Caballero B. (ed.) Encyclopaedia of Food Science, Food Technology and Nutrition. Ten-Volume Set

Подождите немного. Документ загружается.

with single bonds vibrate below 1500 cm

1

. Stretch-

ing vibrations change bond lengths, and deformation

vibrations change bond angles. They occur at

approximately half the stretching frequency.

0005 According to quantum mechanics, vibrations are

excited by energy quanta, hn

, where h is Planck’s

constant, and n

is the vibrational frequency. Radi-

ation quanta in the infrared range, 2.5–1000 mm,

equivalent to 4000–10 cm

1

, can be absorbed by the

molecule and excite it from the vibrational ground

state (v=0) to the vibrational excited state (v=1).

Thus, absorption bands are produced in the infrared

spectrum (Figure 2a).

0006 If the molecules are irradiated with quanta of the

near-infrared range, 0.8–2.5 mm, equivalent to

12 500–4000 cm

1

, the overtones and combination

bands of the infrared spectrum, especially, of the

CH, NH, and OH bonds are excited, giving rise to

the near-infrared spectrum.

Near infrared

0007 When the molecules are irradiated with quanta of

large energy, hn

(by monochromatic, usually laser,

radiation from the ultraviolet, visible or near infrared

region), a scattering process may occur, by which

vibrations with energy hn

are excited and quanta of

lower energy, hn

¼ hn

 hn

, are emitted simultan-

eously. These constitute the so-called Raman

spectrum (Figure 2b). The quantum yield of Raman

scattering is quite low (10

12

), this means that

Raman spectra are of a low intensity. If there are

samples with absorption bands that can be excited

by the quanta hn

, then fluorescence may occur, a

strong emission spectrum with broad bands

(Figure 2c). Since the quantum yield for this process

is nearly 1, Raman spectra can be completely con-

cealed, even when the concentration of the fluores-

cing impurities is very low. Until 1986, Raman

spectroscopy with excitation in the visible region

was only rarely applied to food analysis, since the

spectra were heavily masked by the fluorescence

spectra of some natural compounds: cells can only

live with the help of an enzymatic machinery

with absorption bands in the visible range of the

spectrum – consequently, it shows fluorescence and

is photochemically endangered. Raman spectra of

most food products therefore cannot be investigated

with ‘classical’ excitation in the visible range. How-

ever, when Raman spectra are excited in the near-

infrared range, e.g., with 1064 nm, the energy of

the quanta, hn

, is too small (only about half of

the quanta at 488 nm) to excite fluorescence

(Figure 2c). Excitation with the Nd:YAG laser at

1064 nm marks an optimum for non-destructive

recording of the Raman spectra of living cells and

all kinds of food.

0008The intensities of the vibrations revealed by the

infrared spectrum are proportional to the square of

the change of the molecular dipole moment, m, during

the vibration, described by the ‘normal coordinate,’

q. The intensity of the Raman spectrum is propor-

tional to the square of the change of the molecular

polarizability, a, during the vibration:

Infrared intensity 



Raman intensity 



The polarizability characterizes the elastic deform-

ation of the charge distribution of a molecule by an

electric field. Therefore, infrared spectra usually

show strong bands for vibrations of polar groups.

Raman spectra, however, show strong bands for

vibrations of nonpolar and multiple bonds as well

(a)

(b)

fig0004Figure 4 Entrance optics for an NIR-FT-Raman spectrometer:

(a) SC sample cup with CaF

window and a VITON O-ring, L beam

of a Nd:YAG laser; (b) the sample in a small test tube, TT, in the

center of a hemispherical mirror HK. The exciting radiation,

diffusely scattered by the sample is reflected by the notch filter

N; together with HK this multiple reflection arrangement

enhances the intensity of the observed Raman spectrum and

integrates the spectra of the whole sample.

SPECTROSCOPY/Infrared and Raman 5417

as for symmetric vibrations of collections of similar

bonds, as in ring systems.

