Photonic solution for adulteration detection in edible palm oil

(1)

Dissertations in Forestry and Natural Sciences

DISSERTATIONS | SAMPSON SAJ ANDOH | PHOTONIC SOLUTIONS FOR ADULTERATION DETECTION IN EDIBLE PALM OIL | No 4

SAMPSON SAJ ANDOH

PHOTONIC SOLUTIONS FOR ADULTERATION DETECTION IN EDIBLE

PALM OIL

PUBLICATIONS OF

THE UNIVERSITY OF EASTERN FINLAND

(2)

(3)

PUBLICATIONS OF THE UNIVERSITY OF EASTERN FINLAND DISSERTATIONS IN FORESTRY AND NATURAL SCIENCES

N:o 431

Sampson Saj Andoh

PHOTONIC SOLUTIONS

FOR ADULTERATION DETECTION IN EDIBLE PALM OIL

ACADEMIC DISSERTATION

To be presented by the permission of the Faculty of Science and Forestry for public examination in the Auditorium M100 in Metria Building at the University of Eastern Finland, Joensuu, on October, 15, 2021, at 12 o’clock.

University of Eastern Finland Department of Physics and Mathematics

Joensuu 2021

(4)

PunaMusta Oy Joensuu, 2021

Editors: Pertti Pasanen, Nina Hakulinen, Raine Kortet, Jukka Tuomela, and Matti Tedre

Distribution:

University of Eastern Finland Library / Sales of publications julkaisumyynti@uef.fi

http://www.uef.fi/kirjasto

ISBN: 978-952-61-4299-9 (Print) ISSNL: 1798-5668

ISSN: 1798-5668 ISBN: 978-952-61-4300-2 (PDF)

ISSNL: 1798-5668 ISSN: 1798-5676

(5)

Author’s address: University of Eastern Finland Institute of Photonics

P. O. Box 111 FI-80101 JOENSUU FINLAND

email: sampson.andoh@uef.fi Supervisors: Professor Matthieu Roussey

University of Eastern Finland Institute of Photonics

email: matthieu.roussey@uef.fi Professor Jyrki Saarinen University of Eastern Finland Institute of Photonics

email: jyrki.saarinen@uef.fi

Reviewers: Professor Ilpo Niskanen

University of Oulu Faculty of Technology P. O. Box 7300

FI-90014 OULU FINLAND

email: ilpo.niskanen@oulu.fi Assoc. Professor Erik M. Vartiainen Lappeenranta University of Technology LUT School of Engineering Sciences P. O. Box 20

FI-53851 LAPPEENRANTA FINLAND

email: erik.vartiainen@lut.fi

Opponent: Professor Juha Toivonen

Tampere University

Faculty of Engineering and Natural Sciences P. O. Box 1001

FI-33101 TAMPERE FINLAND

email: juha.toivonen@tuni.fi

(6)

(7)

Sampson Saj Andoh

PHOTONIC SOLUTIONS FOR ADULTERATION DETECTION IN EDIBLE PALM OIL

Joensuu: University of Eastern Finland, 2021 Publications of the University of Eastern Finland Dissertations in Forestry and Natural Sciences N:o 431

ABSTRACT

Edible oils are essential sources of fats and oils, which is a fundamental component of a human balanced diet. It is also an indispensable component in the global food industry, especially in almost all ready-to-eat foods. Palm oil, one of the most useful edible oils in the world currently, has in recent times been found to be adulterated.

Adulteration in edible palm oil are either by addition of other waste edible oil (such as already used frying oil), addition of unauthorized colorants such as Sudan dyes, or addition of residue from the palm oil process. As a remedy, we demonstrate the use of several photonic-based methods for screening in-house adulterated samples as testing kits before application on real edible palm oil.

We achieved this by proposing, developing, and demonstrating different optical and photonic methods; surface-enhanced Raman spectroscopy, refractive indices and spectrophotometric measurements, for authenticating and subsequent detection of Sudan dye adulteration in edible palm oil. These methods opens up a new opportunity for accurate, portable and subsequently on the spot detection of adulteration and provide quality control in edible palm oil adulteration at points of entry, markets, and the factories. This reduces the time to test for adulteration in the mostly laboratory based techniques. These methods also eliminate completely sample preparations, thereby saving time, energy and money. The results obtained are additionally easy to interpret.

Universal Decimal Classification:53.084.85, 535.3, 535.4, 681.7.02 OCIS codes:050.1960, 130.4815

Keywords: Edible oil; palm oil; adulteration detection; food fraud; plasmons; refractive index; surface-enhance Raman spectroscopy; IR spectroscopy; transmittance; optical analysis;

speckle pattern; speckle contrast

(8)

ACKNOWLEDGEMENTS

It is fascinating how it seems nothing changes on daily basis but looking back, so much has.

First and foremost, I would like to thank Jehovah for giving me the breath of life and for His shining light in my life to complete this work.

Secondly, this thesis would not have been impossible if not for the contributions, support, advice and encouragement of certain people. I am eternally grateful to my circle of supervisors, and advisors Professor Matthieu Roussey, Professor Pasi Vahimaa, Professor Seppo Honkanen, Professor Jyrki Saarinen, Professor (Emeritus) Kai-Erik Peiponen, and Dr Tarmo Nuutinen. Sirs, I cannot thank you enough, I say kiitosfor all the insightful guidance, thoughtful support and meaningful encouragement throughout these years and most importantly,’making a doctor out of me’. These lessons will impart the rest of my life. I am thankful to the reviewers for making the work better.

I am also thankful to the Dean and vice dean of the faculty of Science and Forestry; the Head, and Vice Heads of the Department of Physics and Mathematics, for providing the financial assistance throughout my doctoral studies. I am particularly grateful to the Administrative and Technical staff of the Department of Physics and Mathematics;- Katri, Hannele, Marita, Noora, Pertti, Tuire and Tommi, for their immense support and help, whenever I needed administrative and/or technical advice and support.

My sincerest and deepest gratitude belong to my family. I am particularly thankful to my wife, mom, siblings, and my daughter–Michelle Ewura Ama Andoh. I can only say thank you guys for enriching my life and I am happy to be part of your lives as you are of mine.

Lastly, I am thankful to Narrow Way Musical group, friends and other colleagues for the laughter, company, and everything, all this while.

Joensuu, October 15, 2021 Sampson Saj Andoh

(9)

ToMichelle— for changing my perspective of life

and

Abena— for everything

(10)

(11)

LIST OF PUBLICATIONS

This thesis is based on reviewing the following articles by the author.

I S. S. Andoh, K. Nyave, B. Asamoah, B. Kanyathare, T. Nuutinen, C. Mingle, K-E. Peiponen, and M. Roussey, "Optical screening for presence of banned Sudan III and Sudan IV dyes in edible palm oils,"Food Addit. Contam. A37(7), 1049–1060 (2020).

II A. Dutta, A. Matikainen, S. S. Andoh, and T. Nuutinen, "SERS Activity of Photoreduced Silver Chloride Crystals," AIP Conference Proceedings. 2220(1), 050004 (2020).

