Rapid Determination of Sugar and Salt Levels in Snack Products

Using Infrared Spectroscopy

Thesis

Presented in Partial Fulfillment of the Requirements for the Degree

Master of Sciencein the Graduate School of The

Ohio State University

By

Ting Wang

Graduate Program in Food Science and Technology

The Ohio State University

2011

Master’s Examination Committee:

Dr. Luis Rodriguez-Saona, Advisor

Dr. W. James Harper

Dr. Mike Mangino

Abstract

Fourier transform infrared (FT-IR) spectroscopy is a simple, rapid, and

reliable technique that has been widely researched and applied for analysis of

food components, which can identify types of chemical bonds in a molecule by

producing an infrared absorption spectrum that is like a molecular "fingerprint".

FTIR combined with chemometrics, including SIMCA and PLS, makes it a

strong technique to identify unknown materials, determine the quality or

consistency of a sample, and determine the amount of components in a

mixture.

The world snack food market reached an estimated $66 billion in 2003,

with the United States accounting for about a third of it. However the sensor

devices currently available are not suitable in a food processing operation.

There is an urging demand for rapid, high through-put, and real-time

instruments for the food industry. The objective of this research was to

evaluate the feasibility of infrared systems combined with chemometric

analysis for rapid quality assessment in applications to Ohio’s snack industry.

Snack products samples were grounded and spectra were collected using

MIR and NIR equipped with 3-bounce ZnSe MIRacle attenuated total

reflectance accessory or a diffuse reflection-integrating sphere separately.

Partial Least Square Regression (PLSR) was used to construct prediction

ii

models. The spectra showed characteristic absorption bands for sugars (MIR

and NIR) and the salt-induced changes in the NIR water spectrum as affected

by salt content. Best performance PLSR models for salt content were

generated from NIR reflectance (SECV=0.016% and r = 0.96) whereas similar

statistics were obtained for sugar levels (SEP ~0.02mg/g cereal and r > 0.98).

FTIR combined with PLSR analysis provided a rapid quality assessment

technique, which can finally improve the quality control of snack food industry

by saving testing time, decreasing investment and providing reliable results.

iii

Dedication

Dedicated to my mom

iv

Acknowledgments

I would like to thank Dr. Luis Rodriguez-Saona for his excellent

guidance and endless patience. Without your encouragement and

help, this would be impossible.

I would like to thank Dr. W. James Harper and Dr. Mike Mangino

for their consideration and guidance.

I would like to thank the warm and sweet laboratory group for

the help they gave me during the past three years.

I would like to thank my family for their support of every thing in

my life, in every minute.

I would like to specially thank my husband for always standing

beside me, giving me happiness, and love.

v

Vita

2005............................................. B.S. Bioengineering, South

China University of

Technology

2008............................................. M.S. Food Science and

Technology, The Ohio State

University

2009 to present ...............................Graduate Research Associate,

The Ohio State University

Fields of Study

Major Field: Food Science and Technology

vi

Table of Contents

Abstract .....................................................................................................ii

Dedication .................................................................................................iii

Acknowledgments......................................................................................iv

Vita ............................................................................................................v

List of Tables..............................................................................................vii

List of Figures.............................................................................................viii

Chapter 1: Literature Review......................................................................1

Chapter 2: Rapid Determination of Sugar and Salt Levels in Snack Products

Using Infrared Spectroscopy................................................... 30

References ................................................................................................60

vii

List of Tables

Table 1.1 UV absorbance cutoffs of various common solvents....................17

Table 1.2 Three different types of cuvettes commonly used with different

usable wavelengths......................................................................18

Table 2.1 Assignment of Bands in the NIR and MIR-FT Spectra..................43

Table 2.2 PLSR model results for salt, moisture and fat content prediction with

NIR and MIR................................................................................44

Table 2.3 Sugar contents in cereal tested by enzymatic assay kit and HPLC

method...........................................................................................49

Table 2.4 Average sugar content in coated and uncoated base cereal

according to type of flavor............................................................. 50

Table 2.5 Performance statistics for PLSR calibration and validation

models............................................................................................51

viii

ix

List of Figures

Figure 1.1 Instrumentation of high performance liquid chromatography........12

Figure 1.2 Enzymatic determination of sucrose.............................................20

Figure 2.1 Typical Near infrared (A) and ATR-Mid infrared Spectrum of tortilla

chip samples..................................................................................40

Figure 2.2 Partial Least Squares Regression (PLSR) cross-validation plots for

salt, moisture, and oil determination in tortilla chip samples from

Wyandot Inc................................................................................. 46

Figure 2.3 Partial Least Squares Regression (PLSR) cross-validation plots for

sugar determination in cereal samples from Wyandot Inc........... 53

Figure 2.4 Partial Least Squares loading spectra for the first 2 latent variables

for the calibration set measured by FT-NIR reflectance and

ATR-MRI spectroscopy.................................................................55

CHAPTER 1

LITERATURE REVIEW

1.1 Snack food

The world snack food market reached an estimated $66 billion in 2003,

with the United States accounting for about a third of it (Anonymous 2004). As

lifestyles in other parts of the world become more westernized, the global

demand for snack foods continues to increase and evolve (Nikolaou 2006).

According to the U.S. Census Bureau, “the snack food industry comprises

establishments primarily engaged in one or more of the following: (1) salting,

roasting, drying, cooking, or canning nuts; (2) processing grains or seeds into

snacks; (3) manufacturing peanut butter; and (4) manufacturing potato chips,

corn chips, popped popcorn, pretzels (except soft), pork rinds, and similar

snacks (U.S. Census Bureau 2002).” The snack food industry sector also

includes “consumer-ready packaged chocolate and non-chocolate candies,

cookies and crackers, unpopped popcorn and meat snacks (Anonymous

2004).” The list of snack food products is growing steadily, as competition, new

knowledge in nutrition science (Willett and others 2002; Anonymous 2004)

regulatory mandate, and self-imposed guidelines for healthful public nutrition

force companies to introduce snacks with refined features, such as new raw

material basis, improved texture, shape, color, flavor, and nutritional content

1

(Shukla 1994; 2000). The last factor, in particular, has had a significant effect in

recent years, as Americans’ dietary habits fluctuate, in response to scientific

advances, governmental guidelines, commercial offerings, social fads or

personal preferences (Nikolaou 2006).

Along with technology for the manufacture of snack foods, food science is

obviously of paramount importance for this field. By including elements of

physical, chemical and biological sciences, food science is a truly

interdisciplinary field of remarkable complexity and considerable breadth, and

has frequently contributed, rather than simply benefited from its constitutive

scientific fields (Nikolaou 2006).

The technology for the manufacture of snack foods involves a relatively

small number of processes that appear most frequently in snack food

manufacturing plants, including extrusion, frying, baking, and drying (Nikolaou

2006). While the above processes constitute the heart of most snack food

manufacturing lines, where practically all cooking occurs, there are additional

processes and equipment that are important for the operation of an entire line.

For example, oil, powder, and granule applicators, including oil and cheese

sprayers, powder dispensers, electrostatic salters, and coating tumblers. This

equipment is used for snack flavoring, often creating a variety of products from

the same basic substrate (Nikolaou 2006).

One of the main issues associated with the snack foods is the health

concern of adding abundant sugar and salt. When it comes to the American

2

diet, people tend to get too much of the two ingredients. A high-sugar,

high-calorie diet can lead to obesity and to diabetes, both of which put people

at greater risk for heart disease (Roberts 2009). Many studies show a direct

link between sodium intake and high blood pressure (Roberts 2009); the more

salt was consumed, the higher the blood pressure will be (Roberts 2009). This

is because the kidneys cannot process and eliminate salt beyond a certain

amount, which influences blood pressure in the arteries (Roberts 2009). If not

properly controlled, high blood pressure can lead to kidney failure as well as

heart failure, heart attack, and stroke (Roberts 2009).

1.2 Cereal

The grass family has five subfamilies with approximately 10,000 species

(Chapman 1996). Grass species can be found living in a wide range of

altitudes (edge of the sea to 1000's of feet above sea level), temperatures

(equatorial belt to arctic and Antarctic regions), moisture conditions, soil types

and salinity concentrations (Arber 1934; Pool 1948). The grass family offers

many important economic and ecological features including food, forage, cane

sugar, fiber (paper and rope), renewable energy, habitat, development of

ecologically disturbed areas (prevention of soil erosion, extraction of salt and

bioremediation), timber, lawns, ornamentals, and aromatic oils (Arber 1934;

Pool 1948; Chapman 1996).

Cereals such as rice, wheat and maize are members of the grass family

and they are particularly important to humans because of their role as staple

3

food crops in many areas of the world. Cereals are also used to produce

animal feed, oils, starch, flour, sugar, syrup, processed foods, malt, alcoholic

beverages, gluten and renewable energy (Pomeranz and others 1981).

Approximately 50% of the world's calories are provided by rice, wheat and

maize, but in many parts of Africa and Asia, people rely mainly on grains such

as sorghum or millet. Maize, sorghum and barley are important sources of

livestock feed and barley and rice are used in the brewing industry (Chopra

and others 2002). The United States Department of Agriculture recommends

that Americans eat 6-11 servings of grains per day (United States Department

of Agriculture 1996). Grains of all kinds, oats in particular, have been shown to

contain chemical compounds that help to reduce the chance of certain types of

cancer and coronary heart disease (Marquart and others 2002).

Cereal processing technology constitutes a strategic industry worldwide

delivering a range of products from finished items, to raw materials used in

baking, brewing, etc (Solís-Morales and other 2009). An important sector of

the cereal processing industry is formed by the products generally known as

Ready To Eat (RTE breakfast) cereals (Solís-Morales and other 2009). It is a

food made more or less from processed grains often, but not always, eaten

with the first meal of the day or sometimes as a snack later in the day. It is

often eaten cold, usually mixed with milk, water, or yogurt, and sometimes fruit

but sometimes eaten dry. RTE cereals have their beginnings in the vegetarian

movement in the last quarter of the nineteenth century, which influenced

4

members of the Seventh-day Adventist Church in the United States (The

History of Breakfast Cereals 2008).