Typical Features of Infrared and Raman

Spectra

0009 In Table 1, typical properties of the methods of vibra-

tion spectroscopy are collected. In several aspects,

Raman spectroscopy has several advantages com-

pared with infrared spectroscopy: it needs only

very simple or no preparation, uses nonproblematic

sample containers, and needs only one spectrometric

set-up for the whole spectrum. In particular, the

intensity of the bands is proportional to the concen-

tration. Unfortunately, Raman spectrometers are at

present more expensive than the infrared instruments.

The properties of the near-infrared spectrometry are

also listed, for comparison.

0010Table 2 lists the assignment of the bands that can

be observed using both methods. Vibrations of

polar groups are more intense in the infrared spec-

trum, whereas multiple bonds and symmetric

4000

Infrared transmissionRaman intensity

3500 3000

n(CH)

n(NH)

n(CH)

Amide I

Amide II

Amide III

Phe

Tyr

Trp

n(=CH)

δ(CH)

2500

Wavenumber (cm

−1

)

2000 1500 1000 500

fig0005 Figure 5 Infrared and Raman spectra of bovine serum albumin. n, mean stretching vibration; Phe, Tyr and Trp mark typical

vibrations of amino acids.

5418 SPECTROSCOPY/Infrared and Raman

vibrations are usually more intense in the Raman

spectrum. Therefore, it is often useful to evaluate

both spectra.

0011 Every molecule is a vibrating unit. If it contains

n atoms, it has 3n 6 ‘normal vibrations.’ Together,

they represent a unique fingerprint, allowing the

identification of the molecule. For example, Figure 3

shows the infrared spectra (above) and the Raman

spectra (below) of eight important food components.

0012 The spectra of the cystine molecule (Figure 3a)

shows many bands with the same frequencies in the

infrared and the Raman spectrum. However, the in-

tensities clearly demonstrate the nature of the groups

responsible: vibrations of polar groups are strong in

the infrared. The bands due to the COOH and NH

groups with participation of zwitterionic COO



and

groups are strong and broad in the infrared spec-

trum at 3000 and 1600–1400 cm

1

, but they are weak

in the Raman spectrum. Conversely, the stretching

vibrations of the CH

group at 2917 and 2969 cm

1

and especially the S—S stretching vibration at

499 cm

1

are very strong in the Raman spectrum.

0013In Figure 3b, the spectra of linolenic acid methyl

ester are shown. The ester C

—

O group (1743 cm

1

)

produces a strong infrared and a weak Raman band;

the C

—

C group (1654 cm

1

), however, shows a very

strong Raman and a very weak infrared band. These

are useful for the determination of the unsaturated

fatty acids in fats (see below).

0014The spectra of retinol, vitamin A (Figure 3c) show

the typical vibrations of the conjugated C

—

C bonds

at 1586, 1193, and 1155 cm

1

, strong in the Raman

and weak in the infrared spectrum.

0015In Figure 3d and 3e, the spectra of caffeine and

vitamins B

show many strong and sharp bands,

typical for nonflexible ring systems.

4000

Wavenumber (cm

−1

)

3500

Raman intensity Infrared transmission

3000 2500 2000 1500 1000 500

n( CH)

( CH

)

(CH

)

n(C O)

w(CH

)

t(CH

)

n(C O)

n(C 0)

d(CH

)

d(CH

)

n( CH)

fig0006 Figure 6 Infrared and Raman spectra of safflower oil. n, d, o and t mark stretching, deformation, wagging and twisting vibrations.

Note the high intensity of the n (C = C) in the Raman spectrum.

