Food

Chemistry
Food Chemistry 90 (2005) 417–425
www.elsevier.com/locate/foodchem

Identification and olfactometry of French fries flavour extracted

at mouth conditions
Wil A.M. van Loon a, Jozef P.H. Linssen a,*, Aagje Legger a, Maarten A. Posthumus b,
Alphons G.J. Voragen a
a
Department of Agrotechnology and Food Science, Laboratory of Food Chemistry, Wageningen University, Biotechnion 172,
P.O. Box 8129, 6700 EV Wageningen, The Netherlands
b
Laboratory of Organic Chemistry, Wageningen University, P.O. Box 8026, 6700 EG Wageningen, The Netherlands

Abstract

The aim of this study was to isolate and identify odour active compounds from French fries at mouth conditions. Volatile
compounds were released from French fries by purge-and-trap, trapped on Tenax TA, and identified with GC–MS. GC–olfac-
tometry was used to determine odour active compounds with a trained panel using the detection frequency method. A total of 122
compounds were identified of which 85% originated from either sugar degradation and/or Maillard reaction and 15% from lipid
degradation, based on relative areas. About 50 odour active compounds were, due to coelution, responsible for 41 odours perceived
by the panel. 3-Methylbutanal and 2-methylbutanal, hexanal, 2,3-dimethylpyrazine, 2-methylpropanal, 2,3-butanedione, pyridine,
heptanal, 2,5-dimethylpyrazine and/or 2,6-diemethylpyrazine and/or ethylpyrazine, dimethyl trisulfide, octanal, phenylacetaldehyde,
2,5-diethylpyrazine, (E)-2-nonenal, 3-methylbutanoic acid and/or 2-methylbutanoic acid, (E,Z)-2,4-heptadienal, (E)-2-octenal, 5-
ethyl-2,3-dimethylpyrazine and/or 2-ethyl- 3,5-dimethylpyrazine, nonanal, and tentatively 2-methylpyrrole had the highest detection
frequencies. This resulted in a strong malty and fried potato note, combined with caramel/buttery, green, spicy, and deep-fried notes.
Also chemical and sweaty odours were observed.
Ó 2004 Elsevier Ltd. All rights reserved.

Keywords: French fries; Potato; Flavour; Volatile compounds; Dynamic headspace; Olfactometry

1. Introduction Coleman & Ho, 1980; Coleman, Ho, & Chang, 1981;
Duckham, Dodson, Bakker, & Ames, 2001; Duckham,
French fries are popular around the world because of Dodson, Bakker, & Ames, 2002; Ho & Coleman, 1980;
their pleasant taste, which is a combination of a crispy Oruna-Concha, Duckham, & Ames, 2001), microwave
crust, soft inside and typical fried potato flavour. The baked (Oruna-Concha, Bakker, & Ames, 2002a; Oruna-
flavour of potato has been investigated extensively. Concha et al., 2002b), and French fried potatoes (Car-
According to a review by Maga (1994), more than 500 lin, Jin, Huang, Ho, & Chang, 1986; Wagner & Grosch,
volatiles have been identified so far. 1997; Wagner & Grosch, 1998). Lipid oxidation is the
The composition of volatiles obtained from major major source of volatiles in raw potatoes, because the
cooking procedures differs significantly (Whitfield & lipoxygenase content is relatively high (Josephson &
Last, 1991). Therefore, a distinction should be made Lindsay, 1987). In boiled potatoes the concentration of
between raw (Petersen, Poll, & Larsen, 1998), boiled lipid-derived compounds is lower than in raw potatoes.
(Josephson & Lindsay, 1987; Nursten & Sheen, 1974; This is either due to degradation during boiling or to less
Oruna-Concha, Bakker, & Ames, 2002b; Petersen et al., enzymatic activity since shredding is done after boiling,
1998), oven baked (Buttery, Guadagni, & Ling, 1973; when enzymes are inactivated (Petersen et al., 1998).
The composition of volatiles changes upon heating. The
*
Corresponding author. Tel.: +31-317-483548; fax: +31-317-48-48- Maillard reaction becomes predominant during oven-
93. baking of potatoes, while a microwave treatment re-
E-mail address: jozef.linssen@wur.nl (J.P.H. Linssen). sults in a composition of volatiles between boiled and

0308-8146/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.foodchem.2004.05.005
418 W.A.M. van Loon et al. / Food Chemistry 90 (2005) 417–425

