squirt bottle, rinse the sand and basket with 10-20 ml clean DCM and

transfer to the separatory funnel.

• Drain the DCM layer into a 250-ml graduated cylinder, leaving the water layer in the

funnel.

• Bring the extract to a final volume of 150 ml with clean DCM. If a larger volume is used,

record the final volume and adjust the recovery calculations accordingly.

During later stages of testing, the volume of sand used in experiments was reduced from 25 ml to

15 ml to reduce the amount of solvent required for extraction. For 15-ml sand experiments the

following modifications were made:

• Remove sand from the basket and place in the bottom of a 250-ml beaker.

• Rinse the basket with DCM to remove residual sand and oil from the basket.

• Extract the sand with three consecutive 20-ml aliquots of DCM by shaking at 150 rpm for

5 min.

• Decant the DCM extract from the sand and transfer to a separatory funnel to remove

residual water from the sample.

34

• Collect the extract in a graduated cylinder and bring to a final volume of 70 ml. If a larger

volume is used, record the final volume and adjust the recovery calculations accordingly.

2.3 ANALYTICAL METHODS

2.3.1 UV-Vis Spectrophotometry: The final protocol developed by this work will utilize

spectrophotometry as the means of quantifying oil concentrations in DCM extracts. The

instrument used for protocol development and testing was an Agilent 8453 UV/Visible

Spectrophotometer with standard silica 10 mm path length rectangular cells and diode array

detector. Absorbance was measured for all wavelengths and recorded by computer. Oil in DCM

calibration standards were prepared and analyzed with each sample set. Absorbance

measurements at wavelengths 340, 370, and 400 nm were used to determine the area under the

absorbance curve between wavelengths 340 and 400. Samples were quantified against the area

vs. concentration calibration curve.

2.3.1.1 Oil Stock Standard Preparation

• Weigh a clean vial with a loose Teflon and aluminum cap (x grams).

• Add 2 ml of the specific reference oil to the vial and re-weigh vial with the cap (y grams).

• Add 18 ml DCM to the vial and re-weigh the vial with the cap (z grams).

• Crimp the aluminum cap and mix the vial contents by hand shaking.

• Measure the density of the specific reference oil + DCM by either using a density bottle or

a 1 ml gas tight syringe (weigh the empty syringe and the syringe with 1 ml of solution)

(Doil+DCM, g/L).

• Determine the concentration of the oil solution

35

Concentration, g / L =(y - x)

(z - x) / oil+DCMρ

2.3.1.2 Preparation of Standard Solutions from the Stock Standard Solutions

• Add 100 ml synthetic seawater and a specific volume of the oil-DCM stock standard to six

250-ml separatory funnels. For PBC, the volumes of stock standard added to the six

separatory funnels are 22 :l, 40 μl, 100 μl, 200 μl, 250 μl, and 300 μl. The maximum

volume of PBC stock standard that can be added was 300 μL because absorbance saturation

in the spectrophotometer occurs above this concentration value.

• Extract the contents of each separatory funnel with a 20-ml volume of DCM by shaking

vigorously for 15 seconds. Vent as necessary.

• Allow the funnel contents to remain stationary for 2 minutes to permit phase separation.

• Collect the DCM extract in a 50-ml graduated cylinder. For this first extraction, do not

drain the last 1-2 ml of DCM from the separatory funnel as a web-like emulsion may have

formed at the solvent/water interface.

• Extract the contents of each funnel two more times, using a 10-ml portion of DCM for the

second and third extraction.

• Drain to the solvent/water interface, collecting the DCM in the graduated cylinders. Since

DCM is slightly miscible in water, the total collected volume will be less than 40 ml.

Adjust the final extract volume in each graduated cylinder to 40 ml with pure DCM.

• Transfer the DCM extract to a 50-ml crimp style glass vial with an aluminum cap and

Teflon septa. Mix contents of the sealed vial by inverting at least 10 times.

• Store the vials at 4±1 °C until the time of analysis.

36

2.3.1.3 Instrument Conditions and Calibration

• Turn on the tungsten and deuterium lamps of the spectrophotometer and allow a 30-45

minute warm-up period prior to any analysis.

• Remove the standard vials from the refrigerator and allow to equilibrate to laboratory

temperature (23±3 °C).

• Blank the Agilent 8453 spectrophotometer using a pure DCM solvent sample.

• Determine the absorbance of the six standards at analytical wavelengths of 340 nm, 370

nm, and 400 nm. The calibration standards should be introduced in increasing order of

concentration.

• The area under the curve can be calculated using the trapezoidal rule (Burden and Faires,

2000) according to the following equation:

( ) ( )Area

Abs Abs Abs Abs=

++

+340 370 370 40030

2

30

2

* *

Create a calibration curve by plotting concentration versus area under the absorbance curve for

the six calibration standards. Determine the slope of the calibration points by using linear

regression with zero intercept, as follows:

y mx=

where y = area under absorbance curve; x = concentration of oil, g/L; m = slope (Draper and

Smith, 1998).

37

2.3.2 GC/MS: An Agilent 6890 GC with Mass Selective Detector was used to quantify the

alkane and PAH components of oil. DCM extracts were solvent exchanged into hexane prior to

injection by GC/MS to precipitate heavy asphaltenes from the samples. Asphaltenes are heavy

residues in oil that are not resolvable by GC/MS. Injection of this fraction into the GC/MS

would dirty the injection port and shorten column life. The concentrations of 18 alkane and 11

PAH compounds were quantified and summed to give total alkane and total PAH values. Alkane

analytes included C14 through C29 straight chain alkanes plus the branched chains pristine and

phytane. PAHs considered were naphthalene and its C1- to C4- alkylated congeners,

phenanthrene and its C1- to C4- alkylated congeners, and dibenzothiophene. These represent the

saturated hydrocarbons and PAHs that are present in PBC oil at the highest concentrations.

2.3.3 Spectrofluorometry: Oil concentration was measured by spectrofluorometry. The

instrument used for analysis was a Shimadzu RF-5301PC Spectrofluorophotometer with a blazed

holographic grating, photomultiplier and digital signal processing. Calibration standards were

measured and used to create a calibration curve. Samples were diluted 1:40 to avoid quenching

and to fall within the linear calibrated range of the instrument.

2.3.4 Thin Layer Chromatography: An Iatroscan MK6 was used to separate the DCM extracts

into alkane, PAH, resin and asphaltene fractions. Iatroscan combines the techniques of thin layer

chromatography (TLC) for the separation of organic compounds with a flame ionization detector

(FID). It uses Chromarods, quartz rods coated with a thin layer of silica or alumina, to develop

and separate the sample. Spotting of the sample on the chromarod was done using a full

automatic sample spotter. The chromarods were developed in a separate development tank for

38

separation of the fractions, and then introduced into the Iatroscan for analysis. As the chromarod

was advanced at a constant speed through the hydrogen flame of the FID, the substances were

ionized through energy obtained from the hydrogen flame. An electric field applied to the poles

of the FID caused the ions to generate electric current with intensity proportional to the amount

of each organic substance entering the flame. The collector, which was placed above the flame,

generated an analog signal that was recorded by computer.

2.3.5 Surface Tension for Determining Critical Micelle Concentration: Surface tension was

measured using the Wilhelmy plate technique (Adamson and Gast, 1997; Bendure, 1997). Forty

concentrations of each SWA were prepared by incrementally dosing a high concentration stock

solution into pure water. Tests were run in duplicate. CMC was determined by plotting surface

tension vs. log concentration for each surfactant. Surface tension will drop with incremental

increases in solution concentration until the surface of the liquid becomes saturated with

surfactant and the surface tension reaches a plateau. A line is drawn through the slope of the

curve where the surface tension is dropping, and another line is drawn through the plateau

portion of the curve. The intersection of these lines is the CMC. CMC was determined for five

SWAs: Aquaclean, Biosolve, Petroclean, Petrotech and Superall. Corexit 9580 is not water

soluble and does not form micelles. This work is discussed in Chapter 8.

2.3.6 Interfacial Tension: Oil-water equilibrium interfacial tensions between PBC and six

SWAs were measured. Aquaclean, Biosolve, Petroclean, Petrotech, and Superall were evaluated

at the following concentrations: 100% SWA (undiluted), 50% aqueous and 5% aqueous. Corexit

9580 is not water soluble, so only the undiluted product was tested. The experiments were

39

performed by the pendant drop method using a Kruss Drop Shape Analysis System DSA10. This

instrument is capable of measuring interfacial tensions down to 0.01mN/m. A drop of PBC oil

was formed at the end of a capillary tip within the bulk phase of each SWA solution. For SWA

solutions that were heavier in density than the oil, the oil droplet was formed from an upward-

pointing capillary. Corexit 9580 was the only SWA tested that was lighter than oil; for this

SWA, the capillary was downward-pointing. The drop was formed to about 90% of its

detachment volume and then digitally imaged. The drop’s image was fit by a robust

mathematical approach to determine the drop’s mean curvature at over 300 points along its

surface. Interfacial tension was then calculated using the Laplace equation, as described further

in Chapter 8.

2.3.7 Contact Angle Measurement: The three phase contact angle for PBC oil in the presence

of each of six SWAs was measured. Aquaclean, Biosolve, Petroclean, Petrotech, and Superall

were evaluated at the following concentrations: 100% SWA (undiluted), 50% aqueous and 5%

aqueous. Corexit 9580 is not water soluble, so only the undiluted product was tested. For this

analysis, a smooth piece of granite or sandstone could be used. Because the chemical properties

of sandstone are a closer approximation of natural beach substrate, sandstone was selected for

this analysis. The surface was submerged in each cleaning solution, then oil droplets were

introduced via a curved upward-pointing syringe to the underside of the sandstone. For Corexit

9580, which is lighter than oil, the drop of oil was introduced using a downward-pointing syringe

to the top surface of the sandstone. Contact angle was then directly measured on each drop using

a Kruss DSA10 Drop Shape Analysis System. Five replicate drops were analyzed per solution

40

and the results averaged to obtain a single contact angle measurement for each SWA at each

solution concentration. This is discussed further in Chapter 8.

2.4 IDENTIFICATION OF SIGNIFICANT EXPERIMENTAL PARAMETERS

A 35-1 fractional factorial experiment was designed to determine the variables that significantly

affect protocol performance. In this notation, 5 refers to the number of factors considered, 3

represents the number of levels for each factor, and -1 indicates the fraction of the full factorial

over which data are collected (i.e. 3-1 or 1/3 fraction of the full factorial design) (Box et al.,

2005). The resolution was set at 5 to allow estimation of main effects and two-factor interaction

effects unconfounded by higher-order interactions. Three-factor interactions were assumed to be

confounded. The five factors considered were mixing speed, mixing time, oil-SWA contact

time, SWA solution concentration, and SWA-to-oil ratio. Substrate was evaluated as a 2 level

block effect. It was assumed that the two substrates were randomly selected from all types of

substrates and that conclusions based on the averaged responses apply to any randomly picked

substrate. The response variable of interest was the oil mass (mg) in the aqueous extract. This

represents the mass of oil released from the sand or gravel and is a direct measure of SWA

effectiveness.

SAS Proc GLM was used to statistically analyze the data, with terms for substrate, each of the 5

variables, and all 2 way interactions between the 5 variables (SAS/STAT, 1990). The fractional

factorial experiment and results are discussed in Chapter 6.

41

Figure 2.1: Rotational Mixer with Stainless-Steel Shafts, Glass Plates and Baffled Beakers Used in the Glass Plate Protocol.

Figure 2.2: Oiled Glass Plates Before and After Treatment with SWA Solution.

Before After

42

Figure 2.3: Orbital Shaker Table, Beakers and Baskets Used in the Natural Substrate Protocol.

Sand Gravel

43

Figure 2.4: Templates for 5-Drop and 12-Drop Oil Application Pattern on Glass Plates.

Figure 2.5: Template for Application of Oil and SWA to Natural Substrates in 2”x2” Wire Mesh Baskets.

44

Table 2.1: Chemical and Physical Properties of PBC Oil.

Analyte Prudhoe Bay Crude Oil Specific Gravity (at 15 °C) 0.894 kg/L API gravity (at 15 °C) 26.8 ° Saturates (% total weight) 78.3 Aromatics (% total weight) 17.6 Polars (% total weight) 2.5 Asphaltenes (weight %) 2.04 Waxes (weight %) 0.65 Sulfur 1.03 wt% Nitrogen 0.2 wt% Vanadium 21 mg/L Nickel 11 mg/L Pour Point >25 °F Viscosity at 40 °C 100 °C

14.09 cST 4.059 cST

Index 210

45

SW

A

Nam

e

App

lied

Dilu

tion

SW

A:o

il

Flas

h Po

int

Pour

Po

int

Vis

cosi

ty

Sp

ecif

ic

Gra

vity

pH

Su

rfac

e A

ctiv

e A

gent

s S

olve

nts

Add

itive

s S

olub

ilit

y in

w

ater

AQ

UA

CL

EA

N

50%

N

D

-20F

16

.5 c

p at

77F

1.

06

at 7

7F

11.8

(u

ndilu

ted)

A

nion

ic a

nd n

onio

nic

synt

heti

c C

onfi

dent

ial

Con

fide

ntia

l C

ompl

ete

BIO

SOL

VE

3%

to

6%

>20

0F

32.9

F

490

cp a

t 60F

1.

03 a

t 60F

9.

37

Con

fide

ntia

l C

onfi

dent

ial

Con

fide

ntia

l C

ompl

ete

C

LE

AN

-SP

LIT

33

%

Non

e N

A

20 c

ps a

t 73F

1.

075

at 7

3F

7.0

Con

fide

ntia

l W

ater

C

onfi

dent

ial

Com

plet

e

CN

-110

N

one

1gal

/100

ft2

Non

e 30

F

45.7

cst

at 7

8F

1.02

5 11

.4

Tra

ce a

mou

nts

of

sulf

onat

ed c

mpd

s.

Non

e C

ompl

ex

sili

cate

sol

n.

NA

CO

RE

XIT

766

4 1-

3%

116F

7F

25

cst

at 1

00F

34

cst

at 6

0F

1.02

at 6

0F

6.14

C

onfi

dent

ial

Isop

ropa

nol

Wat

er

Non

e N

A

CO

RE

XIT

958

0 N

one

1gal

/100

ft2

174F

-6

5F

3.1

cp a

t 100

F; 1

.7

cp a

t 150

F 0.

810

at 6

0F

NA

C

onfi

dent

ial

Dea

rom

atiz

ed

Hyd

roca

rbon

N

one

NA

CY

TO

SO

L

Non

e 1:

1 36

0F

10F

4.

15 C

ST

at 1

04F

0.88

77 a

t 60F

N

eutr

al

Non

e N

o P

etro

leum

D

isti

llat

es

Con

fide

ntia

l

7 pp

m s

ea

14 p

pm f

resh

D

O A

LL

#18

33

%

>20

0F

32F

1.

8 at

68F

1.

07 a

t 68F

13

.1

Con

fide

ntia

l N

one

Con

fide

ntia

l In

Wat

er

DU

O-S

PL

IT

33%

N

one

N/A

20

cps

-Bro

okfi

eld

#1 S

pind

le @

20

rpm

at 7

3F

1.07

5at 7

3F

7.0

Con

fide

ntia

l W

ater

C

onfi

dent

ial

Com

plet

e

EC

P R

ESP

ON

DE

R-S

W

1%, 3

% o

r 5%

>

200F

25

F

33.8

7 C

ST

1.03

5

9.7

6

(+/-

0.01

) C

onfi

dent

ial

Non

e C

onfi

dent

ial

Com

plet

e

FM

-186

-2

Und

ilut

ed

>20

0F

33.8

F

1.15

0.

995

7.35

C

onfi

dent

ial

Non

e C

onfi

dent

ial

Com

plet

e G

OL

D C

RE

W S

W

1% -

5%

>

200

25F

33

.87

CST

1.

035

9.7

6 (+

/-0.

01)

Con

fide

ntia

l N

one

Con

fide

ntia

l C

ompl

ete

NA

LE

IT

20

%

>21

2F

30F

1.

18

1.02

6.

8-7.

2 C

onfi

dent

ial

Non

e C

onfi

dent

ial

Fre

sh &

Sal

t W

ater

N

AT

UR

E’S

WA

Y H

S U

ndil

uted

N

/A

< 3

2F

< 1

00 C

PS

1.01

8-

9.5

Con

fide

ntia

l C

onfi

dent

ial

Con

fide

ntia

l N

/A

NA

TU

RE

’S W

AY

PC

U

ndil

uted

N

/A

< 3

2F

< 1

00 C

ps

1.01

8-

9.5

Con

fide

ntia

l C

onfi

dent

ial

Con

fide

ntia

l N

/A

PE

TR

O-C

LE

AN

0.

5% to

6%

>

200F

-1

7F

1.26

at 7

5F

0.99

at 7

5F

8.05

(10

% s

ol)

Con

fide

ntia

l N

one

Con

fide

ntia

l 10

0%

PE

TR

O-G

RE

EN

AD

P-7

1

gal/4

2 ga

l >

212

F

22F

10

3.1

at 6

0F

1.03

5 at

60F

10

.5

Con

fide

ntia

l C

onfi

dent

ial

Con

fide

ntia

l N

/A

PE

TR

OT

EC

H 2

5 U

ndil

uted

N

one

0 C

70

0 C

P 1.

03

7.5

Con

fide

ntia

l C

onfi

dent

ial

Con

fide

ntia

l 10

0 %

P

OW

ER

CL

EA

N

Und

ilut

ed

N/A

<

32F

<

100

CPS

1.

01

8-9.

5 C

onfi

dent

ial

Con

fide

ntia

l C

onfi

dent

ial

N/A

PR

EM

IER

99

1:5

160F

N

/A

15 C

PS

1.01

10

-11.

5 A

ctiv

e 80

00*-

Puri

ty

Che

mic

al <

10%

<

5%

<

5%

C

ompl

ete

SC

-100

0 20

%

>21

2F

25F

<

10C

PS

at 2

5 C

1.

009

10.2

-10.

5 C

onfi

dent

ial

Non

e C

onfi

dent

ial

Fre

sh &

Sal

t W

ater

SIM

PL

E G

RE

EN

U

ndilu

ted

> 2

00F

N

one

2.0

Cen

tistr

okes

at

78F

1.

0257

g/m

l at

72F

9.

5 C

onfi

dent

ial

Con

fide

ntia

l C

onfi

dent

ial

Con

fide

ntia

l

SP

LIT

DE

CIS

ION

SC

1:

3 N

one

N/A

20

CP

S B

rook

fiel

d #1

spi

ndle

@20

rp

m a

t 73F

1.

075

at 7

3F

7.0

Con

fide

ntia

l W

ater

C

onfi

dent

ial

Com

plet

e

SU

PER

AL

L #

38

1:5

>21

2F

20F

30

.53

SU

S a

t 100

F

1.06

at 7

0F

12.6

C

onfi

dent

ial

Non

e C

onfi

dent

ial

Mis

cibl

e

SX

-100

U

p to

200

:1

200F

-1

2F

1.81

2 1.

008

5.94

C

onfi

dent

ial

N/A

N

/A

N/A

T

OPS

AL

L #

30

1:5

>21

2F

20F

30

.53

SU

S a

t 100

F

1.06

at 7

0F

12.6

C

onfi

dent

ial

Non

e C

onfi

dent

ial

Mis

cibl

e V

ER

SAC

LE

AN

U

ndil

uted

N

/A

< 3

2F

< 1

00 C

PS

1.01

8-

9.5

Con

fide

ntia

l C

onfi

dent

ial

Con

fide

ntia

l N

/A

Tab

le 2

.2: S

urfa

ce W

ashi

ng A

gent

s L

iste

d on

NC

P Pr

oduc

t Sch

edul

e

46

Table 2.3: Major Ion Composition of “Instant Ocean” Synthetic Sea Salt.

Major Ion % Total Weight Ionic Concentration at 34

ppt salinity (mg/l) Chloride (Cl-) 47.47 18,740 Sodium(Na+) 26.28 10,454 Sulfate (SO4

2-) 6.602 2,631 Magnesium (Mg2+) 3.23 1,256 Calcium (Ca2+) 1.013 400 Potassium (K+) 1.015 401 Bicarbonate (HCO3

-) 0.491 194 Boron (B3+) 0.015 6.0 Strontium (Sr2+) 0.001 7.5 Solids Total 86.11 % 34,089.5 Water 13.88 Total 99.99%

Table 2.4: Experimental Levels to be Tested Under the Natural Substrate Protocol.

Experimental Variables Levels Substrate Sand, gravel Hydration Wet, dry Oil type PBC, IFO-180 Oil volume, μl 90, 180, 360, 450 Oil application pattern, drops 1, 5, 9 drops Oil-substrate adhesion time, hr 0.25, 1, 2, 4, 6, 9, 18 Oil-SWA ratio 1:2, 1:1, 2:1, 4:1, 10:1 SWA solution concentration, % 10, 50, 100 SWA solution volume, μl 90, 180, 360, 450, 720, 900 Oil-SWA contact time, min 5, 15, 30, 60, 120, 300 Rotational mixing speed, rpm 100, 150, 200 Mixing time, min 5, 15, 30

47

Chapter 3

PRELIMINARY TESTING

The objective of this research was to develop a laboratory protocol for determining the

effectiveness of SWAs to release crude oil from a solid substrate. Protocols had been developed

by various researchers and laboratories in the past, and all were similar in approach: oil was

applied to a substrate and weathered, SWA was applied and allowed to contact the oil, the

substrate was washed with water, and the fraction of oil removed from the surface was

quantified. Efficiencies for the SWAs were in the range of 1% to 52%. As described in Chapter

1, the substrates used in these protocols included stainless steel or porcelain troughs or coupons,

glass slides, and aquarium gravel. The major shortcoming of these protocols was the lack of

resemblance to real world conditions, either in terms of substrate type or mixing regime.

Since the protocol developed here will be adopted as an EPA method to be used by independent

laboratories, it is important that the protocol 1) be reproducible in the hands of multiple operators

and 2) provide information that can be used to predict effectiveness in the field. With this in

mind, the goal was to standardize the materials and procedures while mimicking field conditions

to the extent possible. The experimental designs described below attempted to use materials that

are commercially available or that can be custom ordered with standardized dimensions and

properties. The intent was to reduce experimental variability caused by non-standardized

equipment, such as gravel or rock substrates. The experiment had to be simple to perform to

minimize variability due to operator error. It is also had to differentiate among products based

on performance.

48

3.1 GLASS PLATE PROTOCOL

Initial testing began with borosilicate glass as the substrate. Glass was chosen because 1) it is

made from silica sand and should have similar chemical and physical properties to sand and

gravel commonly found on shorelines, 2) its surface can easily be modified to evaluate the

effects of surface characteristics, such as surface roughness and hydrophobicity, and 3) it is

relatively inert. The first prototype glass plate was made from 1/8”-thick glass. For this study,

the surface of the glass was etched using mid-range grinding powder to evaluate the effect of

surface roughness on SWA performance. The glass was cut into 1” x 2” rectangles, and the sides

were heated and shaped to form a lip so that oil would be retained on the plate. The result was a

somewhat irregular shape that would not meet requirements for standardization or

reproducibility.

An alternate design was developed and tested. The proposed design involved slicing a 1 mm

cross-section of a 35 mm x 70 mm rectangular glass tube and fusing this section to a 35 mm x 70

mm flat rectangular glass plate. The resulting plate had a 1 mm walled edge to prevent oil

runoff. A ¼” diameter glass rod was fused to one end so that the plate could be attached to a

rotary flocculation mixer (Figure 3.1 a and b). The plate surface was etched at the following

levels: none, fine, medium, coarse and extra coarse. Porous glass plates were also fabricated by

fusing powdered glass to a non-porous plate at 800°C. The powdered glass was available in four

known particle sizes such that the porosity could be calculated. Due to the excessive heat

required for this process, some warping occurred, which could have affected reproducibility. As

a result, this design was later discarded.

49

The following experimental procedure was used for initial testing:

• Apply 0.5 ml Prudhoe Bay or South Louisiana Crude Oil to the surfaces of 4 replicate

plates and weather for 18 hours at ambient temperature.

• After weathering, apply 0.5 ml 10% SWA solution to the oiled surface and allow the

SWA to contact the oil for 5 minutes.

• Attach the replicate plates to a 6-shaft flocculation mixer via the 1/4" glass rod fused to

each plate.

• Suspend the plate vertically above a 1L square glass jar containing 400 ml artificial

seawater solution.

• Lower the plates into the seawater and rotate at 100 rpm for 2 minutes.

• Elevate the plates above the water surface and allow the plates to drain for 1 minute.

• Transfer a 100 ml aqueous sample from each jar to a separatory funnel. The sample

should be taken from the middle of the water column, avoiding the recoalesced oil on the

water surface. This sample is used to determine the fraction of dispersed oil.

• Extract the dispersed aqueous samples and the remaining water fractions separately with

at least three 10 ml volumes of DCM. Record the final extract volumes. Rinse the plate

with at least three 10 ml volumes of DCM and record the final extract volume.

• Analyze extracts by UV/Vis Spectrophotometer according to the methods outlined in

Chapter 2.

A silanized and unsilanized version of each plate was tested to evaluate the effect of surface

hydrophobicity on SWA performance. Silanization is the chemical conversion of hydroxyl (OH)

50

groups, which often act as adsorption sites on silica or glass, to the inactive -O-SiR3 grouping

through the use of silane coupling agents. Silanization neutralizes surface charges, thus

eliminating non-specific binding.

Six plates were selected for preliminary testing: N (no etching), NS (no etching, silanized), XC

(extra course etching), XCS (extra course etching, silanized), P (porous), and PS (porous,

silanized). South Louisiana crude oil was applied to each plate at a volume of 0.5 ml. For the

porous plates, an additional 0.5 ml was applied to achieve complete coverage of the plate

surface. The oiled plates were weathered at room temperature for 48 hours. A SWA solution

(10% Petroclean) was then applied at a 2:1 SWA-to-oil ratio (SOR). A 10% solution

concentration falls within the range recommended by the manufacturer for extreme oil spill

remediation; a recommended volume and SOR are not specified by the manufacturer. Following

a 5 minute contact time, the plates were swirled in 450 ml artificial seawater for 5 minutes at 100

rpm using a 6-shaft rotational mixer. The seawater and the plates were extracted separately with

dichloromethane (DCM). The total oil in each fraction was determined by spectrophotometric

analysis as described in Chapter 2. This procedure was repeated without the use of a SWA to

determine the amount of oil that is displaced by water alone. Plates not treated with SWA will

hereafter be referred to as controls.

The data are summarized in Figure 3.2. Some increase in oil removal was observed for the 10%

Petroclean treatment over the no-SWA control for the NS and XCS plates; no difference was

observed for the N and XC plates. More oil was released from silanized surfaces than

unsilanized for three of the six conditions: N-control, N-treatment and XC-control. No

51

difference was observed between the silanized and unsilanized porous glass plates. To determine

the effect of weathering, the porous plates, P and PS, were tested according to the above

procedure, except the plates were not weathered prior to testing. The data are summarized in

Figure 3.3. In all cases, more oil was released from non-weathered plates than from weathered

plates. No oil was released from the controls that were weathered for 48 hours, regardless of

silanization. A slight increase in oil release due to silanization was observed for non-weathered

controls and non-weathered treatments.

Due to the heating required during fabrication of the porous plates, it was apparent that it would

be difficult to obtain true replicate plates. Surface heterogeneity would result in highly variable

results. In addition, the plates with the walled edges were fragile and expensive to fabricate. As

a result, this approach was abandoned in favor of a more standardized and commercially

available substrate.

3.2 FRITTED DISC PROTOCOL

The previously proposed glass plate experimental design was evaluated and determined to be

non-ideal due to the lack of uniformity and reproducibility of the plate surface. This design was

abandoned in favor of a design that would utilize a more standardized fritted disc.

Commercially, there are several fritted disc systems available with defined porosities. Porosities

for these discs were determined by the manufacturer according to the ASTM Test Method E128-

99 Standard Test Method for Maximum Pore Diameter and Permeability of Rigid Porous Filters

for Laboratory Use.

52

Ten porous fritted discs were used for preliminary testing. Duplicate 40-mm diameter discs were

ordered from two different manufacturers in each of the following nominal pore sizes: Ace Glass

(145-174 μm; 70-100 μm; 25-50 μm), Kontes (170-220 μm; 40-60 μm). Initial testing of

duplicate glass fritted discs was conducted using 1) a vacuum filtration apparatus and 2) a

chromatography column. Figure 3.4 shows a fritted disc after application of PBC oil.