SPECTROSCOPY/Infrared and Raman 5419

0016 Figure 3f, g, and h show the spectra of carbo-

hydrates: vitamin C, glucose, and sucrose. All show

strong bands in the range 1200–1000 cm

1

, due to

the C—O groups. Hydrogen-bonded O—H groups

show bands around 3400, and ‘free’ OH bonds

above 3500 cm

1

Wavenumber (cm

−1

)

Raman intensity

IR transmission

1800 1600

Amide I

Amide II

Amide III

A,G

n(C C)

n(C O)

Phe

Tyr

Trp

Tyr

Phe

C,U,T

1514

1155

1400

(b)

(a)

1200 1000 800 600 500

n(C C)

fig0007 Figure 7 (a) Infrared and (b) Raman spectrum of Micrococcus roseus. C, cytidine; U, uridine; T, thymidine; A, adenosine;

G, guanosine.

Amide I

Phe

1654

Phe

Tyr

Amide III

Protein, skeletal

backbone

n(C C)

n(C N)

t (CH

)

n(C O)

n(C C)

n(C O)

n(S S)

(a)

(b)

(c)

(a)

(b)

(c)

Raman intensity

1800270029003100 1600 1400 1200 1000 800 600 400

Wavenumber (cm

−1

)Wavenumber (cm

−1

)

fig0008 Figure 8 Raman spectra of typical components of food: (a) banana (carbohydrate); (b) turkey breast (protein); (c) butter (lipid).

5420 SPECTROSCOPY/Infrared and Raman

0017 Identification is made easy by comparison of

the observed spectra with those in atlases of Raman

and infrared spectra of organic and inorganic

compounds and by collections of digitized spectra

supplied by the instrument manufacturers.

Quantitative Analyses

0018 The basic relation for quantitative analyses by infra-

red absorption spectroscopy is the Beer–Lambert law:

F ¼ F

ecl

¼ F

A

Here, F

is the radiant power illuminating the

sample, and F is the transmitted power. The transmit-

tance of the sample, t ¼ F=F

, is usually recorded, e is

the molar absorption coefficient, c is the concentra-

tion and l is the absorbing path length. Most instru-

ments allow quantitative analyses by recording

the absorbance, A, which is proportional to the

concentration.

1800 1600 1400 1200 1000 800 600 400

Wavenumber (cm

−1

)

Raman intensity

(a)

(b)

(c)

(d)

(C C)

fig0010 Figure 10 Raman spectra of food components containing carotinoids: (a) peel of untreated orange; (b) peel of orange, treated with

wax and fungicides; (c) carrot; (d) b-carotene.

3100 2900

Wavenumber (cm

−1

)

Raman intensity

Wavenumber (cm

−1

)

2700

1800 1600 1400

Amide I

(a)

(b)

(c)

(a)

(b)

(c)

Amide II

Phe

Tyr

Trp

(C C)

1200 1000 800 600 400

(C N)

(S S)

fig0009 Figure 9 Raman spectra of fresh meat of (a) turkey, (b) pork, and (c) beef.

SPECTROSCOPY/Infrared and Raman 5421

0019 The intensity of Raman lines is proportional to the

concentration over several orders of magnitude. Non-

linearities due to intermolecular interactions can be

taken into account by calibration curves determined

under conditions similar to those of the practical

application. Depending on the circumstances, ‘inten-

sities’ are measured at the maximum of a band or by

integration of the band contour. The analysis of

spectra is supported by the use of sophisticated multi-

component analysis, e.g., cluster analysis or neuronal

networks.

Special Techniques

Infrared Spectroscopy

0020 For infrared spectroscopy, the common technique

uses sample cells with a lamella of a liquid sample of

a path length between 5 and 20 mm or a solution of

about 100–500 mm, the transmittance of which is

measured as a function of the wavenumber. The la-

mella is supported by infrared-transparent windows,

mostly of (hygroscopic(!)) potassium bromide (KBr).

However, solid samples need special preparation.

They can be mixed with KBr powder and pressed

into disks, or they can be ground together with an

infrared-transparent oil.

0021 The so-called ‘attenuated total reflection’ (ATR) or

‘frustrated multiple internal reflection’ (FMIR) tech-

nique is especially useful for food analyses by infrared

spectroscopy. An infrared beam coming from a ma-

terial with a large refractive index is totally reflected

at a surface adjacent to a sample having a smaller

refractive index. Under the conditions of total reflec-

tion the beam emerges into the adjacent medium as

‘evanescent wave’, thus being partially absorbed by

infrared active vibrations. The ATR technique can

be used for samples which are difficult to prepare

such as meat, cheese, vegetables or dressings.

