**Using Fluorescence Spectroscopy to Diagnose Breast Cancer Provisional chapterUsing Fluorescence Spectroscopy to Diagnose Breast Cancer**

Tatjana Dramićanin and Miroslav Dramićanin Tatjana Dramićanin and Miroslav Dramićanin

Additional information is available at the end of the chapter Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/63534

#### **Abstract**

[87] Liitiä T. Application of modern NMR spectroscopic techniques to structural studies of

[88] Liitia TM, Maunu SL, Hortling B, Toikka M, Kilpelainen I. Analysis of technical lignins by two- and three-dimensional NMR spectroscopy. J Agric Food Chem. 2003;51(8):

[89] Ede RM, Brunow G. Application of two-dimensional homo-and heteronuclear correlation NMR spectroscopy to wood lignin structure determination. J Organic Chem.

[90] Kilpelaeinen I, Sipilae J, Brunow G, Lundquist K, Ede RM. Application of twodimensional NMR spectroscopy to wood lignin structure determination and identification of some minor structural units of hard-and softwood lignins. J Agric Food Chem.

[91] Sette M, Wechselberger R, Crestini C. Elucidation of lignin structure by quantitative 2D

[92] Martínez ÁT, Rencoret J, Marques G, Gutiérrez A, Ibarra D, Jiménez-Barbero J, et al. Monolignol acylation and lignin structure in some nonwoody plants: a 2D NMR study.

[93] Balakshin M, Capanema E, Gracz H, Chang H-m, Jameel H. Quantification of lignin– carbohydrate linkages with high-resolution NMR spectroscopy. Planta. 2011;233(6):

[94] del Río JC, Rencoret J, Marques G, Gutiérrez A, Ibarra D, Santos JI, et al. Highly acylated (acetylated and/or p-coumaroylated) native lignins from diverse herbaceous plants. J

[95] Ämmälahti E, Brunow G, Bardet M, Robert D, Kilpeläinen I. Identification of side-chain structures in a poplar lignin using three-dimensional HMQC-HOHAHA NMR

wood and pulp components. Finland: University of Helsinki; 2002.

260 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

2136–43.

1992;57(5):1477–80.

1994;42(12):2790–4.

1097–110.

NMR. Chem—Eur J. 2011;17(34):9529–35.

Phytochemistry. 2008;69(16):2831–43.

Agric Food Chem. 2008;56(20):9525–34.

spectroscopy. J Agric Food Chem. 1998;46(12):5113–7.

Optical spectroscopy methods have had considerable impact in the field of biomedical diagnostics, providing novel methods for the early or noninvasive diagnosis of various medical conditions. Among them, fluorescence spectroscopy has been the most widely explored mainly because fluorescence is highly sensitive to the biochemical makeup of tissues. It has been shown that tumors were easily detected on account of altered fluorescence properties with respect to fluorescence of ordinary tissue. Breast cancer is one of the most commonly diagnosed cancers among women in the world and also it is one of the leading causes of deaths from cancer for the female population. However, when detected in early stage, it is one of the most treatable forms of cancer. Therefore, fluorescence technologies could be highly beneficial in early detection and timely treatment of cancer. This chapter presents main results and conclusions that have been reported on the use of fluorescence spectroscopy for the investigation of breast cancer. It also gives an overview on the instruments and methodology of measurements, on the main endogenous fluorophores present in tissues, on the tissue fluorescence, and on the statistical methods that aid interpretations of fluorescence spectra. Finally, examples of using various fluorescence techniques, such as excitation, emission and synchronous spectroscopy, excitation-emission matrices, and lifetimes, for the breast cancer diagnosis are presented.

**Keywords:** fluorescence, breast cancer, fluorophores, tissue fluorescence, cancer diagnosis

## **1. Introduction**

According to World Cancer Research Fund International (WCRFI), 1.7 million women have been diagnosed with breast cancer in 2012 [1]. With 25% share of all diagnosed cancers, breast cancer

© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

is a leading malignant disease in female population throughout the world. It is also present in male population, but with about 100 times less incidence that is fairly stable over the last 30 years. Breast cancer incidence is highest in North America, West Europe, and Oceania (in highly developed countries); in 2012, record-high values of age-standardized rate per 100,000 were recorded in Belgium (111.9) and Denmark (105.0). Breast cancer is age- and hormone-status– related, being highest for woman 50–70 years old. Fortunately, prognosis is very good for early diagnosed disease. Five-year survival rate (see **Table 1**) is 100% for 0 and I stage conditions [2]. Therefore, the early diagnosis of breast cancer (and also all cancers) is a key for the successful treatment. On the other hand, in the absence of any early detection or screening and treatment intervention, patients are diagnosed at very late stages when curative treatment is no longer an option [3].


**Table 1.** Five-year survival rate of patients diagnosed with breast cancer of different stage.

Generally, tumor-observing methods can be categorized to screening, diagnostic, and monitoring tests. Clinical breast examination, mammography, and ultrasonography are the most common methods used for screening. Clinical examination by a specialist is usually the first step in breast cancer detection. However, this test is quite subjective and the results depend on the experience and skill of the examiner and most importantly, it does not detect small-size tumors. Mammography is the most cost-effective test, but with moderate sensitivity of 67.8% and specificity of 75%. Mammography combined with clinical breast examination slightly improves sensitivity (77.4%). For screening, ultrasonography is usually recommended to younger persons (premenopause), but it fails to detect microcalcifications and shows poor specificity (34%). Modern imaging techniques, such as positron emission tomography (PET), magnetic resonance imaging (MRI), and computed tomography (CT) are important tools for cancer detection and treatment. Some important features of the most common breast cancer tests are listed in **Table 2**.

Medical diagnostics of breast cancer based on optical spectroscopy and optical imaging are in the early stage of development in comparison with the traditional methods. But the need for sensitive and early cancer detection along with advances in technology played, and still plays, an essential role for the extensive research in this field. For example, lasers have provided a new technology for excitation, and microchips and miniaturized sensors have eased signal detection, while optical fibers have transformed the ways of access to the object of examination. The most attention among a variety of optical techniques is given to fluorescence, Raman spectroscopy, diffuse reflectance, elastic scattering spectroscopy, Fourier transform infrared microspectroscopy, near-infrared imaging, and optoacoustic tomography. These techniques offer several principal advantages over the traditional methods, including (a) noninvasiveness through the use of safe, nonionizing radiation, (b) display of contrast between soft tissues based on optical properties, and (c) a facility for continuous bedside monitoring. Detailed description of optical methods in cancer research is beyond the scope of this chapter, which is devoted to fluorescence techniques. Those interested in more information on this subject can read some of topical books (e.g., Biomedical Photonics Handbook edited by T. Vo-Dinh [4]).


**Table 2.** Some important features of the most common breast cancer tests.

is a leading malignant disease in female population throughout the world. It is also present in male population, but with about 100 times less incidence that is fairly stable over the last 30 years. Breast cancer incidence is highest in North America, West Europe, and Oceania (in highly developed countries); in 2012, record-high values of age-standardized rate per 100,000 were recorded in Belgium (111.9) and Denmark (105.0). Breast cancer is age- and hormone-status– related, being highest for woman 50–70 years old. Fortunately, prognosis is very good for early diagnosed disease. Five-year survival rate (see **Table 1**) is 100% for 0 and I stage conditions [2]. Therefore, the early diagnosis of breast cancer (and also all cancers) is a key for the successful treatment. On the other hand, in the absence of any early detection or screening and treatment intervention, patients are diagnosed at very late stages when curative treatment is no longer an

262 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

**Stage 0 I II III IV 5-year survival rate** 100% 100% 93% 72% 22%

Generally, tumor-observing methods can be categorized to screening, diagnostic, and monitoring tests. Clinical breast examination, mammography, and ultrasonography are the most common methods used for screening. Clinical examination by a specialist is usually the first step in breast cancer detection. However, this test is quite subjective and the results depend on the experience and skill of the examiner and most importantly, it does not detect small-size tumors. Mammography is the most cost-effective test, but with moderate sensitivity of 67.8% and specificity of 75%. Mammography combined with clinical breast examination slightly improves sensitivity (77.4%). For screening, ultrasonography is usually recommended to younger persons (premenopause), but it fails to detect microcalcifications and shows poor specificity (34%). Modern imaging techniques, such as positron emission tomography (PET), magnetic resonance imaging (MRI), and computed tomography (CT) are important tools for cancer detection and treatment. Some important features of the most common breast cancer

Medical diagnostics of breast cancer based on optical spectroscopy and optical imaging are in the early stage of development in comparison with the traditional methods. But the need for sensitive and early cancer detection along with advances in technology played, and still plays, an essential role for the extensive research in this field. For example, lasers have provided a new technology for excitation, and microchips and miniaturized sensors have eased signal detection, while optical fibers have transformed the ways of access to the object of examination. The most attention among a variety of optical techniques is given to fluorescence, Raman spectroscopy, diffuse reflectance, elastic scattering spectroscopy, Fourier transform infrared microspectroscopy, near-infrared imaging, and optoacoustic tomography. These techniques offer several principal advantages over the traditional methods, including (a) noninvasiveness through the use of safe, nonionizing radiation, (b) display of contrast between soft tissues based on optical properties, and (c) a facility for continuous bedside monitoring. Detailed description

**Table 1.** Five-year survival rate of patients diagnosed with breast cancer of different stage.

option [3].

tests are listed in **Table 2**.

In the past several decades, fluorescence spectroscopy has been applied to many different types of samples, ranging from individual biochemical species to organs of living people. It has been applied so far for almost every type of cancer, both *in*- and *ex vivo*, and it has demonstrated advantages over other light-based methods in terms of sensitivity, speed, and safety. Since the early twentieth century, it is known that tissues fluoresce when exposed to light of a suitable wavelength and that infiltrating tumors can be detected on account of altered fluorescence signals. These alternations are the result of the exceptionally high sensitivity of fluorescence on the biochemical makeups of tissues, and are premise for the diagnoses of tissue pathologies by fluorescence. Tissue fluorescence comprises emissions of a number of natural fluorophores (endogenous fluorophores) that have unique spectral characteristics when excited with ultraviolet or visible light. Among them, regarding fluorescence, the most important are tryptophan, reduced nicotinamide adenine dinucleotide (phosphate) (NAD(P)H), elastin, collagen, flavins, and porphyrins. The changes in the concentrations and microenvironments of fluorophores alter tissue fluorescence to a sufficient extent to detect metabolic and pathological changes related to precancerous and cancerous growth, even though fluorescence measurements are not capable of detailing structural changes in tissues. The fluorescence of tissue and tissue fluorophores are discussed later in detail in a separate section.

It is impossible to present beauty, complexity, versatility, and usefulness of fluorescence in a single chapter. Many good treatises could be used for this purpose [5–7]. Thus, with this chapter we intended to provide an overview of applications of fluorescence spectroscopy for the breast cancer diagnosis. This application is a small part of the extremely large and growing field of biomedical fluorescence. To do so, brief descriptions of basic principles of fluorescence, instruments, and techniques are given in Section 2. Tissue fluorescence, management, and interpretation of fluorescence data are explained in Sections 3 and 4. Fluorescence measurements ought to be subjected to mathematical quantification and interpretation for obtaining appropriate data for cancer diagnosis. Useful mathematical tools, mainly modern statistical tools, are described in Section 5. Examples of applications are given in Section 6, and they are selected to cover most of the fluorescence techniques listed in Section 2. These examples are mainly drawn from our own work, simply because we are most familiar with them.

## **2. Fluorescence spectroscopy methods and instrumentation**

Photoluminescence is the physical process of light emission from any substance that has not been heated and takes place from the electronically excited states. Depending upon the nature of the excited state, luminescence can be strictly divided into two types, fluorescence and phosphorescence. Fluorescence is a rapid and spin-allowed emission of light from singlet excited states, while in phosphorescence emission of light comes due to spin-forbidden transitions from triplet excited states to the ground state. The fluorescence process, depicted in **Figure 1**, is ruled by three important events: (1) excitation of a molecule by an incoming photon from the ground state *S*0, (2) vibrational relaxation of *S*1' excited state electrons to the lowest excited energy state *S*1, and, finally, (3) emission of a lower-energy photon and return of the molecule to the ground state.

**Figure 1.** One form of a Jablonski diagram explaining the process of fluorescence.

Fluorescence measurements are the core of any fluorescence technology, and they are presently widely utilized by scientists from many disciplines. Depending on the temporal nature of excitation and detection, fluorescence measurements can be classified as steady-state (implemented with constant illumination and observation) and time-resolved (performed with pulsed excitation). Generally, steady-state measurements are simpler than time-resolved since they require less complex instrumentation and are easier for interpretation. Also, they are sufficient for many applications, so they are more common in practice. However, being simply an average of time-resolved phenomena, steady-state measurements do not account all of the molecular information imparted to fluorescence. Characteristic examples are information on the distribution of emission decays or the nature of fluorescence quenching.

we intended to provide an overview of applications of fluorescence spectroscopy for the breast cancer diagnosis. This application is a small part of the extremely large and growing field of biomedical fluorescence. To do so, brief descriptions of basic principles of fluorescence, instruments, and techniques are given in Section 2. Tissue fluorescence, management, and interpretation of fluorescence data are explained in Sections 3 and 4. Fluorescence measurements ought to be subjected to mathematical quantification and interpretation for obtaining appropriate data for cancer diagnosis. Useful mathematical tools, mainly modern statistical tools, are described in Section 5. Examples of applications are given in Section 6, and they are selected to cover most of the fluorescence techniques listed in Section 2. These examples are

mainly drawn from our own work, simply because we are most familiar with them.

Photoluminescence is the physical process of light emission from any substance that has not been heated and takes place from the electronically excited states. Depending upon the nature of the excited state, luminescence can be strictly divided into two types, fluorescence and phosphorescence. Fluorescence is a rapid and spin-allowed emission of light from singlet excited states, while in phosphorescence emission of light comes due to spin-forbidden transitions from triplet excited states to the ground state. The fluorescence process, depicted in **Figure 1**, is ruled by three important events: (1) excitation of a molecule by an incoming photon from the ground state *S*0, (2) vibrational relaxation of *S*1' excited state electrons to the lowest excited energy state *S*1, and, finally, (3) emission of a lower-energy photon and return

Fluorescence measurements are the core of any fluorescence technology, and they are presently widely utilized by scientists from many disciplines. Depending on the temporal nature of excitation and detection, fluorescence measurements can be classified as steady-state (implemented with constant illumination and observation) and time-resolved (performed with pulsed excitation). Generally, steady-state measurements are simpler than time-resolved since they require less complex instrumentation and are easier for interpretation. Also, they are sufficient for many applications, so they are more common in practice. However, being simply

**2. Fluorescence spectroscopy methods and instrumentation**

264 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

**Figure 1.** One form of a Jablonski diagram explaining the process of fluorescence.

of the molecule to the ground state.

Fluorescence measurements can also be classified according to the type of observed data. Emission spectra are obtained by recording emission intensity over the spectral range of interest for the fixed energy of excitation. Excitation spectra are created by measuring the variation of emission intensity at the fixed energy (wavelength) while changing excitation energy in the desired spectral region. Generally, these spectra are symmetric (mirror image) because the same transitions are involved in both absorption and emission, and similar vibrational levels are present in both ground and excited state. Certainly, many exceptions may occur, but their explanation is beyond the scope of this chapter. Synchronous fluorescence method involves simultaneously scanning both emission and excitation wavelengths while maintaining the interval constant between emission and excitation (constant-wavelength mode) or maintaining the frequency gap constant (constant-energy mode). When selected interval corresponds to the difference between the excitation and emission maxima of a specific molecule, the emission of that molecule is maximally intensified in the measured spectrum. Consequently, synchronous luminescence spectra have sharper spectral characteristics than conventional, characteristics that are particularly valuable for discrimination between different biological tissues.

Biological materials are complex and comprise many different fluorophores, chromophores (molecules that absorb light but do not emit one), and light scatters. Because of that, the measurement of a single spectra, either emission or excitation or synchronous, is sometimes insufficient for the analyses and diagnostics. In this case, more comprehensive measurements are required to thoroughly combine excitation and emission features of tissue; commonly, this type of measurement is referred as multidimensional fluorescence measurement. Excitationemission matrices (EEMs), also termed excitation-emission landscapes, are the most widely used type of multidimensional measurements. They combine in a three dimensional (3D) space the set of emission spectra excited by the light of different wavelengths. One can say the EEM of a sample is its characteristic fluorescent fingerprint, taking that the majority of the fluorescence characteristics of the sample are included in EEM. Alternatively, three-dimensional total synchronous fluorescence spectroscopy (3D-TSFS) can be used for the same purpose. During 3D-TSFS measurements, a series of SFS spectra are recorded for a range of synchronous intervals; therefore, 3D-TSFS is a multidimensional extension of SFS. The term "threedimensional" refers to the space defined by the excitation or emission wavelength, synchronous interval, and the fluorescence emission intensity.

For medical diagnostics, measurements of quantum yield, polarization, and excited state lifetime may also be valuable. Quantum yield is the ratio of the number of photons emitted from fluorophore to the number of photons absorbed. Polarization gives information on the movement of fluorophore, if there is any, during the time between the absorption and emission of light, namely during the excited state lifetime.

To stimulate and measure fluorescence from a sample, one needs to use instrument with five basic components: (1) light source, (2) wavelength selector elements on the excitation and emission paths to/from sample, (3) sample holder/positioner, (4) polarizers, and (5) detector (**Figure 2**).

**Figure 2.** Five basic components of spectrofluorometer.

Polychromatic light from light source is dispersed on a dispersing element from which light beam of selected wavelength is directed on a sample to excite fluorescence. Most frequently, Xe arc lamps are used. These days, deuterium, tungsten, or halogen lamps are rarely utilized for excitation, since they are relatively weak sources for fluorescence. Lasers, laser diodes, and light-emitting diodes (LEDs) are also frequently used. They produce intense monochromatic light, so there is no need for the wavelength-selecting element on the excitation path. The drawback is that excitation spectra cannot be measured with these light sources. Tuneable and supercontinuum lasers are an exception, which can produce emissions over the given spectral interval, but these are rather expensive devices. Light dispersion can be accomplished with prisms and diffraction gratings; the latter is dominantly used in the modern spectrofluorometers. The detector can be either single-channeled or multi-channeled. The single-channeled detector, usually photomultiplier tube (PMT) or semiconductor, detects the intensity of one wavelength at a time. Multi-channeled detectors, such as charge-coupled device cameras (CCDs), record the intensity of emission over the range of wavelengths simultaneously. In addition, modern instruments include some other important components, such as polarizers, filters, and optical fiber connectors. Laboratory spectrometers are complex and robust devices capable of versatile and sensitive measurements. However, for many applications, and for the convenience and mobility of measurements fluorimeters can be designed as miniaturized devices. These devices are particularly suitable for clinical investigations of tissue fluorescence [8].

## **3. Breast tissue fluorescence**

emission paths to/from sample, (3) sample holder/positioner, (4) polarizers, and (5) detector

266 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

Polychromatic light from light source is dispersed on a dispersing element from which light beam of selected wavelength is directed on a sample to excite fluorescence. Most frequently, Xe arc lamps are used. These days, deuterium, tungsten, or halogen lamps are rarely utilized for excitation, since they are relatively weak sources for fluorescence. Lasers, laser diodes, and light-emitting diodes (LEDs) are also frequently used. They produce intense monochromatic light, so there is no need for the wavelength-selecting element on the excitation path. The drawback is that excitation spectra cannot be measured with these light sources. Tuneable and supercontinuum lasers are an exception, which can produce emissions over the given spectral interval, but these are rather expensive devices. Light dispersion can be accomplished with prisms and diffraction gratings; the latter is dominantly used in the modern spectrofluorometers. The detector can be either single-channeled or multi-channeled. The single-channeled detector, usually photomultiplier tube (PMT) or semiconductor, detects the intensity of one wavelength at a time. Multi-channeled detectors, such as charge-coupled device cameras (CCDs), record the intensity of emission over the range of wavelengths simultaneously. In addition, modern instruments include some other important components, such as polarizers, filters, and optical fiber connectors. Laboratory spectrometers are complex and robust devices capable of versatile and sensitive measurements. However, for many applications, and for the convenience and mobility of measurements fluorimeters can be designed as miniaturized devices. These devices are particularly suitable for clinical investigations of tissue fluorescence

(**Figure 2**).

[8].

**Figure 2.** Five basic components of spectrofluorometer.

Native fluorescence (autofluorescence) of breast tissue comes from a number of fluorescent biochemical species (fluorophores) whose excitations and emissions are strongly influenced by the absorption and scattering of tissue components. Since the latter processes are both depth- and wavelength-dependent, the observed tissue fluorescence is subject not only to the concentration and microenvironment of fluorophores, absorbers (chromophores), and scatters but also it substantially depends on the geometry of excitation and observation.

The breast, or mammary gland, is an organ composed of four major structures: skin, subcutaneous tissue, breast tissue, and the nipple centered on the round pigmented skin area (areola). In the breast tissue, distinct optical characteristics show glandular, adipose (fat), and fibrous tissue. The differences in autofluorescence of these tissues come from the structural and compositional differences of which the concentration of fluorophores and their distribution has the largest influence. Most fluorophores are associated with the structural matrix of tissues, such as collagen and elastin, or are involved in cellular metabolic processes such as reduced nicotinamide adenine dinucleotide (NADH) and flavins. Other fluorophores include the aromatic amino acids (e.g., tryptophan, tyrosine, phenylalanine), various porphyrins, and lipopigments (e.g., ceroids, lipofuscin) that are the end products of lipid metabolism [9]. Absorption and fluorescence emission spectra of the most important tissue fluorophores are shown in **Figure 3**. The excitation (*λ*ex), emission (*λ*em), and absorption (*λ*ab) maxima for some important tissue fluorophores are given in **Table 3**.

**Figure 3.** Absorption (A) and fluorescence emission (B) of the most important fluorophores present in tissue. (Adapted from Wagnières et al. [9] with permission of John Wiley and Sons).

#### 268 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences


**Table 3.** Tissue fluorophores and approximate values of their excitation (*λ*ex), emission (*λ*em), and absorption (*λ*ab) maxima.