RTE cereals may be broadly classified into five categories: flaked, puffed,

extruded, whole grain and granolas (Valentas and others 1991). To produce

RTE cereals several unit operations are employed (Solís-Morales and other

2009). In most processes raw grains are first cooked in some manner to

gelatinise the starches present (Culbertson 2004). The grains are then

flattened (flaked), formed (extruded), shredded or expanded (puffed) (Solí

s-Morales and other 2009). Texture in RTE cereals is fundamental for product

acceptance by consumers (Solís-Morales and other 2009). Crisp and crunchy

textures are generally expected, which can be imparted to raw materials in a

number of ways (Solís-Morales and other 2009). Texturising (i.e. controlled

change of texture) processes will provide crispness or crunchiness by

changing, basically, the surface/volume ratio and thickness (Solís-Morales

and other 2009). Texture alteration can be achieved by different operations

such as flaking, puffing or extruding (Solís-Morales and other 2009). Rice and

wheat are the two main cereals whose granules can be processed individually

in order to expand them or inflate them, as to obtain the RTE products known

as puffed cereals (Fast 2000; 2001). Some other grain formulations can be

also expanded to obtain puffed products using extrusion technology (Sevatson

others 2000).

A number of RTE cereals are covered with coats aimed at improving

5

functionality in some manner (Solís-Morales and other 2009). Coating of foods,

also called encapsulation, can be defined as the application of a liquid and/or a

powder in a required pattern onto a food piece in order to enhance

preservation, palatability, nutritional value or some other specific property

(Denoyelle 2008). Coating or encapsulation of food pieces, granules or

particulates, can be carried out by a number of unit operations (Solís-Morales

and other 2009). Some of the operations used for mixing or dry powders can

be equally applied for coating (Solís-Morales and other 2009). Equipment that

can be used for the purpose of mixing powders/coating food pieces include

rotating drums, ribbon mixers and fluidised beds (Ortega- Rivas 2005).

Edible coating has been used in RTE puffed cereals to obtain advantages

such as to enhance flavor, protect the food particulate from oxygen and

moisture, control the release of active ingredients or improve quality

(Torres-Martínez 2001). RTE puffed wheat is normally coated with sweet

coverings to make it more appealing to different tastes and consumers (Solí

s-Morales and other 2009). The coating may also be added with colored

ingredients, such as chocolate or strawberry color (Solís-Morales and other

2009). A typical method to apply sweet coatings in puffed wheat is by using

rotating drums (Solís-Morales and other 2009). Normally, the puffed granules

of wheat are loaded into a tumbling, rotating heated vessel of conical shape

(Solís-Morales and other 2009). The syrup mostly used is prepared from

sucrose, and it may be simply poured onto the puffed granules within the

6

rotating vessel (Solís-Morales and other 2009). In powder coating, the size,

density and charge of particles and air velocity have a significant effect on

coating transfer efficiency and evenness ((Chen and others 1995; Mazumder

and others 1997). Report shows that transfer efficiency increased as particle

size, density or charge increased or as air velocity decreased (Yousuf and

others 2007). Evenness increased as particle size or density decreased, as

velocity increased, or as charge increased to a peak (Yousuf and others 2007).

Electrostatic coating produced higher transfer efficiency and evenness than

nonelectrostatic coating (Yousuf and others 2007).

Since sucrose is normally used as the coating material, it is necessary to

measure the amount remaining on the food after coating processing. Sensory

assessment and instrumental analysis are two methods mostly conducted by

the food industry for quality control purpose.

Sensory evaluation has been defined as a scientific method used to evoke,

measure, analyze and interpret those responses to products as perceived

through the senses of sight, smell, touch, taste, and hearing (Stone and others

1993). The most important criteria that define the quality of a sensory program

are: the training/qualifications of panelists, the type of established

specifications, and the use of controlled test conditions (Mun˜oz 2002). The

quality of a sensory program is to be questioned if any of these components is

only marginal or inappropriate (Mun˜oz 2002). While the sensory program

delivers strong ability to detect sensory problems/issues and control the quality

7

of food products in a comprehensive way, it might show the disadvantages of

time consuming, high expenses, and well trained panelists requirement. Some

companies have preferred to establish analytical/instrumental based sensory

programs to reduce the uncertainty and risks in decision making.

1.3 Instrumental analysis

1.3.1 HPLC

High performance liquid chromatography (HPLC) is today the premier

technique for chemical analysis and related applications, with an ability to

separate, analyze, and/or purify virtually any sample (Snyder and others 2009).

Compared to other separation procedures, HPLC is exceptional in terms of the

following characteristics:

1. Almost universal applicability; Few samples are excluded from the

possibility of HPLC separation.

2. Remarkable assay precision (0.5% or better in many cases)

3. A wide range of equipment, columns, and other materials is commercially

available, allowing the use of HPLC for almost every application.

4. Most laboratories that deal with a need for analyzing chemical mixtures are

equipped for HPLC; It is often the first choice of technique (Snyder and

others 2009).

The underlying principle of chromatography is that molecules not only

dissolve in liquids but can also "dissolve" (absorb) on to or interact with the

surface of solids (Bird 1989). If a molecule dissolved in a liquid is passed down

8

a column of solid particles with which it interacts it will move more slowly than

the solvent, spending some time dissolved in the liquid (liquid phase) and

some on the solid (solid phase) (Bird 1989). As such molecules take longer to

pass down the column than molecules that do not interact with the solid phase

they emerge from the bottom of the column later (Bird 1989). The separation of

molecules by chromatography depends on exploiting the subtle differences in

certain physical properties of the molecules in samples: their solubility in water,

solubility in organic solvents, net positive or negative charge, and size (Bird

1989).

Depending on the HPLC mode, the different types of the adsorption forces

may be included in the retention process: Hydrophobic (non-specific)

interactions are the main ones in reversed-phase (RP) separations

(Anonymous 2011). Dipole-dipole (polar) interactions are dominant in normal

phase (NP) (Anonymous 2011). Ionic interactions are responsible for the

retention in ion-exchange chromatography (Anonymous 2011). All these

interactions are competitive (Anonymous 2011). Analyte molecules are

competing with the eluent molecules for the adsorption sites (Anonymous

2011). So, the stronger analyte molecules interact with the surface, the weaker

the eluent interaction, the longer the analyte will be retained on the surface

(Anonymous 2011). SEC (size-exclusion chromatography) is another case. It

is the separation of the mixture by the molecular size of its components

(Anonymous 2011). The basic principle of SEC separation is that the bigger

9

the molecule, the less possibility there is for it to penetrate into the adsorbent

pore space (Anonymous 2011). So, the bigger the molecule the less it will be

retained (Anonymous 2011).

HPLC instrumentation includes a pump, injector, column, detector and data

system (Anonymous 2011). The heart of the system is the column where

separation occurs (Anonymous 2011). Since the stationary phase is composed

of micrometre size porous particles, a high pressure pump is required to move

the mobile phase through the column (Anonymous 2011). The

chromatographic process begins by injecting the solute onto the top of the

column (Anonymous 2011). Separation of components occurs as the analytes

and mobile phase are pumped through the column (Anonymous 2011).

Eventually, each component elutes from the column as a narrow band (or peak)

on the recorder. Detection of the eluting components is important, and this can

be either selective or universal, depending upon the detector used

(Anonymous 2011). The response of the detector to each component is

displayed on a chart recorder or computer screen and is known as a

chromatogram (Anonymous 2011). To collect, store and analyse the

chromatographic data, computer, integrator, and other data processing

equipment are frequently used (Anonymous 2011).

10

Figure1.1 Instrumentation of High Performance Liquid Chromatography (Clark

2007).

Advances in high performance liquid chromatography (HPLC) systems

have made this technique the method of choice for many researchers

interested in sugar determinations (Conrad and others 1976; Binder 1980).

Picha conducted the quantitative analysis of sugars in raw and baked

sweet potato roots using HPLC (Picha 1985). The major sugars in raw roots:

sucrose, glucose, and fructose, were quantitated with a cation exchange resin

column (Aminex HPX-87C) heated to 75°C using HPLC-grade Hz0 as the

mobile phase and refractive index detection (Picha 1985). The major sugars in

baked roots (convection oven or microwaved) were maltose, sucrose, glucose,

and fructose (Picha 1985). They were quantitated with an aminopropyl bonded

phase column (Bio-Sil Amino 5s) at 21°C using a 70% CHaCN:30% Hz0

mobile phase and refractive index detection (Picha 1985).

11

Lee and others conducted a research to determine the total sugar content

in onions, using an HPLC system with a refractive index detector (LC 25) and a

700CH carbohydrate column (Alltech, Deerfield, Ill., U.S.A.) with a guard

cartridge (Lee and others 2009). Total sugar content was calculated by

summing the sucrose, glucose, and fructose contents. They also identified

glucose as the major sugar content in onions (Lee and others 2009).

Another research focused on the detection of apple juice concentration

(Gomis and others 2004). HPLC method for sugar analyses in cider was used

in order to detect the presence of apple juice concentrate (Gomis and others

2004). Sugars were analyzed by reversed-phase liquid chromatography using

a C8 column and a mobile phase of citrate buffer pH

5.5/tetrahydrofuran/acetonitrile, operated in gradient mode (Gomis and others

2004).

Aminex HPLC columns are one type of normally used analytical columns

for sugar identification. Recent literature suggests that the Aminex

polystyrene-divinylbenzene ion exchange resins are applicable to a wide

variety of water soluble and partially water soluble small organic compounds.

The columns are packed with a polymer-based matrix, offer many advantages

for the analysis of carbohydrates, alcohols, and organic acids in food and

beverage. These columns allow the use of simple isocratic methods, eluting

with water or dilute acid. There is minimal sample preparation, usually just

filtering through a 0.45μm filter with no derivatization necessary.