Especially as ‘circle cell’ this technique is specially

designed for continuous analyses of flowing liquids

such as alcoholic beverages, syrups or milk products.

0022 Powdered materials can be investigated by the ‘dif-

fuse reflection’ method. By making use of multiple

reflections of the infrared beam by the powder,

components in even very low concentrations may

be detected. Analysis of samples on thin-layer chro-

matography (TLC) plates is possible directly after

extraction. (See Chromatography: Thin-layer Chro-

matography.)

0023 Investigation of volatile components is possible by

coupling of an infrared spectrometer to a gas

chromatograph either by making use of internally

reflecting glass capillaries as sample cells or by con-

densing the fractions on a cooled infrared-transparent

support. Aroma components are determined in this

way by headspace analyses. Trace analyses of toxic

components in gases, e.g., in human breath, employ

cells with mirror systems (White cells) allowing large

effective path lengths by multiple pass of the beam.

(See Chromatography: Gas Chromatography.)

0024Infrared microscopes are now available, which

employ the properties of reflecting objectives having

no chromatic aberration. The samples can therefore

be adjusted in the beam by inspection with visible

light, and afterwards, the infrared spectra in either

transmission or reflection can be run from exactly the

same spot. The spatial resolution of infrared micro-

scopes is limited by the wavelengths of the radiation

employed and is of the order of 50 mm.

Raman Spectroscopy

0025The new FT Raman spectrometers usually observe

the spectra of samples with a 180



geometry. This

means that the exciting radiation is directed on to

the sample perpendicularly to its surface. The Raman

radiation emerging from the surface is collected with

1800 1700

Trans

C C

Cis

1600

Butter (1.25)

Raman intensity

Olive oil (3.20)

Sunflower Oil (4.79)

Safflower oil (C C/C O = 5.56)

C O

C C

Sunflower margarine (2.99)

1650

Wavenumber/cm

−1

1750

fig0011Figure 11 Raman spectra of different fats: (a) safflower oil; (b)

sunflower oil; (c) olive oil; (d) sunflower margarine; (e) butter. All

spectra are normalized to the intensity of the C

—

—O vibration at

1746 cm

1

. The trans-C

—

—C is at 1662, and cis-C

—

—C at 1653 cm

1

Intensity ratio of cis-C

—

—C to C

—

—O for (a) 5.56, (b) 4.79, (c) 3.20,

(d) 2.99 and (e) 1.25.

5422 SPECTROSCOPY/Infrared and Raman

a lens system with a large aperture. The samples

are analyzed in cells, allowing multiple reflection

(Figure 4). The amount of sample is about 100 mg,

but 1 mg of a liquid or solid sample in a melting-point

tube is often sufficient; much less is needed when a

Raman microscope is applied.

0026 The Raman spectra were usually recorded with the

Bruker FT-Raman spectrometer (IFS 66 with FRA

106) and a resolution of 4 cm

1

, 350 mW of laser

power at 1064 nm with unfocused beam, and a

recording time of 15 min.

0027 Fiber-optical systems are especially useful for

food analyses by Raman spectroscopy. They connect

the spectrometer with a remote sample. Optical

fibers may have a very high transmittance in the

near-infrared range (about 80% per kilometer), and

therefore an analysis of remote process streams of

powders or fluids is possible, even of the content of

reactors.

0028 A microscope for recording of FT-Raman spectra

of small particles or inclusion is commercially avail-

able as an accessory to the instruments. It is coupled

by fiber optics to the spectrometer, which may be

placed at some distance. Raman spectra of tissues,

microorganisms and fibers, often complementary to

infrared microscopy, can be investigated. One advan-

tage of Raman microscopy is its greater spatial

resolution of the order of 5 mm.

