SOIL REMEDIATION BY SUPERCRITICAL CO EXTRACTION …Soil remediation by supercritical CO 2 extraction: process performance evaluation and cost analysis ACKNOWLEDGMENTS Many people have

SOIL REMEDIATION BY

SUPERCRITICAL CO2 EXTRACTION

Process Performance Evaluation

and Cost Analysis

Laura L. Bretti

Master of Science

Environmental Engineering & Sustainable Infrastructures

Royal Institute of Technology

Stockholm, Sweden

Soil remediation by supercritical CO2 extraction: process performance evaluation and cost analysis

ACKNOWLEDGMENTS Many people have contributed to the present thesis in many different ways, to a

smaller or larger extent, and I am very thankful towards all of them. I especially want to thank my supervisors Bengt Espeby and Tonie Wickman of the

Land and Water Resources Div., Civil and Environmental Engineering Dept., of the Royal Institute of Technology (KTH), Stockholm, Sweden. I would like to further thank Tonie for closely assisting me during the thesis writing and presentation.

I would also like to thank all the people who helped me at the Polytechnic Institute of

Turin, Italy, where all the experimental work was carried out. In particular, Prof. Mariachiara Zanetti for her supervision, and Ada Ferri for her invaluable help, with the actual experiments and analyses as well as morally.

And thanks also to Mattia Cattaneo, my family and my closest friends for their

extraordinary support throughout the whole period. Thank you everybody!

2

Table of contents

ABSTRACT

Aim of the present research is to investigate the application of supercritical CO2 to

naphthalene removal from soils, the final target being the cost estimation of a full-scale plant. The results and the performances obtained are also analysed in order to determine what errors might affect the data obtained by Supercritical Fluid Extraction (SFE) with the laboratory equipment used and the procedure followed. At last, an attempt is made to understand what molecular processes stand behind the relationships between the system parameters and the performances recorded.

The analysis of the results shows an extraction time (defined as time required for the

system to reach the asymptote) rather constant as a function of the operating conditions (pressure, temperature and flow). The extraction efficiency, on the other hand, is found to increase at higher CO2 flows and lower temperatures, while no dependence is recorded from the pressure nor the density.

The relationship between the extraction efficiency and the system influencing

parameters (temperature and flow) is used for a full-scale plant dimensioning, the minimisation of the costs and the cost analysis. The data obtained indicates the CO2-SFE as a highly competitive soil remediation method, compared to other “traditional” methods, with costs as low as 133 $/m3.

The error analysis indicates that the results achieved are precise but not very accurate,

due to the soil contamination procedure followed. The data interpretation suggests that the film transfer resistance is the extraction rate

limiting factor in the flow range considered. At last, the relationships between the extraction efficiency and the operating conditions, together with the non-achievement of 100% efficiency, seems to confirm some authors’ theory on the resistance to desorption due to the presence in the soil of “resistant” and slowly desorbing fractions, constituted by immobile fluid in pores.

3


TABLE OF CONTENTS

1. INTRODUCTION............................................................................................................... 5 1.1. SOIL CONTAMINATION AND PAH .................................................................................. 5 1.2. SOIL DECONTAMINATION METHODS.............................................................................. 6 1.3. SUPERCRITICAL FLUIDS AND SUPERCRITICAL FLUID EXTRACTION................................ 7 1.4. THE STATE OF THE ART ................................................................................................. 9

2. ANALYTICAL INSTRUMENTATION, MATERIALS AND METHOD ................. 11 2.1. ANALYTICAL INSTRUMENTATION ................................................................................ 11 2.2. MATERIALS.................................................................................................................. 11 2.3. METHOD ...................................................................................................................... 13

2.3.1. Pilot plant............................................................................................................ 13 2.3.2. Research layout................................................................................................... 17

3. GAS CHROMATOGRAPH CALIBRATION............................................................... 22

4. THE REPEATABILITY TEST....................................................................................... 25

5. RESULTS .......................................................................................................................... 29

6. ELABORATION OF THE RESULTS ........................................................................... 40 6.1. COMPARISON OF THE RESULTS AT DIFFERENT OPERATING CONDITIONS ..................... 40

6.1.1. Comparison of the extraction results at constant temperature........................... 40 6.1.2. Comparison of the extraction results at constant mass flow .............................. 43 6.1.3. Discussion on the comparisons performed ......................................................... 45

6.2. EXTRACTION EFFICIENCY AT DIFFERENT OPERATING CONDITIONS............................. 46 6.2.1. Extraction efficiency as a function of the operating conditions.......................... 46 6.2.2. Extrapolations..................................................................................................... 49

7. ERROR ANALYSIS AND EVALUATION ................................................................... 53 7.1. ERROR GENERATION .................................................................................................... 53

7.1.1. Soil preparation and contamination ................................................................... 54 7.1.2. Extraction run ..................................................................................................... 56 7.1.3. Sample analyses .................................................................................................. 57

7.2. ERRORS EVALUATION .................................................................................................. 58 7.2.1. External errors.................................................................................................... 58 7.2.2. Intrinsic errors .................................................................................................... 59

8. DATA INTERPRETATION............................................................................................ 63 8.1. EXTRACTION RATE ...................................................................................................... 63 8.2. EXTRACTION EFFICIENCY ............................................................................................ 66

9. COSTS ............................................................................................................................... 67 9.1. DESIGN PARAMETERS .................................................................................................. 67 9.2. VARIABLE COSTS......................................................................................................... 68 9.3. FIXED COSTS................................................................................................................ 69 9.4. TOTAL COSTS............................................................................................................... 70

10. CONCLUSIONS ............................................................................................................... 71

LITERATURE REFERENCES .............................................................................................. 72

4

Introduction

1. INTRODUCTION

1.1. SOIL CONTAMINATION AND PAH Several sites around the world are contaminated by a number of pollutants, brought

there either by a continuous process of contamination that might have taken place throughout many years, or by a sudden dispersion caused by an accident (i.e. oil spilling from a gas station due to a pipe breakage).

The contaminants that can be found in soils are many, inorganic, like heavy metals, as well as organic. Among the organic contaminants, oil and coal refineries are responsible for several cases of soil contamination with Polycyclic Aromatic Hydrocarbons (PAHs). The PAHs are a family of compounds formed by two or more aromatic rings of carbon atoms linked together; the basic aromatic structure is that of benzene (Fig.1.1). Among the PAHs, naphthalene is the simplest molecule, formed by two rings only.

C

C

C

C

C

C

H H

H

H H

H C

C

C

C

H

H H

H

C

C

C

C

C

C

H H

H H

BENZENE (C6H6) NAPHTHALENE (C10H8)

Fig.1.1. Molecular structures of benzene and naphthalene.

Case studies on soil contamination by PAHs show great differences in the actual

concentrations that can be found at specific locations. The concentration might be as low as few ppm to several hundreds or thousands ppm. Contamination at residential sites due to stationary sources (i.e. as a result of a continuous process of hydrocarbon deposition in the surroundings of congested highways or major roads) is usually of the order of magnitude of few ppm, below the maximum acceptable values. On the contrary, heavy contamination is often recorded at industrial sites in case of handling of hydrocarbon mixtures, i.e. oil stocking locations, manufacture of town gas from coal, coke and oil, etc. Studies from many authors [1, 2, 5, 10] refer of contamination by PAHs at former gasworks plant areas with concentration values from tens or hundreds of ppm up to 30.000 ppm at a site near Bedford, IN [10].

5


In some cases, the concentration of the pollutants in the soil might be so high to represent a harm to the vegetation, animals and humans. Limits have been set on the concentration of most contaminants, and growing concern is focusing on soil contamination issues. Limits are set for PAHs as well, as some of them are known to be carcinogenic [9].

USA regulations set the maximum acceptable level of contamination by PAHs equal to 40 ppm at residential sites and 270 ppm at industrial sites. Stricter values are usually adopted in European Countries. The Dutch regulations, for instance, set the limit values of 40 ppm at residential sites and 250 ppm at industrial sites. Similar limit values can be found in the other European Countries.

1.2. SOIL DECONTAMINATION METHODS Remediation of contaminated soils can be carried out according to a number of

methods. Usually, those methods are already long established, and they have been taken from other remediation fields and adapted to soil decontamination with more or less good results. The most important and widely used methods are:

− Landfill disposal: in case of landfill disposal, the recovery of the contaminated site is carried out by removal of the polluted soil. New soil has then to be purchased and transported to the site in order to restore the previous landscape, while the contaminated one is sent to a landfill for hazardous or chemical materials. This system is very costly and rather inadequate from an environmental point of view: the use of landfill for the disposal of otherwise recoverable material is not only a waste of money, but also a waste of natural resources. Plus, it contrasts with the attempt made in the past few years to reduce the number of landfills as far as possible.

− Incineration: soil incineration is also a rather expensive method that implies excavation and transportation of the contaminated soil to often remote locations. Furthermore, while the thermal treatment is capable of destroying all organic pollutants, the extremely high heat leaves the soil free from its natural organic fraction as well, and basically sterile. The treated soil has then become an inert material.

− Biological remediation: the biological remediation is an in situ treatment that shows many advantages: the soil needs not to be excavated and transported; after the treatment, the soil has basically the same characteristics (in terms of organic fraction and other physical and chemical properties) as the natural, uncontaminated soil. On the other hand, soil remediation is a rather long process, with possible logistic and practical disadvantages.

− Solvent extraction: the soil is excavated and then treated on site. The major disadvantage consists in the production of more or less great amounts of contaminated solvent, usually rather diluted, that needs further processing for solvent reclamation and recovery. These secondary operations are responsible for rather high decontamination costs.

New methods are therefore being investigated in order to improve the remediation efficiency, lower the costs or the remediation time. Soil remediation by supercritical fluid extraction is one of those methods. There is general agreement in recognising the supercritical fluid extraction as a rather efficient, fast and low expensive treatment for soil remediation.

6

Introduction

1.3. SUPERCRITICAL FLUIDS AND SUPERCRITICAL FLUID EXTRACTION

Under the supercritical point, a single substance can be found either as a solid, liquid, gas or a as coexistence of two or more states (Fig.1.2). Above a certain temperature, the substance can no longer exists as a solid, but it can be a liquid or a gas. Given the temperature, if the external pressure is equal to the vapour pressure, the liquid is in equilibrium with the gas and the two state coexist. Increasing the temperature the equilibrium pressure arises as well, delineating a curve in a pressure-temperature diagram. Moving along the curve to higher pressure and temperature states, the liquid becomes less dense because of thermal expansion and the gas becomes more dense as pressure rises. Eventually the densities of the two phases become identical, the distinction between the gas and the liquid disappears and the curve comes to an end at the critical point [6]. The disappearance of the distinction between the liquid and the gas phases gives the fluid special characteristics that make it suitable as a solvent for extractions.

T

P

Sol

Liq

Gas

Supercritical fluid

Fig. 1.2. Equilibrium curves of a single substance.

The mechanisms of decontamination by SFE are similar to the solvent extraction ones.

Basically, the extraction fluid is pumped through the contaminated soil. Driven by the concentration gradient, the contaminant to be removed moves from the grains, where it is deposited or adsorbed, to the fluid. The solvent, being it liquid or a supercritical fluid, transports the contaminant out of the matrix. At last, the fluid is usually recovered and the separated contaminant destroyed or disposed (depending on the contaminant nature) (Fig. 1.3).

7


SF

= Contaminant

SF+

ContaminantSF

Contaminantrecovery

SF = Supercritical Fluid

Fig. 1.3. Supercritical fluid extraction (SFE) process. The supercritical fluid (SF) flows through the soil particles, onto which the contaminant is adsorbed or deposited. The contaminant desorbes from the soil into the SF and gets transported out of the solid matrix. At last, the contaminant is separated from the SF

and recovered.

The difference between solvents and supercritical fluids is in the transport properties.

Supercritical fluids are somewhat in between a liquid and a gas, and this gives them special properties: generally, diffusion is faster and viscosity and thermal conductivity are lower than in a liquid. This arises partly from the lower densities used in supercritical fluids, and partly because supercritical fluid substances are composed of relatively small mobile molecules. Faster diffusion can result in more rapid extraction. Lower viscosity means that pumping fluids, particularly through packed beds, is easier. Furthermore, properties are controllable by both pressure and temperature and the extra degree of freedom, compared with a liquid, can mean that more than one property can be optimised. [6]. At last, separation of the contaminant from the extraction fluid usually is much easier in case of a supercritical fluid (generally the contaminant separates easily as the pressure drops since its solubility in the fluid sinks at lower pressures) than a liquid solvent.

Many fluids could in principle be used for supercritical extraction, although one

compound, carbon dioxide, has so far been the most widely used, the reason being its convenient critical conditions (table 1.1): in fact, the carbon dioxide reaches the supercritical state at a pressure of 74 bar and a temperature of 304 K (31°C). This means that the use of CO2 for the extraction does not require too much energy in order to keep the fluid above the supercritical conditions. Other advantages are its cheapness, chemical stability, non-toxicity and non-flammability. Large amounts of CO2 accidentally released could constitute a working hazard, given its tendency to blanket the ground, but hazardous detector can be installed. Furthermore, the CO2 is an environmental friendly substitute for other organic solvents. The CO2 that is used is obtained in large quantities as a by-product of fermentation, combustion, and ammonia synthesis and would be released into the atmosphere sooner or later, if it were not used as a supercritical fluid. Its polar character as a solvent is intermediate between a truly non-polar solvent such as hexane and weakly polar solvents. Because the molecule is non-polar, it is often classified as a non-polar solvent, but it has some limited affinity with polar solutes because of its large molecular quadrupole. To improve its affinity with polar molecules further, polar entrainers and modifiers are sometimes added to CO2.

8

Introduction

However, pure CO2 can be used for many organic solute molecules even if they have some polar character, although in general it is not such a good solvent for hydrocarbon polymers and other hydrocarbons of high molar mass. Ethane, ethene and propane become alternatives for these compounds, although they have the disadvantages of being hazardous because of flammability and of being somewhat less environmental friendly. Water has good environmental and other advantages, although its critical parameters are much less convenient (see table 1.1) and it gives rise to corrosion problems. Supercritical water is therefore used only when the high value of the final products can balance the high installation and operational costs, like for instance in the food industry to extract flavours or for coffee decaffeination. Supercritical water is being used as well, at a research level, as a medium for the oxidative destruction of toxic waste [6].