III S. S. Andoh, T. Nuutinen, C. Mingle, and M. Roussey, "Qualitative analysis of Sudan IV in edible palm oil,"J. Eur. Opt. Soc.-Rapid Publ.15, 21 (2019).

Throughout this overview, these selected papers will be referred to by the Roman numerals (I–III).

International conferences in which results of this research have been presented and the author’s contribution in other scientific publications are listed in Appendix A.

(12)

AUTHOR’S CONTRIBUTION

The publications selected in this dissertation are original research papers on different optical and photonic methods for palm oil authentication and adulteration detection. The author has had a central role in all aspects of these research papers recorded in this thesis. The author was active in developing the initial ideas, planning, and conducting all the experimental work, and data analysis inI. The author was also actively involved in all of the experimental work and data analysis inII.

The author was equally involved in the formulation, planning and executing all the experimental work, data collection, and data analysis in III. The author wrote all the first drafts of the manuscript inI–IIIbefore other authors and collaborators also took part in finalizing the research papers.

(13)

1 Introduction

Fats and oils, which can be found in edible oils, are fundamental elements of a balanced diet for human beings. They are key vectors for vitamins absorption and bring energy in the body. Edible oils are also known as vital sources of energy for the body. Due to relevant reasons such as health, ethical, economical, religious, and many others, edible oils are to remain unadulterated [1]. In this thesis work, we study some photonic solutions for detecting adulteration of palm oil. Palm oil is currently the most utilized edible oil in the world [2, 3].

1.1 PALM OIL

Palm oil is a highly nutritious reddish hued oil produced from the fleshy mesocarp of the oil palm (Elaeis guineensis) fruits, see Figure 1.1. Food and non-food consump- tions necessitates its production [4, 5]. Palm oil is imperative in the food industry for various reasons such as its low cost, easy handling, and convenient utilization.

One other oil from the palm nuts is made from the hard kernel of the palm fruits as illustrated in Figure 1.1. This oil is known as palm kernel oil. This work focuses on the use of palm oil as a sample and not its production procedures or chemical properties.

Fleshy fruit (mesocarp)

Kernel Exocarp

Endocarp

Figure 1.1: Oil palmElaeis guineensistrees. The inset indicates the cross-section of a palm nut and its various parts [6].

Currently, it is estimated that palm oil is found in most of the ready-to-eat foods [4]. This is due to palm oil’s intrinsic properties such as decolorizing easily, and semi-solid quality at temperatures below 25^◦C. Palm oil is also known for its health

(16)

benefits as it is used in treating vitamin A deficiency in some cultures, mainly in Africa and Asia [7, 8]. Palm oil has become a preferential target for adulteration by some fraudulent people to maximize profits due to its wide global usage. This problem is most prevalent in developing countries, where there are weak regulating institutions, and almost nonexistent detection tools and experts. Adulteration of palm oil is usually done to enhance the hue in masking the effect of poorly prepared oils. The most commonly used hue enhancers are Sudan dyes.

1.1.1 Adulteration in edible palm oil

Food adulteration is the addition of one or several unauthorized additives to food.

It is a serious issue with potential public health [9] effects, and even far-reaching economical [10], food safety [11], and ethical [12] consequences. It is the most common form of food fraud. Food adulteration may be deliberate, accidental or negli- gence [13] but in most cases, they are economically motivated [14]. Almost all foods can be adulterated in a way or another [15].

Edible oil adulteration is an unwarranted practice that may have a dramatic effect on the food industry and the economy of a country as a whole. Unfortunately, in recent times, edible oil adulteration is more and more common. These adulterations include the addition of inferior oils to high premium and expensive oils as in the case of olive oil adulteration [16, 17], addition of unwholesome oil, such as used or leftover vegetable oil to edible oil then labeled as fresh oil [18, 19], and addition of unauthorized additives to improve the hue or taste of the edible oil, for example addition of copper chlorophyll to olive oil [20,21] or addition of Sudan dyes to palm oil [22, 23]. Currently, palm oil is the most adulterated edible oil globally [2, 3].

Adulteration of palm oil usually takes place during production of crude palm oil or processed palm oil as illustrated in Figure 1.2. Adulteration can take the form of the addition of sludge palm oil, volatile organic compounds, used cooking oil, recycled frying oil and others such hue enhancers to palm oil. This work only focuses on the use of Sudan dyes to improve the color palm oil samples.

Figure 1.2:Palm oil production circle indicating possible adulteration points.

(17)

1.1.2 Sudan dyes

Sudan dyes are lipophilic azo dyes with reddish hue used for coloring tiles, polishes, and other industrial products [13, 14]. They are banned as food additives by the International Agency for Cancer Research (IARC), and in almost all countries, as there is no tolerable daily dose due to their perceived carcinogenic and mutagenic properties [24–26].

In this work, we focus on their hue enhancement properties and, using different photonic principles, in detecting these adulterations. This work focuses mainly on Sudan III and IV dyes, as they are the two primary palm oil hue enhancers [27]. This is due to their bright reddish color, wide availability and low cost [25]. Sudan III and IV are physically and chemically similar with very close molecular structures.

Sudan IV has a slightly more intense color than Sudan III [28] because they possess two additional methyl groups as compared with Sudan III as depicted in Figure 1.3.

(a) (b)

Figure 1.3:Molecular structures of (a) Sudan III and (b) Sudan IV dyes [29].

1.1.3 Analytical tools for adulteration detection in edible palm oil

Most countries have strict standards concerning the quality of palm oil. For example palm oil must have about 90% triacylglycerols, 3−8% diacylglycerols, and 1%

minor components such as carotenoids [30–32]. In spite of that, the palm oil’s global usage in recent times has attracted the attention of fraudulent people in trying to enhance either its appearance, taste, or quantity in what is now come to be known as commercially motivated adulteration [10]. In such endeavors, other oils such as palm kernel oil, coconut oil or used cooking oil and addition of unauthorized colorants are used.

As the fraud techniques evolve and spread, the need for efficient tools for the authentication of palm oil urges the scientific community in developing new and effective methods and procedures. This pushes for high through-put techniques in detecting adulteration. For example, to detect adulteration of palm oil using other edible or used cooking oil, high-performance liquid chromatography-electrospray ionization-mass spectrometry (HPLC-ESI-MS) is used [33, 34]. This method is relatively fast and simple for individual analysis. In the study of Zhanget al. [3], they differentiated five separate edible oils using gas chromatography coupled with mass spectrometry (GC-MS) based on their fatty acid profiles. Different other analytical techniques are used in detecting adulteration in edible palm oil [4, 13, 14, 26].

It must be noted that depending on the stage of the palm oil production or food processing, different analytical tools are used to either authenticate the palm oil or

(18)

detect the adulteration. For instance, gas chromatography-ion mobility spectrometry (GC-IMS) [35], dielectric spectroscopy [36], and gas chromatography-flame ionization detector (GC-FID) [37] can be used in detecting adulteration of palm oil with used cooking oil at the production stage while Raman spectroscopy [38], Fourier transform infrared (FTIR) [39] and Fourier transform near-infrared (FT-NIR) [40]

spectrometry are used in detecting adulteration of processed palm oil with other already used cooking oil and/or other adulterants such as hue enhancing agents Sudan dyes. The current complex analytical laboratory-based tools lead a cost effect on the products.