oven-baked potatoes (Oruna-Concha et al., 2002b). The aim of this study was to isolate and identify
Duckham et al. (2001) compared volatiles of oven- odour active compounds of French fries at mouth con-
baked potatoes from eleven cultivars and classified them ditions. Mouth conditions will be created to mimic re-
by their origin: lipid degradation, sugar degradation lease of volatile compounds from the food to the nose
and/or Maillard reaction not involving sulfur amino epithelia, where odour is sensed, by taking into account
acids, sulfur amino acid degradation, methoxypyrazines, the amount of product in relation to the mouth volume,
and terpenes. There are great similarities between the the temperature, and by mixing the product with artifi-
composition of oven-baked potatoes and potato chips, cial saliva.
but also some notable differences (Whitfield & Last,
1991). In fried potato products flavour compounds are
not only formed from the potato, but also from the 2. Materials and methods
frying oil and from the interaction of Maillard reaction
compounds and lipids (Maga, 1994; Whitfield, 1992). 2.1. Materials
Frying temperature is an important parameter as it has a
great influence on the formation of pyrazines in potato French fries (10  10 mm, from variety Agria), ob-
chips (Maga & Sizer, 1978; Martin & Ames, 2001a). tained from Boots Frites BV (Purmerend, The Nether-
Brewer, Vega, and Perkins (1999) emphasize that odour lands), were deep-fried in a Princess Classis household
characteristics of French fries reflect the odour charac- fryer (Princess Household Appliances, Middelharnis,
teristics of the frying oils, so that typical lipid oxidation The Netherlands) in partially hydrogenated vegetable
products such as hexanal increase in the product when frying oil (Remia, Den Dolder, The Netherlands, fatty
the same oil is used over a longer period. acid composition: 9.7% C16:0, 5.9% C18:0, 72.3%
Although potato flavour has received much atten- C18:1, 11.5% C18:2, and 0.8% C18:3) for 4 minutes at
tion, the number of papers about flavour compounds of 180 °C.
French fries is limited. The doctoral thesis by Carlin Artificial saliva was prepared in demineralized water
(1983) describes 429 volatiles extracted from French according to Van Ruth et al. (1995) and consisted of
fries, and two papers have been published from this K2 PO4 (1.37 g l 1 ), KCl (0.45 g l 1 ), CaCl2
H2 0 (0.44
thesis: one focusing on alkyloxazoles (Carlin et al., g l 1 ), NaCl (0.88 g l 1 ), NaHCO3 (5.2 g l 1 ), porcine
1986) and the other on 3-(methylthio)alkanals (Carlin, stomach mucine (2.16 g l 1 , type II M2378, Sigma,
Ho, Chang, Velluz, & Pickenhagen, 1990). There is Zwijndrecht, The Netherlands) and a-amylase from A.
however no information about odour impact of the oryzae (10.5 g l 1 , type X-A, 500,000 units, Sigma,
volatiles given. Wagner and Grosch (1997) screened 48 Zwijndrecht, The Netherlands). NaN3 (0.2 g l 1 ) was
compounds on their contribution to French fries fla- used for preservation.
vour using aroma extract dilution analysis (AEDA) and
gas chromatography–olfactometry of headspace sam- 2.2. Isolation of volatile compounds
ples (GCO-H). In a second paper (Wagner & Grosch,
1998) the authors evaluated the flavour profile of a Volatiles were adsorbed on a glass tube (length 100
model system with 21 potent flavour compounds dis- mm, 3.0 mm internal diameter) filled with 100 mg of
solved in sunflower oil in comparison with the original Tenax TA (20/35 mesh, Alltech Nederland BV,
extract. 2-Ethyl-3,5-dimethylpyrazine, 3-ethyl-2,5-dim- Zwijndrecht, The Netherlands) using purge-and-trap. A
ethylpyrazine, 2,3-diethyl-5-methylpyrazine, 3-isobutyl- sample of French fries was cut in pieces ( < 0.5 cm3 ), 3.0
2-methoxypyrazine, 2,4-decadienal (E,E- and E,Z-), g was weighed and transferred into a glass flask (70 ml),
trans -4,5-epoxy-(E)-2-decenal, 4-hydroxy-2,5-dimethyl- and 4 ml of artificial saliva was added. The flask was
3(2H)-furanone, methylpropanal, 2- and 3-methylbut- placed in a water bath of 37 °C and a flow of purified
anal, and methanethiol were identified as character nitrogen (50 ml min 1 ) was passed through the sample
impact odorants. for 30 min. The sample was mixed constantly by a
Starch and lipids, both important constituents of magnetic stirrer. A cold trap with ethanol of )10 °C was
French fries, are known to interact with flavour com- used to prevent water vapour entering the Tenax tube.
pounds (Arvisenet, Voilley, & Cayot, 2002; Doyen,
Carey, Linforth, Marin, & Taylor, 2001). Moreover, 2.3. Gas chromatographic analysis
during consumption food is mixed and diluted with
saliva by mastication, and starch will be degraded to Volatiles were desorbed onto the column using a
some extent by amylase in saliva. The breakdown of thermal desorption unit (5 min at 245 °C) and cold trap
the food matrix affects the food volatile composition ()120 °C/260 °C) device (Carlo Erba TDAS 5000, Inter-
(Van Ruth, Roozen, & Cozijnsen, 1995), but is often science BV, Breda, The Netherlands). Compounds were
not taken into account when key flavour compounds separated on a 60 m  0.25 mm  0.25 lm Supelco MDN-
are determined. 5S capillary column (Sigma, Zwijndrecht, The Nether-
 W.A.M. van Loon et al. / Food Chemistry 90 (2005) 417–425 419