Table 1.1. Critical conditions for various substances.

Substance Critical Temperature

(K)

Critical Pressure

(bar)

Carbon dioxide 304 74

Water 647 221

Ethane 305 49

Ethene 282 50

Propane 370 43

Xenon 290 58

Ammonia 406 114

Nitrous oxide 310 72

Fluoroform 299 49

1.4. THE STATE OF THE ART Much research has been carried out in order to describe various aspects of naphthalene

removal, of more generally PAHs removal, by supercritical CO2. Bartle [3] studies on various substances solubility in pure supercritical CO2 show that naphthalene has little or none solubility in CO2 at temperatures and pressures lower than the supercritical conditions. At a temperature of 308K (35°C), the solubility rises sharply between 80 and 120 BAR, and it then experiences a further 33% increase between 120 and 200 bar. The solubility growth is far more limited at higher pressures. Of course, extraction efficiencies and patterns cannot be directly related to solubility only, since a number of mechanisms take place during the contamination and decontamination processes.

Karimi [8] investigated the interactions between organic compounds and the soil. Basically, the main interaction is adsorption of organics on mineral surfaces, particularly clays. Though, if polar solvents, like water, are present, they occupy the adsorbing sites on mineral surfaces, and non-polar organics such as PAHs are not adsorbed significantly. When mineral surfaces are saturated with water, the main interaction of non-polar organic compounds with soil is partitioning into soil organic matter, proportional to the octanol-water partitioning coefficients. Dissolution into adsorbed water may also take place, with

9


consequent partitioning and interacting with mineral surfaces. Karimi reports also that the presence of water reduces strong PAHs bounding to the soil, enhancing PAHs recovery from the soil itself.

Different results were accomplished by Lee [11], that reports lower PAHs recoveries in case of increased water contents. Lee studies focused on relating the extraction efficiency to the soil properties. Good extraction efficiencies were reached by Lee in case of dry and sandy soils, with a low organic content. The extraction efficiency was lower in case of a higher clay content.

Reduced recoveries were recorded by Karimi as well as by Pignatello [18] for soils that had been contaminated for long times. Connaughton [7] and Burgos [4] disagree from those results, and show similar resistance to desorption both for new-spiked soils and for long contaminated soils (up to 30 years old contamination). Both authors describe also the not complete reversibility of the sorption/desorption process as a consequence of the presence of “resistant” and slowly desorbing fractions. For Connaughton, kinetic limitations are caused by diffusion from immobile fluid in pores. According to Burgos, the contaminant introduced can be divided into contaminant readily extractable, the fraction reversibly adsorbed onto mineral surfaces, and in a fraction that cannot be extracted, either because it is covalently bound to the soil or because it is trapped into micropores, and it is not extracted due to rate-limited processes. In case of naphthalene, the second case applies [4]. On the other hand, the non-complete recovery of naphthalene from the soil is not due to chemical reaction and transformations of the contaminant into other substances [8], as it was reported by Karimi to happen in case of anthracene, which partially converted into molecules with higher molecular weight.

Several studies have also tried to relate the sorption/desorption mechanisms to the operating conditions. Piatt [17] reported of a reduction of the desorption kinetic constants as a consequence of a temperature reduction, although the effect is rather small. The solubility is also affected, decreasing as temperature decreases. On the other hand, a temperature reduction implies higher densities, thus higher extraction efficiencies according to Montero [13]. Montero also studied the CO2 flow effect on the extraction performance. Film transfer resistance was recorded in case of lower flows with a significant reduction of the driving force concentration. At higher superficial velocities, the desorption equilibrium can be considered reached, and the desorption rate controls the overall extraction rate. The extraction rate is in general affected both by external mass transfer, which effect is predominant at lower superficial velocities, and by intraparticle diffusion, that becomes the limiting factor at higher flows. At last, Notar [14] observed an extraction efficiency reduction for 2 and 3-rings PAHs at higher pressures. In fact, even if naphthalene solubility in CO2 increases with density, thus the extraction efficiency should increases, the higher density reduces the diffusion coefficient and the transport properties of the denser fluid.

10

Materials and method

2. ANALYTICAL INSTRUMENTATION, MATERIALS AND METHOD

2.1. ANALYTICAL INSTRUMENTATION The analyses of the collected samples were carried out by means of a gas

chromatograph, operated by helium, as transport medium, air and hydrogen for the flame. The gas chromatograph used is a Hewlett Packard 5890, Series II, equipped with capillary column with the following characteristics:

− stationary phase: SUPELCOWAX 10; − inner diameter ID: 0.053 mm; − length: 30 m; − film thickness: 1 µm.

2.2. MATERIALS The material used for the present research are summarised in table 2.1.

Tab.2.1. Materials used for the experiments.

Material required Material chosen Notes

Contaminant Naphthalene It is the simplest PAH.

Solvent for soil spiking Acetone The solvent chosen has to solve the naphthalene easily and to be far more volatile than the contaminant, in order to evaporate easily and reduce the contemporaneous naphthalene losses.

Solvent for naphthalene collection

Toluene The solvent chosen for the naphthalene collection has to have good affinity with naphthalene and not be too volatile (in order not to evaporate too much during the CO2 bubbling, possibly causing great naphthalene losses), and not very toxic (since its vapours are not collected and could represent a harm for the operators).

Internal Standard (ISTD) for gas chromatography

Octanol The ISTD chosen must be soluble in toluene (the naphthalene collection solvent), and its peak in the gas chromatogram has not to overlap the toluene peak, nor the naphthalene peak, nor any other peak corresponding to impurities).

Soil Silty sand See below

Gas chromatograph operating gases

Helium (transport medium)

Air and hydrogen (flame)

11


The soil type was determined by sieving it by means of sieves with different mashes. The soil fractions F passed through each sieve as a function of the sieve mashes are shown in figure 2.1. The sieves mashes cover the interval between 0.038 mm and 1 mm. Soil fractions with a diameter smaller than 0.038 mm are the 5.8% of the soil sample, while fractions with a diameter bigger than 1 mm are the 9.4% of the soil sample.

0102030405060708090100

Grain diameter (mm)

fract

ion

F (%

)

0.01 0.10 1.00 10.00

Fig. 2.1. Soil analysis result. The soil fractions F passed through each sieve are plotted as a function of the soil mash, representing the grain diameters. The diamonds represent the fractions calculated from the measured weights. The experimental data are interpolated by the continuous line. A dotted line was

drawn to give an idea of the possible grain dimension distribution outside the interval covered by the actual measurements.

The composition of the soil used for the contamination-decontamination process was

therefore the following: − 17% coarse sand (0.5-2 mm) or bigger grained soil − 18% medium sand (0.25-0.5 mm) − 47% fine sand (0.06-0.25 mm) − 18% silt (0.002-0.06 mm) or smaller grained soil The soil used is therefore silty sand.

12


2.3. METHOD The present section describes the pilot plant used for the experimental run and the

scheme followed for the accomplishment of the present research.

2.3.1. Pilot plant The pilot plant used was a continuous flow apparatus. The scheme is shown in fig.2.2.

Cryogenic bath

Pump Mass-Flow

Heater

Extractor cell

Adjustable Restrictor

Collection vessel

Liquid CO2

cylinder

Compressed air

By-pass By-pass

V01

V03

V02

V05 V06 V07 V08

V11 V10V12

V13

V04

V09

TI01

PI01

TI02 03

TI FI01 01

ρI

PI02

04TI

TI05

TI06

Oven

Fig. 2.2. SFE pilot plant scheme.

Shortly, the CO2 is refrigerated in order to be kept liquid prior the compression and it is then compressed to the operating pressure required. The compressed carbon dioxide flows into an oven where it reaches the operating temperature. The density, mass flow, temperature and pressure of the supercritical CO2 are measured before it enters the extraction vessel. The CO2 flows then through the soil packed into the extractor cell extracting the naphthalene. The CO2 is at last expanded (heating is required in order to avoid freezing) and bubbled through a suitable solvent for naphthalene collection.

The various parts of the supercritical extraction apparatus used are here listed and described.

Liquid CO2 cylinder The CO2 is stored in a pressurised cylinder at the CO2 atmospheric temperature vapour

pressure, i.e. 58-60 bar.

13


Cryogenic bath A refrigerated circulating bath ensures refrigeration of the line between the CO2

cylinder and the pump, since the CO2 has to reach the pump as a liquid and not as a gas (being the pump projected for liquid compression, major problems, such as cavitation and possible breakage of the device, would arise if trying to compress a gas).

The cryogenic bath is made of a glycol-water solution (50%wt). A thermometer (TI01) allows control of the refrigerating bath temperature, which is kept at 0°C.

Pump The pump used is the Spe-ed SFE pump of the Applied Separations Inc. A pressure indicator is present on the pump (PI01). The pump used allows a maximum CO2 pressure of 10.000 psig (690 bar). The pump is fed with the liquid CO2 from the cylinder, and operated by compressed

air.

Compressed air Compressed air is used to drive the pump. Minimum air pressure in order to operate

the pump should be 6.2 bar, the maximum 8.6 bar.

Oven The oven maintains the extractor cell and the Mass-Flow at the set operating

temperature. A temperature indicator (TI02), which measures the air temperature inside the oven, is present on the oven front side.

A pressure indicator (PI02) is located downstream the extractor cell (and its by-pass) and refers to the operating pressure in the extractor cell itself when valve V11 is open.

Pressure indicator characteristics: - P max = 600 bar - Accuracy = 0.10% of span - T = -20 to +70°C

Mass-Flow The Mass-Flow apparatus measures the following quantities: mass flow and volume

flow (FI01), temperature (TI03), and density (ρI01). Calibration is performed before each extraction by closing the on-off valves V07 and V08, in order to avoid any CO2 flow through the Mass-Flow.

The Mass-Flow is located in the oven in order to be at the same temperature as the extractor cell, so that the CO2 has the same density as in the extractor cell, thus the same volume flow. Its characteristics are here listed:

Density - Density = 0.1-2.9 g/cm3 - Density accuracy = +/- 0.002 g/cm3 (liquids) +/- 0.008 g/cm3 (gases) - Density repeatability = +/- 0.001 g/cm3 (liquids) +/- 0.004 g/cm3 (gases) Temperature - T max = 204°C - Temperature accuracy = +/- 1°C; +/- 0.5% of reading in °C - Temperature repeatability = +/- 0.2°C

14


Flow - Nominal flow range = 0-82 kg/h - Max flow rate = 108 kg/h - Flow rate accuracy =+/-0.10% +/-[(zero stability/flow rate)*100]% of rate (liquids)

+/-0.50% +/- [(zero stability/flow rate)*100]% of rate (gases) - Flow rate repeatability=+/-0.05% +/-[½(zero stability/flow rate)*100]% of rate (liquids)

+/-0.25% +/-[(zero stability/flow rate)*100]% of rate (gases) - Zero stability = 0.004 kg/h

Extractor cell The extractor cell is a steel vessel with the following characteristics: - P max = 10.000 psig (690 bar) - Outer Diameter (OD) = 1” - Inner Diameter (ID) = 14.4 mm - L = 354 mm - V = 32 ml The volume refers to the vessel internal volume, not the soil volume that was fed into the

vessel for the extraction. The actual volume available for the soil was smaller, since poly-propylene wool was put at the entrance and the exit of the extractor cell in order to avoid the soil from entering the line.

Heater The CO2 is heated before expansion in order to avoid freezing during the expansion

from the operating pressure to the atmospheric pressure. This is performed by heating the line before the expansion valve (restrictor).

The heater is made out of an electric tape surrounding the line and it is combined with a PID (Proportional, Integration, Derivative) temperature controller and a thermometer (Pt100).

Heater characteristics: - T max = 350°C

Adjustable restrictor The restrictor used is the Isco Restrictor Temperature Controller with Adjustable

Restrictor. This device allows expansion of the CO2 from the operating temperature to the atmospheric pressure through a heated regulating valve, which is used in order to vary the CO2 flow.

Adjustable restrictor characteristics: - Flow rate range (5000 psig, 100°C extractor, 100°C restrictor) = 0.5-10 ml/min - Flow rate stability = +/- 20% - Temperature range = 5 – 150°C (the minimum is actually the room temperature,

since the device is not equipped with a cooling system) - Temperature accuracy = +/- 15°C - Temperature repeatability = +/- 10°C

Collection vessel Collection vessels containing a suitable solvent are changed at chosen intervals to

collect the naphthalene extracted. The collection vessels are made out of glass and have an approximate volume of 20 ml.

15


Valves V05 and V13 are regulating valves, all other valves are on-off valves. Regulating valves characteristics: - P max = 1000 bar - Outer Diameter (OD) = 1/8” On-off valves characteristics: - P max = 10000 psig at room temperature - Outer Diameter (OD) = 1/16”

Line - Line material = steel - Outer Diameter (OD) = 1/16” - Inner Diameter (ID)= 0.5 mm

By-Passes The by-passes are made with the same pipes as used for the main line. V05 by-pass allows a higher CO2 flow towards the extractor vessel when required (see

also section 2.3.2). The extractor cell by-pass is used in order to set the operating conditions required,

avoiding naphthalene extraction from the soil before the experimental run. During the simulation, the CO2 is let flow through the by-pass (thus excluding the extractor vessel already filled with soil from the line) while the regulating valves V05 and V13 are adjusted until the approximate operating conditions are reached and kept constant. The by-pass is then closed and the extraction can start, with the required operating conditions set.

16


2.3.2. Research layout The research was carried out according to the following scheme: a. Pre-operations

b. Experiment run b.1. soil preparation

b.2. soil contamination

b.3. preparation of the collection vessels

b.4. pilot plant run

b.5. analyses of the trap samples

c. Repetition of step b under different operating conditions

d. Result analyses

e. Elaboration of the results

f. Error analysis and evaluation

g. Data interpretation

h. Cost analysis

Pre-operations Before the actual run of the experiment, a number of operations have to be carried out.