Although these analytical tools have proven to be essential in detecting adulteration in edible palm oil by providing highly accurate, most efficient, and repeatable results, they possess serious setbacks. Such methods are time-consuming, expensive, require a high level of expertise in operation and most significantly, are laboratory- based.

1.2 MOTIVATION AND SCOPE OF THE THESIS

Palm oil, like any other food products, can be adulterated at any point of the production. Figure 1.2 shows this production line and points out that oils are transferred several times before being in hands of well-controlled organisms. Most of the developed countries have established regulations and protocols to fight against fraud and adulteration. As already stated, this demands expertise, time, and money, which are often weak or entirely non-existent in developing countries where palm trees are cultivated and palm oil is usually manufactured.

Such an adulteration poses a grave menace for all living things, including human, and has the potential to destroy, like a cancer, the economy of a country.

Several adulterants can be mixed with palm oil. In this thesis, we focus on the perceived carcinogenic hue enhancer Sudan dyes III and IV. These dyes are used in infinitesimal quantities difficult to measure without laboratory equipment. How- ever, the problem of adulteration arises in the manufacturing of the oil and in the marketplaces. There is immediate attention for reliable, efficient, and fast methods to determine such dyes in palm oil. Ideally, these methods must be applicable on the field.

The scope of this work focused on Ghana, a developing country yet a significant global palm oil producing country. In 2015, the Food and Drug Authority of Ghana (FDA-Ghana) issued a communique alerting the public of the distribution and sale of palm oil adulterated with Sudan dyes [41]. This was after 98% of all the palm oil samples FDA-Ghana collected from different sellers in 10 major open markets in the Capital Province, the Greater Accra Region, tested positive to Sudan IV dye adulteration.

FDA-Ghana is the central competent authority in matters of food and drugs in the country. We evaluated the ability of different techniques to detect the adulteration. All the samples used in this work were either collected from the open market in Ghana or supplied by the FDA-Ghana. It is important to note that palm oil com- positions differ a lot depending on the production recipe but also on the place of culture, and origin of cultivation of the palm tree. For instance, the concentration of carotene, which depends on the soil of culture affects the color of the palm oil which then cannot be an indicator for adulteration detection.

(19)

One technique we used is using transmission and refractive index measurements of the palm oil samples. Easy to obtain, these data can be processed with any type of computing tool (computer or smartphone) and collected with handheld devices.

The study, performed over real samples from Ghana, has been published in Paper I. This procedure would enable, in a near future, a fast screening of the oil products in local marketplaces.

The other technique is the surface-enhanced Raman spectroscopy using low-cost plates that we developed at the University of Eastern Finland. These novel SERS substrates allow a measurement without any sample preparation and could lead to a collection ready-to-measure already in the field. We have shown such results in PapersIIandIII.

1.3 THESIS OUTLINE

This dissertation is organized into six chapters. The first one introduces the subject and the challenges it addresses, namely the adulteration by Sudan dyes of palm oil and the need for low-cost and simple methods to detect it on-field. Different methods have been evaluated and the background principles of each are summarized into a separate chapter and the results presented in the last one. For each one, we describe the technique itself, then explain how the experiment has been established in the laboratory for this particular thesis work at the University of Eastern Finland (UEF), and finally we give the procedure to prepare the samples or describe the samples that have been used.

In Chapter 2, we describe the propagation of light through a medium and the different phenomena that can result in light interaction with a medium. In Chapter 3, we used these different optical phenomena to detect Sudan dyes in palm oil (PapersI,IIandIII).

In Chapter 4, the main results from the different approaches are presented and discussed, shedding more light on such optical and photonic properties as adulteration detection tools. A comparison of the different methods used in authenticating and detecting adulteration palm oil is also presented in this chapter.

Finally, the overall conclusions in the form a summary and future prospects from this work are detailed in Chapter 5.

(20)

(21)

2 Propagation of light in an oil medium

This chapter presents how light propagates through a medium. As we briefly explain this process, emphasizes will be given to the physical parameters such as refractive index (RI,n) that are directly linked to the inherent properties of the oil.

Transmission and difference of transmission data as well as the phenomenon of Raman scattering as result of light interaction with a medium will be briefly elaborated. Other parameters such as complex refractive index, excess permittivity, and difference of transmission, that can be extracted from direct measurements ofnand Twill be briefly explained. This will enable easy and clear screening of adulteration in edible palm oil in the subsequent sections.

2.1 REFRACTIVE INDEX AND EXCESS REFRACTIVE INDEX

The refractive index of a material is a dimensionless constant that describes the interaction of the medium with incident light. RI can be thought of as a measure of how fast light travels through a material. It is defined mathematically as the ratio of the speed of lightcin a vacuum, to the speed of light ν in the medium, and is expressed as

n= ^c

ν. (2.1)

Considering a light beam incident from air (n=1) with an angleθ_i with respect to the normal to the interface air/medium, one can define a reflected beam at an angle,θr= -θ_iand a transmitted beam at an angleθt, see Figure 2.1.

Figure 2.1: Schematic representation of an incident beam at an interface between vacuum (or air), with refractive indexn=1 and a medium, with refractive indexn.

The Snell’s Law, Eq. 2.2, describes how the beam will be deflected in the medium.

Measuring the angle of the transmitted beam gives a measure of the refractive index

(22)

of the material. Mathematically the Snell’s Law can be written as,

n1sinθ₁=n2sinθ₂. (2.2) This can be rewritten as, sinθ_i=nsinθt, wheren₁=1 andn₂=n(see Figure 2.1).

The refractive index depends on the material properties, such as its composition, density, and homogeneity. It depends on the light properties, e.g., the wavelength (dispersion), the power (in the case of nonlinear materials), and the polarization (in the case of anisotropic materials). It depends on the environmental or surrounding conditions, such as the temperature, pressure, and humidity [42, 43].

The complex refractive index is a quantity, N = n+ik where the real part is indicated by n, and kdenoting the imaginary part. The real part is related to the speed of light in the medium and the imaginary part describes the absorption (or the attenuation of the electric field in the medium).Ncan be expressed as a function of the permittivity and permeability of the medium as indicated by Eq. 2.3,

N=

√εµ

√ε0µ0, (2.3)

where ε and µ are, respectively the permittivity and permeability of the material through which light is traveling, while ε₀ and µ₀ are the corresponding constants for vacuum. For a non-magnetic material, the relative permeability isµr = ^µ

µ0 =1.