lands) in a Carlo Erba MEGA 5300 system (Interscience data was written in Pascal. Assessors recorded the be-
BV, Breda, The Netherlands). The oven temperature was ginning and end of an odour by pressing a key on a
kept at 40 °C for 4 min, and then increased to 270 °C at 4.0 laptop computer. The number of assessors detecting an
°C min 1 with a final hold for 5 min. Helium was used as odour simultaneously is used as a measure for the im-
carrier gas at a constant pressure of 150 kPa and a FID portance and from this data an aromagram can be
detector was used at 300 °C. Retention indices of com- constructed. Three training sessions were carried out
pounds were determined with a homologue series of alk- with the same sample as in the final experiment to make
anes (C5-C16) and relative areas of the compounds were the assessors familiar with the procedure. To generate
calculated based on four replicates. flavour descriptors, assessors were asked to give a de-
Identification was performed with a Varian 3400 gas scription at the end of each odour impression. From the
chromatograph (Varian, Bergen op Zoom, The Neth- list of descriptors twelve groups were created based on
erlands) equipped with a thermo desorption device similarity and occurrence in a panel discussion. In the
(Chrompack TCT injector 16200, Chrompack, Middel- olfactometry experiment assessors were forced to choose
burg, The Netherlands), coupled with a Finnigan the group best describing the odour at the end of each
MAT95 mass spectrometer (Thermo Electron, Bremen, odour impression. Blank Tenax TA tubes were used as
Germany). The mass spectrometer was operated in the dummy samples to determine the noise level of the pa-
70 eV EI ionisation mode and scanned from mass 24 to nel. Retention index, mass spectrum, and odour de-
320 with a cycle time of 0.65 s. Identification was made scription (Maarse, 1991; Sigma–Aldrich, 2003) were
by matching sample spectra against the Wiley/NIST 7th used to identify odour active compounds.
edition library and the Wageningen library, and by re-
tention indices when available (Adams, 2001; Kondjo-
yan & Berdague, 1996). 3. Results and discussion

2.4. Gas chromatography–olfactometry analysis The groups of odour descriptors resulting from the
attribute generation and panel discussion are shown in
Gas chromatography–olfactometry analysis (GC–O) Table 1. Because compounds often eluted closely after
analysis was performed with a comparable system, col- each other and assessors had to react quickly, we de-
umn, and conditions as described in the previous para- cided to create not more than twelve groups. For de-
graph. The effluent was however split 10:45:45 at the end scriptors often mentioned (e.g. French fries, chocolate) a
of the column for FID detector and two sniffing ports, separate group was created, and from some similar de-
respectively. Humidified air ran through stainless steel scriptors a logical group name proceeded (e.g. nutty,
tubing to the sniffing port in order to prevent drying out green). For the remaining descriptors a general group
of assessors’ nasal membranes during the 40 min sniffing name (e.g. spicy, baked flavour) was agreed on.
experiments. A total of 122 compounds identified with GC–MS are
listed in Table 2. Based on the relative areas approxi-
2.5. Olfactometry data acquisition mately 85% of the aroma compounds originated from
sugar degradation and/or Maillard reaction not involv-
Twelve assessors, nine females and three males aged ing sulfur amino acids and 15% were lipid-derived. Less
19–22, were selected for the olfactometry experiment than 0.1% consisted of a number of sulfur compounds
from thirty volunteers. The selection procedure included and one terpene (limonene). Duckham et al. (2001)
a questionnaire and a test for odour sensitivity. The compared volatiles from oven-baked potatoes of eleven
questionnaire was used to test their creativity by de- cultivars, and found that the relative amount of lipid
scribing odours and odour differences between several derived volatiles ranged from 22% to 69% among cul-
products including French fries. Also questions about tivars. For volatiles originating from sugar degradation
their availability and physical condition were included. and/or Maillard reaction not involving sulfur amino
The odour sensitivity test was carried out by putting a acids the range was from 25% to 77%. As the lipid
concentration series of 2-methylbutanal in paraffin content of raw potato is only 2–3 g kg 1 (Duckham
(0.08, 0.13, 0.40, 2.0, and 10 ll l 1 ) in ascending order. et al., 2001) and French fries are prepared in a large
The concentration range was determined in preliminary amount of oil, one might expect the amount of lipid-
experiments. Sixteen volunteers put the concentration derived volatiles to be higher in French fries than in
either in the correct order or mixed up only two con- oven-baked potatoes. It seems however that the high
secutive concentrations. Twelve of them were selected heat transfer from the oil to the product is more fa-
based on the results of the questionnaire. vourable for sugar degradation and/or the Maillard re-
The detection frequency method (Van Ruth et al., action than for lipid degradation. Furthermore,
1995) was used to determine the importance of odour melanoidines formed in the Maillard reaction are
active compounds and a program to acquire sniffing known to have an antioxidative effect (Morales &
420 W.A.M. van Loon et al. / Food Chemistry 90 (2005) 417–425

Table 1
Descriptor groups used for olfactory analysis
Group Descriptor Descriptors from attribute generation

1 French fries French fries, potato chips, baked potato, frying fat, frying odour
2 Baked flavour Popcorn, cookie, baked meat, baked fish, bread
3 Potato, earthy Raw potato, cooked potato, earthy
4 Nutty Nutty, peanut, almond, nougat, marzipan
5 Chocolate Chocolate, cacao
6 Sweet, flowery, fresh Sweet, flowery, mint, anise, apple, lemon, soap, pine, melon
7 Butter, vanilla, caramel Butter, vanilla, caramel
8 Grass, green Vegetable, grass, plant, cucumber, carrots
9 Cheese, sweat Cheese, sweat
10 Spicy Spicy, mushroom, wood, garlic, onion, leather, fungus
11 Chemical, paint, glue Chemical, paint, glue, plastic
12 Metal, burnt Metal, burnt

Table 2 Table 2 (continued)