Namely, those operations are: − Soil analysis: the soil is sieved in order to determine the soil type (see section 2.2) − Solvent choice: the solvents required for the experiment need to be chosen

according to precise criteria. All need to have good affinity and dissolution power towards naphthalene. The solvent used to make the contaminating solution has to be far more volatile than naphthalene, so that it is the solvent rather than the naphthalene that evaporates during the evaporation process. The acetone is thus chosen, being highly volatile and not so toxic as other investigated solvents (i.e. methylene chloride). The solvent used for naphthalene collection after extraction, on the other hand, should not be very volatile in order to reduce the solvent (and possibly the solute) losses during the CO2 bubbling through the solvent itself. Toluene is selected as solvent for the naphthalene collection.

− Gas chromatograph setting: the ISTD (internal standard), temperature and gas flows selected must to allow good resolution of the investigated peaks (i.e. the naphthalene peak and the ISTD peak have not to overlap each other, nor other peaks) and a reasonably short analysis time. Octanol is chosen as ISTD, because of its good solubility into toluene and its good resolution in the gas chromatogram. The flows selected are: air flow, 426ml/min; hydrogen flow, 39.2 ml/min; helium flow, 5.55 ml/min. The split value set is 16.4 ml/min. The temperature profile selected for the analysis is: 130°C for 8 min, 10°C/min for 6 min, 190°C for 1min. Five more minutes are allowed between two analyses to purge the column. Each sample analysis requires therefore 20 min circa.

− Gas chromatograph calibration: the gas chromatograph is calibrated before the experiment by analysing a sufficiently large number of toluene solutions with known concentrations of naphthalene (10, 50, 100, 1000 ppm) and octanol (constant). A calibration line is then drawn, which allows the calculation of the naphthalene concentration in a solution, given the ratio of the naphthalene-octanol peak areas and the octanol amount (chapter 3).

17


− Pressure losses evaluation: clean soil is fed into the extraction vessel and CO2 is let flow through it, simulating an actual extraction. The pressure losses through the soil bed were found to be negligible.

Experiment run Each experiment run can be divided into the following steps: − soil preparation − soil contamination − preparation of the collection vessels − pilot plant run − collection solution sample preparation and analysis

Soil preparation A sample of soil is taken from the whole available amount according to the following

method: the whole amount is divided into two parts by mean of a device that allows the soil to completely casually fall towards one side or the other. Since every soil grain has the same probability to be found in any of the two parts at the end of the process, there is no statistical reason why there should be any composition difference between the two resulting groups. The two samples shall therefore be alike, their composition being the same as the one of the original amount. One of the two samples is then taken and divided again according to the same procedure, until the resulting samples have the required weight. Any of such samples, which shall then be used for the spiking and extraction process, has in principle the same composition as the original soil amount.

The soil sample is dried overnight at 105°C and it is then let cool in contact with the air. This procedure should allow the samples to have similar humidity, around 1% as resulted from the measurement, yet not null.

Soil contamination A contaminating solution is prepared by solving naphthalene into acetone in order to

get a solution with a concentration of about 1000 ppm. The soil is then spiked with then solution, trying to get a concentration as close as possible to 1000 ppm of naphthalene. The vessel is hand-agitated, making a slurry of the soil and the solution, in order to ensure homogenous dispersion of the naphthalene in the soil.

The vessel is then put into a thermostatic bath at 30°C under forced ventilation until the solvent is completely evaporated (about 6 hours). The soil is then transferred into a close container. The container volume must not be far larger than the soil volume, in order to avoid excessive evaporation of the naphthalene from the soil, due to the presence of a significant air volume in the container itself. The extraction takes place about 14 hours after the end of the evaporation process.

In case of the repeatability test, it took about 24 hours for the acetone evaporation at the same temperature, due to lower ventilation around the vessel. The extraction was performed right after the end of the evaporation process.

Preparation of the collection vessels The collection vessels to be used are numbered and weighted; toluene is then fed into

the vessels and weighted. All vessels are closed and kept refrigerated before and after use in order to minimise solvent losses due to evaporation.

Pilot plant run The Supercritical Fluid Extraction (SFE) is carried out according to the following

scheme (see also figure 2.1):

18


− The extraction vessel is opened and cleaned − One end is closed and polypropylene wool is inserted into the vessel in order to

avoid contact between the soil and the vessel inlet/outlet (the aim is to prevent soil particles from entering the line)

− The soil is weighted, fed into the extractor vessel and compacted as much as possible

− PP wool is put on top of the soil before the vessel closure − The extractor vessel is closed and connected to valves V10 and V11, which must

be kept close − The oven temperature is set to the operating temperature − The cooler is switched on and the temperature is set equal to 0°C − When the cooler temperature (TI01) has reached the required value, the CO2

cylinder valve and valve V01 are opened − V04 is opened (while V03 is kept close) − V02 is opened in order to let the compressed air feed the diaphragm pump − the required pressure is set at the pump (PI01) − V06 by-pass valve is closed, while V05 is opened so that the supercritical CO2

flows through V05 − When the temperature TI03 in the Mass-Flow (thus in the oven) has reached the set

value, V07, V08 and V09 are opened − The heater is switched on and the temperature set equal to 100°C − The adjustable restrictor is switched on and V13 is partially opened − When the temperatures at the temperature indicators TI04, TI05 and TI06 (heater

and Adjustable restrictor) have reached the set values, V12 is opened completely in order to let the supercritical CO2 flow along the line (note: the CO2 is not flowing through the extractor vessel, but along its by-pass)

− While the CO2 is flowing through the by-pass, the regulating valves V05 and V13 are adjusted until the approximate operating conditions required are reached

− when the approximate operating conditions are reached V12 is closed, and then V07, V08 and V09 are also closed

− The Mass-Flow is calibrated − The collection vessel 1 is positioned at the pilot plant outlet, ready for CO2

bubbling and naphthalene collection − V07, V08 and V10 are opened, so that the SC-CO2 flows into the extractor cell and

fill it prior the beginning of the extraction. − V06 is opened completely in order to speed up the extractor vessel filling and to

avoid, as far as possible, static extraction before the actual run of the experiment − V10 opening results in a CO2 flow towards the extractor cell, thus a positive flow

through the Mass-Flow, with an initial peak of 20 g/min circa. The flow then decreases until it would eventually become zero and then negative and then positive again and so on, due to pressure waves formation along the line. In order to avoid this phenomenon, V06 is close when Q ≈ 6 g/min (it takes about 30 s – 1 min)

− Straight after, the flow counter on the Mass-Flow is set to zero and, at the same time, V12 is opened and the extraction run starts

− The regulating valves V05 and V13 are continuously regulated during the extraction in order to keep the operating conditions as constant as possible

− At chosen intervals, the collection vessel is substituted by the next vessel. The time passed, the mass flown, the pressure and the CO2 density are recorded

− The used vessels are closed, weighted and kept refrigerated until the analysis

19


Collection solution sample preparation and analysis − About 2 ml of toluene-naphthalene solution are taken from the collection vessels

and put into numbered and weighted vials. − The solution is weighted. − Octanol (ISTD) is added and the solution is weighted again. − The vials are closed and agitated. − The gas chromatograph is switched on, and the gas fluxes (air, hydrogen and

helium) are regulated to the selected values. − 1 µl of the toluene-naphthalene-octanol solution is injected into the gas

chromatograph. − The naphthalene and octanol peak areas are recorded. − From the area ratio, the naphthalene concentration in the toluene-naphthalene

solution is calculated, using the calibration line (chapter 3).

Repetition of the experiment run under different operating conditions Each run as described above is repeated under different operating conditions. The operating conditions investigated are the following:

T = 40°C, 60°C; P = 120 bar, 200 bar; Q = 1 g/min, 1.6 g/min.

A repeatability test is also carried out (chapter 4) in order to determine the reliability of the results obtained.

Result analyses The results obtained from the analyses at the end of each extraction run are elaborated

according to the following curves: − Extraction efficiency vs. time, where the extraction efficiency is defined as below

(Eq.2.1):

η = Nextr (Eq.2.1) Nin

Whereas: Nextr = amount of naphthalene extracted (mg) Nin = amount of naphthalene introduced (mg)

− Percentage of contaminant extracted during a given time interval vs. time, where

the intervals correspond to the time intervals at which the collection vessel are substituted.

While the first curve displays the cumulative percentage of naphthalene extracted as a

function of time, the second curve displays the absolute value corresponding to each single time interval. Basically, the second curve represents the derivative of the first curve vs. time, and therefore it gives a direct information about the extraction rate.

20


Elaboration of the results The results obtained and displayed according to the curves described above are

grouped according to a constant operating condition (T, P or Q) and compared. The comparison aims to reveal anomalies, patterns and other information needed for the data interpretation.

The extraction efficiencies are then analysed as a function of the operating conditions in order to determine what parameters mainly influence the system.

Interpolation of the available data is carried out in order to express the extraction efficiency as a function of the influencing parameters. Given a set of operating conditions, is then possible to assess what extraction efficiency could be obtained. This allows the calculation of the best operating conditions required to bring the soil contamination under target levels.

European regulations set the maximum PAHs concentration in soils at industrial sites equal to 250 ppm (chapter 1). In case of on initial concentration of 1000 ppm (mostly likely found only at industrial sites, chapter 1), 75% extraction efficiency is therefore required to meet the target level of decontamination.

Error analysis and evaluation All possible errors that might affect the final results are investigated and, whenever

possible, evaluated. It is of interest to identify what errors are intrinsic to the method, the procedures and the assumptions made, and what errors are casually committed by the operator during one or more operations. Such analysis allows to determine the precision and accuracy of the results obtained.

Data interpretation An attempt is done to analyse what phenomena and mechanisms lay beyond the

investigated system, based on the results obtained and the elaboration done.

Cost analysis Using the data interpolation done, a field-scale SFE plant is dimensioned. Costs are

then evaluated and compared to the costs of other decontamination methods.

21


3. GAS CHROMATOGRAPH CALIBRATION

A gas chromatographic analysis consists in the detection of all substances present in a solution that is injected into the gas chromatograph, by mean of their separation along a suitable adsorption column, together with their quantitative assessment.

Operatively, a small volume (about 1 µl) of solution is instantly injected at the top of the gas chromatograph column. A suitable substance, called stationary phase, is present as a thin film on the column internal walls. When the solution injected moves along the column transported by an elution gas (helium in this case), the substances present in the solution form more or less strong bounds with the stationary phase, according to their affinity towards it. There is therefore a repartition of the substances between the elution fluid and the stationary phase. A substance with greater affinity towards tends to form stronger bound with it, and thus to show more resistance towards the longitudinal transport. As a result, such substance shall come out of the column later compared to a substance with lower affinity towards the stationary phase (that forms weaker bound, thus flows faster along the column). Different substances shall therefore come out of the column at different times, typical of the various substances and function of the helium flow and the gas chromatograph temperature. Those two parameters can be varied in order to change the exit time times of the substances and thus to allow better resolution of their peaks.

When a substance reaches the exit, it encounters a flame of air and hydrogen and burns with more or less intensity depending on the substance amount.

The gas chromatograph response consists in a gas chromatogram, that is a temporal graph showing the peaks correspondent to all substances present in the analysed solution (Fig.3.1).

Substance D

Substance C

Substance B

time

Substance A

resp

onse

Fig. 3.1. Example of a gas chromatogram.

22

Gas chromatograph calibration

The area of each peak is proportional to the substance amount, which is in turn a function of the substance concentration in the solution and the amount of solution injected into the gas chromatograph. The amount of solution injected, though, can only be known with a rather poor precision.

In order to be able to measure the substance concentration in the analysed solution by mean of a gas chromatograph analysis, without taking into account the injected amount, an internal standard (ISTD) is added to the solution. The ISTD is a substance which amount and concentration in the solution are known.

Given then any peak in the gas chromatogram and the ISTD peak areas, their ratio is proportional to the correspondent substance and the ISTD amount ratio:

amt ratio = k* rsp ratio (Eq.3.1)

Where: amt ratio = amount ratio = m napht / m ISTD rsp ratio = response ratio = Area napht / Area ISDT k = calibration coefficient Known the ratio between the naphthalene and the ISTD amounts (amt ratio), and

known the ISTD weight (mISTD), it is possible to determine the naphthalene amount (mnapht) (Eq.3.2). At last, known the solution amount (msol), the naphthalene concentration (Cnapht) in the collected samples is easily determined (Eq.3.3).

m napht = amt ratio * m ISTD (Eq.3.2)

C napht = m napht / m sol (Eq.3.3)

In order to be able to calculate the naphthalene concentration in the solution samples analysed, the coefficient k has first to be determined by calibration of the gas chromatograph.

The gas chromatograph calibration takes place in two steps: - analysis method setting - construction of the calibration line

Analysis method setting A number of parameters have to be selected before the actual gas chromatograph

calibration since they influence the instrument response. Beside the ISTD, which has to be selected according to the criteria described in

section 2.2, the following parameters have to be set: - air and hydrogen flow for the flame - the helium (elution gas) flow - the temperature

The helium flow and the gas chromatograph temperature influence the exit time of

each substance and thus the resolution of the peaks in the gas chromatogram.

23


Construction of the calibration line After having set the above parameters, toluene solutions with known concentrations of

naphthalene (10, 50, 100, 1000 ppm) and octanol (the ISTD chosen) are injected into the gas chromatograph.

The naphthalene and octanol amount ratio (known) is then plot into a graph as a function of the response ratio, i.e. the ratio between the naphthalene and the octanol areas in the gas chromatogram (Fig.3.2). The experimental points are then interpolated by a line, which is also forced through the axis origin.

The interpolation line is the gas chromatograph calibration line, which coefficient shall be used for the calculation of the naphthalene concentration in the toluene solution samples taken during the extraction run, by mean of equations 3.1, 3.2 and 3.3.

amt ratio = 730.94 * rsp ratioR2 = 0.99996

rsp ratio

amt r

atio

0.0E+00 5.0E-02 1.0E-01 1.5E-01 2.0E-01 2.5E-010.0E+00

2.0E+01

4.0E+01

6.0E+01

8.0E+01

1.0E+02

1.2E+02

1.4E+02

1.6E+02

1.8E+02

Fig. 3.2. Calibration line.

24

The repeatability test

4. THE REPEATABILITY TEST

Errors always affect the results of experimental tests. In particular, during the present research, the errors introduced can be divided into three main groups according to the operations they arise from. Such operations are:

- sample preparation - extraction run - sample analysis

The different types of errors, their relevancy in the data reliability and their interpretation are more deeply discussed in chapter 7.