12

1.3.2 Ultraviolet-visible spectroscopy

Ultraviolet-visible (UV-vis) spectroscopy is used to obtain the absorbance

spectra of a compound in solution or as a solid (Sibilia 1996; Harris 2007;

Skoog and others 2007). What is actually being observed spectroscopically is

the absorbance of light energy or electromagnetic radiation, which excites

electrons from the ground state to the first singlet excited state of the

compound or material (Sibilia 1996; Harris 2007; Skoog and others 2007). The

UV-vis region of energy for the electromagnetic spectrum covers 1.5 - 6.2 eV

which relates to a wavelength range of 800 - 200 nm (Sibilia 1996; Harris 2007;

Skoog and others 2007). The Beer-Lambert Law is the principle behind

absorbance spectroscopy (Sibilia 1996; Harris 2007; Skoog and others 2007).

For a single wavelength, A is absorbance (unitless, usually seen as arbitrary

units), ε is the molar absorptivity of the compound or molecule in solution

(M-1cm-1), b is the path length of the cuvette or sample holder (usually 1 cm),

and c is the concentration of the solution (M) (Sibilia 1996; Harris 2007; Skoog

and others 2007).

A = εbc

UV-vis spectroscopic data can give qualitative and quantitative information

of a given compound or molecule (Sibilia 1996; Harris 2007; Skoog and others

2007). Irrespective of whether quantitative or qualitative information is required

it is important to use a reference cell to zero the instrument for the solvent the

compound is in (Sibilia 1996; Harris 2007; Skoog and others 2007). For

13

quantitative information on the compound, calibrating the instrument using

known concentrations of the compound in question in a solution with the same

solvent as the unknown sample would be required (Sibilia 1996; Harris 2007;

Skoog and others 2007). If the information needed is just proof that a

compound is in the sample being analyzed, a calibration curve will not be

necessary; however, if a degradation study or reaction is being performed, and

concentration of the compound in solution is required, thus a calibration curve

is needed (Sibilia 1996; Harris 2007; Skoog and others 2007).

UV/Vis spectroscopy is routinely used in analytical chemistry for

the quantitative determination of different analytes, such as transition metal

ions, highly conjugated organic compounds, and biological macromolecules.

Solutions of transition metal ions can be colored and absorb visible

light. Organic compounds, especially those with a high degree of conjugation,

also absorb light in the UV or visible regions of the electromagnetic spectrum.

The solvents for these determinations are often water for water soluble

compounds, or ethanol for organic-soluble compounds.

UV-visible spectroscopy does have some limitations when applied to the

practical experiment.

UV-vis spectroscopy works well on liquids and solutions, but if the sample

is more of a suspension of solid particles in liquid, the sample will scatter the

light more than absorb the light and the data will be much skewed (Sibilia 1996;

Harris 2007; Skoog and others 2007). Most UV-vis instruments can analyze

14

solid samples or suspensions with a diffraction apparatus, but this is not

common (Sibilia 1996; Harris 2007; Skoog and others 2007). UV-vis

instruments generally analyze liquids and solutions most efficiently (Sibilia

1996; Harris 2007; Skoog and others 2007).

A blank reference will be needed at the very beginning of the analysis of

the solvent to be used (water, hexanes, etc), and if concentration analysis

needs to be performed, calibration solutions need to be made accurately

(Sibilia 1996; Harris 2007; Skoog and others 2007). If the solutions are not

made accurately enough, the actual concentration of the sample in question

will not be accurately determined (Sibilia 1996; Harris 2007; Skoog and others

2007).

Every solvent has a UV-vis absorbance cutoff wavelength. The solvent

cutoff is the wavelength below which the solvent itself absorbs all of the light

(Sibilia 1996; Harris 2007; Skoog and others 2007).

15

Table 1.1 UV absorbance cutoffs of various common solvents (Sibilia 1996;

Harris 2007; Skoog and others 2007).

Solvent UV absorbance cutoff (nm)

Acetone 329

Benzene 278

Dimethylformamide (DMF)267

Ethanol 205

Toluene 285

Water 180

The material the cuvette (the sample holder) is made from will also have a

UV-vis absorbance cutoff (Sibilia 1996; Harris 2007; Skoog and others 2007).

Glass will absorb all of the light higher in energy starting at about 300 nm, so if

the sample absorbs in the UV, a quartz cuvette will be more practical as the

absorbance cutoff is around 160 nm for quartz (Sibilia 1996; Harris 2007;

Skoog and others 2007).

16

Table 1.2 Three different types of cuvettes commonly used with different

usable wavelengths (Sibilia 1996; Harris 2007; Skoog and others 2007).

Material Wavelength range (nm)

Glass 380 - 780

Plastic 380 - 780

Fused quartzbelow 380

To obtain reliable data, the peak of absorbance of a given compound

needs to be at least three times higher in intensity than the background noise

of the instrument (Sibilia 1996; Harris 2007; Skoog and others 2007).

Obviously using higher concentrations of the compound in solution can combat

this (Sibilia 1996; Harris 2007; Skoog and others 2007). Also, if the sample is

very small and diluting it would not give an acceptable signal, there are

cuvettes that hold smaller sample sizes than the 2.5 mL of a standard cuvettes

(Sibilia 1996; Harris 2007; Skoog and others 2007). Some cuvettes are made

to hold only 100 μL, which would allow for a small sample to be analyzed

without having to dilute it to a larger volume, lowering the signal to noise ratio

(Sibilia 1996; Harris 2007; Skoog and others 2007).

Sugars can be measured by UV-visible spectroscopy via the aids of single

or a series of enzyme catalyzed reactions. Enzymes, as analytical tools, have

found widespread use in the food, biochemical, and pharmaceutical industries

17

(Bergmeyer and others 1974). Enzymatic methods are specific, reproducible,

sensitive, rapid, and therefore, idea for analytical purpose (Bergmeyer and

others 1974). Due to the high specificity and sensitivity of enzymes,

quantitative assays may be done on crude materials with little or no sample

preparation (Bergmeyer and others 1974). In the quantitative, enzymatic

determination of sucrose in food and other materials, Sucrose is hydrolyzed to

glucose and fructose by invertase (Bergmeyer and others 1974). Glucose and

fructose are phosphorylated by adenosine triphosphate (ATP) in the reaction

catalyzed by hexokinase (Bergmeyer and others 1974). Glucose -6-phosphate

(G6P) is then oxidized to 6-phosphogluconate in the presence of nicotinamide

adenine dinucleotide (NAD) in a reaction catalyzed by glucose-6-phosphate

dehydrogenase (G6PDH) (Bergmeyer and others 1974). During this oxidation,

an equimolar amount of NAD is reduced to NADH (Bergmeyer and others

1974). The consequent increase in absorbance at 340nm is directly

proportional to sucrose concentration (Bergmeyer and others 1974).

Sucrose Glucose + fructose

Glucose +2ATP+Fructose Glucose-6-phosphate + Fructose-6-phosphate+2ADP

G6P+NAD NADH + 6-Phosphogluconate

Invertase

Hexokinase

G6PDH

Sucrose Glucose + fructose

Glucose +2ATP+Fructose Glucose-6-phosphate + Fructose-6-phosphate+2ADP

G6P+NAD NADH + 6-Phosphogluconate

Invertase

Hexokinase

G6PDH

Figure 1.2 Enzymatic determination of sucrose (Bergmeyer and others 1974).

18

However in practical experiment, enzymatic assay of sugar does have

some disadvantages. Take sucrose enzymatic assay as an example, sucrose

needs to be extracted from solid samples, filter or deproteinize of extracted

solution is sometimes required to clarify; if glucose amount in the sample is too

large (ratio of glucose to sucrose is larger than 5:1), then the glucose must be

removed before the testing of sucrose.

1.3.3 Fourier transform infrared spectroscopy

Infrared spectroscopy has been a workhorse technique for materials

analysis in the laboratory for over seventy years (Anonymous 2001). An

infrared spectrum represents a fingerprint of a sample with absorption peaks

which correspond to the frequencies of vibrations between the bonds of the

atoms making up the material (Anonymous 2001). Because each different

material is a unique combination of atoms, no two compounds produce the

exact same infrared spectrum (Anonymous 2001). Therefore, infrared

spectroscopy can result in a positive identification (qualitative analysis) of

every different kind of material. In addition, the size of the peaks in the

spectrum is a direct indication of the amount of material present (Anonymous

2001). With modern software algorithms, infrared is an excellent tool for

quantitative analysis (Anonymous 2001).

Fourier transform infrared (FT-IR) spectroscopy is a simple, rapid, and

reliable technique that has been widely researched and applied for analysis of

food components (Subramanian and others 2009). It is based on the principle

19

that different chemical functional groups require different amounts of energy

(different wavelengths) for excitation (Subramanian and others 2009). Fourier

transform infrared spectroscopy is preferred over dispersive or filter methods

of infrared spectral analysis for several reasons:

1. It is a non-destructive technique

2. It provides a precise measurement method which requires no external

calibration

3. It can increase speed, collecting a scan every second

4. It can increase sensitivity – one second scans can be co-added together to

ratio out random noise

5. It has greater optical throughput

6. It is mechanically simple with only one moving part (Anonymous 2001).

FTIR spectroscopy has become an alternative technique for the analysis of

sugars in food samples, having the attractive features of being noninvasive

and potentially more rapid than the methods of HPLC and enzymatic analysis

(Duarte and others 2002).

FTIR was used for quantification of sugars in mango juices as a function of

ripening (Duarte and others 2002). Glucose, fructose, and sucrose are the

most abundant components of mango juice (Duarte and others 2002).

Throughout ripening, sucrose and fructose can be accurately quantified,

whereas the accuracy of glucose determination decreased at later stages,

when concentrations fell to 0.6-1.5 g/L (Duarte and others 2002). The bands

20

related to sucrose were successfully identified as 1149, 1080, 1062, 1024,

1008, 956, and 937 cm-1 (Duarte and others 2002). The researchers

suggested that FTIR can be used as a rapid indicator of ripening degree, which

would help to establish suitable conditions for transport and storage (Duarte

and others 2002).