Relative peak area and retention indices (RI) of volatile compounds
Compounds by main RIexp RIlit Relative peak
identified in French fries, fried for 4 min at 180 °C
origin area (%)
Compounds by main RIexp RIlit Relative peak 2-Hexenal, (E)-a; b;c;d 853 855h <0.1
origin area (%) 2-Heptanone a;b ;c;e
889 892h <0.1
Lipid degradation Styreneb;c;d 893 0.1
Ethanola;b ;c;e <500 503g 3.1 Heptanala; b;c;d; e 902 902h 0.4
2-Propanol 514 524g 4.1 Acetylfuranc; e 910 912h <0.1
Methyl acetate 525 515g <0.1 Propylcyclohexaneb;d 935 939g <0.1
2,3-Butanedionea; b;c;e 585 593g 0.4 Butylcyclopentane 937 941g <0.1
2-Butanonea ;c;d; e 589 597g 0.3 2-Heptenal, (E)-a; b;c;d 959 956g 0.3
2-Methylfuranb;c; d 595 606g <0.1 5-Methylfurfuralc 962 964h <0.1
2-Butanolc; e 601 591g <0.1 1-Heptanola; b;c;e 971 967h <0.1
Methyl 2-propenoate 607 <0.1 Phenole 980 971g <0.1
Tetrahydrofuran 620 628g 0.5 1-Octen-3-ola; b;c;d ;e 981 979h <0.1
2-Methyl-1-propanole 624 619g <0.1 Hexanoic acidb; c; e 987 970g <0.1
Acetic acida;b; c;e 649 660g <0.1 2-Pentylfurana; b;c; d;e 991 993g <0.1
1-Penten-3-ole 682 683g <0.1 2,4-Heptadienal, 998 <0.1
2-Pentanone 685 687g <0.1 (E,Z)-b; c
1-Heptene 689 692g <0.1 Octanala; b;c; d 1005 1004g  0.2
2-Ethylfuranb; c; d 701 704g 0.5 2,4-Heptadienal, 1014 1009g <0.1
2-Vinylfuran 721 <0.1 (E,E)-b;c; d;e
2,5-Dihydro-3,4-dim- 732 <0.1 5-Ethyl-1-cyclopentene- 1035 <0.1
ethylfuranf 1-carboxaldehyde
2-Methyl-2-butenalc;d; e 740 742h <0.1 Benzyl alcohola ;b; c; e 1038 <0.1
2-Pentenal, (E)-a; b;c;d 753 754g <0.1 2-Hydroxybenzaldehyde 1048 1045h <0.1
Methyl 2-butenoate, 761 <0.1 2-Octenal, (E)-a ;b; c; d;e 1060 1060g 0.2
(E)- Acetophenoned;e 1071 1065h 0.1
1-Pentanola; b;e 765 771h <0.1 Nonanalb; c; d;e 1107 1101h 1.0
Tolueneb ;c;d; e 766 770g 0.2 2-Nonenal, (E)-a ;b; c; d;e 1162 1162h <0.1
Methyl 3-methyl- 775 775g <0.1 Benzoic acide 1172 <0.1
butanoate Decanalb; c; d;e 1207 1202h <0.1
2-Ethyl-5-methylfuran 775 <0.1 2,4-Nonadienal, 1229 1212h <0.1
3-Hexanonea ;b; c; d;e 783 787g <0.1 (E,E)-a; b;c;e
2-Methyl-3-hexanone 784 734g <0.1 2-Decenal, (E)-c;e 1265 1264h <0.1
2-Hexanone 787 789g <0.1 2,4-Decadienal, 1298 1293h <0.1
tentativea; c; e (E,Z)-a; b;c;e
2-Propylfuranc;e 788 787g <0.1 2,4-Decadienal, 1323 1317h <0.1
1-Octene 790 792g <0.1 (E,E)-a; b;c;e
Cyclopentanonec;e 791 <0.1 2-Undecenal, (E)-c 1368 1366g <0.1
Hexanala;b; c;d; e 801 802h 1.1 1-Undecanol 1374 1370h <0.1
Butanoic acidc; e 805 779g <0.1 Total 14.6
2-Octene, (E-) 805 818g <0.1 Sugar degradation and/or Maillard reaction not involving sulfur amino
2-Octene, (Z-)b 813 808g <0.1 acids
2-Ethyl-2-butenal, (E)-f 815 <0.1 2-methylpropanala; b;c;e 554 552g 20.3
2,5-Furandionef 830 <0.1 3-Methylbutanala;b; c;d; e 658 654g 29.3
Propylcyclopentane 833 <0.1 2-Methylbutanala;b; c;d; e 669 662g 31.4
 W.A.M. van Loon et al. / Food Chemistry 90 (2005) 417–425 421

Table 2 (continued) Table 2 (continued)