Regarding breast cancer, it has been shown that distinct fluorescence response of tumors compared to one ordinary tissue is a result of notable differences in concentrations of collagen, elastin, NADH, and flavin adenine dinucleotide (FAD) [10]. Generally, fluorescence measurements have indicated lower concentrations of collagen and FAD and increased concentrations of NAD(P)H in malignant tissues compared to normal breast tissue [11]. Transformation from normal to malignant tissue leads to degradation and changes in the cross-links of collagen; breaking the cross-links in collagen is a consequence of the increased presence of collagenase in the tumor cells [12]. Modulation of the extracellular matrix is a common characteristic of invading tumor cells and usually involves increased production of collagenases by the tumor cells or stromal fibroblasts [13]. It has also been shown that the changes in the metabolic status of tissue have influence on the variation in concentrations of both NADPH and NADH. A shift from aerobic to anaerobic metabolism accompanied with damaged mitochondrial metabolism caused by malignant alterations leads to an increased concentrations of electron carriers such as NADH. Increased levels of these coenzymes have been observed in a high-grade malignant tissue [12, 14].

## **4. Interpretation of tissue fluorescence data**

**Endogenous fluorophore λex (nm) λem(nm) λab(nm)**

Tryptophan 280 350 280 Tyrosine 275 300 275 Phenylalanine 260 280 257

268 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

Collagen 325 400,405 325 Elastin 290,325 340,400 325

FAD, flavin 450 535 450 NADH 290,351 440,460 260 NADPH 336 464 340

Vitamin A 327 510 300 - 350 Vitamin K 335 480 249 Vitamin D 390 480 280

Pyridoxine 332,340 400 291 Pyridoxamine 335 400 - Pyridoxal 330 385 - Pyridoxal 5'-phosphate 330 400 - 4-Pyridoxic acid 315 425 307 Vitamin B12 275 305 361

Phospholipid 436 540,560 234 Lipofuscin 340-395 540,430-460 335, 435

Ceroid 340-395 430-460,540 - *Porphyrin* 400-450 630,690 408

**Table 3.** Tissue fluorophores and approximate values of their excitation (*λ*ex), emission (*λ*em), and absorption (*λ*ab)

Regarding breast cancer, it has been shown that distinct fluorescence response of tumors compared to one ordinary tissue is a result of notable differences in concentrations of collagen, elastin, NADH, and flavin adenine dinucleotide (FAD) [10]. Generally, fluorescence measurements have indicated lower concentrations of collagen and FAD and increased concentrations of NAD(P)H in malignant tissues compared to normal breast tissue [11]. Transformation from

*Amino acids*

*Structure proteins*

*Coenzymes*

*Vitamins*

*Lipids*

maxima.

*Vitamin B6 complex*

As in all fluorescence measurements, acquiring experimental data on tissue fluorescence is associated with procedural and instrumental/technical difficulties. One should note that unlike some other types of spectroscopy measurements, for example, absorption measurements, the intensity of recorded fluorescence signal and the shape of spectra strongly depend on the characteristics and settings of the instrument. One can see later in the text that establishing diagnosis from fluorescence measurement can be successfully done only if data were acquired from a statistically significant number of samples. With this in mind, each study related to tissue fluorescence should be done on a single instrument and with exactly the same instrument settings (such as PMT voltage and monochromator slit widths) for each sample. Even so, if measurements are done over the longer times some of instrument characteristics may change, such as the strength of the excitation lamp. Samples most frequently have different surface morphologies and can be differently or partly illuminated when they are small sized. All of this may produce unwanted artificial differences in fluorescence spectra, commonly in the intensity of observed signal. Normalization of measured spectra may resolve this problem; however, there is no universal method or procedure for this task. For *ex vivo* measurements, it is important to note that all excised tissues change over the time. The fluorescence measurements on different tissue samples should be done, in principle, at the similar time interval after excision.

It is also important to keep in mind distinction between technical (uncorrected for instrument characteristics) and molecular (corrected) spectrum. Modern instruments commonly produce both types of spectra. In most cases, technical emission spectra are sufficient for the analysis and presentation. However, when it is important to compare emission spectra obtained on different instruments, corrected spectra must be used. Excitation spectra, and conversely SFS, EEMs, and 3D-TSFS, are more instrument-dependent, being largely influenced with spectral characteristic of excitation sources. These spectra are, therefore, analyzed and presented after corrections. In addition, spectra may contain features that originate from physical processes apart from fluorescence, such as Rayleigh and Raman scatters, which should be excluded from subsequent analyses.

Fluorescence spectra of tissues are generally poorly resolved since they are composed from a broad emission of many fluorophores and affected by strong re-absorption processes. When needed, better-resolved spectra can be achieved with synchronous scanning.

## **5. Mathematical tools for fluorescence-based diagnosis of breast cancer**

To draw conclusions from experimental data of tissue fluorescence measurements, these data ought to be interpreted using adequate mathematical methods, most frequently statistical methods. Even if one analyze fluorescence from a single type of tissue, one must take into account that tissue characteristics considerably differ between individuals, statistically speaking they exhibit "a large in-group variances". For this reason, proper analysis is achieved only if it is done on a statistically significant number of samples and measurements. The differences in fluorescence, for example, differences in wavelengths of emission maxima, integrated emission intensities over some spectral range, excited state lifetimes, etc., between groups of different-type tissues can be asserted by methods of exploratory data analysis, one of which is hypothesis testing. One-tailed and two-tailed *t*-tests are commonly used to find if there is a statistically significant difference between the mean values of analyzed parameters in two data groups. Analysis of variance (ANOVA) and Tukey's test are used when more than two groups of data should be compared and studied. ANOVA is considered the most used and most useful statistical technique in biomedical research, even though this method is not easy to learn and should be implemented with caution.

Frequently, data from measurements of tissue fluorescence are vast, especially in cases of EEM and 3D-TSFS measurements. The size of data can be significantly reduced without loss of variance using some of statistical tools available for reducing the dimensionality. Principal component analysis (PCA) is a common tool for this purpose, and it is also capable of revealing hidden structures in a data set. Generally, PCA is used to recognize the main variations in a data set in an unsupervised way. It is also able to study these variations and to aid in visualizing them. PCA transforms and compresses input variables, which may be correlated, into uncorrelated variables of fewer numbers in such a way that they preserve the majority of variance of the input data. These new variables are called principal components (PCs) and they are obtained by a linear orthogonal transformation of input variables, so that each PC is, in fact, the linear combination of inputs. The largest portion of the input data variance is accumulated in the first PC, then less variance in the second PC, and so on. Then, for further analyses, one can use just few PCs, which help to alleviate numerical problems associated with large data sets and enable efficient visualization and interpretation of data.

Many features of fluorescence differ between healthy and cancer tissues; the selected examples are shown and discussed in the following section. Even though fluorescence measurements contain information needed for the cancer diagnosis, the observed data are subtly related in ways that are often difficult to express in the form of diagnostic rules just by observing spectra and must be processed for tissue classification purposes. To do so, mathematical algorithms ought to be developed and optimized to classify tissues into their respective histological categories. Several methods have been successfully used for this purpose.

Linear discriminant analysis (LDA) can be used for the discrimination of two or more groups from one or more linear functions (latent variables) of input data. As a consequence of singularity problems (caused by fewer samples in group than variables or highly correlated variables), LDA is incapable of solving high-dimensional data problems. In such cases, reduction of data dimensionality is required for the successful application of LDA.

Fluorescence spectra of tissues are generally poorly resolved since they are composed from a broad emission of many fluorophores and affected by strong re-absorption processes. When

**5. Mathematical tools for fluorescence-based diagnosis of breast cancer**

To draw conclusions from experimental data of tissue fluorescence measurements, these data ought to be interpreted using adequate mathematical methods, most frequently statistical methods. Even if one analyze fluorescence from a single type of tissue, one must take into account that tissue characteristics considerably differ between individuals, statistically speaking they exhibit "a large in-group variances". For this reason, proper analysis is achieved only if it is done on a statistically significant number of samples and measurements. The differences in fluorescence, for example, differences in wavelengths of emission maxima, integrated emission intensities over some spectral range, excited state lifetimes, etc., between groups of different-type tissues can be asserted by methods of exploratory data analysis, one of which is hypothesis testing. One-tailed and two-tailed *t*-tests are commonly used to find if there is a statistically significant difference between the mean values of analyzed parameters in two data groups. Analysis of variance (ANOVA) and Tukey's test are used when more than two groups of data should be compared and studied. ANOVA is considered the most used and most useful statistical technique in biomedical research, even though this method is not

Frequently, data from measurements of tissue fluorescence are vast, especially in cases of EEM and 3D-TSFS measurements. The size of data can be significantly reduced without loss of variance using some of statistical tools available for reducing the dimensionality. Principal component analysis (PCA) is a common tool for this purpose, and it is also capable of revealing hidden structures in a data set. Generally, PCA is used to recognize the main variations in a data set in an unsupervised way. It is also able to study these variations and to aid in visualizing them. PCA transforms and compresses input variables, which may be correlated, into uncorrelated variables of fewer numbers in such a way that they preserve the majority of variance of the input data. These new variables are called principal components (PCs) and they are obtained by a linear orthogonal transformation of input variables, so that each PC is, in fact, the linear combination of inputs. The largest portion of the input data variance is accumulated in the first PC, then less variance in the second PC, and so on. Then, for further analyses, one can use just few PCs, which help to alleviate numerical problems associated with large data

Many features of fluorescence differ between healthy and cancer tissues; the selected examples are shown and discussed in the following section. Even though fluorescence measurements contain information needed for the cancer diagnosis, the observed data are subtly related in ways that are often difficult to express in the form of diagnostic rules just by observing spectra and must be processed for tissue classification purposes. To do so, mathematical algorithms ought to be developed and optimized to classify tissues into their respective histological

needed, better-resolved spectra can be achieved with synchronous scanning.

270 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

easy to learn and should be implemented with caution.

sets and enable efficient visualization and interpretation of data.

categories. Several methods have been successfully used for this purpose.

Partial least squares (PLSs) discriminant analysis (DA) is a method based on the PLS regression, which constructs linear discriminant models with no restriction for processing high-dimensional data. Input data are transformed by the PLS-DA into a set of linear components (latent variables) further used to predict the dependent variable. The dependent variable is in fact a dummy variable that only serves to show whether a particular sample belongs to the specific class. With the LDA model, one is able to perform classification of data, that is, to predict the class to which new, unknown samples belong.

Artificial neural network (ANN) is a model used to approximate unspecified relations between a large number of input data in a robust manner. ANN learns from data, so it can be regarded as a machine-learning method and is able to handle a variety of problems including complex ones such as those involved in medical diagnostics. For example, ANNs can perform data classifications in both a supervised and unsupervised way, that is, both with prior knowledge of the data class and without. Unlike human decisions, those made by ANNs are invariably consistent since they are not liable to suffer from fatigue or bias. Kohonen's self-organizing map (SOM) and feed-forward neural network (FFNN) are among the most popular neural network architectures. SOM converts high-dimensional nonlinear statistical relationships into simple geometric relationships in an unsupervised way. On the other hand, FFNN uses a supervised training method which requires data on sample's group membership entered along with other data at the network inputs.

Support vector machine (SVM) is a statistical technology developed by the machine-learning community that can be used for both classification and regression. Compared with other machine-learning methods, SVM has such advantages as it does not require a large number of training samples for developing model and is not affected by the presence of outliers. Having high generalization procedure and feasibility to extract higher order statistics, the SVM has become extremely popular in terms of classification and prediction. In the context of classification, SVM is transforming original data space into a much higher dimension space in which classification groups would be linearly separable. For the data that belong to one of two classes (binary classification), SVM aims to derive the hyperplane in the transformed space so that the data of one class are on one side of the plane, and the data of other class are lying on the other side of the plane. The position of the hyperplane should be such that the greatest possible fraction of data is correctly placed and that the distance of both classes from the hyperplane is maximal. Such conditions minimize the risk of misclassifying data.

The performance of binary classification, the most probable classification case in breast cancer diagnoses, can be evaluated from a receiver operating characteristic (ROC) curves. The ROC curve is in fact graphic that shows the performance of a classification at different discrimination threshold values. It helps in finding how good classification model discriminates samples belonging to the particular group from all other samples. To construct the ROC, sensitivity and specificity of the model are calculated for different threshold values and plotted. Generally, an excellent model has an area under the ROC curve between 0.9 and 1.

Some other mathematical tools may also be useful for the study of breast cancer fluorescence. We believe that it is important to mention parallel factor analysis (PARAFAC) since this method is capable of modeling florescence response of complex fluorescence systems, the category in which all biological systems fall. Nowadays, PARAFAC has commonly used with EEMs in the analysis of fluorescence in many fields since it is a multi-way decomposition method capable of analyzing complex high-dimensional data and perform second-order calibration. Its main advantages are the uniqueness and simplicity of its solutions. When the correct number of components and multilinear data are used to build the PARAFAC model, the true underlying phenomenon is revealed. An additional advantage of this form of analysis is that it can predict the concentrations of different chemical compounds in complex systems, a phenomenon known as second-order advantage. For these reasons, PARAFAC is used to detect fluorophores in multi-component systems and precisely calculate their concentrations.

## **6. Examples of applications of fluorescence spectroscopy in breast cancer research**

So far, different types of fluorescence measurements have been employed for the breast cancer diagnosis and they are listed in **Table 4** along with comments on the used instrumentation and obtained sensitivity and specificity of detection.



Note: Ex – excitation, Em – emission, Δ*λ* – synchronous interval, SE – sensitivity, SP – specificity.

**Table 4.** Fluorescence methods used for breast cancer diagnosis.

specificity of the model are calculated for different threshold values and plotted. Generally, an

Some other mathematical tools may also be useful for the study of breast cancer fluorescence. We believe that it is important to mention parallel factor analysis (PARAFAC) since this method is capable of modeling florescence response of complex fluorescence systems, the category in which all biological systems fall. Nowadays, PARAFAC has commonly used with EEMs in the analysis of fluorescence in many fields since it is a multi-way decomposition method capable of analyzing complex high-dimensional data and perform second-order calibration. Its main advantages are the uniqueness and simplicity of its solutions. When the correct number of components and multilinear data are used to build the PARAFAC model, the true underlying phenomenon is revealed. An additional advantage of this form of analysis is that it can predict the concentrations of different chemical compounds in complex systems, a phenomenon known as second-order advantage. For these reasons, PARAFAC is used to detect fluorophores

**6. Examples of applications of fluorescence spectroscopy in breast cancer**

So far, different types of fluorescence measurements have been employed for the breast cancer diagnosis and they are listed in **Table 4** along with comments on the used instrumentation and

**(nm)**

Em 460-700

Em 360-560

Em 340

Ex 480-700

Ex 300-460 Em 310-600

Em 275-700

**SE, SP (%)**

99.6, 99.6

100, 100

70, 91.7


**Comment Ref.**

*Ex vivo* [16]

*Ex vivo* [18]

*Ex vivo* [19]

[20]


> 90, - *Ex vivo* [17]

excellent model has an area under the ROC curve between 0.9 and 1.

272 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

in multi-component systems and precisely calculate their concentrations.

**Measurement type Instrumentation specifics Measurement setup**

Emission spectra Argon laser Ex 457.9, 488

Emission spectra N2 laser Ex 337

Emission spectra N2 laser Ex 458

EEM lamp-based spectrophotometer with fiber optics probe

Excitation spectra lamp-based spectrophotometer Ex 250-320

EEM lamp-based spectrophotometer Ex 260-540

obtained sensitivity and specificity of detection.

**research**

#### **6.1. Breast cancer diagnosis using emission and excitation spectra**

The first report on the characterization of human breast tissues by fluorescence emission spectroscopy has been published by Alfano et al. [15] They have obtained normal and cancerous breast tissues from two different individuals and have measured emission spectra (460–700 nm) after excitation with an argon ion laser (at 488 and 457.9 nm). However, the assessment of method's capability for diagnosis has been impossible because of small number of investigated samples. Gupta et al. [16] have measured N2 laser-excited (337 nm) emission spectra (360–560 nm) from different sites on benign (fibroadenomas, 35 patients), cancerous (ductal carcinomas, 28 patients), and normal specimens (the uninvolved areas of the resected cancerous specimens). Intensities of emission were much higher from cancerous sites compared to those of benign tumor and normal breast tissue sites (**Figure 4**). In this *ex vivo* study, cancer tissues have been discriminated from benign and normal ones with a sensitivity and specificity of up to 99.6% using absolute intensity of emission (with aid of stepwise multivariate linear regression statistical method).

**Figure 4.** Mean emission spectra from 245 cancerous (C), 230 normal (N), and 436 benign (B) breast tissue sites. (Adapted from Gupta et al. [16] with permission of John Wiley and Sons.)

Yang et al. [17] have reported the use of excitation spectra for breast cancer detection. They have measured excitation spectra in the 250–320-nm spectral region recording emission at 340 nm from 103 malignant and 63 benign breast tissues. The averaged spectra and their difference (**Figure 5**) distinct spectral features can be clearly evidenced. Authors have estimated above 90% sensitivity and specificity of cancer diagnosis for this method.

**Figure 5.** Averaged excitation spectra for emission at 340 nm from 103 malignant and 63 breast tissues. D is the difference spectrum (M – B). B is benign; M is malignant. (Adapted from Yang et al. [17] with permission of John Wiley and Sons.)

#### **6.2. Breast cancer diagnosis using EEMs**

Considering that EEMs incorporate more information pertinent for the breast cancer diagnosis than the single emission or excitation spectra, several studies have been conducted using this method (see **Table 4**). Palmer et al. [19] have measured EEMs (excitation range: 300–460 nm; emission range: 310–600 nm) of 20 malignant, 15 normal/benign fibrous, and 21 adipose tissue samples. They have found four peaks in the spectra of malignant and normal/benign fibrous tissues which occurs in similar locations (excitation/emission pairs of (300, 340), (340, 390), (360, 460), and (440, 520) nm). The spectra of adipose tissue were distinctively different; particularly, the peak at (340, 390) nm was weakly present and the peak at (360, 460) nm has been shifted to approximately (360, 520) nm. Using PCA as a data reduction technique and SVM for classification, authors have showed that EEM was successful in discriminating malignant and nonmalignant tissues with a sensitivity and specificity of 70% and 92%, respectively.

#### **6.3. Breast cancer diagnosis using SFS and 3D-TSFS**

**Figure 4.** Mean emission spectra from 245 cancerous (C), 230 normal (N), and 436 benign (B) breast tissue sites. (Adapt-

Yang et al. [17] have reported the use of excitation spectra for breast cancer detection. They have measured excitation spectra in the 250–320-nm spectral region recording emission at 340 nm from 103 malignant and 63 benign breast tissues. The averaged spectra and their difference (**Figure 5**) distinct spectral features can be clearly evidenced. Authors have estimated above

**Figure 5.** Averaged excitation spectra for emission at 340 nm from 103 malignant and 63 breast tissues. D is the difference spectrum (M – B). B is benign; M is malignant. (Adapted from Yang et al. [17] with permission of John Wiley and

Considering that EEMs incorporate more information pertinent for the breast cancer diagnosis than the single emission or excitation spectra, several studies have been conducted using this method (see **Table 4**). Palmer et al. [19] have measured EEMs (excitation range: 300–460 nm; emission range: 310–600 nm) of 20 malignant, 15 normal/benign fibrous, and 21 adipose tissue samples. They have found four peaks in the spectra of malignant and normal/benign fibrous tissues which occurs in similar locations (excitation/emission pairs of (300, 340), (340, 390), (360,

ed from Gupta et al. [16] with permission of John Wiley and Sons.)

Sons.)

**6.2. Breast cancer diagnosis using EEMs**

90% sensitivity and specificity of cancer diagnosis for this method.

274 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

Synchronous spectrum, in majority of cases, has more features and provides more information than the conventional emission or excitation spectra. This comes as a result of simultaneous acquisition of both excitation and emission spectral characteristics of the sample, which brings more information to a single spectrum. It has been shown that SFS measurements have better selectivity and decreased bandwidths than emission or excitation spectra. For these reasons, they can be used to detect fine differences in the fluorescence of complex systems. Dramićanin et al. [22] have measured SFS with different synchronous intervals on 21 normal and 21 malignant breast tissue samples. The largest differences between spectra have been found for synchronous intervals of 30 and 80 nm, **Figure 6**. Diagnostic criteria have been established from the areas below different spectral peaks and from the ratios of these areas. The first

**Figure 6.** Synchronous fluorescence spectra taken with (a) 30 nm and (b) 80 nm synchronous intervals; the average spectra of 21 normal breast samples and 21 malignant are represented with the red and blue lines, respectively; dashed lines stand for the raw spectra and full lines for those additionally normalized. (Adapted from Dramićanin et al. [22].)

criterion yielded a sensitivity of detection of 77.4% and a specificity of 86.1%, while the second one presented a sensitivity of 90.3% and a specificity of 94.4%.

3D-TSFS and first derivate 3D-TSFS spectra have been measured on the same samples [10], and showed significant differences between normal and malignant tissues. Based on these differences, an artificial neural network (SOM) diagnosis method has been established [26], which provided a sensitivity of 87.1% and a specificity of 91.7% when using 3D-TSFS data, and 100% sensitivity and 94.4% specificity when using first derivate 3D-TSFS data. Diagnosis based on the use of support vector machine algorithm provided 100% specificity and sensitivity for both 3D-TSFS and first derivate 3D-TSFS [27].

#### **6.4. Breast cancer diagnosis using lifetime and polarization measurements**

In contrast to emission or excitation measurements, excited state lifetime measurements are not affected by the morphology of specimen and geometry of the measurements. Therefore, no normalization procedures are needed to compare results obtained with different instruments, which is quite important for the applicability of derived diagnostic methods. Sharma et al. [25] have measured fluorescence lifetimes at multiple emission wavelengths (532, 562, 632, and 644 nm) under excitation at 447 nm on 93 sites from 6 specimens (34 infiltrated ductal carcinoma, 31 benign fibrous, and 28 adipose sites). They have found that lifetime values measured at 532 and 562 nm significantly differ between normal and malignant sites, and achieved 92.3% accuracy for classifying infiltrated ductal carcinoma.

Choi et al. [23] have performed analysis of parallel and perpendicularly polarized fluorescence from 36 normal and 36 cancerous tissue samples. First, they have performed random forest algorithm to the four data sets: the fluorescence intensity and the curvature features for parallel and perpendicular components. Then, SVM classifiers were designed using the significant features from the random forest algorithm, which provided >90% specificity and sensitivity of breast cancer detection.

## **7. Conclusion**

To conclude, fluorescence is sensitive to the biochemical makeup of tissue. There are several natural fluorophores that exist in tissues and cells that, when excited with ultraviolet and visible light, fluoresce over well-defined spectral regions. Among them, the fluorescence of collagen, elastin, NADH, and FAD contribute the most to the different fluorescence of cancerous tissue in respect to the normal because of altered concentrations and local environment of fluorophores. Distinct differences in the fluorescence of cancerous and normal tissues can be observed with almost all conventional fluorescence techniques. For this purpose, up to now, emission spectroscopy and EEMs have been used most extensively. However, even though fluorescence differences between tissues are obvious, the development of a sensitive diagnostic method based on these differences is not an easy task. First, a large number of specimens must be measured under identical conditions. The specimens should be taken from different individuals and some gold diagnostic standard must be provided (e.g., histopathology). Fluorescence data ought to be processed with statistical tools that include analysis of variance, data reduction, and regression to obtain diagnostic criteria, and further with statistical tools to validate diagnostic results. On the other hand, once established, fluorescence methods are a great aid for early cancer detection, since methods can be noninvasive and do not require expensive and sophisticated equipment.