Another experiment used FTIR as a replacement instrument to

Chromatography, which are time consuming and require skilled operators, for

the determination of beet medium invert sugar adulteration in three different

varieties of honey (Sivakesava and others 2001). The spectra in the region

between 950 and 1500 cm-1 were selected for the quantitative analysis and

classification of beet medium invert sugar, most likely due to sugar had highest

negative correlation in this region (Sivakesava and others 2001). The results

demonstrated the potential to distinguish different levels of beet sugar invert in

honey, which is a major problem in the honey industry (Sivakesava and others

2001).

21

1.4 References

1. Anonymous. 2001. Introduction to Fourier transform infrared Spectrometry.

Thermo nicolet corporation.

2. Anonymous. 2004. Global Snack Food Industry Trends. Snack Food &

Wholesale Bakery 93:64.

3. Anonymous. 2004. Snacking, the All-American pastime. Snack Food &

Wholesale Bakery 93:SI-34.

4. Anonymous. 2011. High performance liquid chromatography.

Standardbase techniques. Drenthe College. The Netherlands.

5. Arber A. 1934. The Gramineae: a study of cereal, bamboo, and grass.

University Press, Cambridge.

6. Bergmeyer HU, Bernt E. 1974. Methods of Enzymatic Analysis, 2nd edition.

Bergmeyer HU, eds., Academic Press (New York), 1177-1179.

7. Binder H. 1980. Separation of monosaccharides by high-performance

liquid chromatography: comparison of ultraviolet and refractive index

detection. J. Chromatog 189: 414.

8. Bird IM. 1989. High performance liquid chromatography: principles and

clinical applications. British Medical Journal 299:783-787.

9. Chapman GP. 1996. The Biology of Grasses. CAB International, UK.

10. Chen H, Gatlin B, Burnside G, Reddy RN, Kandarpa VK, Mazumder MK.

22

1995. A transfer efficiency model for powder coating process. In

Conference record of the 1995 IEEE 2:1280 –1283.

11. Chopra VL, Prakash S, eds. 2002. Evolution and Adaptation of Cereal

Crops. Science Publishers Inc, NH, USA.

12. Clark J. 2007. High performance liquid chromatography-HPLC. Chemguid.

Available at

http://www.chemguide.co.uk/analysis/chromatography/hplc.html.

13. Conrad EC, Palmer JK. 1976. Rapid analysis of carbohydrates by

high-pressure liquid chromatography. Food. Technol 30: 84.

14. Culbertson JD. 2004. Grain, cereal: ready-to-eat breakfast cereals. In:

Scott Smith, J., Hui, Y.H. (Eds.), Food Processing: Principles and

Applications. Blackwell Publishing, Oxford, UK, pp. 239–248.

15. Denoyelle L. 2008. Coating processes in food: a rational approach to the

selection and development of a process. AgroFood Industry Hi-Tech 19:28

–30.

16. Duarte IF, Barros AN, Delgadillo I, Almeida CU, Gil AM. 2002. Application

of FTIR spectroscopy for the quantification of sugars in mango juice as a

function of ripening. Journal of agricultural and food chemistry

50:3104-3111.

17. Fast RB. 2000. Manufacturing technology of ready-to-eat breakfast cereals.

In:Fast, R.B., Caldwell, E.F. (Eds.), Breakfast Cereals and How They are

Made. American Association of Cereal Chemists, Paul MN, USA, pp. 29–

23

34.

18. Fast RB. 2001. Breakfast cereals. In: Owen, G. (Ed.), Cereals Processing

Technology. Woodheal Publishing, Cambridge, UK, pp. 158–172.

19. Gomis DB, Tamayo DM, Valles BS, Mangas Alonso JJ. 2004. Detection of

apple juice concentrate in the manufacture of natural and sparkling cider

by means of HPLC chemometric sugar analyses. Journal of agricultural

and food chemistry 52:201-203.

20. Harris DC. 2007. Quantitative chemical analysis, 7th Ed. Freeman, New

York.

21. Lee EJ, Yoo KS, Jifon J, Patil BS. 2009. Characterization of shortday onion

cultivars of 3 pungency levels with flavor precursor, free amino acid, sulfur,

and sugar contents. Journal of food science 74:475-480.

22. Marquart L and others eds. 2002. Whole grain foods in health and disease.

American Association of Cereal Chemists, St. Paul, MN.

23. Mazumder MK, Wankum DL, Sims RA, Mountain JR, Chen H, Pettit P.

1997. Influence of powder properties on the performance of electrostatic

coating process. Journal of Electrostatics 40–41:369–374.

24. Mun˜oz AM. 2002. Sensory evaluation in quality control: an overview, new

developments and future opportunities. Food Quality and Preference

13:329–339.

25. Nikolaou M. 2006. Control of snack food manufacturing systems. Control

Systems, IEEE 26:40-53.

24

26. Ortega-Rivas E. 2005. Handling and processing of food powders and

particulates. In: Onwulata C. (Ed.), Encapsulated and Powdered Foods.

CRC Taylor & Francis, Boca Raton FL, USA, pp. 75–144.

27. Picha DH. 1985. HPLC determination of sugars in raw and baked sweet

potatoes. Journal of food science 50:1189-1190.

28. Pool R. 1948. Marching with the grasses. University of Nebraska Press,

USA.

29. Pomeranz Y, Munck L, eds. 1981. Cereals: A Renewable Resource.

American Association of Cereal Chemists. St. Paul, MN.

30. Roberts E. 2009. The Facts on Heart Disease, Sodium, and Sugar.

EverydayHealth.com.http://www.everydayhealth.com/heart-health/salt-and

-sugar.aspx

31. Sevatson E, Huber GR. 2000. Extruders in the food industry. In: Riaz, M.N.

(Ed.), Extruders in Food Applications. CRC Taylor & Francis, Boca Raton

FL, USA, pp. 167–204.

32. Shukla TP. 1994. Future snacks and snack food-technology. Cereal Foods

World 39:704-705.

33. Shukla TP. 2000. Modern snack foods. Cereal Foods World 45: 477-477.

34. Sibilia JP. 1996. Materials Characterization and chemical analysis, 2nd Ed.

Wiley-VCH, New York.

35. Sivakesava S, Irudayaraj J. 2001. Detection of inverted beet sugar

adulteration of honey by FTIR spectroscopy. Journal of the science of food

25

and agriculture 81:683-690.

36. Skoog DA, Holler FJ, Crouch SR. 2007. Principles of instrumental analysis,

6th Ed. Thomson brooks/cole.

37. Snyder LR, Kirkland JJ, Dolan JW. 2009. Introduction to Modern Liquid

Chromatography. New Jersey:John Wiley & Sons, Inc.

38. Solís-Morales D, Sáenz-Hernández CM, Ortega-Rivas E. 2009. Attrition

reduction and quality improvement of coated puffed wheat by fluidised bed

technology. Journal of Food Engineering 93:236-241.

39. Stone H, Sidel JL. 1993. Sensory evaluations practices, California:

Academic Press.

40. Subramanian A, Harper WJ, Rodriguez-Saona LE. 2009. Cheddar cheese

classification based on flavor quality using a novel extraction method and

Fourier transform infrared spectroscopy. Journal of dairy science

92:87–94.

41. The History of Breakfast Cereals. 2008.

Iventors.about.com. http://inventors.about.com/library/inventors/blcereal.ht

onomous

Available

m.

42. Torres-Mart í nez JL. 2001. Coating of Puffed Wheat Cereal with a

Chocolate Flavoured Cover Using Tumbling. MSc Thesis. Aut

University of Chihuahua (in Spanish), Chihuahua, Chih, Mexico.

43. U.S. Census Bureau. 2002. 2002 Economic Census Industry Series

Report.

26

at http://www.census.gov/epcd/ec97/industry/E311919.HTM.

44. United States Department of Agriculture, 1996. Food Guide Pyramid

ood Processing Operations and

e

cfm?articleID=000A97BD-80FC-1DF6-973

d electrostatic

powder coating. Journal of Food Engineering 83:550–561.

Booklet.

45. Valentas KJ, Levine L, Clark JP. 1991. F

Scale-up. Marcel Dekker, New York, NY.

46. Willett WC, Stampfer MJ. 2002. Rebuilding the Food Pyramid: Overview /

The Food Guide Pyramid (Sidebar). Availabl

at http://www.sciam.com/article.

809EC588EEDF&sc=I100322.

47. Yousuf S, Barringer SA. 2007. Modeling nonelectrostatic an

27

CHAPTER 2

Rapid Determinati Snack Products

Using Infrared Spectroscopy

Ting Wang and Luis E. Rodriguez-Saona

e and Technology

d Science and Technology

olumbus, OH, USA

on of Sugar and Salt Levels in

Department of Food Scienc

The Ohio State University

110 Parker Foo

2015 Fyffe Rd

C

28

2.1 Abstract

The US snack food industry generates billions of dollars in revenue each

year. There is a need for cutting edge sensor technology directed at improving

efficiency, throughput and reliability of critical processes. The aim of the

research was to evaluate the feasibility of infrared systems combined with

chemometric analysis for rapid quality assessment in applications to Ohio’s

snack industry. Samples were ground and spectra were collected using a

mid-IR spectrometer equipped with a 3-bounce ZnSe MIRacle attenuated total

reflectance accessory or FT-NIR system equipped with a diffuse

reflection-integrating sphere. Sugars (cereal) and salt (tortilla chips) content

were determined using a reference HPLC or titrimetric methods, respectively.

Partial least squares regression (PLSR) was used to create cross-validated

calibration models. The prediction ability of the models was evaluated on an

independent set of samples and compared with reference techniques. The

spectra showed characteristic absorption bands for sugars (MIR and NIR) and

the salt-induced changes in the NIR water spectrum as affected by salt content.

