Development of optical measurement techniques and data analysis methods for screening of adulterated diesel oils

(1)

uef.fi

PUBLICATIONS OF

THE UNIVERSITY OF EASTERN FINLAND Dissertations in Forestry and Natural Sciences

ISBN 978-952-61-2893-1 ISSN 1798-5668

Dissertations in Forestry and Natural Sciences

DISSERTATIONS | BONIPHACE ELPHACE KANYATHARE | DEVELOPMENT OF OPTICAL MEASUREMENT... | No 316

BONIPHACE ELPHACE KANYATHARE

DEVELOPMENT OF OPTICAL MEASUREMENT TECHNIQUES AND DATA ANALYSIS METHODS FOR SCREENING OF ADULTERATED DIESEL OILS

PUBLICATIONS OF

THE UNIVERSITY OF EASTERN FINLAND

Fuel adulteration is one of the contributors to climate change. As a partial solution to fuel adulteration, this thesis proposes novel approaches by using handheld refractometer

and novel handheld sensor to combat fuel adulteration especially in field conditions.

Moreover, for the first time in applied optics the concepts of excess permittivity and imaginary excess permittivity are applied to resolve complex fuel adulteration problems.

These novel approaches save as new openings demonstrating the potential of excess permittivity analysis in fraud fuel detection.

BONIPHACE ELPHACE KANYATHARE

30887153_UEF_Vaitoskirja_NO_316_Elphace_Boniphace_LUMET_cover_18_09_12.indd 1 12.9.2018 8.51.23

(2)

(3)

DEVELOPMENT OF OPTICAL MEASUREMENT TECHNIQUES AND

DATA ANALYSIS METHODS FOR SCREENING OF ADULTERATED

DIESEL OILS

(4)

(5)

BONIPHACE ELPHACE KANYATHARE

DEVELOPMENT OF OPTICAL MEASUREMENT TECHNIQUES AND

DATA ANALYSIS METHODS FOR SCREENING OF ADULTERATED

DIESEL OILS

Publications of the University of Eastern Finland Dissertations in Forestry and Natural Sciences

No 316

University of Eastern Finland Joensuu

2018

Academic dissertation

To be presented by permission of the Faculty of Science and Forestry for public examination in the Auditorium M100 in the Metria Building at the University of Eastern Finland, Joensuu, on October 15, 2018, at 12

o’clock noon

(6)

Grano Oy Jyväskylä, 2018

Editors: Pertti Pasanen, Matti Vornanen, Jukka Tuomela, Matti Tedre

Distribution: University of Eastern Finland / Sales of publications www.uef.fi/kirjasto

ISBN: 978-952-61-2893-1 (nid.) ISBN: 978-952-61-2894-8 (PDF)

ISSNL: 1798-5668 ISSN: 1798-5668 ISSN: 1798-5676 (PDF)

(7)

Author’s address: University of Eastern Finland Depart. of Physics and Mathematics P.O. Box 111

80101 JOENSUU, FINLAND email: boniphk@uef.fi

Supervisors: Professor Kai-Erik Peiponen University of Eastern Finland Depart. of Physics and Mathematics P.O. Box 111

80101 JOENSUU, FINLAND email: kai.peiponen@uef.fi

Professor Seppo Honkanen University of Eastern Finland Depart. of Physics and Mathematics P.O. Box 111

80101 JOENSUU, FINLAND email: seppo.honkanen@uef.fi

Reviewers: Associate Professor Juha Toivonen Tampere University of Technology Laboratory of Photonics

P.O.BOX 692

33101 TAMPERE, FINLAND

email: juha.toivonen@tut.fi

Adjunct Professor Ilpo Niskanen University of Oulu

Thule Institute Research Centre P.O. Box 8000

OULU, FINLAND

email: ilpo.niskanen@oulu.fi

Opponent: Professor Valerio Lucarini University of Reading

Department of Mathematics & Statistics Whiteknights, P.O.BOX 220

READING RG6 6AX, UK email: v.lucarini@reading.ac.uk

(8)

(9)

i Kanyathare, Boniphace Elphace

Development of Optical Measurement Techniques and Data analysis Methods for screening of adulterated Diesel oils.

Joensuu: University of Eastern Finland, 2018 Publications of the University of Eastern Finland Dissertations in Forestry and Natural Sciences 2018; 316 ISBN: 978-952-61-2893-1 (print)

ISSNL: 1798-5668 ISSN: 1798-5668

ISBN: 978-952-61-2894-8 (PDF) ISSN: 1798-5676 (PDF)

ABSTRACT

Climate change is one of the major challenges and threats facing our planet, with one of its main contributors being environmental pollution, which is caused in part by fuel adulteration. Because of cheap price of kerosene in developing countries as compared to other fuels, adulteration of diesel oils by kerosene is highly prevalent in those countries, which is also among the difficult cases to screen by conventional techniques in particular when the volume of kerosene is below 20%. As a remedy to this problem, we have developed and demonstrated several optical measurement techniques and data analysis methods for screening very low levels of kerosene adulteration in diesel oils, namely 5%, 10% and 15%.

This we have achieved firstly, by developing an optical sensor which is a modification of commercial handheld gloss meter, by incorporating a removable sensor head with roughened glass for screening of the afore mentioned low levels of adulteration. Secondly, we have developed two data analysis methods that make use of the refractive index measurements as well as transmittance data inversion using singly subtractive Kramers-Kronig (SSKK) relations. The results of SSKK were applied to determine wavelength-dependent relative excess permittivity and wavelength-dependent imaginary excess permittivity, which not only reveal the hidden spectral finger prints, but also discriminate different adulteration levels. Thirdly, we have also demonstrated the application of cheap commercial handheld refractometer (traditionally used for glucose concentration measurements), for fuel adulteration detection as an alternative especially for poor countries.

This thesis demonstrates two novel ideas, namely it reports on excess permittivity study of diesel oils adulterated by kerosene, and on the application of general Kra- mers-Kronig (K-K) analysis which is based on the principle of causality, for assessing both real excess permittivity and imaginary excess permittivity of binary liquid mixtures, especially for the case of adulterated fuels.

(10)

ii

We believe that the concepts demonstrated in this thesis will open more doors for other possible applications far beyond liquid fuels to other areas such as olive oil, palm oil, and even liquid food adulteration. Moreover, we do believe that the results of this work will aid into the development of more affordable and portable field- based sensors for liquid fuel purity measurements.

Universal Decimal Classification: 535.324, 665.7.035.7, 665.753, 681.785.2

Library of Congress Subject Headings: Liquids—Optical properties; Liquid fuels; Diesel fuels; Adulterations; Kerosene; Optics; Optical measurements; Optical detectors; Refractive index; Refractometers; Quality control

Yleinen suomalainen asiasanasto: nesteet; moottoripolttoaineet; dieselöljy;

väärentäminen; petroli; optiikka; optiset ominaisuudet; optiset anturit; laadunvalvonta

(11)

iii

ACKNOWLEDGEMENTS

I have grown old enough to understand and admit the undeniable fact that we as individuals cannot survive and get ahead in life as island, without the support and cooperation of people around us. In fact, the realization of this thesis is a testimony to this fact. Regardless of our nationalities, ethnicities or race, we can all cooperate and work together to make our world a better place for future generations.