## **Acknowledgements**

criterion yielded a sensitivity of detection of 77.4% and a specificity of 86.1%, while the second

3D-TSFS and first derivate 3D-TSFS spectra have been measured on the same samples [10], and showed significant differences between normal and malignant tissues. Based on these differences, an artificial neural network (SOM) diagnosis method has been established [26], which provided a sensitivity of 87.1% and a specificity of 91.7% when using 3D-TSFS data, and 100% sensitivity and 94.4% specificity when using first derivate 3D-TSFS data. Diagnosis based on the use of support vector machine algorithm provided 100% specificity and sensitivity for

In contrast to emission or excitation measurements, excited state lifetime measurements are not affected by the morphology of specimen and geometry of the measurements. Therefore, no normalization procedures are needed to compare results obtained with different instruments, which is quite important for the applicability of derived diagnostic methods. Sharma et al. [25] have measured fluorescence lifetimes at multiple emission wavelengths (532, 562, 632, and 644 nm) under excitation at 447 nm on 93 sites from 6 specimens (34 infiltrated ductal carcinoma, 31 benign fibrous, and 28 adipose sites). They have found that lifetime values measured at 532 and 562 nm significantly differ between normal and malignant sites, and

Choi et al. [23] have performed analysis of parallel and perpendicularly polarized fluorescence from 36 normal and 36 cancerous tissue samples. First, they have performed random forest algorithm to the four data sets: the fluorescence intensity and the curvature features for parallel and perpendicular components. Then, SVM classifiers were designed using the significant features from the random forest algorithm, which provided >90% specificity and sensitivity of

To conclude, fluorescence is sensitive to the biochemical makeup of tissue. There are several natural fluorophores that exist in tissues and cells that, when excited with ultraviolet and visible light, fluoresce over well-defined spectral regions. Among them, the fluorescence of collagen, elastin, NADH, and FAD contribute the most to the different fluorescence of cancerous tissue in respect to the normal because of altered concentrations and local environment of fluorophores. Distinct differences in the fluorescence of cancerous and normal tissues can be observed with almost all conventional fluorescence techniques. For this purpose, up to now, emission spectroscopy and EEMs have been used most extensively. However, even though fluorescence differences between tissues are obvious, the development of a sensitive diagnostic method based on these differences is not an easy task. First, a large number of specimens must be measured under identical conditions. The specimens should be taken from different individuals and some gold diagnostic standard must be provided (e.g., histopathol-

one presented a sensitivity of 90.3% and a specificity of 94.4%.

276 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

**6.4. Breast cancer diagnosis using lifetime and polarization measurements**

achieved 92.3% accuracy for classifying infiltrated ductal carcinoma.

both 3D-TSFS and first derivate 3D-TSFS [27].

breast cancer detection.

**7. Conclusion**

The authors acknowledge the financial support of the Ministry of Education, Science and Technological Development of the Republic of Serbia (Project no 45020).

## **Author details**

Tatjana Dramićanin and Miroslav Dramićanin\*

\*Address all correspondence to: dramican@vinca.rs

Vinča Institute of Nuclear Sciences, University of Belgrade, Belgrade, Serbia

## **References**


fluorescence spectroscopy. Journal of Biophotonics. 2009;1–12. DOI: 10.1002/jbio. 200900070


breast cancer. IEEE Transaction on Biomedical Engineering. 2003;50:1233–1242. DOI: 10.1109/TBME.2003.818488

[20] Palmer GM, Keely PJ, Breslin TM, Ramanujam N. Autofluorescence spectroscopy of normal and malignant human breast cell lines. Photochemistry and Photobiology. 2003;78:462–469. DOI: 10.1562/0031-8655(2003)0780462ASONAM2.0.CO2

fluorescence spectroscopy. Journal of Biophotonics. 2009;1–12. DOI: 10.1002/jbio.

[9] Wagnières GA, Star WM, Wilson BC. In Vivo fluorescence spectroscopy and imaging for oncological applications. Photochemistry and Photobiology. 1998;68:603–632. DOI:

278 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

[10] Dramićanin T, Dramićanin MD, Jokanović V, Nikolić-Vukosavljević D, Dimitrijević B. 3-dimensonal total synchronous luminescence spectroscopy criteria for discrimination between normal and malignant breast tissues. Photochemistry and Photobiology.

[11] Dramićanin T, Lenhardt L, Zeković I, Dramićanin MD. Biophysical characterization of human breast tissues by photoluminescence excitation-emission spectroscopy. Journal

[12] Georgakoudi I, Jacobson BC, Muller MG, Sheets EE, Badizadegan K, Carr-Locke DL. NAD(P)H and collagen as in vivo quantitative fluorescent biomarkers of epithelial

[13] Fenhalls G, Geyp M, Dent MD, Parker MI. Breast tumour cell-induced down-regulation of type I collagen mRNA in fibroblasts. British Journal of Cancer. 1999;81:1142–1149.

[14] Uppal A, Gupta PK. Measurement of NADH concentration in normal and malignant human tissues from breast and oral cavity. Biotechnology and Applied Biochemistry.

[15] Alfano RR, Tang GC, Pradhan A, Lam W, Choy DSJ, Orpher E. Fluorescence spectra from cancerous and normal human breast and lung tissues. IEEE Journal of Quantum

[16] Gupta PK, Majumder SK, Uppal A. Breast cancer diagnosis using nitrogen laser excited autofluorescence spectroscopy. Lasers in Surgery and Medicine. 1997;21:417–422. DOI:

[17] Yang Y, Katz A, Celmer EJ, Zurawska-Szezepaniak M, Alfano RR. Fundamental differences of excitation spectrum between malignant and benign breast tissues. Photochemistry and Photobiology. 1997;66:518–522. DOI: 10.1111/j.

[18] Hage R, Galhanone PR, Zângaro RA, Rodrigues KC, Pacheco MTT, Martin AA, Neto MM, Soares FA, da Cunha IW. Using the laser-induced fluorescence spectroscopy in the differentiation between normal and neoplastic human breast tissue. Lasers in

[19] Palmer GM, Zhu C, Breslin TM, Xu F, Gilchrist KW, Ramanujam N. Comparasion of multiexcitation fluorescence and diffuse reflectance spectroscopy for the diagnosis of

Electronics. 1987;23:1806–1811. DOI: 10.1109/JQE.1987.1073234

10.1002/(SICI)1096-9101(1997)21:5<417::AID-LSM2>3.0.CO;2-T

Medical Science. 2003;18:171–176. DOI 10.1007/s10103-003-0271-9

of Research in Physics. 2012;36:53–62. DOI: 10.2478/v10242-012-0013-z

precancerous changes. Cancer Research. 2002;62:682–687. DOI: NEMA

200900070

10.1111/j.1751-1097.1998.tb02521.x

DOI: 10.1038/sj.bjc.6690821

1751-1097.1997.tb03183.x

2003;37:45–50. DOI: 10.1042/BA20020052

2005;81:1554–1558. DOI: 10.1562/2005-02-15-RA-442


Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

#### **Applications of 1H Nuclear Magnetic Resonance Spectroscopy in Clinical Microbiology Applications of 1H Nuclear Magnetic Resonance Spectroscopy in Clinical Microbiology**

Lara García‐Álvarez, Jesús H. Busto, Jesús M. Peregrina, Alberto Avenoza and José Antonio Oteo Lara García‐Álvarez, Jesús H. Busto, Jesús M. Peregrina, Alberto Avenoza and José Antonio Oteo

Additional information is available at the end of the chapter Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/64450

#### **Abstract**

280 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

Proton nuclear magnetic resonance (1 H NMR) is a spectroscopic technique usually used for structural determination of molecules. In recent years, this technique has been employed for easy and quick recognition of microorganisms, in antimicrobial suscept‐ ibility tests and even for the diagnosis of different infectious conditions. Though 1 H NMR shows great potential for expanded applications in microbiological studies, to date applications of proton NMR to microbiological research are not totally standar‐ dized. In this chapter, we summarize the state of knowledge about 1 H NMR and its current and potential applications in this field.

**Keywords:** nuclear magnetic resonance, 1 H NMR, applications, clinical microbiology, microorganisms

## **1. Introduction**

Scientific progress made in the recent years has enabled the development of new techniques that facilitate and improve microbiological study. In this way, nuclear magnetic resonance (NMR) is a spectroscopic technique easy to use and quick to recognize microorganisms and provides sensitivity to antimicrobials. Anyway, to date we have not consensus about the usefulness of these techniques that are not totally standardized. In this chapter, we summarize the state of knowledge about NMR in microbiological studies.

© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

## **2. Nuclear magnetic resonance (NMR) spectroscopy**

NMR is a spectroscopic technique initially developed by Felix Bloch and Edward M. Purcell that relies on the magnetic properties of the atomic nucleus. Since 1946, it has become a powerful and extremely valuable tool for chemists, physicists, biochemists and more recently for the medical practitioners [1–5].

Although the most widespread application of NMR is the structural determination of mole‐ cules, the technique offers the advantage of direct mixture analysis, and therefore, NMR has demonstrated a unique potential to be used for metabolic mixture analysis, fastidious bacteria included [6]. In comparison with other techniques employed for mixture analysis, NMR can be used to directly investigate biological samples and cell cultures without requiring signifi‐ cant sample preparation. Moreover, the technique allows the determination of compound ratios in a highly reproducible manner. For these reasons, metabolomics and metabonomics are driving new technological advances in the NMR field. Thus, the combination of sophisti‐ cated NMR‐based methods for mixture analysis with the power of statistical and chemometric methods makes NMR spectroscopy the technique of choice for complex biological mixture analysis, especially in clinical and biomedical researches [3, 4].

In the biological field, the NMR technique [7] is employed to determine the structure and function of macromolecules. Additionally, NMR allows the determination of metabolic changes in organisms in response to external stimuli, through the identification and quantifi‐ cation of metabolites (metabolomics/metabonomics). Metabolomics is the study of global metabolite profiles of a cell under a given set of conditions; however, the terms metabolomics and metabonomics are used in the literature interchangeably [8]. Jeremy Nicholson was the pioneer of the approach, 'the distinction between the two terms metabolomics/metabonomics is mainly philosophical rather than technical' [9].

Since NMR is the only technique that allows to carry out *in vivo* analysis, it has applications in medical diagnosis, for example, magnetic resonance tomography (MRT).

The use of NMR in metabolic studies was reported in 1977, by Brown et al., who determined the presence of lactate, pyruvate, creatinine, and alanine in a blood cell suspension by proton NMR (1 H NMR) [10]. Since then, the analysis of biological fluids, tissues, and cell extracts has been carried out successfully by NMR, especially in the context of research on diseases and evaluation of toxicological processes [11, 12].

## **2.1. Basics of NMR spectroscopy**

The NMR phenomenon can only be carefully described and fully understood using quantum mechanics. Therefore, a complete understanding of the technique would require an exhaustive knowledge of the properties of the angular momentum in the quantum mechanics field, along with statistical thermodynamics knowledge to describe the floating processes needed in the liquid state.

However, since the theory and fundamentals of NMR have been fully developed over the last few years [13, 14] and its detailed description would get away from the objective of this chapter, we describe below some aspects of the technique, which could help to understand the equipment and methodologies used in metabolic research.

The basis of the technique is to use the magnetic properties of atomic nuclei that are defined by their angular momentum and their associated magnetic moment. Both moments are vector quantities, and they are related by a constant called gyromagnetic constant (γ), which is specific to each type of atomic nucleus. According to quantum mechanics, both moments are quantized and their value depends on a quantum number called spin (I). Not all nuclei are valid to get NMR signals, only those whose spins are greater than zero are valid, but 1 H or proton is the most used nucleus in NMR studies. Moreover, higher the value of constant γ, more sensitive will be the active nucleus in NMR. So, 1 H is the nucleus used in NMR studies because of its great abundance (100%) and high value of γ. However, other magnetically active nuclei with lower sensitivity may be used, such as 2 H (or D), 13C, 19F, 15N, 31P, and 23Na. Focusing on the proton, in the presence of an external magnetic field (B0), two different energy levels appear. The magnetic moments of these nuclei try to align with B0, resulting in two possible orienta‐ tions at that time. Each of these orientations corresponds to a different energy level. That is, in the presence of B0, the cores can be arranged in two new states with different energies. This phenomenon, from a vectorial viewpoint is known as magnetization.

To obtain a NMR signal, the sample is irradiated with a radiofrequency wave, perpendicular to B0, which compels it to reach the state of resonance, where the nuclei gyrate with a resonance frequency (f0), specific to each atomic nucleus and called Larmor frequency. For this reason, NMR spectrometers are designated by the 1 H resonance frequency instead of the magnetic field (for example, on a 14.1 T field, <sup>1</sup> H resonates at 600 MHz). After the pulse, the excited nuclei return to the initial equilibrium state emitting a radiofrequency signal, which decays with the time, a phenomenon known as relaxation. The resonance of the excited nuclear magnets is detected as an oscillating current in a coil placed around the sample. This signal is the FID (free induction decay), which arrives at the receiver and provides a spectrum formed by lines defining frequencies and widths by a mathematical operation known as Fourier transform. The widths are formed from the contributions of all nuclei of the sample, so that this quality allows quantitative measurements. A line in the NMR spectrum obtained at a certain frequency (or chemical shift) corresponds to an atomic nucleus with a given chemical environment, which allows structural information about the molecule it belongs.

## **2.2. Equipment**

**2. Nuclear magnetic resonance (NMR) spectroscopy**

282 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

analysis, especially in clinical and biomedical researches [3, 4].

is mainly philosophical rather than technical' [9].

evaluation of toxicological processes [11, 12].

**2.1. Basics of NMR spectroscopy**

NMR (1

liquid state.

for the medical practitioners [1–5].

NMR is a spectroscopic technique initially developed by Felix Bloch and Edward M. Purcell that relies on the magnetic properties of the atomic nucleus. Since 1946, it has become a powerful and extremely valuable tool for chemists, physicists, biochemists and more recently

Although the most widespread application of NMR is the structural determination of mole‐ cules, the technique offers the advantage of direct mixture analysis, and therefore, NMR has demonstrated a unique potential to be used for metabolic mixture analysis, fastidious bacteria included [6]. In comparison with other techniques employed for mixture analysis, NMR can be used to directly investigate biological samples and cell cultures without requiring signifi‐ cant sample preparation. Moreover, the technique allows the determination of compound ratios in a highly reproducible manner. For these reasons, metabolomics and metabonomics are driving new technological advances in the NMR field. Thus, the combination of sophisti‐ cated NMR‐based methods for mixture analysis with the power of statistical and chemometric methods makes NMR spectroscopy the technique of choice for complex biological mixture

In the biological field, the NMR technique [7] is employed to determine the structure and function of macromolecules. Additionally, NMR allows the determination of metabolic changes in organisms in response to external stimuli, through the identification and quantifi‐ cation of metabolites (metabolomics/metabonomics). Metabolomics is the study of global metabolite profiles of a cell under a given set of conditions; however, the terms metabolomics and metabonomics are used in the literature interchangeably [8]. Jeremy Nicholson was the pioneer of the approach, 'the distinction between the two terms metabolomics/metabonomics

Since NMR is the only technique that allows to carry out *in vivo* analysis, it has applications in

The use of NMR in metabolic studies was reported in 1977, by Brown et al., who determined the presence of lactate, pyruvate, creatinine, and alanine in a blood cell suspension by proton

been carried out successfully by NMR, especially in the context of research on diseases and

The NMR phenomenon can only be carefully described and fully understood using quantum mechanics. Therefore, a complete understanding of the technique would require an exhaustive knowledge of the properties of the angular momentum in the quantum mechanics field, along with statistical thermodynamics knowledge to describe the floating processes needed in the

However, since the theory and fundamentals of NMR have been fully developed over the last few years [13, 14] and its detailed description would get away from the objective of this chapter,

H NMR) [10]. Since then, the analysis of biological fluids, tissues, and cell extracts has

medical diagnosis, for example, magnetic resonance tomography (MRT).

The NMR spectrometer involves several parts such as a superconducting magnet, a radio transmitter, a probe, a radio receiver, an analog‐to‐digital converter (ADC), and finally, a computer. The magnet is the main element and consists of a solenoid of superconducting Nb/ Ti alloy wire immersed in liquid helium (4 K) that is charged to generate the essential field strength. The helium is protected with a vacuum jacket and further cooled by an outer dewar of liquid nitrogen (77 K).

The probe head is a coil of wire positioned around the sample (NMR tube) that alternately transmits and receives radio frequency signals. The probe head is usually hosted into the magnet from the bottom and is connected to at least three radiofrequency channels provid‐ ing the 2 H lock, 1 H frequency, and one X‐nucleus frequency. In general, devices to control temperature (heater, thermoelement, and air) are needed. New developments include the digital transmission of the probe‐head parameters to the console.

The computer addresses the transmitter to send a high‐power and very short duration pulse of radio frequency to the probe‐head. Instantly, after the pulse of radio frequency, a weak signal (FID) from the sample received by the probe‐head is amplified and sampled at intervals by the ADC to produce a digital FID signal, which is just a list of numbers. The computer automati‐ cally determines the timing and power of pulses output by the transmitter and receives and processes the digitalization. After the computer performs the mathematical processes of Fourier transform in order to convert time domain into frequency domain, the resulting spectrum can be displayed on the computer monitor, transferred to other computers or plotted on a paper.

#### **2.3. Sample preparation**

The sample volume should be about 0.6 mL, which gives a 4.0 cm depth in a standard 5 mm NMR tube. Volume of the samples is less important than the concentration of targets. Very small volumes could not be studied by NMR equipment with low magnetic field but once the volume is established, it is more important to ensure that the metabolites to be studied are in

**Figure 1.** Components of an NMR equipment.

an appropriate concentration. For biological samples, the ideal solvent is D2O or a mixture of H2O/D2O. In this latter case, the 1 H NMR spectra will be recorded using the pulse sequence for presaturation or the equivalent in order to avoid the signals from water. This is the typical and most simple method used to record NMR spectra of biological samples.

Usually, the protocol of preparation for biological and aqueous samples is simple, quick and involves two steps. Samples are prepared from inocula of the microorganisms and the bacterial concentration is adjusted using 0.5–2 McFarland standards [15, 16]. Cultures are incubated at the optimum temperature and time to preserve the same growth conditions than the reference method, if possible. After incubation, the suspensions are removed by centrifugation and the supernatants are decanted and used for NMR experiments. Because the measurements are carried out on supernatants, there is no need to quench cell growth rapidly. Furthermore, pH is measured and fixed by the slow addition of aqueous 1 M HCl and 1 M NaOH solutions or with a buffer solution in order to fix the chemical shifts. Next, a biological sample is added to a 5 mm NMR tube together with D2O with the addition of the sodium salt (trimethyl)propa‐ noic‐2,2,3,3‐d4 acid (TSP) for the chemical shift calibration. **Figure 1** shows an overview of the components of NMR equipment.

## **2.4. Sample treatment**

ing the 2

on a paper.

**2.3. Sample preparation**

**Figure 1.** Components of an NMR equipment.

H lock, 1

H frequency, and one X‐nucleus frequency. In general, devices to control

temperature (heater, thermoelement, and air) are needed. New developments include the

The computer addresses the transmitter to send a high‐power and very short duration pulse of radio frequency to the probe‐head. Instantly, after the pulse of radio frequency, a weak signal (FID) from the sample received by the probe‐head is amplified and sampled at intervals by the ADC to produce a digital FID signal, which is just a list of numbers. The computer automati‐ cally determines the timing and power of pulses output by the transmitter and receives and processes the digitalization. After the computer performs the mathematical processes of Fourier transform in order to convert time domain into frequency domain, the resulting spectrum can be displayed on the computer monitor, transferred to other computers or plotted

The sample volume should be about 0.6 mL, which gives a 4.0 cm depth in a standard 5 mm NMR tube. Volume of the samples is less important than the concentration of targets. Very small volumes could not be studied by NMR equipment with low magnetic field but once the volume is established, it is more important to ensure that the metabolites to be studied are in

digital transmission of the probe‐head parameters to the console.

284 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

The quality of the obtained NMR spectra depends on several variables that influence the process from sample collection to final data collection. The sample collection involves the sample, containers used, additives (preservatives, stabilizers), and time (collection, transport, storage) [17].

Depending of the source of the biological sample, two different methodologies can be used for the experiment acquisition:

(A) Sample concentration by lyophilization and subsequent reconstitution with a deuterated solvent: using this methodology, NMR experiments can be performed without solvent suppression, allowing an increase in sensitivity and stopping enzyme activity. The risks in using this method include the possibility of introducing contaminants into the sample and more importantly, the loss of volatile compounds.

(B) The addition of a small amount of D2O to the aqueous sample: the corresponding NMR experiment is performed with a pulse program that can remove the water signal, which would otherwise mask the signals from the rest of the sample. This method involves minimal sample handling and the ability to detect volatile compounds, making it a more suitable method for metabolite analysis of biological samples.

In addition to these processes, due to the infectious potential of microorganisms contained in the sample, all steps of sample processing must be performed safely, through protocols and in laboratories appropriate to the biosafety level for the organism—until the organism is inactivated.

#### **2.5. Data processing**

Interpretation of the NMR data is essential to complete metabonomics studies to draw conclusions and trends. The first thing is to perform a pre‐processing of NMR data in which NMR spectra are cleaned up in standard ways. After treatment of the spectra, it is possible to get information of metabolites, either through direct quantification by parameters [18, 19] or by applying the methods of data analysis and modelling. In this case, chemometric techniques and multivariate analysis are used to identify and quantify the different metabolites present in the sample. **Figure 2** shows an example of NMR spectrum with the main metabolites obtained. The existence of NMR databases of metabolites can greatly facilitate the latter processes.

**Figure 2.** <sup>1</sup> H NMR spectrum showing the main metabolites.

#### **3. Applications of 1 H NMR spectroscopy in clinical microbiology**

The application of 1 H NMR to living cells is used to determine metabolites in complex mixtures and has been widely used for identification and quantification of the bacterial species [15, 20]. This technique has also been applied for antimicrobial drug susceptibility studies on different species of yeast, and in the last few years, it has also been developed for bacterial studies. Furthermore, other determinations directly in body fluids have emerged to help in the diagnosis of different diseases and conditions.