Best performance PLSR models for salt content were generated from NIR

reflectance (SECV=0.016% and r = 0.96) whereas similar statistics were

obtained for sugar levels (SEP ~0.02mg/g cereal and r > 0.98). Multivariate

models accurately and precisely predicted the salt and sugar levels in snacks

allowing for rapid (<1 min analysis time) analysis. This simple technique allows

29

for reliable prediction of quality parameters, and automation enabling food

manufacturers for early corrective actions that will ultimately save time and

lishing a uniform quality.

2.2

ironment, and regulatory requirements

(Ca

money while estab

Introduction

The snack food market alone is estimated to be roughly $20 billion (USD)

per year with the worldwide market approximately double this amount (U.S.

Census Bureau). Business challenges for food manufacturers are increasing

due to increased regulatory and monitoring requirements, increased economic

pressures, more stringent environmental regulations, and more discriminating

and changing consumer preferences. Cutting edge sensor technologies are

directed at improving efficiency, throughput and reliability of critical processes

(Holm 2003). While many Ohio food processors understand the value of

sensor technology and its potential economic benefits to their businesses, the

sensor devices currently available to them, as well as many under

development, will simply not work in a food processing operation. The food

industry has unique requirements that cannot be met by devices designed for

the processes and products of other industry fields. Food applications

introduce challenges related to unique performance, accuracy, safety, product

composition, processing env

selunghe and others 2000).

Methods available to the food industry are often time-consuming,

expensive, the precision is strongly dependent on the operation of skilled

30

personnel, and they are difficult to adapt into a quality assurance setting

(Halim and others 2006). In addition, unwanted interference from the sample

matrix often requires the use of extensive sample preparation and clean-up

(Halim and others 2006). Field-based devices can streamline quality

assurance, protecting consumers against the risk of purchasing and

consuming nutritionally inadequate, deceptively mislabeled or misbranded,

imp

fiths and others 1986; McKelvy and others 1996; Wilson and

othe

perform to the

target process by the industry.

d Methods

2.3.

ure, or unsafe foods.

Advances in Fourier transform infrared (FT-IR) spectroscopic

instrumentation combined with multivariate data analysis have made this

technology ideal for large volume, rapid screening and identification of various

analytes. Infrared spectroscopy provides valuable information of the

biochemical composition of the samples, especially in the fingerprint region,

which have been used for numerous food applications related to quality control

and safety (Grif

rs 1999).

The objective was to develop rapid methods for the determination of

sugars and salt content in snacks by combining infrared (NIR and mid-IR) and

chemometrics. Also, to validate the infrared technology to

specifications required by the

2.3 Materials an

1 Samples

Tortilla chips samples were obtained from the Snack Foods (Wyandot Inc.)

31

industries, a leading Ohio food manufacturer. Four different types of tortilla

chips were evaluated, PC 035 Yellow Tortilla, PC 349 White Tortilla, PC 231

Dipper Corn Chips, and PC 089 Yellow & White Corn Tortilla, and each type

contained multiple samples. Samples were compressed by using a manual

hydraulic carver press (model 3851 from Carver Inc., Wabash, IN). Ground

chips (10g) were placed in a stainless steel cylinder container, then placed in

the press and pressure was applied at a rate of about one stroke every two

seconds until reaching a final pressure of 15000 pounds. The final tortilla

“coo

terizer

ec. to obtain a homogenous particle size.

2.3.

he reference values for moisture, oil and salt

con

ed through 13 mm Whatman filters

(Ke

kies” were used for spectral data acquisition.

Cereal base and sucrose coated samples were provided by Wyandot Inc.

Two different cereal shapes, puff and round, and flavors, strawberry, banana,

sweet potato, apple cinnamon, peach, apple, and cherry, were included in the

study. The samples were blended using the pulse setting of an Os

blender (Oster 6646) for 8 s

2 Reference Methods

Wyandot Inc provided t

tent of the tortilla chips.

Quantitation of Sugars: Cereal aliquots (0.5g) were mixed with water (5

mL), vortexed and the eluate was filter

nt, England) with 0.2 µm pore size.

High performance liquid chromatography (HPLC): Sucrose, glucose and

fructose contents were analyzed using a Shimadzu high performance liquid

32

chromatography (HPLC) system equipped with LC-6AD pumps, SIL-20AHT

autosampler, CTO-20A column oven, and RID-10A refractive index detector

(Shimadzu Scientific Instruments, Inc. Columbia, MD). Twenty-five microliters

of each filtrate were injected into the HPLC system. The sugars were

separated on a stainless steel, 7.8 mm ID x 300 mm Aminex® HPX-87C

carbohydrate column coupled with a 4.6 mm ID x 30 mm micro-guard Carbo-C

cartridge (Bio-Rad Laboratories, Hercules, CA). Isocratic separation of the

sugars was achieved at 80°C using HPLC grade water with a flow rate of 0.6

mL/min for 30 minutes. LC Solutions software version 3.0 (Shimadzu

Scientific Instruments, Inc. Columbia, MD) was used to collect and analyze the

chromatograms. Sucrose, glucose and fructose were identified by comparing

the peak retention times with those of pure sugar standards (Fisher Scientific,

Fair Lawn, NJ). Calibration curves were prepared for each sugar in

concentrations ranging from 2.5 to 20 mg/ml. Finally, quantitation of sugars

present in each sample was calculated from the computed peak areas and the

corr

esponding standard curves.

Enzymatic Determination of Sugars: Samples for analysis were diluted

with distilled water (1:80 v/v) and analyzed according to the technical product

information provided by Sigma for each enzymatic kit. Sucrose levels were

quantified using Sigma Aldrich enzymatic assay kits (Saint Louis, MO).

Sucrose assay (SCA20-1KT) is based on the hydrolysis of sucrose to glucose

and fructose by invertase. A kexokinase then phosphorylates glucose and

33

fructose to form glucose-6-phosphate that is then oxidized to

6-phosphogluconate in the presence of NAD in a reaction catalyzed by

glucose-6-phosphate dehydrogenase. During this oxidation, an equimolar

amount of NAD is reduced to NADH. The consequent increase in absorbance

at 340 nm is directly proportional to sucrose concentration. Absorbance was

monitored using a UV/Visible spectrophotometer 2450 (Shimadzu, Columbia,

isposable cells.

2.3.

al were placed onto the different spectrometers and spectra were

acq

inst that of the background, a highly reflective gold-coated

refe

MD) with 1-cm path length d

3 FT-IR Spectroscopy

Samples were analyzed by Fourier transform Near-infrared (FT-NIR) and

Fourier transform mid-infrared (FT-IR) spectroscopy. Tortilla cookies and

blended cere

uired.

FT-NIR measurements were collected using an integrating sphere diffuse

reflectance accessory (Integrat IRTM, Pike Technologies, Madison, WI). An

Excalibur 3500 Fourier-Transform IR spectrometer (Varian Inc, Palo Alto, CA)

with a quartz beam splitter and Lead Selenide (PbSe) detector operating at 8

cm-1 resolution were used for all the reading. The spectra were collected in the

range from 10000 to 4000 cm-1, and displayed in terms of absorbance.

Interferograms of 64 scans will be co-added followed by Beer-Norton

apodization. The absorbance spectrum was obtained by ratioing the single

beam spectrum aga

rence material.

34

Ground tortilla cookies (~1g) were pressed onto a 3-bounce ATR-IR

diamond crystal using a high-pressure clamp (Pike Technologies, Madison WI).

An aliquot (0.5g) of the homogenized cereal was mixed with water (1.5mL) and

then placed onto the ATR-IR ZnSe crystal. FT-IR measurements were

collected using a range between 4000 and 700 cm-1 at a resolution of 8 cm-1

on a FTS Excalibur 3500GX FTIR spectrometer (Varian, Randolph, MA).

Interferograms of 64 scans were co-added. The absorbance spectrum was

obtained by ratioing the sample spectrum against that of a blank optical path

(ref

otte, NC). For each sample, a total of 2 independent spectra were

2.3.

erence spectrum).

Spectra of reference blanks and samples were processed using Win-IR

Pro Software 4.05 (Varian Inc, Palo Alto, CA). All the instruments were

continuously purged with CO2 free dry air from a CO2RP140 dryer (Domnick

Hunter, Charl

measured.

4 Multivariate Analysis

Samples spectra collected were analyzed using Multivariate regression

analysis (Partial least squares regression, PLSR), which built calibration

models to simultaneously correlate the reference compositional values to the

infrared spectral information. Partial Least Squares Regression (PLSR) was

cross-validated (leave-one-out approach) to generate calibration models.

PLSR is a bi-linear regression analysis method that extracts a small number of

orthogonal factors that are linear combinations of the spectral (X) variables,

35

and uses these factors as regressors for the analyte’s concentration

(Y-variable). These orthogonal factors (latent variables) explain as much as

possible of the covariance of the X and Y variables. PLSR has been

particularly successful in developing multivariate calibration models for the

spectroscopy field because it uses the concentration information (Y-variable)

actively in determining how the regression factors are computed from the

spectral data matrix (X), reducing the impact of irrelevant X-variations in the

calibration model. This capability provides a more information-rich data set of

reduced dimensionality and eliminates data noise resulting in more accurate

odels (Wold and others 2001).

2.4.

s, resolving overlapping peaks, and reducing variability between

plicates.

and reproducible calibration m

2.4 Results and Discussion

1 Spectral Analysis of Tortilla Chips

The Near- and mid- infrared spectra (Figure 1) were collected from tortilla

chip samples obtained from Wyandot Inc. NIR spectra were collected from

tortilla cookies while for ATR-IR the samples were grounded in order to

improve contact of the sample with the diamond crystal. Mathematical

processing of the raw spectra using a Savitzky-Golay second derivative

algorithm (15-pt gap size) enhanced spectral features through removing of

baseline shift

re

36

Figure 2.1 Typical Near infrared (A) and ATR-Mid infrared Spectrum of tortilla

chip samples.