In the vacuum filtration design, the following procedure was followed: oil was applied to the

fritted disc using a positive displacement micro-pipetter; the oiled disc was weathered at ambient

temperature for 24 hours to allow adhesion of the oil to the frit surface; the disc was inserted into

a vacuum apparatus; SWA was applied to the disc at a 10:1 SOR; the SWA was allowed to

penetrate the oil for 15 minutes prior to washing with seawater; a vacuum was applied to the

filtration apparatus and a volume of seawater was poured through the porous disc to remove any

free oil from the substrate; the wash water and the glass substrate were extracted separately with

DCM to determine the percent oil removed from the substrate.

A modification of this design was also tested, in which the fritted glass disc was placed in a

chromatography column. The application of oil and SWA was performed as stated above. Wash

water was added to the top of the chromatography column and allowed to stand for two minutes.

A valve immediately below the disc was opened and the water passed through the glass disc to

release any free oil. No vacuum or external pressure was required. The water and disc were

extracted to determine oil removal efficiency. Several diameters of column and frit were tested.

These approaches posed several problems. In the vacuum filtration setup, the variability of

53

vacuum pull was a concern with regard to repeatability. In both systems, short-circuiting of the

wash water through non-oiled pores was also a concern. Since oil is lighter than water, any

design that forced the oil to be pulled downward through the frit did not seem logical.

Consequently, this approach was also abandoned.

3.3 REVISED GLASS PLATE PROTOCOL

The glass plate protocol was again explored with an improved design. The new experimental

apparatus consisted of a six-shaft, rotational paddle mixer (Phipps and Bird, model #7790-400)

with custom-designed stainless steel shafts, rectangular glass plates and baffled 1L beakers (see

Figure 3.5). The glass plates were made of 1/4" borosilicate glass, a heavier weight than

previously used. The testing surface was 40 mm x 80 mm with a central tab on one side for

attaching to a shaft. No walls were added to this design as the intent was to dip the glass plate in

oil and allow it to drain. Substrate surface roughness was varied by grinding both sides of the

glass with grinding powders. Five levels of surface roughness were achieved: smooth (S, no

etching), fine (F, 400 grit), medium (M, 220 grit), coarse (C, 80 grit) and extra coarse (XC,

particles used for sand blasting). The surface roughness of each type of plate was estimated by

scanning electron microscopy (SEM) as discussed in Chapter 7. Stainless steel shafts were

constructed of ¼” stainless steel tubing with a ¾” adaptor and a nylon screw to grip the neck of

the glass plate. Four internal baffles were added to 1 liter heavy-weight Pyrex beakers (Fisher

Scientific) to enhance the mixing inside the beaker and prevent vortexing.

3.3.1 Oil Application

Application of the oil to the plate was initially done by dipping the plate in oil in a vertical

54

orientation and lifting the plate above the oil surface to drain and to weather. This approach was

quickly abandoned due to pooling of the oil along the lower edge of the plate. The pooled oil did

not dry completely, even after 24 hours of weathering, and high relative standard deviations

(RSDs) were observed among replicates. For a standardized protocol, the goal was to have

uniform oil coverage across the plate surface to promote uniform oil attachment. Another

problem with this approach was that the total mass of oil applied was not known, so a mass

balance would not be possible. Treatment of the oiled plate with SWA required dipping the plate

into a large volume of SWA solution, which was not practical unless very dilute solutions were

used. Testing of more concentrated SWA solutions would skew analytical results because some

SWAs are highly colored, and the colored fractions could also partition into the DCM, yielding

artificially high spectophotometric measurements. Alternately, SWA could be sprayed onto the

surfaces of the plate, but uniformity of application and determining total mass applied was a

problem.

A new approach for applying oil was tested. This method involved application of the oil to one

side of the plate using a positive displacement pipetter. Two application volumes (50 and 120

μl) and two application patterns were tested (5 and 12 drop). The final decision to use a total

volume of 50 μl was based on the holding capacity of the plate. This volume allowed room for

spreading of the oil without significant overlap of the oil droplets or spilling of oil at plate edges

(which did occur with the 120-μl volume). The purpose of dispensing multiple droplets instead

of a single droplet was to provide greater and more uniform coverage across the plate surface.

Experiments were conducted using 1) five 10-μl drops (total volume 50 μl) in an “X” pattern and

2) twelve 4-μl drops (total volume 48 μl) in a 3 x 4 block pattern to determine whether oil

55

application and spreading affect testing results. The templates for dispensing droplets as 5 and

12 drops are depicted in Figure 3.6. Experimental results for the 5- and 12- drop approach on C

and XC plates using a 50% Biosolve solution at 10:1 SOR are shown in Figure 3.7. The drop

pattern did not have a significant effect on oil removal by the product. Dispensing five droplets

in an “X” pattern gave approximately 60-70% plate coverage. This was determined to be the

optimal approach for this protocol.

The total oil applied to the surface was measured in two ways. Since a known volume of oil was

applied, the mass of oil applied was calculated by multiplying the application volume by the oil

density. However, some heavy crude oils are highly viscous, and incomplete release of the oil

from the pipette tip can occur. To determine whether the actual application volume was the

same as the expected, plates were weighed before and after oiling. The actual mass of oil

dispensed was calculated as the difference between these weights. Analysis of six experiments

(36 samples) revealed that the actual dispensed mass was 89.9±3.1% of the expected mass.

While each drop dispensed was 10% lighter than expected, the deviation from this value was

low, suggesting the pipette dispensed volume was repeatable and consistent. Since 10%

deviation falls within the limits of our acceptance criteria, estimation of recoveries was based on

expected mass for convenience.

3.3.2 Weathering Time

The purpose of a weathering time (WT) is to allow for spreading of oil on the plate surface,

evaporative loss of volatile fractions from the oil, and formation of adhesion bonds between the

oil and the substrate. For the protocol, it is advantageous to use a WT that results in very low

56

removal of oil by seawater alone so that differences between controls and treatments can be seen.

WTs ranging from 5 min to 18 hr were tested for PBC. Results for the removal of PBC from C

plates at varied WT are displayed in Figure 3.8. WTs of 6 hours or less resulted in greater than

10% removal of oil from controls. An 18-hr WT was determined to be sufficient for the oil to

adhere to the plate such that typically less than 8% of the oil was removed with seawater

washing alone. An 18-hr WT was also convenient because plates could be oiled at the end of the

day and treated, washed, and extracted the following morning. The optimal weathering time

may differ for lighter and heavier oils.

The potential for diminished recoveries due to volatility over an 18-hr WT was tested. Oil was

applied to six replicate C plates, with weight measurements taken immediately before and after

oiling. Three plates were extracted immediately with DCM; the other 3 were weathered for 18

hours, reweighed, and extracted with DCM. There were no differences in the measured

recoveries due to weathering. Unweathered recoveries were 93.4%, 86.1%, and 90.4%;

weathered recoveries were 87.3%, 91.1%, and 90.8%. The measured loss in weight of the oil

was 24.0±0.8%. However, it is the less volatile aromatic hydrocarbons that are responsible for

coloration of the oil. As a result, spectrophotometric analysis was not affected by the loss of the

lighter weight alkanes.

3.3.3 SWA Application

SWA was applied to the oiled glass plates using a positive displacement pipetter. Most SWAs

are water-soluble and bead up on an oiled surface rather than spread across it. Therefore, a

larger volume of SWA solution had to be applied to achieve complete coverage of the oiled

57

surface. For initial testing, ten 10-μl drops of SWA were applied to each of the five oil spots, for

a total plate application of 500 μl SWA and a 10:1 SOR. This volume was determined to give

adequate coverage of the oil without spilling from the plate. A 1000-μl volume (ten 20-μl drops

per oil spot) was determined to be too large and resulted in a loss of SWA and oil from the sides

of the plate. A 250-μl volume was insufficient to cover the oiled portion of the plate.

Analysis of data revealed that diluted SWA solutions were losing effectiveness with time.

Therefore, SWA aqueous solutions were prepared as needed from the stock product on the day of

the experiment. All dilutions were made in Milli-Q water. Fresh samples of SWA products

were also periodically ordered from the manufacturer.

3.3.4 Oil-SWA Contact Time

Several experiments were conducted to evaluate the effect of oil-SWA contact time on oil that

had been weathered 18 hours. A range of contact times was tested, from 5 min to 300 min, for

Corexit 9580 on F plates and for Biosolve on C and XC plates. These data are presented in

Figures 3.9, 3.10 and 3.11. Contact times greater than 5 min did not increase SWA efficiency,

and efficiency actually decreased with longer contact times for both SWAs. For Corexit, that

decrease was seen after 90 min; for Biosolve, efficiency declined after 20 to 45 minutes. Based

on this information, a contact time of 15 min was selected for this protocol.

3.3.5 Washing

One liter beakers with baffles were used for washing the glass plates. A volume of 400 ml

seawater was sufficient to cover the plate when suspended vertically inside the beaker, with the

58

bottom plate edge approximately ¼” from the bottom. Rotational mixing speeds of 150 and 200

rpm and mixing times of 5 min and 10 min were tested.

3.3.6 Testing Procedure

As discussed, variables were tested at multiple levels to determine optimal testing conditions for

this protocol. The following is a summary of the generalized testing procedures that were used

based on the conclusions listed above. Adaptations were made as necessary.

• Thoroughly clean and DCM-rinse all glassware. Glass plates should be acid rinsed to

remove residual metals and baffled beakers should be heated in a muffle furnace to remove

residual organics.

• Weigh each of the six glass plates to be used for testing and assign each a number, 1-6.

• Place the plates on a clean, level surface inside a ventilated hood.

• Using a positive displacement Micro/Pettor with capillary sleeve or an Eppendorf positive

displacement repeat pipetter with disposable tip, dispense five 10 µl drops of oil onto each

plate using the pattern shown in Figure 3.6a. For experiments using a 12 drop pattern,

dispense twelve 4 μl drops in a 3 x 4 block pattern as shown in Figure 3.6b.

• Reweigh each plate to determine the mass of oil applied. The RSD for the six replicate oil

application weights should be no greater than 10%.

• Allow the oil to spread and adhere to the plate for 18 hours in a ventilated hood at ambient

temperature. Other weathering times may also be tested.

• Using a positive displacement Micro/Pettor or Eppendorf repeater pipetter, dispense fifty

10 µl drops SWA to each of four plates (total plate application = 500 μl). SWAs are

59

applied neat or at a specified dilution, depending on the experiment. Additional SWA

application volumes and drop patterns are also to be tested. The SWA volume and drop

pattern should allow complete coverage of the oiled portion of the plate. The two plates

that do not receive SWA are untreated controls.

• Allow the SWA to contact the oil for 15 minutes. Additional contact times are to be tested,

ranging from 5 to 300 minutes.

• Meanwhile, attach stainless steel shafts to the 6-shaft rotational mixer and elevate the shafts

to the highest position.

• Dispense 400 ml artificial seawater into each of 6 baffled beakers. Label the beakers 1-6 to

correspond to the plate number.

• After the specified contact time, attach the glass plates to the stainless steel shafts. During

attachment, each glass plate should be held over a baffled beaker so that any oil or SWA

that runs off the plate will fall into the wash water.

• Lower the shafts so that the plates are submerged in seawater. The top of the plate should

be ¼” to ½” below the water surface, and the plate should not touch the bottom or sides of

the beaker.

• Turn the mixer on and allow the plates to mix at 100, 150 or 200 rpm, depending on the

experiment specifications. Mixing times ranging from 5 to 30 minutes will be tested.

• After the specified mixing time, turn the mixer off and immediately elevate the plates

above the surface of the water. Allow the plates to drain in this position for 5 minutes.

• Remove the glass plates from the stainless steel shaft; place each in a separate 250 ml

beaker. Label the beakers 1-6 to correspond to the plate number and heat at 60°C for 15

minutes to remove water. Allow plates to cool completely.

60

• Extract the 400 ml wash water according to the method outlined in Section 2.1.6 (Aqueous

Phase Extraction).

• Extract the glass plates according to the method outlined in Section 2.1.6 (Glass Plate

Extraction).

• Analyze oil concentrations in extracts according to SOP 1: Analysis of Oil Concentration in

DCM by UV/Vis Spectrophotometry.

3.3.7 Effect of Plate Roughness on SWA Performance

Five levels of plate roughness were considered: smooth (S, no etching), fine (F), medium (M),

coarse (C) and extra coarse (XC). Oil was applied in a 5 drop pattern and weathered for 18 hr.

Oil adhered least strongly to the smooth, non-etched plate. Etching aided in the spreading of oil

across the surface of the plate, which resulted in a thinner, more strongly attached film of oil.

Oil applied to the smooth glass did not spread significantly, leaving a thicker droplet that often

did not dry or attach. Thus, treatment as well as control removal efficiencies were highest for the

smooth plates. In general, SWA efficiency decreased with an increase in plate surface etching.

The increase in surface area provided greater opportunity for oil attachment. SWAs were found

to be less effective at removing this oil. Figures 3.12 and 3.13 contain removal efficiency data

for Petroclean and Corexit 9580, respectively, on S, F and C plates at 1:1 SOR. Figure 3.14

presents similar data for five SWAs on C and XC plates at a 10:1 SOR. The efficiency of a

SWA to remove weathered oil decreases with plate etching as follows:

Smooth > Fine > Medium > Coarse > Extra Coarse

61

3.3.8 Effect of SWA Dilution and SWA-Oil Ratio

SWAs were tested at numerous dilutions and SORs. SOR is defined as the ratio (v/v) of SWA

concentrate (as received from the manufacturer) to oil. When dilutions are used, a greater

volume must be applied to maintain the same SOR. Experiments were conducted using PBC oil

and coarse plates with Aquaclean, Biosolve, Petroclean, and Petrotech at the following solution

concentrations and SOR: 5% (1:1), 10% (1:1), 50% (10:1) and 100% (10:1). These data are

shown in Figure 3.15. There was little difference in SWA efficiency between the 5% and 10%

solution concentrations at 1:1 SOR. A decrease in efficiency at 10:1 SOR for all 4 SWAs was

observed when going from a 100% to 50% aqueous solution.

3.3.9 Reduced Efficiency and Reproducibility

Oil removal efficiencies were significantly lower than expected in some cases, and highly

variable in others. Experimental results varied over time and with operator. Many possible

causes for the variability were explored, including glassware contamination, shelf life of the

SWAs, accuracy of pipette dispensing, and analytical anomalies related to the

spectrophotometric analysis. However, no single cause could be identified. Due to variability in

the data collected using glass plates, a new experimental design using wire mesh baskets and

acid washed sand was explored. The design and development of this protocol are described in

detail in the next chapter.

62

Figure 3.1 – Pictures of Experimental Apparatus, Including a) Glass Plates, Rotational Mixer and Wash Jars; b) XCS and P Glass Plates After Washing.

b)

a)

63

Figure 3.2 – Removal of Weathered PBC from Glass Plates in Controls and 10% Petroclean 2:1 SOR Treatments.

0

100

200

300

400

500

600

700

800

900

N -

no

SW

A

N -

10%

Pet

rocl

ean

NS

- n

o S

WA

NS

- 1

0% P

etro

clea

n

XC

- n

o S

WA

XC

- 1

0% P

etro

clea

n

XC

S -

no

SW

A

XC

S -

10%

Pet

rocl

ean

P -

no

SW

A

P -

Pet

rocl

ean

PS

- n

o S

WA

PS

- 1

0% P

etro

clea

n

Oil

mas

s, m

g

aqueous extract plate extract

N no etchingNS no etching, silanizedXC extra course etchingXCS extra course etching, silanizedP pourous glassPS pourous glass, silanized

Applied Oil = 420 mg


Figure 3.3 – Removal of Weathered PBC from Glass Plates in Controls and 10% Petroclean 2:1 SOR Treatments.

0

100

200

300

400

500

600

700

800

900

1000

no S

WA

(48

hr

wea

ther

ing)

10%

Pet

rocl

ean

(48

hrw

eath

erin

g)

no S

WA

(no

wea

ther

ing)

10%

Pet

rocl

ean

(no

wea

ther

ing)

no S

WA

(48

hr

wea

ther

ing)

10%

Pet

rocl

ean

(48

hrw

eath

erin

g)

no S

WA

(no

wea

ther

ing)

10%

Pet

rocl

ean

(no

wea

ther

ing)

oil m

ass,

mg

aqueous extract plate extract


POROUS GLASS PLATE POROUS GLASS PLATE, SILANIZED

64

Figure 3.4 – Fritted Disc After Application of PBC Oil.

Figure 3.5 – Rotational Mixer with Custom Designed Shafts, Etched Glass Plates, and Baffled Beakers.

Figure 3.6 – Templates for a) Five and b) Twelve Drop Oil Application Patterns.

b) a)

65

Figure 3.7 – Removal of Weathered PBC from Coarse and Extra Coarse Plates by Corexit: Comparison of 5- and 12-Drop Oil Application Patterns.

83.4

90.895.1

58.9

89.8 88.8

95.0 94.8

0

10

20

30

40

50

60

70

80

90

100

110

C + 100% Biosolve (10:1) XC + 100% Biosolve (10:1) C + 50% Biosolve (10:1) XC + 50% Biosolve (10:1)

Oil

Rem

ova

l, %

5 drop oil application pattern

12 drop oil application pattern

Figure 3.8 – Removal of PBC from Coarse Glass Plate at Varied Weathering Times.

80.478.6

9.4

13.0

27.9

86.0

68.2

0

10

20

30

40

50

60

70

80

90

100

5 min 30 min 1 hr 2 hr 4 hr 6 hr 18 hr

Weathering Time

Oil

Rem

ova

l, %

66

Figure 3.9 – Effect of Oil-SWA Contact Time on Removal of Weathered PBC Oil from Finely Ground Glass Plates by Corexit at 3:1 SOR.

91.3 90.087.1

90.387.8

73.7

0

10

20

30

40

50

60

70

80

90

100

5 min contact 10 min contact 20 min contact 40 min contact 90 min contact 300 min contact

Oil-SWA Contact Time

Oil

Rem

ova

l, %

Figure 3.10 – Effect of Oil-SWA Contact Time on Removal of Weathered PBC Oil from Coarse Glass Plates by Biosolve at 10:1 SOR.

84.4 85.3 84.6

54.758.2

54.8

0

10

20

30

40

50

60

70

80

90

100

5 min 20 min 45 min 90 min 180 min 270 min


Oil

Rem

ova

l, %

67

Figure 3.11 – Effect of Oil-SWA Contact Time on Removal of Weathered PBC Oil from Extra Coarse Glass Plates by Biosolve at 10:1 SOR.

84.4 84.4

78.6

68.3

60.257.8

0

10

20

30

40

50

60

70

80

90

100



Oil

Rem

ova

l, %

Figure 3.12 – Release of PBC from Smooth, Fine and Coarse Plates in Controls and 20% Petroclean 1:1 SOR Treatments.

41.0

50.4

0.3 0.62.6 3.8

0

10

20

30

40

50

60

70

80

90

100

Control Petroclean Control Petroclean Control Petroclean

Oil

Rem

ova

l, %

Smooth Plate Fine Plate Coarse Plate

68

Figure 3.13 – Release of PBC from Smooth, Fine and Coarse Plates in Controls and Corexit 1:1 SOR Treatments.

41.4

1.3

47.8

0.4

67.9

18.2

0

10

20

30

40

50

60

70

80

90

100

Control Corexit Control Corexit Control Corexit

Oil

Rem

ova

l, %

Smooth Plate Fine Plate Coarse Plate Figure 3.14 – Effect of Plate Etching on PBC Release from SWA Treated Glass Plates.

87.8

64.1

77.783.0

82.2

57.1

73.5

63.1

83.7

70.4

0

10

20

30

40

50

60

70

80

90

100

Aquaclean Biosolve PetroClean Petrotech Corexit

Oil

Rem

ova

l, %

Coarse Plate

Extra Course Plate

69

Figure 3.15 – SWA Efficiency as a Function of Dilution and SOR.

0

10

20

30

40

50

60

70

80

90

100

100 50 10 5

SWA Solution Concentration (v/v), %

Oil

Rem

ove

d, %

Aquaclean

Biosolve

Petroclean

Petrotech

SOR = 10:1 SOR = 1:1

70

Chapter 4

DEVELOPMENT OF THE NATURAL SUBSTRATE PROTOCOL

In developing a laboratory protocol for determining the effectiveness of SWAs, several initial

experimental designs were proposed and discarded for reasons explained in Chapter 3. The

protocol that was ultimately adopted used natural substrates (i.e., sand and gravel) to better

reflect real world applications. Research was conducted to determine the effect of protocol

variables on the performance of surface washing agents in removing Prudhoe Bay Crude (PBC)

oil from these substrates. The following variables were tested at multiple levels: substrate type

(sand and gravel), substrate moisture (wet and dry substrate), substrate drain time (DT, for wet

sand applications), mode/pattern of oil application, oil application volume, oil weathering time

(WT), SWA application volume, SWA dilution, SWA-to-oil ratio (SOR), oil-SWA contact time

(CT), rotational mixing speed (MS), and mixing time (MT). Preliminary testing revealed that

some variables did not significantly affect protocol performance; these variables were fixed at

values convenient for testing purposes, as described below. A fractional factorial experiment

was designed and completed to confirm which variables are significant to surface washing, and

to establish optimal levels to be included in the final protocol. The results of the fractional

factorial experiment and recommendations for the final protocol are discussed in Chapter 6.

4.1 MATERIALS

• ASTM 20/30 silica sand and 1/8 x 1/16 FilPro filter gravel (U.S. Silica Company) were

selected as representative shoreline substrates. Both are >99% whole grain silica.

Substrates were acid washed with a 1M nitric acid solution, rinsed with tap water until the

71

wash water had a neutral pH, rinsed with Milli-Q water, and baked overnight at 500°C

prior to use.

• Prudhoe Bay Crude (PBC), a medium-weight EPA/American Petroleum Institute (API)

standard reference oil, was used in these studies. IFO-180, a heavy fuel oil, was also

tested.

• Six SWAs were selected from those listed on the NCP Product Schedule: Aquaclean,

Biosolve, Corexit 9580, Petroclean, Petrotech25, Superall.

• Artificial seawater was prepared at a concentration of 34 parts per thousand (ppt) using the

synthetic sea salt “Instant Ocean” (Aquarium Systems, Mentor, OH).

• Pesticide quality dichloromethane (DCM) was used as the extraction solvent for standards

and samples.

• The substrates were contained in 125 cm3 baskets constructed of 30-mesh stainless steel

wire cloth and supported by a stainless steel frame (Hillside Wire Cloth Co., Inc.,

Bloomfield, NJ).

• An Eppendorf Repeater Pro positive displacement pipetter and/or a Brinkmann Eppendorf

Research Pro Pipette capable of dispensing 5 to 100 μL was used to dispense the required

volumes of oil and SWA.

• Pyrex 600-mL beakers held the submerged baskets during seawater washing and DCM

extraction.

• A LabLine 3520 orbital platform shaker with adjustable rotational speed provided washing

agitation.

• Aqueous sample and aqueous standard extractions were performed in 250-mL separatory

funnels with ground glass stoppers and Teflon stopcocks.

72

• An Agilent 8453 UV/visible spectrophotometer with standard silica 10 mm path length

rectangular cell was used for quantitation of oil in the sample extracts. This methodology

was chosen to create reproducible and repeatable conditions for a valid testing protocol.

4.2 EXPERIMENTAL DESIGN

The following describes the initial testing procedure under which most of the protocol

development work occurred. Based on testing results, modifications were later made to this

procedure. The final proposed protocol is described at the end of this Chapter.

The SWA effectiveness tests were conducted at room temperature (20±3ºC). Four replicate

treatments and duplicate controls were included in most experiments. A 25-mL volume of sand

was added to each stainless steel basket using a 25-mL graduated cylinder. The amount of sand

used was eventually decreased to 15 ml to reduce consumption of the extraction solvent. This

change had no apparent effect on protocol results, except to limit the volume of oil that could be

applied to the sand bed. For wet sand experiments, the baskets were submerged in artificial

seawater for 30 sec to wet the surface of the sand particles. The substrate was permitted to drain

for 10 min prior to oil application. Using a positive displacement pipetter, PBC oil was applied

to the level surface of the sand as nine 10 μL drops in a 3x3 block pattern for a total application

volume of 90 uL. The oil was weathered on the sand for 18 hr at room temperature in a well

ventilated hood prior to application of SWA. SWA was applied to the oiled sand in the same 9

drop pattern. Following a 15 min oil-SWA contact period, the baskets were submerged in 100

mL seawater and immediately agitated on an orbital shaker platform for 5 min at 150 rpm.

Baskets were subsequently elevated above the water level and allowed to drain for 5 minutes.

73

Figure 4.1 shows oiled sand before and after washing. The wash water was transferred to 250-

mL separatory funnels and extracted with three 15-mL aliquots of DCM. The extracts were

adjusted to a final volume of 50 mL and stored at 5°C in 50-mL glass vials with air-tight caps

and Teflon-lined septa. Baskets were placed in clean 600-mL beakers and extracted with three

consecutive 50-mL aliquots of DCM by shaking at 150 rpm for 5 min per extraction on the

orbital shaker. A 50-mL volume was sufficient to cover the surface of the sand contained in the

basket. For each sand sample, the three consecutive extracts were added to the same 250-mL

graduated cylinder, brought to a final volume of 180 mL, and stored at 5°C in glass vials with

air-tight caps and Teflon-lined septa. For experiments using 15 ml sand, the sand was removed

from the basket and placed in a beaker, extracted with three 20-ml aliquots of DCM and brought

to a final volume of 60 ml. The basket was also rinsed as part of the extraction process. All

extracts were analyzed by UV/visible spectrophotometry within 48 hours of collection, as

outlined in Chapter 2. The efficiency of the SWA was determined based on the mass of oil

released into the wash water relative to the total mass of oil applied.

4.3 EFFECT OF TESTING VARIABLES ON PROTOCOL

4.3.1 Substrate Type

Two substrates were selected for testing: ASTM 20/30 silica sand and FilPro 1/8 x 1/16 filter

gravel (both from U.S. Silica Company). Both were >99% whole grain silica, which is the most

common constituent of inland and non-tropical coastal sand. The properties of these substrates

are described in further detail in Chapter 7. A third substrate, 5/8 x 3/8 common pea gravel

(source unknown) was also tested, but was not included in the fractional factorial experiment.

Experiments conducted during the protocol development phase used only silica sand as the

74

substrate. The two gravel substrates were tested after the other protocol variables had been

largely optimized.

The effect of substrate type on SWA efficiency was tested under wet and dry conditions. PBC

oil was applied to the 3 substrates (sand, FilPro gravel, and pea gravel) as nine 10 :l drops,

weathered for 18 hr, and treated with undiluted Aquaclean at 2:1 SOR. The data are shown if

Figure 4.2. On dry substrates, a decreasing trend in SWA efficiency was observed with particle

size. However, there was no statistically significant difference between sand and FilPro gravel

(p=0.2384) or between the two gravels (p= 0.0726); the difference between sand and pea gravel

was significant (p=0.0012). According to this result, more oil was released from sand than from

pea gravel. Since the surface area per unit mass of substrate decreases with the particle size of

the substrate, this is counter to expectations. This observation could be due to increased

spreading of the oil across the larger surface of the pea gravel and into crevices, resulting in

decreased contact between the SWA and the oil prior to washing. Larger standard deviations

were observed on wet substrates, making it difficult to determine a trend. The average efficiency

for Aquaclean on wet pea gravel was lower than on the other substrates, but analysis of variance

revealed no statistical differences among the three substrates.

Substrate type was also built into the fractional factorial experimental design as a two level block

effect. Sand and FilPro gravel were tested. The effect of substrate type was determined to be

statistically significant (p<0.0001), with a mean response of 109.6 for sand and 96.1 for gravel.

However, the relative significance of all other variables and the conclusions drawn were the

75

same for both substrates. Both substrates will be adopted as part of the final testing protocol.

The fractional factorial experiment is discussed in detail in Chapter 6.

4.3.2 Substrate Hydration and Drainage Time Effects

The effect of substrate hydration on protocol performance was evaluated at numerous stages of

protocol development by conducting experiments on dry and wet substrates. During the early

stages, experiments were conducted on wet and dry sand under the following conditions: 15 min

WT, 5 min CT, 150 rpm MS, 5 min MT, undiluted SWA at 2:1 SOR. Duplicate untreated

controls and four treatment replicates were included in each experimental run. The primary

difference between dry and wet experiments was evident in the untreated control (Figure 4.3).