Measuring the (complex) refractive index, by ellipsometry [44–46], for instance, tells a lot about the material under investigation. In some cases, knowing the real part of the refractive is sufficient. Sincen depends on many factors, including the composition of the medium, we can use it to determine the proportion of the con- stituting materials in a mixture [47–50]. In this work, we use the refractive index to determine how a food product, i.e., palm oil, is adulterated. Refractometry is the measurement of the refractive index. We used this technique, illustrated in Figure 2.2, to determine the RI of all the palm oil samples. Refractometry can be used for the investigation of turbidity and other properties of complex liquids [51, 52].

Figure 2.2: Basic principle and sketch for refractometry. The sample (palm oil in our case) is placed on top of a prism. The deviation of the reflected beam gives a direct measure of the refractive index.

(23)

Excess refractive index, n^E is a measure of the effective refractive index of two liquids mixed together. It is a valuable indicator of the reactivity of two systems or liquids mixed together [52–55]. Excess refractive index can be a positive or negative number, depending on the reactivity of the binary solution, and in some cases it can zero. When it is zero, it means there is no chemical reactions and one gets an ideal homogeneous mixture. Such an understanding is achieved by monitoring the excess refractive indices as function of the volume fraction in adulteration for the binary mixture. In other words we know a priori the RI of both liquids, as well as, their volume fractions. Hence, one gets the RI of the ideal mixture from ne f f = f n1+ (1−f)n2. Due to chemical reactions the truene f f is different from the idealn and hence we call it excess RI. Limited literature is available for the use of the excess refractive index analysis as an anti-counterfeiting technique as elaborated in [52, 56]. For the very first time, we apply this in edible oil adulteration detection.

2.2 TRANSMITTANCE AND DIFFERENCE OF TRANSMISSION Spectrophotometry is a spectroscopy technique implying the measurement of the intensity of transmitted and reflected portion of light through and by a sample as a function of the wavelength of the illumination. It allows also the measurement of the absorbed light. Characteristic features of the spectra, e.g., peaks, dips, give access to information on the vibrational states of the constituents of the sample or analyte. Spectrophotometry is a powerful tool giving an optical signature of chemicals, analytes and constituents of a sample. Typically, the more the concentration of an analyte, the more the absorption increases, too [57, 58]. Spectrophotometry is nowadays an entire field of research applied in many different analytical domains.

Although a direct observation of transmission spectra can give a hint on the properties of the studied sample, it does not provide enough information for an accurate, precise and a stand-alone screening tool without additional and thorough data processing. We propose integrating the area of the transmission spectra over a wavelength range. We obtain for the standard palm oil a reference value,T_refand an unit valueT_sample for each of the other samples. Subtracting this value for each sample from that of the reference sample gives the difference of transmission,∆Tas defined in Eq. 2.4.

∆T= Z _λ₂

λ1

Tref(λ)dλ− Z _λ₂

λ1

Tsample(λ)dλ, (2.4)

whereTrefis the transmittance for the analytical palm oil,Tsampleis the transmittance for the sample and∆Tis the difference of transmittance.

This difference of transmission gives a second reading that could be used to authenticate the quality of the market samples. This improves the markings that can be used in the adulteration detection studies. To the best of our knowledge, the difference of transmission is being applied in the field of palm oil adulteration detection for the very first time. In this work, the transmittance which follows the so- called Beer-Lambert law is measured by ultraviolet/visible (UV/Vis) and infrared (IR) spectrophotometry.

Beer-Lambert’s law describes the absorption of light in a medium by relating the incident light intensity(I0)to the transmitted light intensity(I), see Figure 2.3. This implies that the Beer-Lambert’s law is limited by the medium and the material of the container of the medium.

(24)

Figure 2.3:Beer-Lambert’s law schematic illustration.

Mathematically, the Beer-Lambert’s law is expressed by Eq. 2.5;

I(λ) =_I₀(λ)_e⁻^α(λ)d_, _(2.5) whereαis the coefficient of absorption,λis the wavelength of the illumination light, dis the sample thickness, I0is the intensity of the initial incident light and Iis the intensity transmitted light. Solving for the coefficient of absorptionα(λ), the Eq. 2.5 becomes

α(λ) = ¹ dln 1

T(λ)^, ^(2.6)

whereTis the transmittance and is expressed asT=I/I₀.

Spectrophotometric methods used for food analysis are the various spectrophotometric tools, sensors, devices, testers, and probes for the chemical and/or physical property analysis of various food items. They are usually based on a particular wavelength of the electromagnetic spectrum. Spectrophotometric methods represent a relatively easy set of fast analytical methods in addition to other advantages such as being non-destructive methods requiring minimal to no sample preparations. Moreover, such methods are less expensive than other traditional chemical techniques such as MS, HLPC and others [59–61].

Spectrophotometry is currently widely used for identification, analysis and characterization of various food constituents of diary products, beverages, fruits and edible oils [62]. With these spectrophotometric methods, samples can be discriminated based on geographical origin [63–67] and other composition studies [68–71].

In the case of wines, for instance, spectrophotometry was used to estimate the main oenological parameters such as alcohol content, volatility, pH and so on [72–75].

There are varied applications of spectrophotometry in the case of edible oils, typically for the characterization and the adulteration detection analysis. Recent studies show, for example, how such methods can evaluate the thermal effect in edible oils [76, 77] or adulteration in olive oils [78]. Although UV/Vis spectrophotometry is a simple, cost-effective, and fast technique [77, 78], it remains a laboratory-based technique requiring high level of expertise in interpreting the results.

(25)

The spectral characteristics of edible oils are more visible by IR spectroscopy [79], i.e dealing with the infrared portion of the electromagnetic spectrum. In this regard, there are several studies for characterization, authentication, and adulteration detection in various edible oils such as olive oil, palm oil, sunflower oil and many others [80–85].

Nowadays, UV/Vis/IR spectrophotometry is considered as an ideal tool for the analysis of edible oils as it is rapid, non-destructive, and relatively easier in operation [86, 87]. Combining UV/Vis/IR spectrophotometry with chemometrics analysis also greatly enhances the performance of the spectrophotometry and eases the extraction, handling and interpretation of the obtained results by filtering out the irrelevant information from the data leaving only the most important ones [88–90].

2.3 RAMAN SCATTERING AND SENSING

Light scattering by matter leads to an energy exchange between the photon and the material. An exchange of energy (∆E) occurs pushing an atom or a molecule from a lower energy level to an excited state. The atom or molecule may emit another photon when leaving this unstable state and returning to the previous or a lower energy state. The scattering is said to be elastic if the energy of the emitted photon is∆E. Rayleigh scattering (Mie for large particles) is the most predominant elastic scattering occurring when light hits small particles (at the molecule size). If the energy of the emitted photon is different from the incident one, the scattering is inelastic [91, 92]. Compton, Brillouin and Raman scatterings belong to this category where there is an energy difference in the incident and transmitted photon. In this thesis work, we focus only on the Raman scattering and its application for sensing.

This change in transferred energy leads to a decrease or increase of the frequency of the emitted photon. This very tiny variation is called Raman shift, and is quantified by∆ωin Eq. 2.7

∆ω=10⁷

1

λex

− ¹ λ

, (2.7)

where∆ω is expressed in cm⁻¹and the wavelength,λ, in nm. The laser excitation wavelength (λex) is extremely important and can influence the obtained Raman shift.