Figure 1A shows the characteristic NIR spectrum of tortilla chip “cookies”

with bands representative of specific functional groups associated mainly to

their protein, moisture, carbohydrate and residual oil components. The spectral

bands were matched with the corresponding functional groups that absorbs in

the near infrared region (Table 1). The NIR reflectance spectral features

(Figure 1A) showed the absorption bands at 7190, 5855, 5790, 5680, 4340

37

and 4269 cm-1 associated with C-H structural overtones of aliphatic groups of

oils. Bands near 6960 and 5200 cm-1 were assigned to the first overtones and

combination absorption modes of OH, respectively (Kays and others 1997;

Espinoza and others 1999). Bands in the 5000-4550 cm-1 region have been

related to amides modes (Liu and others 1994) with bands near 4850 and

4600 cm-1 assigned to combinations of amide A/I and amide B/II, respectively

(Wang and others 1994; Robert and others 1999). In addition, bands at 5190

cm-1 and in the region from 4400 - 4033 cm-1 have been related to OH

stretching/HOH deformation modes and C-H groups of carbohydrates,

respectively (Kays and others 1997).

The typical MIR-spectrum (Figure 1B) showed that the prevalent signals

obtained by ATR-IR of tortilla chips were oil-related regardless the fact that the

oil was extracted from the cookies using a Carver hydraulic press. Assignment

of the mid-infrared bands to functional groups of the components of oils (Table

1) was done as reported by Guillen and Cabo (McClure and others 1996;

Guillén and others 1997a). The strong absorption at 2920 and 2860 cm-1 are

associated with asymmetrical and symmetrical stretching modes of the

methylene groups of the fatty acid backbone, respectively. The small band at

3005 cm-1 is characteristic of the cis double-bond stretching of unsaturated fats.

The strongest absorption of the spectra was at 1745 cm-1 related to the ester

carbonyl (C=O) stretching mode of fatty acids. The bands at 1458 and 1375

cm-1 are related to the C-H bending vibrations of CH2 and CH3 aliphatic groups,

38

whereas bands near 1654 cm-1 corresponds to the amide I (N-H) stretching

vibration of proteins and in the fingerprint region (1200 – 900 cm-1) containing

unique and complex information related to C-O, C-C stretching and C-O-H and

C-O-C deformation of carbohydrates.

Table 2.1 Assignment of Bands in the NIR and MIR-FT Spectra

Frequency (cm-1)

Functional Group Assignment

Near Infrared Region

≈ 8600 – 8150

≈ 7200 – 6850

≈ 5950 - 5600

≈ 5200 - 5100

≈ 5000 - 4800

≈ 4350 - 4150

C-H 2nd overtone of -CH3; Fatty acids.

O-H 1st overtone of H2O; water.

C-H 1st overtone of –CH3 and CH2, Fatty acids.

O-H combination of H2O; water.

C-H stretch and C=O stretching; protein.

C-H combination; fatty acids and C-H / C-C stretch;

fatty acids having cis double bonds.

C-H stretch / CH2 deformation combination; starch

Mid Infrared Region

≈ 3040 – 2850

≈ 1744

≈ 1654

≈ 1546

≈ 1500 - 1300

≈ 1480 - 1440

≈ 1400

≈ 1240 - 1180

≈ 1200 – 900

(Fingerprint region)

≈ 750

C-H asymmetrical and symmetrical stretching of CH2

and CH3; Fatty acids.

C=O stretching of esters; free fatty acids

Amide I

Amide II

O-H bending vibration, water

C-H bending vibrations of CH2 and CH3; proteins

C-H bending vibrations of CH2 and CH3; fatty acids

C=O symmetric stretching; COO-

C-O stretching vibration; esters

C-O, C-C stretching and C-O-H and C-O-C deformation

of carbohydrates.

C-H overlapping of CH2 rocking vibration, and out of

plane vibration of cis-disubstituted olefins.

39

2.4.2 Development of Infrared (NIR and mid-IR) models for Salt, Moisture

and Fat content prediction

NIR and MIR techniques allowed the development of calibration models for

quantification of salt, moisture and fat content in tortilla chips. The salt content

ranged from 0.4 to 2.5%, moisture content ranged from 0.1 to 1.5% and the fat

content ranged from 18 to 35%. Cross-validated (leave-one-out) PLSR models

were developed to determine the salt, fat and moisture content in various

tortilla products and to infer relevant chemical information from their spectra by

using the loading vectors (Guillén and others 1997b).

The generated models from NIR and MIR spectral data (Table 2) provided

similar performance statistics for salt, moisture and fat content with standard

error of cross validation (SECV) of ~0.3%, 0.2% and 1.9%, respectively. All

NIR models provided higher correlation coefficients (r) as compared to ATR-IR

(Table 2).

40

Table 2.2 PLSR model results for salt, moisture and fat content prediction with

NIR and MIR

Cumulative

Technique Analyte Factors Variance SEV r Val SEC r Cal

FT-NIR Salt 5 99.00 0.26 0.76 0.25 0.79

Oil 4 98.93 1.80 0.93 1.52 0.95

Moisture 7 98.35 0.16 0.90 0.14 0.92

ATR-IR Salt 7 98.00 0.29 0.55 0.24 0.75

Oil 7 98.61 1.96 0.91 1.62 0.95

Moisture 6 98.17 0.21 0.82 0.19 0.86

Overall, the most accurate models were obtained with NIR spectra,

providing more precise and accurate predictions as compared to the

determined reference values, with SEP of 0.25 for salt obtained from an

independent set of 30 samples not included in the model. The lower prediction

performance of the calibration models by MIR as compared to NIR could be

attributed to sample heterogeneity, particle size, coarseness or preferential

adsorption of fat on the ATR crystal. The PLS regression graph (Figure 2) for

salt, moisture and fat content showed the best correlations between the NIR

estimated concentrations with the reference analysis.

41

Figure 2.2 Partial Least Squares Regression (PLSR) cross-validation plots for

salt, moisture, and oil determination in tortilla chip samples from Wyandot Inc.

Examination of the loading spectra indicated which bands of the spectrum

were associated with the highest variation in the calibration set. Frequencies of

high variation reflect contributions of spectral resolution elements that can be

correlated with combinations of different chemical and physical phenomena

(Bjorsvik and others 1992; Kays and others 1998). The NIR loadings showed

42

that the highest variation in the calibration set for salt content was associated

with the dominant and broad peaks centered at 5181 cm-1. Even though

sodium chloride has no specific absorption bands in the NIR region, salt

analysis is possible because of the effect of the chloride ion on the shift in the

water spectrum. In addition, prediction of fat content was associated with

broad bands at 5945, 5752, and 4344 cm-1 due to C-H overtones and

combination bands of lipids, while moisture content was related with NIR

absorbance near 5200 cm-1 associated to broad OH absorption combination

bands of water (Espinoza and others 1999). Examination of the loadings for

the PLSR models generated with MIR second derivatized spectra showed the

predominance of the fat bands in all models that limited the predictive ability for

salt and moisture content.

2.4.3 Sugar content in Cereals

Addition of sucrose to cereal base resulted in a distinct FT-NIR

absorbance peak at 6955 cm-1 related to the first overtone of the O-H structural

vibration of sucrose (Kays and others 1998). Most of the changes in the

spectrum however, were barely noticeable and chemometric models are

needed to extract the relevant information for quantitative analysis. PLSR was

used to extract pertinent information from complex spectra of several

chemicals by describing the main types of spectral variations and relating it to

the chemical data to generate the calibration models.

The sugar composition of the cereal (coated and uncoated bases) is

43

presented in Table 3. We found discrepancies between the enzymatic sucrose

methods and HPLC analysis (Table 3). Average sucrose contents in coated

cereal of 20.8 g/100g (SD 7.2 and range of 37.7-8.9 g/100g) and 18.4 g/100g

(SD 3.3 and range of 25.7-10.5 g/100g) were determined by enzymatic assays

and HPLC, respectively. Uncoated cereal base showed an average sucrose

content of 6.6 g/100g (SD 5.6 and range of 17.8-1.4 g/100g) by enzymatic kits

and 4.3 g/100g (SD 2.2 and range of 10.0-0.7 g/100g) by HPLC analysis.

Overall, enzymatic determination of sucrose in cereal bases showed higher

data variability as evidenced by the higher standard deviation and range of

values as compared to HPLC analysis. HPLC methods have shown to give

better quantification of sucrose because the enzymatic procedure uses a

two-stage process that introduces more analytical variation (Ashurst 2005).

Table 2.3 Sugar contents in cereal tested by enzymatic assay kit and HPLC

method

44

Table 4 shows the effect of type of base flavor on the sugar levels. Coated

cereals showed that peach and strawberry bases provided the highest and

lowest sucrose levels, respectively, among all samples evaluated. Regarding

the uncoated bases, peach and apple cinnamon (6.9 g/100g) gave the higher

levels of initial sucrose while strawberry and sweet potato bases gave the

lowest (2 g/100g). Levels of fructose and glucose in apple cinnamon (glu 4.4;

fru 5.7 g/100) showed a significant difference from the rest of flavor bases (glu

2.2±0.5; fru 2.6 ±0.3 g/100g). This variability in sugar levels among coated

bases emphasized the need to monitor sugar content of the initial and final

products to standardize the quality of the product. Infrared technologies can

provide simple and real-time analysis of individual sugar levels establishing a

uniform high quality product.

Table 2.4 Average sugar content in coated and uncoated base cereal

according to type of flavor

Flavor [Sucrose] g/100g [Sucrose] g/100g [Fructose] g/100g [Glucose] g/100g Coated Base Uncoated Base

Strawberry 15.6 2.1 3.1 2.3

Banana 17.9 3.1 2.5 2.2

Sweet Potato 17.1 1.9 2.2 1.9

Apple

Cinnamon 19.7 6.5 5.7 4.4

Peach 24.7 7.3 2.4 2.7

Apple 20.7 4.3 2.7 1.4

Cherry 19.5 4.1 2.4 2.9

45

The calibration results for determination of sugars in cereals are presented

in Table 5.