From the bottom of my heart I would like to offer my sincere and deep grounded gratitude to my supervisor Prof. K. -E. Peiponen for enabling me to come this far in such short duration. Dear Kai, to tell it plainly you have played more than a fatherly role, your guidance, coaching, motivation, correction, encouragement, dedication and patience, has played a key and vital role without which I wouldn’t have come this far. All I can say is, God bless you. To my second supervisor and past head of Physics and Mathematics department Prof. S. Honkanen, thank you for your inspi- ration and for accepting to becoming part of this awesome journey.

To co-authors Dr. J. Räty and Dr. K. Kuivalainen, thank you for the samples and for the measurements which contributed to this work. To co-authors Dr. P. Bawuah and P. Silfsten, thank you for your overwhelming support and cooperation from the earliest stages of this work. To Dr. M. Markinnen and Dr. M. Silvennoinen thank you for the countless hours you spent with me in the lab taking measurements. To Msc.

B. Asamoah, thank you for time dependent measurements and for your collaboration.

I wish to express my sincere gratitude and appreciation to other members of the department of physics and mathematics UEF, Noora Heikkilä (coordinator), Prof. P.

Vahimaa (Director of the institute of photonics), Prof. T. Jääskeläinen (Former HOD), Prof. M. Kuittinen, Prof. M. Roussey, Prof. Y. Svirko, and lastly but not least Dr. K.

Saastamoinen. Your support, cooperation and guidance during my studies created a special and awesome environment, that has enabled me to come this far. I say a big thank you.

To my Wife, Mother, and children. Thank you for enriching my life and for put- ting smile on my face constantly.

This thesis is dedicated to Aino Peiponen who passed away during the final stages of this thesis, her memory shall remain in our hearts because it is through her that Kai come into this world.

Finally, I acknowledge the sovereign God for free breath of life, that he has granted us. Moreover, I am greatful for the lessons that he has taught me through challenging as well as wonderful life experiences. These experiences have forever changed my life for the better.

Joensuu, 15^thof October 2018 Boniphace Elphace Kanyathare

(12)

iv

LIST OF ABBREVIATIONS

GC-MS Gas chromatography-mass spectrometry IC-MS Inductively coupled mass spectrometry HPLC High performance liquid chromatography

CD Compact disk

DVD Digital video disk NIR Near infrared

PCA Principal component analysis LDA Linear discriminant nalysis PCR Principal component regression PLS Partial least square

SIMCA Soft independent modelling of class analogy

ASTM American Society of Testing and Materials international

GC Gas chromatography

EN European Norms

Vis Visible light

SSKK Singly subtractive Kramers-Kronig analysis KK Kramers-Kronig analysis

LPG Liquified petroleum gas ATF Aviation turbine fuel PM Particulate matter

PAHS Polycyclic aromatic hydrocarbons VOC Volatile organic compounds MIR Mid Infrared

FTIR Fourier Transform Infrared SFS Synchronous fluorescence scan

EEMF Excitation emission matrices fluorescence ISO International organization for standardization PACS Polycyclic aromatic compounds

MTBE Methyl tetra butyl ether

EWURA Energy and water utilities regulatory authority MSC Multiplicative scatter correction

UV Ultraviolet light AC Alternating current DOE Diffractive optical element CGH Computer generated hologram

(13)

v

LIST OF ORIGINAL PUBLICATIONS

This thesis is based on data presented in the following articles, referred to by the Roman Numerals I-IV.

I. Kanyathare B., Kuivalainen K., Räty J., Silfsten P., Bawuah P. &

Peiponen K.-E. 2018. A prototype of an optical sensor for the identifica- tion of diesel oil adulterated by kerosene. Journal of European Optical So- ciety Rapid Publications 14: 1-6.

II. Kanyathare B. & Peiponen K.-E. 2018. Wavelength-dependent excess permittivity as indicator of kerosene in diesel oil. Applied Optics 57: 2997- 3002.

III. Kanyathare B., Asamoah B. & Peiponen K.-E. 2018. Imaginary excess permittivity in NIR-spectral range for separation and discrimination of adulterated diesel oil binary mixtures (submitted).

IV. Kanyathare B. & Peiponen K.-E. 2018. Handheld Refractometer based measurement and excess permittivity analysis method for detection of diesel oils adulterated by kerosene in field conditions. Sensors 18, 1551.

(14)

vi

AUTRHOR’S CONTRIBUTION

The ideas for all the papers in this thesis were developed based on intensive discussions between the author and his Supervisor. The first paper was a result of nice group effort where the author played a key role in all the measurements as well as manuscript writing. In the subsequent three works, the author did most part of the measurements including all computations and wrote the articles in collaboration with and guidance of the supervisor. The author dealt with all submissions and all the collaborations during the entire review and publication process, as the main and corresponding author.

(15)

vii

1 INTRODUCTION

The problem of fuel adulteration continues to negatively impact our planet and it is present at alarming levels not only in developing countries and third world countries, but also in some parts of Europe (Bhanu P., 2017; Kalligeros S., 2003). Petroleum products are the backbone of any economy in the world because of high demand, among these, kerosene is usually subsidized especially for poor and developing countries. For these countries, kerosene is therefore a lucrative product for adulteration, because it has similar chemical properties with diesel oil and gasoline (Majhi A., 2012; Bhanu P., 2017).

Liquid fuel adulteration results into environmental pollution which is the big challenge facing the globe currently. It is one of the major contributors to atmospheric ozone pollution (Marais E. A., 2014), and soot carbon (Schuster G. L., 2016). Moreo- ver, adulteration contributes to black carbon, which further increases earths average temperature inducing more far reaching negative impacts (Majhi A., 2012; Speight J.

G, 2015; Gupta A. & Sharma R, 2010; Sinha S. N. & Shivgotra V, 2012; Sadat A, 2014).

Liquid fuel adulteration is not the only type of adulteration affecting the world community, the other more severe category of liquid adulteration has direct impact on human health because it deals with liquid foods as well as oils that are consumed by humans. These includes, olive oil adulteration (Isabel D. M., 2018), alcohol products such as wines and whiskey (Lachenmeier D. W. & Rehm J, 2016; Lachenmeier D. W. & Rehm J, 2013; Lachenmeier D. W., 2007). Moreover, beverage adulteration is also a big issue (Maireva S., 2013; Ogrinc N., 2003), and milk adulteration (Tanzina A. & Shoeb A, 2016; Moore J. C., 2012). Milk adulteration is even more severe, because it affects the lives of infants. Even though, the mentioned types of adulteration are not dealt with in this work, we do believe that the findings of this work can aid in resolving some of those.

Traditionally liquid measurements and identification is based on density utilizing hydrometers (Gupta S, 2002; Guay R. R., 1983; Demin V, 2007). However, simple density of liquids measurement can be misleading in the case of sensitive measurement requirements like adulterated diesel oil by kerosene, because of very close density values between the two constituent liquids. Other than density measurements, other traditional measurements include gas chromatography-mass spectrometry (GC-MS), inductively coupled-mass spectrometry (IC-MS), and high performance-liquid chromatography (HPLC) (Bhanu P., 2017), these methods are invasive, laborious and time consuming. To overcome these challenges and enable timely and accurate measurement of liquids, optical based measurement methods such as refractive index and optical spectroscopy are potential candidates.

The refractive index of a material whether solid or liquid is not only directly related to the density of that material, but also related to temperature and pressure.

This implies that, the purity of material as well as other external factors aﬀects the

(18)

2

refractive index. Therefore, when the adulterants are added into pure fuel samples, purity of the samples is altered, leading to change in refractive index (Räty J. & Peipo- nen K. -E, 2015). Refractive index readings can be utilized to characterize materials, either on their own or can further be applied in advanced data analysis methods for purity identification. This quantity has been utilized for purity measurement of liquids in different studies (Payri R., 2013; Geacai S., 2012; Polynkin P., 2005; Kim C. -B.

