This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

Analytica Chimica Acta 699 (2011) 66– 72

Contents lists available at ScienceDirect

Analytica Chimica Acta

jou rn al hom epa ge: www.elsev ier .com/ locate /aca

Comparison of silica gel modified with three different functional groups withC-18 and styrene–divinylbenzene adsorbents for the analysis of selected volatileflavor compounds

Mariusz Dziadasa, Magdalena Nowackab, Teofil Jesionowskib, Henryk H. Jelen a,∗

a Faculty of Food Science and Nutrition, Poznan University of Life Sciences, Poznan, Polandb Faculty of Chemical Technology, Poznan University of Technology, Poznan Poland

a r t i c l e i n f o

Article history:Received 16 February 2011Received in revised form 5 May 2011Accepted 8 May 2011Available online 14 May 2011

Keywords:Solid phase extractionGrapeWineFlavor compoundsSilica

a b s t r a c t

A comparison of SPE cartridges produced in authors laboratory containing silica modified by addition ofthree functional moieties with standard C-18 and SDVB cartridges was made in terms of their applicabilityfor the isolation of flavor compounds. Compounds found in wine and grapes were used for model mixture,which was spiked into a grape juice. Functionalized phases for SPE were prepared modifying silica gelwith alkoxysilanes with different functional groups: (3-(phenylamino)-propyltrimethoxysilane, octyl-triethoxysilane and octadecyl-silane. The functionalization was carried out by the dry method, whichresulted 5, 10 and 20 weight parts of initial support. Functionalized phases were characterized usingFT-IR, elemental analysis and NMR.

Performance for new phases compared to “standard” ones (C-18 and SDVB (styrene–divinylbenzene)varied, depending on the group or type of analyzed compound. They were more efficient in extrac-tion of methyl anthranilate and vanilins. For extraction of terpenes, C-6 alcohols, isoprenoids, benzenederivatives and phenols their efficiency was comparable to that of C-18.

Functionalized laboratory-made mixed phases are suitable for extraction of flavor compounds fromgrape juice. They are suitable for extraction of compounds belonging to different chemical classes withthe efficiency comparable to C-18 and SDVB phases. The production of such functionalized phases canbe easily performed in the laboratory, at a very low cost, comparing to C-18 or SDVB cartridges. Thismakes the proposed functionalized phases an interesting alternative, in sample preparation for analysisand particularly in preparative/flash chromatography.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

Wine volatiles are a group of several hundred compounds. Manyof them influence wine flavor and are present in the concentra-tion equal or exceeding their odor thresholds. They are presentin grapes, but are also formed during fermentation processesand during wine ageing and storage [1]. Concentration of volatilecompounds in wine range from mg/L down to ng/L. Apart from com-pounds that are present in a free form, many, especially terpenes,isprenoids and volatile phenols can be present also in a boundform. Volatile/odoriferous aglycon is bound with mono or disac-charide. They are present in both forms in wine as well as in grapes[2–4]. Glycosides can undergo hydrolysis as a result of low pH envi-ronment in wine during its storage and of enzymatic hydrolysis[5]. Flavor aglycons released from glycosidally bound precursorsinfluence wine flavor and in grapes flavor potential. Therefore, the

∗ Corresponding author. Tel.: +48 61 8487273; fax: +48 61 8487314.E-mail address: henrykj@up.poznan.pl (H.H. Jelen).

analysis of grapes and wine aroma compounds comprise of analysisof free and bound odorants.

For the analysis of food aroma compounds—their extrac-tion and preconcentration, headspace methods based oncompounds volatility, solvent extraction and methodsbased on adsorption/absorption mechanisms are used. Theextraction methods determine often the profile of isolatedcompounds [6,7].

Among sorbents used in extraction of food aroma compoundsseveral groups can be distinguished: silica gels (polar due to theirhydroxyl groups), activated aluminas (polar), activated carbon(apolar), zeolites and polymers, such as polystyrene, polyacrylilcesters, PDMS and phenolic resins [8]. Extraction and preconcen-tration of wine flavor volatile compounds are challenging tasksfor the analyst, therefore various isolation techniques are used:liquid–liquid extraction, solid phase microextraction (SPME), stirbar sorptive extraction (SBSE) and solid phase extraction (SPE). Foran overview of extraction methods used for wine and other oeno-logical products a recent review by Castro et al. [9] can supply moreinsight into these applications.

0003-2670/$ – see front matter © 2011 Elsevier B.V. All rights reserved.doi:10.1016/j.aca.2011.05.011

Author's personal copy

M. Dziadas et al. / Analytica Chimica Acta 699 (2011) 66– 72 67

Due to the nature of wine and grape volatile compounds—theirpresence in free and bound forms—extraction methods basedon sorption are commonly used. The SPE approach allows frac-tionation of free forms of volatiles from nonvolatiles and theirbound forms and also permits a concentration of volatiles. Pre-concentration is an important step, as many of the odoriferousvolatiles occur in low concentrations. Generally, sample intro-duced into SPE cartridge is eluted with a nonpolar solvent, whichelutes free compounds. Glycosidally bound flavor compounds,which are polar fraction are retained on the column and elutedfrom it using methanol or methanol/water. To analyze agly-cons in bound form fast acid hydrolysis is usually performedand aglycons separated from sugars on a second SPE cartridgeand subsequently analyzed by GC/MS [10]. Due to the range ofconcentrations in which flavor compounds of wine are presentsometimes a fractionation using solvents of different polaritiesis advised. Compounds present usually in very high concentra-tions such as ethyl esters are separated from those occurring intrace amounts, though important from the sensory point of view[11].