A loss in energy corresponds to the Stokes scattering, a gain of energy to the anti- Stokes scattering [93]. The schematic presented in Figure 2.4 illustrates Rayleigh and Raman scattering mechanisms from the point of view of the energy diagram of a molecule.

∆Ein Figure 2.4 corresponds to the actual energy of specific vibrational mode of the molecule under study, which is also related to the chemical bonds and structure of the molecule. Raman scattering is therefore a vibrational spectroscopic technique that can give a unique signature for a particular molecule, revealing the chemical composition of the analyte [94–96]. Raman signals are extremely weak i.e. less than 0.001% of an appropriate incident light results in Raman scattering. Such effect led to different techniques such as coherent anti-Stokes Raman scattering [97] (CARS), surface-enhanced Raman scattering [91] (SERS) and many others as detailed in [98], to increase the low Raman signal intensities.

(26)

ΔΕ

Ener

gy

Ground State

hv hv- hv+

Energy

Raman shift Vibrational state anti-Stokes

Stokes Rayleigh

Virtual state

Figure 2.4:Energy diagram illustrating Rayleigh, Stokes and anti-Stokes scattering.

2.4 SURFACE-ENHANCED RAMAN SPECTROSCOPY

Surface-enhanced Raman spectroscopy (SERS) couples nanotechnology and Raman scattering to present a robust technique in chemical analysis and detection even in trace quantities [96,99,100]. With the aid of roughened or nanostructured metal surface, the weak Raman signals are remarkably enhanced when the light of adequate frequency and intensity is incident on the material.

This phenomenal enhancement relies on two mechanisms: the electromagnetic (EM) [101] and the chemical enhancement (CE) [102]. While the CE is generated by the formation of charge-transfer complexes between the metal and the analytes on the surfaces, the EM enhancement mechanism is based on the excitation of localized surface plasmons (LSP). It is established that enhancement factors of ∼ 10⁵−10⁷ and∼ 10⁰−10² are observed for EM and CE mechanisms, respectively [103–105].

Although CE mechanism contribution is small, as compared to the EM enhancement mechanism, it is extremely significant in the overall enhancement of the Raman signals [91, 106].

The LSP, which generates the majority of the enhancement, is thought of as col- lective oscillations of surface electrons. A particle is dipole induced when light of suitable frequency is incident on it. The Valence electrons are set in oscillations and those electrons in close proximity to each other can couple their individual oscillations together creating a surface plasmon. Analyte molecules in close proximity to these oscillations also experience induced high fields. A counter resonance frequency is created by the analyte molecules on the substrate to nullify the effect of the oscillations of the molecules. This frequency is the so-called surface plasmon resonance frequency.

(27)

2.4.1 Surface plasmon polaritons and localized surface plasmons

Nanoparticles become dipole induced when light wave of appropriate frequency is incident on them [107] and may experience forced high fields. In order to nullify the effect of the experienced high fields, the molecules induce a resonance frequency that matches that of the incident light. This creates what is known as surface plasmon resonance (SPR) [107, 108]. This SPR can propagate as in the case of surface plasmon polaritons (SPP) or localized as observed in localized surface plasmons (LSP). This is illustrated in Figure 2.5 below.

(a) (b)

Figure 2.5: Schematic illustration of (a) localized surface plasmons (LSP) and (b) surface plasmons polaritons

SPPs possess attractive properties that have been applied in numerous domains such as sensing [109–111], photovoltaics [112, 113], and lasers [114, 115]. As the LSPs are confined to the surface of the nanoparticles, they generate high local electric fields in between the intermolecular spaces of the nanoparticles. This causes a restoring force on the driven electron cloud leading to a field enhancement both inside and in proximate surroundings of the particle [116]. A direct observation of the SPP is a dip in the reflected spectrum. The position and amplitude of such a dip depends on the metal used, the thickness of the layer, and the environmental surrounding, namely the refractive index of the medium above the metal layer.

2.4.2 Nanoparticles as enhancement agents

Nanoparticles (NPs) are particles with sizes ranging from 1 to 100 nm [117]. Nanopar- ticles possess unique properties that are virtually absent in the bulk material due to their large surface-to-volume ratio [118, 119]. NPs of the same material exhibit different optical, physical, and chemical properties depending on the size [120], shape (rods, spheres, wires) [121–123], and surface modifications (smooth or rough) [102,124]. Their presence generates a metal/dielectric interface that enables EM and chemical enhancements. In manipulating the above-mentioned parameters, one can fine-tune the properties required for improved performance in the SERS signal [125].

Silver (Ag) and gold (Au) NPs are the two commonly used NPs to modify a substrate’s surface for enhancing Raman signals. Other metals used for enhancement purposes include copper, zinc, and their alloys [122, 123]. In addition to the geometry of the nanoparticles, light/metal interactions are highly influenced by the surrounding medium [126]. Silver and gold nanoparticles provide a sharp and deep

(28)

plasmon resonance in the visible (Vis) and near infra-red (NIR) regions, respectively. AgNPs usually possess superior optical properties than AuNPs having the same size, shape, and in the same surroundings [126].

(29)

3 Determination of adulteration in edible palm oil

In this chapter, the use of basic intrinsic parameters such as RI,T, SERS, and other parameters dependant on these, in authenticating and subsequent adulteration detection edible palm oil is presented.

The experimental procedure, and samples used together with the equipment requirement for each measurement is also outlined in this chapter.

3.1 REFRACTIVE INDEX AND SPECTROPHOTOMETRIC METHOD The techniques we previously described have been used to analyze adulterated palm oils samples collected in Ghana. We established a procedure to allow fast screening of the oil, prior to further heavy and robust measurements, limiting then the number of resources required in the fight against counterfeiting.

3.1.1 Samples

The Food Physico-Chemical Laboratories of the Ghana Food and Drugs Authority (FDA-Ghana) provided us with 39 palm oil samples from the eight largest producing regions among the sixteen in the country, see Figure 3.1.

FDA-Ghana is the central competent authority in charge of food and drugs in Ghana. Some of the samples were collected from open markets, while others were ready for export according to FDA-Ghana. It must be noted that, the samples were collected before the referendum on 27 December 2019 by the Government of Ghana to divide the Administrative regions of Ghana from 10 to 16. There was a landslide approval in the referendum [128]. These eight (previously five) regions are known to be palm oil producing regions in Ghana.

The samples were sorted and grouped according to their geographical origin.

Three sample items were indicated as unadulterated and thus were separated and labeled accordingly as standard palm oil (STDPO). Then the rest of the samples were labeled as FDAx, where x = 1−5 depending on the region of origin (before the 2019 referendum changes) as indicated in Fig. 3.1. Ten other in-house adulterated samples were also prepared, in volume ratios of 5−25% of known adulteration concentration of Sudan III and IV dyes. These were labeled as ST3POiand ST4POi for Sudan III and IV dye respectively wherei= 5, 10, 15, 20, and 25 indicates the volume fraction of the dye added. This was used as a test and reference kit for our method. Let us remark that the high percentage considered here is due to the dilution of the dyes in acetone, used for an easier mixing and homogeneity in the palm oil. It corresponds to a very tiny amount of dyes for example in mass, the actual concentration of Sudan III and IV is 0.0022% wt to 0.014% wt and 0.0024% wt to 0.015% wt for Sudan III and IV dyes, respectively. Details about their preparation are given in detail in PaperI.