& Su C. B, 2004; Magnusson R., 2010; Fernandes V. H, 2008; Mishra V., 2008;

Ariponnammal S., 2012).

The other very important arm of optical measurement techniques is optical spectroscopy, thanks to the presence of organic molecules fingerprints in infrared spectral region, the quality of liquids such as fuels can be tested (Workman J, 1996; Keifer J, 2015). This is possible because material molecules such as hydrocarbons and other functional groups, absorb the incident electromagnetic radiation especially in the near infrared (NIR) spectral range.Recently, the applications of optical spectroscopy have moved further beyond the early traditional medical and industrial ﬁelds, to se- curity both in civil and military environments (Peiponen K. -E. & Saarinen J. J, 2009).

Moreover, these techniques are also prevalent in consumer devices such as computer mice, ﬁngerprint readers, barcode scanners as well as CD and DVD’s (Oksman A, 2008).

The NIR spectra is achieved through spectrophotometric measurements which scans the wavelength of the incident radiation either in transmission, reflectance or absorbance mode. these types of measurements have been demonstrated and applied not only in the rapid measurement of diesel engine oil quality, but also in the investigation of chain of custody for crude oil samples (SongQing Z., 2012; Soraya S. B., 2006). Diesel oils and kerosene have overlapping spectral fingerprints in the NIR region which complicates the adulteration detection process and is probably the reason for the failure of techniques for screening adulteration levels below 20%. Therefore, the process of differentiating adulterated fuels only based on either transmittance or reflectance measurement data alone is usually impossible. Not long ago, several interesting measurements and data analysis methods were proposed for screening of fake liquid fuels (Taksande A. & Hariharan C, 2006; Bassbasi M., 2013; Paiva E. M., 2015). Moreover, the liquid purity studies detection using optical spectroscopy have grown because of studies of biodiesel and bioethanol products (Golebiowski J. & Pro- hun T, 2008; Kontturi V., 2011; Ventura M., 2013).

The potential of optical spectroscopy in liquid fuel studies is unquestionable. In recent years terahertz time-domain spectroscopy has also found way in these studies.

It has been applied to quantify gasoline-diesel mixtures (Yi-nan L., 2013), in the analysis of petroleum products and their mixtures (Yun-Sik J., 2008), in discriminating gasoline fuel contamination in engine oils (Abdul-Munaim A. M., 2018). Moreover, it has been applied for studying the dielectric properties of diesel and gasoline (Arik E., 2014), for quantitative distinction of gasoline mixtures (Li Y., 2013), in sensing of petroleum industrial applications (Al-Douseri F. M., 2006), and finally in determining spectral features of commercial derivative fuel oils (Zhao H., 2012).

(19)

3 Another method based on spectroscopy that is gaining acceptance in fuel adulteration studies is multivariate analysis such as PCA, PCR, PLS, SIMCA, and many others. Recently, the potential of these methods was demonstrated by several re- searchers (Marcio J. C. P., 2011; Fazal M., 2017; Silva A.C., 2012) to uncover liquid fuel adulteration. One of the reasons why multivariate analysis is necessary in fuel adulteration studies is that, the traditional testing methods such as ASTM 4052, and EN 14078 measure the absorbance at only one wavelength (Marcio J. C. P., 2011).

These methods can be compromised especially for the case of diesel and kerosene, because they have the same chemical composition. Multivariate analysis has lots of potential in solving different liquid measurement challenges as well as solid samples and other materials. However, it requires many samples and knowledge about sta- tistical methods (Swierenga H, 1995). These methods are very efficient especially under laboratory conditions but requires a broad spectral range which is a drawback especially in field conditions.

This thesis focuses on studying adulteration of diesel with kerosene, because most of the existing methods in literature for fuel adulteration deal with a relatively simple case in the sense of measurement, namely, gasoline adulteration using kerosene. For this purpose, the complex excess relative permittivity was introduced. The excess relative permittivity of liquid mixtures is highly important in assessing the interactions between different molecules of different liquid components in the liquid mixture (Reis J. C., 2009). Depending whether the excess permittivity value is positive, negative or zero, it can be exploited for estimating the presence or absence of dipole- dipole interactions between the liquid molecules (Ahire S., 1998). Theoretical foun- dations on the complex excess permittivity that depends on the frequency of the incident electromagnetic field has been derived quite recently for liquid mixtures (Ig- lesias T. P. & Reis J. C. R, 2016). Moreover, sensor solutions based on the measurement of the complex permittivity has been suggested, such as, for the inspection of contamination of motor oil (Perez A. T. & Hadfield M, 2011). Likewise, similar concepts were applied for investigation of excess absorbance and the utilization of classical Lorentz oscillator model of binary liquid mixtures (Baranovic G, 2017). Further- more, the study of terahertz time-domain spectroscopy of water and alcohol mixtures (McGregor J., 2015), and the study of ionic liquid mixtures incorporated with quan- tum mechanical interpretation by the concept of dampened harmonic oscillator (Mou S., 2017). On the second hand, interesting studies on dynamic behavior of molecular two-layered nanofilms have been studied in NIR region for accessing permittivity and other optical properties of such structures (Rodic D., 2013; Setrajcic J. P, 2017).

For this research we utilize the idea of light dispersion of authentic diesel samples namely diesel samples for varying climatic conditions, and kerosene. Then we study the dispersion properties of fake fuels, this involves mixtures of various diesel samples and kerosene. We investigated dispersion properties by measuring transmission spectrum in the Vis-NIR using spectrophotometer, and refractive index using Abbe refractometer. In this thesis it is demonstrated that, the refractive index data obtained using Abbe refractometer which is accurate table measurement device, and Vis-NIR

(20)

4

spectrophotometer, can be combined and used for novel type of diesel oils adulteration detection by utilizing the excess permittivity property. For this purpose, the singly-subtractive Kramers-Kronig (SSKK) dispersion relation which is based on transmittance data inversion is applied, in the calculation of wavelength-dependent refractive index of liquids (Lucarini V., 2005). SSKK was investigated and applied in several studies where its potential was demonstrated (Heijin S., 2017; Herrmann M., 2012; Unuma T., 2011). In this thesis based on our knowledge, for the first time we report on the application of Kramers-Kronig (KK) analysis, which is based on causality principle, to assess excess permittivity of adulterated fuels which are binary liquid mixtures.

This dissertation reviews and discuss about liquid fuels as well as their adulteration in chapter 2. The theory of light interaction with the roughened glass-fuel interface, the wetting property of liquids (contact angle), Beer-Lambert’s law, complex refractive index, the SSKK relations, complex excess permittivity, Lorentz-Lorenz formula, as well as the modified ideal law of binary mixtures, are discussed in chapter 3. In chapter 4, the mechanism of protype sensor signal measurements, training set description, refractive index measurements as well as transmittance measure- ments are presented (Paper I, II, III and IV). In chapter 5 we briefly describe the data analysis methods such as extinction coefficient, real refractive index, excess permit- tivity, imaginary excess permittivity (Paper II and III), and the method of volume increase (Paper IV). Chapter 6 delves into the results and discussions concerning the major findings of this thesis. There we deal with the results of the prototype of an optical sensor (Paper I), results of extinction coefficient calculations, Wavelength de- pendent refractive index, and excess optical properties (Paper II and III). Moreover, the results of handheld refractometer method are also presented (Paper IV).