For the extraction of small molecules from aqueous solu-tion chemically bonded sorbents having large surface areaswith long alkyl chains and high phase loading maximizetheir retention. Macroreticular porous polymers represented bycopolymers of styrene–divinylbenzene (SDVB)—particles consist-ing of agglomerates of randomly packed microspheres with anetwork of holes and channels. They are usually more reten-tive than chemically bonded phases and have a high samplecapacity.

The SPE analysis of free and bound flavor compounds in grapeand wine is usually done using reversed phase SPE [12–14] orAmberlite XAD-2 resins [15]. The latter one and Lichrolut EN is themost effective for the extraction of acids, benzene derivatives andvolatile phenols [10], whereas RP SPE proved to be the most effec-tive for extraction of terpenes [10]. When Pineiro et al. [16] testedvarious C-18 and SDVB cartridges for the extraction of terpenesfrom wine when using dichloromethane for compounds elutionnoticed better recoveries for these compounds for SDVB cartridge.Macroporous polydivinylbenzene-N-vinylpirrolidone copolymertested for the extraction of chloroanisols from water and 12%ethanol was much more efficient than C-18 [17]. As observedby Cullere et al. [11] silica sorbents present low capacity andsome compounds can bind irreversibly, whereas styrene divinyl-benzene have better loading capacity, stability and ability to beused in normal as well as reversed mode. Lopez et al. [18] foundSDVB copolymer (LiChrolut-EN) satisfactory for analysis of 27volatiles from wine. Extraction of polar compounds together withmid polar and nonpolar in alcoholic beverages often cannot beperformed satisfactory due to the different interactions of thesecompounds with SPE beds. As tested by Lukic [19] higher alcohols(butanols) and some short chain esters provided low recoverieson C-18.

As the aroma volatiles in wine comprise compounds of differentmechanism of binding, ranging from polar to nonpolar, aromaticand aliphatic, the development of functionalized adsorbents withdifferent mechanisms of volatiles binding can provide more effi-cient extraction of volatiles. Our goal was to develop a SPE phases,containing one cartridge different functional groups to enhancespectrum of extracted compounds and use it for extraction of com-pounds of different characters that are found in wine, grapes andother fruit in a model system of grape juice spiked with mixture of53 aroma compounds. To verify the effectiveness of modified sil-ica produced in our lab in aroma compounds extraction they werecompared with the two most popular phases commercially avail-able for this type of analysis—C-18 and styrene–divinylbenzenecartridges.

2. Materials and methods

2.1. Volatile compounds

All standards of volatile compounds used for method perfor-mance evaluation were purchased from Sigma–Aldrich, Poznan,Poland and were of highest available purity. Compounds wereselected to reflect those present in grape juice and wine, but alsoin other fruits that could influence their flavor. Compounds ana-lyzed represented the following classes: monoterpenes (�-pinene,�-pinene, limonene, �-trans-ocimene, �-terpinene, �-cis-ocimene,terpinolene, trans-rose oxide, cis-rose oxide, trans-linalool oxide,cis-linalool oxide, citronellal, linalool, linalyl acetate, menthylacetate, geranyl acetate, terpinyl acetate, terpinen-4-ol, men-thol, �-terpineol, citral, �-citronellol, nerol, myrtenol, geraniol);isoprenoids (�-damascenone, �-damascone, �-ionone, �-ionone,cis-pseudoionone, trans-pseudoionone); benzene derivatives (ben-zaldehyde, benzyl alcohol, benzeneacetaldehyde, phenylethan-2-ol, 1,2-methyl-4-propenyl benzene); volatile phenols (guaiacol,2,6-dimethylphenol, o-cresol, p-cresol, m-cresol, ethylguaiacol,eugenol, 4-ethyl phenol, trans-isoeugenol, 2,6-dimethoxy phe-nol); and miscellaneous compounds not included in previouslymentioned classes (methyl anthranilate, vanillin, methylvanilin, �-decalactone, �-nonanactone, cis-3-hexene-1-ol, cis-hexene-2-ol).As internal standards [2H7]-geraniol, [2H3]-�-ionone and [2H8]-napthalene were used. Deuterated geraniol was synthesized inauthors laboratory according to procedure described by Peder-sen et al. [20] and deuterated �-ionone–according to proceduredescribed by Kotseridis et al. [21]. Deuterated naphthalene waspurchased at Sigma (Poznan, Poland).

2.2. Preparation of sorbents

For the experiments reference sorbents were purchased fromBiotage, Sweden (ISOLUTE cartridges). They were 500 mg/6 mL car-tridges filled with C-18 and SDVB. For the preparation of selectiveadsorbents two silica gels—CCH Silica gel 60, 0.05–0.2 mm, and HRSilica gel 60, TLC HR “high purity” Merck, Darmstadt, Germany werepurchased.

The adsorbent’s silica surface was modified with selectedalkoxysilanes with different functional groups (3-(phenylamino)-propyltrimethoxysilane, octyltriethoxysilane and octadecyl-silanepurchesd from Fluka. The appropriate modifying agents were sep-arately hydrolyzed with methanol/water solutions (4:1, v/v) andsprayed onto silica gel surface applying “dry method” describedin detail in [22,23]. The silane coupling agents were deposited inthe amount of 5, 10, and 20 weight parts by mass of initial sup-port (CCH-5, CCH-10, CCH-20, HR-5, HR-10, HR-20). The solventswere subsequently removed by distillation. Scheme of adsorbentspreparation is shown in Fig. 1.