Selected Applications

0029Figure 5 shows the infrared and Raman spectra of

bovine serum albumin. The typical vibrations of the

amide groups of proteins can be seen in the infrared as

well as in the Raman spectra (Table 2). (See Amino

Acids: Determination.)

0030Figure 6 shows the spectra of safflower oil. As

mentioned in the discussion of Figure 3b, the esters

of unsaturated fatty acids show complementary in-

tensities of the C

—

O and the C

—

C stretching vibra-

tions in the IR and Raman spectra. (See Fatty Acids:

Analysis.)

0031Microorganisms can be characterized by Raman as

well as infrared spectroscopy. Employing multivariate

statistical analysis of the second derivative of the

spectra, different strains of bacteria can be classified.

With FTIR and especially near infrared FT Raman

microscopes the identification of bacteria, viruses,

fungi, and even mammalian cells is possible (Figure 7).

0032As already discussed with Table 1 for several appli-

cations, Raman spectroscopy is better than infrared

spectroscopy. This is partially due to the fact that the

water content of food shows a very strong absorption

in the infrared range and also, since, for Raman

spectroscopy, only minimal sample preparation is

necessary. Therefore, the examples discussed below

concentrate on NIR-FT-Raman spectra.

3500

0.0000

0.0025

0.0050

0.0075

0.0100

Wavenumber (cm

−1

)

Raman units

3250 3000 2750

(a)

(b)

(c)

(d)

(e)

(f)

(g)

2500 2250 2000 1750 1500 1250 1000

fig0012 Figure 12 Raman spectra of cheeses with different concentration of fat in solid matter (%): (a) Wo

rrishofer (60), (b) Pikantje (48),

(c) Gouda (48), (d) Edam I (40), (e) Edam II (40), (f) Westlite (30), and (g) Harzer Roller (c. 3). Spectra are not processed.

SPECTROSCOPY/Infrared and Raman 5423

0033 Figure 8 represents typical spectra of important

food components: carbohydrates (banana), proteins

(turkey breast), and lipids (butter).

0034 Carbohydrates are poor Raman scatterers, revealing

vibrations at 2902 and 2941 cm

1

, based on their

large quantity of methine groups. The spectral range

from 1200 to 800 cm

1

is dominated by stretching

vibrations resulting from the carbon skeleton strongly

coupled with C—O bonding.

0035 Protein spectra demonstrate a stretching vibration

at 2939 cm

1

. The amide I-band, amide III-band, and

the vibrations originating from the protein backbone,

allow an analysis of the secondary protein structure

(see Table 2). The Raman spectra show specific vibra-

tions of amino acids with ring systems: phenylalan-

ine; 1607, 1003, 621 cm

1

; tyrosine 1611, 854,

829 cm

1

; tryptophan; 1556, 1424, 1356, 1009,

873, 755 cm

1

. The very strong S—S vibration in

the Raman spectrum between 540 and 520 cm

1

cystine is shown in Figure 3a. Cysteine shows a strong

S—H vibration at 2543 cm

1

0036 Lipids are dominated by strong stretching vibra-

tions of the methylene group at 2850 and

2884 cm

1

, indicating a large quantity of coupled

methylene groups. Several typical spectral features

are: the vibration of the ester group at 1744 cm

1

an intense bending vibration at 1440 cm

1

, a twisting

vibration at 1299 cm

1

, and vibrations of the carbon

chain between 1150 and 800 cm

1

. If unsaturated

fatty acids are present, the spectra show an additional

band, a sensitive marker of conjugation, ranging from

1655 cm

1

—

C, cisoid form) to 1670 cm

1

—

transoid form) (Figure 9).

0037Fresh meat can be distinguished by general differ-

ences of the spectra (Figure 10).

0038Carotenes show strong vibrations in the Raman

spectrum at about 1150 and 1500 cm

1

, due to the

—

C and the C—C vibrations of the whole chain.