(30)

Figure 3.1: The Administrative map of Ghana, indicating the regions from which the samples were collected [127].

3.1.2 Experimental procedure

The samples were characterized at 45 ± 2 ^◦C due to the semi-solid nature of the palm oil. The refractive index and transmission of each sample from the market (FDAx), in house prepared sample adulterated in laboratory (ST3POiand ST4POi), and considered unadulterated by the Ghanaian authorities (STDPO) have been measured systematically.

All the refractive indices of all the samples were measured with J57WR-S2 Abbe refractometer (Rudolph Research Analytical, USA 2013) atλ = 589 nm. It must be emphasized that the law of critical angle of total reflection is relevant although the Abbe refractometer is used.

The transmission spectra of all the samples were measured with Perkin Elmer Lambda 9 UV/VIS/NIR spectrophotometer, with a quartz cuvette the length of 10 mm. The spectrophotometric measurements were obtained atλ = 589 nm and for the wavelength range 400 nm <λ< 2500 nm.

An adulterant alters the physical, chemical, optical, and other properties of the original sample. Each chemical has a spectral signature and each medium has a refractive index. Mixtures of oil and dyes can be here considered as homogeneous media because of the high solubility of the dyes in the oil. RI and transmittance of the mixture depend on the ratio of the two chemicals and therefore RI and spec-

(31)

trophotometric properties can be used to detect adulteration in these samples. RI and transmittance measurements offer a way to classify adulteration [52, 54, 55, 129]

as an already extensive literature [58,60,130–134] has established. The nature and degree of change in these optical properties invariably depend on the constituents and the adulterants [53,135]. To analyze the data from the transmission spectra, we used chemometrics analysis tools to support and complement the RI results. The analysis is performed by combining the excess refractive index (n^E) [52, 53], difference of transmission (∆T) [136, 137], and principal component analysis (PCA) [88, 138] to distinguish adulterated palm oils from unadulterated samples. This method, used for the first time in food analysis opens new possibilities for discriminatory applications in other sectors such as pharmaceuticals, winery, and cosmetics industries.

This can also be implemented in handheld devices to be used in field measurement conditions.

Principal component analysis (PCA) was used to create a map of all the in-house adulterated samples and then, used to estimate the degree of adulteration in the FDA-Ghana samples. PCA is a well-established data-representative tool used in analyzing complex and heavy datasets [88,89]. It allows for a deeper and qualitative examination of the differences between authentic and adulterated palm oil samples.

Unscrambler X v10.5 (CAMO ASA, Oslo, Norway) was used to perform the PCA and verified with Origin Pro 2019.

The results of this study are presented in Chapter 4, Section 4.1 and in more detail in PaperI.

3.2 SERS APPLICATION FOR ADULTERATION DETECTION

In this section, we present the details on the experiments we performed during the PhD studies to measure SERS signals in order to detect adulteration in palm oil samples.

SERS, which is a rapid and sensitive spectroscopic technique based on Raman scattering [139, 140], has been demonstrated to be a potential alternative method for in situ measurement. In this work, all experiments have been carried in laboratory conditions.

3.2.1 Simulation of silver nanoparticulate plate

AgCl nanoparticles (AgCl) were grown as an intermediary step onto a silicon wafer to minimize oxidization of AgNPs. AgCl are then photoreduced, as and when measurement is to be made, to reveal fresh AgNPs.

In approximating the shape both the AgCl and AgNPs on the plate, each crystal was considered as a solid Ag sphere of an effective diameter of 350 nm in air. The dispersive dielectric function of Ag was extracted from Johnsonet al[141] and the nondispersive refractive index of air considered as unity.

Figure 3.2 shows the cross-section profile of a single photoreduced AgNP simu- lated in air and normalized geometrical cross-section.

Using MiePlot v4614, the scattering response at the far field for the unpolarized plane wave excitation was calculated. From the Figure 3.2, it can be observed easily that the AgNPs show a high scattering nature in the 533−564 nm for the 514 nm excitation laser and 830−906 nm for the 785 nm laser. This result indicate the effi- ciency of these AgNPs for SERS activity for both the 514 nm and 785 nm excitation wavelengths.

(32)

Figure 3.2: Normalized scattering profile of a single photobleached AgNP sim- ulated in air. The vertical dotted lines (red and green) show the excitation wavelengths of 785 nm and 514 nm respectively. The block colored regions indicate the Stokes signal windows of the analytes, which are 533-564 nm for the 514 nm laser and 830-906 nm for the 785 nm excitation wavelength.

Using SEMLeo 1550Gemini microscope, SEM characterization before and after photoreduction was measured. Three different analytes (rhodamin 6G, adenine, and riboflavin) were used to characterize the plate after simulation and fabrication for analysis. The results of these three analytes are presented in Figure 4.6. A more detailed information on the substrate fabrication and application can be found in PapersIIandIII.

3.2.2 Substrate fabrication

According to ref. [142], the AgNPs substrate was prepared for use. A schematic of the fabrication and the photobleaching processes is illustrated in Figure 3.3.

Figure 3.3:Schematic of the (a) fabrication and (b) photobleaching process [142].

In summary, silver chloride (AgCl) nanoparticles were grown as an intermediary step on a silicon wafer by repeatedly and periodically dipping a silicon wafer in 5 mM precursor solutions of silver nitrate (AgNO3) and sodium chloride (NaCl). This

(33)

step was to prevent the AgNPs from tarnishing and loosing its enhancing prowess.

A section of the plate is photoreduced for measurement when need be.

3.2.3 Samples

For this study, five sets of edible palm oil samples were supplied by the Food and Drugs Authority of Ghana (FDA-Ghana). Four of them were indicated to be adulterated with Sudan IV dye after FDA-Ghana has conducted their own analysis (these samples were labeled FDA1−4) and one was indicated as unadulterated (labeled Sample K). All chemicals and reagents used in this work were purchased from Sigma-Aldrich (St. Louis, Mo., USA).

In addition, three other in-house samples were prepared by diluting the appropriate concentration of Sudan IV dye in Sample K as presented in Table 3.1. Only the Sudan IV dye was used due to the fact it is the most widely used palm oil hue enhancer, according to the FDA-Ghana.

Table 3.1: Prepared palm oil samples with the volumes of unadulterated palm oil and Sudan IV dye to obtain the required concentrations.

Volume[_µL] Sudan IV concentration[_µM] Sample K Sudan IV dye

800 200 200

950 50 50

990 10 10

A 1µM stock solution of Sudan IV dye in acetone was prepared and used for spiking the analytical grade palm oil to prepare these in-house samples. These samples were used as learning tests for the method before using the real samples from the FDA-Ghana.