The degree of modification and the type of reaction betweenmodifiers and silica surface were evaluated using FT-IR EQUINOX55 spectrophotometer made by Bruker. Elemental analysis of thesamples studied (C, H, N contents) was performed on an instrumentElementar model Vario EL III, Germany. NMR analysis of unmodifiedand modified silicas was conducted in DSX spectrometer (Bruker).For the determination of NMR spectra a sample of about 100 mgwas placed in a rotator made of ZrO2 with 4 mm of diameter, whichpermitted spinning of the sample. Centrifugation at the magic anglewas performed at the spinning frequency of 8 kHz. 29Si CP MAS NMRspectra were recorded at the pulse duration of 4.5 �s, contact timeof 1.5 ms and pulse spacing of 6 s. The 13C CP MAS NMR spectra wererecorded at 100.63 MHz in a standard 4 mm MAS probe by applica-tion of single pulse excitation with high power proton decoupling(pulse repetition 10 s and spinning speed 8 kHz).

Author's personal copy

68 M. Dziadas et al. / Analytica Chimica Acta 699 (2011) 66– 72

Fig. 1. The scheme of sorbents preparation procedure.

2.3. Sample preparation for volatile compounds analysis

A stock solution of volatile compounds was prepared in ethanol.From the stock solution working solutions of pasteurized grapejuice produced from concentrated juice (bought in a local super-market) were prepared by spiking it with the stock solution. Thefinal concentration of analyzed compounds in spiked grape juiceranged from 3 to 450 �g/20 mL. Working solutions were spikedwith internal standards. As internal standards [2H7]-geraniol,[2H3]-�-ionone and [2H8]-naphthalene were used to quantify ter-penes, isoprenoids and remaining compounds, respectively. Avolume of 20 mL of the spiked juice was introduced into a SPEcartridge containing 500 mg of tested sorbent in a 6 mL reser-voir previously preconditioned. Preconditioning was performedby purging at 3 mL/min the column with 25 mL portions of pen-tane/dichloromethane (2/1, v/v) followed by 25 mL of methanoland 25 mL of water. After loading a sample onto the columnit was washed with 150 mL of water. Fraction containing com-pounds of interest was eluted using 25 mL of a mixture ofpentane/dichloromethane (2/1, v/v) and evaporated to approxi-mately 500 �L firstly heating it at 30 ◦C water bath without mixingor stirring for 30 min, then in a delicate stream of nitrogen. Theanalysis was performed with 1 �L of the concentrated extract insplit less mode into GC/MS system.

2.4. GC/MS analysis of investigated volatiles

Analyses were performed using Agilent 7890GC A witha split/splitless injector coupled to Agilent 5975 C VL MSDquadrupole mass spectrometer equipped with a triple axis detec-tor (TAD) and a diffusion pump. Samples were injected usingAgilent 7683B autosampler in a splitless mode at 220 ◦C usingdouble gooseneck liner. Analyzed compounds were resolved ona Supelcowax-10 column (60 m × 0.200 mm × 0.200 �m, Supelco,Bellefonte, PA). Analyses were performed at programmed temper-ature starting from 40 ◦C for 1 min, then increased at 4 ◦C/min to180 ◦C (5 min), followed by an increase of 4 ◦C/min to 240 ◦C, atwhich the temperature was kept constant for 20 min. The total runtime was 76 min. GC/MS transfer line temperature was 240 ◦C, ionsource temperature was 230 ◦C and quadrupole at 150 ◦C. Spectrawere acquired using electron impact ionization (EI, 70 eV) in a fullscan mode, using scan range of m/z 33–333 Da. Detector was run at

a gain factor (GF = 1) mode at EMV voltage of 1760 V. Helium flowwas 0.8 mL/min (32.4 cm/s).

Quantitation of compounds of interest was based on usinginternal standards—[2H7]-geraniol, [2H3]-�-ionone and [2H8]-naphtalene (Sigma). Monoterpenes were quantified using deuter-ated geraniol, isoprenoids—using deuterated �-ionone and ben-zene derivatives and volatile phenols using deuterated naphtha-lene. Modified silica phases and also SDVB ability to extract volatileswas compared with reference one (C-18) in the following way:standards mixture was spiked into the grape juice in various con-centrations, calibration curve was calculated for each compoundfor C-18 phase covering concentration range of approximately 3orders of magnitude. Compounds were quantified based on theirtarget ions, in few cases using TIC (total ion chromatograms). Cal-culated calibration curves equations and their linearities, togetherwith targeted ions used for quantitation for analyzed compoundsare shown in Table 1. Extraction efficiencies using other sorbentswere determined quantifying compounds using data provided inthis table, based on C-18 calibration curve.

2.5. Statistical analysis

All samples were run in triplicates. For the determination ofthe least differences Tukey HSD test was performed (Statistica 9.0,Statsoft Inc., Tulsa, USA).

3. Results and discussion

3.1. Characterization of mixed phase adsorbents

The adsorbents obtained in the author’s laboratory were char-acterized in terms of composition and structure, using FT-IR,elemental analysis and NMR. All the measurements were per-formed on HR silica subjected to modifications.