When oil was applied directly to dry sand, the oil bonded to the sand particles and was not

released by washing with water alone. This gave a clear differentiation between treated and

untreated samples, which is an important consideration when evaluating a testing protocol.

When oil was applied to wet sand, the oil did not strongly adhere to the sand particles, and

significant amounts of oil were released in the untreated controls. This difference was not

observed in samples treated with SWA. The mass of oil removed in SWA-treated samples was

similar for both dry and wet sand experiments.

For wet sand experiments, drain time (DT), the length of time between wetting the sand and

applying the oil, was tested to evaluate its effect on oil release from controls. Baskets were

submerged in seawater for 5 min and allowed to drain by gravity for 1, 5, 10, 15 or 20 min

before oil application. The protocol was then performed under the following conditions: 15 min

WT, 150 rpm MS, and 5 min MT. SWA was not applied. Drain times from 1 to 20 min did not

76

have a significant effect on oil release (Figure 4.4). The relative standard deviation (RSD) across

samples was 8.9% with no apparent trend. Therefore, a 5 min drainage time was chosen for the

wet sand experiments.

During later stages of protocol development, the effect of applying oil to wet and dry sand was

evaluated using 18-hr WT. Other operational variables were as follows: 5 min DT (wet sand

only), undiluted SWA at 2:1 SOR, 150 rpm MS, 5 min MT. With the longer WT, oil released

from untreated controls was reduced to 22.2 ± 2.8% on wet sand; oil released from dry sand was

4.1 ± 0.8% (Figure 4.5). While the 15 min WT yielded similar results for SWA treatments on

wet and dry substrates, a significant difference was observed at 18 hr WT. Oil released from

treated wet substrates was higher than from dry by 14% to 38% (absolute percent).

These experiments were repeated by a different operator to determine the repeatability of the

results. As seen in Figures 4.6 and 4.7, there was better agreement between the two operators for

dry sand experiments than for wet. The average relative percent difference between operators

was 12.6% for dry sand, 26.7% for wet sand. The largest RPD was 53% for Petroclean on wet

sand.

Wet and dry conditions were also used in testing the three substrate types: sand, FilPro gravel,

and pea gravel. Results were discussed in the previous section and are shown in Figure 4.2.

Controls with no SWA treatment had average removals of 4.4% for dry substrates and 12.4% for

wet. Standard deviations were again much higher for wet substrates. RSDs for dry substrates

77

were between 2.4% and 8.6%, while RSDs for wet substrates were 17.1% (sand), 31.0%

(gravel), and 18.8% (pea gravel).

An argument can be made for both wet and dry sand applications. The decision to apply the oil

to wet sand is reasonable since interstitial water will be a factor in sandy beach applications.

However, oil that deposits on large rocks, rip rap, walls, piers, etc. may dry between tidal cycles.

In order to be useful, the testing protocol must be able to 1) distinguish between good and bad

performance relative to controls that receive no treatment and 2) give reproducible results. Wet

sand applications resulted in higher release of oil from controls and had higher standard

deviations among sample replicates and between operators. Therefore, application of oil to dry

sand is recommended for this protocol.

4.3.3 Mode and Pattern of Oil Application

There are numerous possible methods for applying oil to the substrate, including pulling the

basket of sand through a slick on the surface of the water, using a specific weight or volume of

pre-oiled substrate, and dispensing a specific volume using a positive displacement pipetter. For

a standardized protocol, it is important to know the total mass of oil applied so that a mass

balance can be performed. Therefore, the pipetter approach is preferred. However, due to the

viscosity of oil, pipette dispensing can be a source of error. It is recommended that an electronic

pipetter or a positive displacement pipetter be used for increased accuracy.

During early stages of protocol development, the mode of oil application was tested using a total

applied volume of 100 μL PBC dispensed as 1, 2, 5, or 10 drops. Oil was applied to wet sand

78

and weathered for 15 min. The removal of PBC from sand by washing without SWA treatment

was 57.7%, 59.5%, 56.3% and 54.1% for the 1, 2, 5 and 10 drop application methods,

respectively. No significant difference was observed among the application methods (RSD =

3.5%). However, applying the oil as multiple drops across the surface of the sand serves the

primary purpose of ensuring increased contact surface area between the sand and the oil. For the

protocol, a 9 drop pattern in a 3 x 3 block design was chosen.

Other approaches were also tested, including stirring the oil into the sand and application of the

oil in a solvent carrier. However, there was no advantage to these approaches over direct

application of the oil to the sand surface. Stirring the oil into the sand left an oil residue on the

stirrer, which had to be extracted and accounted for in the mass balance. SWA applied to the

surface without additional stirring (as it would be on a shoreline) would contact only the top

layers of the oiled sand prior to washing. This resulted in decreased efficiency and increased

variability. Dissolving the oil in a solvent prior to application would provide a more even

coverage of the sand particles throughout the depth of the bed. However, the solvent could affect

how the oil attaches to the surface of the particles. Since there is no real world justification for

this approach, it was tested but not adopted.

4.3.4 Oil Weathering Time

Weathering time, the time between oil application and SWA application, allows for (1)

evaporation and drainage of some interstitial water, (2) spreading of oil on the substrate surface,

(3) evaporative loss of volatile fractions from the oil, and (4) formation of adhesive bonds

between the oil and substrate. The length of time the oil contacts the sand prior to treatment may

79

affect the release of oil from the sand. It is desirable to minimize this release in untreated

controls in order to maximize the differences between control and treated samples. The effect of

oil weathering time on the protocol was tested by varying the amount of time the oil remained on

the sand prior to SWA application. Oil weathering times of 5, 15, 60, 360 and 1080 min were

examined. Experiments were conducted on wet sand with 5 min CT, 150 rpm MS, and 5 min

MT. Only control samples with no SWA application were included in this experiment since the

primary interest was to reduce the oil removed in the controls.

Weathering times less than 60 min did not appear to alter oil release from the controls (Figure

4.8). At 5, 15 and 60 min, the oil recovered from the wash water was 52% to 56% of the total oil

applied. After 6 hours, a small decrease was noticed, but 43% of the applied oil was still

released. At 18 hrs, the oil released was reduced to 22%, and the mass of oil remaining on the

sand was significantly greater than the mass released into the wash water. Thus, all subsequent

experiments were conducted with a WT of 18 hr.

4.3.5 Mode and Volume of SWA Application

Several approaches for application of SWA to the sand surface were proposed. On shorelines,

SWA would be applied through high pressure sprayers, usually in diluted form. However, for

laboratory testing, it is difficult to control the volume and uniformity of application using a spray

nozzle. Therefore, SWA was applied drop-wise in the same manner as the oil. Volumes of

applied SWA solution varied with each experiment, depending on the solution concentration and

SOR being tested. The maximum volume that could be applied to the 25-ml sand bed without

breakthrough was 720 μl. The effect of product dilution and SOR are discussed below.

80

4.3.6 Oil-SWA Contact Time

The effect of oil-SWA contact time was examined for Corexit (oil-soluble) and Petroclean

(water-soluble) on wet sand under the following operating conditions: 18 hr WT, undiluted SWA

at 2:1 SOR,150 rpm MS, and 5 min MT. Samples were then washed at 5, 15, 30, 60, 180, and

360 min after SWA application. For both SWAs, the length of time the SWA contacted the oil

prior to washing did not have an effect on SWA effectiveness (Figure 4.9). Based on analysis of

variance (ANOVA), the differences in mean values among the washing times are not great

enough to exclude the possibility that the difference is due to random sampling variability.

Therefore, the conclusion is that contact time between SWA and oil does not play a significant

role for these two SWAs (p = 0.616 for Corexit; p = 0.849 for Petroclean). Based on these

findings, an oil-SWA contact time of 15 min was chosen as a reasonable contact time for

subsequent experiments.

Under the glass plate protocol, contact times greater than 5 min did not increase SWA efficiency,

and efficiency actually decreased at contact times greater than 90 min for Corexit and 45 min for

Biosolve. These findings are discussed in Chapter 3 and are consistent with the findings here.

4.3.7 SWA Dilution Effects

The NCP Product Schedule provides information about the manufacturer recommended

application rate and/or procedure for every SWA listed. Usually the manufacturer specifies a

product dilution rate or range, which may vary depending on the severity of the cleaning

application. However, a recommended application volume per unit area or product to oil ratio is

81

given for only a few of these products. SWAs are typically applied to shorelines in diluted form

using high pressure hoses until a desired result is achieved. The volume required will depend on

the substrate being cleaned, the extent of oil weathering, the forcefulness of the spray, and

environmental conditions. Some of these conditions can be evaluated in the lab (e.g. substrate

type, oil weathering, temperature); others cannot (e.g. the effect of high pressure application).

With such a broad range of application instructions and without specific recommendations for

volumes or ratios, it is difficult to objectively quantify and compare effectiveness for these

products under standard laboratory conditions. Extensive testing had to be done to determine the

effects of product dilution, volume, and SOR on performance.

Following are the NCP Product Schedule recommendations for the SWAs in this study:

• Aquaclean should be diluted to 50% with fresh water and applied to shorelines and beaches

using a pressure spray to cover the entire contaminated area. Rinse with fresh water.

• BioSolve® is a highly concentrated product and must be diluted with water before use.

Typical dilution rates are 6%, 3%, or 1%. Apply through fire hose, power washers, steam

powered units, or chemical boom sprayers with nozzles that produce a shearing action. For

shoreline cleanup involving heavy or weathered crude, presoak with a 6% solution as

necessary.

• Corexit 9580 is sprayed directly on oiled rocky shorelines full strength as supplied. After a

soak time of zero to thirty minutes, rinse with water. Apply 1 gallon per 100 square feet.

Diluting with water will reduce effectiveness.

82

• Petro-Clean is a highly concentrated product and must be diluted before use. Normal

recommended dilutions are from 0.5% to 6%. Apply through power washers or garden type

sprayers. On heavy or weathered crude, pre-soak with 6% as necessary.

• Petrotech 25 can be applied neat in its factory supplied concentrate form for surface

scrubbing or in diluted form through high pressure nozzles. The application rate is

unaffected by neat or diluted application as long as the concentrate to oil ratio remains the

same.

• Superall should be diluted 1 part to 5 for cleaning of crude oil and 1 part to 30 for pressure

washing applications.

The effect of diluting a SWA during the application process was tested 1) by applying 100%,

50%, and 10% solutions of SWA using a fixed solution volume (variable SOR), and 2) by

applying the same mass of SWA using multiple concentration-volume combinations (constant

SOR). The first set of experiments was conducted on wet and dry sand under the following

conditions: 90 μL PBC, 10 min DT, 18 hr WT, 180 μL neat SWA or diluted SWA solution, 15

min CT, 150 rpm MS, and 5 min MT. Quadruplicate treatment samples and duplicate controls

were included in each experiment. The shortcoming of this design was that SWA concentration

could not be evaluated separately from SOR. SORs for the 100%, 50%, and 10% solutions were

2:1, 1:1, and 1:5 SOR, respectively.

Data from these experiments are presented in Figures 4.10 through 4.13. In each graph, the

control data points represent the averages of three sets of duplicate samples run simultaneously

with each of the three treatment levels (6 samples total). The average removal of oil from

83

controls was less than 5.2% from dry sand and less than 25.3% from wet sand in all experiments.

As with other wet-dry comparisons done previously, release of oil from wet sand was greater

than from dry sand by an average of 21.8 ± 9.4% for treatments and controls. In general, SWA

efficiency increased with solution concentration and SOR. However, there was no difference

between the 50% and 100% treatments for Aquaclean and Superall on wet sand and for

Petroclean, Petrotech and Superall on dry sand. The highest removals were achieved by

Petrotech at 100% on wet sand (75.8%) and by Aquaclean at 100% on dry sand (61.5%).

Superall was the least effective SWA overall, with effectiveness not exceeding 48% on wet sand

and 23% on dry.

In the second set of experiments, four SWAs were tested undiluted and at 3%, 10%, and 30%

aqueous solution concentrations. This covers the range of dilution recommendations for all

products. The amount of SWA concentrate applied was the same in all treatments (i.e., constant

SOR), but the dilution of the product varied. Thus, concentration could be evaluated

independently from SOR. Treatment concentrations and volumes were as follows: 180 μL

undiluted SWA; 600 μL 30% solution; 1800 μL 10% solution; 6000 μL 3% solution.

Experiments were conducted on wet sand under the following conditions: 90 μL PBC, 10 min

DT, 18 hr WT, 2:1 SOR, 15 min CT, 150 rpm MS, and 5 min MT. Samples were run in

quadruplicate and controls were run in duplicate at every dilution level.

The results of this dilution study are shown in Figures 4.14 through 4.17. The average release of

oil from the controls across all experiments was 19.8 ± 5.3%. Data indicate that diluting the

SWA did not impair efficiency as long as the total mass of SWA (i.e., product concentrate)

84

applied was constant. Trend lines through the averaged oil removal efficiencies at each dilution

level reveal a slight negative slope (decrease in efficiency with increase in dilution) for all

products; however, the slopes across controls associated with the same treatment levels were also

negative for 3 of the 4 sets. For each product, an analysis of variance (ANOVA) was performed

for the 4 replicate samples at each dilution level to determine if there was any statistical

difference due to treatment. No significance was found across concentration level (p = 0.815,

0.569, 0.089 and 0.367 for Aquaclean, Petroclean, Petrotech and Superall, respectively). Thus, it

was concluded that SWA solution concentration is less critical than the overall mass of SWA

applied to treat a given amount of oil.

Results of the fractional factorial experiment confirm that SWA solution concentration is not a

significant factor in surface washing. Those findings are discussed in Chapter 6.

4.3.8 Oil Volume and SWA-to-Oil Ratios

The objective of a laboratory effectiveness protocol is to provide data that will be predictive of

SWA performance in the field. Since real world applications will involve treating large volumes

of oil, it is important to demonstrate that oil volume does not significantly affect protocol

performance for a given SOR condition. The relative ratio of SWA to oil has been determined to

be significant. However, since preliminary testing had been primarily conducted using an oil

volume of 90 uL, this had only been tested on a small volume scale. Additional testing was done

to determine the applicability of the protocol to larger oil volumes.

85

Two sets of experiments were conducted to determine the importance of SOR and oil volume on

protocol performance. In the first set of experiments, a fixed volume of undiluted SWA (180 uL)

was applied to a range of PBC volumes (90, 180, 360, and 450 uL) on wet sand, resulting in

SORs of 2:1, 1:1, 1:2, and 1:2.5. Larger volumes of oil were attempted, but the oil penetrated the

sand matrix and spilled out of the basket during application. Untreated controls were run for

each PBC volume, with averaged oil releases ranging from 17.2% to 23.5%. As expected, SWA

efficiency increased with SOR (Figure 4.18). Aquaclean, Petroclean, and Petrotech performed

better than Superall at every SOR, but their relative rankings changed with SOR. Aquaclean and

Petroclean were comparable in effectiveness at every SOR; the relative standing of Petrotech

improved significantly at the highest SOR, causing it to surpass Aquaclean and Petroclean.

While Superall was inferior at every level, its performance also increased significantly when the

SOR was increased from 1:1 to 2:1. The relative increase in percent oil removal in going from

1:2.5 to 1:2 and from 1:2 to 1:1 was 4.0% and 3.8%, respectively; the increase in going from 1:1

to 2:1 was 19.5%. This suggests that Petrotech and Superall may require higher SOR than the

other products to be effective. Overall, these data confirm that the ratio of SWA to oil is

important and that the relative percent release of oil from untreated controls is independent of oil

volume for the volumes tested.

The next set of experiments tested four oil volumes (90, 180, 360, 450 uL) each at three SORs

(2:1, 1:1, 1:2). Thus, the effect of oil volume was evaluated at each SOR. Aquaclean and

Petrotech were applied to PBC on wet sand and the protocol was executed under the standard

operating conditions. As before, removal efficiency increased with SOR for each level of oil

volume tested. Oil removal vs. SOR data are presented in Figure 4.19 for Aquaclean and Figure

86

4.21 for Petrotech. The same data are replotted in Figure 4.20 (Aquaclean) and Figure 4.22

(Petrotech) as percent removal vs. oil volume to highlight the significance of oil volume. Within

a given SOR, no trend was observed with oil volume. Volumes that produced the highest

removal efficiencies at one SOR had lower relative efficiencies at other SORs. For Aquaclean,

the greatest spread in data occurred at 2:1 SOR, with the highest effectiveness occurring for 180

µl oil and the lowest for 360 µl oil. ANOVAs were run on replicate sample responses at four oil

volumes to determine if oil volume significantly affects SWA performance. These analyses

reveal that oil volume is significant for the 2:1 (p = 0.039) and 1:1 (p = 0.027) ratios, but not for

the 1:2 (p = 0.115) ratio. However, this significance is due to the 360 µl oil condition at 2:1 and

the 180 µl oil condition at 1:1. The other three oil volume levels were not statistically different

from each other. Because the levels that cause significance are intermediary levels, this suggests

that oil volume does not significantly affect oil removal. For Petrotech, ANOVAs revealed no

significant differences caused by oil volume at each SOR (p = 0.111 at 2:1; p = 0.100 at 1:1; p =

0.148 at 1:2). The conclusion drawn from these data is that SOR is more important than oil

volume in determining SWA performance.

The importance of SOR was examined using undiluted SWAs at ratios of 1:1, 2:1, 4:1 and 10:1

for dry sand applications. Data are shown in Figure 4.23. An increasing trend was observed

with increasing SOR for all products except Corexit, which did poorer at the 10:1 ratio than at

4:1. Efficiencies greater than 90% were achieved by Biosolve, Aquaclean, and Petroclean at

10:1 SOR. The ranking of Petroclean relative to the other SWAs increased dramatically when

increased from 2:1 to 10:1 SOR. At the 2:1 level, Petroclean ranked 5th with an effectiveness of

39.6%; at 10:1, it ranked 2nd with 100.6% oil removal. The effectiveness of Petrotech remained

87

relatively unchanged at 1:1, 2:1 and 4:1 SOR, with efficiencies of 39.7%, 52.0%, and 51.6%,

respectively. At 10:1, its effectiveness jumped to 73.2%. Higher ratios would have to be tested

to determine the optimal SOR for this product.

4.3.9 Mixing Speed and Time

The effects of rotational mixing speed and mixing time were evaluated for Corexit and untreated

controls at 3 levels in a factorial experimental design. The levels tested were 100, 150, and 200

rpm MS and 5, 10, and 20 min MT. Other operational variables were held constant as follows:

wet sand, 18 hr WT, 2:1 SOR, and 5 min CT.

Mixing speed had the greatest effect on oil removal for both controls and treatments (Figures

4.24 and 4.25). The release of oil in the control averaged 4.6 ± 2.9% at 100 rpm and 71.0 ± 1.2%

at 200 rpm across the three mixing times. For SWA treatments, the release ranged from 20.9 ±

4.4% at 100 rpm to 89.4±1.1% at 200 rpm. Differentiation in performance between the

treatment and control was evident at each level, with 15% to 28% more oil released by Corexit.

200 rpm was an excessively rigorous mixing speed and resulted in the breakage of some beakers.

Thus, it would be an impractical operating speed for the protocol. A mixing speed of 100 rpm

yielded low release of oil from the controls, but also lower efficiencies for Corexit. The

maximum difference between Corexit and the control occurred at 150 rpm at each of the three

mixing times.

Differences in protocol performance for the control and Corexit were not strongly linked to the

length of mixing time. The relative standard deviation (RSD) for Corexit performance across the

88

three mixing times was less than 2% for 150 and 200 rpm conditions. For the control under the

same conditions, the RSD was less than 6 %. The greatest variability across mixing times was

observed at 100 rpm where high RSD values were a result of lower oil removals on average and

a 10 min mixing time yielded the highest removal. Based on these preliminary data, a rotational

mixing speed of 150 rpm and mixing time of 10 minutes is proposed.

4.3.10 Oil Type

All protocol development work was done using Prudhoe Bay Crude, a medium weight crude oil.

However, some preliminary testing was done using South Louisiana crude, Bunker C, and IFO-

180 refined oils. South Louisiana Crude (SLC) is lighter than PBC and is included as one of the

test oils in the Baffled Flask Test for determining the effectiveness of oil dispersants. Bunker C

and IFO-180 are heavier oils. The effectiveness of SWAs may differ depending on oil type, and

at least two oils may be specified as part of the final protocol.

4.3.10.1 SLC and Bunker C

The release of oil from wet sand after 18 hr WT was evaluated for SLC and Bunker C. Drain

times of 1, 2, 5, 10, 15, and 20 minutes were tested without replication (Figure 4.26). For the

lighter crude oil, high release of SLC from untreated controls was observed (57.3 ± 6.1% across

all drain times). It was concluded that SLC will not be used in the protocol because high release

of oil from controls will make it difficult to determine effectiveness in SWA treatments. For

Bunker C, minimal oil was released from the controls regardless of drain time (average 4.0 ±

0.5%). When treated with undiluted Aquaclean at 2:1 SOR, Bunker C was removed from wet

sand with an efficiency of 33.9 ± 3.2% (Figure 4.27). This value is less than the removal for

89

either PBC (69.0 ± 7.6%) or IFO-180 (54.4 ± 6.7%) under the same conditions, suggesting that

Bunker C is harder to remediate than the other oils tested.

4.3.10.2 IFO-180

Preliminary experiments were conducted using IFO-180, and results were compared to those

obtained for PBC oil. The variables WT, MT, MS, and CT were evaluated on wet sand using a

fixed oil volume of 90 μl. Undiluted Aquaclean and Petroclean were tested at 2:1 SOR. Aqueous

and sand extracts were analyzed by UV-vis spectrophotometry at three different wavelengths and

quantified using IFO-180 calibration standards.

Weathering times of 5 min, 15 min, 1 hr, 6 hr and 18 hr were tested for IFO-180 on wet sand and

with no SWA. Minimal amounts of IFO (5.4 ± 1.7%) were released into the aqueous solution

regardless of weathering (Figure 4.28). After 5 min weathering, only 5.5% of the oil was

removed with seawater washing; 2.7% was removed after 18 hr weathering. When treated with

Aquaclean at 2:1 SOR (Figure 4.29), no difference in oil removal was found for WTs of 5 min, 1

hr or 18 hr (56.6%, 55.2% and 57.5%, respectively). An anomalous value of 46.1% was

observed at 6 hr; however, differences in oil removal due to weathering time were found to be

non-significant by ANOVA analysis (p = 0.464).

Oil-SWA contact times for Aquaclean and Petroclean were tested at the following levels: 5 min,

15 min, 30 min, 1 hr, 3 hr and 6 hr. The SWAs were applied to IFO-180 at 2:1 SOR after 18 hr

WT. As shown in Figure 4.30, CTs greater than 5 min did not change SWA effectiveness. The

highest release of IFO-180 was achieved by Aquaclean at 5 min CT (53.0 ± 2.5%). However, no

90

difference was observed for this treatment at CTs ranging from 15 min to 6 hr (35.9 ± 1.9%).

Petroclean was ineffective at treating IFO-180, with an average removal of 6.5 ± 1.9% across

CT.

Three levels of MS and MT were evaluated in a full factorial experimental design. Mixing

speeds of 125, 150, and 200 rpm and mixing times of 5, 15, and 30 min were tested to determine

their effects on removal of IFO-180 from wet sand in controls and Aquaclean treatments (2:1

SOR). IFO-180 was applied to wet sand and weathered for 18 hours. The results for untreated

controls are shown in Figure 4.31. MS was the most significant factor, with an increasing trend

observed at every level. The highest MS level resulted in absolute increases of 9.5%, 33.6%, and

38.7% over the lowest level at 5 min, 15 min and 30 min, respectively. Mixing time did not have

a significant effect on oil release at 125 rpm or 150 rpm, but it was significant at the 175 MS

level. A MT increase from 5 min to 30 min resulted in the release of an additional 28.7% of the

total applied oil. That represents a SWA efficiency increase of 157% over the 5 min response.

The highest release of oil in untreated controls was at 175 rpm and 30 min; the lowest release

was at 125 rpm and 15 min.

In Aquaclean treatments, an increasing trend was again observed with MS. At the 5 min, 15

min, and 30 min MTs, the highest MS released an additional 26.1%, 22.9%, and 31.6% over the

lowest MS (Figure 4.32). These represent relative increases in SWA efficiency of 66.0%,

70.0%, and 128.0%, respectively. However, the greatest increase was found in going from 125

rpm to 150 rpm, with little additional benefit gained by increasing the speed to 175 rpm. MT

was significant only at the lowest rpm level, where oil release actually decreased with MT. One

91

explanation for this is that Aquaclean may disperse or emulsify the oil into the water column,

causing the oil to penetrate into the sand bed rather than lift off the surface. The extent of

dispersion and penetration may increase with MT, thereby causing the trend seen here. This

phenomenon was observed previously with Petroclean and PBC; oil quantified in the aqueous

phase was actually lower for some Petroclean treatments than for untreated controls. However,

in these experiments, Aquaclean performed better than the controls at all levels. The maximum

release of oil was seen at 175 rpm and 5 min MT, but no statistical difference was found across

MTs or MSs at the 150 rpm and 175 rpm levels. Since it is desirable to have low oil release from

controls and high release from treatments, a MS of 150 rpm is recommended; it gives the

greatest difference between controls and treatments. A MT greater than 5 min is not necessary.

4.4 OTHER CONSIDERATIONS FOR USE: DISPERSABILITY AND TOXICITY

SWAs are used on shorelines to release oil from substrates. Due to the biodiversity within near

shore waters, it is not desirable to disperse the oil into the water column where it may have a

toxic effect on fish or other native organisms. The oil that is released should recoalesce on the

surface of the receiving water and be collected through physical containment and skimming.

Thus, a good SWA should also be a poor dispersant.

The U.S. EPA developed the Baffled Flask Test (BFT) for determining the effectiveness of

chemical agents to disperse oil (Sorial et al., Venosa et al.). The SWAs used in these studies

were tested for their ability to disperse PBC oil under the BFT. Results of duplicate experiments

are shown in Table 4.1. SWAs that dispersed the most oil were Petrotech (73%) and Biosolve

(53%). Superall dispersed the least (4%).

92

In addition to effectiveness and dispersability, product toxicity must also be considered when

making a use/no-use decision. Older formulations were much more toxic than the new

generation of SWAs. Table 4.2 summarizes effectiveness, dispersability and toxicity data for our

six SWAs. Toxicity data were taken from the U.S. EPA Oil Program website and are reported as

ppm. Products in the table are listed in order of decreasing effectiveness based on performance

under the developed SWA testing protocol at 10:1 SOR. While Biosolve is the top performer in

terms of effectiveness, it is a relatively strong dispersant. Petroclean is equally as effective, with

lower rates for dispersability and toxicity. The decision to use one product over another will be

the decision of the On-Scene Coordinator and will depend on site conditions.

4.5 CONCLUSIONS

Protocol variables were tested at multiple levels to determine their effects on SWA performance.

The protocol was most sensitive to SOR and MS. SWA effectiveness increased with SOR for

ratios ranging from 1:2.5 to 10:1, and the relative ranking of SWAs changed as SOR increased.

Multiple oil volumes were tested at each SOR, but no trend was observed with changes in oil

volume. Diluting the SWA did not affect performance as long as the total mass of applied SWA

was constant.

Oil removals from controls and treatments increased with rotational MS. The maximum

difference between treatments and controls occurred at 150 rpm, while 200 rpm was excessively

rigorous, and 100 rpm yielded low release from treatments and controls. Differences in oil

release from treatments and controls were not strongly linked to mixing time.

93

Oil was applied to wet and dry sand and gravel to determine the effect of substrate type and

hydration. Statistical differences were found between sand and gravel substrates, but the relative

significance of all other variables was not affected by substrate type. Two substrates will be

included in the final protocol. DT had no effect. The decision to use dry substrates was made

based on high replicate RSD and high release of oil from controls in wet substrate treatments.

WT was set at 18 hr to reduce oil release from controls and maximize the difference between

treatments and controls. SWA performance was insensitive to CT.

SLC, Bunker C and IFO-180 oils were also tested. The lighter weight oil was easily released

from sand without SWA treatment and will not be used in the protocol. The heavy weight oil

was retained on the sand, with only 5% release of IFO-180 after 5 min WT for controls. IFO-

180 was insensitive to DT, WT and CT, but sensitive to MS and MT.