(21)

5

2 LIQUID FUEL ADULTERATION

Fuels are materials such as coal, gas or oil, which when burnt in the presence of air or oxygen produces energy in form of heat or power. The material must contain one or more combustible elements such as carbon, hydrogen, sulphur, and other components for it to be classified as fuel. In its natural state fuel does not contain any usable energy, for it to be useful it must undergo a transformation process from one form to another. Like other matter fuels exist in three states, which are classified either as solid fuel such as wood or coal, liquid fuel such as diesel oil, and gasoline or kerosene and gaseous fuel such as liquiﬁed petroleum gas (LPG). We focus our attention on liquid fuels which are also sub classified as natural, which means crude oils before any processing or artificial, meaning manufactured or processed fuels.

The common fuels are the final products of crude oil refinery process, these includes gaseous fuels, liquid fuels, lubricants, solvents, waxes and asphalts. Crude oil is a complex mixture of many hydrocarbon groups such as butane, pentane, propane, methane and ethane. Moreover, crude oils contain other hydrocarbon groups namely, as paraffinic, naphthenic and aromatic, which in most cases are the main building blocks of organic industry (Chaudhuri U. R, 2016). It is necessary to separate different components from the original crude oil to obtain a usable fuel. To achieve this, crude oil must go through a refinery process which involve several stages that must be undertaken in succession before the final fuel product is obtained these are, distillation, thermal cracking, catalytic process, treatment, formulation and blending (US Patent No. 2,914,457, 1959). The results of a refinery process are liquified petroleum gas (LPG), gasoline, naphtha, kerosene, aviation turbine fuel (ATF), diesel oil, lubrication oils, in the respective order, and many other products (Chaudhuri U. R, 2016; Speight J. G, 2006; Chaudhuri U. R., 2011). The variation of refineries influences the appearance and concentration of the end products, therefore chemical compounds of ﬁnal fuel products vary depending on geographic origin of crude oil. In this work we focus our attention on diesel oil and kerosene whose mixtures represent a difficult case of fuel adulteration.

2.1 DIESEL OIL

One of the many products of crude oil refinery process is diesel, which is also called petrol diesel. This is necessary to different it from bio diesels, un like bio diesel, petro diesel is a by-product of crude oil, which is obtained through fractional distillation process with burning temperature ranges of 250˚C – 350˚C. In the refinery process temperature is the chief method through which different components are extracted and separated from the rest. The chemical composition of diesel is 75% hydrocarbons which are mainly paraﬃns, and 25% aromatic hydrocarbons such as naphthalenes and alkylbenzenes (Cooke R. A. & Ide R. H, 1985). For common diesel oils, the aver- age chemical formula ranges from C¹⁰H²² to C¹⁵H³², and their densities has a range of

(22)

6

0.820−0.850 kg/l, this density variation imply a consequent refractive index variation across the span of different diesel samples, especially those coming from different oil fields as well as the differences in the refinery process (Aleme H. G., 2010). Diesel fuel is more or less superior to gasoline because it is more eﬃcient, much safer in operation, and oﬀers a wide range of performance as a transport fuel for many different types of engines. Literature claims that diesel oil contains 32-40 MJ/L of energy, about 18% to 30% more energy per gallon as compared to gasoline (Murago E. N. M, 2013). Because of this diesel fuels are widely used in many types of transportation vehicles, with the exception of the gasoline-powered passenger automobile (Aleme H. G., 2010). In many developing countries the proportion of automobiles utilizing diesel is larger compared to those utilizing gasoline, making it a lucrative target for adulteration practices.

2.2 KEROSENE

There are two ways by which kerosene is obtained, the first one is from distillation of crude oil under atmospheric pressure, and the second one is from thermal or steam cracking of heavier petroleum streams. After this initial process further treatment is necessary where by kerosene is treated by a couple of different processes to reduce the level of materials such as sulphur, nitrogen and olefin (Klimisch H. -J., 1997). Pre- dominantly kerosene contains C⁹ to C¹⁶ hydrocarbons with a boiling range of 145˚C – 300 ˚C, moreover, the density of kerosene varies from 0.760 − 0.800 kg/l. This implies a consequent variation in refractive index for different kerosene samples. The intrin- sic composition of kerosene is 70% paraffins and naphthenes, 25% aromatic hydrocarbons and less than 5% olefins. Kerosene is mainly utilized for powering jet engines of aircraft (jet fuel), and is also utilized in some rocket engines, never the less it is utilized domestically for cooking and lighting fuel (US Patent No. 2,914,457, 1959).

The cost for kerosene is usually lower compared to diesel, this makes it lucrative product for malpractice such as adulteration.

2.3 ADULTERATION

Adulteration is actually illegal practice which involves addition of unauthorized substances into pure products, this practice leads to a final product which does not com- ply to specifications or standards, and products that might be harmful to humans or to the environment (US Patent No. 2,914,457, 1959; Obeidat S. M., 2014). For the case of diesel adulteration foreign substances such as kerosene, naphtha, as well as other chemicals resulting or originating from petroleum refining processes are utilized as adulterants for diesel oils. Furthermore, in certain cases the gasoline boiling range hydrocarbons are added into automotive diesel, also west industrial solvents such as lubricants or heavier fuel oils in small portions are added into diesel fuels (Murago E. N. M, 2013; Cunha D. H., 2016). Likewise, other adulterants for diesel oils are biodiesel, vegetable oil, residual oil and sulphur (Majhi A., 2012; Flumignan D. L., 2010).

(23)

7 Among these, kerosene is highly preferred because, kerosene and diesel both have overlapping hydrocarbons namely, C¹⁰ to C¹⁵for diesel and C⁹ to C¹⁶for kerosene (Tharby R, 2002; Mendes G. & Barbeira P. J. S, 2013; Pedrosso M. P., 2008; Dey R. &

Dwidevi A, 2014). In areas where adulteration practice is prevalent 10% to 15% adulteration is usually proﬁtable while adulteration level below 10% becomes less ﬁnan- cially attractive. Moreover, the adulteration levels above 30% can easily be detected due to its direct impact on the emitted gases, and because of malfunction of engines.

Despite this fact, adulteration even below 10% is still practiced. In recent years, major techniques for screening adulterated fuel have been unable to detect adulteration levels below 20% (Mishra V., 2008). It is therefore necessary to consider adulteration at 5%, 10% and 15%, and this we have done in this thesis.

Fuel adulteration practice in different parts of the world is based on very similar reasons such as greed caused by different tax systems for different fuels where, kerosene is always subsidized by the government in the attempt to lower its price for poor population. Gasoline is taxed highest followed by diesel, while kerosene is taxed least (Gupta A. & Sharma R, 2010). On the other hand, profit making in oil business is also one of the major reasons for adulteration practice (Sinha S. N., 2005).

Despite these reasons lack of monitoring and consumer awareness is also the major reason for adulteration, because similarities in properties of these petroleum products make it diﬃcult for lay individuals to notice the diﬀerence when adulterated.

Moreover, the lack of transparency and noncontrolled regulations especially in pro- duction, supply and market chains and lack of simple equipment for detection and identiﬁcation of fake fuels (Murago E. N. M, 2013).