Fig. 2 presents the FT-IR spectra of unmodified HR sil-ica gel (sample 1) and silica modified with a mixture ofthe following silanes: 3-(phenylamino)propyl-trimethoxysilane,octyltriethoxysilane and octadecylsilane used in the amounts ofeither 5, 10 or 20 wt. parts per 100 wt. parts of SiO2 each (sam-ples 2, 3 and 4 respectively). The spectra show broad bands in therange 3200–3600 cm−1 coming from the OH stretching vibrationscharacteristic of physisorbed water. As a result of modification, the

Author's personal copy

M. Dziadas et al. / Analytica Chimica Acta 699 (2011) 66– 72 69Ta

ble

1C

omp

oun

ds

use

d

in

exp

erim

ents

, th

eir

cali

brat

ion

curv

es

equ

atio

ns

use

d

for

ph

ases

com

par

ison

(C-1

8), l

inea

rity

and

ion

s

use

d

for

com

pou

nd

s

quan

tita

tion

and

iden

tifi

cati

on. C

alib

rati

on

curv

es

for

SPE

met

hod

usi

ng

HR

-10

labo

rato

ry-m

ade

mod

ified

sili

ca

gel,

lin

eari

ty

(for

the

sam

e

con

cen

trat

ion

ran

ge

as

C-1

8)

lim

its

of

det

ecti

on

(LO

D)

and

pre

cisi

on

(RSD

)

are

also

incl

ud

ed

in

the

tabl

e.

No

Com

pou

nd

sC

-18

equ

atio

nR

2Ta

rget

ion

Qu

al

1Q

ual

2H

R10

equ

atio

nR

2LO

D

(pp

b)R

SD

1�

-Pin

ene

1.05

e

+

000

×x

−9.

70e

−00

30.

999

9377

121

1.17

e

−

001

×

x

−

1.12

e

−

003

0.99

9

11

3.9

2�

-Pin

ene

1.09

e

+

000

×x

+

2.29

e

−00

30.

999

9312

113

69.

95e

−00

2

×x

+

3.21

e

−00

40.

998

133.

73

Lim

onen

e

7.78

e

−

001

×

x

+

5.43

e

−

003

0.99

9

68

93

107

8.50

e

−

002

×

x

−

5.64

e

−

004

0.99

9

10

2.9

4�

-tra

ns-O

cim

ene

5.23

e

−00

1

×x

−2.

07e

−00

30.

999

9379

105

5.45

e

−00

2

×x

+

4.49

e

−00

50.

997

83.

75

�-T

erp

inen

e

1.25

e

+

000

×x

+

4.81

e

−00

10.

961

9312

113

61.

28e

−00

1

×x

−3.

54e

−00

30.

998

21.

26

�-c

is-O

cim

ene

3.39

e

−00

1

×x

+

2.38

e

−00

20.

997

9379

105

4.71

e

−00

2

×x

−3.

18e

−00

31

92.

27

Terp

inol

ene

7.55

e

−00

1

×x

−1.

29e

−00

20.

999

9312

113

67.

90e

−00

2

×x

−4.

67e

−00

30.

997

161.

18

cis-

Ros

e

oxid

e2.

50e

+

000

×x

+

2.02

e

−00

20.

999

139

154

2.92

e

−00

1

×x

−6.

49e

−00

30.

997

81.

69

tran

s-R

ose

oxid

e2.

62e

+

000

×x

+

2.12

e

−00

20.

999

139

832.

97e

−00

1

×x

−1.

42e

−00

20.

997

42.

310

cis-

Lin

aloo

l oxi

de

6.75

e

−00

1

×x

+

1.22

e

−00

20.

999

5994

111

8.85

e

−00

2

×x

+

4.75

e

−00

30.

999

112.

711

tran

s-Li

nal

ool o

xid

e6.

89e

−00

1

×x

+

1.29

e

−00

20.

999

5911

194

1.08

e

−00

1

×x

+

8.90

e

−00

30.

999

64.

412

Cit

ron

ella

l6.

06e

−00

1

×x

−3.

22e

−00

20.

999

4195

121

8.14

e

−00

2

×x

−3.

79e

−00

30.

999

71.

413

Lin

aloo

l6.

40e

−00

1

×x

−2.

36e

−00

30.

999

7193

121

8.14

e

−00

2

×x

−3.

79e

−00

30.

999

83.

314

Lin

alyl

y

acet

ate

1.02

e

+

000

×x

−2.

02e

−00

30.

999

9380

121

1.20

e

−00

1

×x

−6.

55e

−00

30.

998

113.

315

Men

thyl

acet

ate

9.99

e

−00

1

×x

+

4.49

e

−00

20.

999

9513

812

31.

14e

−00

1

×x

+

2.79

e

−00

40.

998

61.

916

Terp

inen

-4-o

l6.

55e

−00

1

×x

+

1.14

e

−00

20.

999

7111

193

7.99

e

−00

2

×x

+

2.52

e

−00

30.

999

62.

417

Terp

inyl

acet

ate

9.00

e

−00

1

×x

+

3.15

e

−00

20.

999

121

9313

68.

86e

−00

2

×x

−3.

93e

−00

40.

999

63.

218

Men

thol

6.13

e

−00

1

×x

+

4.65

e

−00

20.

999

7181

123

7.65

e

−00

2

×x

+

6.09

e −

003

0.99

88

2.0

19�

-Ter

pin

eol

7.68

e

−00

1

×x

+

3.04

e

−00

20.

999

5993

121

1.00

e

−00

1

×x

+ 4.

68e

−00

30.

999

92.

620

Cit

ral

7.64

e

−00

1

×x

−1.

10e

−00

20.

998

6984

942.

