Food Additives & Contaminants: Part A

ISSN: 1944-0049 (Print) 1944-0057 (Online) Journal homepage: http://www.tandfonline.com/loi/tfac20

Rapid and Sensitive Method for the Determination

of Four EU Marker Polycyclic Aromatic
Hydrocarbons in Cereal based Foods Using Isotope
Dilution GC/MS

Sibel Kacmaz, Zuzana Zelinkova & Thomas Wenzl

To cite this article: Sibel Kacmaz, Zuzana Zelinkova & Thomas Wenzl (2016): Rapid and
Sensitive Method for the Determination of Four EU Marker Polycyclic Aromatic Hydrocarbons
in Cereal based Foods Using Isotope Dilution GC/MS, Food Additives & Contaminants: Part A,
DOI: 10.1080/19440049.2016.1162032

To link to this article: http://dx.doi.org/10.1080/19440049.2016.1162032

Accepted author version posted online: 07

Mar 2016.

Submit your article to this journal

Article views: 6

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=tfac20

Download by: [University of Tennessee, Knoxville] Date: 15 March 2016, At: 12:50
 Publisher: Taylor & Francis

Journal: Food Additives & Contaminants: Part A

DOI: 10.1080/19440049.2016.1162032
Rapid and Sensitive Method for the Determination of Four EU Marker Polycyclic
Aromatic Hydrocarbons in Cereal based Foods Using Isotope Dilution GC/MS

Sibel Kacmaza *, Zuzana Zelinkovab, and Thomas Wenzlb

a
Giresun University, Faculty of Engineering, Department of Food Engineering, 28200, Güre
Downloaded by [University of Tennessee, Knoxville] at 12:50 15 March 2016

Campus, Giresun, Turkey; bEuropean Commission, Joint Research Centre, Institute for
Reference Materials and Measurements, Retieseweg 111, 2440 Geel, Belgium

* Corresponding author; Sibel Kacmaz

Address: Giresun University, Faculty of Engineering, Department of Food Engineering,

28200, Güre Campus, Giresun, Turkey

Tel.: +90 454 310 41 73 Fax: +90 454 310 1749

E-mail addresses: sibel.kacmaz@giresun.edu.tr

 Rapid and Sensitive Method for the Determination of Four EU Marker Polycyclic
Aromatic Hydrocarbons in Cereal based Foods Using Isotope Dilution GC/MS

Abstract

A rapid and sensitive method has been developed for the determination of the four EU
marker polycyclic aromatic hydrocarbons (PAHs; benz[a]anthracene, chrysene,
benzo[b]fluoranthene and benzo[a]pyrene) in some cereal based foods. The method is based
on pressurized liquid extraction (PLE), solid phase extraction clean-up (SPE) and isotope
dilution gas chromatography with mass-spectrometric detection (GC/MS). The developed
method was calibrated for the content range of 0.05 µg kg-1 to 12.5 µg kg-1 (expressed on
Downloaded by [University of Tennessee, Knoxville] at 12:50 15 March 2016

13
product basis). Recoveries of PAH were monitored in each sample via the recovery of C-
labelled PAHs. Recovery values were in the range between 86 % and 91 %, with relative
standard deviations (RSDs) between 5 % and 9 %. The achieved limits of detection for all
analytes were below 0.05 µg kg-1. The applicability of the method for the analysis of routine
samples was studied by the analysis of a set of commercial bread and breakfast cereal
samples. In all analysed samples, benzo[a]pyrene (BAP) was the most prevalent PAH with
the content between 0.09 µg kg-1 and 0.30 µg kg-1. On average, samples showed low levels of
the sum of the four EU marker PAHs (ΣPAH4) that ranged between 0.11 µg kg-1 and 0.22 µg
kg-1 for bread samples and 0.23 µg kg-1 and 0.87 µg kg-1 for breakfast cereal samples. The
developed method was found suitable for the determination of PAHs in cereal based foods
like cornflakes and breads with total relative fat contents below 3.5%.