#### **3.1. Bacterial identification and metabolic studies**

1 H NMR spectroscopy has been used for bacterial identification and quantification and for metabolic pathways studies. Several studies have been conducted for the diagnosis of the bacteria that cause urinary tract infections (UTI). These focus on the use of 1 H NMR spectro‐ scopy for the identification and quantification of common uropathogens such as *Pseudomonas* *aeruginosa*, *Klebsiella pneumoniae*, *Escherichia coli,* and *Proteus mirabilis* in urine samples. These studies are based on specific properties of the metabolism of the studied bacteria, and the results showed that 1 H NMR is a simple and fast tool compared with the traditional methods [16, 21, 22].

The qualitative and quantitative determination of *P. aeruginosa* using NMR spectroscopy is based on the specific property of the bacteria to metabolize nicotinic acid (NA) to 6‐hydroxy‐ nicotinic acid (6‐OHNA). Only this bacterium can produce this reaction. The addition of NA to urine samples after incubation and the subsequent analysis by 1 H NMR spectroscopy showed that NA signals disappeared from the medium after some time, while the appearance of new signals of the metabolite 6‐OHNA indicated the presence of *P. aeruginosa*. The increase in the intensity of the metabolite signals, together with the decrease in the NA signals, involved a proportional increase in the number of bacteria. This shows the potential offered by this technique for quantitative and qualitative identification, simultaneously, on the bacteria [21].

A similar process occurs in the determination of *K. pneumoniae*. In this case, the specific metabolic reaction is the transformation of glycerol into 1,3‐propanediol, so the substance that is added into the medium is glycerol. Despite *Citrobacter freundii* also being capable of carrying out this reaction, both bacteria can be easily differentiated using microscopic examination by observing their motility. *K. pneumoniae* is not mobile while *C. freundii* is. In addition, *C. freundii* is not a common nosocomial urinary tract infection agent. The combination of both methods showed very good sensitivity and specificity (90 and 100%, respectively), suggesting the potential usefulness of NMR for bacterial diagnosis [22].

The same experiment carried out on *E. coli* and *P. mirabilis* revealed that the specific metabolites of the bacteria are lactate and 2‐hydroxy‐4‐(methylthio)butyric acid (MOBA) after incubation with lactose and methionine, respectively [16].

The results obtained using this alternative technique provided the warrant for the development of this method in bacterial identification and quantification and the technical development with other microorganisms. With experience, spectral analysis and data interpretation could be quick and reliable [16].

## **3.2. Antimicrobial susceptibility assays**

**2.5. Data processing**

processes.

**Figure 2.** <sup>1</sup>

1

**3. Applications of 1**

The application of 1

H NMR spectrum showing the main metabolites.

diagnosis of different diseases and conditions.

**3.1. Bacterial identification and metabolic studies**

Interpretation of the NMR data is essential to complete metabonomics studies to draw conclusions and trends. The first thing is to perform a pre‐processing of NMR data in which NMR spectra are cleaned up in standard ways. After treatment of the spectra, it is possible to get information of metabolites, either through direct quantification by parameters [18, 19] or by applying the methods of data analysis and modelling. In this case, chemometric techniques and multivariate analysis are used to identify and quantify the different metabolites present in the sample. **Figure 2** shows an example of NMR spectrum with the main metabolites obtained. The existence of NMR databases of metabolites can greatly facilitate the latter

286 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

**H NMR spectroscopy in clinical microbiology**

and has been widely used for identification and quantification of the bacterial species [15, 20]. This technique has also been applied for antimicrobial drug susceptibility studies on different species of yeast, and in the last few years, it has also been developed for bacterial studies. Furthermore, other determinations directly in body fluids have emerged to help in the

H NMR spectroscopy has been used for bacterial identification and quantification and for metabolic pathways studies. Several studies have been conducted for the diagnosis of the

scopy for the identification and quantification of common uropathogens such as *Pseudomonas*

bacteria that cause urinary tract infections (UTI). These focus on the use of 1

H NMR to living cells is used to determine metabolites in complex mixtures

H NMR spectro‐

The use of 1 H NMR spectroscopy for antimicrobial susceptibly tests has been not highly studied despite its powerful utility in this area of study [5]. Application of 1 H NMR spectroscopy to antimicrobial susceptibility studies was first carried out on different species of yeast. The standardized methods currently available for fungal susceptibility studies are unreliable and relatively slow, so, 1 H NMR spectroscopy can be a simple indicator, an objective and fast method (metabolic changes detected by this method are more easily observed than growth inhibition in broth). 1 H NMR spectroscopy is potentially valuable in determining the metabolic composition of yeast suspensions incubated with a drug. In addition, it is a high performance automated method with low operating costs, so that both operator time and reagent cost are greatly reduced. Therefore, it has great potential to emerge as an alternative method for the antifungal drug susceptibility determination of different yeast species [15, 20].

One of the few studies in which the 1 H NMR technique has been applied to observe the effect on microorganisms upon exposure to several drugs, has been carried out with medically relevant fungi. The fungal species analysed were *Cryptococcus neoformans*, different species of *Candida* and *Aspergillus* spp. These studies are based on the identification of the fungal metabolites produced, on their comparative profile implementation and on the monitoring of the nutrient utilization of the incubation medium in the presence of certain drug concentrations (caspofungin, amphotericin B, and voriconazole). The spectra obtained after subjecting the sample to the 1 H NMR were interpreted based on the metabolites produced (fumarate, malate, ethanol, etc.) and/or the metabolites consumed (tyrosine, phenylalanine, valine, etc.). This interpretation established a measurable parameter, the metabolic end point (MEP), from the spectral peak area. The MEP is defined as the lowest drug concentration at which nutrient utilization from the medium or the production of fungal metabolites is inhibited ≥ 50% and compared with minimum inhibitory concentrations (MIC) used in the reference method. The results of MEP generally showed a good correlation with MICs, which were determined by a modification of the reference method in broth microdilution M27‐A of the CLSI. Discrepancies, which may arise between MEPs and MICs, could be due to differences in the culture medium and incubation time. In addition, 1 H NMR spectroscopy is a potentially valuable method for determining the metabolism of microorganisms incubated with the drug because it is a reproducible and relatively quick method (it takes 16 h versus 48 h required by the reference method) suggesting its consolidation as a platform for rapid determination of antifungal susceptibility [15, 20].

In reference to bacteria, there are very few studies based on their antimicrobial susceptibility according to metabolic profiling by 1 H NMR. One of these studies focused on a bacterial disease called 'Withering Syndrome' of abalone (a type of mollusc belonging to *Haliotis* spp. important in aquaculture). It is caused by a pathogen of the Rickettsiaceae family, '*Candidatus Xenohaliotis californiensis*' that infects digestive epithelial cells [23, 24]. On this basis, the effects of the antibiotic oxytetracycline in the metabolic profiles were observed by 1 H NMR. This drug, used to treat bacterial infections in aquatic species, reduces the severity and mortality of the Withering Syndrome. The aim of this study was to observe whether the recovery of the metabolism during treatment with oxytetracycline coincided with the disappearance of the disease caused by the bacteria. To this end, they examined the metabolic constituents present in the foot muscle of the mollusc after oxytetracycline treatment during several established days at two different seawater temperatures (13.4 and 17.3°C) [24]. Metabolic changes were observed at both temperatures: levels of taurine, glycine‐betaine, and homarine increased and the amino acid and carbohydrate levels decreased. The detection of metabolic differences between animals treated and untreated with antibiotics was observed only at the highest temperature. Therefore, oxytetracycline eradicated the infection and at the highest tempera‐ ture it reduced the metabolic changes due to the syndrome. The conclusions drawn from these experiments drive the development of 1 H NMR based on metabolic studies and its comple‐ mentarity with other techniques, such as the histology for the identification of pathological processes in the aquatic species and for the optimization of drug therapy. This tool displays the performance by analysing quickly and cheaply the functional status of an organism [23, 24]. We have analysed the metabolism and antimicrobial susceptibility of *Escherichia coli* ATCC 25922 in the presence of gentamicin using 1 H NMR and compared with a reference method [25]. The MIC, determined by the reference method used in this study, would correspond to the termination of the bacterial metabolism observed using NMR. To carry out these experi‐ ments, serial dilutions of gentamicin were tested. Furthermore, two controls were also analysed (one was the medium with an inoculum of bacterium (control I), and the other was only the medium (control II)). The comparison of the two control 1 H NMR spectra showed different signals. Succinic acid, acetic acid, and ethanol were only detected in the control I spectra and threonine was only detected in the control II medium. According to the results obtained by visual turbidity, the lowest concentration of drug that completely inhibited visible growth (MIC) was 0.5 μg/ml (MIC: 0.25–1 μg/ml) [26]. When we registered antibiotic spectra at different concentrations, we detected the presence of succinic acid, acetic acid, and ethanol only in samples with concentrations of gentamicin lower than the MIC. Moreover, when the concentration of gentamicin was greater than the MIC, we detected the presence of threonine. These data suggested that the results obtained by 1 H NMR spectroscopy were in agreement with those obtained by visual turbidity. These results confirm that *E. coli* is able to metabolize components of the medium to produce succinic acid, acetic acid, and ethanol. Furthermore, threonine only appeared in the spectra of those samples with gentamicin concentrations of ≥0.5 μg/ml. Differences in peak intensities for the metabolites observed in spectra allowed the determination of the MIC for gentamicin using 1 H NMR spectroscopy. Consumption of the amino acid threonine, present in the culture medium, was interrupted when MIC was performed. Therefore, we assume that succinic acid, acetic acid, and ethanol are metabolites produced by the bacteria and threonine is the amino acid consumed by *E. coli*.

Furthermore, to evaluate the potential of this tool, we also performed the same biological experiments but using an NMR tube as the incubation reactor. The bacterial activity occurred effectively within the NMR tube, and the metabolic process started around 3 h 20 min and ended at 6 h. Moreover, when samples containing gentamicin were analysed in the same way the ethanol signal appeared later using the lowest concentration of gentamicin (4 h 40 min) compared with the experiments performed in the absence of the antibiotic (3 h 20 min) and much later (8 h 40 min) when the gentamicin concentration used was close to MIC. Similarly, threonine consumption by bacteria was delayed when the concentration of antibiotic in the medium was higher [25].

#### **3.3. Applications in biofluids**

One of the few studies in which the 1

and incubation time. In addition, 1

according to metabolic profiling by 1

experiments drive the development of 1

susceptibility [15, 20].

sample to the 1

H NMR technique has been applied to observe the effect

H NMR spectroscopy is a potentially valuable method for

H NMR. One of these studies focused on a bacterial disease

H NMR based on metabolic studies and its comple‐

H NMR. This drug, used

on microorganisms upon exposure to several drugs, has been carried out with medically relevant fungi. The fungal species analysed were *Cryptococcus neoformans*, different species of *Candida* and *Aspergillus* spp. These studies are based on the identification of the fungal metabolites produced, on their comparative profile implementation and on the monitoring of the nutrient utilization of the incubation medium in the presence of certain drug concentrations (caspofungin, amphotericin B, and voriconazole). The spectra obtained after subjecting the

288 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

ethanol, etc.) and/or the metabolites consumed (tyrosine, phenylalanine, valine, etc.). This interpretation established a measurable parameter, the metabolic end point (MEP), from the spectral peak area. The MEP is defined as the lowest drug concentration at which nutrient utilization from the medium or the production of fungal metabolites is inhibited ≥ 50% and compared with minimum inhibitory concentrations (MIC) used in the reference method. The results of MEP generally showed a good correlation with MICs, which were determined by a modification of the reference method in broth microdilution M27‐A of the CLSI. Discrepancies, which may arise between MEPs and MICs, could be due to differences in the culture medium

determining the metabolism of microorganisms incubated with the drug because it is a reproducible and relatively quick method (it takes 16 h versus 48 h required by the reference method) suggesting its consolidation as a platform for rapid determination of antifungal

In reference to bacteria, there are very few studies based on their antimicrobial susceptibility

called 'Withering Syndrome' of abalone (a type of mollusc belonging to *Haliotis* spp. important in aquaculture). It is caused by a pathogen of the Rickettsiaceae family, '*Candidatus Xenohaliotis californiensis*' that infects digestive epithelial cells [23, 24]. On this basis, the effects of the

to treat bacterial infections in aquatic species, reduces the severity and mortality of the Withering Syndrome. The aim of this study was to observe whether the recovery of the metabolism during treatment with oxytetracycline coincided with the disappearance of the disease caused by the bacteria. To this end, they examined the metabolic constituents present in the foot muscle of the mollusc after oxytetracycline treatment during several established days at two different seawater temperatures (13.4 and 17.3°C) [24]. Metabolic changes were observed at both temperatures: levels of taurine, glycine‐betaine, and homarine increased and the amino acid and carbohydrate levels decreased. The detection of metabolic differences between animals treated and untreated with antibiotics was observed only at the highest temperature. Therefore, oxytetracycline eradicated the infection and at the highest tempera‐ ture it reduced the metabolic changes due to the syndrome. The conclusions drawn from these

mentarity with other techniques, such as the histology for the identification of pathological processes in the aquatic species and for the optimization of drug therapy. This tool displays the performance by analysing quickly and cheaply the functional status of an organism [23, 24].

antibiotic oxytetracycline in the metabolic profiles were observed by 1

H NMR were interpreted based on the metabolites produced (fumarate, malate,

In the last few years, 1 H NMR has been used to directly analyse biofluids and to diagnose different diseases directly from body fluids. In this sense, it has been applied to analyse human microbiota from faeces and urine samples, to study the metabolic implications that take place in sepsis, or even to diagnose hepatitis C virus infection, distinguish HIV‐1 positive patients from negative individuals or to diagnose pneumonia from urine [27–33].

As mentioned above, 1 H NMR has been employed to study gut microbiome focusing on the metabolite profiling obtained by the analysis of different human samples. In this sense, Jacobs et al. [27] studied faeces from healthy subjects after consuming placebo, grape juice, or a mix of grape juice and wine during four weeks by 1 H NMR. The comparison of the NMR profiling of the samples indicated that only the mixture of wine and grape juice was able to modulate gut microbiota assessed by the reduction in isobutyrate observed only in the samples from subjects who had consumed the mixture. These results confirm that 1 H NMR could determine the impact of the nutrition in human microbiome [27]. Furthermore, the use of 1 H NMR to understand gut microbiota has been also extended to the study its impact on obesity by analysing metabolite profiling obtained from urine samples [28, 34].

1 H NMR has been also used to study different conditions as sepsis. Stringer et al. used 1 H NMR to compare the metabolic profile of whole blood from serum. This study revealed that the use of 1 H NMR in whole blood allowed obtaining more metabolic details that serum, and, in some cases, the metabolites detected were in higher concentrations in whole blood than in serum. Furthermore, whole blood allowed the determination of metabolites associated with red blood cell metabolism and observed that alterations in their metabolism could be in relationship with sepsis due to the haemolysis they cause [29]. The same authors have carried out several experiments in the same way. In these experiments, they were able to observe that blood samples of sepsis‐induced acute lung injury patients were measurable and distinguishable from healthy blood due to differences in metabolites using 1 H NMR [35]. Other studies have been carried out to evaluate sepsis by 1 H NMR, but in rats [30]. Metabolic profiles obtain from the 1 H NMR analysis reveal changes in the metabolites involved in energy metabolism, especially in the serum of septic rats. From these results, authors concluded that according to the metabonomic approach, 1 H NMR has the potential for the early prognostic evaluation of sepsis [30].

As discussed above, 1 H NMR has been applied to the diagnosis of hepatitis C virus infection [31]. This study performed in urine samples was able to identify infected patients and negative individuals with good sensitivity and specificity using a metabonomics model based on the spectra obtained from the urine samples analysis. In this study, although the differences observed in the spectra allowed comparison of both groups, the chemical structures showed in the spectra are still being analysed [31]. In a similar way, 1 H NMR has been also used to distinguish between HIV‐1 positive/AIDS patients on antiretroviral treatment and HIV‐1 negative individuals [32]. These experiments were carried out in serum samples and differ‐ ences in the metabolite profiling showed the distinction between the two groups. The authors also suggested that ARV‐associated side effects could be monitored using 1 H NMR [32].

Pneumococcal pneumonia is another condition that has been diagnosed using 1 H NMR [33]. The use of the technique applied to urine samples of patients diagnosed with pneumonia has enable to distinguish pneumonia caused by *Streptococcus pneumoniae* from that caused by other microbes such as viruses or other bacteria. This distinction is observed due to the differences in the metabolomic profiles. So the use of 1 H NMR‐metabolite profiling could result in a rapid, specific and sensitive tool for the diagnosis of pneumococcal pneumonia. In this study, it is also observed that the metabolic profile shown in the samples of patients with pneumonia caused by *S. pneumoniae* changed to a more normal metabolite profiling when specific treatment is administrated, suggesting that the urinary profiles were specific to the infection [33].

#### **3.4. Other types of analyses**

of grape juice and wine during four weeks by 1

1

of 1

the 1

[33].

sepsis [30].

subjects who had consumed the mixture. These results confirm that 1

290 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

analysing metabolite profiling obtained from urine samples [28, 34].

from healthy blood due to differences in metabolites using 1

in the spectra are still being analysed [31]. In a similar way, 1

also suggested that ARV‐associated side effects could be monitored using 1

Pneumococcal pneumonia is another condition that has been diagnosed using 1

been carried out to evaluate sepsis by 1

in the metabolomic profiles. So the use of 1

the metabonomic approach, 1

As discussed above, 1

H NMR. The comparison of the NMR profiling

H NMR could determine

H NMR [35]. Other studies have

H NMR has been also used to

H NMR [32].

H NMR [33].

H NMR, but in rats [30]. Metabolic profiles obtain from

H NMR‐metabolite profiling could result in a rapid,

H NMR has the potential for the early prognostic evaluation of

H NMR has been applied to the diagnosis of hepatitis C virus infection

H NMR to

H NMR

of the samples indicated that only the mixture of wine and grape juice was able to modulate gut microbiota assessed by the reduction in isobutyrate observed only in the samples from

understand gut microbiota has been also extended to the study its impact on obesity by

to compare the metabolic profile of whole blood from serum. This study revealed that the use

H NMR in whole blood allowed obtaining more metabolic details that serum, and, in some cases, the metabolites detected were in higher concentrations in whole blood than in serum. Furthermore, whole blood allowed the determination of metabolites associated with red blood cell metabolism and observed that alterations in their metabolism could be in relationship with sepsis due to the haemolysis they cause [29]. The same authors have carried out several experiments in the same way. In these experiments, they were able to observe that blood samples of sepsis‐induced acute lung injury patients were measurable and distinguishable

H NMR analysis reveal changes in the metabolites involved in energy metabolism, especially in the serum of septic rats. From these results, authors concluded that according to

[31]. This study performed in urine samples was able to identify infected patients and negative individuals with good sensitivity and specificity using a metabonomics model based on the spectra obtained from the urine samples analysis. In this study, although the differences observed in the spectra allowed comparison of both groups, the chemical structures showed

distinguish between HIV‐1 positive/AIDS patients on antiretroviral treatment and HIV‐1 negative individuals [32]. These experiments were carried out in serum samples and differ‐ ences in the metabolite profiling showed the distinction between the two groups. The authors

The use of the technique applied to urine samples of patients diagnosed with pneumonia has enable to distinguish pneumonia caused by *Streptococcus pneumoniae* from that caused by other microbes such as viruses or other bacteria. This distinction is observed due to the differences

specific and sensitive tool for the diagnosis of pneumococcal pneumonia. In this study, it is also observed that the metabolic profile shown in the samples of patients with pneumonia caused by *S. pneumoniae* changed to a more normal metabolite profiling when specific treatment is administrated, suggesting that the urinary profiles were specific to the infection

the impact of the nutrition in human microbiome [27]. Furthermore, the use of 1

H NMR has been also used to study different conditions as sepsis. Stringer et al. used 1

The combination of NMR spectroscopy, with the use of isotopically substituted molecules as tracers is a well‐established protocol in microbiology. These NMR analyses appear to be the most appropriate for such studies because of their analytical power (provided that the labelling of products can be easily monitored non‐invasively), their non‐destructive features, and the large number of compounds that can be analysed simultaneously [6, 36–40]. However, despite the great potential of this combination in clinical practice, these analyses are out of the aim of this chapter.

## **4. Conclusion**

In conclusion, 1 H NMR is an emerging technique in the microbiological field that promises to be a useful tool for the diagnosis of a broad range of conditions, including rapid identification of microorganisms, antimicrobial susceptibility and infectious‐related syndromes. It can be also employed for knowing the metabolic pathways used by microorganisms, allowing the performance strategies for fighting against the infection.

## **Author details**

Lara García‐Álvarez1 , Jesús H. Busto2 , Jesús M. Peregrina2 , Alberto Avenoza2 and José Antonio Oteo1\*

\*Address all correspondence to: jaoteo@riojasalud.es

1 Infectious Diseases Department, San Pedro Hospital‐Center for Biomedical Research of La Rioja (CIBIR), Logroño, Spain

2 Department of Chemistry, Center for Research in Chemical Synthesis, University of la Rioja, Logroño, Spain

## **References**


[17] Price KE, Vandaveer SS, Lunte CE, Larive CK. Tissue targeted metabonomics: metabolic profiling by microdialysis sampling and microcoil NMR. J Pharm Biomed Anal. 2005; 38: 904–909.

[3] Lindon JC, Nicholson JK, Holmes E. The Handbook of Metabonomics and Metabolo‐

292 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

[4] Powers R. NMR metabolomics and drug discovery. Magn Reson Chem. 2009; 47: S2–

[5] García‐Álvarez L, Busto JH, Peregrina JM, Fernández Recio MA, Avenoza A, Oteo JA. Nuclear magnetic resonance applied to antimicrobial drug susceptibility. Future

[6] García‐Álvarez L, Busto JH, Peregrina JM, Santibáñez S, Portillo A, Avenoza A, et al. Nuclear Magnetic Resonance (NMR) as a tool for the study of the metabolism of Rickettsia slovaca. Microbes Infect. 2015; 17: 850–855. DOI: 10.1016/

[7] Gadian DG. NMR and its applications to living systems. Oxford University Press.