18e

−00

2 ×

x +

2.80

e

−00

30.

964

72.

721

Ger

anyl

acet

ate

1.41

e

+

000

×x

+

1.74

e

−00

20.

999

69

93

136

9.56

e

− 00

2 ×

x

−

9.79

e

−

004

0.99

8

12

3.2

22�

-Cit

ron

ello

l

1.83

e

+

000

×x

+

1.74

e

−00

20.

999

6982

123

7.50

e

−00

2 ×

x

+

3.04

e

−00

30.

997

81.

323

Ner

ol9.

33e

−

001

×

x

+

2.26

e

−

002

0.99

9

69

93

121

1.19

e −

001

×

x

+

6.56

e

−

003

0.99

8

7

3.0

24M

yrte

nol

1.05

e

+

000

×x

+

4.04

e

−00

20.

999

7991

108

1.36

e −

001

×

x

+

6.25

e

−

003

0.99

8

7

2.3

25

Ger

anio

l

2.30

e

+

000

×

x

+

6.65

e

−

002

0.99

9

69

93

123

1.35

e

−

001

×

x

+

6.75

e

−

003

0.99

8

7

3.2

27�

-Dam

asco

ne

1.06

e

+

000

×x

+

6.25

e

−00

20.

999

177

192

123

9.09

e

−00

2

×x

−

1.28

e

−

002

0.99

98

1.5

28�

-Dam

asce

non

e

1.08

e

+

000

×

x

+

1.01

e

−

001

0.99

9

121

190

175

9.42

e

−

002

×

x

−

1.54

e

−

002

0.99

3

1.1

29

�-I

onon

e

1.10

e

+

000

×

x

+

9.65

e

−

002

0.99

9

121

93

136

9.52

e

−

002

×

x

−

1.23

e

−

002

0.99

9

9

1.2

30�

-Ion

one

2.04

e

+

000

×x

+

1.66

e

−00

10.

999

177

135

911.

77e

−00

1

×x

−2.

11e

−00

20.

998

102.

831

tran

s-Ps

eud

oion

one

1.17

e

+

000

×x

+

2.19

e

−00

30.

999

69

81

124

1.11

e

−

001

×

x

−

2.08

e

−

002

0.99

8

15

1.7

32

cis-

Pseu

doi

onon

e

8.21

e

−

001

×

x

+

2.86

e

−

002

0.99

9

69

81

109

7.50

e

−

002

×

x

−

7.07

e

−

003

0.99

9

9

2.7

34B

enze

nal

deh

yde

2.40

e

−00

1

×x

+

2.00

e

−00

20.

999

7710

53.

42e

−00

2

×x

−3.

80e

−00

20.

995

112.

335

Ben

zen

acet

alal

deh

yd3.

89e

−00

1

×x

−1.

13e

−00

10.

999

9112

065

5.00

e

−

002

×

x

−

2.02

e

−

002

0.99

5

6

3.7

36

Ben

zyl a

lcoh

ol

1.99

e

−

001

×

x

+

3.34

e

−

002

0.99

9

139

83

111

1.70

e

−

002

×

x

+

5.35

e

−

003

0.98

9

12

2.2

37Ph

enyl

eth

an-2

-ol

4.07

e

−00

1

×x

+

4.59

e

−00

20.

999

9112

25.

94e

−00

2

×x

−1.

88e

−00

20.

996

91.

238

1,2-

Met

hyl

-4-p

rop

enyl

ben

zen

e2.

38e

−00

1

×x

+

3.06

e

−00

20.

999

162

104

3.23

e

−00

2

×

x

−

1.49

e

−

002

1

21

2.3

40

Gu

aiac

ol

3.07

e

−

001

×

x

+

3.96

e

−

002

0.99

9

109

124

81

318e

−

002

×

x

−

6.61

e

−

003

0.99

8

8

3.6

412,

6-D

imet

hyl

ph

enol

2.46

e

−00

1

×x

+

1.97

e

−00

20.

999

122

107

3.46

e

−00

2

×x

−1.

31e

−00

20.

996

171.

842

o-C

reso

l2.

66e

−00

1

×x

+

2.44

e

−00

20.

999

108

793.

29e

−00

2

×x

−5.

89e

−

003

0.99

87

3.1

43

Eth

ylgu

aiac

ol

5.03

e

−

001

×

x

+

3.11

e

−

002

0.99

9

137

152

6.64

e

−

002

×

x

−

2.76

e

−

002

0.99

9

11

2.1

44

p-C

reso

l

3.64

e

−

001

×

x

+

4.41

e

−

002

0.99

9

107

77

3.96

e

−

002

×

x

−

2.50

e

−

003

0.99

6

9

3.8

45m

-Cre

sol

2.69

e

−00

1

×x

+

5.53

e

−00

20.

999

108

772.

84e

−00

2

×x

−1.

90e

−00

30.

998

123.

346

Euge

nol

2.25

e

−

001

×

x

+

3.14

e

−

002

0.99

9

164

103

2.62

e

−

002

×

x

−

8.79

e

−

003

0.99

8

9

2.3

47

2,6-

Dim

eth

oxy

ph

enol

2.65

e

−

001

×

x

−

2.41

e

−

003

0.99

9 15

4

139

3.25

e

−

002

×

x

−

1.79

e

−

002

0.99

7

24

3.7

484-

Eth

ylp

hen

ol6.

42e

−00

2

×x

−5.

00e

−00

40.

998

107

122

6.42

e

−00

2

×x

−5.