The repeatability test concerns the errors generated during the extraction run. In fact,

even if the operating conditions set are known, the parameters fluctuate during the extraction due to the pilot plant system functioning. The fluctuations might be or be not responsible for significant variations of the extraction efficiency and the overall performance of the system. In order the final results to be reliable, it is therefore necessary to empirically verify the extraction repeatability, i.e. that the errors introduced by the operating conditions fluctuations do not affect the results so much that, if the extraction were repeated, the results obtained would no longer be the previous ones.

The repeatability test consists in the run of two extractions at the same operating conditions (temperature, pressure and mass flow) on ideally the same soil. If the two extractions give the same results, it is then possible to state that the plant and the method used guarantee extraction repeatability, and thus the reliability of the obtained results.

In order to have two soil samples as similar as possible, a sufficient amount of soil has to be dried, spiked and evaporated together; the soil is then divided into the two samples, which are extracted at the same operating conditions (Fig.4.1). This procedure reduces the errors introduced during the contamination phase, so that the differences between the soil samples can be assumed to be negligible.

25


26

Acetone Naphthalene

Contaminatingsolution preparation

Contaminatingsolution Soil

Soil contamination

Acetone evaporation

Contaminatedsoil

Acetone Emissioninto air

Soil sampling

Sample 1

SFE(P1, T1, Q1)

SC-CO2

Naphthalenedissolution into

toluene

Toluene

CO2

Toluenevapours

Emissioninto air

Emissioninto air

Toluene+

Naphthalene

Samplepreparation

ISTD

Sampleanalysis

Sample 2

SFE(P2, T2, Q2)

SC-CO2

Naphthalenedissolution into

toluene

Toluene

CO2

Toluenevapours

Emissioninto air

Emissioninto air

Toluene+

Naphthalene

Samplepreparation

ISTD

Sampleanalysis

P1 = P2T1 = T2Q1 = Q2

Fig.4.1. Scheme of the repeatability test operations.


The soil samples and extraction conditions used for the repeatability test are summarised in tables 4.1 and 4.2. The extraction efficiency difference recorded is 0.49%.

Tab. 4.1. Extraction 1 data.

Extraction number - 1 Set temperature °C 40

Extraction duration min 40 Set pressure bar 200

Number of samples - 10 Set mass flow g/min 1.6 Sample amount g 60.55 Measured temperature (av.) °C 39.83

Naphthalene concentration ppm 1000.24 Measured pressure (av.) bar 201.1

Naphthalene amount mg 60.56 Measured mass flow (av.) g/min 1.608

Naphthalene extracted mg 53.84 Measured density (av.) g/cm3 0.8484

Extraction efficiency % 88.89



Extraction duration min 40 Set pressure bar 200





Extraction efficiency % 88.45

In order to make a comparison not only of the efficiency asymptotic value but also of

the extraction curve shape, and in order to detect any extraction pattern, the curves are displayed in figures 4.2 and 4.3.

The extraction efficiency curves (Fig.4.2) of the repeatability test extractions have similar shapes and basically the same asymptote. The only difference is a small delay of curve 1 compared to curve 2. The observed delay might be due to a slower opening of the on-off valve V12 at the beginning of the extraction run (see chapter 2). If extraction 1 is shifted backwards of one minute, though, the two curves overlap perfectly.

27


0102030405060708090

100

0 10 20 30 40 5

Time (min)

Ext

ract

ion

effic

ienc

y (%

)

Extraction 1

Extraction 2

Extraction 1 anticipated

0

Fig. 4.2. Extraction efficiency curves. Extractions 1, 1 anticipated and 2.

Figure 4.3 compares the amount of naphthalene extracted at given time intervals vs. time in case of the two extractions. Again, best overlapping is obtained introducing one minute backward shift of curve 1.

0

5

10

15

20

25

30

0 10 20 30 40 50Time (min)

Nap

htha

lene

ext

ract

ed (m

g)

Extraction 1

Extraction 1 anticipated

Extraction 2

Fig. 4.3. Naphthalene extraction vs. time. Extractions 1, 1 anticipated and 2.

The curves overlap rather well, although some differences can be observed. Curve 2 has a better shape according to CO2 extraction models, while curve 1 is

slightly irregular. Curve 1 has in fact a lower peak, while slightly more extraction takes place after 15-20 minutes from the beginning of the run.

28


5. RESULTS

The extractions were carried out varying the operating conditions pressure, temperature, and mass flow. Two values were assigned to each parameter, for a total number of 8 extractions. The actual total number is 10, in order to take into account extractions with excessive anomalies, that had to be repeated. The operating conditions investigated are summarised in the schemes below.

Tab. 5.1 and 5.2. Operating conditions investigated and correspondent extractions.

Q = 1 g/min Q = 1.6 g/min

T (°C) T (°C)

P (bar) 40 60 P (bar) 40 60

120 Ex. 3 Ex. 5 / Ex. 9 120 Ex. 8 Ex.11 / Ex.12

200 Ex. 4 Ex. 6 200 Ex. 7 Ex. 10

The present chapter summarises the extraction runs performed and the results

obtained. For each extraction, the following data are displayed: − extraction information − result graphs − brief comments on the result curves profiles and/or anomalies recorded

Extraction information: it is summarised in tables, and includes: Extraction general information: − Extraction number − Extraction duration − Number of naphthalene collection solution samples taken − Operating conditions set: temperature, pressure, mass flow Measured parameters: − Average operating conditions: temperature, pressure, mass flow − CO2 average density Soil and contaminant information: − Soil sample amount − Naphthalene concentration in the soil sample neglecting losses during the

contamination procedure − Naphthalene amount in the soil sample calculated from the two quantities above − Naphthalene extracted − Extraction efficiency

Result graphs: the following graphs are used to display the results (section 2.3.2): − extraction efficiency vs. time (cumulative) − percentage of naphthalene extracted during a given time interval vs. time

(absolute)

29


EXTRACTION 3 Tab. 5.3. Extraction 3 data.


Extraction duration min 126 Set pressure bar 40Extraction efficiency 68.62%

Number of samples - 17 Set mass flow g/min 1 Sample amount g 62.25 Measured temperature (av.) °C n.a.*

Naphthalene concentration ppm 999.87 Measured pressure (av.) bar n.a.*



Notes: A delay was recorded at the beginning of the extraction, due to the fact that the extraction vessel was not filled with pressurised CO2 when the extraction started (V10 had not been opened); before the beginning of the actual extraction, then, the CO2 had to fill the vessel volume, quite relevant compared to the total line volume. A pressure drop of 40-50 bar at PI02 was also recorded.

The double peak in fig. 5.2 is due to the “channelling effect” caused by poor compacting of the soil (Ch. 7).

* The average operating conditions have not been measured during the first runs. This limits the interpretation of eventual anomalies.

02468

10121416

0 50 100 150time [min]

Extra

cted

nap

htha

lene

(%)

01020304050607080

0 20 40 60 80 100 120 140

time [min]

Extra

ctio

n ef

ficie

ncy

(%)

Fig. 5.1. Extraction efficiency vs. time. Extraction 3. Fig. 5.2. Naphthalene extraction vs. time. Extraction 3.

30

Results

EXTRACTION 4




Number of samples - 17 Set mass flow g/min 1 Sample amount g 64.724 Measured temperature (av.) °C n.a.

Naphthalene concentration ppm 1011.45 Measured pressure (av.) bar n.a.



Notes: The anomalies shown by the graphs in figure 5.3 and 5.4 are common to those recorded in case of extraction 3, and namely: extraction starting time delay and a double peak, although far less pronounced. For the explanation, see extraction 3.

01020304050607080

0 20 40 60 80 100 120 140time [min]

Extra

ctio

n ef

ficie

ncy

(%)

0

5

10

15

20

0 50 100 150time [min]

Extra

cted

nap

htha

lene

(%)


31


EXTRACTION 5





Naphthalene concentration ppm 1000.57 Measured pressure (av.) bar n.a.



Notes: A small inflection is present in both curves at about 6 minutes. This might be due to poor control of the plant operating conditions at the beginning of the extraction. The remaining part of the curves is smooth and regular.

The overall efficiency is very high, out of the possible range. The extraction is therefore repeated (extraction 9).

0

510

1520

2530

35

0 20 40 60 80 100 120 140time [min]

Extra

cted

nap

htha

lene

(%)

0

20

40

60

80

100

0 50 100 150time [min]

Ext

ract

ion

effic

ienc

y (%

)


32

Results

EXTRACTION 6




Number of samples - 15 Set mass flow g/min 1 Sample amount g 63.87 Measured temperature (av.) °C n.a.*




Notes: The extraction efficiency curve is rather smooth and regular. A minor anomaly is present in the extraction curve of figure 5.8 at about 20 minutes.

0

5

10

15

20

25

0 20 40 60 80 100time [min]

Ext

ract

ed n

apht

hale

ne (%

)

0

10

20

30

40

50

60

70

0 20 40 60 80 100time [min]

Extra

ctio

n ef

ficie

ncy

(%)


33


EXTRACTION 7




Number of samples - 12 Set mass flow g/min 1.6 Sample amount g 63.63 Measured temperature (av.) °C n.a.




Notes: The extraction curves are regular and smooth.

05

10152025303540

0 10 20 30 40 50 60 70time [min]

Extra

cted

nap

htha

lene

(%)

0102030405060708090

0 10 20 30 40 50 60 70time [min]

Extra

ctio

n ef

ficie

ncy

(%)


34

Results

EXTRACTION 8




Number of samples - 8 Set mass flow g/min 1.6 Sample amount g 63.14 Measured temperature (av.) °C n.a.





05

10152025303540

0 5 10 15 20 25 30 35time [min]

Extra

cted

nap

htha

lene

(%)

0102030405060708090

0 5 10 15 20 25 30 35time [min]

Extra

ctio

n ef

ficie

ncy

(%)


35


EXTRACTION 9








Notes: Extraction 9 is a repetition of extraction 5. The extraction curves are regular and smooth.

0

5

10

15

20

25

0 20 40 60 80 100time [min]

Extra

cted

nap

htha

lene

(%)

0

10

20

30

40

50

60

0 20 40 60 80 100time [min]

Extra

ctio

n ef

ficie

ncy

(%)


36

Results

EXTRACTION 10









010203040506070

0 20 40 60 80 100time [min]

Extra

ctio

n ef

ficie

ncy

(%)

0

5

10

15

20

25

30

0 20 40 60 80 100time [min]

Extra

cted

nap

htha

lene

(%)


37


EXTRACTION 11








Notes: Double peaks can be noted at 12 and 20 minutes. Beside the “channelling effect” (see section 7.1.2), a poor control of the operating conditions might be the cause.

The overall efficiency is very low, out of the possible range. The extraction is repeated (extraction 12).

-5

0

5

10

15

20

0 20 40 60 80 100

time [min]Ex

tract

ed n

apht

hale

ne (%

)

05

10152025303540

0 20 40 60 80 100time [min]

Ext

ract

ion

effic

ienc

y (%

)


38

Results

EXTRACTION 12









0

5

10

15

20

25

30

0 20 40 60 80 100time [min]

Extra

cted

nap

htha

lene

(%)

010203040506070

0 20 40 60 80 100time [min]

Extra

ctio

n ef

ficie

ncy

(%)


39


6. ELABORATION OF THE RESULTS

6.1. COMPARISON OF THE RESULTS AT DIFFERENT OPERATING CONDITIONS

This section includes the comparison of all extraction curves at different operating conditions. Aim of the comparison is to find patterns and possible dependence of the variables from the controlled parameters.

6.1.1. Comparison of the extraction results at constant temperature

Extractions at 40°C The efficiency increases in case of higher mass flow, varying from about 69% to 80%

for a flow varying between 1 and 1.6 g/min (Fig.6.1). As the mass flow increases, the time needed to reach the asymptote decreases from 40 minutes to 20-25 minutes.

Small differences in density and no difference in extraction efficiency were recorded between 120 and 200 bar.

0102030405060708090

0 50 100 150time [min]

Ext

ract

ion

effic

ienc

y (%

)

Extr.3 (0.7307 g/cm3; 120 bar; 1g/min)Extr.4 (0.8461 g/cm3; 200 bar; 1 g/min)

Extr.7 (0.8446 g/cm3; 200 bar; 1.6 g/min)Extr.8 (0.7278 g/cm3; 120 bar; 1.6 g/min)

Fig. 6.1. Comparison of the efficiency curves at 40°C.

The comparison of the naphthalene extraction curves at 40°C (Fig.6.2) is complicated by major anomalies recorded in case of extraction 3 and 4 (extractions at 1 g/min). Difficulties in controlling the operating conditions during the extractions and probably imperfections in the extraction cell filling and soil compacting resulted in a time delay of the curves and in double peaks, with lower height of the main peak. It is therefore impossible to estimate the peak value of extractions at 40°C and 1 g/min without

40

Elaboration of the results

anomalies. It is though likely that the peak position would be the same as for the remaining curves.

In case of a mass flow of 1.6 g/min, the curve reaches the peak after 6 minutes from the beginning of the run. From the 3rd to the 6th minute, 35% circa of the naphthalene is recovered from the soil.

0

5

10

15

20

25

30

35

40

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140time (min)

Ext

ract

ed n

apht

hale

ne (%

) Extr.3 (0,7307 g/cm3; 120 bar; 1 g/min)

Extr.4 (0,8461 g/cm3; 200 bar; 1 g/min)

Extr.7 (0,8448 g/cm3; 200 bar; 1,6 g/min)

Extr.8 (0,7285 g/cm3; 120 bar; 1,6 g/min)

Fig. 6.2. Comparison of the naphthalene extraction curves at 40°C.

Extractions at 60°C Extraction 5 and 11 efficiencies, which values are 95% and 34% respectively, are

doubtlessly out of the possible efficiency range. Errors occurred during the contamination process are the most likely explanation. Because of the meaningless of the correspondent results, the runs were repeated (extraction 9 and 12 respectively).

The comparison of extractions 6, 9, 10 and 12 shows on the other hand similar extraction efficiencies, with values ranging from about 58 to 60%.

Higher mass flows resulted in steeper curves (therefore the asymptote is reached in a shorter time), and in a slight efficiency increase. Higher extraction efficiencies are observed in case of higher pressures as well, although the variation is rather limited whether compared to the correspondent density increase.