[8] Tang J. Microbial metabolomics. Curr Genomics, 2011; 12: 391–403. DOI:

[9] Nicholson JK, Lindon JC. Systems biology: metabonomics. Nature. 2008; 455: 1054–

[10] Brown FF, Campbell ID, Kuchel PW, Rabenstein DC. Human erythrocyte metabolism

[11] Nicholson JK, Connelly J, Lindon JC, Holmes H. Metabonomics: a platform for studying

[12] Lindon JC, Nicholson JK, Holmes E, Antti H, Bollard ME, Keun H, et al. Contemporary issues in toxicology—the role of metabonomics in toxicology and its evaluation by the

[13] Ernst R, Bodenhausen G, Wokaun A. Principles of nuclear magnetic resonance in one

[14] Goldman M. Quantum description of high‐resolution NMR in liquids. Oxford Univer‐

[15] Coen M, Bodkin J, Power D, Bubb WA, Himmelreich U, Kuchel PW, et al. Antifungal effects on metabolite profiles of medically important yeast species measured by nuclear magnetic resonance spectroscopy. Antimicrob Agents

[16] Gupta A, Dwivedi M, Mahdi AA, Gowda GA, Khetrapal CL, Bhandari M. 1H nuclear magnetic resonance spectroscopy for identifying and quantifying common uropatho‐ gens: a metabolic approach to the urinary tract infection. BJU Int. 2009; 104: 236–244.

drug toxicity and gene function. Nat Rev Drug Discov. 2002; 1: 153–161.

studies by H‐1 spin‐echo NMR. FEBS Lett. 1977; 82: 12–16.

COMET project. Toxicol Appl Pharm. 2003; 187: 137–146.

and two dimensions. Oxford University Press. Oxford, 1990.

mics. Elsevier. Amsterdam, 2007.

Microbiol. 2013; 8: 537–547. DOI: 10.2217/fmb.13.8.

S11. DOI: 10.1002/mrc.2461

j.micinf.2015.09.022.

10.2174/138920211797248619

1056. DOI: 10.1038/4551054a.

sity Press. Oxford, 1991.

Chemother. 2006; 50: 4018–4026.

DOI: 10.1111/j.1464‐410X.2009.08448.x.

Oxford, 1995.


**Analysis and Characterization of Foods and Beverages Using Molecular Spectroscopy**

[29] Stringer KA, Younger JG, McHugh C, Yeomans L, Finkel MA, Puskarich MA, et al. Whole blood reveals more metabolic detail of the human metabolome than serum as measured by 1H‐NMR spectroscopy: implications for sepsis metabolomics. Shock.

[30] Lin ZY, Xu PB, Yan SK, Meng HB, Yang GJ, Dai WX, et al. A metabonomic approach to early prognostic evaluation of experimental sepsis by (1)H NMR and pattern recogni‐

[31] Godoy MM, Lopes EP, Silva RO, Hallwass F, Koury LC, Moura IM, et al. Hepatitis C virus infection diagnosis using metabonomics. J Viral Hepat. 2010; 17: 854–858. DOI:

[32] Hewer R, Vorster J, Steffens FE, Meyer D. Applying biofluid 1H NMR‐based metabo‐ nomic techniques to distinguish between HIV‐1 positive/AIDS patients on antiretrovi‐ ral treatment and HIV‐1 negative individuals. J Pharm Biomed Anal. 2006; 41: 1442–

[33] Slupsky CM, Rankin KN, Fu H, Chang D, Rowe BH, Charles PG, et al. Pneumococcal pneumonia: potential for diagnosis through a urinary metabolic profile. J Proteome

[34] Calvani R, Miccheli A, Capuani G, Tomassini Miccheli A, Puccetti C, Delfini M, et al. Gut microbiome‐derived metabolites characterize a peculiar obese urinary metabo‐

[35] Stringer KA, Serkova NJ, Karnovsky A, Guire K, Paine R 3rd, Standiford TJ. Metabolic consequences of sepsis‐induced acute lung injury revealed by plasma 1

[36] Mobashery S, Lerner SA, Johnston M. Monitoring beta‐lactamase activity in vivo by 13C nuclear magnetic resonance spectroscopy. Antimicrob Agents Chemother. 1988:

[37] Aguayo JB, Gamcsik MP, Dick JD. High resolution deuterium NMR studies of bacterial

[38] Ramos A, Jordan KN, Cogan TM, Santos H. C Nuclear Magnetic Resonance Studies of Citrate and Glucose Cometabolism by *Lactococcus lactis*. Appl Environ Microbiol. 1994;

[39] Neves AR, Ramos A, Shearman C, Gasson MJ, Almeida JS, Santos H. Metabolic characterization of Lactococcus lactis deficient in lactate dehydrogenase using in vivo

[40] Chalk PA, Roberts AD, Blows WM. Metabolism of pyruvate and glucose by intact cells of *Helicobacter pylori* studied by 13C NMR spectroscopy. Microbiology. 1994; 140:

magnetic resonance quantitative metabolomics and computational analysis. Am J Physiol Lung Cell Mol Physiol. 2011; 300: L4–L11. DOI: 0.1152/ajplung.00231.2010.

H‐nuclear

type. Int J Obes (Lond). 2010; 34: 1095–1098. DOI: 10.1038/ijo.2010.44.

2015; 44: 200–208. DOI: 10.1097/SHK.0000000000000406.

10.1111/j.1365‐2893.2009.01252.x.

Res. 2009; 8: 5550–5558. DOI: 10.1021/pr9006427.

metabolism. J Biol Chem. 1988; 263: 19552–19557.

13C‐NMR. Eur J Biochem. 2000; 267: 3859–3868.

1446.

32: 1196–1203.

60: 1739–1748.

2085–2092.

tion. NMR Biomed. 2009; 22: 601–608. DOI: 10.1002/nbm.1373.

294 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

## **Improving Food Safety by Using One- and Two-Photon-Induced Fluorescence Spectroscopy for the Detection of Mycotoxins Improving Food Safety by Using One- and Two-Photon-Induced Fluorescence Spectroscopy for the Detection of Mycotoxins**

Lien Smeesters, Wendy Meulebroeck and Hugo Thienpont Lien Smeesters, Wendy Meulebroeck and Hugo Thienpont

Additional information is available at the end of the chapter Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/103856

#### **Abstract**

The presence of mycotoxins in food products is a major worldwide problem. Nowadays, mycotoxins can only be detected by the use of sample-based chemical analyses. Therefore, we demonstrate the use of one- and two-photon-induced fluorescence spectroscopy for the non-destructive detection of mycotoxins in unprocessed food products. We first explain our optical set-up, which is able to measure the localized oneand two-photon-induced fluorescence spectra. Following, as a case study, the detection of aflatoxin in maize kernels is discussed. We present our research methodology, from the characterization of the fluorescence of pure aflatoxin, to the study of the one- and two- photon-induced fluorescence spectra of maize kernels and the development of an optical detection criterion. During both one- and two-photon-induced fluorescence processes, the fluorescence of the aflatoxin influences the intrinsic fluorescence of the maize. Based on the fluorescence spectrum between 400 and 550 nm, a detection criterion to sense the contaminated kernels is defined. Furthermore, we successfully monitored the localized contamination level on the kernel's surface, showing both contaminated kernels with a high contamination in a limited surface area (a few square millimetres) and kernels with a low contamination spread over a large surface area (up to 20 mm2 ). Finally, the extensibility of our research methodology to other fluorescent mycotoxins is discussed.

**Keywords:** spectroscopy, fluorescence, two-photon-induced fluorescence, optical sensing, aflatoxin, multiphoton processes, spectrum analyses

© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

## **1. Introduction**

Fluorescence can be defined as the emission of electromagnetic radiation during the relaxation process of an excited electron that was excited from the ground state to a higher energetic state by the absorption of light. Fluorescence spectroscopy studies the fluorescence light emission as function of the wavelength. To date, fluorescence spectroscopy is already integrated in an extensive range of applications in, amongst others, the biochemical, medical and petrochemical industry [1–5]. In biology and chemistry, it facilitates the study of proteins and the tracking of biochemical reactions. In the medical industry, it is used to follow smart drug delivery systems that map the drug interaction with diseased tissue, while in the petroleum industry, it enables to characterize crude oils during refining processes.

Considering food applications, fluorescence spectroscopy has currently been demonstrated for the screening of products on basis of their chlorophyll concentration [6–8]. Particularly, the detection of foreign objects in food streams has been presented, by comparing the fluorescence spectra of the food products and the foreign objects. For example, when considering a mixture of green glasses and peas, a distinction can be made on basis of the chlorophyll fluorescence intensity, since the peas emit strong fluorescence signals while the glasses emit no fluorescence. In addition, also a distinction between peas, garden beans and sprouts has been demonstrated by evaluating the shape of the chlorophyll fluorescence spectrum [6]. However, toxic contaminants in food and feed products can hardly be optically detected due to the presence of large background fluorescent signals emitted by natural food constituents, like proteins, and the localized presence of the toxins. The current published fluorescence measurements only allow the identification of mycotoxins in homogeneous liquids, like beer or wine, in which no or very low background fluorescent elements are present [9, 10]. Therefore, we investigate the use of fluorescence spectroscopy for the detection of mycotoxins in unprocessed, solid food products.

Mycotoxins are secondary metabolites of toxic fungi, produced by various moulds on a wide range of food products, like nuts, maize, pistachios and peanuts [11]. Six types of mycotoxins are identified as major threats to food safety, of which we focus on the fluorescent mycotoxins, being aflatoxin, ochratoxin and zearalenone. Mycotoxins are observed under a diverse range of environments, both before and after the harvest [11]. Moreover, they appear both in the raw as processed products, since they cannot be destroyed during food processing, like cooking, freezing and roasting. The Food and Agriculture Organization (FAO) estimates that 25% of the world's food crops are affected by mycotoxin-producing fungi. The accumulation of mycotoxins in food and feed products represents a major threat to human and animal health, because they can induce cancer, liver diseases, immune-system suppression, mutagenicity and nervous disorders [9]. Therefore, to decrease the exposure to mycotoxins, international regulations were established [12, 13]. For example, for aflatoxins, the European Commission stated maximum allowed concentrations in the range between 2 and 15 ppb, depending on the commodity, while the USA food safety regulations included a maximum contamination level of 20 ppb for all food products. To fulfil these limitations, mycotoxins are nowadays detected by using time-consuming, sample-based chemical analyses, like liquid chromatography-dual mass spectrometry (LC-MS/MS). However, due to the localized presence of the toxin in the food products and crops, these analyses often give a limited view on the degree of contamination, inducing a large amount of food waste, without entirely preventing the toxins to enter the food chain [11]. Consequently, to increase food safety, the development of a non-destructive detection methodology, which is able to identify the localized contamination on the food products, is indispensable.

**1. Introduction**

to characterize crude oils during refining processes.

Fluorescence can be defined as the emission of electromagnetic radiation during the relaxation process of an excited electron that was excited from the ground state to a higher energetic state by the absorption of light. Fluorescence spectroscopy studies the fluorescence light emission as function of the wavelength. To date, fluorescence spectroscopy is already integrated in an extensive range of applications in, amongst others, the biochemical, medical and petrochemical industry [1–5]. In biology and chemistry, it facilitates the study of proteins and the tracking of biochemical reactions. In the medical industry, it is used to follow smart drug delivery systems that map the drug interaction with diseased tissue, while in the petroleum industry, it enables

298 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

Considering food applications, fluorescence spectroscopy has currently been demonstrated for the screening of products on basis of their chlorophyll concentration [6–8]. Particularly, the detection of foreign objects in food streams has been presented, by comparing the fluorescence spectra of the food products and the foreign objects. For example, when considering a mixture of green glasses and peas, a distinction can be made on basis of the chlorophyll fluorescence intensity, since the peas emit strong fluorescence signals while the glasses emit no fluorescence. In addition, also a distinction between peas, garden beans and sprouts has been demonstrated by evaluating the shape of the chlorophyll fluorescence spectrum [6]. However, toxic contaminants in food and feed products can hardly be optically detected due to the presence of large background fluorescent signals emitted by natural food constituents, like proteins, and the localized presence of the toxins. The current published fluorescence measurements only allow the identification of mycotoxins in homogeneous liquids, like beer or wine, in which no or very low background fluorescent elements are present [9, 10]. Therefore, we investigate the use of fluorescence spectroscopy for the detection of mycotoxins in unprocessed, solid food products.

Mycotoxins are secondary metabolites of toxic fungi, produced by various moulds on a wide range of food products, like nuts, maize, pistachios and peanuts [11]. Six types of mycotoxins are identified as major threats to food safety, of which we focus on the fluorescent mycotoxins, being aflatoxin, ochratoxin and zearalenone. Mycotoxins are observed under a diverse range of environments, both before and after the harvest [11]. Moreover, they appear both in the raw as processed products, since they cannot be destroyed during food processing, like cooking, freezing and roasting. The Food and Agriculture Organization (FAO) estimates that 25% of the world's food crops are affected by mycotoxin-producing fungi. The accumulation of mycotoxins in food and feed products represents a major threat to human and animal health, because they can induce cancer, liver diseases, immune-system suppression, mutagenicity and nervous disorders [9]. Therefore, to decrease the exposure to mycotoxins, international regulations were established [12, 13]. For example, for aflatoxins, the European Commission stated maximum allowed concentrations in the range between 2 and 15 ppb, depending on the commodity, while the USA food safety regulations included a maximum contamination level of 20 ppb for all food products. To fulfil these limitations, mycotoxins are nowadays detected by using time-consuming, sample-based chemical analyses, like liquid chromatography-dual mass spectrometry (LC-MS/MS). However, due to the localized presence of the toxin in the We investigate the use of fluorescence spectroscopy, including one- and two-photon-induced fluorescence (OPIF and TPIF) spectroscopy, as a non-destructive, optical detection technique for the sensing of fluorescent mycotoxins. OPIF and TPIF are two types of fluorescent processes, both resulting in the emission of a fluorescent photon during the relaxation process of an excited electron, but with a different excitation process [14]. To generate an OPIF photon, only a single excitation photon needs to be absorbed, while during the generation of a TPIF photon, two excitation photons need to be absorbed simultaneously to excite the electron and generate the fluorescence signal (**Figure 1**). In the case of OPIF, the energy of the incident photon equals the energy difference between the electronic states, such that the electron obtains sufficient energy to bridge the energy gap between the ground state and the excited state (**Figure 1(a)**). Considering TPIF, the sum of the energies of the two incident photons must equal the bandgap energy (**Figure 1(b)**). The first incident photon excites the electron to a virtual state, which does not need to correspond to any electronic or vibrational energy eigenstate. The second incident photon excites the electron from the virtual state to its higherenergy excited state. Because for both OPIF and TPIF the electron is excited to the same energetic state, they give rise to the same fluorescence wavelengths, but with the use of another excitation wavelength. Generally, OPIF requires ultraviolet (UV) excitation wavelengths, while TPIF uses near-infrared (NIR) light. If the energy of the incident photons does not match with the bandgap, no fluorescence emission will occur. As a result, because the bandgap is a molecular specific property, fluorescence is known to be a selective process. In addition, because TPIF is a non-linear process that requires the simultaneous absorption of two photons instead of one, it is considered to be more selective than OPIF [10].

**Figure 1.** Jabloński diagram for (a) OPIF, in which the electron is excited by the absorption of a single incident photon; (b) TPIF, in which the electron is excited by the simultaneous absorption of two incident photons; (c) second harmonic generation, in which two incident photons recombine to a new photon with the double energy.

Two-photon excitation was first theoretically described by Mario Göppert-Mayer in 1931. However, two-photon fluorescence has only been experimentally demonstrated in the 1960s by Kaiser and Garret when exciting an europium-doped crystal [15]. To date, two-photon fluorescence is mostly used in biomedical engineering, when studying living cells and tissue by using two-photon fluorescence microscopy [16–18]. Intuitively, we would expect that all fluorescent molecules, able to emit OPIF, generate a TPIF signal if they are excited with laser light with the double wavelength and a sufficiently high excitation power density. However, according to the quantum physics selection rules that describe the electron transitions in molecules, one- and two- photon absorption follow different selection rules [19]. More specifically, the occurrence of both OPIF and TPIF electronic transitions is only allowed in noncentrosymmetric molecules. When considering centrosymmetric molecules, the electron transition between the ground state and the excited state is allowed for either OPIF or TPIF. For these molecular structures, no TPIF signals can be observed if OPIF signals can be emitted.

Two-photon-induced fluorescence should not be confused with second-harmonic generation (SHG) (**Figure 1(c)**). SHG is generally defined as a non-linear optical process in which photons with the same energy interacting with a non-linear material are combined to generate new photons with the double energy and thus half the wavelength of the initial photons [20]. Considering the Jabloński diagram for SHG, the two incident photons do not excite the electron but recombine to a new photon with an energy equal to the total energy of the two incident photons. Since no electron is excited, the process does not contain any information about the molecular structure of the products, making it unusable for the detection of toxins.

OPIF and TPIF spectroscopy are both considered as promising optical detection techniques. OPIF shows stronger fluorescence signals than TPIF, giving rise to a strong mycotoxin signal of the solid food products that enables the detection of low contamination levels. Particularly, the OPIF intensity increases linearly with the excitation power, while the TPIF intensity increases with the square of the excitation power. However, the TPIF process features a more selective excitation of the mycotoxins, which can reduce the influence of the background fluorescent signals emitted by the natural fluorescent substituents of the food products. Moreover, TPIF is obtained after excitation with NIR laser light that is more widely commercially available than the required UV laserlines used during OPIF. In addition, compact NIR lasers generally feature higher output powers than UV lasers, enhancing the fluorescence signal. When using UV light in an optical set-up, the optical mirrors and lenses need to be made from or coated with fused silica, resulting in a more expensive set-up than for NIR wavelengths.

In this chapter, we demonstrate the use of OPIF and TPIF as valuable tools for the nondestructive detection of mycotoxins. We first present our measurement configuration, able to study both OPIF and TPIF. Second, we discuss our measurement methodology, including the selection of the excitation wavelengths. Following, as a case study, we investigate the detection of aflatoxin B1 in individual maize kernels. This is because aflatoxin is considered as one of the most dominant mycotoxins and maize cultivates in climates that show an extensive presence of the fungi, giving rise to permanent high aflatoxin contamination levels. We first characterize the fluorescence of pure aflatoxin B1. Subsequently, we study the OPIF and TPIF spectra of healthy and contaminated maize kernels, containing an aflatoxin B1 contamination lower than 1 ppb and higher than 70 ppb, respectively. The emission wavelengths and intensities of the obtained spectra are investigated and a comparison between the performance of OPIF and TPIF is given. Afterwards, we examine the development of an optical detection criterion and study the localized contamination on the samples surfaces. Finally, we discuss the extensibility of our research methodology to other fluorescent mycotoxins.

## **2. Methodology and measurement set-up**

by Kaiser and Garret when exciting an europium-doped crystal [15]. To date, two-photon fluorescence is mostly used in biomedical engineering, when studying living cells and tissue by using two-photon fluorescence microscopy [16–18]. Intuitively, we would expect that all fluorescent molecules, able to emit OPIF, generate a TPIF signal if they are excited with laser light with the double wavelength and a sufficiently high excitation power density. However, according to the quantum physics selection rules that describe the electron transitions in molecules, one- and two- photon absorption follow different selection rules [19]. More specifically, the occurrence of both OPIF and TPIF electronic transitions is only allowed in noncentrosymmetric molecules. When considering centrosymmetric molecules, the electron transition between the ground state and the excited state is allowed for either OPIF or TPIF. For these molecular structures, no TPIF signals can be observed if OPIF signals can be emitted.

300 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

Two-photon-induced fluorescence should not be confused with second-harmonic generation (SHG) (**Figure 1(c)**). SHG is generally defined as a non-linear optical process in which photons with the same energy interacting with a non-linear material are combined to generate new photons with the double energy and thus half the wavelength of the initial photons [20]. Considering the Jabloński diagram for SHG, the two incident photons do not excite the electron but recombine to a new photon with an energy equal to the total energy of the two incident photons. Since no electron is excited, the process does not contain any information about the

OPIF and TPIF spectroscopy are both considered as promising optical detection techniques. OPIF shows stronger fluorescence signals than TPIF, giving rise to a strong mycotoxin signal of the solid food products that enables the detection of low contamination levels. Particularly, the OPIF intensity increases linearly with the excitation power, while the TPIF intensity increases with the square of the excitation power. However, the TPIF process features a more selective excitation of the mycotoxins, which can reduce the influence of the background fluorescent signals emitted by the natural fluorescent substituents of the food products. Moreover, TPIF is obtained after excitation with NIR laser light that is more widely commercially available than the required UV laserlines used during OPIF. In addition, compact NIR lasers generally feature higher output powers than UV lasers, enhancing the fluorescence signal. When using UV light in an optical set-up, the optical mirrors and lenses need to be made from or coated with fused silica, resulting in a more expensive set-up than for NIR

In this chapter, we demonstrate the use of OPIF and TPIF as valuable tools for the nondestructive detection of mycotoxins. We first present our measurement configuration, able to study both OPIF and TPIF. Second, we discuss our measurement methodology, including the selection of the excitation wavelengths. Following, as a case study, we investigate the detection of aflatoxin B1 in individual maize kernels. This is because aflatoxin is considered as one of the most dominant mycotoxins and maize cultivates in climates that show an extensive presence of the fungi, giving rise to permanent high aflatoxin contamination levels. We first characterize the fluorescence of pure aflatoxin B1. Subsequently, we study the OPIF and TPIF spectra of healthy and contaminated maize kernels, containing an aflatoxin B1 contamination lower than 1 ppb and higher than 70 ppb, respectively. The emission wavelengths and

molecular structure of the products, making it unusable for the detection of toxins.

wavelengths.

We pursue a fluorescence measurement set-up suitable for both OPIF and TPIF fluorescence measurements. However, to obtain reliable measurement data, four challenges need to be tackled. (1) The selection of an optimal excitation wavelength is of major importance, to maximize the mycotoxin fluorescence and to minimize the influence of the intrinsic fluorescence signal. (2) The excitation spot size and excitation laser power should be optimized to enable the measurement of the weak mycotoxin fluorescence without damaging the food samples. (3) The natural variation within food products should be monitored, to assure we are measuring the influence of the toxin instead of the optical contrast between different product batches. (4) A careful selection of the products and the illumination positions on their surfaces is indispensable to deal with the localized presence of the mycotoxins. Due to the inhomogeneous presence of the toxin, measurements on individual products are much harder than optical measurements on homogeneous solutions and powders.

Keeping these challenges in mind, we first present our fluorescence measurement configuration, after which we explain the selection of the optimal excitation wavelengths.

## **2.1. OPIF and TPIF measurement set-up**

To measure the OPIF and TPIF spectra of the localized contaminants, we require the generation of both UV and NIR illumination wavelengths, in combination with an accurate detection of the fluorescence signal while scanning the sample surface.