Keywords: Four EU marker PAHs, bread, cornflakes, GC/MS analysis, method validation

Introduction

Polycyclic aromatic hydrocarbons (PAHs) comprise a group of about ten thousand organic
compounds; some of them, particularly those with low molecular weight, occur in
considerable amounts in food. PAHs consist of fused aromatic rings and do not contain
heteroatoms or carry substituents. Some PAHs are classified carcinogenic and mutagenic.
Exposure of humans to PAHs occurs mostly by intake of food. Food contamination with
 PAHs might originate from processes during manufacturing of food or environmental
contamination, since PAHs are formed via the pyrolysis of organic matter and through
incomplete combustion during food processing, such as charcoal grilling, roasting, and
smoking. (CCFAC 2005; SCF 2002a; SCF 2002b; JECFA 2005).

Based on the outcome of dedicated risk assessments conducted by the Scientific Committee
on Food (SCF) and the International Agency for Research on Cancer (IARC), the European
Commission recommended further investigation into the levels of benzo[a]pyrene (BaP) and
other genotoxic compounds in foodstuffs (SCF 2002a; IARC 2010; IARC 2012; CEC 2005).
The European Commission issued Commission Regulation (EC) No 1881/2006 setting
Downloaded by [University of Tennessee, Knoxville] at 12:50 15 March 2016

maximum levels of BaP in food and Commission Regulation (EC) No 333/2007 laying down
sampling methods and performance criteria for methods of analysis for the official control of
BaP levels in foodstuffs (CEC 2006; CEC 2007). A more recent risk assessment conducted
by the European Food Safety Authority (EFSA) identified a set of four PAHs
(benzo[a]pyrene, benz[a]anthracene, benzo[b]fluoranthene and chrysene) as suitable markers
for the total PAH content in food (EFSA 2008). Usually they are referred to as PAH4 or as
"the four EU marker PAHs". The current maximum limits for BaP and for the sum of the four
EU marker PAHs (ΣPAH4) are listed in Commission Regulation (EU) No 835/2011 (EC
2011a) for foodstuffs such as smoked meat and smoked meat products, smoked fish, oils,
fats, infant formula and baby food. However, this legislation emphasized that for setting
maximum levels current occurrence data on the levels of PAHs in cereal products are limited,
and requested further monitoring of PAHs in cereal products. Therefore, the need for reliable
data on the content of PAHs in cereal foods is increasing in order to establish maximum
permitted levels.

Analytical data on cereal products available to EFSA were highly left-censored, i.e. with a
vast majority of values reported as below the limit of detection (LOD) (EFSA 2008). Few
samples exceeded the respective LODs indicated rather low levels of PAHs in cereal
products. At the same time, EFSA identified cereal products as important contributor to
human exposure, due to their high consumption volumes. Therefore, reliable analysis
methods are required to determine low level contamination of cereal based foods with PAHs.
The relevance of the contribution of cereal based food to the total dietary exposure to PAHs
was recently confirmed for the French and Swedish populations (Veyrand et al. 2013;
 Abramsson-Zetterberg et al. 2014). Like EFSA, they had also to deal with many left-censored
data.
Literature contains limited information regarding the determination of PAHs in breads and
other cereal products. Ciecieska and Obiedzinski (2013) studied the occurrence of PAHs in
bread ingredients and the potential formation of PAHs during baking of two types of bread.
They applied ultrasonic extraction, followed by gel-permeation chromatography prior to
measurement of 19 PAHs by HPLC-FD. The reported LODs ranged for the four PAHs
between 0.05 µg kg-1 (BaA) and 0.15 µg kg-1 (BaP). Germán-Hernández et al. (2011; 2012)
applied ionic liquids for the extraction of PAHs from "gofios", a toasted cereal specialty,
which is typically consumed in the Canary Islands. The extraction of 15+1 EU priority PAHs
Downloaded by [University of Tennessee, Knoxville] at 12:50 15 March 2016