0102030405060708090

100

0 20 40 60 80 100 120 140time [min]

Extra

ctio

n ef

ficie

ncy

(%)

Extr.5 (0,4495 g/cm3; 120 bar; 1 g/min)Extr.6 (0.7294 g/cm3; 200 bar; 1 g/min)Extr.9 (0,4606 g/cm3; 120 bar; 1 g/min)Extr.10 (0,7307 g/cm3; 200 bar; 1.6 g/min)Extr.11 (0,4404 g/cm3; 120 bar;1.6 g/min)Extr.12 (0,4508 g/cm3; 120 bar; 1.6 g/min)

Fig. 6.3. Comparison of the efficiency curves at 60°C.

41


0

5

10

15

20

25

30

0 20 40 60 80time (min)

Extra

cted

nap

htha

lene

(%)

Extr.6 (0,7294 g/cm3; 200 bar; 1 g/min)

Extr.9 (0,4606 g/cm3; 120 bar; 1 g/min)

Extr.10 (0,7307 g/cm3; 200 bar; 1,6 g/min)

Extr.12 (0,4508 g/cm3; 120 bar; 1,6 g/min)

100

Fig. 6.4. Comparison of the naphthalene extraction curves at 60°C.

Because of their major anomalies, extractions 5 and 11 are not included in figure 6.4 since they would not be significant for any further consideration and analysis.

The peaks occur after 6 minutes from the beginning of the run for all extractions. Peaks are higher at higher mass flows, while the pressure seems not to have any relevant effect. The peak height is about 26% of the total naphthalene present in the soil at 1.6 g/min, 23% at 1 g/min. Extraction 6 shows a lower peak due to the presence of another peak after 20 minutes.

42


6.1.2. Comparison of the extraction results at constant mass flow

Extractions at 1 g/min

0

10

20

30

40

50

60

70

80

0 20 40 60 80 100 120 140time [min]

Extra

ctio

n ef

ficie

ncy

(%)

Extr.3 (0.7307 g/cm3; 120 bar;40°C)Extr.4 (0.8461 g/cm3; 200 bar; 40°C)Extr.6 (0.7294 g/cm3; 200 bar; 60°C)Extr.9 (0,4606 g/cm3; 120 bar; 60°C)

Fig. 6.5. Comparison of the efficiency curves at 1 g/min.

The extraction efficiency increases from about 57% to 69% as the temperature decreases from 60°C to 40°C, while the pressure has a modest influence on it, and only at higher temperatures (i.e. when the extraction efficiency is lower) (Fig.6.5).

On the contrary, the temperature is the only extraction time influencing parameter, causing an extraction time decrease from 40 to 20 minutes at higher pressures. An apparent pressure effect on the extraction time at lower temperatures is most likely due to the major anomalies affecting extractions 3 and 4, which are better displayed by figure 6.6.

0

5

10

15

20

25

0 20 40 60 80 100 120 140time (min)

Extra

cted

nap

htha

lene

(%)

Extr.3 (0.7307 g/cm3; 40 °C; 120 bar)

Extr.4 (0.8461 g/cm3; 40 °C; 200 bar)

Extr.6 (0.7294 g/cm3; 60 °C; 200 bar)

Extr.9 (0.4606 g/cm3; 60°C; 120 bar)

Fig. 6.6. Comparison of the naphthalene extraction curves at 1 g/min.

Because of the major anomalies recorded in case of extractions 3 and 4, no certain conclusion can be drawn on the curves of figure 6.6.

43


Extractions at 1,6 g/min

Fig. 6.7. Comparison of the efficiency curves at 1,6 g/min.

As for a mass flow of 1 g/min, the extraction efficiency increases as the temperature decreases, varying between 59% and 80 % for 60°C and 40 °C respectively. The pressure effect is rather small and limited to extractions at 60°C.

The asymptote is reached after 20 minutes for all extractions.

0102030405060708090

0 20 40 60 80time [min]

Extra

ctio

n ef

ficie

ncy

(%)

Extr.7 (0.8448 g/cm3; 40°C; 200 bar)Extr.8 (0.7285 g/cm3; 40°C; 120 bar)Extr.10 (0.7307 g/cm3; 60°C; 200 bar)Extr.12 (0.4508 g/cm3; 60°C; 120 bar)

100

0

5

10

15

20

25

30

35

40

0 20 40 60 80 0time (min)

Extra

cted

nap

htha

lene

(%)

Extr.7 (0,8448 g/cm3; 40°C; 200 bar)

Extr.8 (0,7285 g/cm3; 40°C; 120 bar)

Extr.10 (0,7307 g/cm3; 60°C; 200 bar)

Extr.12 (0,4508 g/cm3; 60°C; 120 bar)

10

Fig. 6.8. Comparison of the naphthalene extraction curves at 1,6 g/min.

All curves have the same shape and width, but different peak height. The peaks are higher for lower temperatures, reaching 35% of the naphthalene present in the soil at 40°C against 26% at 60°C. The peaks occur 6 minutes after the beginning of the extractions. Nor the pressure nor the CO2 density seem to influence the extraction performance.

44


6.1.3. Discussion on the comparisons performed The extraction curves obtained (Fig. 6.1, 6.3, 6.5, 6.7) show no clear patterns relating

the operating conditions to the extraction time (i.e. the time required to reach the asymptote). This is mainly due to anomalies occurred during some extraction runs, which limit the meaningfulness of the results (section 7.1.2).

In some cases, though, the data seem to suggest that the extraction time decreases as the mass flow and the temperature increase, while there seems to be no relation with the CO2 pressure and density. Anyhow, an extraction time of 30-40 minutes allows the system to reach the asymptote under all investigated operating conditions.

More certainly, the extraction efficiency increases as the temperature decreases and

the flow increases. Higher pressures seem to be responsible of a small efficiency increase, but only at higher temperatures (thus lower overall efficiencies).

The percentage of naphthalene extracted vs. time follows a curve that has the same

shape for all the extractions, beside the cases where a poor control of the operating conditions resulted in anomalies. The extraction percentage peaks always occur after six minutes from the beginning of the extraction run, whenever the operating conditions are sufficiently constant. The peak height increases as the temperature decreases, while the dependence from the flow appears more uncertain. Again, pressure and density have no influence on the peak height.

Considering the complexity of the operating conditions interactions, further

elaboration of the data is necessary in order to be able to express the extraction efficiency as a function of the relevant parameters. Such elaboration is carried out in the following sections.

45


6.2. EXTRACTION EFFICIENCY AT DIFFERENT OPERATING CONDITIONS

6.2.1. Extraction efficiency as a function of the operating conditions

In the present section, the extraction efficiency is analysed as a function of the operating conditions, in order to determine the relationship between the variables.

The operating conditions set for each extraction run together with the recorded CO2 densities and the extraction efficiencies are summarised in table 6.1. Extractions 5 and 11 showed major anomalies, and they shall not be considered for further elaboration.

Tab. 6.1. Summary of the operating conditions and the extraction efficiencies of each extraction. Extractions displaying major anomalies are highlighted and shall not be included in further elaboration.

Extraction Temperature Density Pressure Mass Flow Efficiency

°C g/cm3 bar g/min %

3 40 0.7307 120 1 68.62

4 40 0.8461 200 1 68.62

5 60 0.4495 120 1 95.53

6 60 0.7294 200 1 57.54

7 40 0.8446 200 1.6 79.83

8 40 0.7285 120 1.6 80.06

9 60 0.4606 120 1 56.06

10 60 0.7307 200 1.6 61.19

11 60 0.4404 120 1.6 34.39

12 60 0.4508 120 1.6 57.28

The relationship between the extraction efficiency and the operating conditions is

investigated by mean of the graphs in figures 6.9, 6.10, 6.11 and 6.12. Each graph displays the efficiencies recorded as a function of a single parameter (pressure, temperature, density and mass flow respectively). The relations between the variable are highlighted by interpolation lines. The dotted lines represent the target recovery sought, i.e. the recovery needed in order to reduce the naphthalene concentration under the law limits. In this case, the naphthalene initial concentration is 1000 ppm. Such concentration could be found at industrial sites, but usually not at residential sites (see chapter 1). A value of 250 ppm is taken as law limit in the EU for PAHs concentration at industrial sites (chapter 2). The target recovery is therefore 75%.

46


0

20

40

60

80

100

100 120 140 160 180 200 220

Pressure (bar)

Ext

ract

ion

effic

ienc

y %

Extraction efficiency %Target recovery

Interpolation (Efficiency)

Fig. 6.9. Extraction efficiency as a function of pressure.

In the interval considered, the average extraction efficiency (interpolation line) is rather constant as a function of pressure. Its value is around 70%, thus lower than the target recovery (Fig.6.9).

0

20

40

60

80

100

30 40 50 60 70Temperature (°C)

Ext

ract

ion

effic

ienc

y %

Extraction efficiency %

Target recovery


Fig. 6.10. Extraction efficiency as a function of temperature.

The extraction efficiency increases as the temperature decreases. On average, less than 40°C are required to meet the target recovery (Fig.6.10).

Figure 6.11 shows an efficiency increase at higher CO2 densities. The relation between

density and efficiency is though only apparent. In fact, density is a function of pressure and temperature. The extraction efficiency

recorded was shown to be independent from pressure (Fig.6.9), while it increases as the temperature decreases (fig.6.10). Given a constant pressure, lower temperatures imply higher densities, as well as higher extraction efficiencies. The efficiency increase observed in case of higher densities is therefore a mere effect of the temperature variation.

On average, the target recovery is approached but not met at higher experimental densities (0.85 g/cm3).

47


0

20

40

60

80

100

Density (g/cm3)

Ext

ract

ion

effic

ienc

y %

Extraction efficiency %Target recovery


0.4 0.5 0.6 0.7 0.8 0.9

Fig. 6.11. Extraction efficiency as a function of density.

10

20

40

60

80

100

Mass flow (g/min)

Ext

ract

ion

effic

ienc

y %

Extraction efficiency %

Target recoveryInterpolation (Efficiency)

0.6 0.8 1.0 .2 1.4 1.6 1.8

Fig. 6.12. Extraction efficiency as a function of flow.

The extraction efficiency increases as the mass flow increases. The target recovery is not met, on average, not even at higher flows (Fig.6.12).

48


6.2.2. Extrapolations The analysis of the extraction efficiency as a function of the operating conditions

(section 6.2.1) shows dependence from temperature and mass flow only. In the present section, the relations between efficiency, temperature and mass flow are further analysed in order to extrapolate general equations for the description of the extraction.

Due to the limited number of extraction runs, only linear interpolations and extrapolations can be carried out. This limits the reliability of the interpolation results to a close interval around the measured data, since the error between the interpolated results and the actual extraction efficiency increases with the distance from the measured data.

In particular, the linear interpolation may not be very accurate in describing the relations between the extraction efficiency and the operating conditions when approaching 100% recovery. Actually, the resistance of naphthalene towards a complete desorption, as described by Connaughton [7] and Burgos [4] (chapter 1), certainly implicates a non-linear correlation at high extraction efficiencies.

Extrapolation of the extraction efficiency vs. temperature Figure 6.13 describes the relations between efficiency and temperature at different

mass flow conditions. The data are linearly interpolated.

E = -0.591 T + 92.26

E = -1.0355 T + 121.37

0

20

40

60

80

100


Extra

ctio

n ef

ficie

ncy

%

1 g/min1.6 g/minInterpolation (1 g/min)Interpolation (1.6 g/min)

Fig. 6.13. Extraction efficiency vs. temperature at different mass flows. Linear interpolation.

Known the equations of the extraction efficiency interpolation lines at 1 and 1.6 g/min, it is possible to extrapolate the general equation that expresses the efficiency as a function of temperature at different flows. The general equation describes an infinite number of lines that pass through the intersection of the lines in figure 6.13, and with a coefficient depending from the flow:

E = m(Q) * (T – T1) + E1 Where (T1; E1) is the intersection point of lines in figure 6.13. Being: E = - 1.0355 T + 121.37 (Q = 1.6 g/min) E = - 0.591 T + 92.26 (Q = 1 g/min)

49


The intersection point is then: T1 = 65.4893°C E1 = 53.5558 % Hp.: m(Q) is a linear function of Q in the surrounding of the interval examined:

m(Q) = n * Q + q Known that: Q = 1 g/min m = - 0.591 Q = 1.6 g/min m = - 1.0355 Then: n = -0.7408 q = 0.1498 Therefore:

E = (- 0.7408 Q + 0.1498) * (T - 65.4893) + 53.5558 The formula, as well as its graphic representation below, are valid for temperature

expressed in Celsius degrees and for mass flow expressed in g/min. Both the formula and the graph refer to a soil amount of 63 g circa.

50

60

70

80

90

100


Ext

ract

ion

effic

ienc

y (%

) 1.0 g/min 1.3 g/min1.6 g/min 1.9 g/min

Target recovery

Fig. 6.14. Extraction efficiency vs. temperature at different mass flows. Linear extrapolation. The CO2 flow refers to a soil amount of 63 g circa.

Note: The lowest temperature possible is 31°C. Below this temperature, the CO2 loses its supercritical properties.

The formula above can also be more generally expressed as:

E = (- 0.3899 Q + 0.1498) * (T - 65.4893) + 53.5558 (Eq. 6.1) Where: [T] = °C [Q] = m3

CO2/h m3soil

50


Extrapolation of the extraction efficiency vs. flow Figure 6.15 describes the relations between efficiency and mass flow at different

temperature conditions. The data are linearly interpolated.

0

20

40

60

80

100

Flow (g/min)

Extra

ctio

n ef

ficie

ncy

(%)

40°C60°CInterpolation (40°C)Interpolation (60°C)

0.8 1.0 1.2 1.4 1.6 1.8

E = 18.875 Q + 49.745

E = 4.0583 Q + 52.742

Fig. 6.15. Extraction efficiency vs. mass flow at different temperatures. Linear interpolation.

Known the equations of the extraction efficiency interpolation lines at 40°C and 60°C, it is possible to extrapolate the general equation that expresses the efficiency as a function of mass flow at different temperatures.