Compounds by main RIexp RIlit Relative peak Compounds by main RIexp RIlit Relative peak
origin area (%) origin area (%)
2,3-Pentanedioneb; c; e 698 700h 0.7 Isopentylmethylpyrazine 1254 <0.1
Pyrazine 732 <0.1 isomere
3-Methyl-1-butanola; b;c 734 737g <0.1 Isopentyldimethyl- 1316 <0.1
1-Methylpyrrolec;d 736 749g <0.1 pyrazine isomere
Pyridinea; b;c 746 751g <0.1 Total 85.4
Pyrrole 748 762g 0.2 Sulfur compounds
4,5-Dimethyloxazolef 755 <0.1 Dimethyl disulfideb; c; d;e 743 744g <0.1
2,4,6-Trimethyl-1,3,5- 781 776g <0.1 2-Methylthiophene 770 779g <0.1
trioxanec 3-Methylthiophene 779 786g <0.1
2-Methylpropanoic 797 0.8 Dimethyl trisulfideb;c;d; e 975 950g <0.1
acidc;e Dimethyl 1241 1215g <0.1
Dihydro-2-methyl- 807 <0.1 tetrasulfideb;c;e
3(2H)-furanone.c; e Total <0.1
1-Ethylpyrrole 811 820g <0.1 Terpenes
Methylpyrazinec; e 825 826h 0.7 Limonenea; b;c;d; e 1034 1031g <0.1
Furfurala; b;c; d;e 832 836h <0.1 Grand total 100
2-Methylpyrroled 843 <0.1 a
Found in raw potatoes (Adams, 2001; Berdague, Denoyer, Le
3-Methylbutanoic 878 843g <0.1 Quere, & Semon, 1991).
acidc;e b
Found in boiled potatoes (Adams, 2001; Berdague et al., 1991;
2-Methylbutanoic acide 888 838g 0.2 Brewer et al., 1999; Buttery et al., 1973; Carlin, 1983).
2,5- and/or 2,6-dimethyl- 915 913h 0.4 c
Found in oven-baked potatoes (Adams, 2001; Carlin et al., 1986,
pyrazinec; e 1990; Coleman & Ho, 1980; Coleman et al., 1981; Doyen et al., 2001;
Ethylpyrazinec 918 916h 0.1 Duckham et al., 2001, 2002).
2,3-Dimethylpyrazinec; e 921 920h 0.3 d
Found in microwave baked potatoes (Oruna-Concha et al., 2002a,
Vinylpyrazinee 934 <0.1 2002b).
Benzaldehydea ;b; c; d;e 966 960h 0.1 e
Found in French fries (Maga, 1994; Carlin et al., 1986; Wagner &
2-Ethyl-6-methyl- 1000 0.2 Grosch, 1997).
pyrazinec f
Tentative identification.
2-Ethyl-3-methyl- 1004 1003h  0.2 g
According to Adams (2001).
pyrazinec h
According to Kondjoyan and Berdague (1996).
2-Ethyl-5-methyl- 1006  0.2
pyrazinec; e
2-Methyl-6-vinyl- 1020 <0.1 Jimenez-Perez, 2001). This may have an effect on the
pyrazinee
Isopropenylpyrazine 1021 <0.1
amount of lipid-derived compounds, because the ma-
2-Methyl-5-vinyl- 1025 <0.1 jority of these compounds is formed through oxidation.
pyrazinec; e In this study fresh frying oil was used for all experi-
1-Methyl-2-pyrrolidi- 1043 1042g <0.1 ments. The composition of volatiles may however
none change as the frying oil gets older, because the odour
Phenylacetalde- 1048 1049g 0.1
hydea ;b; c; d;e
characteristics of French fries reflect the odour charac-
3-Ethyl-2,5-dimethyl- 1079 1060g <0.1 teristics of the oils in which they are fried (Brewer et al.,
pyrazinec; e 1999).
2,6-Diethylpyrazine 1081 <0.1 Many of the identified compounds have been de-
2,3-Diethylpyrazinec 1082 1085h <0.1 scribed previously as volatiles from French fries or
2-Ethyl-3,5-dimethyl- 1085 <0.1
pyrazine
(processed) potato (see Table 2). The three main con-
5-Ethyl-2,3-dimethyl- 1088 <0.1 tributors of sugar degradation and/or Maillard reaction
pyrazine were 3-methylbutanal, 2-methylbutanal, and 2-methyl-
2,5-Diethylpyrazine 1094 <0.1 propanal, Strecker aldehydes of respectively isoleucine,
Dimethylvinylpyrazine 1100 <0.1 leucine, and valine (Whitfield, 1992), and they accounted
isomerc;e;f
2,3-Dihydroindole 1120 <0.1
for 81% of the total yield of volatiles. Other contributors
Isobutylmethylpyrazine 1139 1137h <0.1 were phenylacetaldehyde, the Strecker aldehyde of
isomerc;e phenylalanine (Whitfield, 1992), 2,3-pentanedione, and
2,3-Diethyl-5-methyl- 1153 <0.1 many heterocyclic nitrogen containing compounds of
pyrazinec; e which pyrazines were most abundant. A total of 26
3,5-Diethyl-2-methyl- 1157 <0.1
pyrazinec; e
pyrazines were found of which eleven have not been
Methylpropenylpyrazine 1179 <0.1 reported as volatiles from French fries, and six not from
isomere potato previously. Pyrazines are formed mainly from
Dimethylisobutyl- 1201 <0.1 glutamine and asparagine, because these amino acids are
pyrazine isomerc present in potato in large quantities (Martin & Ames,
422 W.A.M. van Loon et al. / Food Chemistry 90 (2005) 417–425