from the toasted cereals was carried out by focused-microwave assisted extraction (MAE)
using ionic liquid aggregates of 1-hexadecyl-3-butyl imidazolium bromide (Germán-
Hernándeza et al. 2011), including optional preconcentration with ionic liquid-based
surfactants (Germán-Hernándeza et al. 2012). They reported LODs for the four PAHs studied
in this paper in the range between 2 µg kg-1 and 189 µg kg-1 for the method without pre-
concentration, and 0.07 µg kg-1 to 3.9 µg kg-1with preconcentration. Al-Rashdan et al. applied
Soxhlet extraction, followed by clean up on silica gel and aluminum oxide columns for the
determination of 16 EPA PAHs in baked bread (Al-Rashdan et al. 2010). The measurement
by GC-MS revealed for eight out of 18 test samples BaP levels in the range between 2.8 µg
kg-1 and 16.5 µg kg-1, which is far above the data reported to EFSA. The performance of this
and the other mentioned analytical methods regarding detection capability is in the view of
the authors not sufficient for the generation of reliable occurrence data, which are fit for use
in the assessment of the contribution to human exposure to PAHs of food commodities
containing PAHs only at low levels. Veyrand et al. reported recently a gas chromatography
tandem mass spectrometry (GC-MS/MS) based analysis method for the determination of
PAHs in total diet samples (Veyrand et al. 2013). They published LOD values for solid
samples, but not specifically for bread samples, in the range of 0.01 to 0.02 µg/kg and LOQ
values in the range of 0.03 to 0.06 µg/kg.
The aim of the present work was to design an analytical method that could be used by routine
food control laboratories for the determination of low PAH levels in cereal based foods. The
developed GC/MS method based on pressurized liquid extraction (PLE) followed by SPE on
silica is environmentally friendly due to low solvent consumption, requires short sample
preparation time, and allows many samples to be prepared in parallel without the need of any
time-consuming sample preparation (e.g. gel permeation chromatography (GPC)). The
 analytical procedure is closely related to an analytical method that was recently standardized
by the European Committee for Standardization (CEN) (CEN 2015). However, the
collaborative trial, preceding the standardization of the method, revealed limitations in
performance for low-contaminated flour samples. The modified analytical method presented
in this paper was single-laboratory validated, taking into account compliance with
performance characteristics specified in EU legislation. A special focus was given in the
development of the method to achieving low limits of detection. The studied performance
parameters comprised linearity, repeatability and intermediate precision, limits of detection
and quantification, as well as recovery, and measurement uncertainty. The proposed method
was successfully applied to the determination of PAHs in commercial bread and breakfast
Downloaded by [University of Tennessee, Knoxville] at 12:50 15 March 2016

cereal samples.

Materials and Method

Instruments, chemicals, reagents, and consumables

The GC-MS system used for this study was an Agilent HP 6890N gas chromatograph
coupled to an Agilent 5875B XL mass selective detector (Agilent Technologies S.A. Diegem,
Belgium). The analytical column was a Select PAHs® 15 m length, 0.15 mm internal
diameter and 0.10 µm film thickness which was provided by Agilent Technologies.

Samples were extracted by pressurized liquid extraction with an ASE 300 instrument using
33 mL PLE extraction cell (Dionex Benelux B.V., Tienen, Belgium). Solvent evaporation
was achieved with a stream of nitrogen by applying both a Turbo Vap II sample concentrator
(200 mL capacity, Calliper Life Sciences NV/SA, Teralfene, Belgium) and a Techne
FSC400D sample concentrator (VWR, Leuven, Belgium).

The PAHs investigated in this study were benz[a]anthracene (BaA), chrysene (CHR),
benzo[b]fluoranthene (BbF), and benzo[a]pyrene (BaP). Neat certified reference materials of
native PAHs were obtained from the Joint Research Centre, Institute for Reference Materials
and Measurements (Geel, Belgium). The isotope labelled standards and 9-
fluorobenzo[k]fluoranthene (9-FBkF) were obtained from Chiron AS (Trondheim, Norway).