00e

−00

40.

998

133.

849

tran

s-Is

oeu

gen

ol2.

57e

−00

1

×x

−2.

82e

−

003

0.99

916

414

9

2.65

e

−

002

×

x

−

4.50

e

−

002

0.99

7

13

4.0

50

cis-

3-H

exen

e-1-

ol

5.99

e

−

001

×

x

+

2.45

e

−

002

0.99

9

41

67

82

7.44

e

−

002

×

x

+

8.56

e

−

003

0.99

8

4

2.2

51ci

s-2-

Hex

ene-

1-ol

8.05

e

−00

1

×x

+

2.89

e

−00

20.

999

5785

991.

08e

−00

1

×x

+

8.90

e

−00

30.

999

102.

052

�-N

onal

acto

ne

2.24

e

+

000

×x

+

9.69

e

−00

20.

999

TIC

2.89

e

−00

1

×

x

+

8.49

e

−

003

0.99

9

8

1.6

53

�-D

ecal

acto

ne

2.30

e

+

000

×

x

+

1.12

e

− 00

1 0.

999

TIC

2.85

e

−

001

×

x

+

6.36

e

−

003

0.99

9

9

1.6

54M

eth

yl

anth

ran

ilat

e4.

32e

−00

2

×x

+

1.06

e −

002

0.99

911

915

192

4.09

e

−00

2

×x

+

1.63

e

−00

20.

998

612.

255

Van

illi

n7.

72e

−00

1

×x

+

1.79

e −

001

0.99

9TI

C9.

33e

−00

2

×

x

−

3.74

e

−

002

0.99

1

23

6.1

56

Met

hyl

van

ilin

3.31

e

−

001

×

x

+ 7.

12e

−

002

0.99

9

TIC

1.21

e

−

001

×

x

−

1.96

e

−

002

0.99

7

77

5.9

Author's personal copy

70 M. Dziadas et al. / Analytica Chimica Acta 699 (2011) 66– 72

Fig. 2. FT-IR spectra of HR silica: unmodified (sample 1) and modified with a mixtureof silanes (samples 2–4).

intensity of the band assigned to silanol groups decreases withrespect to that in the spectrum of the unmodified silica, which isa result of condensation of these groups with the hydroxyl groups.The spectrum of the unmodified silica also reveals a low-intensityband characteristic of even silanol groups at 3748 cm−1 (isolatedsilanols). In the spectra of the modified silica the intensity of thisband decreases. The spectrum of the unmodified silica also showsthe band assigned to the C–H stretching vibrations in the range2965–2850 cm−1. The intensity of this band increases significantlywith increasing amount of the modifying mixture. The best effectof modification is observed when the modifier is 20 w/w of silane,because when this amount of modifier was used the most pro-nounced disappearance of silanol groups was observed. Moreover,the intensity of the band assigned to C–H considerably increases.The spectrum of unmodified silica also shows a band assigned to thehydroxyl groups present at the silica surface at 1640 cm−1, a bandassigned to the stretching vibration of Si–O–Si bond at 1106 cm−1

and a band assigned to the NH bending vibration in the range1500–1600 cm−1.

Table 2 provides the elemental composition and the degreeof silica surface coverage calculated for different amounts of themodifiers used. The silica surface coverage degree was evaluatedaccording Berendsen and de Golan equation [24]. The results showa significant increase in the content of carbon, hydrogen and nitro-gen with increasing amount of the silanes used for modification.The degree of coverage increases from 0.49 �mol/m2 for the silicamodified with the mixture of silanes in the amount of 5 w/w to2.85 �mol/m2 for the silica modified with the mixture of silanes inthe amount of 20 w/w.

Fig. 3a and b presents spectra recorded by two techniquesof the atomic magnetic resonance: 13C and 29Si CP MAS NMRof the silica unmodified and modified with a mixture of 3-(phenylamino)propyltrimethoxysilane, octyltriethoxysilane andoctadecylsilane used in the amounts 5, 10 or 20 wt. parts per 100 wt.

Table 2Elemental composition of laboratory-made SPE phases (HR Silica gel with modifi-cations. Sample names explained in the text.

Sample name Elemental content (%) Coverage degree (�mol/m2)

C H N

1 0.029 0.804 – –2 5.405 1.094 0.232 0.493 11.18 2.265 0.433 1.114 22.58 4.327 0.658 2.85

Fig. 3. (a) 29Si CP MAS NMR and (b) 13C CP MAS NMR spectra of HR silica unmodified(sample 1) and modified with a mixture of silanes (samples 2–4).

parts of SiO2. The results confirm the chemical character of interac-tions between the silica surface and the mixture of silanes used formodification. The spectrum 29Si CP MAS NMR of the unmodifiedsilica shows three bands at −110, −100 and −99 ppm. The mainband at ı = −110 ppm is assigned to a structure of three siloxanegroups and one silanol group—isolated silanols (Q3), the band at−100 ppm—to four siloxane bridges (Q4), while the one at −99 ppm(the so-called silicon site Q2) is assigned to double silanol groups.In the spectrum taken after modification of silica the intensities ofthe signals Q4 and Q3 significantly increase. The presence of bandscorresponding to the structures T2 (−56.6 ppm) and T3 (−66.0 ppm)informing about the condensation of two or three alkoxyl groupsof a given silane to the silica matrix, respectively, particularly wellvisible for the sample modified with the silanes used in the amountof 20 w/w, is the evidence for the chemisorption of the modifiersapplied and a proof of effective modification.