The general equation describes an infinite number of lines that pass through the intersection of the lines in figure 6.15, and with a coefficient depending from the temperature:

E = m(T) * (Q – Q1) + E1 Where (Q1; E1) is the intersection point of lines in figure 6.15. Being: E = 18.875 Q + 49.745 (T = 40°C) E = 4.0583 Q + 52.742 (T = 60°C) The intersection point is then: Q1 = 0.20223 g/min E1 = 53.5629% Hp.: m(T) is a linear function of T in the surrounding of the interval examined:

m(T) = n * T + q Known that: T = 40°C m = 18.875 T = 60°C m = 4.0583 Then: n = -0.7399 q = 48.4544

51


Therefore: E = (- 0.7399 T + 48.4544) * (Q - 0.2023) + 53.5629

The formula, as well as its graphic representation below, are valid for temperature

expressed in Celsius degrees and for mass flow expressed in g/min. Both the formula and the graph refer to a soil amount of 63 g circa.

50

60

70

80

90

100

Flow (g/min)

Ext

ract

ion

effic

ienc

y (%

) 31°C 40°C50°C 60°C

0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0

Target recovery

Fig. 6.16. Extraction efficiency vs. mass flow at different temperatures. Linear extrapolation. The CO2 flow refers to a soil amount of 63 g circa.

Note: The line at 31°C is the upper limit for naphthalene recovery by SC-CO2, since

CO2 loses its supercritical properties at lower temperatures. The formula above can also be more generally expressed as:

E = (- 0.7399T + 48.4544) * (0.5249Q - 0.2023) + 53.5629 (Eq. 6.2) Where: [T] = °C [Q] = m3

CO2/h m3soil

52

Error analysis and evaluation

7. ERROR ANALYSIS AND EVALUATION

7.1. ERROR GENERATION The errors affecting the results of any experimental work may arise either from the

hypotheses on which the research is based, or from the actual execution of each experimental step. In the present research, the extraction efficiency is calculated as ratio between the naphthalene extracted and the naphthalene introduced into the soil by mean of the contaminating solution (Eq.7.1).

η = Nextr (Eq.7.1) Nin

Whereas: Nextr = amount of naphthalene extracted (mg) Nin = amount of naphthalene introduced (mg) This implies the hypothesis that all the naphthalene solved into the solution stays in

the soil after the contamination process, and also that all the naphthalene extracted by mean of the CO2 flow gets trapped into the collecting solution that is afterwards analysed (Fig. 7.1). Naphthalene is a rather volatile substance, and might therefore partially evaporate during the described steps. The assumptions above might therefore not be true, or they might be acceptable only within a certain range of error.

It is of interest to identify and possibly to estimate all possible errors and their

sources, in order to be able to judge about the reliability and the meaningfulness of the results themselves. In principle, every single operation carried out during the experiment run introduces errors in the final results. Some operations are, from this point of view, more relevant than others.

The error investigation and analysis is carried out by dividing them into groups, according to the steps from which they arise. Particularly, the following steps can be highlighted:

− soil preparation and contamination − extraction run − sample analyses

53


Naphthalene Acetone

Soil+

Solution

Soil+

Naphthalene

Acetone

Evaporation

Soil+

Residual Naphthalene

CO2

CO2+

Naphthalene

Extraction

DissolutionCO2

Toluene

Toluene+

Naphthalene

Fig. 7.1. Steps undergone by the naphthalene.

7.1.1. Soil preparation and contamination The soil samples were taken from the available amount of soil according to a method

supposed to ensure that they statistically represent the original sample (section 2.3.2). Of course, two given soil samples shall always display some differences. Still, if those differences are within a sufficiently limited range, they might be assumed not to affect the overall extraction performance and results.

The contamination process might be a significant source of error: even if the

contaminant solution used to spike the soil is accurately prepared, the real concentration is only known with a certain approximation dependent on the scale precision. Moreover, major errors can occur, leading to extremely inaccurate results: this is for instance the case of extractions 5 and 11, where the concentration of the solutions used to spike the soil were probably very different from the supposed ones, due to mistakes done during the solution preparation. As a consequence, the results obtained were completely meaningless, and the extractions had to be repeated.

The procedure adopted for the solvent evaporation used during the soil contamination

is another key issue. As already stated, the extraction efficiency is calculated as ratio between the naphthalene extracted and the naphthalene introduced into the soil by mean of the contaminating solution, implying the hypothesis that all the naphthalene solved into the solution stays in the soil after the contamination process (Fig.7.1). The results obtained, though, disagree from this assumption.

The extraction efficiencies of the extraction runs 3 to 12 range from 56% to 80%, neglecting extraction runs 5 and 11 where major differences have probably occurred during the contaminating solution preparation. The repeatability test extraction efficiencies are though sensibly higher, with values around 88.5%.

54


The operating conditions at which the repeatability test extractions were carried out correspond to those of extraction 7. In that case the efficiency recorded was lower (79.83%). No major anomalies were observed in case of extraction 7, which could relate the lower extraction efficiency to errors introduced during the extraction run or the sample analysis, and the efficiency recorded is coherent to those of extractions 3-12. The lower efficiency is therefore most probably common to all extraction runs from 3 to 12.

Changes in the soil contamination procedures are very likely the reason for the difference. The soil was spiked with a naphthalene-acetone solution for all extractions. In case of extractions 3-12, the acetone was let evaporate in a thermostatic bath at 30°C for 6 hours circa. In case of the repeatability test, the solution was let evaporate slower, for a total time of about 24 hours. A longer time might have allowed a better deposition and adsorption of the naphthalene on the soil, so that the acetone only evaporated. In case of faster evaporation, the naphthalene did not probably have sufficient time for adsorption on the soil and partially evaporated with the acetone, resulting in apparent lower recovery by supercritical extraction. This hypothesis could be verified analysing the contaminated soil with a method of known efficiency. Whether this were verified, it would imply that the extraction efficiencies recorded for extractions 3-12 were underestimated of about 9%. The actual efficiencies would therefore vary between 65% and 89% at the considered operating conditions1.

The evaporation time allowed is indeed a very delicate matter. On one hand, if evaporation is too fast, the evaporating solvent might drag along some contaminant and remove it from the soil (or simply, too fast evaporation means that the evaporating conditions are too favourable, both for the solvent and for the contaminant). On the other hand, if the evaporation time is too long, the naphthalene independent evaporation has more time to take place. There is probably a best evaporating time, as well as evaporating conditions, that minimise the naphthalene losses during the solvent evaporation itself. Those parameters would need to be investigated thoroughly in order to set the best values.

It should be noted that another factor has been regarded as possibly critical by many authors: the contact time between the pollutant and the soil before the extraction. Results are not univocal, as some authors report of reduced recoveries for soils that had been contaminated for long times (ref. Karimi [8] and Pignatello [18]), while others show similar resistance to desorption for both new-spiked spiked soils and for long-contaminated soils (up to 30 years contamination) (ref. Burgos [4] and Connaughton [7]). For the present study, the extraction runs were carried out about 14 hours after the end of the contamination process in case of extractions 3-12, straight after the end of the contamination process in case of extractions 1 and 2 (repeatability test).

The reasons why the naphthalene-soil contact time should influence the overall extraction efficiency (defined as above) is based on the fact that a short contact-time between the naphthalene and the soil might not allow a sufficient time for the contaminant adsorption and penetration into the solid grains; only a mere deposition onto the solid surface might take place, leading to higher contaminant recoveries due to the weaker bounds of the naphthalene to the soil. This would explain Karimi and Pignatello results, but it is not confirmed by other studies. Again, only soil analyses with methods of known efficiency could confirm and quantify the above phenomena.

1 Since the hypotheses discussed were not experimentally verified, comparisons and any further elaboration were though based on the recorded data, regardless of the considerations above.

55


The soil preparation and contamination operations that might lead to error generation are summarised in table 7.1, together with the operations related to the other steps.

Tab. 7.1. Operations that might lead to error generation.

Step Critical operations for error generation

Soil Sampling

Contaminating solution preparation

Solvent evaporation procedure and conditions Soil preparation and contamination

Naphthalene and soil contact time

Soil packing Extraction run

Operating conditions stability

Gas chromatograph calibration

Gas chromatograph precision

Sample preparation Sample analyses

Sample injection

7.1.2. Extraction run The errors possibly generated during the extraction runs mainly relate to the soil

packing into the extraction vessel and to the operating conditions stability (Table 7.1). Uneven compacting of the soil is most likely the cause of major anomalies that

affected some of the extraction runs (chapter 5). In fact, in case of a non-homogeneous compacting of the soil, the CO2 might flow mainly through the looser soil, along preferential routes that prevent it from reaching all parts of the soil. The extraction efficiency recorded is therefore lower than the potential one, since the extraction has taken places only in part of the soil.

This effect was enhanced by the horizontal position of the extractor vessel, required for it in order to fit into the oven. The horizontal position favours a tighter compacting of the soil in the lower half of the vessel, and looser compacting in the upper half.

Furthermore, small soil movements might take place during the extraction, closing some of the preferential routes described above and allowing the supercritical CO2 to get in contact with new soil areas and to extract the contaminant kept shielded until then. This might explain the extraction anomalies observed during some extraction runs (see, for instance, extractions 3 and 4, chapter 5). The phenomenon can be reduced by a better compacting of the soil during the vessel filling or by putting the extraction cell in vertical position. In case of field scale extraction, a vertical flow through the soil together with even distribution over the section should be ensured.

The operating conditions stability is another relevant issue, since the extraction

efficiency is a function of at least some of the set parameters. Major anomalies in some extraction results were observed, most likely related to a poor control of the operating conditions due to instability of the experimental apparatus. Temporary but great changes in the operating conditions were in fact recorded in a few cases, not related to any external measurable cause. In those cases, the extraction properties of the supercritical fluid were affected by wide fluctuations of pressure, mass flow, and density occurred. As it was seen in chapter 6, the extraction efficiency is a function of the mass flow, but

56


shows no significant dependence from pressure or density. The equipment instability is therefore likely to affect the extraction performance because of the mass flow fluctuations created.

Temperature did not suffer from the fluctuations, likely because of the significant thermal capacity of the apparatus, that was able to balance temporary tendencies to changes.

7.1.3. Sample analyses The errors introduced during the sample analyses might be due to the analysis

instrumentation itself (gas chromatograph), or arise from the sample preparation and injection (Table 7.1).

The errors due to the use of a measure instrument are proportional to the instrument precision, in this case the gas chromatograph precision, and to the measuring accuracy, which in turn is dependent from the instrument calibration.

The main errors associated to the sample preparation relate to the weighing of the solution to be analysed and of the ISTD (internal standard) added to the solution itself. Such errors result in a too low or too high estimated naphthalene concentration in the collecting solvent.

The errors generated during the sample injection into the gas chromatograph might lead to double peaks in the chromatogram or to lower estimated concentrations whenever the injection is not instantaneous.

57


7.2. ERRORS EVALUATION For the present analysis, all possible errors shall be divided into two main types: errors

intrinsic to the method, procedure or precision of the instruments, and external errors, defined as errors introduced by the operator during the completion of one operation.

7.2.1. External errors The external errors are due to mistakes done during one or more process operation by

the operator. In principle, these errors can be of any order of magnitude: since they do not depend from the procedure adopted, they do not have any relation with the result obtained.

The external errors may affect the information collected to a more or less high degree. For instance, it may be that a single and limited error occurs and that it is possible to correct the result, eliminating the anomaly. It is the case of the repeatability test, where extraction 1 suffered from a delay, likely due to a slower valve opening at the beginning of the extraction run (see chapter 4). The comparison with the extraction 2 curve allowed the correction of such error. In general, the correction of an external error requires one or more results comparable to the one affected by the anomaly, together with a knowledge of the error that might have occurred.

More often, this type of errors causes non-correctable anomalies that result in loss of all or part of the information otherwise collected.

Total loss of information occurred in the present research in case of extractions 5 and 11. Probable errors during the contaminating solution preparation or the soil spiking resulted in extraction efficiencies outside the possible range (as concluded by comparison with the other extractions performed). In those cases, the extraction runs had to be repeated.

Partial loss of information occurred in case of extractions 3 and 4. Imperfections during the extraction vessel filling and soil compacting, together with poor operating conditions control (see section 7.2.2), caused time delays and double peaks, which reduce the possibility of comparison with other extraction curves and the related eventual considerations and conclusions.

Table 7.2 summarises the external errors recorded during the present research and

what kind of impact they had on further results elaboration and conclusions.

Table 7.2. External errors recorded during the present research, their probable cause and their impact on the results and further elaboration.

External error Probable cause Correction Impact on results

Time delay (1) Slower valve opening Yes None

Time delay (2) Poor operating conditions control Partial Some uncertainties in data elaboration

Double peaks Insufficient soil compacting Partial Some uncertainties in data elaboration

Extraction efficiency out of the possible range

Mistakes during contaminating solution preparation and/or soil spiking

No Total loss of data. Extraction runs re-done.

58


7.2.2. Intrinsic errors The intrinsic errors are those errors that affect the experimental results even if no

mistake is made by the operator. The measure device precision, the stability of the instruments and the procedure adopted determine how precise and how accurate the results are.

It should be noted that the results could be accurate but not precise, the viceversa, neither of the two or both. A result is defined accurate whether the error between the mean value recorded and the actual value is small; it is defined precise whether several results show small differences one from another (Fig.7.2). A calibrated instrument, for instance, is supposed to give accurate results, with a precision depending on the type of measurement and on the instrument itself. But if the instrument has not been calibrated, it will have the same precision but lower accuracy.

a) accurate but imprecise a) inaccurate and imprecise

a) precise but inaccurate a) accurate and precise

Fig.7.2. Graphical representation of accuracy and precision referred to experimental results.

Intrinsic errors affect all steps of the present research, from the soil preparation and contamination, to the performance of the extraction, to the sample analyses (see section 7.1.1). In particular, intrinsic errors are always associated to measuring operations. The measured quantities always differ from the real values, the difference depending on the instrument precision and the measuring accuracy.

The instrument precision is a characteristic of the instrument utilised, and it is usually given. It could also be verified by comparing several results of the same sample measures. The precision of the instrumentation used for the present research is given in chapter 2.

In order to ensure accuracy of the results obtained from measuring operations, the instruments have to be calibrated using samples of known weight, concentration, etc (depending on the measured quantity). During the present research, the scale used was calibrated before each weighing session and during the sessions themselves whether many weighing were required. The gas chromatograph was also calibrated (chapter 3), using a large number of injections of solutions of known concentration. The calibration line obtained fitted the experimental points rather well (R2 = 0.99996).