Based on these tests and the results of the fractional factorial experiment discussed in Chapter 6,

the following protocol conditions are recommended:

Substrate type: Sand and gravel

Substrate moisture: Dry substrate application

Oil type: PBC and IFO-180

Oil volume and application: 180 μl

Weathering time: 18 hr

SWA solution concentration: Manufacturer recommended

SOR: 10:1

Contact time: 5 min

Mixing speed: 150 rpm

Mixing time: 15 min

94

Figure 4.1 – Oiled Sand Before and After Washing for Control and Treatment.

Control Before Washing Treatment Before Washing

Control After Washing Treatment After Washing

Control and Treatment Wash Water

95

Figure 4.2 – Effect of Substrate Type on Aquaclean Effectiveness to Remove PBC from Dry and Wet Substrates

0

10

20

30

40

50

60

70

80

90

100

20/30Sand

FilProGravel

PeaGravel

Oil

Rem

ova

l, %

0

10

20

30

40

50

60

70

80

90

100

20/30Sand

FilProGravel

PeaGravel

Oil

Rem

ova

l, %

WET SUBSTRATESDRY SUBSTRATES

Figure 4.3 – Removal of PBC Oil from Dry and Wet Sand for Untreated Control and Three SWAs after 15 min Weathering.

0

50

100

Control SWA A SWA B SWA C

Oil

Rem

ove

d, %

Dry Sand

Wet Sand

Control Petroclean Aquaclean Corexit

96

Figure 4.4 – Effect of Substrate Drain Time on PBC Removal from Wet Sand after 15 min Weathering.

47.8 48.0 49.4

38.2

48.9

0

10

20

30

40

50

60

70

80

90

100

1 5 10 15 20

Drain Time, min

Oil

Rem

ova

l, %

Figure 4.5 – PBC Removal by SWAs Under Dry and Wet Sand Applications after 18 hr Weathering.

0

10

20

30

40

50

60

70

80

90

100

Control Corexit Petroclean Aquaclean Petrotech Superall

Oil

Rem

ova

l, %

Dry sand

Wet sand

(Recov.%, Std Dev.%)(22.3%, 2.5%)

(71.8%, 3.4%)

(41.2%, 2.6%)(42.1%, 0.7%)

(30.4%, 5.3%)

(69.0%, 7.6%)

(55.3%, 3.0%)

(68.8%, 3.4%)

(24.5%, 3.0%)

(41.0%, 2.4%)

(22.2%, 2.8%)

(4.1%, 0.8%)

97

Figure 4.6 – Comparison of Effectiveness Data for Two Operators Using Undiluted SWAs on Wet and Dry Sand at 2:1 SOR.

0

10

20

30

40

50

60

70

80

90

100

100% Aquaclean 100% Petroclean 100% Petrotech 100% Superall

Oil

Rem

ova

l, %

Operator 1 - Dry Sand

Operator 1 - Wet Sand



Figure 4.7 – Comparison of Effectiveness Data for Two Operators Using 50% Aqueous SWA Solutions on Wet and Dry Sand at 1:1 SOR.

0

10

20

30

40

50

60

70

80

90

100

50% Aquaclean 50% Petroclean 50% Petrotech 50% Superall

Oil

Rem

ova

l, %





98

Figure 4.8 – Effect of Oil Weathering Time on PBC Release from Wet Sand.

0

10

20

30

40

50

60

70

80

90

100

5 min 15 min 1 hr 6hr 18 hr

Weathering Time

Oil

Dis

trib

uti

on

, %

Wash Water

Sand

(Avg, Stdev)

(55.5, 10.4)

(52.7, 4.2)(55.8, 4.9)

(48.3, 5.3)

(75.5, 4.3)

(50.6, 13.5)(35.6, 1.4)

(39.8, 2.6)(43.3, 3.6)

(22.3, 2.4)

Figure 4.9 - Effect of Contact Time on Removal of PBC by Corexit and Petroclean.

0

10

20

30

40

50

60

70

80

90

100

5 min 15 min 30 min 1 hr 3 hr 6 hrContact Time

Oil

Rem

ove

d, %

CorexitPetroclean

(65.6%, 5.3%) (68.3%, 2.4%)

(73.4%, 2.0%)

(76.9%, 5.9%)

(68.1%, 3.3%)(65.2%, 2.2%)

(28.3%, 2.3%)(25.5%, 2.3%) (26.0%, 0.8%) (25.1%, 0.3%) (25.5%, 0.1%) (24.7%, 0.4%)

(Avg, Std Dev)

99

Figure 4.10 – Effect of SWA Dilution with Different Application Mass for Aquaclean.

0

10

20

30

40

50

60

70

80

90

100

Oil

Rem

ova

l, %

Dry sand application

Wet sand application

(Avg, Std Dev)

10% Aquaclean180 ul

100% Aquaclean180 ul

50% Aquaclean180 ul

0% (Control)

(3.2%, 0.7%)

(69.0%, 7.6%)

(61.5%, 8.1%)

(66.7%, 3.1%)

(43.1%, 2.9%)(40.1%, 3.1%)

(22.2%, 4.0%)

(17.8%, 2.8%)

Figure 4.11 – Effect of SWA Dilution with Different Application Mass for Petroclean

0

10

20

30

40

50

60

70

80

90

100

Oil

Rem

ova

l, %



(68.8%, 3.4%)

(25.3%, 2.5%)

(32.8%, 2.6%)

(28.0%, 2.8%)

(44.2%, 2.4%)

(29.7%, 3.6%)

(3.8%, 0.4%)

(12.2%, 1.1%)

10% Petroclean180 ul



0% (Control)

(Avg, Std Dev)

100

Figure 4.12 – Effect of SWA Dilution with Different Application Mass for Petrotech.

0

10

20

30

40

50

60

70

80

90

100

Oil

Rem

ova

l, %



10% Petrotech180 ul

100% Petrotech180 ul

50% Petrotech180 ul

0% (Control)

(2.7%, 0.3%)

(75.8%, 3.1%)

(41.2%, 2.6%)

(54.2%, 8.2%)(56.2%, 3.5%)

(15.4%, 1.4%)

(44.1%, 9.6%)

(17.4%, 1.0%)

(Avg, Std Dev)

Figure 4.13 – Effect of SWA Dilution with Different Application Mass for Superall.

0

10

20

30

40

50

60

70

80

90

100

Oil

Rem

ova

l, %



10% Superall180 ul

100% Superall180 ul

50% Superall180 ul

0% (Control)

(27.9%, 3.3%)

(21.2%, 0.5%)(22.3%, 2.5%)(24.8%, 1.2%)

(47.7%, 1.1%)(43.8%, 1.6%)

(5.2%, 0.6%) (16.8%, 1.5%)

(Avg, Std Dev)

101

Figure 4.14 – Effect of SWA Dilution with Same Application Mass for Aquaclean.

0

10

20

30

40

50

60

70

80

90

100

180 ul 100% SWA 600 ul 30% SWA 1800 ul 10% SWA 6000 ul 3% SWA

Oil

Rem

ova

l, %

Aquaclean

Controls(Avg, Std Dev)

(69.0%, 7.6%)

(66.9%, 4.4%)(64.7%, 4.8%)

(66.8%, 4.6%)

(24.6%, 5.6%)

(14.9%, 0.01%)(16.1%, 1.6%) (17.4%, 1.1%)

Figure 4.15 – Effect of SWA Dilution with Same Application Mass for Petroclean.

0

10

20

30

40

50

60

70

80

90

100


Oil

Rem

ova

l, %

Petroclean

Controls

(68.8%, 3.4%)(64.1%, 4.6%) (63.5%, 6.8%) (63.7%, 5.6%)

(23.6%, 3.4%)(20.6%, 1.8%) (22.8%, 1.7%)

(26.7%, 1.4%)

(Avg, Std Dev)

102

Figure 4.16 – Effect of SWA Dilution with Same Application Mass for Petrotech.

0

10

20

30

40

50

60

70

80

90

100


Oil

Rem

ova

l, %

Petrotech

Controls

(Avg, Std Dev)

(67.8%, 3.8%)

(61.3%, 3.4%)(58.2%, 1.6%)

(63.0%, 5.4%)

(19.6%, 0.6%) (18.9%, 2.2%)

(25.3%, 0.9%)

(17.1%, 1.8%)

Figure 4.17 – Effect of SWA Dilution with Same Application Mass for Superall.

0

10

20

30

40

50

60

70

80

90

100


SWA Application

Oil

Rem

ova

l, %

Superall

Controls

(43.8%, 1.6%)

(Avg, Std Dev) (48.4%, 4.1%) (47.7%, 5.4%)

(43.6%, 4.0%)

(25.6%, 1.7%)

(21.3%, 1.4%)

(8.7%, 0.06%) (13.6%, 0.6%)

103

Figure 4.18 – Effect of Oil Volume on SWA Effectiveness at Four SOR. 180 µl SWA Applied to Oil on Wet Sand.

0

10

20

30

40

50

60

70

80

90

100

SOR 2:1 (90 ul PBC) SOR 1:1 (180 ul PBC) SOR 1:2 (360 ul PBC) SOR 1:2.5 (450 ul PBC)

Oil

Rem

ova

l, %

AquacleanPetrotechPetrocleanSuperallControl

104

Figure 4.19 – Effect of SOR on Aquaclean Effectiveness at Four Oil Volumes

0

10

20

30

40

50

60

70

80

90

100

2:1 1:1 1:2

SOR

Oil

Rem

ova

l, %

90 uL oil

180 uL oil

360 uL oil

450 uL oil

Figure 4.20 - Effect of Oil Volume on Aquaclean Effectiveness at Three SOR.

0

10

20

30

40

50

60

70

80

90

100

90 uL oil 180 uL oil 360 uL oil 450 uL oil

Oil Volume

Oil

Rem

ova

l, %

2:11:11:2

105

Figure 4.21 – Effect of SOR on Petrotech Effectiveness at Four Oil Volumes.

0

10

20

30

40

50

60

70

80

90

100

2:1 1:1 1:2

SOR

Oil

Rem

ova

l, %

90 uL oil

180 uL oil

360 uL oil

450 uL oil

Figure 4.22 - Effect of Oil Volume on Petrotech Effectiveness at Three SOR.

0

10

20

30

40

50

60

70

80

90

100

90 uL oil 180 uL oil 360 uL oil 450 uL oil

Oil Volume

Oil

Rem

ova

l, %

2:11:11:2

106

Figure 4.23 – Effect of SOR on Oil Removal from Dry Sand.

0

20

40

60

80

100

120

1:1 SOR 2:1 SOR 4:1 SOR 10:1 SOR

Oil

Rem

ova

l, %

Aquaclean

Biosolve

Corexit

Petroclean

Petrotech

Superall

107

Figure 4.24 – Effect of Mixing Speed and Mixing Time on Removal of PBC from Wet Sand in Untreated Controls after 18 Hours Weathering.

5 min MT10 min MT

20 min MT

100 rpm

150 rpm

200 rpm

0

10

20

30

40

50

60

70

80

90

100

Oil

Rem

ova

l, %

Figure 4.25 – Effect of Mixing Speed and Mixing Time on Removal of PBC from Wet Sand in Corexit Treatments after 18 Hours Weathering.

5 min MT10 min MT

20 min MT

100 rpm

150 rpm

200 rpm

0

10

20

30

40

50

60

70

80

90

100

Oil

Rem

ova

l, %

108

Figure 4.26 – Effect of Drain time on Release of Bunker C and South Louisiana Crude Oils from Wet Sand after 18 Hours Weathering.

0

10

20

30

40

50

60

70

80

90

100


Drain Time

Oil

Rem

ova

l, %

Bunker C

South Louisiana

Figure 4.27 - Effectiveness of Undiluted Aquaclean to Remove PBC, IFO-180 and Bunker C from Wet Sand after 18 Hours Weathering.

0

10

20

30

40

50

60

70

80

90

100

PBC IFO-180 Bunker C

Oil

Rem

ova

l, %

Controls

100% Aquaclean at 2:1

109

Figure 4.28 – Comparison of Weathering Times for Untreated IFO-180 Controls.

0

20

40

60

80

100

120

5min 15min 1hr 6hr 18hr

Weathering Time

Oil

Rem

ova

l, %

Aqueous Extract

Sand Extract

Figure 4.29 – Comparison of Weathering Times for IFO-180 Treated with Undiluted Aquaclean at 2:1 SOR.

0

10

20

30

40

50

60

70

80

90

100

5min 1hr 6hr 18hr

Weathering Time

Oil

Rem

ova

l, %

Aqueous Extract

Sand Extract

110

Figure 4.30 – Effect of Contact Time on Release of IFO-180 from Wet Sand by Aquaclean and Petroclean.

0

10

20

30

40

50

60

70

80

90

100

5min 15min 30min 1hr 3hr 6hr


Oil

Rem

ova

l, %

Aquaclean

Petroclean

111

Figure 4.31 – Effect of Mixing Speed and Mixing Time on Removal of IFO-180 from Wet Sand in Untreated Controls after 18 Hours Weathering.

5 min15 min

30 min

125

150

175

0

10

20

30

40

50

60

70

80

90

100

Oil

Rem

ova

l, %

Figure 4.32 - Effect of Mixing Speed and Mixing Time on Removal of IFO-180 from Wet Sand in Aquaclean Treatments after 18 Hours Weathering.

5 min15 min

30 min

125

150

175

0

10

20

30

40

50

60

70

80

90

100

Oil

Rem

ova

l, %

112

Table 4.1 – Results of Baffled Flask Test for Dispersability.

Extract Oil Extract Mass in extract Total dispersed Total PercentOil SWA 340nm 370nm 400nm Dilution Area Conc. volume Vol. (30 ml aq sample) (120 ml aq sample) Dispersed

Factor mg/ml added, ml ml mg mg %Aquaclean 0.8035 0.4441 0.2941 1 29.8 0.2459 0.10 20 4.92 19.67 22.01Biosolve 0.9561 0.5312 0.3521 2 71.1 0.5872 0.10 20 11.74 46.97 52.54Corexit 0.6224 0.3450 0.2282 2 46.2 0.3816 0.10 20 7.63 30.52 34.14Superall 0.1577 0.0870 0.0576 1 5.8 0.0482 0.10 20 0.96 3.86 4.32

Petrotech 1.4438 0.7987 0.5282 2 107.1 0.8841 0.10 20 17.68 70.73 79.11Petroclean 0.7595 0.4199 0.2776 1 28.2 0.2324 0.10 20 4.65 18.59 20.80

PBC

Extract Oil Extract Mass in extract Total dispersed Total PercentOil SWA 340nm 370nm 400nm Dilution Area Conc. volume Vol. (30 ml aq sample) (120 ml aq sample) Dispersed

Factor mg/ml added, ml ml mg mg %Aquaclean 1.1096 0.6138 0.4061 1 41.2 0.3397 0.10 20 6.79 27.18 30.40Biosolve 0.9824 0.5437 0.3592 2 72.9 0.6016 0.10 20 12.03 48.13 53.84Corexit 0.6353 0.3545 0.2349 2 47.4 0.3911 0.10 20 7.82 31.29 35.00Superall 0.1429 0.0795 0.0531 1 5.3 0.0440 0.10 20 0.88 3.52 3.94

Petrotech 1.5663 0.8673 0.5730 2 116.2 0.9595 0.10 20 19.19 76.76 85.86Petroclean 0.6813 0.3754 0.2470 1 25.2 0.2079 0.10 20 4.16 16.64 18.61

PBC

Table 4.2 – Effectiveness, Dispersability and Toxicity for Five SWAs.

Effectiveness DispersabilitySWA 10:1 SOR (%) BFT Result (%) Menidia (96-hr) Mysidopsis (48-hr)Biosolve 109.78 53.19 7.40 1.30Petroclean 100.61 19.70 115.00 105.00Aquaclean 92.33 26.20 6.50 2.10Superall 74.15 4.13 4.60 5.00Petrotech 73.22 82.49 3.40 1.00Corexit 45.73 34.57 13.20 9.06

Toxicity (ppm)

113

Chapter 5

ANALYTICAL METHODOLOGIES: COMPARISONS AND LIMITATIONS

The measurement of oil concentration in DCM extracts was done according to the protocol

outlined in the U.S. EPA Baffled Flask Test (BFT) for testing oil spill dispersants (Sorial et al.,

2004a, b). The method entails using a UV/visible spectrophotometer to measure absorbance

values at three specified wavelengths. Oil concentration is quantified using a six-point standard

calibration curve plotted as concentration versus area under the absorbance curve. The point of

variation from this protocol is in the preparation of standards. The dispersant protocol is a two-

phase (oil-water) system in which dispersant is added to an oil slick floating on the water

surface. After dispersion, a single sample is removed from the vessel and extracted to quantify

the dispersed oil. The dispersant and oil are added at a specified ratio (1:25), and every sample is

assumed to contain dispersant and oil at this ratio. Standards are also prepared at this ratio, so

the effect of the dispersant on analysis is factored out. Because the volume of dispersant is

small, any absorbance in the wavelengths of interest due to the dispersant is negligible.

In the SWA protocol, we have three phases to consider: oil, water, and substrate. The ratio of

SWA to oil in any given experiment is known, but the partitioning of SWA molecules between

the aqueous and solid phases is not. SWAs contain surfactants that will align at the oil-water

interface. A good SWA will release most of the oil into the aqueous phase; a poor SWA will

leave the oil (and some surfactant) adhered to the substrate. When the substrate is removed from

the wash water, the relative amount of SWA in each phase is unknown. For water soluble

surfactants, the bulk of SWA may remain with the water regardless of whether the oil is

114

predominantly associated with one phase or the other. However, this will depend on the

hydrophilic-lipophilic balance of the surface active agents. Since the SWA formulations are

confidential, it is not possible to predict exactly how the SWA will partition. Thus, SWA

standards cannot be prepared at a specified SWA-oil ratio that would be appropriate for all

samples. The best solution is to use oil-only standards without SWA for analysis of all samples,

but this may lead to overestimation for samples treated with highly colored SWAs. Because

SWAs in this protocol are applied at a much higher rate than dispersants, the analytical error

incurred by this approach could be significant.

Numerous steps were taken to evaluate this problem. First, the DCM extractable fraction of each

SWA was analyzed spectrophotometrically to determine whether absorption would occur in the

wavelength range of interest. Extractions were done with and without oil, and the extracts were

measured at 340, 370 and 400 nm. Second, other analytical methods were considered that would

not have the same colorimetric bias. These include 1) gas chromatorgraphy/mass spectrometry

(GC/MS) to quantify individual alkane and PAH constituents, 2) spectrofluorometry to quantify

oil based on its fluorescence, and 3) thin layer chromatography by Iatroscan to separate and

quantify the total alkane, PAH, resin and asphaltene fractions.

5.1 SPECTROPHOTOMETRY

5.1.1 Theory

Spectrophotometry is the measure of light intensity as a function of wavelength. A

spectrophotometer measures quantitatively the fraction of light that passes through a given

solution. Deuterium and tungsten lamps transmit UV and visible light, respectively, through a

115

monochromator, which picks light of one particular wavelength out of the continuous spectrum.

The intensity (I) of light passing through a sample is measured by a photodiode and compared to

the intensity of a reference beam (Io). The transmittance at each wavelength is calculated as the

ratio I/Io. Transmittance is converted to absorbance (A) by the following equation:

A=log(Io/I)

The instrument used for this work was an Agilent 8453 UV/Visible Spectrophotometer with

diode array detector. Absorbance was measured for all wavelengths and recorded by computer.

Oil in DCM calibration standards were prepared and analyzed with each sample set. Absorbance

measurements at wavelengths 340, 370, and 400 nm were used to determine the area under the

absorbance curve between wavelengths 340 and 400. Samples were quantified against the area

vs. concentration calibration curve.

5.1.2 Analytical Biases and Interferences

5.1.2.1 Absorbance Interference Due to SWAs

For this protocol, absorbance was measured at wavelengths of 340, 370, and 400 nm. These

wavelengths are in the near UV range, with 400 nm being on the edge of the visible light range.

Some SWA may absorb within the wavelength range of interest. To determine potential

interferences caused by the products, each SWA was added to a separatory funnel at a final

concentration of 500 µl SWA concentrate in 400 ml seawater. This level is higher than used

experimentally. These SWA solutions were extracted with DCM, and extracts were measured

spectrophotometrically. The area under the absorbance curve between 340 and 400 nm was used

to calculate the equivalent oil recovery represented by that area. At the time of this testing, a

total volume of 50 µl PBC was used for SWA experiments. Thus, the percent interference is

116

based on a 50 µl oil volume. Control samples containing 400 ml seawater with no SWA were

also extracted into 50 ml DCM to determine interferences due to water extraction. Pure DCM

was used to zero the baseline and absorbance values for pure DCM (no aqueous extraction) were

included for reference. The results are shown in Table 5.1. The interferences for all SWA were

below 1% and thus were not significant for this work.

These experiments were repeated with the addition of 50 µl PBC to each separatory funnel to

determine if the presence of oil would enhance the solubility of SWA in the DCM extract.

Recovery of oil was calculated based on the oil added. The control was PBC in water with no

SWA. Mass balance recoveries are shown in Table 5.2. No significant interferences were

observed for any SWA, with recoveries ranging from 94% to 95.5%. The recovery for the

control was 90.7%, which is lower than the SWA treatments. However, since the bias was of

similar magnitude for all SWA relative to the control (3%-5%), this is not a concern. Biosolve,

which is strongly magenta colored, tended to produce high mass balance recoveries (110% to

120%) in protocol experiments. However, Biosolve resulted in an increase of only 1.5% over the

other SWA treatments. Thus it can be concluded that there were no significant

spectrophotometric interferences that would bias the protocol against or in favor of any one

product.

5.1.2.2 Water in DCM Emulsions

Water is slightly miscible with DCM and can interfere with spectrophotometric analysis. The

surfactants contained in SWA products can also promote water in solvent emulsions during the

extraction process. Control and sample extracts are typically clear, but when an emulsion forms,

117

the extract will appear slightly cloudy. These emulsions will typically break when the extract is

cooled during refrigerated storage, and small droplets of water will appear on the surface of the

extract. When an emulsified sample is placed into a cuvette, the water will tend to stick to the

surface of the cuvette and alter the passage of light through the sample.

To determine the effect of water in DCM emulsions on spectrophotometric results, several

samples were split and analyzed 1) directly and 2) after filtering through sodium sulfate to

remove the water. Control samples were visibly clear, while treatment samples were slightly

emulsified. The absorbance values and measured masses of oil in a typical sample and control

before and after filtering are shown in Table 5.3. The unfiltered extract was 3.6% and 4.6%

higher than the filtered extract for the sample and control, respectively. The bias caused by not

filtering the sample was higher for the non-emulsified control than for the emulsified sample.

Thus, this was determined to be an insignificant source of error.

5.2 GAS CHROMATOGRAPHY/MASS SPECTROMETRY (GC/MS)

5.2.1 Theory

GC/MS combines gas phase chromatography, which separates compounds based on their

physical properties, and mass spectrometry, which ionizes the molecules into fragments that are

quantified based on their mass to charge (M/Z) ratio. As the individual compounds elute from the

GC column, they are bombarded with a stream of electrons causing them to break apart into

fragments. These fragments are charged ions, typically with a charge of +1. Thus, the M/Z

usually represents the molecular weight of the fragment. The quadrupole contains 4

electromagnets that focus each of the fragments and allow only one M/Z at a time to pass into the

118

detector. This signal intensity or abundance of each M/Z detected during every scan is recorded

as the mass spectrum. For this work, a DB-5 capillary GC column was used to measure non-

polar oil hydrocarbons. Separation was achieved based on the boiling point of the hydrocarbons.

5.2.2 Bias

Only the alkane and PAH components of oil can be quantified by GC/MS. DCM extracts were

solvent exchanged into hexane prior to injection by GC/MS to precipitate heavy asphaltenes

from the samples. Asphaltenes are heavy residues in oil that are not resolvable by GC/MS.

Injection of this fraction into the GC/MS would dirty the injection port and clog the column.

However, it is the asphaltene fraction of oil that is least volatile and most difficult to remove

from a substrate. Biases may occur if lighter weight and more water-soluble compounds are

preferentially removed from surfaces before this fraction, or if the oil is highly weathered.

5.2.3 Results

Sample extracts were exchanged into hexane and analyzed by GC/MS. Data were then

compared to spectrophotometric results for the same samples. The concentrations of 22 alkane

and 14 PAH compounds were quantified and summed to give total alkane and total PAH values.

Alkane analytes included C16 through C35 straight chain alkanes plus pristine and phytane.

PAHs considered were naphthalene and its C1- to C4-alkylated congeners, phenanthrene and its

C1- to C4-alkylated congeners, and dibenzothiophene and its C1- to C3-alkylated congeners.

Concentrations were multiplied by extract volumes to give total mass extracted from aqueous

and substrate samples. The total extracted alkane and total extracted PAH masses were added to

give a total oil mass for each sample.

119

The mass of these alkane and PAH compounds in the volume of oil applied experimentally (180

μl) was determined by analyzing a standard containing 180 μl PBC in hexane. The sum of the

alkane and PAH compounds considered were 6.5 mg and 3.1 mg, respectively. This represents

approximately 6% of the total mass of oil applied experimentally. Thus, the measured masses of

these alkanes and PAHs by GC/MS were more than an order of magnitude lower than the mass

of total oil measured by spectrophotometer. The mass balance recoveries calculated based on the

combined recovery from aqueous and substrate extracts was in the range of 73% to 78%.

The total mass of each group (alkanes, PAHs and total oil) measured by GC/MS are plotted

against the mass of total oil measured by spectrophotometer in Figure 5.1. Alkane, PAH and

total oil results by GC/MS were highly correlated to spectrophotometer results (R2 > 0.98 for

all). Mass data for each class of compound were converted to percent recovery based on the mass

of oil added experimentally. The percent recovery of each class is plotted against the percent

recovery measured by spectrophotometer and the data are highly correlated (R2 > 0.98 for all)

(Figure 5.2). When percent PAH recoveries by GC/MS are plotted against the corresponding

percent alkanes by GC/MS, the data follow a linear trend (R2=0.996) with a slope equal to 1.0

(Figure 5.3). This suggests that both classes of compounds are released from the surface at the

same rate.

5.3 SPECTROFLUOROMETRY

5.3.1 Theory

Spectrofluorometry (or fluorescence spectroscopy) is a type of electromagnetic spectroscopy that

measures fluorescence from a sample. A beam of light, usually ultraviolet light, excites the

120

electrons in molecules of certain compounds causing them to emit light of a lower energy,

usually visible light. To observe this effect, spectrofluorometers have dispersive optics, usually

grating monochromators, both between the source and the sample for the incident light and

between the sample and the detector for the exitent light. This permits the identification of the

wavelengths that excite the sample and the wavelengths at which that excitation generates the

fluorescence. In oil, it is the C-H stretch of aromatic compounds that causes fluorescence, so the

fluorometric signal is a measure of the level of PAHs in the sample. The instrument used for

analysis was a Shimadzu RF-5301PC Spectrofluorophotometer with a blazed holographic

grating, photomultiplier and digital signal processing.

5.3.2 Bias

Spectrofluorometry is a very sensitive analytical tool. Due to its high sensitivity, experimental

samples had to be diluted 1:40 to avoid self-quenching and to fall within the linear calibrated

range of the instrument. Analytical errors in measurement or sample preparation become

magnified when high dilution factors are used. Additionally, this method is not appropriate for

oils with low fluorescence.

5.3.3 Results

Thirty six samples were measured by UV/visible spectrophotometry and by spectrofluorometry.

The measured percent oil recoveries for samples under each method are displayed in Table 5.4

along with the relative percent difference (RPD) between the two measured values. Thirty four

of the 36 samples had an RPD less than 14%, indicating that there was excellent agreement

between the two methods. The two samples with high RSD had very low (<2 %) values for oil

121

recovery. These values are below the reporting limit for the instruments, which accounts for the

high variability. In most cases (28 of 36 samples), the spectrofluorometer gave a higher result

than the UV/vis spectrophotometer. Because there was good agreement between the analytical

methods, there was no advantage to using the spectrofluorometer for this work.

5.4 THIN LAYER CHROMATOGRAPHY

5.4.1 Theory

An Iatroscan MK6 was used to separate the DCM extracts into alkane, PAH, resin, and

asphaltene fractions. The Iatroscan instrument was developed for the analysis of organic

substances that show no UV-absorption and no fluorescence and are difficult to resolve by GC.