2001b; Oruna-Concha et al., 2001), and pyrazines are

considered to be key components of baked potato fla-
vour (Duckham et al., 2002). Benzaldehyde was not put
in the list of lipid-derived products as was done previ-
ously (Duckham et al., 2001; Oruna-Concha et al.,
2002a), because it is formed from phenylacetaldehyde
(Martin & Ames, 2001b). The formation of 2- and 3-
methylbutanoic acid and 2-methylpropanoic acid can be
explained by further oxidation of their corresponding
aldehydes, the three main contributors. Lipid degrada-
tion resulted in the formation of several aldehydes, ke-
tones, alcohols, hydrocarbons, acids, esters, and furans,
of which many are secondary oxidation products. The
highest relative areas were obtained by ethanol, 2-pro-
panol, nonanal, and hexanal. Nonanal is formed from
oleic acid, which is the main fatty acid of the frying oil
used, and hexanal is a typical oxidation product of li-
noleic acid (Brewer et al., 1999; Whitfield, 1992). Etha-
nol and 2-propanol are probably formed after several
degradation steps, induced by the high frying tempera-
ture. The three sulfides and two thiophenes that were
found in this study were formed from sulfur containing
amino acids. As these sulphides are formed in the
Maillard reaction they have been found in boiled,
baked, and fried potato, but not in raw potato. 2- and
3-Methylthiophene were suggested to be formed
from the interaction between fatty aldehydes and
cysteine (Macku & Shibamoto, 1991). 2-Pentanone and
1-methyl-2-pyrrolidinone have not been found in po-
tato, but they have been reported as volatiles from cured
ham (Berdague et al., 1991). Methional and metha-
Fig. 1. Aromagram of dynamic headspace samples. FID response (a)
nethiol, both appointed as potent odorants of French and detection frequency of assessors at sniffing port (b). Horizontal
fries flavour (Wagner & Grosch, 1997), were not line in (b) represents noise level.
detected.
In Fig. 1 the chromatogram from the FID-detector
and the aromagram resulting from the olfactometry 3- and 2-methylbutanal, resulting in a large peak with
experiments is shown, and identification of the 41 peaks a small shoulder. The odour is usually referred to as
is explained in Table 3. Not all aromagram peaks could chocolate or malty, but some assessors described it as
be identified conclusively, because sometimes com- sweaty, which corresponds to what Berdague et al.
pounds having similar odours eluted closely after each (1991) found. Pyrazines were responsible for eleven
other. Compounds with the highest detection frequency peaks in the aromagram and resulted in the dominant
contribute most to French fries flavour. In this way 2- fried potato note. C2- and C4-substituted pyrazines
methylbutanal and/or 3-methylbutanal, hexanal, 2,3- eluted closely after each other making it difficult to
dimethylpyrazine, 2-methylpropanal, 2,3-butanedione, identify the one responsible for peak 13 and peaks 29–
pyridine, heptanal, 2,5-dimethylpyrazine and/or 2,6- 31, respectively. 2,3-Butanedione and to a lesser extent
dimethylpyrazine and/or ethylpyrazine, dimethyl trisul- 2,3-pentanedione contributed to the caramel or buttery
fide, octanal, phenylacetaldehyde, 2,5-diethylpyrazine, note. Fatty aldehydes, of which hexanal, heptanal,
(E)-2-nonenal, 2-methylbutanoic acid and/or 3-meth- octanal, (E)-2-nonenal, and nonanal were the most
ylbutanoic acid, (E,Z)-2,4-heptadienal, (E)-2-octenal, important representatives, were perceived as green
5-ethyl-2,3-dimethylpyrazine and/or 2-ethyl-3,5-dimeth- odours. These compounds are however known to
ylpyrazine, nonanal, and tentatively 2-methylpyrrole cause a rancid off-flavour (Brewer et al., 1999; Whit-
were found to be the most important odour active field, 1992). We confirm that alkadienals, such as
compounds of French fries. (E,E)-2,4-decadienal, (E,E)-2,4-nonadienal, and (E,Z)-
2-Methylpropanal, 3-methylbutanal and 2-methyl- 2,4-heptadienal, contribute to the deep-fried note
butanal were perceived by almost all assessors. It was (Wagner & Grosch, 1997). Dimethyl trisulfide, phe-
difficult for the assessors to notice a difference between nylacetaldehyde, and possibly 2-methylpyrrole cause a
 W.A.M. van Loon et al. / Food Chemistry 90 (2005) 417–425 423