Generally, our measurement set-up can be divided into different building blocks, comprising an illumination laser system, the sample objective and the detecting optical spectrum analyser (**Figure 2**). A frequency-doubled Nd:YAG pump laser (Spectra-Physics Millennia Prime 10sJSPG) pumps a tunable titanium-sapphire laser (Spectra-Physics Tsunami laser). The wavelength of the titanium-sapphire laser can be tuned from 710 to 835 nm by the use of an internal slit, which selects the preferred wavelength after the dispersion of the generated laser light by internal prisms. The maximum output power ranges between 1.20 and 1.50 W, depending on the selected wavelength. The laser light of the titanium-sapphire laser can follow one out of two optical paths, depending on whether OPIF or TPIF measurements are performed (**Figure 2(a)**). For TPIF, the selected laser light is immediately directed towards the sample. In the case of OPIF, the fundamental laser light of the titanium-sapphire laser is directed towards a harmonic generating unit, comprising a second- and third-harmonic-generating crystal to generate the UV excitation wavelengths (**Figure 2(b)**). The second-harmonic-generating crystal is able to generate wavelengths between 355 and 417 nm, with a maximal output power between 200 and 450 mW. The third-harmonic-generating crystal enables us to use wavelengths between 240 and 275 nm, with a maximal output power between 60 and 200 mW. After the generation

of the UV light, by either the second- or third-harmonic-generating crystal, the laser light is directed towards the sample. During both OPIF and TPIF measurements, the sample is illuminated with a circular beam, with a spot diameter of 951 and 231 µm, respectively. To maximize the TPIF irradiance, the spot size of the illumination laser beam was minimized by the use of an additional focusing lens positioned in front of the sample (**Figure 2(d)**).

**Figure 2.** Measurement set-up that allows the investigation of both OPIF and TPIF: (a) schematic representation of the set-up; (b) tunable titanium-sapphire laser (710–835 nm) and harmonic generating unit with the frequency doubling (355–417 nm) and tripling crystals (240–275 nm); (c) automated translation stages on which the sample is mounted, enabling an accurate scanning in the X and Y direction; (d) optical path for the excitation of the sample, after which the fluorescence spectrum is captured by the detecting fibre. The focusing lens minimizes the spot size during the TPIF measurements.

The sample is positioned on a sample holder, containing a circular aperture with a diameter of 7 mm, enabling to position the kernels directly on this aperture. The sample holder is mounted on two automated translation stages (Newport 850G linear actuators) to accurately scan the product in both the X and Y direction (**Figure 2(c)**). For each X and Y position, the incident laser beam illuminates different parts of the product, allowing to study the localized contamination of the sample. Both automated translation stages have a travel range of 5 cm, scanning a maximum surface area of 25 cm2 . The translation stages are driven by a motion controller (Newport ESP300), enabling a high movement speed while assuring a high movement accuracy of 1 µm. During the scanning measurements, we use movement steps between 0.5 and 1.0 mm, as a trade-off between the measurement resolution and the illumination time of the sample. In order to avoid damage to the samples, the illumination time for each scanning measurement was restricted to 3 min, limiting the scanning resolution of the surface measurements.

After the excitation of the sample, the fluorescent signals are captured by a collimating lens, coupled into a broadband optical fibre (UVIR600 fibre of Avantes, transmitting light between 250 and 2500 nm) and guided towards the spectrum analyser (**Figure 2(d)**). We use a collimating lens in combination with a large fibre core diameter (600 µm) to obtain a total acceptance angle of 4.1°, allowing to capture the weak fluorescence signals of a surface area of 39 mm2 (corresponding to the area within the circular aperture of the sample holder). In front of the detecting fibre, we additionally mounted an interference-based filter to eliminate the excitation light. During the OPIF measurements, we implement a long-wave pass filter transmitting from 405 nm onwards (Semrock 405 nm EdgeBasic filter BLP01-405R-25), while during the TPIF measurements, a short-wave pass filter with a cut-on wavelength of 650 ± 5 nm is used (Newport 10SWF-650-B). Without the use of an optical filter, the excitation signals would saturate the measured fluorescence spectrum, leading to incorrect measurement data. Furthermore, we make use of the AvaSpec2048 spectrum analyser, which is able to measure the spectrum between 300 and 1100 nm with a resolution of 8 nm. This instrument has a wide entrance slit of 200 µm that allows capturing the weak fluorescence signals.

To obtain the absolute fluorescence spectrum of a sample, the measured spectra are corrected by a transfer function, taking the wavelength-dependent transmittance of the optical fibre and the sensitivity of the detector inside the spectrum analyser into account. This transfer function was obtained after the measurement of a calibrated light source with an a priori known absolute spectrum. We first connected the calibration light source to the spectrum analyser by using the UVIR600 fibre. Following, we measured the spectrum of the calibration light source and compared the measured photon counts for each wavelength with the specified output power of the manufacturer, in µW/cm2 . Using this measurement data, the Avasoft 8 software is able to calculate the absolute fluorescence spectrum (in µW/cm2 /nm) of a sample (*I*sample), by using the following equation [21]:

$$I\_{\text{sample}}\left(\mathcal{\lambda}\right) = \text{Caldata}\left(\mathcal{\lambda}\right) \frac{\text{sample}\left(\mathcal{\lambda}\right) - \text{dark}\left(\mathcal{\lambda}\right)}{\text{refcal}\left(\mathcal{\lambda}\right) - \text{dark}\text{cal}\left(\mathcal{\lambda}\right)} \frac{\text{Int}\_{\text{cal}}}{\text{Int}\_{\text{sample}}} \tag{1}$$

With Caldata(λ) the intensity of the calibrated light source in µW/cm2 , obtained from the manufacturer; sample(λ) the measured spectrum of the sample, in A/D counts; dark(λ) the dark signal during the sample measurements, in A/D counts; refcal(λ) the measured spectrum of the calibrated light source, in A/D counts; darkcal(λ) the dark signal during the measurements with the calibrated light source, in A/D counts; Intcal the integration time during the measurements of the calibrated light source; and Intsample the integration time during the measurements of the sample.

During the fluorescence characterization measurements, a study of the fluorescence intensity as function of the excitation power is essential. Therefore, the excitation laser power is tuned by mounting attenuation filters behind the laser output, in front of the tilted mirror (mirror indicated in **Figure 2(d)**; the attenuation filters are not presented in this picture).

#### **2.2. Selection of the excitation wavelengths**

of the UV light, by either the second- or third-harmonic-generating crystal, the laser light is directed towards the sample. During both OPIF and TPIF measurements, the sample is illuminated with a circular beam, with a spot diameter of 951 and 231 µm, respectively. To maximize the TPIF irradiance, the spot size of the illumination laser beam was minimized by

**Figure 2.** Measurement set-up that allows the investigation of both OPIF and TPIF: (a) schematic representation of the set-up; (b) tunable titanium-sapphire laser (710–835 nm) and harmonic generating unit with the frequency doubling (355–417 nm) and tripling crystals (240–275 nm); (c) automated translation stages on which the sample is mounted, enabling an accurate scanning in the X and Y direction; (d) optical path for the excitation of the sample, after which the fluorescence spectrum is captured by the detecting fibre. The focusing lens minimizes the spot size during the TPIF

The sample is positioned on a sample holder, containing a circular aperture with a diameter of 7 mm, enabling to position the kernels directly on this aperture. The sample holder is mounted on two automated translation stages (Newport 850G linear actuators) to accurately scan the product in both the X and Y direction (**Figure 2(c)**). For each X and Y position, the incident laser beam illuminates different parts of the product, allowing to study the localized contamination of the sample. Both automated translation stages have a travel range of 5 cm,

controller (Newport ESP300), enabling a high movement speed while assuring a high movement accuracy of 1 µm. During the scanning measurements, we use movement steps between 0.5 and 1.0 mm, as a trade-off between the measurement resolution and the illumination time of the sample. In order to avoid damage to the samples, the illumination time for each scanning measurement was restricted to 3 min, limiting the scanning resolution of the surface meas-

After the excitation of the sample, the fluorescent signals are captured by a collimating lens, coupled into a broadband optical fibre (UVIR600 fibre of Avantes, transmitting light between 250 and 2500 nm) and guided towards the spectrum analyser (**Figure 2(d)**). We use a collimating lens in combination with a large fibre core diameter (600 µm) to obtain a total acceptance angle of 4.1°, allowing to capture the weak fluorescence signals of a surface area of 39 mm2 (corresponding to the area within the circular aperture of the sample holder). In front of the

. The translation stages are driven by a motion

measurements.

urements.

scanning a maximum surface area of 25 cm2

the use of an additional focusing lens positioned in front of the sample (**Figure 2(d)**).

302 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

To maximize the mycotoxin fluorescence intensity, we select an excitation wavelength that is strongly absorbed by the mycotoxins, since an increasing amount of absorbed photons gives rise to an increasing amount of excited electrons and thus also to an enlarged number of fluorescent photons. Furthermore, by using an excitation wavelength with a strong mycotoxin absorbance, the influence of the mycotoxin fluorescence emission onto the natural fluorescence spectrum of the food products can be maximized. When considering the absorption spectrum of aflatoxin B1, ochratoxin A and zearalenone, the strongest absorbances are observed within the 200–400-nm spectral region (**Figure 3(a)**) [9]. Specifically, aflatoxin B1 shows the strongest absorbance in the range between 200 and 275 nm and around 365 nm. Ochratoxin A has a strong absorbance between 200 and 250 nm and around 335 nm, while zearalenone shows a strong absorbance in the 200–350 nm spectral range.

**Figure 3.** Absorbance and fluorescence spectrum of fluorescent mycotoxins: (a) aflatoxin B1, (b) ochratoxin A, (c) zearalenone.

In addition to the absorbance of the mycotoxins, the intrinsic fluorescence of the food products needs to be taken into account. Food products often contain several proteins that show fluorescence emission after excitation with certain wavelengths [22]. For example, maize contains the fluorescent proteins tryptophan (Trp), tyrosine (Tyr) and phenylalanine (Phe), which all show a strong absorbance in the 200–300 nm wavelength range (**Figure 4**). If we would excite maize kernels with a wavelength between 240 and 275 nm, the fluorescence of these proteins would disturb our measurements. The proteins show no fluorescence after excitation with 355–417 nm light, since they only show a weak absorbance from 300 nm onwards.

absorbance, the influence of the mycotoxin fluorescence emission onto the natural fluorescence spectrum of the food products can be maximized. When considering the absorption spectrum of aflatoxin B1, ochratoxin A and zearalenone, the strongest absorbances are observed within the 200–400-nm spectral region (**Figure 3(a)**) [9]. Specifically, aflatoxin B1 shows the strongest absorbance in the range between 200 and 275 nm and around 365 nm. Ochratoxin A has a strong absorbance between 200 and 250 nm and around 335 nm, while zearalenone shows a

304 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

**Figure 3.** Absorbance and fluorescence spectrum of fluorescent mycotoxins: (a) aflatoxin B1, (b) ochratoxin A, (c) zeara-

In addition to the absorbance of the mycotoxins, the intrinsic fluorescence of the food products needs to be taken into account. Food products often contain several proteins that show fluorescence emission after excitation with certain wavelengths [22]. For example, maize contains the fluorescent proteins tryptophan (Trp), tyrosine (Tyr) and phenylalanine (Phe),

strong absorbance in the 200–350 nm spectral range.

lenone.

**Figure 4.** Absorbance spectrum of different fluorescent proteins in maize: phenylalenine (Phe), tryptophan (Trp) and tyrosine (Tyr) [22].

**Figure 5.** Intrinsic fluorescence of healthy food products: (a) OPIF of coffee beans after excitation with 265 and 398 nm, (b) OPIF of pistachio nuts after excitation with 365 nm, (c) TPIF of pistachio shells after excitation with 780 nm.

Furthermore, also the natural variation within food products should be examined during the fluorescence measurements. Due to the large internal variation in density, texture and substituents concentration, various batches of a product type can show different fluorescence spectra or intensities. As an illustration, we present the intrinsic fluorescence of healthy coffee beans and pistachio nuts, showing significant differences within the fluorescence spectra of different batches (**Figure 5**). When illuminating two batches of coffee beans with 265 and 398 nm laser light, different ratios of the local fluorescence maxima at 480 and 680 nm can be observed for each batch (**Figure 5(a)**). Two varieties of pistachio nuts, after illumination with 365 nm, show a different fluorescence spectrum between 425 and 625 nm, with a maximal difference at 530 nm (**Figure 5(b)**). Both types of pistachio nuts do show chlorophyll fluorescence, present between 625 and 825 nm. Considering the outer pistachio shells, after illumination with 780 nm, a clear contrast between the fluorescence of both pistachio types can be identified between 425 and 625 nm (**Figure 5(c)**). The fluorescence spectrum of the first type shows a clear two-photon-induced fluorescence signal between 425 and 625 nm and a chlorophyll fluorescence maximum at 645 nm, while the second type only shows chlorophyll fluorescence. As a result, to account for the natural variation, different independent sets of products should always be compared.

Finally, when selecting the excitation wavelengths, we keep in mind the commercial available laserlines. We prefer using commercially available wavelengths to enable the integration of our developed optical detection criterion into industrial scanning-based sorting devices.

When considering our case study, the sensing of the aflatoxin contamination in maize kernels, we use an excitation wavelength of 365 nm during the OPIF measurements, since at this wavelength, the absorbance of the aflatoxin B1 is maximized, while the absorbance of the proteins is minimized. Consequently, to study the TPIF spectrum, excitation wavelengths close to 730 nm are required. In addition, when studying the maize kernels, we also investigate the fluorescence spectrum after excitation with 750 and 780 nm, to study the influence of the matrix constituents onto the fluorescence spectrum.

## **3. Case-study: localized detection of the aflatoxin contamination**

As a case study, we focus on the detection of the localized aflatoxin contamination in individual maize kernels. Specifically, within this section, we present an in-depth explanation of our research methodology, giving rise to the development of an accurate optical sensing criterion. We first characterize the OPIF and TPIF spectra of pure aflatoxin, after which we investigate the fluorescence properties of the healthy and contaminated maize kernels.

## **3.1. Characterization of the aflatoxin fluorescence spectrum**

Aflatoxin is one of the most dominant and toxic mycotoxins. It is produced by the fungi *Asperigullus flavus* and *Asperigullus parasiticus*. To perform a basic characterization of the fluorescence properties of aflatoxin, we measured the OPIF and TPIF spectrum of pure aflatoxin B1 powder (98% or better purity, produced by the *A. flavus* fungi) that we purchased from Sigma-Aldrich. It is a white to yellow crystalline powder that we measured in its solid state.

We present the mean OPIF and TPIF spectra of pure aflatoxin B1 powder (**Figure 6**). The OPIF spectrum is obtained after excitation with 365 nm, with an excitation power density of 42 mW/mm2 , while the TPIF spectrum is measured after excitation with 730 nm, with an excitation power density of 14.3 W/mm2 . Both fluorescence spectra show their maximal fluorescence intensity at 428 nm, which corresponds with the expected aflatoxin B1 fluorescence maximum (**Figure 3(a)**). The OPIF spectrum shows higher fluorescence intensities than the TPIF spectrum. The maximal OPIF intensity (at 428 nm) is equal to 1.68 ± 0.93 µW/cm2 , while we observe a maximal TPIF intensity of 0.20 ± 0.09 µW/cm2 . The measured TPIF intensity is hence 10 times weaker than the OPIF intensity. The shape of the TPIF spectrum shows a narrower peak than the OPIF spectrum due to the more selective excitation during two-photon absorption than during one-photon absorption. Furthermore, in the TPIF spectrum at 365 nm, we observe a second-harmonic generation signal. This peak is created by the recombination of two illumination photons to a new photon with the double energy. Around 400 nm, the OPIF spectrum shows a steeper slope than the TPIF spectrum, because of the presence of a long-wave pass filter suppressing light below 405 nm during the OPIF measurements.

nation with 780 nm, a clear contrast between the fluorescence of both pistachio types can be identified between 425 and 625 nm (**Figure 5(c)**). The fluorescence spectrum of the first type shows a clear two-photon-induced fluorescence signal between 425 and 625 nm and a chlorophyll fluorescence maximum at 645 nm, while the second type only shows chlorophyll fluorescence. As a result, to account for the natural variation, different independent sets of

306 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

Finally, when selecting the excitation wavelengths, we keep in mind the commercial available laserlines. We prefer using commercially available wavelengths to enable the integration of our developed optical detection criterion into industrial scanning-based sorting devices.

When considering our case study, the sensing of the aflatoxin contamination in maize kernels, we use an excitation wavelength of 365 nm during the OPIF measurements, since at this wavelength, the absorbance of the aflatoxin B1 is maximized, while the absorbance of the proteins is minimized. Consequently, to study the TPIF spectrum, excitation wavelengths close to 730 nm are required. In addition, when studying the maize kernels, we also investigate the fluorescence spectrum after excitation with 750 and 780 nm, to study the influence of the matrix

As a case study, we focus on the detection of the localized aflatoxin contamination in individual maize kernels. Specifically, within this section, we present an in-depth explanation of our research methodology, giving rise to the development of an accurate optical sensing criterion. We first characterize the OPIF and TPIF spectra of pure aflatoxin, after which we investigate

Aflatoxin is one of the most dominant and toxic mycotoxins. It is produced by the fungi *Asperigullus flavus* and *Asperigullus parasiticus*. To perform a basic characterization of the fluorescence properties of aflatoxin, we measured the OPIF and TPIF spectrum of pure aflatoxin B1 powder (98% or better purity, produced by the *A. flavus* fungi) that we purchased from Sigma-Aldrich. It is a white to yellow crystalline powder that we measured in its solid

We present the mean OPIF and TPIF spectra of pure aflatoxin B1 powder (**Figure 6**). The OPIF spectrum is obtained after excitation with 365 nm, with an excitation power density of 42

intensity at 428 nm, which corresponds with the expected aflatoxin B1 fluorescence maximum (**Figure 3(a)**). The OPIF spectrum shows higher fluorescence intensities than the TPIF spec-

, while the TPIF spectrum is measured after excitation with 730 nm, with an excitation

. Both fluorescence spectra show their maximal fluorescence

. The measured TPIF intensity is hence 10 times

, while we observe

**3. Case-study: localized detection of the aflatoxin contamination**

the fluorescence properties of the healthy and contaminated maize kernels.

trum. The maximal OPIF intensity (at 428 nm) is equal to 1.68 ± 0.93 µW/cm2

**3.1. Characterization of the aflatoxin fluorescence spectrum**

products should always be compared.

constituents onto the fluorescence spectrum.

state.

mW/mm2

power density of 14.3 W/mm2

a maximal TPIF intensity of 0.20 ± 0.09 µW/cm2

**Figure 6.** OPIF and TPIF spectra of the pure aflatoxin B1 powder, after excitation with 365 and 730 nm, respectively.

To characterize the aflatoxin B1 fluorescence spectrum, we studied the integrated fluorescence intensity as function of the excitation power, while maintaining a constant spot diameter of 951 µm during the OPIF measurements and 231 µm during the TPIF measurements (**Figure 7**). For each excitation power, we integrated the measured mean fluorescence spectrum between 400 and 600 nm to obtain the depicted integrated fluorescence intensity. Studying the integrated OPIF intensity as function of the excitation laser power, we observe a linear relationship, representing the linear one-photon absorption. However, at higher excitation powers, starting from 35 mW onwards, the integrated OPIF intensity deviates from the linear relationship and starts saturating. In the saturated region, the maximum fluorescence intensity is reached since then all electrons are excited to the higher energy state. The TPIF intensity shows a quadratic dependence on the excitation laser power, confirming the occurrence of non-linear two-photon absorption. Moreover, we can observe that TPIF requires much higher excitation powers, starting from 100 mW onwards. The variation of the measured data is less than 10 and 12% during the OPIF and TPIF measurements, respectively. Both the linear and exponential fits show an adjusted r-square value of 0.98 and 0.97, respectively, ensuring that the fitted function is a good representation of the measured data.

TPIF requires a higher excitation power density than OPIF [23]. The minimal excitation power density for OPIF is 5 mW/mm2 , while TPIF requires a minimal power density of 2 W/mm2 . As a result, the measurement constraints are more severe for TPIF than for OPIF. A high excitation power, a small spot size and a sensitive detector are indispensable to measure TPIF signals. However, considering the implementation in practical applications, TPIF uses NIR excitation wavelengths which are more widely commercially available than the required UV excitation wavelengths for OPIF.

**Figure 7.** Intensity of the integrated OPIF spectrum increases linearly with the excitation power, while the intensity of the integrated TPIF spectrum shows a quadratic dependence on the excitation power.

**Figure 8.** Scanning of the aflatoxin B1 powder, visualizing the location of the emitted fluorescence signals: (a) aflatoxin B1 power, (b) OPIF surface plot and (c) TPIF surface plot, indicating the localized fluorescence intensity at 428 nm.

To validate the automated screening of the samples, a scanning of the pure aflatoxin B1 powder is performed (**Figure 8**). Particularly, we consider a Petri dish on which the aflatoxin B1 powder is disposed in two separate areas, of which we screen an area of 20.0 mm by 25.0 mm, with a resolution of 0.5 mm. By adjusting the X and Y position of the automated translation stages, while measuring the fluorescence signal for every position, the aflatoxin powder can be identified (indicated by the solid and dotted circles in **Figure 8**). During both the OPIF and TPIF measurements, the aflatoxin B1 fluorescence is detected. The encircled regions in the surface plots indicate the areas from which a minimum fluorescence intensity of 0.8 and 0.1 µW/cm2 at 428 nm is detected during the OPIF and TPIF measurements, respectively. The aflatoxin powder consists of granules positioned at two areas on the Petri dish (**Figure 8(a)**). If a small aflatoxin granule is illuminated, we detect a fluorescence signal. However, we obtain a smoother surface plot for OPIF than for TPIF, due to the larger spot diameter of the illumination beam (951 µm) during the OPIF measurements. During the OPIF measurements, most illumination positions around the powder excite one or multiple aflatoxin granules, since the surface area in-between the granules is smaller than the spot diameter. When using a smaller spot diameter, like during the TPIF measurements (231 µm), no fluorescence signal can be captured when the illuminating beam is positioned in-between the aflatoxin granules. Consequently, a more fragmented surface plot is obtained (**Figure 8(c)**).

Generally, our fluorescence measurements of the pure aflatoxin B1 powder correspond with the theoretical characteristics, validating a correct measurement of the fluorescence signals. In addition, the powder could be successfully visualized, indicating a good operation of the automated scanning of the samples.

## **3.2. Sensing of the localized aflatoxin contamination in maize kernels**

We investigate the sensing of aflatoxin-contaminated maize kernels by using fluorescence spectroscopy. We first give an overview of the investigated samples. Following, we study the OPIF and TPIF spectra of healthy and contaminated maize kernels, including the investigation of the intrinsic fluorescence spectra and its dependency on the excitation wavelength. Based on this study, we develop a detection criterion enabling to sense the contaminated kernels. Finally, the detection of the localized contamination areas onto the kernel's surfaces are studied and visualized with surface plots.