It combines the techniques of thin layer chromatography (TLC) for the separation of organic

compounds with a flame ionization detector (FID). It uses Chromarods, quartz rods coated with

a thin layer of silica or alumina, to develop and separate the sample. Spotting of the sample on

the chromarod is done using a full automatic sample spotter. The chromarods are developed in a

separate development tank for separation of the fractions, and then introduced into the Iatroscan

for analysis. As the chromarod is advanced at a constant speed through the hydrogen flame of

the FID, the substances are ionized through energy obtained from the hydrogen flame. An

electric field applied to the poles of the FID cause the ions to generate electric current with

intensity proportional to the amount of each organic substance entering the flame. The collector,

which is placed above the flame generates an analogue signal, which is evaluated with a PC.

5.4.2 Bias

The composition of PBC oil is 78.3% aliphatic hydrocarbons, 17.6% aromatic hydrocarbons,

2.5% polars, 2.0% asphaltenes, and 0.6% resins (Clayton, 1993). The low concentrations of

122

asphaltenes and resins may make these inaccurate measures of overall oil concentration.

Additionally, asphaltenes have low solubility in most solvents. While they are soluble in DCM,

the extraction efficiency for these heavy weight residues may be lower than for lighter weight

fractions.

5.4.3 Results

DCM extracts from two experiments were analyzed by Iatroscan and Spectrophotometer. The

alkane, PAH, resin, and asphaltene fractions of the PBC extracts were separated by thin layer

chromatography and quantified using Iatroscan. The percent recovery of each of these oil

components was plotted against the total oil recovery measured by spectrophotometer for the

same extracts. The data are shown in Figure 5.5. The spectrophotometer recovery is represented

by the line, and component recoveries are depicted by symbols. Good correlation was found

between the two analytical methods for alkanes, PAHs, and resins (depicted with open symbols);

very little correlation was found for asphaltenes (filled circles). The average absolute deviations

from the spectrophotometer results were 3.2 ± 1.5% for alkanes, 7.2 ± 4.2% for PAHs, 3.5 ±

3.4% for resins, and 24.6 ± 14.7% for asphaltenes. Separate plots of the individual components

against the spectrophotometer results are shown in Figures 5.6 through 5.9. Correlations were

0.978 for alkanes, 0.901 for PAHs, 0.957 for resins, and 0.261 for asphaltenes. Thus, the

Iatroscan and spectrophotometer produce results that are most similar for alkanes and resins,

with little correlation for asphaltenes. When alkane results are plotted against PAH results, the

regression line through these data has a slope less than 1 (Figure 5.9). This suggests that when

little oil is released from a surface by treatment with an SWA, PAHs are preferentially removed,

while with high release of oil, alkane concentrations are higher.

123

5.5 CONCLUSIONS

Concerns over possible analytical interferences caused by the SWA products led to testing of

other analytical methods for their potential use in the protocol. Extracts generated using the

developed protocol were analyzed using UV/vis spectrophotometer, GC/MS, spectrofluorometer,

and TLC/FID (Iatroscan). GC/MS percent recoveries for select alkane and PAH constituents

were found to be well correlated to spectrophotometric results. However, these compounds

represent a small fraction of the total oil mass, and biases could result from not considering the

heavier fractions and residues associated with highly weathered oil. Spectrofluorometry yielded

similar results to spectrophotometry, but the required dilution factor was another source of

potential error. Results obtained by Iatroscan for alkane, PAH, and resin constituents were highly

correlated with spectrophotometer results. However, asphaltenes were found to be an inaccurate

measure of total oil concentration in this study. In general, GC/MS, spectrofluorometry, and thin

layer chromatography were more labor intensive and offered no benefit over spectrophotometry.

Due to biases associated with some methods and good correlations with other methods, oil

analyses for the protocol will be done by spectrophotometry.

124

Figure 5.1 – Mass Alkane or PAH by GC/MS.vs. Mass of Oil by Spetrophotometer.

y = 0.0358x - 0.0317

R2 = 0.9961

y = 0.0124x - 0.0756R2 = 0.9829

y = 0.0482x - 0.1073

R2 = 0.9947

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0

Mass Oil (mg), Spectrophotometer

Mas

s A

lkan

e o

r P

AH

(m

g),

GC

/MS

Alkane GC/MS

PAH GC/MS

Total Oil GC/MS

Linear (Alkane GC/MS)

Linear (PAH GC/MS)

Linear (Total Oil GC/MS)

Figure 5.2 –Percent Recovery of Alkanes and PAHs by GC/MS vs. Percent Recovery of Oil by Spectrophotometer.

y = 1.013x - 0.6481

R2 = 0.9983

y = 1.0829x - 4.1461

R2 = 0.9864

y = 1.0296x - 1.4795

R2 = 0.997

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0

Percent Oil (%), Spectrophotometer

Per

cen

t A

lkan

e o

r P

AH

(%

), G

C/M

S

Alkane GC/MS

PAH GC/MS

Total Oil GC/MS

Linear (Alkane GC/MS)

Linear (PAH GC/MS)

Linear (Total Oil GC/MS)

Alkanes

PAHs

Total Oil

125

Figure 5.3 – Percent PAH Recovery vs. Percent Alkane Recovery by GC/MS.

y = 1.0429x - 1.7887R2 = 0.9928

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0

Percent Alkane (%), GC/MS

Per

cen

t P

AH

(%

), G

C/M

S

Figure 5.4 - Percent Recovery of Alkanes, PAHs, Resins, and Asphaltenes by Iatroscan versus Percent Recovery of Total Oil by Spectrophotometer.

0

10

20

30

40

50

60

70

80

90

100

0 10 20 30 40 50 60 70 80 90 100

% Oil Recovery by Spectrophotometry

% C

on

stit

uen

t R

eco

very

by

Iatr

osc

an

Alkanes

PAHs

Resins

Asphaltenes

Spectrophotometer

126

Figure 5.5 - Percent Recovery of Alkanes by Iatroscan versus Percent Recovery of Total Oil by Spectrophotometer.

y = 0.9361x + 3.193

R2 = 0.9784

0

10

20

30

40

50

60

70

80

90

100

0 10 20 30 40 50 60 70 80 90 100


% C

on

stit

uen

t R

eco

very

by

Iatr

osc

an

Alkanes

Figure 5.6 - Percent Recovery of PAHs by Iatroscan versus Percent Recovery of Total Oil by Spectrophotometer.

y = 0.7378x + 13.112

R2 = 0.901

0

10

20

30

40

50

60

70

80

90

100

0 10 20 30 40 50 60 70 80 90 100


% C

on

stit

uen

t R

eco

very

by

Iatr

osc

an

PAHs

127

Figure5.7 - Percent Recovery of Resins by Iatroscan versus Percent Recovery of Total Oil by Spectrophotometer.

y = 0.9687x + 1.5647

R2 = 0.9568

0

10

20

30

40

50

60

70

80

90

100

0 10 20 30 40 50 60 70 80 90 100


% C

on

stitu

ent

Rec

ove

ry b

y Ia

tro

scan

Resins

Figure5.8 - Percent Recovery of Asphaltenes by Iatroscan versus Percent Recovery of Total Oil by Spectrophotometer.

y = 0.7136x + 14.321

R2 = 0.2606

0

10

20

30

40

50

60

70

80

90

100

0 10 20 30 40 50 60 70 80 90 100


% C

on

stit

uen

t R

eco

very

by

Iatr

osc

an

Asphaltenes

128

Figure 5.9 – Percent Recovery of Alkanes by Iatroscan vs. Percent Recovery of PAHs by Iatroscan.

y = 0.7932x + 10.341

R2 = 0.9329

0

10

20

30

40

50

60

70

80

90

100

0 10 20 30 40 50 60 70 80 90 100

% alkanes, iatroscan

% P

AH

s, ia

tro

scan

129

Table 5.1 – Interferences in Spectrophotometric Analysis Caused by 500 µl SWAs.

Sample Average Interference, % DCM blank -0.02 Control (water only) 0.21 Aquaclean 0.33 Biosolve 0.54 Corexit 0.85 Petroclean 0.73 Petrotech 0.52

Table 5.2 – Recovery of 50 µl PBC from Seawater in the Presence of 500 µl SWA.

Sample PBC Recovery, % Control (no SWA) 90.7 Aquaclean + PBC 94.2 Biosolve + PBC 95.5 Corexit + PBC 94.1 Petroclean + PBC 94.0 Petrotech + PBC 95.0

Table 5.3 - Absorbance Values and Measured Masses of Oil in a Typical Sample and Control Before and After Filtering with Sodium Sulfate.

Extract Extract Mass in Conc. Vol. extract Percent

340nm 370nm 400nm mg/ml ml mg differencePBC sample, unfiltered 1.984 1.119 0.758 74.70 0.614 20 12.28PBC sample, filtered through sodium sulfate 1.940 1.076 0.708 72.00 0.592 20 11.84 3.6PBC standard, unfiltered 0.794 0.440 0.284 29.37 0.241 20 4.83PBC standard, filtered through sodium sulfate 0.772 0.416 0.264 28.02 0.230 20 4.61 4.6

ABSORBANCEArea

130

Table 5.4 – Comparison of Analytical Results Measured by UV/Visible Spectrophotometry and Spectrofluorometry.

UV/Vis SpectrofluormeterSample 3 42.5 38.9 -8.8Sample 5 50.0 47.2 -5.7Sample 29 58.9 56.6 -4.0Sample 4 39.0 37.0 -5.4Sample 6 32.3 30.3 -6.2Sample 35 36.2 35.7 -1.3Sample 31 98.3 97.9 -0.4Sample 8 16.3 16.0 -2.1Sample 13 1.2 1.6 28.1Sample 14 1.6 2.3 36.7Sample 36 30.3 31.1 2.4Sample 16 47.7 49.3 3.3Sample 10 23.7 25.4 6.6Sample 34 47.3 49.0 3.6Sample 1 17.4 19.2 9.8Sample 24 48.0 50.1 4.2Sample 2 25.5 27.7 8.1Sample 23 38.4 40.8 6.1Sample 21 46.8 49.8 6.4Sample 7 26.7 29.9 11.4Sample 9 28.2 31.4 10.8Sample 12 27.6 30.8 11.1Sample 11 28.9 32.4 11.3Sample 18 54.3 58.5 7.4Sample 27 58.9 63.3 7.2Sample 26 80.9 85.5 5.5Sample 19 51.8 56.5 8.8Sample 33 49.5 54.5 9.7Sample 20 46.0 51.1 10.5Sample 22 43.3 48.5 11.3Sample 25 71.7 77.2 7.4Sample 15 44.5 50.4 12.4Sample 17 53.9 61.8 13.8Sample 28 59.9 67.9 12.6Sample 30 67.0 75.2 11.4Sample 32 97.8 111.9 13.5

Measured Oil in Extract, %RPD

131

Chapter 6

TESTING OF VARIABLES: THE FRACTIONAL FACTORIAL EXPERIMENT

Fractional factorial designs are experimental designs consisting of a carefully chosen subset of

the experimental runs of a full factorial design. The subset is chosen to exploit the sparsity-of-

effects principle, which states that a system is usually dominated by main effects and low-order

interactions (Box et al., 2005). It is most likely that main (single factor) effects and two-factor

interactions are the most significant responses; higher order interactions such as three-factor

interactions are very rare. The goal is to obtain information about the most important factors

while using a fraction of the effort of a full factorial design (Box et al., 2005; Wu and Hamada,

2000).

For this work, a fractional factorial experiment was designed to determine the most important

factors that contribute to the performance of SWAs in removing crude oil from surfaces under

the developed protocol. The effective levels of each factor were determined. Main effects and

two-way interactions were evaluated for significance.

Based on results of exploratory experiments as discussed in previous chapters, the following

conditions were fixed at the levels indicated:

substrate hydration: none (dry substrate)

oil volume: 180 uL

oil application pattern: nine 20 uL drops

weathering time: 18 hours

132

SWA application pattern: nine drops

seawater volume: 100 mL

oil type: PBC

SWA type: Aquaclean

temperature: room temp

The following variables were considered at multiple levels:

substrate: sand, gravel

SWA solution concentration: 100%, 50%, 25%

SOR: 1:2, 1:1, 2:1

oil-SWA contact time: 15 min, 30 min, 45 min

mixing speed: 125, 150, 175 rpm

mixing time: 5, 15, 30 min

6.1 DESIGN AND METHODS

A 35-1 fractional factorial design with 5 factors and 3 levels was utilized. This resulted in a 3-1 or

1/3 fraction of the full 35 factorial design. The resolution, which is the ability to separate main

effects and low-order interactions from one another, was set at 5. This allowed for estimation of

main effects and two-factor interaction effects unconfounded by higher-order interactions; three-

factor interactions may be confounded by other three-factor interactions. The five factors

considered were mixing speed (MS), mixing time (MT), oil-SWA contact time (CT), SWA

solution concentration (SC), and SWA-to-oil (v/v) ratio (SOR). The three levels for each factor

are listed in the preceding section. Substrate was considered as a 2 level block effect, and the

133

assumption was made that the two substrates were randomly selected from all types of substrates

and that our conclusions based on the averaged responses apply to any randomly picked

substrate. The response variable of interest (hereafter referred to as ‘response’) was the oil mass

(mg) in the aqueous extract. This represents the fraction of oil released from the sand or gravel

and is a direct measure of SWA effectiveness.

SAS Proc GLM was used to perform the statistical analyses, with terms for substrate, each of the

5 variables, and all 2 way interactions between the 5 variables. The statistical results are

presented in Tables 6.1-6.11 and in Figures 6.1-6.10. Because significance was found between

the two substrates, analyses were also performed for both substrates separately to confirm the

conclusions found in the combined model. The results of the separate analyses yielded the same

conclusions as the combined model, so the results presented in Tables 1-11 are from the

combined analysis.

6.2 RESULTS

6.2.1 Main Effects

The means and standard deviations for all single-factors at each level are presented in Table 6.1.

The mean responses represent the mass (mg) of oil released into the wash water and are a direct

measure of SWA effectiveness. P-values denoting significant differences among the three levels

for each factor are also given. All factors were found to be significant with the exception of SC

(p = 0.69). The mean responses at 25%, 50%, and 100% solution concentrations were 103.6,

102.0 and 102.9, respectively. This suggests that dilution of the SWA product does not affect

performance as long as the total mass of active product applied is constant. In other words,

134

dilution is not significant within a given SOR. This is in agreement with experimental findings

presented in Chapter 4.

All other factors were found to significantly affect response across the levels. However, ceiling

effects were observed for MT and CT with no additional benefit provided by the highest level

over the middle level for these factors. Thus, the middle level can be considered to be the

maximum value needed to achieve the best response. For MT, this value is 15 min; for CT, 30

min. Positive correlations were observed for MS and for SOR. Response values increased as the

level increased. Therefore, additional testing at higher levels would have to be done to

determine the best response possible for these factors. From a practical standpoint, a MS of 175

rpm was found to be too rigorous, and 200 rpm resulted in beaker breakage. As a result, 150 rpm

is recommended even though a larger response could be achieved at higher rotational speeds.

There is no restriction on SOR except what is reasonable from a cost and toxicological

standpoint. SORs of 4:1 and 10:1 were tested and shown to give higher oil releases than 2:1.

Recoveries of almost 100% were achieved at 10:1 SOR for Aquaclean, Biosolve and PetroClean.

Data are presented in Chapter 4.

The substrate was also found to be a significant factor. Both substrates are crystalline silica and

should have similar chemical and surface properties. The mean response for sand was 109.6 and

for gravel, 96.1, suggesting sand is easier to clean than gravel. Since sand has a higher surface

area per unit mass, this is counter to what would be expected. However, this can be explained by

the increased spreading of the oil that occurs over the larger surface of the gravel. When a

droplet of oil is applied to sand, the oil coats the individual granules that it first comes in contact

135

with and penetrates into the bed. The spread of the oil affects approximately 5-6 granules of

sand radially from the point of dispensing. When SWA is applied, the coverage of the affected

particles with SWA is easily achieved because the oil has remained within a 1/8” radius of the

point of dispensing. However, when a droplet of oil is applied to gravel, it spreads over the

surface of 2-3 pieces of gravel with an affected radius of approximately ¼”. Due to the larger

spread, it is difficult to achieve 100% coverage of the area with SWA unless the volume of SWA

is large.

6.2.2 Interactions

An analysis of interactions was done to determine whether two-way interactions occur that will

augment or diminish the result obtained by either factor independently. Within this experimental

design, only 2-way interactions were considered; 3-way interactions were assumed to be

confounded.

6.2.2.1 SWA Solution Concentration Interactions

As discussed above, SC main effects were found not to be significant. Dilution of the product

does not change its effectiveness as long as the applied mass is constant. Thus it would seem

that SC can be eliminated as a factor, and a value can be assigned for use in the final protocol.

However, before taking that step it was necessary to determine that the potency of the other

factors was not materially affected by SC. To do this, the two-way interactions between SC and

the other four factors were evaluated. The results are given in Tables 6.2-6.5 and Figures 6.1-

6.4. Of the four interactions, only the SC-MT interaction was statistically significant (p =

0.0159). However, a ceiling effect was observed for MT, and there was no appreciable

136

difference in response between the 15 min and 30 min MT at any of the SC levels. Responses

for MT at 15 min and 30 min were, respectively, 106.0 and 105.9 at 25% SC, 104.4 and 102.0 at

50% SC, and 107.2 and 109.6 at 100% SC. The slight differences revealed no consistent trend

across SC, with the lowest SC-MT responses at 50% SC. Thus, while the p-value is significant

for this two-way interaction, the interaction is caused by the lowest MT level (5 min), which is

the ineffective level. Therefore, the conclusion is that SC is not statistically significant and it

does not affect the potency of the other factors. SC is not an important factor in surface washing.

6.2.2.2 SOR Interactions

The two-way interactions for SOR were evaluated. As stated above and shown in Table 6.2 and

Figure 6.1, the SOR-SC interactions were found not to be significant. Tables 6.6-6.8 and Figures

6.5-6.7 contain the results for SOR-CT, SOR-MS and SOR-MT. Interactions were not

significant for SOR-CT (p = 0.4312) and SOR-MT (p = 0.0541). As stated previously, a ceiling

effect was observed for MT in Table 6.1. The 30 min MT did not increase the response over the

15 min MT. When the two-way interactions were considered for SOR and MT, a slight increase

in response was seen at the highest SOR level for the 30 min MT (130.0) over the 15 min MT

(126.2). This represents a 3% increase for this SOR level. However, at the lower SOR levels, a

decrease in response was observed when going from a 15 min MT to a 30 min MT. Thus, it does

not appear that MT significantly enhances the response for SOR.

The two-way interaction between SOR and MS was significant (p = 0.007). Responses increased

for both factors as the level increased. To determine whether the interaction had a positive or

negative effect on either factor, the responses for the SOR and MS main effects were compared

137

to the SOR-MS interaction effects at the low, middle, and high levels. The main effects and

interaction effects for these factors are as follows:

Level SOR MS SOR-MS Low 80.7 80.2 51.6 Med 104.6 108.5 108.9 High 123.3 119.8 136.5

The main effects for SOR and MS at the low level were 80.7 and 80.2, respectively, and the

interaction effect was 51.6. Thus, a negative effect was seen for the interaction of these two

factors at the low level. At the middle level, the interaction enhanced the SOR response by 4%

and had no effect on MS. At the highest level, a positive effect was observed, with a higher

response for the interaction (136.5) than for the main effects (SOR 123.3, MS 119.8). Since

SOR had no effect on CS, CT, and MT and it had a negative effect at the lowest level for MS, we

can eliminate 1:2 SOR and recommend 1:1 or 2:1 SOR for the protocol. Since no ceiling effect

was observed for SOR, higher SOR can also be considered.

6.2.2.3 Oil-SWA Contact Time

As shown in Table 6.1, a 5% increase in response was observed when CT increased from 15 min

(99.4) to 30 min (104.7). However, CT was found to have a ceiling effect, with no benefit

gained by increasing CT from 30 min to 45 min (104.4). When considered in conjunction with

SC and with SOR, the interaction effects for CT were not significant, as discussed above. The

interaction effects for CT with MS and MT are shown in Tables 6.9-6.10 and Figures 6.8-6.9.

No significant interactions were found. CT-MS interactions had a ceiling effect for the low MS

level, a slight positive effect at the middle MS level, and a slight negative effect at the high level.

CT-MT interactions had a ceiling effect at all levels; responses were not improved by increasing

CT from 30 min to 45 min. Therefore, the recommended CT is 30 min.

138

6.2.2.4 Mixing Speed by Mixing Time Interaction

The last two-way interaction to be considered is the MS-MT interaction. MS had no significant

interactions with CS (p = 0.0622) or CT (p = 0.0806), but it did significantly affect SOR (p =

0.0007). MT had no interactions with CT (p = 0.8958) or SOR (p = 0.0541); while the MT-CS

interaction was statistically significant, it had a ceiling effect so the lowest and least effective

MT level was disregarded. The results for the MS-MT interaction are shown in Table 6.11 and

Figure 6.10. The interaction is significant (p = 0.0001); however, the response is not improved

by going from a 15 min MT to a 30 min MT. When main effects are compared to interaction

effects for these factors, we see a negative interaction effect at the low level (MS 80.2, MT 96.8,

MS-MT 68.0), a positive effect at the middle level (MS 108.5, MT 105.9, MS-MT 113.2), and an

augmenting of MT at the high level (MS 119.8, MT 105.8, MS-MT 119.4). An increasing

positive effect on response was observed for MS. However, mixing speeds greater than 175 are

not recommended due to the potential for glassware breakage.

6.3 CONCLUSIONS

Five factors were evaluated at three levels using a resolution 5, 35-1 fractional factorial design.

Main effects were considered for the five factors. CT was the only factor that was non-

significant. All other factors were found to positively and significantly affect response.

However, ceiling effects were observed for MT and CT, with equivalent responses for the middle

and high levels. Levels over 15 min MT and 30 min CT did not improve the response. A

positive correlation was seen for MS and SOR; response increased with level for these factors.

139

Additional testing at higher SOR can be done. However, excessively rigorous mixing can

become experimentally impractical, so MS greater than 175 is not recommended.

Substrate was also found to be statistically significant. Analyses were performed for each of the

substrates separately and for the two substrates combined. The separate analyses yielded the

same conclusions as the combined model, so the results from the combined model were

presented here.

Two-way interactions were also considered for each combination of the five factors. CS was not

significant as a single-factor and it was not responsible for any significant two-way interactions.

While the CS-MT interaction was significant, it had a ceiling effect and no consistent trend

across the levels. Therefore, SC was determined not to be a significant factor in surface

washing. SOR had a statistically significant interaction with MS, and at the highest level their

combined effect was greater than either one alone. However, at the lowest level, a diminishing

effect occurred. The lowest level is not recommended for either SOR or MS. CT had a ceiling

effect at 30 min and no significant two-way interactions. A ceiling effect was also observed for

MT at 15 min, and a positive interaction was seen for MT and SOR only at the highest level.

Thus an argument could be made for using either the 15 min or 30 min MT. The MT-MS

interaction was significant with positive reinforcement occurring at the middle and high levels.

An increasing positive response was observed for MS, but mixing speeds greater than 175 are

not practical experimentally. The following recommendations can be made based on this

analysis:

140

Factor Recommendation Reason SC Use manufacturers’ recommended

dilution procedures Not a significant factor in surface washing

MT 15 min A ceiling effect observed at 15 min; no benefit gained by 30 min

CT 30 min A ceiling effect observed at 30 min; no benefit gained by 45 min

MS 150 rpm Negative interaction with MT at 125 rpm; 175 rpm too rigorous for this experimental design

SOR 2:1 or higher; appropriateness of SOR for field application depends on cost and toxicity

Increasing positive response with level; testing at higher SOR can be done

141

Figure 6.1- Interaction of SWA Solution Concentration and SOR.

0

20

40

60

80

100

120

140

1:2 1:1 2:1

SOR

Oil

Mas

s in

Aq

ueo

us

Ext

ract

, mg

25% Concentration50% Concentration100% Concentration

Figure 6.2- Interaction of SWA Solution Concentration and Oil-SWA Contact Time.

0

20

40

60

80

100

120

140

15 30 45

Oil-SWA Contact Time, min

Oil

Mas

s in

Aq

ueo

us

Ext

ract

, mg


142

Figure 6.3 - Interaction of SWA Solution Concentration and Mixing Speed.

0

20

40

60

80

100

120

140

125 150 175

Mixing Speed, rpm

Oil

Mas

s in

Aq

ueo

us

Ext

ract

, mg


Figure 6.4 - Interaction of SWA Solution Concentration and Mixing Time.

0

20

40

60

80

100

120

140

5 15 30

Mixing Time, min

Oil

Mas

s in

Aq

ueo

us

Ext

ract

, mg


143

Figure 6.5 - Interaction of SOR and Oil-SWA Contact Time.

0

20

40

60

80

100

120

140

15 30 45

Oil-SWA Contact Time, min

Oil

Mas

s in

Aq

ueo

us

Ext

ract

, mg

1:2 SOR1:1 SOR2:1 SOR

Figure 6.6 - Interaction of SOR and Mixing Speed.

0

20

40

60

80

100

120

140

125 150 175

Mixing Speed, rpm

Oil

Mas

s in

Aq

ueo

us

Ext

ract

, mg


144

Figure 6.7 - Interaction of SOR and Mixing Time.

0

20

40

60

80

100

120

140

5 15 30

Mixing Time, min

Oil

Mas

s in

Aq

ueo

us

Ext

ract

, mg


Figure 6.8 - Interaction of Oil-SWA Contact Time and Mixing Speed.

0

20

40

60

80

100

120

140

125 150 175

Mixing Speed, rpm

Oil

Mas

s in

Aq

ueo

us

Ext

ract

, mg

15 min Contact Time30 min Contact Time45 min Contact Time

145

Figure 6.9 - Interaction of Oil-SWA Contact Time and Mixing Time.

0

20

40

60

80

100

120

140

5 15 30

Mixing Time, min

Oil

Mas

s in

Aq

ueo

us

Ext

ract

, mg

15 min Contact Time30 min Contact Time45 min Contact Time

Figure 6.10 - Interaction of Mixing Speed and Mixing Time.

0

20

40

60

80

100

120

140

5 15 30

Mixing Time, min

Oil

Mas

s in

Aq

ueo

us

Ext

ract

, mg

125 rpm150 rpm175 rpm

146

Table 6.1 – Main Effects.

Effect level

Mean

(mg) Std Dev p-Value

Mix Speed 125 80.2 27. 7 < 0.0001

150 108.5 21.5

175 119.8 17.6

Mix Time 5 96.8 29.5 < 0.0001

15 105.9 25.5

30 105.8 28.4

Oil-SWA Contact Time 15 99.4 29.0 0.0077

30 104.7 28.0

45 104. 4 27.1

SWA Solution Concentration 25 103.6 28.3 0.6904

50 102.0 28.1

100 102.9 28.1

Substrate gravel 96.1 26.3 < 0.0001

sand 109.6 28.2

SOR 1:2 80.7 24.7 < 0.0001

1:1 104.6 20.7

2:1 123. 3 20.4

Table 6.2 – SWA Solution Concentration by SOR Interaction.

SWA Solution

Concentration SOR

Mean

(mg) Std Dev

p-Value

25 1:2 78.2 26.1 0.0874

1:1 108.6 16. 7

2:1 124.1 19.4

50 1:2 80.3 22.8

1:1 104.2 24.5

2:1 121.6 20.7

100 1:2 83.5 26.1

1:1 101.0 20.6

2:1 124.1 21.9

147

Table 6.3 – SWA Solution Concentration by Oil-SWA Contact Time Interaction.

SWA Solution

Concentration

Oil-SWA Contact

Time

Mean

(mg) Std Dev

p-Value

25 15 101.5 29.8 0.2485

30 103.1 30.1

45 106.2 26.3

50 15 95.6 28.6

30 105.9 27.3

45 104.5 28.9

100 15 101.0 30.1

30 105.2 28.1

45 102.4 27.6

Table 6.4 – SWA Solution Concentration by Mixing Speed Interaction.

SWA Solution

Concentration Mix

Speed

Mean

(mg) Std Dev

p-Value

25 125 84.6 30.7 0.0622

150 105.6 24.1

175 120.6 16.5

50 125 75.9 25.6

150 110.3 17.6

175 119.9 19.0

100 125 80.3 27.4

150 109.6 23.0

175 118.8 18.2

148

Table 6.5 – SWA Solution Concentration by Mixing Time Interaction.