Fuel adulteration has negative impacts which are widely visible in our daily lives, the environment surrounding us, and economy. These include pollution crisis, res- piratory infections and more poisonous exhaust gases (Taksande A. & Hariharan C, 2006). Moreover, there are several ambient air pollutants such as SO², NO², particulate matter (PM), polycyclic aromatic hydrocarbons (PAHS), as well as volatile organic compounds (VOC), which are emitted from automobile exhaust and industrial activity. These pollutants lead to morbidity and mortality especially in developing and third world countries (Sinha S. N. & Shivgotra V, 2012; Sinha S.N., 2010; Roy S, 1999). In some cases when high boiling compounds are utilized as adulterants in fuel, it causes increase in knock, engine wear and in certain circumstances engine starting problems. On the other hand, when low boiling point compounds are applied as adulterants it leads to vapor lock. The other serious effect for governments is tax eva- sion (Payri R., 2013). Furthermore, adulteration leads to pollution, poor engine performance, failure of machine components, lower returns for buyer’s money, and de- crease of the availability of kerosene to the needy, especially poor communities (Mishra V., 2008). On the other hand, fuel adulteration may lead to environmental hazards, increase in tail pipe emission, ill eﬀects on public health, and engine malfunction (Obeidat S. M., 2014). Finally, nonconformity resulting from adulteration of fuel can cause damages that are difficult to repair in cars, these includes engine’s

(24)

8

sudden drops, possibility of traﬃc accidents, over fuel consumption and more emission of exhaust gases as well as particulate matter (Cunha D. H., 2016).

To control the problem of fuel adulteration, monitoring of fuel quality is inevita- ble especially at different distribution chain stages. In developing countries this is even more difficult because there are multiple importers and distributors of fuel, therefore complicating the process of monitoring the quality of fuel. The task of fuel quality monitoring can be accomplished either by taking the sample to the laboratory for measurements, or by performing field screening using portable equipment. Many existing methods for fake fuel screening are meant for laboratory purposes, however, few methods do exist that can be used for field measurements. The challenge with field measurement is the lack of more portable and cheaper equipment, especially for end consumers like car drivers who have little or completely lack knowledge about neither physics nor analytical chemistry.

Currently several methods have been deployed by regulating agencies for adulteration detection, among them are, test based on evaporation (ASTM D380), based on distillation (ASTM D86), based on gas chromatography (GC), by using fiber optics sensor, and using ultrasound. This is based on American Society of Testing and Ma- terials international (ASTM) (Gupta A. & Sharma R, 2010; Vogt T. K., 2004; Teixeira L. S., 2008). Other methods for laboratory and field screening include near and mid infrared spectroscopy NIR and MIR, Fourier transform infrared spectroscopy FTIR (Roy S, 1999; Vogt T. K., 2004; Gupta A., 1992), filter based method, test based on specific gravity, viscosity, odour test, ultrasonic, and titration techniques (Sadat A., 2014) (Bhatnagar V, 1981; Shahrubahari M., 1990; Perreira R. C., 2006). Furthermore, methods such as synchronous fluorescence scan (SFS) (Sinha S.N., 2010), by measuring refractive index (Ariponnammal S., 2012), by FTIR spectra coupled with principle component analysis (PCA) as well as linear discriminant analysis (LDA)(Perreira R.

C., 2006). Also, by gas chromatography and gas chromatography spectroscopy (Moreira L., 2003), by using photothermal detection method (Lima J., 2004), by filter paper method (Majhi A., 2012), by using excitation emission matrices fluorescence spectroscopy (EEMF), and by multiway principle component analysis (Obeidat S. M., 2014; Patra D. & Mishra A, 2001), moreover, it is done by using optical sensors (Sri- vastava A., 1997), and finally by fiber optics sensors (Eriksson M. & Iqbal Z, 2014).

There are several measurement methods that deal with portable liquid measurements such as, mobile phone based optical sensing (Felix V. J., 2015), by using portable electronic techniques (Wiziack N. K., 2011), by image processing method (Ram- mohan V. M, 2010), and by using a chemical sensor array (Cozzolino D., 2006).

With respect to adulteration detection policy, several standards such as ASTM D4052, ASTM D3810, ASTM D86, and ASTM D842, exist to facilitate the fuel adulteration detection both in US and Europe (Gupta A. & Sharma R, 2010). Some of the specific fuel compositions that are tested by the standards are cetane number (ISO 5165, EN15195, EN16444 and ASTM D613). In addition to this cetane index is also tested (ISO 4264/ASTM D976). Consequently, the density tests are (ISO 3675/ISO 12185), aromaticity test (EN 12916, ASTM D 1319), testing the content of Sulphur (ISO

(25)

9 20846), testing flash point (ISO 2719/ASTM D93), testing kinematic viscosity (ASTM D445/ISO 3104) (Bhanu P., 2017). However, these standards cannot be adopted in the entire world for checking adulteration, the reason being that, petroleum products are mixtures constituting of several polycyclic aromatic compounds (PACS). The composition of these compounds is affected by the origin of oil field, which vary from one location and geographic origin to the other. Because of this, it is necessary to enact standards for different countries, however, this has not been achieved.

(26)

10

3 THEORY

This thesis studies the optical properties of diesel oils and kerosene, and thereafter use these properties in data analysis, to identify and possibly quantify the level of adulteration. This is possible if we consider the interactions between electromagnetic radiations and the liquid samples by optical spectroscopy. Optical spectroscopy measures the spectral distribution of electromagnetic radiation(light) analytically.

The electromagnetic radiation is very broad, it spans all the way from gamma rays (10^-4- 10^-1nm) to microwaves (1 mm-100 cm). The spectral region between 380 nm and 800 nm represents the visible part of electromagnetic spectrum which is sensitive to our vision system, while the region spanning from 700 nm to 3000 nm is the near- infrared region (Nicolai B. M., 2007). Analysis of NIR spectra is characterized by low molar absorption as well as less scattering, which enables evaluation of pure materials with less effort. It is interesting that although it was discovered by Herschel as early as 1800, it was initially ignored by the scientists, who thought it lacked analytical capabilities (Roggo Y., 2007; McClure W. F., 2003; Wiedemann L. S. M., 2005).

Despite that, in recent years NIR spectral region has become dominant especially in molecular spectroscopy, thanks to powerful and inexpensive computers. The computers can give quantitative information on the major organic components and functional groups of fuels such as hydrocarbons (Takeshita E.V., 2008). Recently the significance of NIR region especially for fuel studies has become popular for determination of the quality of liquid fuels. This it achieves through determination of the octane number, ethanol contents, methyl tetra-butyl ether (MTBE) content, distilla- tion points, aromatic as well as saturated contents, and Reid vapor pressure. Reid vapor pressure is the absolute vapour pressure exerted by a liquid at 100˚F, it is a common measure of volatility of gasoline as determined by the test method ASTM- D-323 (Teixeira L. S., 2008; Rammohan V. M, 2010; Barbeira P. J. S, 2002; Wooten F, 2013; Wiener O, 1912).

When photons of varying wavelengths strike the matter, the transmission properties of the medium are influenced by reflection, scattering, refraction as well as absorption. Part of the light incident on the material may be deflected back to the direc- tion from which it come, that is called reflection. Part of light might be lost within the material due to interaction with molecules in the material, which leads to specific electron transition in the material, or the lost energy may be converted to heat, this is known as absorption. Part of incident light might end up being deflected randomly or even spread randomly in all directions this is the so-called scattering, while the light that emerges to the other side is the transmitted light (Nicolai B. M., 2007). For the case of diesel oils and kerosene major part of light is transmitted to the other side because the fuels are highly transparent. Moreover, the phenomenon of refraction is exhibited at the interface between liquid and cuvette which further leads to total internal reﬂec-

(27)

11 tion and scattering. This takes place due to non-absorbing characteristics of these sam- ples. Finally, the other challenge to worry about is the cuvette reflections, which we have dealt with rigorously in chapter 4.