## *3.2.1. Overview of the investigated samples*

TPIF requires a higher excitation power density than OPIF [23]. The minimal excitation power

308 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

a result, the measurement constraints are more severe for TPIF than for OPIF. A high excitation power, a small spot size and a sensitive detector are indispensable to measure TPIF signals. However, considering the implementation in practical applications, TPIF uses NIR excitation wavelengths which are more widely commercially available than the required UV excitation

**Figure 7.** Intensity of the integrated OPIF spectrum increases linearly with the excitation power, while the intensity of

**Figure 8.** Scanning of the aflatoxin B1 powder, visualizing the location of the emitted fluorescence signals: (a) aflatoxin B1 power, (b) OPIF surface plot and (c) TPIF surface plot, indicating the localized fluorescence intensity at 428 nm.

To validate the automated screening of the samples, a scanning of the pure aflatoxin B1 powder is performed (**Figure 8**). Particularly, we consider a Petri dish on which the aflatoxin B1 powder is disposed in two separate areas, of which we screen an area of 20.0 mm by 25.0 mm, with a resolution of 0.5 mm. By adjusting the X and Y position of the automated translation stages, while measuring the fluorescence signal for every position, the aflatoxin powder can be identified (indicated by the solid and dotted circles in **Figure 8**). During both the OPIF and TPIF measurements, the aflatoxin B1 fluorescence is detected. The encircled regions in the surface plots indicate the areas from which a minimum fluorescence intensity of 0.8 and 0.1

at 428 nm is detected during the OPIF and TPIF measurements, respectively. The

the integrated TPIF spectrum shows a quadratic dependence on the excitation power.

, while TPIF requires a minimal power density of 2 W/mm2

. As

density for OPIF is 5 mW/mm2

wavelengths for OPIF.

µW/cm2

We consider two different independent maize batches, each with a healthy and contaminated subsample. One healthy and one contaminated maize batch were harvested in 2012 and provided by an Italian company. The second set of healthy and contaminated maize samples was collected from Croatian farmers, after the harvest in 2013. From each maize batch, a subsample of 25 g was drawn for the analytical determination of the aflatoxin contamination level. Each sample was chemically analysed using the ToxiQuant mycotoxin testing system of ToxiMet [24]. Considering the Italian maize kernels, the contaminated sample shows 72.1 ppb aflatoxin B1 and the healthy one 0.0 ppb. The maize samples from Croatia show approximately the same aflatoxin B1 contamination level, being 78.9 ppb for the contaminated sample and 0.8 ppb for the healthy sample. After our measurements were performed, the contamination level of the maize samples was confirmed by the CODA-CERVA, the Belgian Reference Laboratory for Mycotoxins. Comparing the contamination levels with the international regulations, both the Italian and Croatian contaminated batches can be considered as highly contaminated. The European Commission states the maximum allowed total aflatoxin concentration in maize to be 10 ppb, while the USA food safety regulations included a limit of 20 ppb of total aflatoxins in all food products [12].

During our fluorescence measurements, we investigated the fluorescence spectra of 45 healthy and contaminated Croatian maize kernels to obtain a statistic relevant distribution of our measurement data. To observe the influence of the sample type and the harvest environments on the fluorescence spectra, we also measure the fluorescence spectra of 15 healthy and contaminated Italian maize kernels. The Italian company provided only small maize batches, which limits the number of studied maize kernels, but is sufficient to monitor the environmental influences on the spectra. Because aflatoxin is sensitive to light, the samples are permanently stored in a dark enclosure to reduce the environmental influences onto the measurements. In-between the measurements, the samples are stored in a fridge to avoid crosscontamination.

#### *3.2.2. OPIF and TPIF spectra of the maize kernels*

To investigate the optical detection of aflatoxin B1, we study the OPIF and TPIF spectra of healthy and contaminated maize kernels of both the Croatian and Italian maize batches. We measured the OPIF spectrum after excitation with 365 nm, with an excitation power density of 317 mW/mm2 (**Figure 9(a)**). The TPIF spectra are measured after excitation with 730, 750

**Figure 9.** Fluorescence spectra of the healthy and contaminated maize kernels of the Croatian and Italian maize batches: (a) OPIF spectrum after excitation with 365 nm; TPIF spectrum after excitation with (b) 730 nm, (c) 750 nm and (d) 780 nm.

and 780 nm, with an excitation power density of 26.2, 29.1 and 36.0 W/mm2 , respectively (**Figure 9(b)**, **(c)** and **(d)**). During the TPIF measurements, we investigate the fluorescence spectrum after excitation with multiple wavelengths to monitor the influence of the illumination wavelength onto the fluorescence emission of the maize kernels. To maximize the signal to noise ratio, we illuminate the maize kernels with the maximal output power of the titaniumsapphire laser and harmonic generating unit. However, we limited the measurement time to 20 and 200 ms for the OPIF and TPIF measurements, respectively, to avoid damage to the sample surface. For every maize batch, the mean fluorescence spectrum is depicted. Both the healthy and the contaminated samples show a fluorescence signal due to the intrinsic fluorescence of the maize kernels (**Figure 9**). On the basis of these measurements, a comprehensive evaluation can be made by comparing the fluorescence spectra of the Italian and Croatian maize batches, by evaluating the performance of OPIF and TPIF and by defining the spectral differences between the healthy and contaminated samples.

During our fluorescence measurements, we investigated the fluorescence spectra of 45 healthy and contaminated Croatian maize kernels to obtain a statistic relevant distribution of our measurement data. To observe the influence of the sample type and the harvest environments on the fluorescence spectra, we also measure the fluorescence spectra of 15 healthy and contaminated Italian maize kernels. The Italian company provided only small maize batches, which limits the number of studied maize kernels, but is sufficient to monitor the environmental influences on the spectra. Because aflatoxin is sensitive to light, the samples are permanently stored in a dark enclosure to reduce the environmental influences onto the measurements. In-between the measurements, the samples are stored in a fridge to avoid cross-

310 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

To investigate the optical detection of aflatoxin B1, we study the OPIF and TPIF spectra of healthy and contaminated maize kernels of both the Croatian and Italian maize batches. We measured the OPIF spectrum after excitation with 365 nm, with an excitation power density

**Figure 9.** Fluorescence spectra of the healthy and contaminated maize kernels of the Croatian and Italian maize batches: (a) OPIF spectrum after excitation with 365 nm; TPIF spectrum after excitation with (b) 730 nm, (c) 750 nm and (d)

(**Figure 9(a)**). The TPIF spectra are measured after excitation with 730, 750

contamination.

of 317 mW/mm2

780 nm.

*3.2.2. OPIF and TPIF spectra of the maize kernels*


**Table 1.** Measured maximum fluorescence intensity and its variation for the different excitation wavelengths.

Considering the different maize batches, the fluorescence spectra of the Croatian and Italian maize correspond well within the variances of the measurements (**Table 1**). For all excitation wavelengths, both maize types emit corresponding fluorescence signals, showing a similar spectral shape in the same wavelength regions. The present intensity variances are caused by the differences within the molecular structure of the maize kernels. Because maize is a natural product, the maize kernels show a large internal variation in surface shape, density and natural composition. In correspondence with the measurements on the pure aflatoxin B1 powder, the fluorescence intensity induced by one-photon excitation is much stronger than the intensity induced by two-photon excitation. Specifically, the OPIF signal is approximately 500 times stronger than the TPIF signal.

Comparing the fluorescence spectra of the healthy and contaminated maize kernels, we observe significant differences in intensity and emission wavelength. We do not observe the aflatoxin fluorescence directly, but we measure its influence on the intrinsic fluorescence of the maize. In the contaminated samples, the aflatoxin is bonded to the different natural constituents of the healthy maize, changing the molecular structure of the constituents and therefore influencing the intensity and wavelength of the emitted fluorescence signals. During both the OPIF and TPIF measurements, we observe higher fluorescence intensities for the healthy maize kernels than for the contaminated ones (**Table 1**). The intensity contrast is the largest for the TPIF spectra, since the different bonds inside the maize are more selectively excited during the TPIF process than during the OPIF one. The largest intensity differences between the healthy and contaminated samples are observed after excitation with 730 nm, where the mean fluorescence intensity of the healthy maize is four times stronger than the mean fluorescence intensity of the contaminated maize. In contrast, the minimum intensity differences are observed after excitation with 365 nm. The observed large intensity variations within the healthy and contaminated samples are caused by a combination of the curved surface of the maize kernels and the non-homogeneous presence of the toxins inside the maize.

In addition to the fluorescence intensity differences, we observe a wavelength shift between the fluorescence maxima of the healthy and contaminated maize samples (**Table 2**). Generally, the OPIF and TPIF spectra show a wavelength shift of approximately 50 nm. To compare the obtained wavelength shift for the different excitation wavelengths, we calculate the class difference for both the Croatian and Italian maize. The class difference (*D*) is a measure for the difference between the average values (*µ*) of two product types, taking the standard deviation (*σ*) and the amount of measured samples (*N*) into account [25]:

$$D = \frac{\left| \mu\_{\text{containated}} - \mu\_{\text{heathty}} \right|}{\sqrt{\frac{\sigma^2\_{\text{containated}}}{N\_{\text{containated}}} + \frac{\sigma^2\_{\text{heathty}}}{N\_{\text{heathty}}}}} \tag{2}$$


**Table 2.** Emission wavelength at maximum fluorescence intensity and its variation for the different excitation wavelengths.

The OPIF measurements show the largest class differences, indicating the largest optical contrast between the healthy and contaminated maize kernels. The variation of the class differences between the OPIF and TPIF measurements are mainly caused by their different variances. The TPIF spectra show a larger variance than the OPIF spectra, which decreases their class difference. TPIF signals show a weaker intensity, resulting in a stronger relative noise signal and therefore a lower signal to noise ratio.

To validate the influence of the aflatoxin onto the intrinsic fluorescence spectrum of the maize, we tune the excitation wavelength towards the wavelengths for which aflatoxin B1 shows a weaker absorbance. We measured the fluorescence spectrum of healthy and contaminated maize kernels after excitation with 405 and 810 nm laser light, during the OPIF and TPIF measurements, respectively (**Figure 10**). Both the OPIF and TPIF measurements show a small difference between the fluorescence spectrum of the healthy and contaminated samples. We measure a wavelength shift of 8 ± 6 nm and 26 ± 12 nm, after excitation with 405 and 810 nm, respectively. Moreover, we observe only small differences between the fluorescence intensities. Because aflatoxin B1 shows a weak absorbance at 405 nm (**Figure 3(a)**), it only minimally influences the intrinsic fluorescence spectrum of the maize kernels.

**Figure 10.** Fluorescence spectra of the healthy and contaminated maize kernels: (a) OPIF spectrum after excitation with 405 nm and (b) TPIF spectrum after excitation with 810 nm.

Summarized, both OPIF and TPIF show a spectroscopic contrast between the healthy and contaminated maize kernels. In the next step, we use the above quantitative evaluation to define and test our optical detection criterion.

### *3.2.3. Development of an optical detection criterion*

TPIF spectra, since the different bonds inside the maize are more selectively excited during the TPIF process than during the OPIF one. The largest intensity differences between the healthy and contaminated samples are observed after excitation with 730 nm, where the mean fluorescence intensity of the healthy maize is four times stronger than the mean fluorescence intensity of the contaminated maize. In contrast, the minimum intensity differences are observed after excitation with 365 nm. The observed large intensity variations within the healthy and contaminated samples are caused by a combination of the curved surface of the

In addition to the fluorescence intensity differences, we observe a wavelength shift between the fluorescence maxima of the healthy and contaminated maize samples (**Table 2**). Generally, the OPIF and TPIF spectra show a wavelength shift of approximately 50 nm. To compare the obtained wavelength shift for the different excitation wavelengths, we calculate the class difference for both the Croatian and Italian maize. The class difference (*D*) is a measure for the difference between the average values (*µ*) of two product types, taking the standard deviation

> contaminated healthy contaminated healthy contaminated healthy

**Class difference Croatian maize** 

**Healthy Contaminated Healthy Contaminated** 

365 nm 443 ± 5 497 ± 15 9.7 442 ± 5 486 ± 6 10.8 730 nm 441 ± 3 496 ± 20 6.1 441 ± 3 496 ± 16 4.1 750 nm 444 ± 7 497 ± 24 3.8 440 ± 5 490 ± 17 2.7 780 nm 450 ± 10 503 ± 19 5.2 450 ± 7 498 ± 10 4.8

**Table 2.** Emission wavelength at maximum fluorescence intensity and its variation for the different excitation

The OPIF measurements show the largest class differences, indicating the largest optical contrast between the healthy and contaminated maize kernels. The variation of the class differences between the OPIF and TPIF measurements are mainly caused by their different variances. The TPIF spectra show a larger variance than the OPIF spectra, which decreases their class difference. TPIF signals show a weaker intensity, resulting in a stronger relative

To validate the influence of the aflatoxin onto the intrinsic fluorescence spectrum of the maize, we tune the excitation wavelength towards the wavelengths for which aflatoxin B1 shows a

s

<sup>+</sup> (2)

**Class difference Italian maize**

**Dominant emission wavelength Italian maize (nm)**

<sup>²</sup> ²

*N N*

*µ µ <sup>D</sup>*


s maize kernels and the non-homogeneous presence of the toxins inside the maize.

312 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

(*σ*) and the amount of measured samples (*N*) into account [25]:

**Dominant emission wavelength Croatian**

noise signal and therefore a lower signal to noise ratio.

**maize (nm)**

**Excitation wavelength** 

wavelengths.

To obtain an accurate optical detection criterion, taking the shape, intensity and maximum emission wavelength of the fluorescence spectra into account, we examine the integral of the fluorescence spectra in various wavelength intervals. Keeping in mind the wavelength ranges of the fluorescence spectra and the maximum fluorescence intensities, both depicted in **Figure 9** and presented in **Tables 1** and **2**, we obtain an optimal contrast between the healthy and contaminated samples by considering the ratio of the integrated fluorescence intensity from 475 to 550 nm and the integrated fluorescence intensity from 400 to 475 nm. When visualizing the ratio of these integrals for each maize kernel, a clear distinction between the healthy and contaminated samples is visible (**Figure 11**). Particularly, the ratios show a minimum contrast of 0.24, 0.13, 0.19, 0.25 and a maximum contrast of 2.15, 1.93, 2.87, 3.39 for excitation with 365, 730, 750 and 780 nm, respectively. The largest difference between the mean ratio of the healthy and contaminated samples is observed after excitation with 780 nm. Considering the class differences of the integral ratio between the healthy and contaminated samples, we obtain a value of 116.2, 111.1, 48.8 and 54.9 for excitation with 365, 730, 750 and 780 nm, respectively. The largest class difference can thus be found after excitation with 365 nm. Regarding the TPIF measurements, the largest class difference is obtained after excitation with 780 nm. The TPIF measurements show generally a smaller class difference than the OPIF ones, due to their larger measurement variation. However, the contaminated samples can still be properly identified with TPIF, allowing using NIR excitation wavelengths, instead of the UV wavelengths for OPIF.

**Figure 11.** Contrast between the fluorescence spectra of the healthy and contaminated maize samples, visualized by the ratio of the integrated fluorescence spectrum from 475 nm until 550 nm to the integrated fluorescence spectrum from 400 nm until 475 nm, after excitation with (a) 365 nm, (b) 730 nm, (c) 750 nm and (d) 780 nm.

The ratios of the integrated fluorescence intensities of the contaminated samples generally show a larger variation than the ratios of the healthy samples, because of the localized presence of the aflatoxin and the variable aflatoxin concentration in the maize kernels. Depending on the illumination position on the maize kernel's surface, a different contamination level may be present, influencing the wavelength and intensity of the fluorescence spectrum. To visualize the localized aflatoxin contamination on the maize kernel's surfaces, scanning measurements needed to be performed.

#### *3.2.4. Monitoring of the localized presence of the aflatoxin*

365, 730, 750 and 780 nm, respectively. The largest class difference can thus be found after excitation with 365 nm. Regarding the TPIF measurements, the largest class difference is obtained after excitation with 780 nm. The TPIF measurements show generally a smaller class difference than the OPIF ones, due to their larger measurement variation. However, the contaminated samples can still be properly identified with TPIF, allowing using NIR excitation

314 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

**Figure 11.** Contrast between the fluorescence spectra of the healthy and contaminated maize samples, visualized by the ratio of the integrated fluorescence spectrum from 475 nm until 550 nm to the integrated fluorescence spectrum

The ratios of the integrated fluorescence intensities of the contaminated samples generally show a larger variation than the ratios of the healthy samples, because of the localized presence of the aflatoxin and the variable aflatoxin concentration in the maize kernels. Depending on the illumination position on the maize kernel's surface, a different contamination level may be present, influencing the wavelength and intensity of the fluorescence spectrum. To visualize

from 400 nm until 475 nm, after excitation with (a) 365 nm, (b) 730 nm, (c) 750 nm and (d) 780 nm.

wavelengths, instead of the UV wavelengths for OPIF.

To monitor the localized presence of the aflatoxin, we scanned the maize kernels by using the automated translation stages. We investigated the localized aflatoxin contamination when illuminating the samples with 365 and 780 nm laser light, since these wavelengths showed the largest spectroscopic differences during the OPIF and TPIF measurements, respectively [26].

**Figure 12.** Visualization of the localized aflatoxin contamination, measured during OPIF, by mapping the dominant fluorescent emission wavelength along the kernel's surface: (a) healthy kernel, (b) kernel with a small localized contamination area, (c) kernel with homogeneous contamination and (d) kernel with a medium to large contamination at the edge.

During the OPIF measurements, we scanned the maize kernels with a resolution of 0.5 mm. For every sample, we recorded 405 spectra, each of which corresponding with a different illumination position of the laser beam. Considering the fluorescence wavelength showing the maximum fluorescence intensity, different regions in the kernels can be identified (**Figure 12**). In correspondence with the previous measurements, the fluorescence spectra of the healthy maize kernels show their maximum fluorescence intensity between 400 and 450 nm (**Figure 12(a)**). The contaminated areas show their maximum fluorescence intensity at wavelengths longer than 450 nm. By interpreting the surface plots, various areas containing different contamination levels can be observed on different positions of the kernel's surface (**Figure 12(b)**, **(c)** and **(d)**). As an example, the contaminated kernels can show a small localized contaminated area, a homogeneous contamination along the maize kernel or a medium to large contamination at the edge of the kernel. When considering the ratio of the integrated fluorescence spectra, specifically the ratio of the integrated fluorescence intensity between 475 and 550 nm and the integrated fluorescence intensity between 400 and 475 nm, a similar observation can be made (**Figure 13**). In correspondence with **Figure 11**, the high-contaminated areas show the largest ratios. Furthermore, the gradient of the contamination level is clearly visible. When considering **Figure 13(b)**, two highly contaminated areas surrounded by a medium-contaminated area can be identified, while **Figure 13(d)** shows a highly contaminated area at the edge that flows into a healthy area in the centre of the kernel. Concluding, localized contaminated areas can be accurately identified and characterized by mapping the integral ratio on the screened kernel's surfaces.

**Figure 13.** Visualization of the localized aflatoxin contamination during OPIF, by mapping fluorescence ratios (integrated fluorescence intensity between 475 and 550 nm divided by the integrated fluorescence intensity between 400 and 475 nm) along the kernel's surface: (a) healthy kernel, (b) kernel with small localized contamination area, (c) kernel with homogeneous contamination and (d) kernel with a medium to large contamination at the edge.

Considering the TPIF measurements, the maize kernels could only be partially scanned with a lower resolution of 1 mm (instead of the 0.5 mm resolution used during the OPIF scanning measurements). When accurately scanning a complete maize kernel's surface during the TPIF measurements, the kernel's surface would be damaged, due to the required long illumination time and high illumination power density. Specifically, the illumination time is significantly longer during the TPIF measurements than during the OPIF measurements (approximately 10 times longer) due to the longer measurement time of the spectrum analyser required to capture the weak TPIF signals. Consequently, to avoid measuring the influence of damaging effects, the kernels are measured with a lower accuracy (**Figure 14**). Studying the corresponding surface plots, regions with a high and low contamination level can still be observed. Specifi‐ cally, the regions with a dominant fluorescence wavelength longer than 450 nm can be considered as contaminated. Furthermore, the longer the fluorescence wavelength, the higher the contamination level on the kernel's surface. The maize kernel depicted in **Figure 14(a)** shows generally a lower contamination than the one depicted in **Figure 14(b)**.

ent positions of the kernel's surface (**Figure 12(b)**, **(c)** and **(d)**). As an example, the contaminated kernels can show a small localized contaminated area, a homogeneous contamination along the maize kernel or a medium to large contamination at the edge of the kernel. When considering the ratio of the integrated fluorescence spectra, specifically the ratio of the integrated fluorescence intensity between 475 and 550 nm and the integrated fluorescence intensity between 400 and 475 nm, a similar observation can be made (**Figure 13**). In correspondence with **Figure 11**, the high-contaminated areas show the largest ratios. Furthermore, the gradient of the contamination level is clearly visible. When considering **Figure 13(b)**, two highly contaminated areas surrounded by a medium-contaminated area can be identified, while **Figure 13(d)** shows a highly contaminated area at the edge that flows into a healthy area in the centre of the kernel. Concluding, localized contaminated areas can be accurately identified and characterized by mapping

316 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

**Figure 13.** Visualization of the localized aflatoxin contamination during OPIF, by mapping fluorescence ratios (integrated fluorescence intensity between 475 and 550 nm divided by the integrated fluorescence intensity between 400 and 475 nm) along the kernel's surface: (a) healthy kernel, (b) kernel with small localized contamination area, (c) kernel

Considering the TPIF measurements, the maize kernels could only be partially scanned with a lower resolution of 1 mm (instead of the 0.5 mm resolution used during the OPIF scanning measurements). When accurately scanning a complete maize kernel's surface during the TPIF measurements, the kernel's surface would be damaged, due to the required long illumination time and high illumination power density. Specifically, the illumination time is significantly

with homogeneous contamination and (d) kernel with a medium to large contamination at the edge.

the integral ratio on the screened kernel's surfaces.

**Figure 14.** Visualization of the localized aflatoxin contamination during TPIF, by mapping the dominant fluorescent emission wavelength along the kernel's surface: (a) low‐contaminated kernel and (b) high‐contaminated kernel.