SWA Solution

Concentration Mix

Time

Mean

(mg) Std Dev

p-Value

25 5 99.0 29.7 0.0159

15 106.0 23.2

30 105.9 32.2

50 5 99.7 31.0

15 104.4 27.7

30 102.0 26.8

100 5 91.8 28.6

15 107.2 26.9

30 109.6 27.1

Table 6.6 – SOR by Oil-SWA Contact Time Interaction.

SOR

Oil-SWA Contact

Time

Mean

(mg) Std Dev

p-Value

1:2 15 79.3 28.2 0.4312

30 80.0 24.9

45 82.7 21.9

1:1 15 101.2 23.6

30 107.2 18.4

45 105.3 20.4

2:1 15 117.7 22.1

30 127.0 17.6

45 125.2 21.0

149

Table 6.7 – SOR by Mixing Speed Interaction.

SOR Mix

Speed

Mean

(mg) Std Dev

p-Value

1:2 125 51.6 12.6 0.0007

150 89.2 14.9

175 101.2 10.1

1:1 125 83.2 15.2

150 108.9 12.7

175 121.6 11.5

2:1 125 105.9 20.5

150 127.4 16.9

175 136.5 8.1

Table 6.8 – SOR by Mixing Time Interaction.

SOR Mix

Time

Mean

(mg) Std Dev

p-Value

1:2 5 78.1 27.1 0.0541

15 83.6 21.2

30 80.3 26.6

1:1 5 98.8 24.7

15 107.7 19.9

30 107.2 16.6

2:1 5 113.6 26.2

15 126.2 15.0

30 130.0 14.9

150

Table 6.9 – Oil-SWA Contact Time by Mixing Speed Interaction.

Oil-SWA Contact

Time Mix

Speed

Mean

(mg) Std Dev

p-Value

15 125 74.2 29.0 0.0806

150 103.5 19.5

175 120.5 15.8

30 125 83.3 29.0

150 110.1 21.6

175 120.9 18.5

45 125 83.2 25.4

150 112.0 23.4

175 118.0 19.2

Table 6.10 – Oil-SWA Contact Time by Mixing Time Interaction.

Oil-SWA Contact

Time Mix

Time

Mean

(mg) Std Dev

p-Value

15 5 94.9 32.7 0.8958

15 101.9 28.3

30 101.4 26.9

30 5 98.2 27.6

15 107.6 25.3

30 108.4 31.3

45 5 97.4 29.5

15 108.1 23.8

30 107.6 28.0

151

Table 6.11 – Mixing Speed by Mixing Time Interaction.

Mix Speed

Mix Time

Mean

(mg) Std Dev

p-Value

125 5 68.0 22.6 0.0001

15 86.2 25.5

30 86.5 31.4

150 5 100.7 17.8

15 113.3 21.7

30 111.5 23.5

175 5 121.7 18.1

15 118.1 16.8

30 119.4 18.7

152

Chapter 7

COMPARISON OF GLASS PLATE AND NATURAL SUBSTRATE PROTOCOLS

During the course of developing a standardized protocol for testing SWA effectiveness, several

prototype designs were proposed. The two that were tested the most rigorously were the glass

plate protocol and the natural substrate (sand/gravel) protocol, which was ultimately adopted.

Details of these protocols are discussed in Chapters 3 and 4. The glass plate protocol was

abandoned because of high variability in the data collected and concerns over standardization if

plates are fabricated by different companies. However, it is interesting to compare the results

obtained under each protocol and to evaluate whether they give similar predictions for SWA

ranking and efficiency.

7.1 PHYSICAL AND CHEMICAL PROPERTIES OF SUBSTRATES

Glass is composed primarily of silica sand (silicon dioxide, SiO2) and an alkali. These materials

fuse together at high temperatures and form a rigid, disordered (non-crystalline) and non-porous

structure when cooled rapidly. Glass is inert, except to hydrofluoric acid.

Silica is found in nature in several forms, including quartz and opal. Thirty five crystalline forms

have been identified. The most common constituent of sand in inland continental settings and

non-tropical coastal settings is silica, usually in the form of quartz because the considerable

hardness of this mineral resists erosion. However, the composition of sand varies according to

local rock sources and conditions.

153

For this study, ASTM 20/30 silica sand and FilPro Filter Gravel from U.S. Silica Company were

used as the natural substrates. ASTM 20/30 sand (plant: Ottawa, Illinois) is 99.8% whole grain,

unground silicon dioxide (quartz) with other trace metal oxides comprising the remaining 0.2%.

The sieve sizes 20 and 30 correspond to particle sizes of 0.850 mm and 0.600 mm, respectively;

97% of the particles are retained within this range. Sand grains are round with a hardness of 7 on

the Mohs scale. The FilPro Filter Gravel (plant: Mauricetown, New Jersey) is also whole grain,

crystalline silica. Particle sizes range from 1/16” to 1/8” and have a hardness of 7 on the Mohs

scale.

Because glass is made from silica sand, these materials should behave similarly with respect to

surface washing. However, surface area and surface roughness will play a key role in how

strongly the oil attaches to the surface and how easily it is lifted away. The depth of oil

penetration will also be a factor because oil that penetrates into the pore spaces of the sand/gravel

bed may not come in contact with the SWA prior to submerging the substrate in water. The

mixing energy will also be reduced within the protected pore spaces of the bed.

7.2 SURFACE ROUGHNESS

The effect of surface roughness on SWA performance was tested by using etched glass and two

natural substrates, sand and gravel. The glass was etched using three sizes of grinding powders

(80, 220, and 400 grit) to achieve three levels of surface roughness (coarse, medium, and smooth,

respectively). A smooth, unetched glass plate was also tested. For the natural substrate protocol,

silica sand and silica gravel were used to represent two levels of grain size and surface area.

154

7.2.1 Scanning Electron Microscopy

The glass plates, sand, and gravel were analyzed by Scanning Electron Microscopy to evaluate

the surface topography of each substrate. The images are shown in Figures 7.1 and 7.2.

Magnification levels of 100x, 150x, and 40x were used for the glass, sand, and gravel images,

respectively. The surface of the sand is more heterogeneous than the glass plates, with areas that

appear somewhat smooth and areas that are textured or gouged. Overall, its surface

characteristics are probably most similar to the coarse ground plate. For the four levels of glass

etching, ten images were taken at random locations across the surface of the plate. The ten

replicate sample images had similar topographies based on visual examination, indicating a

relatively uniform etching across the plates. The replicate images for the fine, medium and

coarse etched plates are displayed in Figures 7.3, 7.4 and 7.5, respectively.

SEM can provide information about surface topography, but it cannot directly measure surface

area or surface roughness. However, a software program was identified that could digitize the

SEM image and estimate various surface parameters from the 2D image. TrueMap by

TrueGaugeTM Surface Metrology provides visualization and analysis of surface topography,

including estimates for surface roughness, peak-to-valley height and surface area across the

entire surface (S parameter) or any x- or y- cross-sectional profile of the image (P parameter).

The parameter definitions and calculations specified by the software program are as follows:

The average roughness parameter, Sa, is the most used surface roughness parameter. It is the

arithmetic mean or average of the absolute distances of the surface points from the mean plane.

The equation for this parameter is as follows:

155

where M is the number of columns in the surface, N is the number of rows in the surface, and x,

y, z are directional coordinates.

The Root Mean Square (RMS) roughness parameter, Sq, is the root mean square of the surface

departures from the mean plane within the sampling area. It is defined as follows:

where M is the number of columns in the surface and N is the number of rows in the surface. Sq

is equivalent to the sample standard deviation in statistical terms.

The Peak to Valley Height, St, is defined as the sum of the largest peak height (Sp) value and the

largest valley depth (Sv) from the mean plane within the sampling area.

Surface area can be calculated using the Sdr parameter, which is the ratio of the surface area, Sa

(including peaks and valleys) to the nominal area, Sn (width x length).

156

Sdr = Sa/Sn * 100

The limitation of this software application is that the user must input dimensions for the SEM

image in the x-, y-, and z-direction. For a 2D image, the x- and y-dimensions are known, but the

z-dimension is not. Attempts were made to measure the z-height by rotating the plate 90 degrees

and taking a SEM image of the side of the plate. However, an accurate measurement could not

be made due to chipping that occurred along plate edges during fabrication. These chips are not

visible to the eye, but they confounded attempts to accurately measure the depth of etching. The

z-height measured at the plate edges could not be assumed to be the same as the z-height across

the interior of the plate.

In order to estimate an approximate roughness value, it was assumed that the average z-height on

the etched plates is proportional to the average particle size of the grinding powder used for

etching. For the three levels of etching, fine, medium, and coarse, respective grit sizes of 400,

220 and 80 were used. The average particle sizes for these grinding powders are as follows: 22

μm (400 grit); 63 μm (220 grit); 163 μm (80 grit). These values were entered as the z-height

corresponding to each plate type. Values for Sa, Sq, St, and Sdr for the entire plate surface and

Pa, Pq, and Pt for one cross-sectional plane across the center of the plate were calculated.

Figures 7.6, 7.7 and 7.8 contain the 2-D and 3-D digital representations of the plate surface and

the cross-sectional profiles for fine, medium, and coarse plates, respectively. Color on the

digitized image represents depth, which is estimated from the light and dark areas on the SEM

157

image. Blue represents the deepest valleys, red the highest peaks. The cross-sectional profiles

are shown again in Figure 7.9 scaled relative to each other according to their y-axis dimensions.

This was done so that a visual comparison can be made with regard to surface roughness along

this profile. The parameters for the three plate types are summarized in Table 7.1. These values

should be considered relative, not absolute, due to the assumption made for z-height. The values

obtained for the medium plate are approximately 3-4 times those of the fine plate; the coarse

plate values are 8-10 times those of the fine plate, and 2.5 times those of the medium plate.

While the coarse plate has the highest surface area and roughness, the image reveals it also has

larger, deeper gouges with significant smooth areas in the valleys and plateaus.

7.2.2 Other Methods for Determining Surface Area

Atomic Force Microscopy (AFM) was also attempted as a means of estimating surface area.

However, AFM is designed for nano-level surface analysis and would not give meaningful

results for any of the substrates because of the macro-level surface heterogeneity. Nano-scale

measurements would differ across the surface, depending on where the measurement was taken.

Only the fine etched plate could potentially be evaluated by this method. Therefore, this analysis

was not pursued further.

Determination of the surface area of the sand and gravel was attempted by BET isotherms.

Triplicate samples were run for each substrate on two separate occasions using a Tristar 3000

(Micromeritics) porosimetry analyzer. The theory of operation is summarized by the

manufacturer as follows:

158

“The instrument uses physical adsorption and capillary condensation principles to obtain information about the surface area and porosity of a solid material. A sample is contained in an evacuated sample tube and cooled to cryogenic temperatures. It is then exposed to N2 gas at a series of precisely controlled pressures. With each incremental pressure increase, the number of gas molecules adsorbed on the surface increases. The equilibrated pressure (P) is compared to the saturation pressure (Po) and their relative pressure ratio (P/Po) is recorded along with the quantity of gas adsorbed by the sample at each equilibrated pressure. As adsorption proceeds, the thickness of the adsorbed film increases. Any micropores in the surface are filled first, then the free surface becomes completely covered, and finally the larger pores are filled by capillary condensation. The process may continue to the point of bulk condensation of the analysis gas. Then, the desorption process may begin in which pressure systematically is reduced resulting in liberation of the adsorbed molecules. As with the adsorption process, the changing quantity of gas on the solid surface at each decreasing equilibrium pressure is quantified. These two sets of data describe the adsorption and desorption isotherms. Analysis of the shape of the isotherm yields information about the surface and internal pore characteristics of the material.”

Sand and gravel samples were purged with nitrogen gas for 2 hrs at 150°C using Flow prep 060

(Micromeritics). The method utilized five points between 0.1 and 0.3 relative pressure to

calculate the BET N2 gas adsorption isotherm. The results were as follows:

Run number 1:

ASTM20/30 sample: 0.0459, 0.0486, 0.0446 m2/g

FilPro sample: 0.2359, 0.2369, 0.2343 m2/g

Run number 2:

ASTM20/30 sample: 0.0859 m2/g

FilPro sample: 0.2645m2/g.

Based on these data, it would appear that the gravel has a larger surface area per unit mass than

the sand, which is not reasonable. Both substrates are inert with no internal porosity, so it is

assumed that the smaller particle should have higher surface area per gram. However, BET

159

analysis is accurate only for surfaces with very high surface area and internal porosity.

Therefore it is assumed that BET analysis is not appropriate for these low surface area substrates.

7.2.3 Substrate Sphericity and Shape Factor

From SEM images of the sand and gravel particles used in this study, an average sphericity and

shape factor can be assigned. Figure 7.10 shows these SEM images, as well as typical sphericity

and shape factors of granular materials for the purpose of comparison (Fair, et al., 1968). The

ASTM 20/30 sand used in our studies can be characterized as worn with an average sphericity of

0.94, shape factor of 6.4, and porosity of 0.39. The gravel particles are also worn, but slightly

more elongated on average.

7.2.4 Sand Particle Size Distribution

Particle size distribution (PSD) for the sand substrate was measured on ten replicate, well-mixed

samples to confirm the specifications provided on the Product Data Sheet. A Beckman Coulter

LS Particle Size Analyzer was used for this analysis. The PSD for the ten replicate samples is

shown in Figure 7.11. The size range predicted by the Product Data Sheet is designated by the

rectangular box. Particle diameters for all samples fell between 320 μm and 1100 μm, with a

mean particles size of 815.8 μm. This range is much wider than indicated by the Product Data

Sheet (600 μm - 850 μm); however, the measured mean particle size does fall within the

expected range.

160

7.3 COMPARISON OF EXPERIMENTAL RESULTS

The glass plate protocol and the natural substrate protocol differ in many respects. Most notably

is the mixing regime that is applied during the washing stage. In the glass plate protocol, the

plates were suspended from a shaft and spun around a vertical axis at a rotational speed of 200

rpm. Square jars with no baffles were used. In the natural substrate protocol, the baskets of sand

were placed in round, baffled beakers and agitated on an orbital shaker table at 150 rpm. Both

approaches utilized a rotational mixing scheme, but the flow of water across the substrate was

very different. This along with differences in substrate surface characteristics naturally led to

differences in results when the two protocols were compared.

Experiments were conducted at 1:1 and 10:1 SOR for Aquaclean, Biosolve, Corexit, Petroclean,

and Petrotech under the glass plate protocol and the natural substrate protocol. Figure 7.12

shows the results for the coarse plate and for sand at 1:1 SOR. In every case, SWAs were more

effective at removing PBC from sand than from the coarse plate. This difference was greatest

for Aquaclean, which achieved 60.3% removal from sand and only 12.0% removal from the

plate, and least for Corexit, which had 41% to 44% removal under both protocols. These

removals for Corexit are comparable to those obtained by Fingas, et al. (1990) in the

Environment Canada (42%). Rankings of the SWAs were also different under the two testing

procedures. For the natural substrate protocol, the ranking was as follows:

Aquaclean > Biosolve > Corexit > Petrotech > Petroclean

In the glass plate protocol, the following ranking was seen:

Corexit > Petrotech > Biosolve > Aquaclean > Petroclean

161

Figure 7.13 shows the results for the coarse plate, extra coarse plate, and sand at 10:1 SOR.

There is good agreement between the two plate types for all SWAs. Once again there was more

oil removed from sand than from the plates for all SWAs except Corexit, which performed

significantly poorer under the sand protocol than under the glass plate protocol. However, with

the exception of Corexit, the relative rankings were the same under both protocols:

Biosolve > Petroclean > Aquaclean > Petrotech

Efficiencies close to 100% were observed for the top three SWAs at this SOR under the sand

protocol, while the removals from glass plates ranged from 63% to 84% for these same SWA

under the glass plate protocol.

7.4 CONCLUSIONS

Due to variability of data and concerns over standardized equipment, the glass plate protocol was

abandoned in favour of the natural substrate protocol. The results collected under each protocol

differed in the measured efficiency and relative ranking of SWA. The natural substrate protocol

was tested for reproducibility and repeatability and found to be superior. Thus, this protocol will

be recommended for adoption by U.S. EPA.

162

SM

OO

TH

FIN

E -

400

ME

DIU

M -

220

CO

AR

SE

-80

Figu

re 7

.1 –

SE

M I

mag

es o

f G

lass

Pla

tes

at 1

00x

Mag

nifi

cati

on

800

µm

800

µm80

0 µm

800

µm

SM

OO

TH

FIN

E -

400

ME

DIU

M -

220

CO

AR

SE

-80

Figu

re 7

.1 –

SE

M I

mag

es o

f G

lass

Pla

tes

at 1

00x

Mag

nifi

cati

on

800

µm80

0 µm

800

µm80

0 µm

800

µm80

0 µm

800

µm80

0 µm

163

Figure 7.2 – SEM Images of a) ASTM 20/30 Sand at 150x Magnification and b) FilPro Gravel at 40x Magnification.

a

b

500 µm

2 mm

a

b

500 µm

2 mm

164

Figu

re 7

.3 –

Rep

licat

e SE

M I

mag

es o

f Fi

ne G

lass

Pla

tes

at 1

00x

Res

olut

ion

Figu

re 7

.3 –

Rep

licat

e SE

M I

mag

es o

f Fi

ne G

lass

Pla

tes

at 1

00x

Res

olut

ion

165

Figu

re 7

.4 –

Rep

licat

e SE

M I

mag

es o

f M

ediu

m G

lass

Pla

tes

at 1

00x

Res

olut

ion

Figu

re 7

.4 –

Rep

licat

e SE

M I

mag

es o

f M

ediu

m G

lass

Pla

tes

at 1

00x

Res

olut

ion

166

Figu

re 7

.5 –

Rep

lica

te S

EM

Im

ages

of

Coa

rse

Gla

ss P

late

s at

100

x R

esol

utio

nFi

gure

7.5

–R

epli

cate

SE

M I

mag

es o

f C

oars

e G

lass

Pla

tes

at 1

00x

Res

olut

ion

167

Figu

re 7

.6 –

Dig

itize

d 2-

D a

nd 3

-D I

mag

es a

nd C

ross

-Sec

tiona

l Pro

file

of

Fine

Gla

ss P

late

at 1

00x

Mag

nifi

catio

n.Fi

gure

7.6

–D

igiti

zed

2-D

and

3-D

Im

ages

and

Cro

ss-S

ectio

nal P

rofi

le o

f Fi

ne G

lass

Pla

te a

t 100

x M

agni

fica

tion.

168

Figu

re 7

.7 –

Dig

itiz

ed 2

-D a

nd 3

-D I

mag

es a

nd C

ross

-Sec

tiona

l Pro

file

of

Med

ium

Gla

ss P

late

at 1

00x

Mag

nifi

cati

on.

Figu

re 7

.7 –

Dig

itiz

ed 2

-D a

nd 3

-D I

mag

es a

nd C

ross

-Sec

tiona

l Pro

file

of

Med

ium

Gla

ss P

late

at 1

00x

Mag

nifi

cati

on.

169

Figu

re 7

.8 –

Dig

itize

d 2-

D a

nd 3

-D I

mag

es a

nd C

ross

-Sec

tiona

l Pro

file

of

Coa

rse

Gla

ss P

late

at 1

00x

Mag

nifi

catio

n.Fi

gure

7.8

–D

igiti

zed

2-D

and

3-D

Im

ages

and

Cro

ss-S

ectio

nal P

rofi

le o

f C

oars

e G

lass

Pla

te a

t 100

x M

agni

fica

tion.

170

Fin

e

Med

ium

Coa

rse

Fig

ure

7.9

–Cro

ss-S

ecti

onal

Pro

file

of

Fin

e, M

ediu

m, a

nd C

oars

e G

lass

Pla

tes

Scal

ed to

Rel

ativ

e Si

ze.

Fin

e

Med

ium

Coa

rse

Fin

e

Med

ium

Coa

rse

Fig

ure

7.9

–Cro

ss-S

ecti

onal

Pro

file

of

Fin

e, M

ediu

m, a

nd C

oars

e G

lass

Pla

tes

Scal

ed to

Rel

ativ

e Si

ze.

171

500

µm

a) S

and

150

x

2 m

m

2 m

m

c) G

rave

l 40x

b)

San

d 4

0x

Figu

re 7

.10

–P

artic

le C

hara

cter

istic

s an

d SE

M I

mag

es f

or A

ST

M 2

0/30

San

d an

dFi

lPro

Gra

vel.

500

µm

a) S

and

150

x

500

µm

a) S

and

150

x

2 m

m

2 m

m

c) G

rave

l 40x

2 m

m

c) G

rave

l 40x

b)

San

d 4

0x

Figu

re 7

.10

–P

artic

le C

hara

cter

istic

s an

d SE

M I

mag

es f

or A

ST

M 2

0/30

San

d an

dFi

lPro

Gra

vel.

172

Figure 7.11 – Particle Size Distribution for Ten Replicate Sand Subsamples. Particle Size Distribution for Ten Replicate Sand Subsamples

0

5

10

15

20

25

30

0 200 400 600 800 1000 1200 1400 1600 1800

Particle Diameter, micron

Vo

lum

e, %

173

Figure 7.12 – Removal of PBC from C Plate and Sand at 1:1 SOR.

0

20

40

60

80

100

120

Aquaclean Biosolve Petroclean Petrotech Corexit

Oil

Rem

ova

l, %

1:1 sand1:1 C plate

Figure 7.13 – Removal of PBC from C and XC Plates and Sand at 10:1 SOR.

0

20

40

60

80

100

120

Aquaclean Biosolve Petroclean Petrotech Corexit

Oil

Rem

ova

l, %

10:1 sand10:1 C plate10:1 XC plate

174

Table 7.1 – Surface Characteristics for Fine, Medium and Coarse Glass Plates. Parameter Description Fine (400 grit) Medium (220 grit) Coarse (80 grit)

z=22 μm z=63 μm z=163 μmComputed from one cross-sectional profilePa average roughness, μm 1.6 6.2 15.3Pq RMS roughness, μm 2.0 7.8 19.7Pt peak to valley height, μm 12.2 47.2 104.6Computed from entire surfaceSa average roughness, μm 2.1 6.8 17.4Sq RMS roughness, μm 2.6 8.5 21.6St peak to valley height, μm 22.0 63.0 163.0Sdr ratio of surface area to nominal area 206.2 983.5 4147.5

175

Chapter 8

INTERFACIAL CHEMISTRY FOR SURFACE WASHING AGENTS

In order to understand what makes a surface washing agent (SWA) effective, it is necessary to

examine the chemical and physical interactions that occur at the system surfaces. A surface

washing system contains three phases (oil [O], water [W], and solid substrate [S]) and three

interfaces (oil-water [OW], oil-substrate [OS], and water-substrate [WS]). Due to its unique

amphipathic nature, a pure surfactant will align at these interfaces, especially at the OW

interface, and reduce the interfacial tension. This is the primary mechanism of action for surface

washing agents, even though their formulations are proprietary and may also contain solvents

and additives. Key factors in detergency include the surfactant’s ability to 1) form micelles, 2)

lower the OW interfacial tension, 3) increase the contact angle between the oil and the substrate,

and 4) preferentially wet the substrate surface (Thompson, 1994). These factors are examined in

this chapter for six SWAs and a relative ranking of SWAs is assigned based on these criteria.

8.1 CRITICAL MICELLE CONCENTRATION

Surfactants reduce the surface tension of water by adsorbing at the liquid-gas interface, and they

reduce OW interfacial tension by adsorbing at the liquid-liquid interface. They can also assemble

in the bulk solution into soluble aggregates known as micelles. When micelles form in water,

their hydrophobic tails form a core that can encapsulate an oil droplet, and their heads, which are

typically ionic or polar, form an outer shell that maintains favorable contact with water. A

schematic picture of a micelle is shown in Figure 8.1. The external “spheres” represent the

176

hydrophilic portions of the molecules and the internal “tails” represent the hydrophobic portions,

which are responsible for the aggregation.

At low concentrations, surfactants will favor arrangement on the surface. As the surface

becomes crowded with surfactant molecules, some molecules will arrange into micelles. The

concentration at which surfactants begin to form micelles is known as the critical micelle

concentration (CMC). This is also the point at which the surface has become completely loaded

with surfactant. Since it is desirable from a surface cleaning point of view to maximize the

molecules at the surface, concentrations at or above the CMC should be used.

The CMC for a surfactant in solution can be determined by measuring the surface tension of the

solution as a function of surfactant concentration. As mentioned, surfactant molecules will align

at the surface until the surface becomes saturated and micelles begin to form. The CMC can be

identified by determining the concentration at which the surface tension stops dropping with an

incremental increase in concentration. If the solubility limit is reached before the surface tension

reaches a plateau, then the compound has no CMC. A plot of surface tension vs. log of

concentration for a typical surfactant will have look like the graph in Figure 8.2.

8.1.1 Method

For this work, CMC was determined by Dr. Christopher Rulison of Augustine Scientific for five

SWAs: Aquaclean, Biosolve, Petroclean, Petrotech, and Superall. Corexit 9580 is not water

soluble and does not form micelles. Surface tension was measured by the Wilhelmy plate

177

technique for a minimum of 40 concentrations of each SWA. This was done by incrementally

dosing higher concentration stock solutions into initially pure water. Tests were run in duplicate.

8.1.2 Results

The plots of surface tension vs. log concentration for each surfactant are shown in Figure 8.3. A

line is drawn through the slope of the curve where the surface tension is dropping, and another

line is drawn through the plateau of the curve. The intersection of these lines is the CMC. To

maximize the concentration of free surfactant at the OW interface, it is necessary to work at

concentrations at or above the CMC.

The CMC values for the five SWAs in pure water are shown in Table 8.1. They are listed in

order of increasing CMC:

Biosolve < Aquaclean < Petrotech < Petroclean < Superall

This suggests that the onset of detergency will begin at much lower SWA concentrations for

Biosolve than for Superall. The variance in CMC between the samples is 13.5 from the lowest

CMC product (Biosolve) to the highest (Superall). However, all the CMC’s are in the 100ths of a

percent range for all samples. It is common for non-ionic surfactants, particularly blended

surfactants, to have CMC in this range. These concentrations are much lower than the

concentrations that had been used in the testing protocol. Most experiments were conducted at

2:1 SOR with total SWA application volume of 180 µl. Thus, the typical concentration in the

wash water was 180 µl per 100 ml, or 0.180%, which is above the CMC for all products.

178

8.2 INTERFACIAL TENSION

As discussed in Chapter 1, commercial SWAs contain a mixture of surfactants, solvents, and/or

additives. The exact formulation of each SWA is not known because this information is

proprietary. However, the primary mechanism of action is considered to be the alignment of

surface active agents at the oil-water interface, which causes a lowering of OW interfacial

tension and an increase in contact angle between the oil and the substrate. As the contact angle

is increased, the oil film rolls up into droplets and is released. This is called the roll-up

mechanism.

The relationship between interfacial tension and contact angle is defined by Young’s equation

for a three phase system:

σSC = σSO + σCO cos θ

where σSC = the interfacial tension between the cleaning solution and the substrate; σSO = the

interfacial tension between the oil and the substrate; σCO = the interfacial tension between the

cleaning solution and the oil; and θ = the contact angle between the oil and the substrate (Young,

1805). Surface wetting occurs when the contact angle is less than 90°; detachment occurs when

the contact angle equals 180°.

As stated previously, the efficiency of a SWA is related to its ability 1) to lower interfacial

tension, 2) to increase contact angle, and 3) to preferentially wet the substrate surface better than

oil. These criteria are inter-related, but each does not necessarily predict SWA efficiency when

evaluated by itself, as will be discussed below.

179

8.2.1 Method

For this work, the OW equilibrium interfacial tensions for six SWA against Prudhoe Bay Crude

oil were measured by Dr. Christopher Rulison of Augustine Scientific. Aquaclean, Biosolve,

Petroclean, Petrotech, and Superall were evaluated at the following concentrations: 100% SWA

(undiluted), 50% aqueous, and 5% aqueous. Corexit 9580 is not water soluble, so only the

undiluted product was tested.

The experiments were performed by the pendant drop method using a Kruss Drop Shape

Analysis System DSA10. This instrument is capable of measuring interfacial tensions down to

0.01 mN/m. According to this method, a drop of liquid is formed on the end of a 254-μm

capillary tip within a bulk phase of a second liquid. For this work, a drop of PBC oil was formed

within each bulk SWA solution. For SWA solutions that were heavier in density than the oil, the

oil droplet was formed from an upward-pointing capillary. Corexit 9580 was the only SWA

tested that was lighter than oil; for this SWA, the capillary was downward-pointing. The drop is

formed to about 90% of its detachment volume. The drop is then digitally imaged. The drop’s

image is fit by a robust mathematical approach to determine the drop’s mean curvature at over

300 points along its surface.