3.1 LIGHT INTERACTION WITH THE ROUGHENED GLASS-FUEL INTERFACE

Our first publication (Paper I) was solely devoted to report on practical experimental aspects rather than theory. Therefore, in this section a qualitative theoretical picture about a stable rough glass-liquid system is given, followed by liquid drop spreading which is unstable system. Firstly, several phenomena that are present during the process of light interaction with the roughened glass-fuel interface are considered. These includes the effects originating from the nature of incident light, and the effects resulting from the surfaces in contact, namely fuel and roughened glass. Secondly, phenomena that affect the time development, which is one indicator of the drop spreading and which affect the scattering of light thus influencing the measured signal are described.

3.1.1 Description adopted from the pigment model

The interaction between electromagnetic radiation with roughened glass-fuel interface is a rather complex phenomenon. For a better understanding of the complex processes taking place, the theoretical setup utilized by (Niskanen I., 2010; Beckmann P., 1963) is adopted, with slight modification of replacing a pigment with roughened glass window. In the setup the pigment is transparent, same case holds for a roughened silica glass, moreover, the pigment surface is considered rough, same case holds for the roughened glass window.

If a light ray strikes the interface between roughened glass and liquid, it will always be scattered from the glass provided that, there is a mismatch of refractive index between the fuel and rough glass. The scattering process is a complex phenomenon since it incorporates both reflection, refraction, and diffraction phenomena which are inseparable in the process (Niskanen I., 2012; Nussbaumer R. J., 2005). In the theory it was assumed that, the plane wave is incident on the suspension. How- ever, in this thesis the case of handheld device is qualitive because the laser beam is focused and not a plane wave. Let us consider a hypothetically oversimplified model in Fig. 3.1.

(28)

12

Figure 3.1. Optical paths of transmitted and refracted rays (Niskanen I., 2012).

From (Fig. 3.1), the optical pathlength difference between the two rays is given by

(n^glass – n^liquid) z, where n^glass and n^liquid are the refractive indices of glass and liquid respec-

tively, and z is the separation between the two surfaces. This difference causes the phase shift in the electric field between the two light rays, provided that the glass size is larger in comparison to the wavelength of incident light. The resulting phase shift is given by:

∆𝜑 =2𝜋

𝜆 (𝑛_{𝑔𝑙𝑎𝑠𝑠}− 𝑛_{𝑙𝑖𝑞𝑢𝑖𝑑})𝑧, ( 3.1)

where λ is the wavelength of light. Next, upon further mathematical manipulations (Niskanen I., 2012), one arrives at the expression for intensity of transmitted light which is detected at the far field region. This is given by the power density function as;

𝐼 = |𝐸02∫ 𝑤(𝑧) 𝑒𝑥𝑝 {𝑖 [2𝜋𝑐𝑜𝑠𝜃

𝜆 (𝑛𝑔𝑙𝑎𝑠𝑠− 𝑛𝑙𝑖𝑞𝑢𝑖𝑑)] 𝑧}

∞

−∞

𝑑𝑧|

2

, (3.2)

where E⁰is the electric field amplitude of incident light. Next, if Eq. (3.2) is taken further by determining its Fourier transform, the final equation is a function with a Gaussian shape. Moreover, if the distribution of the transmitted light follows the Gaussian distribution, the final resulting equation is given as:

𝑇 = 𝐼 𝐼⁄ = exp {− [_𝑂 2𝜋𝑐𝑜𝑠𝜃

𝜆 (𝑛_{𝑔𝑙𝑎𝑠𝑠}− 𝑛_{𝑙𝑖𝑞𝑢𝑖𝑑})]

2

}, (3.3)

where I^O and I are the incident and transmitted light intensities respectively.

(29)

13 According to (Niskanen I., 2010), the treatment above strongly resembles light scattering from a rough surface that is subjected to the statistics of a Gaussian surface, and therefore, from this observation it is logical if we conclude that, the light scattering obeys the gaussian distribution (Beckmann P., 1963). Moreover, since absorption is very low in diesel oils, it is possible to obtain the reflectance by R=1-T, which is the first approximation. The real phenomenon taking place is rather complex because of back scattered light which is reflected from the smooth side of the rough glass, leading eventually to multiple scattering and transmission, thus the model is only quali- tative. The simple treatment of R=1-T implies that, reflectance also obeys the Gauss- ian distribution properties. The average surface roughness of the roughened glass utilized in this work was Ra = 0.48 μm which was measured by a stylus profilometer, the surface was finished by diamond grinding pads with criss-cross surface finishing style which obeys gaussian distribution according to (Tanner L. T, 1976; Whitley J.

Q., 1987; Richard B. Z, 1983; Wolfgang S. & Nico C, 2011). This is very important, because it leads us to expect normally distributed signals resulting from back scattering. On the other hand, as the difference or variation of refractive index becomes high between the liquid and glass, the scattering signal also becomes high and vice versa (Paper I).

Facet model along Gaussian surface height distribution was investigated for transmission of light through a roughened glass slide immersed in liquid (Nuss- baumer R. J., 2005).

3.2 THE WETTING PROPERTY OF LIQUIDS (CONTACT ANGLE)

3.2.1 Ideal wetting process

In the liquid-glass interface, another very important factor which affects the scattering of light is the wetting which dictates the contact-angle between the liquid and solid. There are two types of wetting processes namely hydrophilic, when the liquid and the glass come into contact spontaneously resulting into a film, and hydrophobic where there is no any contact between liquid and glass. The basic laws of wetting were first developed by Laplace and Young, for solids which are ideal, namely smooth solid surfaces (Quere D, 2008).

According to Laplace and Young, the material surface carries a certain amount of energy known as surface tension which is denoted by γIJ representing energy per unit area for an interface with varying phases I and J. If γ is the surface energy (surface tension) at the interface between liquid and air, the two leads to a relation which was first imagined by Marangoni (Quere D, 2008), the relation considers the spreading of a film as demonstrated in Fig. 3.2(a). The relation of Marangoni is known as spreading parameter (S) which is given by S= γ^SA – γ^SL – γ, where γ^SAand γ^SL are the surface tensions between the solid-air, and solid-liquid interfaces respectively.

(30)

14

Figure 3.2. Two classical wetting situations for an ideal material (Quere D, 2008).

It is the sign of parameter S that dictates the behavior of the liquid drop on the surface of the solid, if S> 0 spreading takes place, while for the opposite case a lens like shape is formed which does not spread. This opposite case is demonstrated in Fig. 3.2(b), where unlike for the case of (Fig. 3.2(a)) here a contact angle β exist be- tween the solid and liquid drop. Moreover, the equilibrium condition is achieved by the liquid drop on the contact line due to the actions of varying surface tensions. The balance equation for such equilibrium condition is given as:

𝛾_𝑆𝐴= 𝛾_𝑆𝐿+ 𝛾_{𝐶𝑂𝑆𝛽}, (3.4)

this theoretical model is only valid for smooth surfaces (Quere D, 2008).

3.2.2 Contact angle hysteresis

Most glasses are said to be naturally rough, even those glasses which looks smooth to our eyes are usually rough at micrometric dimensions. There are several reasons for the roughness, these includes lamination process during fabrication which gen- erates micro grooves, grain compaction a process which results into roughness of grains, as well as coating (Quere D, 2008). According to (Quere D, 2008), the contact line on the solid surface can be pinned due to surface defects, consequently causing drops on an inclined plane to either remain stationery or causing wetting and non- wetting defects on both sides of the surface (Furmidge C. G. L, 1962). This leads to asymmetry in contact angle which causes pressure to build up. The built-up pressure further leads to a force which is capable of resisting gravity especially when the drop is small. Therefore, both the diversity in chemical composition of the material and normal surface roughness affects the contact angle.