3.2.4 Equipmental procedure

A commercially available Renishaw inVia Raman microscope [143] connected to a Renishaw enabled program WIRE^R on a computer was used to obtain all the SERS spectra of all the samples. A schematic of this Raman microscope is shown in Figure 3.4. The sample mount stage was modified to inculcate a heating system to have the palm oil samples in liquid form. This was to mimic the field conditions in temperate regions and also to have the palm oil samples in liquid form during measurement.

All the spectra were obtained with an excitation wavelength of 514 nm, power of 50 µW and integration time of 10 s. The SERS signals were measured for Raman shifts between 1200 cm⁻¹ to 1800 cm⁻¹, which is the region where the signature of palm oil and its adulteration can be found.

In this work, principal component analysis (PCA), a well known data representative technique [88,144], was applied to the SERS spectra to distinguish adulterated samples, the FDA and the in-house prepared samples. With this, a deeper dis- crimination between the authentic and the adulterated palm oil samples could be established. The PCA analysis was performed using the OriginPro 2019 software.

The results are presented in Chapter 4.

The use of SERS in trace analysis is wide spread but its application in adulteration detection in edible oils is mainly focused on olive oil analysis and minimum

(34)

Figure 3.4:Schematic of the Renishaw inVia Raman microscope.

on the most used oil in the world. This part of the work, however, re-enforces the viability of SERS application in palm oil adulteration considering the added advan- tage of no sample preparation, relatively easy interpretation of results and minimal time for operation it presents, thanks to our SERS substrate. This study led to the publication of PapersIIandIII.

(35)

4 Results and discussion

In this chapter, the main and relevant findings of Papers I–III are summarized.

Firstly, the results based on the use of refractive index and transmission measurements in authenticating and detecting adulteration in edible palm oil are presented (PaperI). Secondly, we present the use of surface-enhanced Raman spectroscopy in palm oil adulteration detection. An optimized substrate fabrication method is presented (PaperII) then its subsequent use in adulteration detection (PaperIII).

4.1 REFRACTIVE INDEX AND SPECTROPHOTOMETRIC MEASURE- MENTS

A summary of the results from PaperIis presented in this section. The objective is to determine the degree of adulteration in commercially available palm oil samples from different regions of Ghana. The origin of each sample is essential since manufacturing processes and amount differ from one region to another.

4.1.1 Results

Refractive index measurements

The refractive indices distribution for each region for all the 36 samples from the FDA-Ghana were measured. We then calculated an average refractive index per region. This is indicated in Figure 4.1(a). The refractive indices as a function of the volume fraction of dye in the palm oil is also reported in Figure 4.1 (b) below.

a)

FDA1FDA2FDA3FDA4FDA5

1.466 1.468 1.470 1.472 1.474 1.476 1.478

Refractive indexn

Figure 4.1: Refractive index analysis of Ghanaian palm oil samples. a) Distribution of the refractive index for the palm oil samples. The average refractive index of each region represented the grey dots, and the black dashed line is the refractive index of STDPO. b) Refractive index of the prepared adulterated palm oil samples by spiking Sudan dyes solution in STDPO (ST3POi in blue and ST4POi in red).

The green dashed lines represent the Regional average refractive indices. The grey- shaded area corresponds to the refractive index of the STDPO, (n_STDPO±σ) where σis the standard deviation of the measurement.

(36)

The refractive indices of prepared samples are observed to range between that of standard palm oil and that of the adulterant (SUDIII and SUDIV) depending on the one used. The change is observed to be in the second-order decimal which is huge and clearly detectable. We also observed a linear decrease in the RI with an increase in the volume fraction of the adulterant. This linearity holds for both SUDIII and SUDIV volume fraction adulteration of the authentic palm oil, with a correlation coefficient of 0.992 and 0.995, respectively.

A great disparity is observed amongst the studied palm oil samples based on their RI measurements in the different regions. The calculated average standard deviation (σ) over all the samples was plotted in a grey shaded zone around STDPO corresponding ton_STDPO±σ. Almost all the samples studied fell within this range.

The intersection of the shaded zone and the linear regression of SUDIII and SUDIV gives the threshold of this method. This is determined from Figure 4.1 (b) as 2.24%

and 1.90% in fraction volume for SUDIII and SUDIV, respectively.

Transmission spectra

Figure 4.2 shows the transmission spectra of the authentic palm oil (STDPO), SUDIII and SUDIV, acetone, and a sample of the palm oil sample adulterated with SUDIII (ST3PO20) and SUDIV (ST4PO20).

Figure 4.2: VIS-NIR transmission spectra of STDPO (black curve), SUDIII (purple curve), SUDIV (yellow curve), and acetone (green curve). One SUDIII (ST3PO20, in red) in-house prepared sample and SUDIV (ST3PO20, in blue) is also shown indicating the 1200 nm, 1400 nm and 1700 nm (in a black dashed lines) absorption bands.

It can be easily observed that all the spectra have a similar shape with three dis- tinct absorption bands atλ∼1200 nm,λ∼1400 nm andλ∼1700 nm as indicated.

An overall higher transmission for SUDIII and SUDIV is observed compared to the oil samples. There is also a bending of the spectra in the visible part of the adulter-

(37)

ated palm samples. Secondly, the NIR part of the spectra drastically differs for the prepared samples than for the acetone, SUDIII and SUDIV. This huge drop in the peaks corresponds to the chemical activity between the dyes, acetone and the oils.

The transmission spectra of all the other SUDIII and SUDIV adulterated palm oil samples are presented in Figures 4.3.

1800 1900 2000 2100 2200 2300

0,000 0,025 0,050 0,075 0,100 0,125 0,150

Transmittance

Wavelength (nm)

STDPMOH ST3PM5H ST3PM10H ST3PM15H ST3PM20H ST3PM25H

500 1000 1500 2000 2500

0,0 0,2 0,4 0,6 0,8 1,0

Transmittance

Wavelength (nm)

500 1000 1500 2000 2500

0,0 0,2 0,4 0,6 0,8 1,0

Transmittance

Wavelength (nm)

1800 1900 2000 2100 2200 2300

0,000 0,025 0,050 0,075 0,100 0,125 0,150

Transmittance

Wavelength (nm)

a) b)

c) d)

Figure 4.3: Transmission spectra of the prepared samples with the adulterants. a) with SUDIII as adulterant in the range 400 nm <λ < 2500 nm. b) with SUDIII as adulterant in the range 1750 nm <λ< 2300 nm. c) with SUDIV as an adulterant in the range 400 nm <λ< 2500 nm. d) with SUDIV as an adulterant in the range 1750 nm <λ< 2300 nm.

It can be observed from Figure 4.3 that there is no evident dependence of the adulterated samples and the volume fractions of SUDIII and SUDIV. It is observed that ST3PO15 and ST3PO25 spectra present a higher transmittance than ST3PO5 and ST3PO20 samples. The transmission spectra of ST3PO10 is below all the others. An identical behaviour is observed for SUDIV adulterated samples. Such behaviour in both SUDIII and SUDIV adulterated samples is a form of chemical activity of the liquid mixture showing excess property in transmittance such as excess refractive index.