Table 2.5 Performance statistics for PLSR calibration and validation models

The optimal number of factor to retain in the calibration model should not be

too few that can leave important structure un-modeled, and it should not be too

46

many that draws too much measurement noise. In general, mid-infrared (MIR)

PLSR models incorporated higher number of factors as compared to FT-NIR.

FT-NIR and MIR techniques allowed the quantification of individual sugars in

cereals. The cross-validated calibration models from derivatized spectral data

gave SECV of ~0.4% for glucose and fructose and ~1.3% for sucrose

determinations, explaining > 99% of the variability. Overall, FT-NIR and MIR

gave accurate predictions from an independent set of samples not included in

the development of the calibration model. MIR (SEP 1.4%) showed a slightly

improved sucrose prediction compared to FT-NIR (1.7%) probably because by

incorporating more factors is able to include more relevant information-rich

data and eliminates data noise resulting in robust calibration models. PLS

correlation graph for individual sugar content are presented in Figure 3,

showing the good correlation of the models with the reference HPLC analysis,

both with NIR and MIR.

47

Figure 2.3 Partial Least Squares Regression (PLSR) cross-validation plots for

sugar determination in cereal samples from Wyandot Inc.

Examination of the loadings for cereals coated with sugars (Figure 4), the

FT-NIR spectral features were associated the characteristic absorbances of

carbohydrate O-H and C-H groups. The FT-NIR spectrum of crystalline sugar

was very similar to that reported by McClure et al. (McClure and others 1996)

with a sharp band at 6963 cm-1 dominated the first loading, which is indicative

of the importance of O-H groups in crystalline sucrose, as previously reported

48

by Kays et al. (Kays and others 1998). The bands near 6960, 6595 and 6325

cm-1 has been assigned to the first overtones of stretching modes of free OH,

intramolecular hydrogen-bonded OH, and intermolecular hydrogen-bonded

OH groups, respectively (McClure and others 1996). Bands in the 5900-5700

cm-1 region have been associated with first overtones of CH stretching modes.

In addition, bands in the information-rich region from 4600-4000 cm-1 can be

ascribed to combinations of O-H bend/ hydrogen-bonded O-H stretch (~ 4428

cm-1), O-H stretch/C-C stretch (~ 4393 cm-1) and combinations of C-H/C-C

(4385-4063 cm-1) vibrations of the sugar molecules (Osborne and others

1986).

For MIR, first and second PC loadings obtained showed many of the

features that can be attributed to C–O and C–C stretching modes in the

900–1150 cm-1 region, with a band around 1020-1060 cm-1 due to O-H

vibrations. Bands appearing between 1150-1470cm-1 are attributed to bending

modes of C-C-H, C-O-H and O-C-H groups while more intense peaks in the

region between 900 and 1150 cm-1 arise mainly from C-O and C-C stretching

modes (Sivakesava and others 2002).

49

Figure 2.4 Partial Least Squares loading spectra for the first 2 latent variables

for the calibration set measured by FT-NIR reflectance and ATR-MRI

spectroscopy.

2.5 Conclusions

FT-NIR and MIR spectroscopy and multivariate techniques (PLSR)

allowed for rapid, accurate and non-destructive determination of salt, fat and

moisture content in tortilla chips and sugar levels in cereals. Overall, near

infrared results were convenient for quantification purposes due to the faster

50

sample preparation and ease of use. The total time required for sample

preparation and analysis was less than 2 minutes per sample. The infrared

technology can provide sensing capabilities for rapid monitoring of quality

assurance, providing the industry with a tool for timely correction measures

during manufacture. Handheld portable infrared systems could provide

greater versatility, ruggedness, and portability as a real-time infrared sensor for

direct fingerprinting strategies of food components for the reliable assessment

of quality and safety

51

2.6 References

1. Ashurst PR. 2005. Chemistry and technology of soft drinks and fruit juices.

2. Bjorsvik HR, Martens H. 1992. Data analysis: Calibration of NIR

instruments by PLS regression. In:Burns D, Ciurczak E, editors.

Handbook of Near-Infrared Analysis. New York:Marcel Dekker, Inc.

3. Caselunghe M, Lindeberg J. 2000. Biosensor-based determination of folic

acid in fortified food. Food Chem 70:523-532.

4. Espinoza LH, Lucas D, Littlejohn D. 1999. Characterization of hazardous

aqueous samples by Near-IR spectroscopy. Appl. Spectrosc 53:97-102.

5. Griffiths PR, De Haseth JA.1986. Chemical Analysis 83: Fourier

Transform Infrared Spectrometry. 656 p.

6. Guillén MD, Cabo N. 1997a. Infrared spectroscopy in the study of edible

oils and fats. Journal of the Science of Food and Agriculture 75:1-11.

7. Guillén MD, Cabo N. 1997b. Characterization of edible oils and lard by

Fourier transform infrared spectroscopy. Relationships between

composition and frequency of concrete bands in the fingerprint region.

Journal of the American Oil Chemists’ Society 74:1281-1286.

8. Halim Y, Schwartz S, Francis D, Baldauf N, Rodriguez-Saona L. 2006.

Direct determination of lycopene content in tomatoes (Lycopersicon

52

esculentum) by attenuated total reflectance infrared spectroscopy and

multivariate analysis. J AOAC Int 89:1257-1262.

9. Holm F. 2003. Food Quality Sensors. Institut National de la Recherche

Agronomique. Paris, France. ISBN : 2-7380-1076-8.

10. Kays SE, Windham WR, Barton II FE. 1997. Effect of cereal product

residual moisture content on total dietary fiber determined by near-infrared

reflectance spectroscopy. J. Agric. Food Chem 45:140-144.

11. Kays SE, Windham WR, Barton II FE. 1998. Prediction of total dietary fiber

by Near-Infrared reflectance spectroscopy in high-fat- and

high-sugar-containing cereal. J. Agric. Food Chem 46:854-861.

12. Liu Y, Cho R, Sakurai K, Miura T, Ozaki Y. 1994. Studies on

spectra/structure correlations in Near-Infrared spectra of proteins and

polypeptides. Part I. A marker for hydrogen bonds. Appl. Spectrosc

48:1249-1254.

13. McClure WF, Maeda H, Dong J, Liu Y, Ozaki Y. 1996. Two-Dimensional

correlation of Fourier Transform Near-Infrared and Fourier Transform

Raman spectra I: Mixtures of sugar and protein. Appl. Spectrosc

50:467-475.

14. McKelvy M, Britt T, Davis B, Gillie J, Lentz L, Leugers A, Nyquist R, Putzig

C. 1996. Infrared spectroscopy. Analytical Chemistry 68:93R-160R.

53

15. Osborne BG, Fearn T. 1986. Theory of near infrared spectrophotometry. In:

Osborne BG, Fearn T, editors. Near Infrared Spectroscopy in Food

Analysis. Essex :Longman Scientific & Technical.

16. Robert P, Devaux MF, Mouhous N, Dufour E. 1999. Monitoring the

secondary structure of proteins by Near-Infrared spectroscopy. Appl.

Spectrosc 53:226-232.

17. Sivakesava S, Irudayaraj J. 2002. Classification of simple and complex

sugar adulterants in honey by mid-infrared spectroscopy. International

Journal of Food Science and Technology 37:351–360.

18. U.S. Census Bureau, Company Statistics Division, Statistics of U.S.

Businesses, Tabulations by Enterprise Size, U.S. all industries.

19. Wang J, Sowa MG, Ahmed MK, Mantsch HH. 1994. Photoacustic

Near-Infrared investigation of homo-polypeptides. J. Phys. Chem

98:4748-4755.

20. Wilson RH, Tapp HS. 1999. Trends in Analytical Chemistry. 18:85-93(9).

21. Wold S, Sjostrom M, Eriksson L. 2001. PLS-regression: a basic tool of

chemometrics. Chem and Intell Lab Systems 58:109-30.

54

References

1. Anonymous. 2001. Introduction to Fourier transform infrared Spectrometry.

Thermo nicolet corporation.

2. Anonymous. 2004. Global Snack Food Industry Trends. Snack Food &

Wholesale Bakery 93:64.

3. Anonymous. 2004. Snacking, the All-American pastime. Snack Food &

Wholesale Bakery 93:SI-34.

4. Anonymous. 2011. High performance liquid chromatography.

Standardbase techniques. Drenthe College. The Netherlands.

5. Arber A. 1934. The Gramineae: a study of cereal, bamboo, and grass.

University Press, Cambridge.

6. Ashurst PR. 2005. Chemistry and technology of soft drinks and fruit juices.

7. Bergmeyer HU, Bernt E. 1974. Methods of Enzymatic Analysis, 2nd edition.

Bergmeyer HU, eds., Academic Press (New York), 1177-1179.

8. Binder H. 1980. Separation of monosaccharides by high-performance

liquid chromatography: comparison of ultraviolet and refractive index

detection. J. Chromatog 189: 414.

9. Bird IM. 1989. High performance liquid chromatography: principles and

clinical applications. British Medical Journal 299:783-787.

55

10. Bjorsvik HR, Martens H. 1992. Data analysis: Calibration of NIR

instruments by PLS regression. In:Burns D, Ciurczak E, editors.

Handbook of Near-Infrared Analysis. New York:Marcel Dekker, Inc.

11. Caselunghe M, Lindeberg J. 2000. Biosensor-based determination of folic

acid in fortified food. Food Chem 70:523-532.

12. Chapman GP. 1996. The Biology of Grasses. CAB International, UK.

13. Chen H, Gatlin B, Burnside G, Reddy RN, Kandarpa VK, Mazumder MK.

1995. A transfer efficiency model for powder coating process. In

Conference record of the 1995 IEEE 2:1280 –1283.

14. Chopra VL, Prakash S, eds. 2002. Evolution and Adaptation of Cereal

Crops. Science Publishers Inc, NH, USA.

15. Clark J. 2007. High performance liquid chromatography-HPLC. Chemguid.

http://www.chemguide.co.uk/analysis/chromatography/hplc.html.