Another proof of the chemisorption of the modifiers is given bythe spectra presented in Fig. 3b. The signals appearing at 21 and29 ppm are assigned to –CH3 groups, while that at 49 ppm to –CH2groups from the mixture of silanes used. A distinct band at 75 ppmis assigned to the Si−O−CH3 group, while the signals at 128 and147 ppm—to the ring of the aromatic modifier.

To assess the reproducibility of procedure of obtaining func-tionalized silica adsorbent HR silica gel was chosen with 10 w/wof (3-phenylamino)-propyltrimethoxysilane plus 10 w/w of octyl-triethoxysilane and plus 10 w/w of octadecylsilane for testing.

Author's personal copy

M. Dziadas et al. / Analytica Chimica Acta 699 (2011) 66– 72 71

Fig. 4. Comparison of extraction efficiencies of produced modified silica phases with commercially available C-18 and SDVB. For better illustration C-18 extraction efficiencyis projected as 100%, remaining phases are compared to C-18. Bars show the relative standard deviations. Phases codes are the same as in Fig. 1.

Apart from initial batch used for experiments (Batch A) three otherbatches were produced separately (B, C, D). For all the batches ele-mental content was determined as well as coverage degree. Table 3illustrates batch-to-batch reproducibility of adsorbents produc-tion. The coverage degree between batches differed less than 5% andthe elemental composition measured from batch to batch differedat the maximum (for nitrogen) at nearly 9% (Table 3)

3.2. Analysis of flavor compounds

Volatile compounds were resolved on the 60 m Wax column,which provided satisfactory separation of analyzed compoundsspiked into grape juice. In case of partially overlapping peaks useof mass spectrometer as detector allowed their quantitation basedon extracted ions.

The results obtained for the C-18 cartridge were used as a refer-ence and scaled to 100% (Fig. 4). Adsorption of particular groups ofcompounds on different phases differed when compared to C-18.For terpenes the performance of SDVB was about 20% lower thanfor C-18. Similar effect was observed for CCH 20. The best extrac-tion efficiencies were obtained for Ultra Pure for TLC 60 silica gel(HR 5, HR 10 and HR 20) although they were not better than for C-18. For C-6 alcohols CCH 5 was the most efficient phase with 130%of extraction efficiency of C-18. For isoprenoids extraction efficien-cies were similar for all tested phases except HR 5, which showed116% efficiency compared to C-18. For benzene derivatives C-18provided the best extraction efficiency, whereas CCH 5 the low-est. For volatile phenols only CCH 20 and HR 20 provided slightlybetter responses than C-18 (106% and 105% respectively). For theextraction of methyl anthranilate CCH 20 provided response of

Table 3Batch-to-batch reproducibility of modified HR silica with 10 w/w of each modifier,produced in authors laboratory.

Batch Elemental content (%) Coverage degree (�mol/m2)

C H N

A 11.180 2.265 0.433 1.11B 12.174 2.699 0.468 1.21C 12.570 2.650 0.396 1.25D 11.719 2.351 0.385 1.18Average 11.911 2.491 0.420 1.187RSD (%) 5.02 8.64 8.97 4.97

129% compared to C-18, similarly for the extraction of vanilins itwas 138% of C-18. For the extraction of vanillins all tested phasesperformed better than C-18.

For the determination of the least differences Tukey HSD testwas used. Samples were tested at a significance level = 0.05.Formed uniform groups indicate the smallest differences betweentested adsorbents within first group—SDVB, CCH 5 and CCH 20,second group (HR 5, HR 10, HR 20), third group (CCH 10, SDVB)and the forth one (C-18 and HR5) (Table 4). Moreover, the clus-ter analysis (raw data similarities) confirmed Tukey test results.Two main clusters were distinguished: 1—CCH 20, CCH 10, CCH 5,SDVB and cluster 2—HR 10, HR 20, HR 5, C-18 (Fig. 5). The rankingof the results may indicate that the coating of functional groupsonto silica was effective—the efficiency of the extraction of volatilecompounds as a result of adequate binding of the functional groupson silica yielded very similar results to C-18 in case of HR 5. Forthe performance evaluation of functionalized silica HR with 10 w/wgroups (HR 10) a basic SPE method characteristics were providedin Table 1. Data provided in this table referring to C-18 cartridgeswas used for determination of produced adsorbents (their recov-ery compared to C-18). Data provided for HR-10 column provideinformation on its linearity limits of detection and precision. Forall analyzed compounds in a broad range of analyte concentration(Section 2) linearity was satisfactory. Limits of detection rangedfrom 2 �g/L (gamma terpinene) to 77 �g/L (methyl vanillin). Forthe majority of compounds it was 10 �g/L or less. Also the RSD val-ues (n = 3) for peak areas of analyzed compounds was in majorityof compounds <5%. This data indicate the usefulness of producedadsorbents for the analysis of volatile flavor compounds.

Table 4Determination of the differences in adsorbents analyzed using Tukey HSD test,a variable replication, homogenous groups, alpha = 0.05, error: MS betweengroups = 185.94, df = 896.00.

No Adsorbents# Replicates## 1 2 3 4

7 CCH10 201,6792 ****1 SDVB 206,1015 **** ****5 CCH5 207,5462 ****6 CCH20 208,1948 ****4 HR10 224,2168 ****3 HR20 225,1493 ****2 HR5 228,0322 **** ****8 C-18 229,8502 ****

Author's personal copy

72 M. Dziadas et al. / Analytica Chimica Acta 699 (2011) 66– 72

Tree Diagra m for 8 Variabl esWard`s metho d

Squared Eu clidean di stan ces

c18hr5hr20hr10sdvbcch5cch10cch200

1E5

2E5

3E5

4E5

5E5

6E5

7E5

Link

age

Dis

tanc

e

Fig. 5. Dendrogram of obtained functionalized silica phases in comparison withSDVB and C-18 phases.