The curvature of a drop that is pendant to a capillary tip, at any given point on its interface with

the continuous phase, is dependent on two opposing forces. Buoyancy works to make the drop

elongated or “drip-like.” The greater the difference in density between the drop liquid and the

continuous phase, the greater this force. Interfacial tension works to keep the drop spherical

180

since a sphere has the lowest surface to volume ratio of any shape, and interfacial tension is, by

definition, the amount of work necessary to create a unit area of interface. Pendant drop

interfacial tension evaluation involves observing the balance that exists between these two forces

on a pendant drop, in the form of the drop’s mean curvature at various points along its interface

with the continuous phase. Lower interfacial tension will have a more “drip-like” shape, while

higher interfacial tension will have a more spherical shape.

The mathematics of pendant drop analysis is based on the Young-Laplace equation, which

describes the equilibrium pressure balance at the interface between two static fluids. The

pressure difference at any given point on the surface (ΔP) is equal to mean curvature of the

interface at that point multiplied by twice the tension (σ) contained in the interface.

where Δp is the pressure difference over the interface, γ the surface tension, H is the mean

curvature, and R1 and R2 are the principal radii of curvature at the interface (Laplace, 1806).

For a pendant drop, the pressure difference within the drop, between two vertical positions is:

Δρ g Z

where Δρ = difference in density between the liquid that is forming the drop and the bulk phase,

g = gravity, and Z = vertical distance between the two positions, as shown in Figure 8.4.

181

The measurement of interfacial tension is made by determining the mean curvature on the drop

at over 300 points (like those labeled A and B in Figure 8.4). Those points are then used in pairs,

with the equations given above, to solve for interfacial tension in the following manner:

((1/r1 +1/r2)at A - (1/r1 +1/r2)at B ) 2 σ = Δρ g Zbetween A and B

Therefore, from one drop image, interfacial tension is determined at least 150 times. These

interfacial tension values are averaged to give a single value for the overall interfacial tension of

the drop. This technique has been found to be extremely accurate for determining surface

tensions of liquids with known surface tension (typical errors of less than 0.1%), as well as

interfacial tensions that are known by other means.

8.2.2 Results

The pendant drop technique requires accurate values for the densities of all liquids to be

analyzed. Densities were measured for PBC and each SWA solution by weighing fixed volumes

of each solution. The results are contained in Table 8.2. Corexit 9580 was the only SWA

solution that was lighter than oil. As mentioned, the capillary had to be positioned downward in

the SWA solution for this sample.

Interfacial tensions were then determined for each SWA solution with five drop replication.

These data are presented in Table 8.3 for undiluted SWAs, Table 8.4 for 50% solution

concentrations, and Table 8.5 for 5% solution concentrations. The data in these tables are

presented in order of increasing interfacial tension, which would be thought to correlate with

poorer performance as you go from left to right. Thus, the relative ranking of SWAs based on

ability to lower interfacial tension is as follows:

182

Corexit > Aquaclean > Biosolve > Petroclean > Superall > Petrotech

The ordering among the samples did not change with dilution, although obviously greater

dilution means higher interfacial tension. If we assume a SWA’s ability to lower OW interfacial

tension is directly related to its ability to remove oil from a substrate, then this should also

predict cleaning effectiveness. This assumption is based on the concept that lower interfacial

tension between the oil and the solution means that less energy is required to emulsify the oil

into solution. It is not surprising that Corexit, which is hydrophobic, should have the lowest

interfacial tension with the oil. However, there is more to cleaning oil from a surface than

emulsification of the oil. Specifically, the cleaning solution ought to be favored over the oil in a

competitive wetting situation for a given substrate. These concepts are discussed in the next

section.

8.3 CONTACT ANGLE AND COMPETITIVE WETTING

In order to understand a SWA’s ability to remove oil from a surface, it is important to consider

more than the interfacial tension between the two liquids. Displacement of the oil from the

surface must be thermodynamically favored. For this to be true, the SWA must have a lower

interfacial tension with the substrate than does the oil to be removed. This is known as

competition for wetting of a surface.

Complete determinations of solid-liquid interfacial tensions cannot be done directly. They must

be calculated based on determining the polar and non-polar components of the surface tension of

183

each of the liquids involved, as well as those of the surface energy of the solid surface. A more

practical means of studying competitive wetting is by three phase contact angle experimentation.

Contact angle is the angle formed by a liquid at the three phase contact line where a solid and

two liquids intersect. To determine contact angle, a substrate is pre-wet by submerging it in one

of the liquids of interest, a drop of the other liquid is applied, and the contact angle between the

solid and the drop is measured. The shape of the drop is controlled by the three forces of

interfacial tension shown in Figure 8.5. The contact angle is a quantitative measure of the

wetting of a surface by a liquid. A contact angle of 0 degrees represents the complete wetting of

a surface, while a contact angle of 180 degrees represents the complete repulsion of the liquid by

the surface.

Contact angles and wettability are interrelated components. The angle of incidence between a

solid surface and a liquid can predict how wettable, or non-wettable the solid is with respect to

that particular liquid. Generally the higher the contact angle, the less wettable the solid. If water

is used as the test liquid, this can also give insight into how hydrophobic or hydrophilic the solid

in question.

8.3.1 Method

For this study, the three phase contact angle for PBC oil in the presence of each of six SWAs was

measured by Dr. Christopher Rulison of Augustine Scientific. Aquaclean, Biosolve, Petroclean,

Petrotech, and Superall were evaluated at the following concentrations: 100% SWA (undiluted),

184

50% aqueous, and 5% aqueous. Corexit 9580 is not water soluble, so only the undiluted product

was tested.

For this analysis, a smooth piece of granite or sandstone could be used. Because the chemical

properties of sandstone are a closer approximation of natural beach substrate, sandstone was

selected for this analysis. The surface was submerged in each cleaning solution, then oil droplets

were introduced via a curved upward-pointing syringe to the underside of the sandstone. For

Corexit 9580, which was the only SWA tested that is lighter than oil, the drop of oil was

introduced using a downward-pointing syringe to the top surface of the sandstone. Contact angle

was then directly measured on each drop using a Kruss DSA10 Drop Shape Analysis System.

Five replicate drops were analyzed per solution and the results averaged to obtain a single

contact angle measurement for each SWA at each solution concentration.

8.3.2 Results

The results of three phase contact angle measurements for oil droplets on sandstone in the

various SWA solutions are given in Tables 8.6, 8.7, and 8.8 for 100%, 50%, and 5% SWA

concentrations, respectively. Pictures of drop #5 from each of the above contact angle studies

are shown in Figures 8.6 through 8.8. An examination of these data shows that the trend among

the products from highest to lowest oil contact angle is as follows:

Biosolve > Petroclean > Superall > Petrotech > Aquaclean > Corexit 9580

185

This trend is significantly different from the ranking predicted by interfacial tension. In fact, the

lowest interfacial tension producing SWAs – Corexit 9580 and Aquaclean – actually cause the

lowest contact angles. Thus, while they are predicted by interfacial tension to be good

emulsifiers, they do not provide the best competition for the crude oil in terms of competing for

the sandstone surface. The relative rankings of Biosolve, Petroclean, Superall, and Petrotech are

the same for both interfacial tension and contact angle. Biosolve, which has the highest contact

angles at each concentration studied and also the second lowest interfacial tension, appears to

have the best combination of properties for removal of crude oil from sandstone based on these

data.

While interfacial tension and contact angle are both key elements in surface cleaning, cleaning

efficiency is ultimately determined by the phase (oil or SWA) that has a lower interfacial tension

with the substrate. For cleaning to be thermodynamically favored, the SWA must have a lower

interfacial tension with the substrate than does the oil to be removed. Solid-liquid interfacial

tensions cannot be measured directly, but the difference between the two, σSC - σSO, can be

calculated by Young’s equation (Young, 1805), as follows:

σSC - σSO = σCO cos θ

We know the interfacial tension between the oil and the cleaning solution and the three phase

contact angle, so only the left side of this equation is unknown. This quantity, σSC - σSO, should

ideally be negative, indicating that the interfacial tension between the substrate and the oil is

greater than the interfacial tension between the substrate and the cleaning solution. In other

words, if the quantity is negative, the cleaning solution is favored to interact with the substrate in

preference to the oil, and should lift the oil from the surface.

186

For that to happen the contact angle for the oil drop on the sandstone, within the cleaning

solution, has to be greater than 90o. It also makes intuitive sense that if the contact angle for the

oil is greater than 90o then the oil is wetting the sandstone to a lesser extent than is the cleaning

solution, and we see that is the case for some, but not all, of the SWA solutions.

Table 8.9 contains values for the σSC - σSO difference. Negative values indicate that the SWA is

more attractive to the sandstone surface than is the oil; therefore, oil removal is favored. As

predicted, Biosolve (BS) produces the most favorable oil removal values both as 100% product

and as a 50% dilution. Petroclean (PC) ranks second at the 50% dilution level, but it is slightly

less negative than Superall (SA) at the 100% level. All SWAs have positive (non-favorable)

interfacial tension differences in the 5% aqueous condition. However, Aquaclean (AC) and

Petrotech (PT) also have positive values at the 50% aqueous condition, and Corexit 9580 (C) at

the 100% condition. The relative ranking at each level is as follows:

100% SWA: BS > SA > PC > PT > AC > C

50% SWA: BS > PC > SA > AC > PT

5% SWA: BS > PC > AC > SA > PT

If we rank these products based on their manufacturer recommended dilution rates for field

applications, Petrotech and Corexit would be considered at 100%, Aquaclean at 50%, and the

other three products at the lowest concentration level. The relative ranking would be as follows:

187

Manufacturer Recommended Rate: PT > C > AC > BS > PC > SA

Alternately, the σSC - σSO differences for the three concentration levels can either be summed or

averaged, and the following ranking results:

Summed or Averaged σSC - σSO: BS > PC > AC > SA > PT

8.4 SWA PERFORMANCE UNDER PROPOSED PROTOCOL

The relative ranking of SWAs based on performance under the proposed protocol was found to

be SOR dependent. Certain products performed consistently across SOR, while others were far

superior at high SOR than at low. Figure 8.9 shows the relative performance of SWAs at each

SOR. The most notable change with SOR was for Petroclean, which was ranked 5th at 1:1 SOR

but moved up to 2nd at 10:1. Superall also improved significantly with SOR, moving from last

ranking at 1:1 to being tied with Petrotech for 4th ranking at 10:1 SOR. Biosolve and Aquaclean

performed well at all levels. Observed rankings at each SOR were as follows:

1:1 SOR AC > BS > C > PT > PC > SA

2:1 SOR BS > AC > PT > C > PC > SA

4:1 SOR BS > AC > PC > C > SA > PT

10:1 SOR BS > PC > AC > SA > PT > C

Predicted rankings based on interfacial chemistry analyses discussed above can be summarized

as follows:

188

CMC: BS > AC > PT > PC > SA

Interfacial Tension: C > AC > BS > PC > SA > PT

Contact Angle: BS > PC > SA > PT > AC > C

Summed or Averaged σSC - σSO: BS > PC > AC > SA > PT

At high SOR, the best predictor of SWA performance was the summed or averaged σSC - σSO

difference. A high correlation was found between this variable and the measured efficiency

under the protocol at 10:1 SOR. These data are plotted in Figure 8.10 and 8.11 for the summed

and averaged σSC - σSO differences, respectively. The linear fit through the data gave R2 values

of 0.93 for both sets of data. When SWA efficiency at 10:1 SOR is plotted against the σSC - σSO

difference for each solution concentration individually, we see the data are highly correlated for

the 5% concentration (R2 = 0.94) and not correlated at all for the undiluted σSC - σSO difference

(R2=0.003) (Figure 8.12). In the protocol experimental system, the σSC - σSO difference is a

measure of the preferential wetting of the substrate surface by either the oil or the wash water

that contains the surfactant. In the 10:1 SOR studies, SWA was added at a rate of 0.9 ml SWA

concentrate per 100 ml wash water (final wash concentration = 0.9%). Thus, it is not surprising

that the σSC - σSO difference for the 5% aqueous solution would be a better measure of SWA

efficiency under the protocol. No significant relationship was observed between the σSC - σSO

difference and SWA performance at lower SOR, indicating that certain SWAs must be applied at

higher applications rates to be effective.

189

At low SOR, CMC was a better predictor of SWA efficiency. Figure 8.13 shows the relationship

between SWA efficiency at 1:1 SOR and CMC. The linear fit through these data gave an R2 of

0.82. In the 1:1 experiments, SWA was added at a rate of 0.09 ml SWA per 100 ml wash water

(final wash concentration = 0.09%). This is above the CMC for Biosolve (0.01%), Aquaclean

(0.02%) and Petrotech (0.05%); close to the CMC for Petroclean (0.08%); and below the CMC

for Superall (0.15%). At 10:1 SOR, the correlation was low between efficiency and CMC (R2 =

0.35), suggesting that CMC is only a predictor of efficiency at concentrations close to the CMC.

When surfactants are used at concentrations many times greater than CMC, other interfacial

measurements become more important.

8.5 CONCLUSIONS

A SWA’s ability to lower interfacial tension is an important aspect of surface cleaning in that

low interfacial tension means less energy is needed to emulsify crude oil in the cleaning solution.

However, a look at the relative abilities of the SWAs to hold off or displace oil from a sandstone

surface reveals a different order of predicted cleaning effectiveness. This is common in hard

surface cleaning, with the best cleaners being designed to both have low interfacial tension and

the ability to displace the boundary layer of oil from a surface. The two properties are

independent to some extent and also surface dependent.

In general, the best cleaner for a given oil will be one that matches the oil in terms of the

hydrophilic/lipophilic balance. A surfactant with hydrophobic tail lengths similar to the chain

lengths in the oil will promote packing of the surfactant at the oil-water interface and optimize its

cleaning effectiveness. For high oil loads, the OW interfacial tension is key because oil is being

190

emulsified away from itself rather than from the substrate. For removing the last layer of oil

from the substrate surface, contact angle and competitive wetting become more important.

In these studies, the best predictor of SWA performance at high SOR was the term representing

the difference between SWA-substrate and oil-substrate interfacial tensions (σSC - σSO). This

term is a measure of the preferential wetting of the substrate surface by either the oil or the wash

water. It was calculated from measured values for interfacial tension and contact angle using

Young’s equation. A high correlation was observed between SWA efficiency at 10:1 SOR and

the σSC - σSO difference at 5% SWA solution concentration. Using the σSC - σSO difference at the

lowest concentration level is logical because the SWA in the final wash water is less than 1%.

The correlation applies to high SOR, indicating that SWA concentrations many times greater

than the CMC must be used for optimal SWA performance. At low SOR, CMC was a better

predictor of SWA performance.

191

Figure 8.1 – Schematic Drawing of a Micelle.

192

Figure 8.2 – Method for Determining CMC from Surface Tension Data with Incremental Changes in Solution Concentrations. Figure 8.3 – Determination of CMC for Five SWAs.

25

30

35

40

45

50

0.001 0.010 0.100 1.000

Concentration (%)

Sur

face

Ten

sion

(m

N/m

)

Bio

solv

e

Aqu

acle

an

Pet

rote

ch

Pet

rocl

ean

Sup

eral

l

Concentration [mg/ml]

Surf

ace

Ten

sion

[mN

/m]

0.001 0.01 0.1 1.0 10

20.0

25.0

30.0

35.0

40.0

45.0

50.0

55.0

60.0

CMC

193

Figure 8.4 – Schematic Drawing of Pendant Drop Method for Determining Interfacial Tension. Figure 8.5 – Measurement of Contact Angle for Oil on Sandstone Surface.

Sandstone

Cleaning Solution

Oil θ

σco

σso σsc

z

A

B

Sandstone

Cleaning Solution

Oil θ

σco

σso σsc

194

PB

C O

il D

ropl

et o

n S

ands

tone

in

100

% B

ioso

lve

Dro

p #5

Con

tact

Ang

le =

124

.8o

PB

C O

il D

ropl

et o

n S

ands

tone

in

100

% P

etro

clea

n D

rop

#5 C

onta

ct A

ngle

= 1

10.5

o

PB

C O

il D

ropl

et o

n S

ands

tone

in

100

% A

quac

lean

D

rop

#5 C

onta

ct A

ngle

= 9

2.1o

PB

C O

il D

ropl

et o

n S

ands

tone

in

100

% S

uper

all

Dro

p #5

Con

tact

Ang

le =

104

.1o

PB

C O

il D

ropl

et o

n S

ands

tone

in

100

% C

orex

it 95

80D

rop

#5 C

onta

ct A

ngle

= 7

6.0o

PB

C O

il D

ropl

et o

n S

ands

tone

in

100

% P

etro

tech

D

rop

#5 C

onta

ct A

ngle

= 9

8.9o

Bio

solv

e >

Pet

rocl

ean

> S

uper

all>

Pet

rote

ch >

Aqu

acle

an >

Cor

exit

PB

C O

il D

ropl

et o

n S

ands

tone

in

100

% B

ioso

lve

Dro

p #5

Con

tact

Ang

le =

124

.8o

PB

C O

il D

ropl

et o

n S

ands

tone

in

100

% P

etro

clea

n D

rop

#5 C

onta

ct A

ngle

= 1

10.5

o

PB

C O

il D

ropl

et o

n S

ands

tone

in

100

% A

quac

lean

D

rop

#5 C

onta

ct A

ngle

= 9

2.1o

PB

C O

il D

ropl

et o

n S

ands

tone

in

100

% S

uper

all

Dro

p #5

Con

tact

Ang

le =

104

.1o

PB

C O

il D

ropl

et o

n S

ands

tone

in

100

% C

orex

it 95

80D

rop

#5 C

onta

ct A

ngle

= 7

6.0o

PB

C O

il D

ropl

et o

n S

ands

tone

in

100

% P

etro

tech

D

rop

#5 C

onta

ct A

ngle

= 9

8.9o

Bio

solv

e >

Pet

rocl

ean

> S

uper

all>

Pet

rote

ch >

Aqu

acle

an >

Cor

exit

Fig

ure

8.6:

Con

tact

Ang

les

for

PB

C O

il o

n S

ands

tone

in 1

00%

SW

A.

195

PB

C O

il D

ropl

et o

n S

ands

tone

in

50%

Aqu

eous

Aqu

acle

an

Dro

p #5

Con

tact

Ang

le =

80.

9o

PB

C O

il D

ropl

et o

n S

ands

tone

in

50%

Aqu

eous

Bio

solv

e

Dro

p #5

Con

tact

Ang

le =

108

.5o

PB

C O

il D

ropl

et o

n S

ands

tone

in

50%

Aqu

eous

Sup

eral

lD

rop

#5 C

onta

ct A

ngle

= 9

0.2o

PB

C O

il D

ropl

et o

n S

ands

tone

in

50%

Aqu

eous

Pet

rote

ch

Dro

p #5

Con

tact

Ang

le =

86.

2o

PB

C O

il D

ropl

et o

n S

ands

tone

in

50%

Aqu

eous

Pet

rocl

ean

D

rop

#5 C

onta

ct A

ngle

= 1

02.2

o

Bio

solv

e >

Pet

rocl

ean

> S

uper

all>

Pet

rote

ch >

Aqu

acle

an

PB

C O

il D

ropl

et o

n S

ands

tone

in

50%

Aqu

eous

Aqu

acle

an

Dro

p #5

Con

tact

Ang

le =

80.

9o

PB

C O

il D

ropl

et o

n S

ands

tone

in

50%

Aqu

eous

Bio

solv

e

Dro

p #5

Con

tact

Ang

le =

108

.5o

PB

C O

il D

ropl

et o

n S

ands

tone

in

50%

Aqu

eous

Sup

eral

lD

rop

#5 C

onta

ct A

ngle

= 9

0.2o

PB

C O

il D

ropl

et o

n S

ands

tone

in

50%

Aqu

eous

Pet

rote

ch

Dro

p #5

Con

tact

Ang

le =

86.

2o

PB

C O

il D

ropl

et o

n S

ands

tone

in

50%

Aqu

eous

Pet

rocl

ean

D

rop

#5 C

onta

ct A

ngle

= 1

02.2

o

Bio

solv

e >

Pet

rocl

ean

> S

uper

all>

Pet

rote

ch >

Aqu

acle

an

Fig

ure

8.7:

Con

tact

Ang

les

for

PB

C O

il o

n S

ands

tone

in 5

0% A

queo

us S

olut

ion.

196

PB

C O

il D

ropl

et o

n S

ands

tone

in

5%

Aqu

eous

Aqu

acle

an

Dro

p #5

Con

tact

Ang

le =

62.

8o

PB

C O

il D

ropl

et o

n S

ands

tone

in

5%

Aqu

eous

Pet

rote

ch

Dro

p #5

Con

tact

Ang

le =

69.

1o

PB

C O

il D

ropl

et o

n S

ands

tone

in

5%

Aqu

eous

Pet

rocl

ean

D

rop

#5 C

onta

ct A

ngle

= 7

6.3o

PB

C O

il D

ropl

et o

n S

ands

tone

in

5%

Aqu

eous

Sup

eral

lD

rop

#5 C

onta

ct A

ngle

= 7

0.1o

PB

C O

il D

ropl

et o

n S

ands

tone

in

5%

Aqu

eous

Bio

solv

e

Dro

p #5

Con

tact

Ang

le =

88.

2o

Bio

solv

e >

Pet

rocl

ean

> S

uper

all>

Pet

rote

ch >

Aqu

acle

an

PB

C O

il D

ropl

et o

n S

ands

tone

in

5%

Aqu

eous

Aqu

acle

an

Dro

p #5

Con

tact

Ang

le =

62.

8o

PB

C O

il D

ropl

et o

n S

ands

tone

in

5%

Aqu

eous

Pet

rote

ch

Dro

p #5

Con

tact

Ang

le =

69.

1o

PB

C O

il D

ropl

et o

n S

ands

tone

in

5%

Aqu

eous

Pet

rocl

ean

D

rop

#5 C

onta

ct A

ngle

= 7

6.3o

PB

C O

il D

ropl

et o

n S

ands

tone

in

5%

Aqu

eous

Sup

eral

lD

rop

#5 C

onta

ct A

ngle

= 7

0.1o

PB

C O

il D

ropl

et o

n S

ands

tone

in

5%

Aqu

eous

Bio

solv

e

Dro

p #5

Con

tact

Ang

le =

88.

2o

Bio

solv

e >

Pet

rocl

ean

> S

uper

all>

Pet

rote

ch >

Aqu

acle

an

Fig

ure

8.8:

Con

tact

Ang

les

for

PB

C O

il o

n S

ands

tone

in 5

% A

queo

us S

olut

ion.

197

Figure 8.9 – Performance and Relative Ranking of SWAs at Four SOR.

0

20

40

60

80

100

120

1:1 SOR 2:1 SOR 4:1 SOR 10:1 SOR

Oil

Rem

oval

, %

Aquaclean

Biosolve

Corexit

Petroclean

Petrotech

Superall

198

Figure 8.10 – SWA Efficiency as a Function of the Summed Difference Between Substrate-SWA and Substrate-Oil Interfacial Tensions at Three SWA Concentrations.

y = -40.168x + 106.93

R2 = 0.9331

0

20

40

60

80

100

120

-0.2 0 0.2 0.4 0.6 0.8 1

Summed IFT Difference

SW

A E

ffici

ency

, %

Figure 8.11 – SWA Efficiency as a Function of the Averaged Difference Between Substrate-SWA and Substrate-Oil Interfacial Tensions at Three SWA Concentrations.

y = -120.48x + 106.92

R2 = 0.9334

0

20

40

60

80

100

120

-0.1 -0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35

Averaged IFT Difference

SW

A E

ffici

ency

, %, %

199

Figure 8.12 – SWA Efficiency at 10:1 SOR as a Function of IFT Difference at Three SWA Concentrations.

y = -25.116x + 88.286

R2 = 0.0032

0

20

40

60

80

100

120

-0.1 -0.09 -0.08 -0.07 -0.06 -0.05 -0.04 -0.03 -0.02 -0.01 0

IFT Difference at 100% SWA Concentration

SW

A E

ffici

ency

at 1

0:1

SO

R, %

y = -230.74x + 84.631

R2 = 0.6446

0

20

40

60

80

100

120

-0.1 -0.08 -0.06 -0.04 -0.02 0 0.02 0.04 0.06


SW

A E

ffici

ency

at 1

0:1

SO

R, %

y = -46.418x + 113.78

R2 = 0.9428

0

20

40

60

80

100

120

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1


SW

A E

ffici

ency

at 1

0:1

SO

R, %

200

Figure 8.13 - SWA Efficiency at 1:1 SOR as a Function of Critical Micelle Concentration.

y = -233.99x + 56.206R2 = 0.82

0

10

20

30

40

50

60

70

80

90

100

0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16

Critical Micelle Concentration of SWA Solution, % Solution

SW

A E

ffici

ency

, %

201

Table 8.1 – Critical Micelle Concentrations for Five Water Soluble SWAs.

Solution CMC

Test #1 (%)

CMC Test #2

(%)

CMC Average

(%) Biosolve 0.0110 0.0113 0.0112

Aquaclean 0.0190 0.0193 0.0192 Petrotech 0.0520 0.0522 0.0521 Petroclean 0.0803 0.0809 0.0806 Superall 0.1521 0.1514 0.1518

Table 8.2 – Measured Densities for Prudhoe Bay Crude Oil and SWA Solutions.

Oil Density (g/cm3) Prudhoe Bay Crude 0.8851

Solution/Concentration 100% 50% 5% Aquaclean 1.0337 1.0164 1.0001 Biosolve 0.9929 0.9955 0.9981

Petroclean 1.0019 0.9997 0.9978 Superall 1.0411 1.0195 0.9999 Petrotech 1.0275 1.0130 0.9997

Corexit 9580 0.8082 N/A N/A

202

Table 8.3 – Oil-Water Interfacial Tension Values for Undiluted SWAs.

Prudhoe Bay Crude Interfacial Tension Values Against 100% (Pure) Products

Drop #

100% Corexit (mN/m)

100% Aquaclean (mN/m)

100% Biosolve (mN/m)

100% Petroclean (mN/m)

100% Superall (mN/m)

100% Petrotech (mN/m)

1 0.031 0.089 0.167 0.238 0.377 0.472 2 0.034 0.088 0.164 0.229 0.383 0.465 3 0.035 0.083 0.164 0.231 0.384 0.470 4 0.034 0.082 0.165 0.230 0.381 0.470 5 0.036 0.088 0.165 0.232 0.378 0.466

Average 0.034 0.086 0.165 0.232 0.381 0.469 Std. Dev. 0.002 0.003 0.001 0.004 0.003 0.003

Table 8.4 – Oil-Water Interfacial Tension Values for 50% Aqueous SWA Solutions.

Prudhoe Bay Crude Interfacial Tension Values Against 50%wt Aqueous Solutions

Drop #

50% Aquaclean (mN/m)

50% Biosolve (mN/m)


50% Superall (mN/m)

50% Petrotech (mN/m)

1 0.154 0.268 0.365 0.574 0.691 2 0.154 0.265 0.372 0.577 0.687 3 0.148 0.269 0.369 0.570 0.685 4 0.150 0.264 0.365 0.575 0.688 5 0.150 0.273 0.373 0.576 0.686

Average 0.151 0.268 0.369 0.574 0.687 Std. Dev. 0.003 0.004 0.004 0.003 0.002

Table 8.5 – Oil-Water Interfacial Tension Values for 5% Aqueous SWA Solutions.

Prudhoe Bay Crude Interfacial Tension Values Against 5%wt Aqueous Solutions

Drop #

5% Aquaclean (mN/m)

5% Biosolve (mN/m)


5% Superall (mN/m)

5% Petrotech (mN/m)

1 0.944 1.374 1.680 2.203 2.479 2 0.950 1.374 1.670 2.203 2.479 3 0.949 1.372 1.677 2.204 2.485 4 0.950 1.370 1.676 2.204 2.479 5 0.953 1.375 1.672 2.204 2.482

Average 0.949 1.373 1.675 2.204 2.481 Std. Dev. 0.003 0.002 0.004 0.001 0.003

203

Table 8.6 – Contact Angles for PBC Oil Submerged in Undiluted SWA.