The scanning plots demonstrate a successful monitoring of the localized aflatoxin contamina‐ tion. In contrast to the destructive chemical analyses that measure the mean contamination of a certain number of maize kernels, we are able to measure the localized contamination in individual unprocessed food products. By monitoring the localized contamination level of the kernels, we allow to sort‐out the maize kernels that contain a low mean contamination level but feature a small area with a high contamination level, which would not be detected by any type of chemical analyses. As a result, our developed optical detection methodology contrib‐ utes to an improved food safety.

## **4. Extensibility of our research methodology to the detection of ochratoxin and zearalenone**

Next to aflatoxins, also ochratoxin and zearalenone may be present in contaminated food products. Ochratoxins are mainly produced by the fungi *Penicillium verrucosum* and *Aspergillus ochraceus*, and occur most frequently in cereals, fruits and coffee beans. Cereal and cereal products are considered as the main contributors to the ochratoxin intake in Europe [27]. There exist five types of ochratoxins, of which only ochratoxin A and ochratoxin B occur naturally in food products. Particularly, ochratoxin A is considered as being the most common and most toxic ochratoxin [28]. The European regulations limit the maximum allowed ochratoxin A concentration from 2 to 15 ppb, depending on the commodity [12]. Zearalenone, on the other hand, is a mycotoxin that can be produced by several fungi of the genus *Fusarium*, like *Fusarium graminearum, Fusarium culmorum* and *Fusarium cerealis* [29]. It mostly occurs in cereals as barley, oats, wheat, rice and maize. The European limitation on the zearalenone concentration ranges between 20 and 400 ppb [12].

In this section, we investigate the fluorescence characteristics of ochratoxin and zearalenone and discuss the applicability of our developed measurement methodology onto these toxins. Specifically, we performed measurements on pure ochratoxin A and zearalenone powder, purchased from Sigma-Aldrich.

#### **4.1. Ochratoxin A detection**

Ochratoxin A shows theoretically a strong absorbance in-between 200 and 275 nm and around 335 nm, while its theoretical fluorescence emission wavelength is situated around 475 nm (**Figure 3(b)**). The chemical analyses of ochratoxin A by using liquid chromatography in combination with fluorescence detection is widely discussed in literature [30–33]. During the execution of the fluorescence analysis, excitation wavelengths of 330, 332, 333 and 390 nm are commonly used, while the fluorescent detection occurs at 440, 460 and 476 nm.

Based on the absorbance spectrum of ochratoxin, while keeping in mind the wavelengths that are available with our fluorescence measurement configuration, we excite the ochratoxin A powder with 244, 250, 260, 356, 365, 375 and 395 nm during the OPIF measurements (**Figure 15**). The 244, 250 and 260 nm wavelengths are generated with the third-harmonicgenerating crystal, while the 356, 365, 375 and 395 nm wavelengths are generated by the second-harmonic-generating crystal. To be able to compare the captured fluorescence intensities, we rescaled the fluorescence spectra to an excitation power density of 1 mW/mm2 for all excitation wavelengths (illumination powers used during the measurements varied between 60 and 450 mW, depending on the excitation wavelength). For all excitation wavelengths, the measured fluorescence spectra indicate a maximum around 480 nm. Furthermore, the highest fluorescence intensity is obtained after excitation with 356 nm. In correspondence with the theoretical absorbance spectrum, the fluorescence intensity decreases when tuning the excitation wavelength from 244 to 260 nm and from 356 to 395 nm.

To measure the TPIF spectrum of the ochratoxin A powder, we used an excitation wavelength of 710 nm (with an excitation power of 562 mW and an excitation spot diameter of 375 µm), since the OPIF measurements showed the strongest fluorescence intensities after excitation with 356 nm. The obtained TPIF spectrum occurs in the same wavelength region as the OPIF spectrum, with a maximum fluorescence intensity at 480 nm (**Figure 16(a)**). However, the TPIF intensity is 1000 times weaker than the OPIF intensity and shows a larger noise contribution. To experimentally validate that we are measuring the TPIF spectrum of ochratoxin A, the integrated fluorescence intensity, obtained by integrating the fluorescence spectrum between 400 and 600 nm, is studied as a function of the excitation power (**Figure 16(b)**). Similar as for aflatoxin B1, the integrated fluorescence intensity shows a quadratic dependence on the excitation power, corresponding with the non-linear behaviour of two-photon absorption.

**Figure 15.** OPIF spectra of ochratoxin A: (a) spectra after excitation with 244, 250, 260 nm and (b) spectra after excitation with 356, 365, 375 and 395 nm.

**Figure 16.** Study of TPIF of ochratoxin A: (a) TPIF spectrum after excitation with 710 nm and (b) quadratic relationship between the integrated fluorescence intensity and the excitation power, validating the two-photon absorption.

The fluorescence signal of the ochratoxin A powder shows a similar performance as the fluorescence of the aflatoxin B1 powder, but with excitation wavelengths of 356 and 710 nm during the OPIF and TPIF processes, respectively, instead of the 365 and 730 nm excitation wavelengths used during the aflatoxin characterization. As a result, the same research methodology as used during the aflatoxin detection on maize kernels can be applied to ochratoxin-contaminated food products.

#### **4.2. Zearalenone detection**

toxic ochratoxin [28]. The European regulations limit the maximum allowed ochratoxin A concentration from 2 to 15 ppb, depending on the commodity [12]. Zearalenone, on the other hand, is a mycotoxin that can be produced by several fungi of the genus *Fusarium*, like *Fusarium graminearum, Fusarium culmorum* and *Fusarium cerealis* [29]. It mostly occurs in cereals as barley, oats, wheat, rice and maize. The European limitation on the zearalenone concentration ranges

318 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

In this section, we investigate the fluorescence characteristics of ochratoxin and zearalenone and discuss the applicability of our developed measurement methodology onto these toxins. Specifically, we performed measurements on pure ochratoxin A and zearalenone powder,

Ochratoxin A shows theoretically a strong absorbance in-between 200 and 275 nm and around 335 nm, while its theoretical fluorescence emission wavelength is situated around 475 nm (**Figure 3(b)**). The chemical analyses of ochratoxin A by using liquid chromatography in combination with fluorescence detection is widely discussed in literature [30–33]. During the execution of the fluorescence analysis, excitation wavelengths of 330, 332, 333 and 390 nm are

Based on the absorbance spectrum of ochratoxin, while keeping in mind the wavelengths that are available with our fluorescence measurement configuration, we excite the ochratoxin A powder with 244, 250, 260, 356, 365, 375 and 395 nm during the OPIF measurements (**Figure 15**). The 244, 250 and 260 nm wavelengths are generated with the third-harmonicgenerating crystal, while the 356, 365, 375 and 395 nm wavelengths are generated by the second-harmonic-generating crystal. To be able to compare the captured fluorescence intensities, we rescaled the fluorescence spectra to an excitation power density of 1 mW/mm2

all excitation wavelengths (illumination powers used during the measurements varied between 60 and 450 mW, depending on the excitation wavelength). For all excitation wavelengths, the measured fluorescence spectra indicate a maximum around 480 nm. Furthermore, the highest fluorescence intensity is obtained after excitation with 356 nm. In correspondence with the theoretical absorbance spectrum, the fluorescence intensity decreases when tuning the excitation wavelength from 244 to 260 nm and from 356 to 395 nm.

To measure the TPIF spectrum of the ochratoxin A powder, we used an excitation wavelength of 710 nm (with an excitation power of 562 mW and an excitation spot diameter of 375 µm), since the OPIF measurements showed the strongest fluorescence intensities after excitation with 356 nm. The obtained TPIF spectrum occurs in the same wavelength region as the OPIF spectrum, with a maximum fluorescence intensity at 480 nm (**Figure 16(a)**). However, the TPIF intensity is 1000 times weaker than the OPIF intensity and shows a larger noise contribution. To experimentally validate that we are measuring the TPIF spectrum of ochratoxin A, the integrated fluorescence intensity, obtained by integrating the fluorescence spectrum between 400 and 600 nm, is studied as a function of the excitation power (**Figure 16(b)**). Similar as for

for

commonly used, while the fluorescent detection occurs at 440, 460 and 476 nm.

between 20 and 400 ppb [12].

purchased from Sigma-Aldrich.

**4.1. Ochratoxin A detection**

Zearalenone shows theoretically a strong absorbance between 200 and 350 nm and a maximum fluorescence intensity around 475 nm (**Figure 3(c)**). Similar to ochratoxin, the analysis of zearalenone by using high-performance liquid chromatography in combination with fluorescence detection is widely discussed in literature [34–37]. During the execution of the fluorescence step, an excitation wavelength of 254, 274, 275 or 356 nm is often used, while the detection of the fluorescence signal occurs at 440, 446 or 470 nm. Applying the same measurement procedure to the zearalenone powder as to the ochratoxin A, we studied the zearalenone OPIF signal after excitation with 244, 250, 260, 365, 366, 375 and 395 nm (**Figure 17**). Particularly, to compare the fluorescence intensity for all excitation wavelengths, we present the fluorescence spectrum rescaled to an excitation power density of 1 mW/mm2 . The zearalenone fluorescence spectrum reaches its maximum between 425 and 500 nm. Its fluorescence spectrum shows a broader peak than the aflatoxin and ochratoxin fluorescence spectra. Furthermore, the zearalenone fluorescence emission shows approximately a 10 times weaker intensity than for ochratoxin A. However, similar to the ochratoxin A fluorescence, the largest fluorescence intensities are obtained after excitation with 244 and 356 nm.

**Figure 17.** OPIF spectra of zearalenone: (a) spectra after excitation with 244, 250, 260 nm and (b) spectra after excitation with 356, 365, 375 and 395 nm.

To study the TPIF signal of zearalenone, we illuminated the zearalenone powder with 730, 750 and 780 nm, with a spot diameter of 231 µm and an excitation laser power of 1.10, 1.22 and 1.51 W, respectively (**Figure 18**). For all three excitation wavelengths, the measured spectra show two light signals: one at the excitation wavelength and one at half the excitation wavelength. The signal at the excitation wavelength represents the illumination laser light that reflects on the sample and reaches the spectrum analyser. Due to imperfections of the optical short-wave pass filter, small fractions of the high-power incident light are still able to reach the spectrum analyser. The signal at half the excitation wavelength, at 365, 375 and 385 nm for excitation with 730, 750 and 780 nm, respectively, represents SHG instead of TPIF. The absence of the TPIF signal can be explained by the study of the molecular structure of the zearalenone (**Figure 19**). Comparing the different molecular structures, zearalenone shows a significant different structure than aflatoxin B1 and ochratoxin A. Generally, molecules show a strong, two-photon absorption if they feature a long conjugation system in combination with strong donor and acceptor groups, because this induces non-linearity in the system and increases the potential for charge transfer [19]. Consequently, the molecular structure of zearalenone favours SHG instead of TPIF. In comparison with TPIF, the SHG shows a smaller linewidth. During SHG, only one wavelength will be emitted for each illumination wavelength, because two incident photons will always recombine to the double energy. In contrast, during the TPIF process, different emission wavelengths are captured, since the relaxation of the exited electrons can start from different excited energy levels, dependant on their vibrational energy loss. Furthermore, the stronger the captured illumination signal, the weaker the generated SHG signal. When illuminating the sample, part of the excitation photons recombine to the SHG signal, while another part of the excitation signal reflects on the samples surface. The higher the amount of photons that recombine, the stronger the SHG signal and the lower the reflection on the sample.

zearalenone by using high-performance liquid chromatography in combination with fluorescence detection is widely discussed in literature [34–37]. During the execution of the fluorescence step, an excitation wavelength of 254, 274, 275 or 356 nm is often used, while the detection of the fluorescence signal occurs at 440, 446 or 470 nm. Applying the same measurement procedure to the zearalenone powder as to the ochratoxin A, we studied the zearalenone OPIF signal after excitation with 244, 250, 260, 365, 366, 375 and 395 nm (**Figure 17**). Particularly, to compare the fluorescence intensity for all excitation wavelengths, we present the fluorescence

spectrum reaches its maximum between 425 and 500 nm. Its fluorescence spectrum shows a broader peak than the aflatoxin and ochratoxin fluorescence spectra. Furthermore, the zearalenone fluorescence emission shows approximately a 10 times weaker intensity than for ochratoxin A. However, similar to the ochratoxin A fluorescence, the largest fluorescence

**Figure 17.** OPIF spectra of zearalenone: (a) spectra after excitation with 244, 250, 260 nm and (b) spectra after excitation

To study the TPIF signal of zearalenone, we illuminated the zearalenone powder with 730, 750 and 780 nm, with a spot diameter of 231 µm and an excitation laser power of 1.10, 1.22 and 1.51 W, respectively (**Figure 18**). For all three excitation wavelengths, the measured spectra show two light signals: one at the excitation wavelength and one at half the excitation wavelength. The signal at the excitation wavelength represents the illumination laser light that reflects on the sample and reaches the spectrum analyser. Due to imperfections of the optical short-wave pass filter, small fractions of the high-power incident light are still able to reach the spectrum analyser. The signal at half the excitation wavelength, at 365, 375 and 385 nm for excitation with 730, 750 and 780 nm, respectively, represents SHG instead of TPIF. The absence of the TPIF signal can be explained by the study of the molecular structure of the zearalenone (**Figure 19**). Comparing the different molecular structures, zearalenone shows a significant different structure than aflatoxin B1 and ochratoxin A. Generally, molecules show a strong,

. The zearalenone fluorescence

spectrum rescaled to an excitation power density of 1 mW/mm2

320 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

intensities are obtained after excitation with 244 and 356 nm.

with 356, 365, 375 and 395 nm.

**Figure 18.** Second harmonic generation signal at 365, 375 and 385 nm, observed after illumination of the zearalenone powder with 730, 750 and 780 nm laser light, respectively. The inner graph presents a close-up on the weakest captured signals.

**Figure 19.** Chemical structure of (a) aflatoxin B1, (b) ochratoxin A and (c) zearalenone.

As a result, OPIF seems the most promising technique for the detection of zearalenone. In contrast to aflatoxin and ochratoxin, zearalenone does not show TPIF. Consequently, we consider TPIF as a less promising technique for the zearalenone detection. However, the influence of the zearalenone onto the intrinsic fluorescence of food products should be further investigated.

## **5. Conclusion**

We successfully demonstrated the use of one- and two-photon-induced fluorescence spectroscopy for the detection of fluorescent mycotoxins in solid, unprocessed food products. First, we developed a sensitive measurement configuration able to study the localized one- and twophoton-induced fluorescence spectra of food products. Afterwards, as a case study, we investigated the detection of aflatoxin in individual maize kernels. We presented our research methodology, starting from the characterization of the aflatoxin fluorescence, to the measurement of the one- and two-photon-induced fluorescence spectra of the maize kernels and the development of a spectroscopic detection criterion. During both one- and two-photon-induced fluorescence processes, the fluorescence of the aflatoxin influenced the intensity and wavelength of the intrinsic fluorescence of the maize kernels. Based on the fluorescence spectrum between 400 and 550 nm, we defined an optical detection criterion, indicating a maximal contrast between the healthy and contaminated maize kernels for excitation with 365 and 780 nm, during one- and two-photon-induced fluorescence, respectively. Both one- and twophoton-induced fluorescence processes show a similar contrast. However, two-photoninduced fluorescence requires higher excitation power densities and a more sensitive detector than one-photon-induced fluorescence, but uses NIR laser wavelengths that are widely commercially available. Besides, in contrast to the chemical analyses that investigate the mean contamination of a batch of maize kernels, we successfully monitored the localized contamination level on the maize kernel surfaces. Both kernels containing a small area with a high contamination level and kernels containing a large region with a medium contamination level could be identified. Finally, we discussed the extensibility of our research methodology to the detection of ochratoxin A and zearalenone. Ochratoxin A showed a characteristic one- and two-photon-induced fluorescence signal, while for zearalenone only OPIF could be observed. Generally, we can conclude that we demonstrated the use of fluorescence spectroscopy as a valuable tool for the sensitive optical detection of fluorescent mycotoxins, paving the way for a non-destructive, real-time and high-sensitive industrial scanning-based detection.

## **Acknowledgements**

This work was supported in part by FWO (G008413N), IWT (IWT120528), the COST Action (MP1205), the Methusalem and Hercules Foundations and the OZR of the Vrije Universiteit Brussel (VUB). The authors would also like to thank Dr Ir Alfons Callebaut and Dr Ir Bart Huybrechts of the CODA-CERVA, the National Reference Laboratory for Mycotoxins, for their valuable feedback and know-how about mycotoxins. We are very grateful to them for the execution of the liquid chromatography-tandem mass spectroscopy analytical technique. Parts of this chapter are reproduced from authors' recent conference publication, cited in reference number [26].

## **Author details**

As a result, OPIF seems the most promising technique for the detection of zearalenone. In contrast to aflatoxin and ochratoxin, zearalenone does not show TPIF. Consequently, we consider TPIF as a less promising technique for the zearalenone detection. However, the influence of the zearalenone onto the intrinsic fluorescence of food products should be further

322 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

We successfully demonstrated the use of one- and two-photon-induced fluorescence spectroscopy for the detection of fluorescent mycotoxins in solid, unprocessed food products. First, we developed a sensitive measurement configuration able to study the localized one- and twophoton-induced fluorescence spectra of food products. Afterwards, as a case study, we investigated the detection of aflatoxin in individual maize kernels. We presented our research methodology, starting from the characterization of the aflatoxin fluorescence, to the measurement of the one- and two-photon-induced fluorescence spectra of the maize kernels and the development of a spectroscopic detection criterion. During both one- and two-photon-induced fluorescence processes, the fluorescence of the aflatoxin influenced the intensity and wavelength of the intrinsic fluorescence of the maize kernels. Based on the fluorescence spectrum between 400 and 550 nm, we defined an optical detection criterion, indicating a maximal contrast between the healthy and contaminated maize kernels for excitation with 365 and 780 nm, during one- and two-photon-induced fluorescence, respectively. Both one- and twophoton-induced fluorescence processes show a similar contrast. However, two-photoninduced fluorescence requires higher excitation power densities and a more sensitive detector than one-photon-induced fluorescence, but uses NIR laser wavelengths that are widely commercially available. Besides, in contrast to the chemical analyses that investigate the mean contamination of a batch of maize kernels, we successfully monitored the localized contamination level on the maize kernel surfaces. Both kernels containing a small area with a high contamination level and kernels containing a large region with a medium contamination level could be identified. Finally, we discussed the extensibility of our research methodology to the detection of ochratoxin A and zearalenone. Ochratoxin A showed a characteristic one- and two-photon-induced fluorescence signal, while for zearalenone only OPIF could be observed. Generally, we can conclude that we demonstrated the use of fluorescence spectroscopy as a valuable tool for the sensitive optical detection of fluorescent mycotoxins, paving the way for

a non-destructive, real-time and high-sensitive industrial scanning-based detection.

This work was supported in part by FWO (G008413N), IWT (IWT120528), the COST Action (MP1205), the Methusalem and Hercules Foundations and the OZR of the Vrije Universiteit Brussel (VUB). The authors would also like to thank Dr Ir Alfons Callebaut and Dr Ir Bart Huybrechts of the CODA-CERVA, the National Reference Laboratory for Mycotoxins, for their

investigated.

**5. Conclusion**

**Acknowledgements**

Lien Smeesters\* , Wendy Meulebroeck and Hugo Thienpont

\*Address all correspondence to: lsmeeste@b-phot.org

Department of Applied Physics and Photonics (TONA), Brussels Photonics Team (B-PHOT), Faculty of Engineering, Vrije Universiteit Brussel, Brussels, Belgium

## **References**


[22] Held P. Quantitation of Peptides and Amino Acids with a SynergyTM HT using UV Fluorescence. BioTek Instruments. 2006. p. 1–8.

[8] Nawrocka A, Lamorska J. Determination of Food Quality by Using Spectroscopic Methods. In: Grundas S, Stepniewski A, editors. Advances in Agrophysical Research.

324 Applications of Molecular Spectroscopy to Current Research in the Chemical and Biological Sciences

[9] Rasch C, Kumke M, Löhmannsröben HGG. Sensing of Mycotoxin Producing Fungi in the Processing of Grains. Food Bioprocess Technol [Internet]. Springer New York; 2010 Apr 29 [cited 2013 Jan 31];3(6):908–16. Available from: http://dx.doi.org/10.1007/

[10] Rasch C, Böttcher M, Kumke M. Determination of aflatoxin B1 in alcoholic beverages: Comparison of one‐ and two‐photon‐induced fluorescence. Anal Bioanal Chem.

[11] Hruska Z, Yao H, Kincaid R, Darlington D, Brown RL, Bhatnagar D, et al. Fluorescence imaging spectroscopy (FIS) for comparing spectra from corn ears naturally and

[12] Romer Labs. Mycotoxin regulations [Internet]. 2012. p. 1–5. Available from: http:// www.romerlabs.com/en/knowledge/mycotoxin‐regulations/regulations‐usa/

[13] European Commission. Commission regulation (EC) No 1881/2006 setting maximum levels for certain contaminants in foodstuffs. Official Journal of the European Union.

[14] Lakowicz JR. Principles of Fluorescence Spectroscopy. Third Edit. New York: Springer

[15] Kaiser W, Garrett CGB. Two‐Photon Excitation in Ca:F2 Eu2+. Phys Rev Lett. 1961;7(6):

[16] Benninger RKP, Piston DW. Two‐Photon Excitation Microscopy for the Study of Living Cells and Tissues. In: Current protocols in cell biology. Canada: John Wiley & Sons, Inc.;

[17] Mulligan S, MacVicar B. Two‐photon fluorescence microscopy: basic principles, advantages and risks. Mod Res Educ Top Microsc [Internet]. 2007;881–8. Available

[18] Svoboda K, Yasuda R. Principles of Two‐Photon Excitation Microscopy and Its Appli‐

[19] Pawlicki M, Collins HA, Denning RG, Anderson HL. Two‐photon absorption and the

[20] Krueger A. MULTIPHOTON MICROSCOPY?: Turnkey femtosecond lasers fuel

[21] Avantes. Irradiance. In: AvaSoft for AvaSpec‐USB2 User's Manual. Avantes; 2012. p. 63–

design of two‐photon dyes. Angew Chemie–Int Ed. 2009;48(18):3244–66.

growth of multiphoton imaging. LaserFocusWorld. 2010;46(10):1–7.

from: http://formatex.org/microscopy3/pdf/pp881‐889.pdf

cations to Neuroscience. Neuron. 2006;50(6):823–39.

artificially infected with aflatoxin producing fungus. J Food Sci. 2013;78(8).

Rijeka: InTech; 2013. p. 347–68.

s11947‐010‐0364‐y

2010;397(1):87–92.

Brussels; 2006 Dec.

2013. p. 4.11.1–4.11.24.

229–31.

4.

Science + Business Media; 2006. 607‐613 p.