16. Conrad EC, Palmer JK. 1976. Rapid analysis of carbohydrates by

high-pressure liquid chromatography. Food. Technol 30: 84.

17. Culbertson JD. 2004. Grain, cereal: ready-to-eat breakfast cereals. In:

Scott Smith, J., Hui, Y.H. (Eds.), Food Processing: Principles and

Applications. Blackwell Publishing, Oxford, UK, pp. 239–248.

18. Denoyelle L. 2008. Coating processes in food: a rational approach to the

selection and development of a process. AgroFood Industry Hi-Tech 19:28

–30.

19. Duarte IF, Barros AN, Delgadillo I, Almeida CU, Gil AM. 2002. Application

56

of FTIR spectroscopy for the quantification of sugars in mango juice as a

function of ripening. Journal of agricultural and food chemistry

50:3104-3111.

20. Espinoza LH, Lucas D, Littlejohn D. 1999. Characterization of hazardous

aqueous samples by Near-IR spectroscopy. Appl. Spectrosc 53:97-102.

21. Fast RB. 2000. Manufacturing technology of ready-to-eat breakfast cereals.

In:Fast, R.B., Caldwell, E.F. (Eds.), Breakfast Cereals and How They are

Made. American Association of Cereal Chemists, Paul MN, USA, pp. 29–

34.

22. Fast RB. 2001. Breakfast cereals. In: Owen, G. (Ed.), Cereals Processing

Technology. Woodheal Publishing, Cambridge, UK, pp. 158–172.

23. Gomis DB, Tamayo DM, Valles BS, Mangas Alonso JJ. 2004. Detection of

apple juice concentrate in the manufacture of natural and sparkling cider

by means of HPLC chemometric sugar analyses. Journal of agricultural

and food chemistry 52:201-203.

24. Griffiths PR, De Haseth JA.1986. Chemical Analysis 83: Fourier

Transform Infrared Spectrometry. 656 p.

25. Guillén MD, Cabo N. 1997a. Infrared spectroscopy in the study of edible

oils and fats. Journal of the Science of Food and Agriculture 75:1-11.

26. Guillén MD, Cabo N. 1997b. Characterization of edible oils and lard by

Fourier transform infrared spectroscopy. Relationships between

57

composition and frequency of concrete bands in the fingerprint region.

Journal of the American Oil Chemists’ Society 74:1281-1286.

27. Halim Y, Schwartz S, Francis D, Baldauf N, Rodriguez-Saona L. 2006.

Direct determination of lycopene content in tomatoes (Lycopersicon

esculentum) by attenuated total reflectance infrared spectroscopy and

multivariate analysis. J AOAC Int 89:1257-1262.

28. Harris DC. 2007. Quantitative chemical analysis, 7th Ed. Freeman, New

York.

29. Holm F. 2003. Food Quality Sensors. Institut National de la Recherche

Agronomique. Paris, France. ISBN : 2-7380-1076-8.

30. Kays SE, Windham WR, Barton II FE. 1997. Effect of cereal product

residual moisture content on total dietary fiber determined by near-infrared

reflectance spectroscopy. J. Agric. Food Chem 45:140-144.

31. Kays SE, Windham WR, Barton II FE. 1998. Prediction of total dietary fiber

by Near-Infrared reflectance spectroscopy in high-fat- and

high-sugar-containing cereal. J. Agric. Food Chem 46:854-861.

32. Lee EJ, Yoo KS, Jifon J, Patil BS. 2009. Characterization of shortday onion

cultivars of 3 pungency levels with flavor precursor, free amino acid, sulfur,

and sugar contents. Journal of food science 74:475-480.

33. Liu Y, Cho R, Sakurai K, Miura T, Ozaki Y. 1994. Studies on

spectra/structure correlations in Near-Infrared spectra of proteins and

58

polypeptides. Part I. A marker for hydrogen bonds. Appl. Spectrosc

48:1249-1254.

34. Marquart L and others eds. 2002. Whole grain foods in health and disease.

American Association of Cereal Chemists, St. Paul, MN.

35. Mazumder MK, Wankum DL, Sims RA, Mountain JR, Chen H, Pettit P.

1997. Influence of powder properties on the performance of electrostatic

coating process. Journal of Electrostatics 40–41:369–374.

36. McClure WF, Maeda H, Dong J, Liu Y, Ozaki Y. 1996. Two-Dimensional

correlation of Fourier Transform Near-Infrared and Fourier Transform

Raman spectra I: Mixtures of sugar and protein. Appl. Spectrosc

50:467-475.

37. McKelvy M, Britt T, Davis B, Gillie J, Lentz L, Leugers A, Nyquist R, Putzig

C. 1996. Infrared spectroscopy. Analytical Chemistry 68:93R-160R.

38. Mun˜oz AM. 2002. Sensory evaluation in quality control: an overview, new

developments and future opportunities. Food Quality and Preference

13:329–339.

39. Nikolaou M. 2006. Control of snack food manufacturing systems. Control

Systems, IEEE 26:40-53.

40. Osborne BG, Fearn T. 1986. Theory of near infrared spectrophotometry. In:

Osborne BG, Fearn T, editors. Near Infrared Spectroscopy in Food

Analysis. Essex :Longman Scientific & Technical.

41. Ortega-Rivas E. 2005. Handling and processing of food powders and

59

particulates. In: Onwulata C. (Ed.), Encapsulated and Powdered Foods.

CRC Taylor & Francis, Boca Raton FL, USA, pp. 75–144.

42. Picha DH. 1985. HPLC determination of sugars in raw and baked sweet

potatoes. Journal of food science 50:1189-1190.

43. Pool R. 1948. Marching with the grasses. University of Nebraska Press,

USA.

44. Pomeranz Y, Munck L, eds. 1981. Cereals: A Renewable Resource.

American Association of Cereal Chemists. St. Paul, MN.

45. Roberts E. 2009. The Facts on Heart Disease, Sodium, and Sugar.

EverydayHealth.com.http://www.everydayhealth.com/heart-health/salt-and

-sugar.aspx

46. Robert P, Devaux MF, Mouhous N, Dufour E. 1999. Monitoring the

secondary structure of proteins by Near-Infrared spectroscopy. Appl.

Spectrosc 53:226-232.

47. Sevatson E, Huber GR. 2000. Extruders in the food industry. In: Riaz, M.N.

(Ed.), Extruders in Food Applications. CRC Taylor & Francis, Boca Raton

FL, USA, pp. 167–204.

48. Shukla TP. 1994. Future snacks and snack food-technology. Cereal Foods

World 39:704-705.

49. Shukla TP. 2000. Modern snack foods. Cereal Foods World 45: 477-477.

50. Sibilia JP. 1996. Materials Characterization and chemical analysis, 2nd Ed.

Wiley-VCH, New York.

60

51. Sivakesava S, Irudayaraj J. 2001. Detection of inverted beet sugar

adulteration of honey by FTIR spectroscopy. Journal of the science of food

and agriculture 81:683-690.

52. Sivakesava S, Irudayaraj J. 2002. Classification of simple and complex

sugar adulterants in honey by mid-infrared spectroscopy. International

Journal of Food Science and Technology 37:351–360.

53. Skoog DA, Holler FJ, Crouch SR. 2007. Principles of instrumental analysis,

6th Ed. Thomson brooks/cole.

54. Snyder LR, Kirkland JJ, Dolan JW. 2009. Introduction to Modern Liquid

Chromatography. New Jersey:John Wiley & Sons, Inc.

55. Solís-Morales D, Sáenz-Hernández CM, Ortega-Rivas E. 2009. Attrition

reduction and quality improvement of coated puffed wheat by fluidised bed

technology. Journal of Food Engineering 93:236-241.

56. Stone H, Sidel JL. 1993. Sensory evaluations practices, California:

Academic Press.

57. Subramanian A, Harper WJ, Rodriguez-Saona LE. 2009. Cheddar cheese

classification based on flavor quality using a novel extraction method and

Fourier transform infrared spectroscopy. Journal of dairy science

92:87–94.

58. The History of Breakfast Cereals. 2008.

Iventors.about.com. http://inventors.about.com/library/inventors/blcereal.ht

m.

61

59. Torres-Mart í nez JL. 2001. Coating of Puffed Wheat Cereal with a

Chocolate Flavoured Cover Using Tumbling. MSc Thesis. Autonomous

Series

s of U.S.

tates Department of Agriculture, 1996. Food Guide Pyramid

ood Processing Operations and

investigation of homo-polypeptides. J. Phys. Chem

e

cfm?articleID=000A97BD-80FC-1DF6-973

: a basic tool of

d electrostatic

University of Chihuahua (in Spanish), Chihuahua, Chih, Mexico.

60. U.S. Census Bureau. 2002. 2002 Economic Census Industry

Report. http://www.census.gov/epcd/ec97/industry/E311919.HTM.

61. U.S. Census Bureau, Company Statistics Division, Statistic

Businesses, Tabulations by Enterprise Size, U.S. all industries.

62. United S

Booklet.

63. Valentas KJ, Levine L, Clark JP. 1991. F

Scale-up. Marcel Dekker, New York, NY.

64. Wang J, Sowa MG, Ahmed MK, Mantsch HH. 1994. Photoacustic

Near-Infrared

98:4748-4755.

65. Willett WC, Stampfer MJ. 2002. Rebuilding the Food Pyramid: Overview /

The Food Guide Pyramid (Sidebar). Availabl

at http://www.sciam.com/article.

809EC588EEDF&sc=I100322.

66. Wilson RH, Tapp HS. 1999. Trends in Analytical Chemistry. 18:85-93(9).

67. Wold S, Sjostrom M, Eriksson L. 2001. PLS-regression

chemometrics. Chem and Intell Lab Systems 58:109-30.

68. Yousuf S, Barringer SA. 2007. Modeling nonelectrostatic an

62

63 63

powder coating. Journal of Food Engineering 83:550–561.