Since the absorption bands of carotenes are not far

from the position of the exciting line, the observed

Raman spectrum show a ‘pre-resonance’ enhance-

ment. They allow the detection of carotenes, even in

low concentrations in plant and animal cells.

0039As demonstrated in Figure 11, the Raman spectra

of fats, when the spectra are normalized to the inten-

sity of the C

—

O vibration, show the relative amount

of unsaturation. The spectra demonstrate the high

number of cis-substituted double bonds in safflower

and sunflower oil but demonstrate that the fat

hardening for preparing sunflower margarine reduces

the concentration of the cis bonds and enhances the

trans-substituted double bonds.

0040Figure 12 shows spectra of cheeses with different

concentration of fat in solid matter, recorded using

the original samples. The bands due to the fat and

protein content reduce their intensity with a lower

concentration, whereas the band at 3200 cm

1

, due

to water, is nearly constant.

0041Typical noodle spectra reveal the spectral feature

of starch (amylose and amylopectin) (Figure 13 and

14). The intense band at 479 cm

1

is assigned to a

C—O bending mode. The vibrations at 845 + 5cm

1

32503500 27503000

Raman units

22502500 17502000

(a)

(b)

(c)

(d)

Wavenumber (cm

−1

)

12501500 7501000 250 0

0.0

2.5

5.0

7.5

10.0

2913.3

1659.9

1459.5

1340.0

1262.9

1126.2

940.5

865.7

479.2

355.3

12.5

500

fig0013 Figure 13 Raman spectra of noodles: (a) Spa

tzle; (b) Gold-Ei Landnudeln (durum wheat semolina, whole eggs 15%); (c) Bamboo

Garden Chinese Rice Noodles; (d) Reine Hartweizen-Schnittnudeln (100 % durum wheat semolina, no egg). All spectra are normalized

to the carbohydrate band at 1162 cm

1

5424 SPECTROSCOPY/Infrared and Raman

(a-anomers) and at 900 + 10cm

1

( b-anomers) are

marker bands indicating the type of configuration at

the anomeric C

position. However, the spectra struc-

ture correlations show that the band at 900 cm

1

not attributed exclusively to the b-anomeric configur-

ation. The variation in protein content can be moni-

tored by the band at 1660 cm

1

(amide I) in relation

to the band at 1459 cm

1

( d(C—H)).

0042The spectra of noodles are an example of a

nondestructive investigation of packaged food.

There are only small reductions of the intensity of

all bands visible. Food, even in glass containers, can

be investigated by Raman spectroscopy.

0043Finally, Figure 15 gives examples of food contain-

ing fat, carbohydrates and protein in different rela-

tive concentration with the typical bands already

1800

0.00

0.01

0.02

0.03

0.04

1747.5

1654.8

1606.1

1460.1

1440.9

1301.6

1160.4

1125.0

1084.0

1038.5

921.6

849.4

641.4

525.2

443.1

402.3

1700160015001400130012001100

Wavenumber (cm

−1

)

Raman Units

1000900800

(a)

(b)

(c)

(d)

(e)

(f)

700600500400300

fig00 15 Figure 15 Ramanspectraofchocolatesand marzipan:(a)SchwarzeHerrenschokolad e(Chocolatepuroextrafino),60%cocoa;(b)

chocolatefordiabetics,50%cocoa;(c)nougatchocolate,42%cocoa;(d)superiorsemibitterchocolate,55%cocoa;(e)milk

chocolate,30%cocoa;(d)marzipan.

35003250300027502500

0.0000

0.0025

0.0050

Raman units

0.0075

0.0100

2250 2000

(a)

(b)

Wavenumber (cm

−1

)

1750 1500 1250 1000 750 500 250

fig00 14 Figure 14 RamanspectraofnoodlesasinFigure10b: (a) in test tube; (b) directly through the original plastic bag.

SPECTROSCOPY/Infrared and Raman 5425

discussed. Of special interest is chocolate for diabetics

(Figure 15b with the typical bands of sucrose with

very low intensity).