A similar drop of the transmission spectra in the near-infrared region was also observed for the FDA samples. It is noteworthy that as the adulteration was prepared in acetone, which is not the case in the FDA samples, there is little to no effect of the presence of acetone in the samples. This was observed from the transmission spectra as there no change, even after heat treating the samples for 1.5 hours.

Sufficient conclusion could not be drawn by a simple observation of the refractive index and the transmittance spectra in discriminating adulterated palm oil samples from authentic ones. We therefore proposed as a novel detection tool a combination

(38)

of the use of excess refractive index and the difference which is based on the direct measurement of the refractive index and the transmission spectra.

4.1.2 Discussion

Excess refractive index and difference of transmission

Analysis of the excess refractive index, n^E results for all the spiked samples with SUDIII and SUDIV, and the FDA samples are presented in Figure 4.4. This is presented for both the original samples and after heat treating the samples

5 10 15 20 25

0,000 0,005 0,010 0,015 0,020 0,025 0,030

ExcessrefractiveindexnE

Volume fraction f[%]

n^E_SUDIII

n^E_SUDIII(after heat treatment) n^E_SUDIV

n^E_SUDIV(after heat treatment)

FDA1 FDA2

FDA3/FDA5 FDA4

Figure 4.4: Excess refractive index analysis of the FDA and in-house SUDIII and SUDIV adulterated samples compared to the reference sample, STDPO. The dashed lines represent the FDA sample values. SUDIII spiked samples before and after heat treatment in black and red, respectively, whiles SUDIV spiked samples are represented in blue and green, respectively, for before and after heat treatment.

As noted, the acetone may vary the chemical activity of the dye and the palm oil, although no drastic change was observed in the spectra of the prepared samples. As there is no acetone in the FDA samples it is essential to consider the effect of acetone. We also noted from the RI measurements that heating the samples essentially removed the acetone.

It is observed from Figure 4.4, that the curves keep changing to monotonic after heat treatment. The monotonous is clearly observable and nearly linear with a corresponding increase in n^E. This is more visible in SUDIII than in SUDIV. This may be from the delocalization of the electrons inside the SUDIII molecule. The two additional methyl groups may, however, block the reactivity in the case of SUDIV hence the observed chart in Figure 4.4.

It is further observed that the excess RI values of FDA1 and FDA4 are very close to the STDPO value. Based on FDA3 and FDA5 values, they seem to be adulterated while FDA2 shows a high excess RI value indicating a high volume of adulterants.

(39)

The difference of transmission was calculated by integrating the signals on the wavelength range where acetone has no influence. The result is presented in blue for SUDIII adulterated samples and red for SUDIV based samples in Figure 4.5.

STDPM OH ST3PM5H

ST3PM10H ST3PM15H

ST3PM20H ST3PM25H

ST4PM5H ST4PM10H

ST4PM15H ST4PM20H

ST4PM25H STAPM

5H STAPM10H

STAPM15H STAPM20H

STAPM25H FDA1

FDA2 FDA3

FDA4 FDA5 0

50 100 150 200 250

None Unclear Sudan III Sudan IV Adulteration

Ref. Prepared as adulterated with

Sudan III

Prepared as adulterated with

Sudan IV

Reference mixed with acetone

(no dyes)

Unknown adulteration (real samples)

T[a.u.]

Figure 4.5: Difference of transmission analysis for SUDIII (blue bars), SUDIV (red bars) and FDA samples (green bars). Standard palm oil with acetone only at different volumes is shown in orange.

The same trend is, however, observed in both SUDIII and SUDIV. In this instance, FDA2, FDA3 and FDA1 are closest to STDPO. This means FDA1, FDA2, and FDA3 are most probably unadulterated with either SUDIII and SUDIV. FDA4 and FDA5 can clearly be discriminated as adulterated with SUDIV, according to this method. This result also reaffirm that SUDIV is a predominantly used palm oil hue adulterant.

From the above results, it is observed that a simple refractive index and transmittance study of the sample may not provide sufficient data for analysis. It must be stated that combining excess refractive index and the difference of transmittance, however, provides a good analytical tool for adulteration detection analysis.

4.2 SURFACE ENHANCED RAMAN SPECTROSCOPY

SERS profiles are key tools enabling the authentication and successive detection of adulteration in edible palm oil. In this section, a summary of the results of Papers IIandIIIare presented.

(40)

4.2.1 Results

Characterization of the AgNP plate

From the SEM characterization, it is observed that after photoreduction, the AgCl particles collapse into irregularly-shaped AgNPs. The SERS spectra of the three analytes (Rh6G, adenine, and riboflavin) are presented in Figure 4.6 for the wave number range 1200 – 1800 cm⁻¹.

Figure 4.6: SERS spectra of a) Adenine, (b) rhodamine, Rh6G and (c) riboflavin.

Profiles were obtained with 514 nm lasing wavelength (power of 50µW, acquisition time of 10 s), and 785 nm lasing wavelength (power of 3.61 mW, acquisition time of 10 s).

From Figure 4.6, the distinguishable peak at 734 cm⁻¹ [145] of adenine is observed in (a), for both the 514 and 785 nm excitation wavelength. In Figure 4.6(b), identifier peaks of Rh6G at 1311 cm⁻¹, 1362 cm⁻¹, 1509 cm⁻¹, 1575 cm⁻¹, 1600 cm⁻¹, and 1650 cm⁻¹[146, 147] are easily recognized for both lasing wavelengths.

Then in Figure 4.6(c), the descriptive peaks of riboflavin at 1155 cm⁻¹, 1223 cm⁻¹, 1346 cm⁻¹, 1401 cm⁻¹, 1463 cm⁻¹, 1578 cm⁻¹, and 1630 cm⁻¹[148] are seen. These peaks are observable for both wavelengths.

From these results, one can remark that this AgCl NP plate, which shows irreg- ular AgNPs after the photoreduction, presents an effective and efficient plasmonic platform for SERS characterization of analytes at both VIS and NIR regions with high signal strength. Furthermore, a strong scattering is observed at multiple wavelengths. This makes the photoreduced AgNPs an ideal alternative in spectroscopic measurements where multi-wavelength excitation is paramount.

Photonic solution for adulteration detection in edible palm oil

Dissertations in Forestry and Natural Sciences

SAMPSON SAJ ANDOH

PHOTONIC SOLUTIONS FOR ADULTERATION DETECTION IN EDIBLE

PALM OIL

Sampson Saj Andoh

PHOTONIC SOLUTIONS

FOR ADULTERATION DETECTION IN EDIBLE PALM OIL

TABLE OF CONTENTS

1 Introduction

Fleshy fruit (mesocarp)

Kernel Exocarp

Endocarp

2 Propagation of light in an oil medium

3 Determination of adulteration in edible palm oil

4 Results and discussion

a)