Table 3
Detection frequency and odour description of odour active compounds from French fries
Peak no. Retention time on GC–O RIa Detection Odour groupb Compound
frequency
1 6.19 <500 4 5, 6, 11 Ethanol
2 7.25 550 10 7, 5, 9, 11 2-Methylpropanal
3 7.83 585 10 7 2,3-Butanedione
4 9.44 660 11 11, 9, 5, 7 3-Methylbutanal and 2-methylbutanal
5 10.76 700 5 7 2,3-Pentaandion
6 11.92 730 5 6 2,5-Dihydro-3,4-dimethylfuran tentative
7 12.51 745 10 11 Pyridine
8 12.84 755 4 11, 6, 1 Pyrrole
9 13.26 760 7 6 Methyl 2-butenoate, (E)-
10 14.07 775 6 11, 9 2-Methylpropanoic acid
11 15.21 800 11 6, 8 Hexanal
12 15.57 810 7 1, 10 1-Ethylpyrrole tentative
13 17.42 845 9 9, 3 2-Methylpyrrole tentative
14 17.75 855 9 11, 9 3-Methylbutanoic acid and/or 2-methylbu-
tanoic acid
15 18.97 880 6 3 Unknown
16 20.07 900 10 8, 3, 6 Heptanal
17 21.00 920 10 3, 1, 12 2,5-Dimethylpyrazine, 2,6-dimethylpyrazine
and/or ethylpyrazine
18 21.30 925 11 2, 4 2,3-Dimethylpyrazine
19 21.70 935 5 1, 2 Vinylpyrazine
20 22.97 960 8 3 2-Heptenal, (E)-
21 23.87 980 10 9, 10, 2, 12 Dimethyl trisulfide
22 24.67 995 5 6 2-Pentylfuran
23 24.90 1000 9 6, 1 2,4-Heptadienal, (E,Z)-
24 25.23 1005 10 6, 1 Octanal
25 26.29 1025 7 2, 12, 4 2-Methyl-5-vinylpyrazine
26 27.01 1045 5 3, 11 1-Methyl-2-pyrrolidinone tentative
27 27.36 1050 10 3, 12, 6, 8 Phenylacetaldehyde
28 27.87 1060 9 8, 11, 10, 7 2-Octenal, (E)-
29 28.67 1080 8 3, 11, 2, 10 3-Ethyl-2,5-dimethylpyrazine, 2,3-diethyl-
pyrazine and/or 2,6-diethylpyrazine
30 29.09 1090 9 6, 2, 10 5-Ethyl-2,3-dimethylpyrazine and/or 2-eth-
yl-3,5-dimethylpyrazine
31 29.27 1095 10 6, 3, 12, 11 2,5-Diethylpyrazine
32 29.76 1105 9 1, 8 Nonanal
33 32.11 1155 7 3, 6, 9 3,5-Diethyl-2-methylpyrazine and/or 2,3-di-
ethyl-5-methylpyrazine
34 32.48 1170 10 8 2-Nonenal, (E)-
35 33.22 1180 7 2, 4 Pyrazine, methylpropenyl isomer
36 33.89 1200 6 8, 6, 3, 12 Pyrazine, dimethylisobutyl isomer
37 34.67 1215 7 8 Decanal
38 35.16 1230 7 2, 1, 4, 11 2,4-Nonadienal, (E,E)-
39 36.26 1255 4 8, 2, 3 Pyrazine, isopentylmethyl isomer
40 37.02 1275 5 6, 7, 2, 12 2-Decenal, (E)-
41 39.68 1330 5 1 2,4-Decadienal, (E,E)-
a
Retention index estimated by 30–40 s delay between FID-signal and flavour perception.
b
See Table 1.

spicy note. Based on retention index and mass spec- tention index is higher than the literature value in
trum peak 13 could be 2-methylpyrrole, however no Table 2. As these compounds were dominating the
information about the odour description could be mass spectra around peak 15, it was not possible to
found in literature. Although the chemical odour from identify the compound responsible for the potato or
pyridine, and sweaty odour from 3-methylbutanoic earthy odour.
acid and 2-methylbutanoic acid were clearly noticed by Although the potato variety used was the same, the
the panel, they do not seem to give a distinct note, but results were different from what Wagner and Grosch
may influence the perceived aroma as a whole. 2- (1997) found. They did not find chemical or sweaty
Methylbutanoic acid and 3-methylbutanoic acid had a odours such as pyridine and 3-methylbutanoic acid, and
very broad peak and that is the reason why the re- that we did not find methional and methanethiol. This
424 W.A.M. van Loon et al. / Food Chemistry 90 (2005) 417–425