Prudhoe Bay Crude Contact Angles on Sandstone Submerged in 100% Pure SWAs

Drop #

100% Aquaclean

(deg)

100% Biosolve

(deg)

100% Petroclean

(deg)

100% Superall

(deg)

100% Petrotech

(deg)

100% Corexit (deg)

1 92.2 123.2 112.0 104.1 98.6 74.7 2 93.2 123.2 110.8 103.4 98.2 75.7 3 92.2 123.2 112.4 102.0 98.5 75.1 4 92.8 122.7 111.5 103.3 98.8 74.4 5 92.1 124.8 110.5 104.1 98.9 76.0

Average 92.5 123.4 111.4 103.4 98.6 75.2 Std. Dev. 0.5 0.8 0.8 0.9 0.3 0.7

Table 8.7 – Contact Angles for PBC Oil Submerged in 50% Aqueous SWA Solutions.

Prudhoe Bay Crude Contact Angles on Sandstone Submerged in 50%wt Aqueous Solutions

Drop #

50% Aquaclean

(deg)

50% Biosolve

(deg)

50% Petroclean

(deg)

50% Superall

(deg)

50% Petrotech

(deg)

1 82.2 109.3 101.9 90.5 87.1 2 83.1 107.5 103.4 90.4 86.6 3 82.1 108.6 101.9 91.7 86.6 4 81.1 107.2 103.1 92.1 88.0 5 80.9 108.5 102.2 90.2 86.2

Average 81.9 108.2 102.5 91.0 86.9 Std. Dev. 0.9 0.9 0.7 0.9 0.7

Table 8.8 – Contact Angles for PBC Oil Submerged in 5% Aqueous SWA Solutions.

Prudhoe Bay Crude Contact Angles on Sandstone Submerged in 5%wt Aqueous Solutions

Drop #

5% Aquaclean

(deg)

5% Biosolve

(deg)

5% Petroclean

(deg)

5% Superall

(deg)

5% Petrotech

(deg)

1 62.7 88.1 76.3 69.4 68.0 2 63.0 87.0 77.1 70.3 67.0 3 64.2 87.1 76.1 70.6 67.7 4 63.6 87.1 76.0 69.5 68.1 5 62.8 88.2 76.3 70.1 69.1

Average 63.3 87.5 76.4 70.0 68.0 Std. Dev. 0.6 0.6 0.4 0.5 0.8

204

Table 8.9 – Difference Between Substrate-SWA and Substrate-PBC Interfacial Tensions.

Interfacial Tension Difference σSC - σSO

Solution

Aquaclean IFT Diff. (mN/m)

Biosolve IFT Diff. (mN/m)

Petroclean IFT Diff (mN/m)

Superall IFT Diff (mN/m)

Petrotech IFT Diff (mN/m)

Corexit IFT Diff (mN/m)

100% Product -0.004 -0.091 -0.085 -0.088 -0.070 0.009

50% Aqueous 0.021 -0.084 -0.080 -0.010 0.037

5% Aqueous 0.426 0.060 0.394 0.754 0.929

205

Chapter 9

CONCLUSIONS

9.1 PROTOCOL COMPARISON

Research was conducted to develop a laboratory testing protocol for evaluating effectiveness of

surface washing agents (SWAs) to remove crude oil from a solid substrate. There are several

SWA effectiveness protocols described in the literature, including Environment Canada’s

Inclined Trough Test and SAIC’s Swirling Coupon Test developed for U.S. EPA (Clayton et al.

1995). A major shortcoming of these protocols was the lack of similarity of the test substrate to

natural shoreline substrates. The Inclined Trough Test used stainless steel and porcelain tile

troughs, while the Swirling Coupon Test used stainless steel and porcelain coupons. Stainless

steel and porcelain are smooth surfaces with low surface area and should be easier to clean than

higher surface area substrates, such as shoreline sand or gravel. Other disadvantages to the

Inclined Trough Test included the following:

• No mixing energy was applied during the washing step to facilitate the removal of oil that

had been released from the substrate surface;

• A short weathering time resulted in high release of oil from untreated controls;

• A potential source of error was introduced by blotting the troughs with the tip of a paper

towel after treatment and washing;

• Oil released from the surface was determined gravimetrically based on the weight of the

substrate before and after treatment; this method was insensitive to small weight

differences and could be biased by the weight of the water that adheres to the surface

after washing.

206

In the current study, two prototype experimental designs were proposed that improved upon the

designs set forth by Environment Canada and SAIC. The first design that was tested extensively

was the Glass Plate Protocol. Glass was proposed as the substrate because it is made from silica

sand and has physical and chemical properties similar to those of natural shoreline substrates. In

addition, its surface can be easily modified to study the effect of surface roughness and

hydrophobicity on SWA performance. Limitations to this design were noted as follows:

• Most SWAs were water soluble and tended to bead up rather than spread across the oiled

surface; therefore, a large volume of SWA solution was required to achieve complete

coverage of the oiled area;

• The plates were rotated in a single direction around a central axis during washing; this

washing regime did not simulate wave action, and the flow of water across the plate was

not uniform due to unidirectional rotation;

• The grinding of the plate surface with grinding powders could not be standardized.

This protocol was ultimately abandoned due to high variability in data, lack of reproducibility

and repeatability, and concerns over standardizing the surface grinding.

The second design that was tested and ultimately proposed as the final protocol was the Natural

Substrate Protocol. Silica sand served as the representative shoreline substrate and PBC oil as

the representative oil for all protocol development experiments. Gravel and IFO-180 (a heavy

oil) were also tested and will be included in the final protocol. Mass balances on oil distributed

between aqueous and solid phases were typically greater than 90%, indicating good extraction

efficiencies and appropriate analytical methodologies. The test was robust with respect to

207

reproducibility and repeatability. Details of the development and validation of this protocol are

summarized in Section 9.2.

Table 9.1 presents data collected under the Inclined Trough Test, Swirling Coupon Test, Glass

Plate Protocol, and Natural Substrate Protocol for Corexit 9580 and PBC oil (the only SWA and

oil tested under all 4 protocols). Corexit achieved the highest release of PBC from the substrate

under the Inclined Trough Test. However, untreated controls under the same test also had high

release of PBC, so the overall effectiveness due to Corexit was 32-50%. The Swirling Coupon

Test had low release of PBC from controls and the greatest overall effectiveness for Corexit (55-

63%). Overall effectiveness for Corexit under the Glass Plate Protocol ranged from 40-50%.

RSDs among replicates were highest for the Swirling Coupon Test and Glass Plate Protocol and

indicate poor repeatability for these tests. Overall effectiveness under the Natural Substrate

Protocol was lower than for the other tests, but the RSDs were also lower indicating better

repeatability under this protocol. Differences in cleaning effectiveness between artificial

substrates, such as stainless steel or porcelain, and natural substrates could result from

differences in wetting and adhesion properties of oil to different types of substrates. Thus, the

effectiveness of a SWA in the field will be best predicted using oil and substrate from the

impacted site. When possible, it is desirable to perform on-site field testing of SWAs using the

specific oiled substrates under spill-specific conditions.

9.2 TESTING OF THE NATURAL SUBSTRATE PROTOCOL

Under the Natural Substrate Protocol, variables were tested at multiple levels to determine their

effects on SWA performance. The protocol was most sensitive to SWA-to-oil ratio (SOR) and

208

mixing speed (MS). SWA effectiveness increased with SOR for ratios ranging from 1:2.5 to

10:1, and the relative ranking of SWAs changed as SOR increased. Multiple oil volumes were

tested at each SOR, but no trend was observed with changes in oil volume. Diluting the SWA

did not affect performance as long as the total mass of applied SWA was constant. Oil removals

from controls and treatments increased with rotational MS. The maximum difference between

treatments and controls occurred at 150 rpm, while 200 rpm was excessively rigorous and 100

rpm yielded low release from treatments and controls. Differences in oil release from treatments

and controls were not strongly linked to mixing time (MT).

Oil was applied to wet and dry sand and gravel to determine the effect of substrate type and

hydration. Statistical differences were found between sand and gravel substrates, but the relative

significance of all other variables was not affected by substrate type. Two substrates, sand and

gravel, will be included in the final protocol. Drain time (DT) had no effect. The decision to use

dry substrates was made based on high replicate RSD and high release of oil from controls in wet

substrate treatments. Weathering time (WT) was set at 18 hr to reduce oil release from controls

and maximize the difference between treatments and controls. SWA performance was

insensitive to oil-SWA contact time (CT).

A resolution 5, 35-1 fractional factorial experiment was designed to determine the factors that

contribute to the performance of SWAs in removing crude oil from surfaces such as sand and

gravel. The factors considered were MS, MT, CT, SWA solution concentration (SC), and SOR.

Main effects were considered for the five factors. CT was the only factor that was non-

significant. All other factors were found to positively and significantly affect response.

209

However, ceiling effects were observed for MT and CT, with equivalent responses for the middle

and high levels. Levels over 15 min MT and 30 min CT did not improve the response. A

positive correlation was seen for MS and SOR; response increased with level for these factors.

Additional testing at higher SOR can be done. However, excessively rigorous mixing can

become experimentally impractical, so MS greater than 175 rpm is not recommended.

Substrate was also found to be statistically significant. Analyses were performed for each of the

substrates separately and for the two substrates combined. The separate analyses yielded the

same conclusions as the combined model, so the results from the combined model were

presented here.

Two-way interactions were also considered for each combination of the five factors. CS was not

significant as a single-factor and it was not responsible for any significant two-way interactions.

While the CS-MT interaction was significant, it had a ceiling effect and no consistent trend

across the levels. Therefore SC was determined not to be a significant factor in surface washing.

SOR had a statistically significant interaction with MS, and at the highest level their combined

effect was greater than either one alone. However, at the lowest level, a diminishing effect

occurred. The lowest level is not recommended for SOR or MS. CT had a ceiling effect at 30

min and no significant two-way interactions. A ceiling effect was also observed for MT at 15

min, and a positive interaction was seen for MT and SOR only at the highest level. Thus an

argument could be made for using either the 15 min or 30 min MT. The MT-MS interaction was

significant with positive reinforcement occurring at the middle and high levels. An increasing

positive response was observed for MS, but mixing speeds greater than 175 rpm are not practical

210

experimentally. Thus, the following recommendations can be made based on the fractional

factorial analysis:

SC Recommendation: follow manufacturers’ instructions for dilution Reason: not a significant factor in surface washing MT Recommendation: 15 min Reason: a ceiling effect was observed at 15 min; no benefit gained by 30 min CT Recommendation: 30 min Reason: A ceiling effect observed at 30 min; no benefit gained by 45 min MS Recommendation: 150 rpm Reason: Negative interaction with MT at 125 rpm; 175 rpm too rigorous

SOR Recommendation: 2:1 or higher; appropriateness of SOR for field application depends on cost and toxicity

Reason: Increasing positive response with level; test at higher SOR. 9.3 INTERFACIAL ACTIVITY OF SWAs

A SWA’s ability to lower interfacial tension is an important aspect of surface cleaning in that

low interfacial tension means less energy is needed to emulsify crude oil in the cleaning solution.

However, a look at the relative abilities of the cleaners to hold off or displace oil from a

sandstone surface reveals a different order of predicted cleaning effectiveness for the cleaners.

This is common in hard surface cleaning, with the best cleaners being designed to both have low

interfacial tension and the ability to displace the boundary layer of oil from a surface. The two

properties are independent to some extent and also surface dependent.

In general, the best cleaner for a particular oil will be one that matches the oil in terms of the

hydrophilic/lipophilic balance. A surfactant with hydrophobic tail lengths similar to the chain

lengths in the oil will promote packing of the surfactant at the oil-water interface and optimize its

211

cleaning effectiveness. For high oil loads, the OW interfacial tension is key because oil is being

emulsified away from itself rather than from the substrate. For removing the last layer of oil

from the substrate surface, contact angle and competitive wetting become more important.

In these studies, the best predictor of SWA performance at high SOR was the term representing

the difference between SWA-substrate and oil-substrate interfacial tensions (σSC - σSO). This

term is a measure of the preferential wetting of the substrate surface by either the oil or the wash

water. It was calculated from measured values for interfacial tension and contact angle using the

Young equation. A high correlation was observed between SWA efficiency at 10:1 SOR and the

σSC - σSO difference at 5% SWA solution concentration. Using the σSC - σSO difference at the

lowest concentration level is logical because the SWA in the final wash water is less than 1%.

The correlation applies to high SOR, indicating that SWA concentrations many times greater

than the critical micelle concentration (CMC) must be used for optimal SWA performance. At

low SOR, CMC was a better predictor of SWA performance.

9.4 RECOMMENDATIONS FOR FUTURE WORK

The next step in the development of the effectiveness protocol will be to evaluate protocol

sensitivity to operator error through round robin testing. At least three independent laboratories

will perform the protocol using specified levels for all variables. Variation across operators

(reproducibility) and within replicate samples by a single operator (repeatability) will be

evaluated. If the protocol is determined to be reproducible and insensitive to the operator

conducting the test, then variables can be fixed, a final protocol established, and the procedure

recommended to U.S. EPA Headquarters for adoption as an Agency standard. Acceptance

212

criteria must also be established and should be based on statistically significant differences in

performance between the SWA treatment and the untreated controls.

To date, the development and testing of the effectiveness protocol has been performed using

PBC as a representative oil. Additional testing of the protocol needs to be done using a heavy

weight fuel oil, such as Bunker C or IFO-180. The fractional factorial experiment should be

repeated using one of these oils. SC can be fixed in the factorial design because it was

determined to be a non-significant factor in surface washing. Other variables can be tested at

multiple levels as before and their relative significance determined.

Additionally, testing should be done to evaluate the effect of salinity and temperature on SWA

performance. These are expected to be significant variables affecting on-scene response efforts.

Other factors for consideration include 1) the pH of the wash water, 2) the hydrophilic-lipophilic

balance of the oil and surfactants, 3) the acid/base composition of the oil, and 4) the relative

amount of polar/non-polar fractions in the oil. Substrates such as rip rap and natural vegetation

should also be tested under the protocol. Additional SWAs listed on the NCP Product Schedule

may also be selected for testing.

213

Table 9.1 – Comparison of Testing Protocols Using Corexit 9580 and PBC Oil.

Test

Substrate

SOR

% of Total Oil Released Into Wash Water Overall

Performance (%)

Corexit 9580 Controls

Mean Stdev RSD Mean Stdev RSD

Inclined Trough Steel 1:5

76.3 4.9 6 44.1 8.8 20 32.2 Tile 88.5 9.0 10 37.9 8.1 21 50.6

Swirling Coupon Steel 1:3

57.6 11.2 19 2.2 0.6 27 55.4 Tile 65.9 11.9 18 3.4 2.1 61 62.5

Glass Plate Smooth 1:1 67.9 12.6 19 18.2 2.5 14 49.7 Fine 48.7 6.4 13 0.4 0.1 24 48.3

Coarse 41.4 4.6 11 1.3 0.2 15 40.1 Natural Substrate Dry Sand 1:1 35.5 0.6 2 4.1 0.8 20 31.4

Wet Sand 45.7 2.1 5 22.2 2.8 13 23.5

214

REFERENCES

Adamson, A.W. and A.P. Gast. 1997. Physical Chemistry of Surfaces. Sixth Edition. Wiley-

Interscience Publication, John Wiley & Sons, Inc. 784 p.

Alaska Department of Environmental Conservation. 1993. The Exxon Valdez Oil Spill: Final

Report, State of Alaska Response (Chapter 2: Technology, pp 61-87).

Aveyard, R., B.P. Binks, S. Clark and J. Mead. 1986. J. Chem. Soc. Faraday Trans. 1, 82, 125.

Aveyard, R., B.P. Binks, T.A. Lawless and J. Mead. 1985. J. Chem. Soc. Faraday Trans. 1, 81,

2155.

Bendure, R.L. Dynamic Adhesion Tension Measurement. 1973. Journal of Colloid Interface

Science, 42(1), 137-144.

Box,G. E, Hunter,W.G., Hunter, J.S., Hunter,W.G., 2005. Statistics for Experimenters: Design,

Innovation, and Discovery, 2nd Edition, Wiley.

Buckley, J. S. 1994. Chemistry of the Crude Oil/ Brine Interface. Proceedings of the 3rd

International Symposium on Evaluation of Reservoir Wettability and its Effect on Oil

Recovery, 33-38.

Buckley, J.S., Liu, Y., Monsterleet, S., 1998. Mechanism of wetting alteration by crude oils.

SPEJ 3, 54– 61.

Buckley, J.S., Liu, Y., Xie, X., Morrow, N.R., 1997. Asphaltenes and crude oil wetting—the

effect of oil composition. SPEJ 2, 107–122.

215

Buckley, J.S., Takamura, K., Morrow, N.R., 1989. Influence of electrical surface charges on the

wetting properties of crude oils. SPE Reservoir Eval. 4, 332–340.

Burden, R.L., J.D. Faires. 2000. Numerical Analysis, 7th Ed., Brooks/Cole.

Clayton, J.R., Jr., B.C. Stransky, M.J. Schwartz, D.C. Lees, J. Michel, B.J. Snyder, and A.C.

Adkins. 1995. Development of Protocols for Testing Cleaning Effectiveness and Toxicity of

Shoreline Cleaning Agents in the Field. Marine Spill Response Corporation, Washington,

D.C. MSRC Technical Report Series 95-020.1, 180 p.

Clayton, J.R., Jr., J. Michel, B.J. Snyder, and D.C. Lees. 1995. Utility of Current Shoreline

Cleaning Agent Tests in Field Testing. Marine Spill Response Corporation, Washington,

D.C. MSRC Technical Report Series 95-004, 53 p.

Clayton, J.R., S.F. Tsang, V. Frank, P. Marsden, N. Chau, and J. Harrington. 1993. Chemical

Shoreline Cleaning Agents: Evaluation of Two Laboratory Procedures for Estimating

Performance. Science Applications International Corporation, San Diego, CA, EPA Report

No. EPA/600/R-93/113a, 137 p.

Clayton, Jr., J. Chemical Shoreline Cleaning Agents for Oil Spills - Update State-of-the-Art on

Mechanisms of Action and Factors Influencing Performance. U.S. Environmental Protection

Agency, Washington, D.C., EPA/600/R-93/113b.

Craig Jr., F.F., 1971. The Reservoir Engineering Aspects of Waterflooding. Am. Inst. Min.

Metall. and Pet. Engrs., New York.

DeLaune, R.D., S.R. Pezeshki, and J.A. Nyman. 1993. Investigation of Corexit 9580 for

Removing Oil from Marsh Grass. Exxon Research and Engineering, Co., Florham Park, NJ,

36 p.

Draper, N.R. and Smith, H. (1998). Applied Regression Analysis. Wiley Series in Probability

and Statistics.

216

Drummond, C., J. Israelachvili. Surface Forces and Wettability. Journal of Petroleum Science

and Engineering. 33 (2002) 123– 133.

Fair, Geyer and Okun. 1968. Water and Wastewater Engineering: Volume 2. Water Purification

and Wastewater Treatment and Disposal. John Wiley & Sons, Inc. New York.

Fingas, M.F., R. Stoodley, and N. Laroche. 1990. Effectiveness Testing of Spill-Treating

Agents. Oil & Chemical Pollution, 7, 337-348.

Fiocco, R.J. et al. 1991. Development of Corexit 9580:A chemical beach cleaner. Proceedings

of the 1991 International Oil Spill Conference, March 4-7, 1991, San Diego, California,

American Petroleum Institute, Washington, D.C. pp. 395-400.

Gibbs, Josiah Willard. On the Equilibrium of Heterogeneous Substances. 1876.

Gonzalez, M.B., R. Salas, N. Pereira, and H. Briceno. 1996. Effect of an Oil Spill on Fungal

Biomass in an Estuarine Mangrove of Lake Maracaibo, Venezuela, and it’s Bioremediation.

International Biodeterioration and Biodegradation, 37(1-2), 130-136.

Haines, J.R., E.J. Kleiner, K.A. McClellan, K.M. Koran, E.L. Holder, D.W. King and A.D.

Venosa. "Laboratory Evaluation of Oil Spill Bioremediation Products in Salt and Freshwater

Systems." Journal of Industrial Microbiology. 2005. 32. 171-185.

Haines, J.R., K.M. Koran, E.L. Holder and A.D. Venosa. "Protocol for Laboratory Testing of

Crude Oil Bioremediation Products in Freshwater Conditions." Journal of Industrial

Microbiology. 2003. 30. 107-113.

Hoeiland, S., T. Barth, A.M. Blokhus, A. Skauge. 2001. The Effect of Crude Oil Acid

Fractions on Wettability as Studied by Interfacial Tension and Contact Angles. Journal of

Petroleum Science & Engineering, 30: 91-103.

217

Karsa, D.R. 2000. Surface Active Behaviour of Performance Surfactants: Annual Surfactants

Review. Volume 3. Sheffield Academic Press. 205 p.

Koran, K.M., A.D. Venosa, and S.J. Rodriguez. “A New Screening Protocol for Testing the

Effectiveness of Surface Washing Agents.” Proceedings of Interspill 2006 Conference.

London, England.

Koran, K.M., A.D. Venosa, S. Vyas, and G.A. Sorial. “Development of a Surface Washing

Agent Effectiveness Protocol.” Proceedings of 2005 International Oil Spill Conference.

Miami Beach, FL.

Kowalewski, E., T. Holt, O. Torsaeter. Wettability alterations due to an oil soluble additive.

Journal of Petroleum Science and Engineering, 33 (2002) 19– 28.

Laplace, Pierre Simon. Mécanique Céleste, Supplement to the tenth edition, 1806

Le Floch, S., F.X. Merlin, M. Guillerme, C. Dalmazzone, and P. Le Corre. 1999. A Field

Experimentation on Bioremediation: BIOREN. Environmental Technology, 20(8), 897-907.

Lee, K. and F.X. Merlin. 1999. Bioremediation of Oil on Shoreline Environments:

Development of Techniques and Guidelines. Pure and Applied Chemistry, 71(1), 161-171.

Lee, K. and S. de Mora. 1999. In Situ Bioremediation Strategies for Oiled Shoreline

Environments. Environmental Technology, 20(8), 783-794.

Liu, L., J.S. Buckley. “Alteration of wetting of mica surfaces.” Journal of Petroleum Science

and Engineering, 24. 1999. 75–83.

Merlin, F.X. and P. Le Guerroue. 1994. The New French Approval Procedure for Shoreline

Cleaning Agents. Proceedings of the Seventh Arctic Marine Oil Spill Program Technical

Seminar. Environment Canada, Ottawa, Ontario. pp. 943-949.

218

Michel, J. and B.L. Benggio, 1994, Testing and use of shoreline cleaning agents during the

MORRIS J. BERMAN oil spill, Proceedings of 1995 International Oil Spill Conf., API Publ.

No. 4620, American Petroleum Institute, Wash., D.C., pp. 197-202.

Michel, J. and J.A. Galt, 1995, Conditions under which floating slicks can sink in marine

settings. Proceedings of1995 International Oil Spill Conference, API Publ. No. 4620,

American Petroleum Institute, Wash., D.C. pp. 573-576.

National Research Council, Marine Board. Free Executive Summary: Using Oil Spill Disersants

on the Sea. Committee on Effectiveness of Oil Spill Dispersants. 1989.

Owens, E.O., R.A. Davis, J. Michel, and K. Stritzke. 1995. Beach Cleaning and the Role of

Technical Support in the 1993 Tampa Bay Spill. Proceedings of the 1995 International Oil

Spill Conference, February 27-March 2, Long Beach, CA. American Petroleum Institute,

Washington D.C., 627-634.

Pezeshki, S.R., A. Jugsujinda, and R.D. Delaune. 1998. Responses of Selected U.S. Gulf Coast

Marsh Macrophyte Species to Oiling and Commercial Cleaners. Water Air and Soil

Pollution, 107(1-4), 185-195.

Pezeshki, S.R., R.D. Delaune, and A. Jugsujinda. 2001. The Effects of Crude Oil and the

Effectiveness of Cleaner Application Following Oiling on U.S. Gulf of Mexico Coastal

Marsh Plants. Environmental Pollution, 112(3), 483-489.

Prince, R.C., R. Varadaraj, R.J. Fiocco, and R.R. Lessard. 1999. Bioremediation as an Oil Spill

Response Tool. Environmental Technology, 20(8), 891-896.

Rosen, M. J. 1978. Surfactants and Interfacial Phenomena. New York, John Wiley & Sons.

SAS/STAT User's Guide, volume 2,'' version 6 fourth edition, SAS Institute (1990)

219

Sorial, G.A., Venosa, A.D., Koran, K.M., Holder, E., and King, D. "Oil Spill Dispersant

Effectiveness Protocol - Part I Impact of Operational Variables." ASCE Journal of

Environmental Engineering. 2004. 130(10). 1073-1084.

Sorial, G.A., Venosa, A.D., Koran, K.M., Holder, E., and King, D. "Oil Spill Dispersant

Effectiveness Protocol - Part II Performance of the Revised Protocol." ASCE Journal of

Environmental Engineering. 2004. 130(10). 1085-1093.

Standal, S. H., A.M. Blokhus, J. Haavik, A. Skauge, T. Barth. 1999. Partition Coefficients and

Interfacial Activity for Polar Components in Oil/ Water Model Systems. Journal of Colloid

and Interface Science, 212: 33-41.

Standal, S., J. Haavik, A.M. Blokhus, A. Skauge. 1999. Effect of Polar Organic Compounds on

Wettability as Studied by Adsorption and Contact Angles. Journal of Petroleum Science &

Engineering, 24: 131-144.

Teas, H.J., R.R. Lessard, G.P. Canevari, C.D. Brown, and R. Glenn. 1993. Saving Oiled

Mangroves Using a New Non-Dispersing Shoreline Cleaner. Proceedings of the 1993 Oil

Spill Conference, March 29-April 1, Tampa, FL. Washington, D.C., American Petroleum

Institute, 146-151.

Thompson, L. 1994. The Role of Oil Detachment Mechanisms in Determining Optimum

Detergency Conditions. Journal of Colloid and Interface Science, 163, 61-73.

Tumeo, M.A. and A. Cote. 1998. Effect of Flush Water Temperature and Type on Commercial

Shoreline Cleaning Agent Efficiency. Journal of Soil Contamination, 7(2), 213-226.

U.S. EPA, 1994. National Oil and Hazardous Substances Pollution Contingency Plan, Subpart

J, Section 300.310, Phase III.

U.S. EPA. 1999. Understanding Oil Spills And Oil Spill Response: Shoreline Cleanup of Oil

Spills (Chapter 4). EPA 540-K-99-007. OSWER 9200.5-104A.

220

U.S. EPA. Region IV. Shoreline Cleaner Test and Evaluation Protocol. 2002.

Venosa, A.D., D.W. King, and G.A. Sorial. 2002. The Baffled Flask Test for Dispersant

Effectiveness: A Round Robin Evaluation of Reproducibility and Repeatability. Spill Sci. &

Technol. Bulletin, 7(5-6), 299-308.

Walker, A.H., J. Michel, G. Canevari, J. Kucklick, D. Scholz, C.A. Benson, E. Overton, and B.

Shane. 1993. Chemical Oil Spill Treating Agents. Marine Spill Response Corporation,

Washington, D.C. MSRC Technical Report Series 93-015, 328 p.

Walker, A.H., J.H. Kucklick, and J. Michel. 1999. Effectiveness and Environmental

Considerations for Non-Dispersant Chemical Countermeasures. Pure and Applied

Chemistry, 71(1), 67-81.

Walker, A.H., J.H. Kucklick, J. Michel, D.K. Scholz, and T. Reilly, 1994, Chemical treating

agents: response niches and research and developments needs. Proceedings of 1995

International Oil Spill Conf., API Publ. No. 4620, American Petroleum Institute, Wash., D.C.

pp. 211-217.

Weise, A.M., C. Nalewajko, and K. Lee. 1999. Oil-Mineral Fine Interactions Facilitate Oil

Biodegradation in Seawater. Environmental Technology, 20(8), 811-824.

Wu, C. F. and M. Hamada. (2000) Experiments: Planning, Analysis, and Parameter Design

Optimization, New York: Wiley

Yang, S.-Y., G.J. Hirasaki, S. Basu, R. Vaidya. Statistical Analysis on Parameters that Affect

Wetting for the Crude Oil/Brine/Mica System. Journal of Petroleum Science and

Engineering, 33 (2002) 203– 215.

Young, Thomas. An Essay on the Cohesion of Fluids. Philosophical Transactions of the Royal

Society of London. 1805. p. 65.

Documents

Surface Interactions of Surface Washing Agentshomepages.uc.edu/~sorialga/Karen PhD Dissertation.pdf · Interfacial tension and contact angle were measured for PBC in the presence