The measuring results obtained during the present research can therefore be said to be accurate and precise (of course whether no external errors are introduced by the operator).

59


The intrinsic errors due to the weighing operations and to the sample preparation and analyses can therefore be assumed to negligibly affect the final results (Table 7.3).

Many other intrinsic errors may though originate from the remaining operations. In

order to estimate most of such errors, a repeatability test was carried out (chapter 4). The test ideally compares the results of the same extraction done twice on the same soil sample. Theoretically, the same extraction efficiency should be recorded, while in fact there is a difference due to the intrinsic errors.

Operatively, the test consists in the run of two extractions at the same operating conditions on two soil samples spiked together, so that the differences between the samples due to the soil preparation and contamination procedures can be assumed negligible (see also chapter 4). The differences in the results obtained therefore arise only from the following operations (fig.4.1):

− soil sampling − extraction run − sample preparation and analyses − measuring operations The last two issues have already been reported to be sources of negligible intrinsic

errors. The differences due to the soil sampling are in general inversely proportional to the

sample dimension. Because of the soil partitioning technique adopted (section 2.3.2), 3 to 5 g can be considered sufficient to guarantee the sample is representative of the whole soil. In this case, the intrinsic errors associated to the soil sampling operations can thus be assumed to be small, being the soil samples filled into the extraction vessel of about 63 g.

The repeatability test therefore allows an estimation of the intrinsic errors associated to the extraction run. More specifically, such errors concern(Table 7.3):

− the plant structure, i.e. the equipment used (whether it is suitable for the extraction), the warming systems used, etc.

− the operating conditions control, i.e. what devices are used to control, select and adjust the operating conditions, the response time to any operating condition variation, etc.

− the operating conditions stability The operating conditions stability represented a significant problem during some

extraction runs (chapter 5). In those cases, wide fluctuations of the operating conditions were recorded as an effect of little or none operating conditions adjustments by the operator.

The fluctuations were found to be mainly related to pressure sudden peaks and drops generated by the pump. The selected working pressure was set at the pump. As the CO2 was pumped towards the extraction vessel, the pressure at the pump (indicated by the PI01, fig.2.1) slowly decreased until the actual pressure was too low compared to the set value. While trying to bring the pressure back to the set value, the pump generated a sudden pressure drop followed by a fast pressure increase. This caused longitudinal pressure waves that propagated along the line back and forth, thus leading to significant pressure, as well as density and mass flow, fluctuations. The regulating valve V05 (fig. 2.1) was in fact introduced in order to create a local pressure drop and to “break” the pressure waves, preventing them from further propagation along the line to the extraction vessel. Such effect was accomplished introducing local pressure drops higher than 15-20 bars.

60


During the repeatability test, anyway, the operating conditions showed a good stability in case of both extractions, so that no intrinsic errors due to the operating conditions stability presumably affected the final results. In this case, the difference between the extraction efficiency recorded was of 0.49%, i.e. 4.9 ppm for an initial concentration of 1000 ppm (Table 7.3).

It is therefore possible to state that the plant structure and the operating conditions control system selected are able to guarantee, in case of absence of operating conditions fluctuations, precise results.

Whilst the method selected, from the soil sampling to the sample analysis, can thus

ensure precise and repeatable results, worse conclusions can be drawn on the accuracy of such results, affected by the contamination procedure selected.

The comparison of the repeatability test extraction efficiencies to the other extraction runs’ efficiencies (and specifically to the efficiency of extraction 7, that was carried out at the same operating conditions as the repeatability test extraction runs) shows relevant differences, being the repeatability test efficiencies fairly higher than the other efficiencies (section 7.1.1). As already discussed, those differences are most likely due to the soil contamination procedures: different evaporating conditions were in fact set in case of the repeatability test, allowing a longer contact time between the soil and the naphthalene.

The extraction efficiency differences recorded clearly imply that the assumption that all the naphthalene present in the contaminating solution stays in the soil after the solvent evaporation (section 7.1) is not consistent with the results. Compared to the procedure utilised for the repeatability test, the one used for all other extractions caused a significant 9% loss by mean of evaporation of the naphthalene introduced into the soil by the contaminating solution. It is therefore very likely that the extraction efficiencies recorded and utilised for the data elaboration were underestimated of at least 9%. The figure might be even higher since some naphthalene “escape” during the solvent evaporation might have taken place in case of the procedure used for the repeatability test as well.

Table 7.3 summarises the possible sources of intrinsic errors and their relevancy on

the final results.

Tab. 7.3. Intrinsic error sources, and relevancy of each intrinsic error, grouped according to the step where they are generated.

Step Error source Error relevancy

Soil Sampling Negligible

Contaminating solution preparation Negligible Soil preparation and contamination

Solvent evaporation procedure and conditions

Relevant effect on accuracy. Intrinsic errors around 9% or higher.

Plant structure

Operating conditions control Precision around 0.5%.

Extraction run Operating conditions stability Possibly relevant effects in case of

uncontrolled fluctuations.

Gas chromatograph calibration Negligible

Gas chromatograph precision Negligible

Sample preparation Negligible Sample analyses

Sample injection Negligible

61


Summarising, whenever no major errors occurred as a result of operating conditions fluctuations or of external errors introduced by the operator, the results obtained were precise but rather inaccurate. In order to get more accurate results, the extraction efficiency should be calculated as ratio between the naphthalene extracted and the naphthalene present in the soil before the extraction (Eq.7.2):

η = Nextr (Eq.7.2) Nsoil

where Nsoil should be measured each time using a method of known efficiency. As an alternative, chosen any procedure for the soil spiking, the naphthalene losses

during the solvent evaporation could be calculated measuring the naphthalene concentration in the soil with a method of known efficiency (Eq.7.3):

L = Nlost = Nin – Nsoil = 1 _ Nsoil (Eq.7.3) Nin Nin Nin

Whereas: Nlost = amount of naphthalene lost during the solvent evaporation (mg) Nin = amount of naphthalene introduced into the soil with the spiking solution (mg) Nsoil = amount of naphthalene present in the soil after solvent evaporation (mg) and where the naphthalene losses L depend from the soil spiking and evaporating

procedure selected, and could therefore be determined once and then be used as a correction factor for the extraction efficiency calculation. The extraction efficiency as defined in eq.7.2 can then be modified as follows (Eq.7.4):

η = Nextr = Nextr = Nextr (Eq.7.4) Nsoil Nin – Nlost (1 – L) Nin

62

Data interpretation

8. DATA INTERPRETATION

8.1. EXTRACTION RATE The extraction process of a substance from a solid matrix can be divided into three

main consecutive processes, namely: − the desorption of the substance from the solid grain − the diffusion of the substance through the fluid − the advective transport of the substance along the solid bed The adsorption-desorption is a reversible chemical-physical process by which a

substance moves from a fluid phase onto the solid surface or viceversa, according to the substance concentration in the fluid and on the solid surface. In general, a substance desorbes from a solid grain whenever the substance concentration in the surrounding fluid is lower than the one that would be in equilibrium with the substance concentration on the solid surface. In that case, the substance moves from the solid surface into the fluid and could then be transported away.

The pollutant transportation along the solid bed and out of the matrix requires the presence of an advective movement of the extracting fluid through the solid. Even if this is true from a macroscopic point of view, though, the situation might be locally different. Given a constant flow, the instant velocity of the fluid at any given point of the system changes in time, around the mean value. Neglecting this phenomenon and considering the velocity constant at a given point around a solid grain, it though changes as a function of space from a maximum (bulk velocity vb) to zero on the grain surface. Figure 8.1 shows the fluid velocity distribution in the proximity of the solid grain.

v

r

vb

Solid Grain

Fig. 8.1. Velocity distribution of the fluid in the proximity of the solid grain.

Basically, there is a thin region of fluid by the grain (film) that is still or almost still. In that region there cannot be any advective transport due to the fluid movement. In order to get extracted from the matrix, then, the pollutant needs to migrate through the immobile fluid film to regions where there is fluid movement. Such migration takes place by mean of diffusion.

63


The desorption of the substance from the solid surface can be expressed by the following equation (Eq. 8.1):

Φ = k A DF (Eq. 8.1) Φ = substance flux A = solid surface area DF = driving force, function of the concentration difference between the actual concentration

and the equilibrium concentration k = mass transfer coefficient of the desorption process The same equation describes also the diffusion process through the immobile fluid

film. In that case, A refers to the cross-sectional area orthogonal to the diffusion flow direction, DF is a function of the substance concentration difference in the fluid film by the solid surface and the bulk fluid, and k is the mass transfer coefficient of the diffusion process.In both cases, the process rates are determined by the mass transfer coefficients, which are in turn a function of the system conditions.

In the present case, the mass transport rates of the desorption process and of the diffusion process are directly proportional to temperature, and inversely proportional to pressure. Neither process is in principle affected by the fluid flow, since there is no interaction of the fluid movement in the diffusion or in the desorption processes (Eq.8.1). On the other hand, the immobile fluid film layer becomes thinner as the fluid velocity increases. As a consequence, the pollutant that moves through the fluid film by diffusion needs to travel a shorter distance before reaching the region of fluid where the advective movement takes place, therefore crossing the immobile fluid region in a shorter time. Moreover, given the same pollutant concentrations in the fluid film by the grain surface and by the moving region, a shorter distance results in a steeper concentration gradient, thus driving force DF, thus pollutant flux Φ (Eq.8.1).

The advective transport rate increases at higher flows and temperatures. This second effect is due to the lower viscosity, thus better kinetic properties, of the fluid as a result of the temperature increase. On the contrary, higher pressures reduce the advective transport properties of the fluid because of the higher density, and thus worse kinetic properties.

The operating conditions influence on the transport processes that take place during the supercritical fluid extraction are summarised in table 8.1.

Table 8.1. Operating conditions variations and their effect on the transport processes analysed.

Process Process rate

variation Operating conditions Operating conditions variation

Temperature

Pressure Desorption

Flow

Temperature

Pressure External mass transfer diffusion

Flow

Temperature

Pressure Advective transport

Flow

= increase; = decrease; = no (relevant) variation.

64

Data interpretation

The overall extraction rate corresponds to the rate of the slowest process that is said to control or to limit the overall process. For instance, if no fluid flows through the matrix, no extraction takes place because there is no advective transport. In that particular case, the advective transport is the controlling process. As the flow increases, the extraction rate associated to the advective transport increases rather fast, until it eventually becomes higher than the extraction rates associated to the remaining processes. At lower flows, therefore fluid velocities, the film transfer resistance is relevant and represents the controlling factor. If the fluid velocity increases, the film transfer resistance decreases until it eventually is no longer the limiting factor.

Figure 8.2 graphically represents what mass transport process is the extraction rate limiting factor as the supercritical CO2 flow increases.

0 Supercritical CO2 flow

Advectivetransport Diffusion Desorption

Fig.8.2. Extraction rate limiting processes as a function of the supercritical CO2 flow.

In the present study, the CO2 velocities investigated (1 and 1.6 g/min) are already out of the range where the advective transport is the controlling process [13]. That implies also that the pressure and the temperature influence on the advective transport (Tab.8.1) do not transfer into the overall process response to those parameters’ variations, since the advective transport is no longer the overall process limiting factor in the flow range analysed.

The extraction times observed, which are inversely proportional to the overall

extraction rate, displayed moderate differences at different operating conditions (chapter 6.1). It should be noted, though, that the operation conditions intervals investigated were rather limited excepted than for pressure.

An extraction rate increase was recorded at higher temperatures, as expectable because of the temperature favourable effect on all mass transport processes investigated (Tab.8.1), as well as at higher flows.

Since the overall process rate corresponds to the rate of the slowest (limiting) process, if the flow affects the extraction rate it implies that the limiting process is dependent on the mass flow. Only the advective transport and the diffusion through the fluid film are related to the fluid flow (Tab.8.1). In the flow range investigated, though, the advective transport rate is no longer the extraction rate limiting factor. The film transfer resistance is therefore the overall process limiting factor in the flow interval studied. As a consequence, the extraction rate could be further increased by increasing the flow, until the desorption process become the limiting factor (Fig.8.2). This conclusion is a little uncertain because of anomalies occurred during some extraction runs and because of the small flow interval considered (60% variation).

At last, no extraction rate variation was recorded as a function of pressure. This means that the pressure influence on both the desorption and the diffusion mass transfer coefficients can be considered negligible within the pressure range considered.

65


8.2. EXTRACTION EFFICIENCY The extraction efficiencies obtained range from 56.05% to 80.06%. The repeatability

test extractions constitute an exception, with efficiencies around 88.5%. The reasons for this difference are discussed in chapter 7. In all cases, though, the experimental results show a maximum possible extraction efficiency lower than 100%.

The data seem therefore to confirm Connaughton and Burgos conclusions on the resistance to desorption. Both authors recorded a not completely reversibility of the sorption/desorption process both for new spiked soils and for long-contaminated soils. The phenomenon is interpreted as a consequence of the presence of “resistant” and slowly desorbing fractions of pollutant in the soil bed. Basically, part of the pollutant is trapped in immobile fluid present in micropores. In order to get extracted, the pollutant has to migrate to regions where there is fluid movement, and such migration of the pollutant can only be accomplished by mean of diffusion through the immobile fluid present in the pores. This causes kinetic limitation to the extraction process, being the diffusion a low rate process compared to the advective transport.

The results show an extraction efficiency increase at higher CO2 flows. Higher flows

mean higher turbolence: this might allow the fluid to penetrate otherwise “hidden regions” (small pores), therefore allowing more naphthalene extraction from the soil.

Although higher temperatures increase the fluid turbolence and enhance the diffusion processes, the extraction efficiency was though found to increase at lower temperatures. In those conditions, the fluid density is higher, with an increase in naphthalene solubility into supercritical CO2 [3]. This probably favours the pollutant migration towards the bulk fluid, and this effect prevails on the negative ones.

The density, and thus the naphthalene solubility into supercritical CO2, increases as well at higher pressures. But pressure changes within the interval studied seemed not to have any effect on the extraction efficiency. The pressure has in fact a number of counteracting effect on naphthalene extraction from micropores. As already stated, an increase in pressure reduces diffusion rates, thus preventing the naphthalene from extraction. Additionally, higher pressure mean worse transport properties. The most significant effect is though likely to be the tighter compacting of the soil at higher pressures, resulting in the creation of more and smaller pores where the fluid cannot flow. On average, the positive and negative pressure effects on the naphthalene extraction from the micropores balance.