Generally, the contact angle depends on the previous status of liquid deposition on the surface. Provided that the drop was deposited gently, it will continue to spread and will only stop due to new wetting defects. After some time, evaporation will take place and will change the configuration to resemble that of a pinned drop as illustrated in Fig. 3.3.

(31)

15 Figure 3.3. Illustration of the apparent pinning of a contact line on an edge (Quere D., 2008).

According to basic laws of wetting developed by Laplace and Young, when the liquid meets the solid surface, the contact angle β is formed. It is possible for β to assume various values between β and ℼ-ɸ + β at the edge of the solid, where ɸ is the interior angle between the two surfaces forming an edge. If the horizontal is considered as the reference, the contact angle hysteresis can be quantified. Contact angle hysteresis phenomenon should be considered for positive reasons such as, to assist in guiding the flow of liquid along a specified line of defects. This will allow the liquid to follow a specific path. On the contrary this phenomenon can be detrimental, for instance when water droplets remain stationery on glass surfaces leading to their deterioration (Quere D, 2008; Furmidge C. G. L, 1962).

3.2.3 Behavior of liquid over a rough surface

Unlike smooth surfaces, the contact angle is a local quantity for rough surfaces. Pro- vided the surface gradient is very minimal in comparison to the angle of contact β, the spreading of liquid over the glass surface is not influenced by roughness. To the contrary, when the gradient is significant which is usually the case for spreading drops, the spreading rate is altered (Raltson J., 2008). The rate of drop spreading on rough surfaces changes with time. The changes are described by two regimes and defined by shape of drop as it spreads over the surface, namely, the vary fast region where the drop assumes the shape of a Mexican-hat (MH), where the drop has a cap and a foot, and the vanishing of the cap as it is consumed by the foot. Mathematically, in the MH region, the rate of change of the spreading rate is given by:

𝛥𝑅 (𝑡) = √𝐷𝑡, (3.5)

where D = C/ η is the coefficient of diffusion, 𝜂 is the viscosity, and C is a property of the surface proportional to roughness height, and t, is the time. The growth of the foot then reduces the spreading rate to

𝑅 (𝑡) ≈ √𝑡⁴ , (3.6)

(32)

16

3.3 BEER-LAMBERT’S LAW

In optical transmission spectroscopy the most important law governing the interaction of light and liquid is the Beer-Lambert’s law, especially for homogeneous and isotropic liquids such as diesel oils and kerosene. If the beam of light with intensity I^O strikes the sample and get transmitted through the sample, emerging to the other side with intensity I, the extent of absorption of light intensity in the sample can be treated with the equation:

𝑑𝐼

𝑑𝑥= −𝛼𝐼(𝑥), (3.7)

by integrating Eq. (3.7), we obtain the wavelength-dependent Beer-Lambert’s law given as:

𝐼(𝜆) = 𝐼_𝑂(𝜆)𝑒^{−𝛼(𝜆)𝑑}, (3.8)

where α, λ and d represents the absorption coefficient, wavelength of incident light, and sample thickness respectively. If the absorption coefficient is represented in terms of wavelength, the resulting equation becomes:

𝛼(𝜆) =1 𝑑𝑙𝑛 1

𝑇(𝜆) , (3.9)

where T = I/I^O is the transmittance. The schematic diagram depicting light matter interaction process is illustrated in Fig. 3.4.

Figure 3.4. Schematic illustration of Beer- Lambert’s Law.

(33)

17

3.4 COMPLEX REFRACTIVE INDEX

In cases when the medium exhibits both absorption and refraction, it is possible to represent both quantities with the so called complex refractive index (N) (Wooten F., 2013). This quantity is represented in terms of the real and the imaginary refractive indices n and k respectively. The wavelength dependent complex refractive index is given by Eq. (3.10) as:

𝑁(𝜆) = 𝑛(𝜆) − 𝑖𝑘(𝜆), (3.10)

where i is the imaginary unit, n can be measured by several methods. However, in this work the measurement was done by an abbe refractometer. The relationship be- tween the extinction coefficient k and the absorption coefficient α is given by:

𝑘(𝜆) =𝛼(𝜆)𝜆

4𝜋 . (3.11)

3.5 SINGLY SUBTRACTIVE KRAMERS-KRONIG RELATION (SSKK)

In optical spectroscopy, the refractive index of a particular sample at a specified wavelength cannot be obtained directly based on transmittance measurement. How- ever, this quantity is directly related to the density of liquid samples. Moreover, by Abbe refractometers it is usually possible to find the value of refractive index at a single specified wavelength. Therefore, to find this value at different wavelengths, it is imperative to utilize the absorption and dispersion phenomena which are usually related. In linear optical spectroscopy light absorption and dispersion is guided by the causality principle. Therefore, it is usually possible to estimate one quantity provided that the other is known, namely absorption and dispersion. These quantities are connected by a pair of equations called Kramers-Kronig (KK) relations (Peiponen K. -E. & Saarinen J. J, 2009; Ahrenkiel R, 1971; Peiponen K. -E., 1998).

If ﬁnite range integration is desired, the conventional K-K relation is modiﬁed to include the anchor point obtained by the Abbe refractometer, result of such modification is the SSKK. We use this relation to extrapolate data especially in refractive index calculations. The significance of SSKK comes from the fact that, devices utilized for refractive index measurements, such as Abbe refractometer, give the refractive index reading only under one wavelength. Therefore, we can extrapolate the values of refractive indices over a wide wavelength range by using SSKK (Peiponen K. -E., 1998). In this thesis the wavelength of Vis-NIR radiation is exploited rather than an- gular frequency. Because we can measure a discrete refractive index value for an anchor point with the Abbe refractometer, we further utilize SSKK dispersion relation to find out n as:

(34)

18

𝑛(𝜆^′) − 𝑛(𝜆^′′) =2(𝜆^′2− 𝜆^′′2)

𝜋 𝑃 ∫ 𝜆𝑘(𝜆)𝑑𝜆

(𝜆²− 𝜆^′2)(𝜆²− 𝜆^′′2)

∞

0

, (3.12)

where n(λ’) is the wavelength-dependent refractive index, n(λ’’) is a priori known refractive index value at an anchor point λ’’ and P denotes a constant known as the Cauchy principal value. For practical purposes, the integral especially in the case of SSKK analysis, can be usually truncated to correspond to the initial and final point of the measured transmittance. Here we use the method of optimal choice of the anchor point λ’’ to minimize possible data inversion error. This method is based on the properties of a Chebyshev polynomial as shown by Palmer et al (Palmer K. F., 1998).

We are not free to choose the anchor point because of the fixed wavelength of the Abbe refractometer. Nevertheless, we obeyed a novel approach by the choice of the optimal anchor point by fixing the wavelength of the final point (λ^f), of the transmittance spectrum and we calculated the initial wavelength (λⁱ) that corresponds to the choice of the optimal anchor point. The wavelength λⁱ is then obtained from a zero of a Chebyshev polynomial of the first kind as follows (Palmer K. F., 1998):

𝜆_𝑖= 𝜆𝑓𝜆^′′

√2𝜆_𝑓²− 𝜆^′′2

. (3.13)

Therefore, we initiated scanning of transmittance spectrum from this calculated wavelength. The SSKK method has been used, e.g., to analyze properties of diesel oil droplets (Dombrovsky L. A., 2003). In the case of SSKK relation the choice of the location if a single anchor point usually is not a critical factor for successful inversion of the extinction data (Paper II and III).