can be explained by using different methods for sample Doyen, K., Carey, M., Linforth, R. S. T., Marin, M., & Taylor, A. J.
preparation and extraction. Other factors such as stor- (2001). Volatile release from an emulsion: Headspace and in-
mouth studies. Journal of Agricultural and Food Chemistry, 49,
age conditions of the potatoes (Duckham et al., 2002), 804–810.
the type of frying oil (Brewer et al., 1999), and the frying Duckham, S. C., Dodson, A. T., Bakker, J., & Ames, J. M. (2001).
temperature (Maga & Sizer, 1978; Martin & Ames, Volatile flavour components of baked potato flesh. A comparison
2001b) may play a role as well. of eleven potato cultivars. Nahrung, 45(5), 317–323.
Duckham, S. C., Dodson, A. T., Bakker, J., & Ames, J. M. (2002).
Effect of cultivar and storage time on the volatile flavor compo-
nents of baked potato. Journal of Agricultural and Food Chemistry,
4. Conclusion
50, 5640–5648.
Ho, C.-T., & Coleman, E. C. (1980). Chemistry of baked potato flour:
Isolation of French fries flavour by purge-and-trap Further identification of heterocyclic compounds in the volatile
using conditions during consumption yielded 122 iden- flavor of baked potato. Critical Reviews in Food Science and
tified volatiles of which 85% of the relative area origi- Nutrition, 45, 1094–1095.
Josephson, D. B., & Lindsay, R. C. (1987). c4-Heptenal: An influential
nated from sugar degradation and/or Maillard reaction.
volatile compound in boiled potato flavor. Journal of Food Science,
2-Methylpropanal, 2-methylbutanal, 3-methylbutanal, 52(2), 328–331.
and 26 pyrazines were the main representatives. 15% of Kondjoyan, N., & Berdague, J.-L. (1996). A compilation of relative
the volatiles were lipid-derived and ethanol, 2-propanol, retention indices for the analysis of aromatic compounds. Dijon:
hexanal, and nonanal showed the highest relative areas Laboratoire Flaveur INRA.
Maarse, H. (1991). Volatile compounds in food and beverages. New
of this group. About 50 odour active compounds were,
York: Marcel Dekker Inc..
due to coelution, responsible for 41 odours perceived by Macku, C., & Shibamoto, T. (1991). Volatile sulfur-containing com-
the panel. The compounds with the highest detection pounds generated from the thermal interaction of corn oil and
frequencies caused a strong malty note and fried potato cysteine. Journal of Agricultural and Food Chemistry, 39, 1987–1989.
note, combined with caramel/buttery, green, spicy, and Maga, J. A. (1994). Potato flavor. Food Reviews International, 10(1), 1–
48.
deep-fried notes. Also chemical and sweaty odours were
Maga, J. A., & Sizer, C. E. (1978). Flavour preferences for potato chips
observed. as influenced by time and temperature of chipping and total
pyrazine concentration. Lebensmittel Wissenschaft und Technologie,
11, 181–182.
References Martin, F. L., & Ames, J. M. (2001a). Comparison of flavor
compounds of potato chips fried in palmolein and silicone fluid.
Adams, R. P. (2001). Identification of essential oil components by gas Journal of the American Oil Chemists Society, 78(8), 863–866.
chromatography/quadrupole mass spectroscopy. Carol Stream, IL: Martin, F. L., & Ames, J. M. (2001b). Formation of strecker aldehydes
Allured. and pyrazines in a fried potato model system. Journal of Agricul-
Arvisenet, G., Voilley, A., & Cayot, N. (2002). Retention of aroma tural and Food Chemistry, 49, 3885–3892.
compounds in starch matrices: Competitions between aroma Morales, F. J., & Jimenez-Perez, S. (2001). Free radical scavenging
compounds toward amylose and amylopectin. Journal of Agricul- capacity of Maillard reaction products as related to colour and
tural and Food Chemistry, 50, 7345–7349. fluorescence. Food Chemistry, 72, 119–125.
Berdague, J.-L., Denoyer, C., Le Quere, J.-L., & Semon, E. (1991). Nursten, H. E., & Sheen, M. R. (1974). Volatile flavour components of
Volatile components of dry-cured ham. Journal of Agricultural and cooked potato. Journal of the Science of Food and Agriculture, 25,
Food Chemistry, 39, 1257–1261. 643–663.
Brewer, M. S., Vega, J. D., & Perkins, E. G. (1999). Volatile Oruna-Concha, M. J., Bakker, J., & Ames, J. M. (2002a). Comparison
compounds and sensory characteristics of frying fats. Journal of of the volatile components of eight cultivars of potato after
Food Lipids, 6, 47–61. microwave baking. Lebensmittel Wissenschaft und Technologie, 35,
Buttery, R., Guadagni, D. G., & Ling, L. C. (1973). Volatile 80–86.
components of baked potatoes. Journal of the Science of Food Oruna-Concha, M. J., Bakker, J., & Ames, J. M. (2002b). Comparison
and Agriculture, 24, 1125–1131. of the volatile components of two cultivars of potato cooked by
Carlin, J. T. (1983). French fry flavor project. New Brunswick, NJ: boiling, conventional baking and microwave baking. Journal of the
Rutgers University. Science of Food and Agriculture, 82, 1080–1087.
Carlin, J. T., Ho, C.-T., Chang, S. S., Velluz, A., & Pickenhagen, W. Oruna-Concha, M. J., Duckham, S. C., & Ames, J. M. (2001).
(1990). Analysis of French fried potato flavor: Identification of 3- Comparison of volatile compounds isolated from the skin and flesh
(methylthio)alkanals. Lebensmittel Wissenschaft und Technologie, of four potato cultivars after baking. Journal of Agricultural and
23, 276. Food Chemistry, 49, 2414–2421.
Carlin, J. T., Jin, Q. Z., Huang, T.-C., Ho, C.-T., & Chang, S. S. Petersen, M. A., Poll, L., & Larsen, L. M. (1998). Comparison of
(1986). Identification of alkyloxazoles in the volatile compounds volatiles in raw and boiled potatoes using a mild extraction
from French-fried potatoes. Journal of Agricultural and Food technique combined with GC odour profiling and GC–MS. Food
Chemistry, 34, 621–623. Chemistry, 61(4), 461–466.
Coleman, E. C., & Ho, C.-T. (1980). Chemistry of baked potato Sigma–Aldrich, 2003. Flavors and fragrances. International Edition,
flavor. 1. Pyrazines and thiazoles identified in the volatile flavor of Milwaukee, USA.
baked potato. Journal of Agricultural and Food Chemistry, 28, 66– Van Ruth, S. M., Roozen, J. P., & Cozijnsen, J. L. (1995). Changes in
68. flavour release from rehydrated diced bell peppers (Capsicum
Coleman, E. C., Ho, C.-T., & Chang, S. S. (1981). Isolation and annuum) by artificial saliva components in three mouth model
identification of volatile compounds from baked potatoes. Journal systems. Journal of the Science of Food and Agriculture, 67, 189–
of Agricultural and Food Chemistry, 29, 42–48. 196.
 W.A.M. van Loon et al. / Food Chemistry 90 (2005) 417–425 425

Wagner, R. K., & Grosch, W. (1997). Evaluation of potent odorants of Whitfield, F. B. (1992). Volatiles from interactions of Maillard
French fries. Lebensmittel Wissenschaft und Technologie, 30(2), reactions and lipids. Critical Reviews in Food Science and Nutrition,
164–169. 31(1/2), 1–58.
Wagner, R. K., & Grosch, W. (1998). Key odorants of french fries. Whitfield, F. B., & Last, J. H. (1991). Tuber vegetables. In H. Maarse
Journal of the American Oil Chemists Society, 75(10), 1385– (Ed.), Volatile compounds in foods and beverages (pp. 222–231).
1392. New York: Marcel Dekker.