66

Costs

9. COSTS

The costs related to soil decontamination by supercritical CO2 extraction can be divided into fixed costs and variable costs. The variable costs are a direct function of the operating conditions, while the fixed costs mainly relate to the investment for the extraction plant construction. The knowledge of the operating condition range at which the plant has to work is fundamental for the calculation not only of the variable costs, but of the fixed costs as well. In fact, the extraction equipment is designed to suit specific parameter ranges. It is therefore necessary to define what operating conditions will be adopted before the calculations of the costs.

9.1. DESIGN PARAMETERS The variable costs are related to the operating conditions and to the extraction time. In

general, such costs vary as any of the parameters is modified, within the range allowed by the extraction plant design (section 9.2). Because of the considerations carried out at chapters 6 and 8, the best operating conditions for cost minimisation are the following:

− Temperature: the extraction efficiency showed a sensible increase at lower temperatures. Since the costs decrease at lower temperatures as well, it is then recommended to adopt the lowest temperature possible, considering that the carbon dioxide reaches supercritical conditions at about 31°C. An operating temperature of 33-35°C seems thus suitable to ensure supercritical conditions.

− Mass flow: the extraction efficiency rises at higher flows (thus superficial velocities), as well as the costs. In order to find the best mass flow that ensures sufficient contaminant recovery at the lowest costs, it is first necessary to define the target recovery. The recovery required is calculated on the basis of the initial contaminant concentration and the maximum allowed level (250 ppm in an industrial site). Known the target extraction efficiency and the operating temperature (33-35°C), the mass flow can be calculated by mean of equation 6.2.

− Pressure: in the range considered, the pressure seems not to influence the extraction efficiency, nor the extraction time. It is therefore recommended to work at the lowest pressure possible. This would reduce the energy consume (variable cost), but most of all it would reduce the fixed costs, since the equipment cost rises significantly with the pressure that the materials will have to withstand. Although pressure changes does not affect the extraction in the examined interval, it is in principle possible that the system behaviour outside that range would be different. The use of the lowest investigated pressure, i.e. 120 bar, is therefore suggested.

− Extraction time: the contaminant recovery percentage reaches the asymptote value after a maximum time of 40 minutes during all extractions.

The design parameters for a supercritical extraction plant, suitable for recovery of soil

with an initial concentration up to 7500 ppm circa, are summarised in table 9.1.

67


Tab. 9.1. Design parameters for a supercritical CO2 extraction plant.

Parameter Measure unit Value MEDIUM PROPERTIES

Treated medium - Contaminated soil

Contaminant - Naphthalene

Contamination level (max) Ppm up to 7500

Textural analysis of soil - Sand

Bed porosity - 0.4

Bed density Kg/m3 1500

OPERATING CONDITIONS

Extraction temperature °C 33-35

Extraction pressure bar 120

CO2 density G/cm3 0.75

Volumetric flow m3/h up to 38

Extraction time min 40

Operating time 8 h/d, 360 d/y

9.2. VARIABLE COSTS The costs are defined variable whether they are a function of the operating conditions.

Since the temperature, the pressure and the extraction time required to maximise the extraction efficiency and to reach its asymptote are though constant (section 9.1), all variable costs that are a function of these three parameters only are constant in turn.

On the contrary, the flow varies as a function of the initial concentration. A maximum

volumetric flow of 38 m3/h guarantees the decontamination of 10 m3 of soil (see plant construction design, section 9.3) with an initial naphthalene concentration up to 7500 ppm. Lower flows can be used in case of lower initial concentrations, resulting in reduced costs, while higher flows will be selected in case of higher required extraction efficiency, thus in case of higher initial concentrations.

The flow variation influences a number of costs, i.e. the electricity cost for the pump, the heating (as the mass to warm up changes), etc. The cost variation related to them is though negligible if compared to the effect on cost variation due to the CO2 purchase, being the cost for the CO2 supply linear with the mass flow.

Table 9.2 summarises the operating costs for the supercritical extraction plant based

on the design parameters of table 9.1 and for a vessel extraction volume of 10 m3 (Tab.9.3). The CO2 supply cost calculated is the maximum possible cost, based on the hypothesis of maximum flow, which would be required for the treatment of a soil with an initial contaminant concentration of 7500 ppm, and CO2 discharge after the extraction. The major operating cost is the CO2 supply. The CO2 clean up and recycling, i.e. with activated carbon, would significantly reduce the CO2 supply cost, limited to the make-up only (see also section 9.4).

68

Costs

Tab.9.2. Operating cost for a supercritical extraction plant.

Item description Remarks Cost ($/m3)

Transfer of soil into vessel Excavation & Transportation 40

CO2 supply 0.1 $/kg 493 (max)

Operating labour (L) 1 per shift, 15 $/h 2.56

Direct supervision (S) 15 $/h 2.56

Utilities (electricity, heating) 5 $/h 0.85

Maintenance (M) 4% of T.E.C. (Tab.9.3) 2.02

Operating supplies 15 % of maintenance 0.3

Total operating costs 541.66

9.3. FIXED COSTS The fixed costs refer to the capital costs for the extraction plant construction, plus the

administrative costs and other costs not directly dependent from the operating conditions. Concerning the equipment construction, two extraction vessels are employed in order

to allow “continuous” operation, i.e. one vessel performs the extraction while the second gets emptied from the cleaned soil and loaded with new contaminated soil.

Cost calculations for the extraction vessel and the pump (Tab.9.3) are based on the

data by Peters and Timmerhaus [16]. Since the data refer to year 1979, the calculated value have been transported to year 2001 by considering the money depreciation in the USA between the mentioned years. All other costs were taken from Montero et al. [12].

Tab. 9.3. Equipment costs for a transportable supercritical extraction plant.

Equipment description Design pressure Size Estimated cost ($)

Extraction vessels 206 bar 2 × 10 m3 236,700

High pressure liquid pump 206 bar 15,900

Valves and accessories 206 bar 5,400

Instrumentation 206 bar 36,000

Total equipment costs (T.E.C.) 294,000

The total fixed costs (table 9.4) are obtained by calculating the equipment depreciation

over 15 years, plus the administrative costs and the plant overhead (heating, light, rent, etc.) The equipment depreciation is calculated on the hypothesis of 8 h/d 360 d/y plant operation. Data from Montero at al. [12].

Tab. 9.4. Fixed costs for supercritical extraction process.

Remarks Cost ($/m3)

Depreciation Equipment costs over 15 years 19.6

Plant overhead 60% of (L + S + M) (Tab.9.2) 4.284

Administrative cost 15% of (L + S + M) (Tab.9.2) 1.071

Total fixed costs 24.955

69


9.4. TOTAL COSTS The total cost for naphthalene soil remediation, with an initial contaminant

concentration of 7500 ppm, is equal to:

Total Cost = Variable Cost + Fixed Costs

Total Cost = 541.66 $/m3 + 24.96 $/m3 = 566.62 $/m3 The CO2 supply cost, equal to 493 $/m3, represents the 87% of the total cost. The

introduction of a CO2 recycling step would reduce to a great extent this cost. For instance, if 90% of the CO2 used is recovered (for instance by adsorption of the impurities on activated carbon) and recycled, the CO2 supply cost would be reduced to 49.3 $/m3.

The introduction of a recycling step requires the introduction of suitable equipment in the extraction plant, that would increase the fixed costs. The increase was calculated by Montero, and it amounts to 50% of the other investment costs. The total cost calculation would therefore be modified as follows:

Total Cost (with CO2 recycling) = 97.96 $/m3 + 34.76 $/m3 = 132.72 $/m3

If compared to other remediation processes (Tab.9.5), the supercritical CO2 extraction

shows relevant economical advantages.

Tab.9.5. Comparison of the soil remediation costs of different methods.

Remediation process Cost ($/m3)

Supercritical fluid extraction (without CO2 recycling)2 567 (max)

Supercritical fluid extraction (with CO2 recycling)2 133 (max)

Bio-clean 191 – 370

Acurex solvent wash 196 – 569

KPEG 211 – 378

Supercritical water oxidation 250 – 733

Vitrification 255 – 548

Chemical waste landfill 260 – 490

O.H.M. methanol extraction 400 – 514

Soilex solvent extraction 856 – 913

Incineration 1713 – 1826

2 The cost refers to remediation of soil contaminated by a naphthalene concentration of 7500 ppm. Lower concentrations imply lower CO2 flow thus lower costs (the major cost is in fact due the CO2 supply).

70

Conclusions

10. CONCLUSIONS

The results obtained during the present research indicate the supercritical CO2 extraction as an adequate and cost effective method for naphthalene removal from contaminated soils. In particular, remediation costs as low as 133 $/m3 make the process very competitive compared to more traditional methods, such as landfilling, solvent extraction or biological remediation.

The extraction efficiencies achievable by means of SFE (supercritical fluid extraction) would ensure, in case of most contaminated industrial sites, the recovery required to bring the naphthalene residual concentration below the law limits. This result refers to the treatment of sandy soil with limited moisture content, and for naphthalene initial concentrations up to 7500 ppm.

The operating conditions that influence the process performances were found to be temperature and mass flow, with an extraction efficiency increase at lower temperatures and higher flows respectively. The limited number of the process controlling parameters would imply a rather easy operability of the process itself in an hypothetical full-scale application.

The error analysis and evaluation shows that the method followed and the extraction

equipment used for the present research were able to guarantee a good precision yet a rather low accuracy of the results. The data elaboration done is therefore affected by a not negligible degree of uncertainty, which limits to some extents the validity of the calculations performed.

71


LITERATURE REFERENCES

[1] Andersson, B., E., Tornberg, K., Henrysson, T., Olsson, S., Bioresource Technology – Vol. 78 – Three dimensional outgrowth of a wood-rotting fungus added to a contaminated soil from a former gasworks site, pages 37-45, 2001.

[2] Andersson, B., E., Welinder, L., Olsson, P., A., Olsson, S., Henrysson, T., Bioresource Technology – Vol. 73 – Growth of inoculated white-rot fungi and their interactions with the bacterial community in soil contaminated with PAH, as measured by phospholipid fatty acids, pages 29-36, 2000.

[3] Bartle, K., D., Clifford, A., A., Jafar., S., A., Shilstone, G., F., Journal of Physical Chemistry Reference Data – Vol. 20, no. 4 – Solubilities of Solids and Liquids of Low Volatility In Supercritical Carbon Dioxide, 1991.

[4] Burgos, W., D., Novak, J., T., Berry, D., F., Environmental Science & Technology – Vol. 30, no. 4 – Reversible Sorption and Irreversible Binding of Naphthalene and α-Naphthol to Soil: Elucidation of Processes, pages 1205-1211, 1996.

[5] Canet, R., Birnstingl, J., G., Malcolm, D., G., Lopez-Real, J., M., Beck, A., J., Bioresource Technology – Vol. 76 – Biodegradation of PAH by native microflora and combinations of white-rot fungi in a coal-tar contaminated soil, pages 113-117, 2001.

[6] Clifford, T., Fundamentals of Supercritical Fluids, Oxford University Press, 1999.

[7] Connaughton, D., F., Stedinger, J., R., Lion, L., W., Shuler, M., L., Environmental Science & Technology – Vol. 27, no. 12 – Description of Time-Varying Desorption Kinetics: release of Naphthalene from Contaminated Soils, pages 2397-2403, 1993.

[8] Karimi-Lotfabad, S., Pickard, M., A., Gray, M., R., Environmental Science & Technology – Vol. 30, no. 4 – Reactions of Polynuclear Aromatic Hydrocarbons on Soil, pages 1145-1151, 1996.

[9] Kiely, G., Environmental Engineering. Mc Graw Hill International (UK) Limited, 1997.

[10] Khodadoust, A., P., Bagchi, R., Suidan, M., T., Brenner, R., C., Sellers, N., G., Journal of Hazardous Materials – B80 – Removal of PAHs from highly contaminated soils found at prior manufactured gas operations, pages 159-174, 2000.

[11] Lee, C. M. and Gongaware, D. F., Environmental Technology – Vol. 19 – Effects of Selected Soil Characteristics on the Removal of Diesel Fuel by Supercritical Carbon Dioxide, Publication Division Selper Ltd., 1998.

[12] Montero, G.A., Giorgio, T. D., Schnelle, K. B. Jr., Environmental Progress – Vol. 15, No. 2 – Scale-Up and Economic Analysis for the Design of Supercritical Fluid Extraction Equipment for Remediation Soil, 1996.

72

Literature references

[13] Montero, G.A., Giorgio, T. D., Schnelle, K. B. Jr., Innovations in Supercritical Fluid Science and Technology – Removal of Hazardous Contaminants from Soils by Supercritical Fluid Extraction, pages 281-297, Edited by Keith W. Hutchenson and Neil R. Foster, 1995.

[14] Notar, M., Leskovsek, H., Fresenius' Journal of Analytical Chemistry, No. 358 – Optimisation of supercritical fluid extraction of polynuclear aromatic hydrocarbons from spiked soil and marine sediment standard reference material, pages 623-629, 1997.

[15] Perry, R., H., Chilton, C., H., Chemical Engineers' Handbook, fifth edition, McGraw-Hill Kogakuska, Ltd., 1979.

[16] Peters, M., S., Timmerhaus, K., D., Plant Design and Economics for Chemical Engineers - Third Edition, McGraw-Hill International Editions, Singapore, Chemical Engineering Series, 1988.

[17] Piatt, J., J., Backhus, D., A., Capel, P., D., Eisenreich, S., J., Environmental Science & Technology – Vol. 30, no. 3 – Temperature-Dependent Sorption of Naphthalene, Phenanthrene, and Pyrene to Low Organic Carbon Aquifer Sediments, pages 751-760, 1996.

[18] Pignatello, J., Journal of Environmental Toxicology and Chemistry, 9, 1117-1126, 1990.

73

Documents

SOIL REMEDIATION BY SUPERCRITICAL CO EXTRACTION …Soil remediation by supercritical CO 2 extraction: process performance evaluation and cost analysis ACKNOWLEDGMENTS Many people have