3.6 COMPLEX EXCESS PERMITTIVITY

In this study we deal with binary liquid fuel mixtures. We denote relative permittivity of diesel oil by ε^D and corresponding permittivity of kerosene by ε^K. According to the definition, and allowing losses (Iglesias T. P. & Reis J. C. R, 2016), the excess relative permittivity in our case is given by the expression:

ε^E= ε − ε^{𝑖𝑑𝑒𝑎𝑙}= ε − 𝜀_𝐾[1 + f_D((𝜀𝐷

𝜀𝐾

) − 1)], (3.14)

where ε is the measured permittivity, ε^ideal is the permittivity of binary mixture with- out any interaction processes between the mixture components, f^D is the volume fill fraction of diesel oil present in the mixture, and the permittivity ratio in the paren- thesis is known as permittivity contrast (Papers II, III and IV).

(35)

19 The ideal permittivity concept assumes that there is no chemical activity when- ever two different liquids are mixed, some liquids fulfil this requirement. However, due to the presence of chemical activity in some liquids, there will be intermolecular interactions which will contribute to the molecular polarization of electric charges, causing changes to optical properties. Eventually, the ideal permittivity is the same as the upper Wiener bound (Wiener O, 1912) of an effective medium which is given as:

𝜀^{𝑖𝑑𝑒𝑎𝑙}= 𝑓𝐷𝜀𝐷+ (1 − 𝑓𝐷)𝜀_𝐾. (3.15)

In general case if the permittivity is a complex number, then the excess permittivity is also complex. The frequency dependent complex permittivity in a homogeneous medium as a function of wavelength is given by the expression:

ε(𝜆) = 𝜀1(𝜆) + 𝑖𝜀₂(𝜆) = 𝑛²(𝜆) − 𝑘²(𝜆) + 𝑖2𝑘(𝜆)𝑛(𝜆).

(3.16)

The complex relative permittivity of a medium is given by the relation ε = N ², this was utilized in (Eq. (3.16)) and in (Papers II, III and IV).

Next, by utilizing Eqs. (3.14) and (3.16), the equation for imaginary excess permittivity is given by:

𝐼𝑚𝜀^𝐸(𝜆) = 2𝑛(𝜆)𝑘(𝜆) − 2𝑛_{𝑖𝑑𝑒𝑎𝑙}(𝜆)𝑘_{𝑖𝑑𝑒𝑎𝑙}(𝜆) =

2𝑛(𝜆)𝑘(𝜆) − 2(𝑓_𝐷𝑛_𝐷(𝜆)𝑘_𝐷(𝜆) + (1 − 𝑓_𝐷)𝑛_𝑘(𝜆)𝑘_𝑘(𝜆)), (3.17)

where n(λ) and k(λ) are values of adulterated samples, while in our case n^ideal(λ) and k^ideal(λ) are the values of ideal binary mixtures, n^D(λ) and k^D(λ) are the values of au- thentic diesel oil, and nK(λ) and kK(λ) are the values of kerosene respectively. All these were obtained using the measured refractive indices and the transmitted data inver- sion using SSKK (Paper III).

3.7 LORENTZ-LORENZ FORMULA

The Lorentz oscillator model is the highly preferred formula for calculating the refractive index of binary mixtures, such as adulterated diesel oils. This is given in (Bar- anovic G, 2017), and is expressed as:

𝑛²−1 𝑛²+2= 𝑓𝐷

𝑛_𝐷²−1

𝑛_𝐷²+2+^(1−𝑓^𝐷^)(𝑛^𝐾²⁻¹⁾

𝑛_𝐾²+2 , (3.18)

where n is the refractive index of the resulting mixture, nD is the refractive index of diesel oils,and n^K is for kerosene. We can define an ideal mixture using this model

(36)

20

because, it gives good estimate of volume fill fraction for ideal mixtures when there are no interactions between molecules of the participant liquids. However, when the interactions exist, this formula gives erroneous volume estimates (Paper IV).

3.8 MODIFIED IDEAL LAW OF BINARY MITURES

The ideal mixture equation (Eq. 3.16) can be re-written so that the total volume expression includes the volumes of the individual binary mixture constituents, namely diesel oil as well as kerosene. This leads to a modified equation given as:

𝜀^{𝑖𝑑𝑒𝑎𝑙}=𝑉_𝐷

𝑉 𝜀𝐷+𝑉_𝐾

𝑉 𝜀𝐾, (3.19)

where V = VD + VK is the total volume of the mixture.

Next, Eq. (3.16) is further modified to incorporate the novel concept of increase of volume of pure diesel oil in the suspected sample. This was achieved by introducing another variable V’, this leads to the modified formula given by:

𝜀^{𝑖𝑑𝑒𝑎𝑙}(𝑉^′) =^𝑉^𝐷^+𝑉′

𝑉+𝑉′ 𝜀𝐷+ ^𝑉^𝐾

𝑉+𝑉′𝜀𝐾. (3.20)

The added volume V’ should be with respect to the magnitude of the initial vol- ume V of the suspected sample. If the volume V’ is continuously increased, the sam- ple approaches the case of ideal mixture as the concentration of kerosene continues to diminish. The limiting value for this mixing procedure is given by Eq. (3.21) (Paper IV).

𝑉′→∞lim 𝜀^{𝑖𝑑𝑒𝑎𝑙}(𝑉^′) = 𝜀_𝐷. (3.21)

Development of optical measurement techniques and data analysis methods for screening of adulterated diesel oils

uef.fi

Dissertations in Forestry and Natural Sciences

BONIPHACE ELPHACE KANYATHARE

DEVELOPMENT OF OPTICAL MEASUREMENT TECHNIQUES AND DATA ANALYSIS METHODS FOR SCREENING OF ADULTERATED DIESEL OILS

BONIPHACE ELPHACE KANYATHARE

DEVELOPMENT OF OPTICAL MEASUREMENT TECHNIQUES AND

DATA ANALYSIS METHODS FOR SCREENING OF ADULTERATED

DIESEL OILS

BONIPHACE ELPHACE KANYATHARE

DEVELOPMENT OF OPTICAL MEASUREMENT TECHNIQUES AND

DATA ANALYSIS METHODS FOR SCREENING OF ADULTERATED

DIESEL OILS

ABSTRACT

ACKNOWLEDGEMENTS

LIST OF ABBREVIATIONS

LIST OF ORIGINAL PUBLICATIONS

AUTRHOR’S CONTRIBUTION

TABLE OF CONTENTS

1 INTRODUCTION

2 LIQUID FUEL ADULTERATION

2.1 DIESEL OIL

2.2 KEROSENE

2.3 ADULTERATION

3 THEORY

3.1 LIGHT INTERACTION WITH THE ROUGHENED GLASS-FUEL INTERFACE

3.2 THE WETTING PROPERTY OF LIQUIDS (CONTACT ANGLE)

3.3 BEER-LAMBERT’S LAW

3.4 COMPLEX REFRACTIVE INDEX

3.5 SINGLY SUBTRACTIVE KRAMERS-KRONIG RELATION (SSKK)

3.6 COMPLEX EXCESS PERMITTIVITY

3.7 LORENTZ-LORENZ FORMULA

3.8 MODIFIED IDEAL LAW OF BINARY MITURES