4. Conclusion

Phases used for SPE, produced in authors laboratory whencompared to a reference C-18 and SDVB ones showed different effi-ciencies in extraction of volatile flavor compounds from grape juice.Usually the efficiencies were similar to commercially available C-18 or SDVB phases with several exceptions (methyl anthranilate,vanillins, C-6 alcohols and isoprenoids). The modified adsorbentscan provide better performance for the analysis of selected volatileflavor compounds where the extraction efficiency increased at itsmaximum to 130% in relation to C-18. For benzene derivativeslaboratory-made phases were inferior to C-18 and SDVB. Of all pro-duced and tested phases these produced using ultra pure silica 60gel for TLC (HR) provided results on average the most similar toC-18 and SDVB phases tested. Therefore, this silica gel can be rec-ommended for use for flavor compounds isolation, despite slightlybetter performance for other silica for selected compounds.

The idea of using three modifiers of silica gel usually not usedtogether in one cartridge allowed a broad spectrum of analytesreact with the modified silica surface yielding a universal phasesfor extraction of flavor compounds of different character from grapejuice.

However, the most important outcome is a simple method offunctionalization of silica gel phases that can be produced, due toits simplicity, in any laboratory, at the very low cost compared to

C-18 and especially to SDVB cartridges. As presented in Fig. 4 andin Table 1. The repeatability for all groups of compounds is satisfac-tory, and one has to remember that in-laboratory-made and packedcartridges were used for experiments. It has proven to be veryreproducible, observing batch-to-batch elemental composition andsurface coverage. The very low cost of produced modified silicamakes it especially attractive, besides analytical size cartridges,for the flash chromatography or preparative chromatography pur-poses.

Acknowledgement

Polish Ministry of Science and Higher Education is acknowl-edged for financing the project under project number N N312215538

References

[1] S. Ebeler, Food Rev. Int. 17 (2001) 45.[2] M. Vilanova, C. Sieiro, J. Food Composition Anal. 19 (2006) 694.[3] R. Di Stefano, J.J. Mateo, M. Jimenez, N. Gentilini, T. Huerta, J. Chromatogr. A

778 (1997) 219.[4] Z. Gunata, C.L. Bayonove, R.L. Baumes, R.E. Cordonnier, J. Chromatogr. A 331

(1985) 83.[5] S.J. Macais, J.J. Mateo, Appl. Microb. Biotechnol. 67 (2005) 322.[6] M. Ortega-Heras, Beltrán Gonzales-SanJosè, Anal. Chim. Acta 458

(2002) 85.[7] C. Dominguez, D.A. Guillén, C.G. Barroso, Anal Chim. Acta 458 (2002) 95.[8] A. Nongonierma, P. Cayot, J.-L. Le Quéré, M. Springett, A. Voilley, Food Rev. Int.

22 (2006) 51.[9] R. Castro, R. Natera, E. Durán, C. Garcia-Baroso, Eur. Food Res. Technol. 228

(2008) 1.[10] M.J. Ibarz, V. Ferreira, P. Hernandez-Orte, N. Loscos, J. Cacho, J. Chromatogr. A

1116 (2006) 217.[11] L. Culleré, M. Aznar, J. Cacho, V. Ferreira, J. Chromatogr. A 1017 (2003) 17.[12] P.J. Williams, W. Cynkar, I.L. Francis, J.D. Gray, P.G. Iland, B.G. Coombe, J. Agric.

Food Chem. 43 (1995) 121.[13] J.J. Mateo, N. Gentilini, T. Huerta, M. Jimenez, R. Di Stefano, J. Chromatogr. A.

778 (1997) 219.[14] M. Esti, P. Tamborra, Anal. Chim. Acta 563 (2006) 173.[15] N. Carro, E. Lopez, Z.Y. Günata, R.L. Baumes, C.L. Bayonove, Analysis 24 (1996)

254.[16] Z. Pineiro, M. Palma, C.G. Barroso, Anal. Chim. Acta 513 (2004) 209.[17] S. Insa, V. Salvadó, E. Anticó, J. Chromatogr. A 1047 (2004) 15.[18] R. López, M. Aznar, J. Cacho, V. Ferreira, J. Chromatogr. A 966 (2002) 167.[19] I. Lukic, M. Banovic, d. Persuric, S. Radeka, B. Sladonja, J. Chromatogr. A 1101

(2006) 238.[20] D.S. Pedersen, D.L. Capone, G.K. Skouroumounis, A.P. Pollnitz, M.A. Sefton, Anal.

Bioanal. Chem. 375 (2003) 517.[21] Y. Kotseridis, R. Baumes, G.K. Skouroumounis, J. Chromatogr. A 824

(1998) 71.[22] T. Jesionowski, A. Krysztafkiewicz, J. Non Cryst. Solids 277 (2000) 45.[23] T. Jesionowski, F. Ciesielczyk, A. Krysztafkiewicz, Mater. Chem. Phys. 119 (2010)

65.[24] G.E. Berendsen, L. de Golan, J. Liq., Chromatogr. Relat. Technol. 1 (1978) 561.