Ottawa sea sand and cross linked polyacrylic acid (partial sodium salt- graft- poly(ethylene
oxide)) were purchased Sigma-Aldrich (Diegem, Belgium), respectively. Anhydrous sodium
sulphate was from Merck (Darmstadt, Germany).
 All solvents, n-hexane, cyclohexane, toluene (all of them of grade for residue analysis) were
supplied by VWR. The polytetrafluoroethylene (PTFE) membrane filters (1 µm pore size, 25
mm i.d.) were purchased from Sigma Aldrich.

For solid phase extraction clean-up, silica column cartridges of 500 mg/4ml (50 µm particle
size and surface area around 500 m2/g) were purchased from GRACE (Grace Davison
Discovery Science, Columbia, USA).

Test samples used for method validation comprised brown bread, which was purchased at a
local supermarket.
Downloaded by [University of Tennessee, Knoxville] at 12:50 15 March 2016

Commercial bread and breakfast cereal samples with relative fat content levels below 3.5 %
were collected in local stores from Belgium. Major types of breads (wheat bread, whole grain
bread, and brown bread) and breakfast cereals (wheat frosted and rice flakes, whole grain
cereals and coco hoops) were included.

Standard preparation and calibration

Standard stock solutions of the four PAHs were prepared gravimetrically from neat certified
reference materials (100 µg/ml each). A mixed stock solution of 1000 ng/g of the four PAHs
was prepared in toluene. The labelled mixed standard solution containing 150 ng/g in toluene
of benz[a]anthracene-13C 6 , benzo[a]pyrene-13C 4 , benzo[b]fluoranthene-13C 6 , and chrysene-
13
C 12 were prepared from commercial single substance standard solutions in toluene (100
µg/ml). They were employed for the preparation of calibration standards and were added as
internal standards to test samples. An injection standard solution of 400 ng/g in toluene of
9-fluorobenzo[k]fluoranthene was added to the sample extract prior to injection and served as
a reference for assessing recovery of labelled standards.

The final calibration standards were prepared in toluene by dilution of the standard stock
solutions to provide final contents of native PAHs of 1, 10, 25, 40, 55, 70, 85, 100, and 125
ng/g. The isotope labelled analogues were in each calibration standard at a content of 60 ng/g,
and the injection standard solution, 9-fluorobenzo[k]fluoranthene, at a content of 40 ng/g.

Extraction and clean-up

 A method described by Gomez-Ruiz et al. (2009) and Lerda et al. (2011) was used with some
modifications for sample extraction and clean up. Aliquots of 5 g ± 0.1 g of the homogenized
sample (previously ground to a fine powder) were weighed into aluminum weighing boats,
and 200 µl of 13C-labelled PAHs solution were added to each test portion for quantification of
the analytes by isotope dilution. The sample was transferred into a mortar and was mixed
with 5 g of polyacrylic acid and 15 g of pre-cleaned Ottawa sand. Pre-cleaning was carried
out by extraction of the sand by PLE with n-hexane, which aimed to reduce the risk of
interferences. Samples, polyacrylic acid and sand were mixed until visually homogeneous.
Extraction was carried out by pressurised liquid extraction with n-hexane at 100°C in 2
Downloaded by [University of Tennessee, Knoxville] at 12:50 15 March 2016

cycles of 10 min static time with 60% purge volume and 120 s purge time. (Jira et al. 2008,
Gomez-Ruiz et al. 2009, Lerda et al. 2011b). About 5 grams of anhydrous Na 2 SO 4 were put
into each collection bottle before starting extraction to bind the water eventually extracted
with PLE. Afterwards, the extractant was evaporated in the TurboVap® workstation at 40 °C
using a stream of nitrogen.
The evaporated PLE-extract was dissolved in approximately 1 mL of cyclohexane and
filtered through PTFE filters of 1 µm pore size. Solid phase extraction (SPE) on silica was
applied for sample clean up. The silica cartridge was conditioned with 2 mL of cyclohexane.
Consecutively the extract was loaded onto the SPE column. The replaced solvent was
discarded. PAHs were eluted thereafter with 10 mL of cyclohexane, which were entirely
collected.

Toluene (200 µL) was added as a "keeper" to the collected SPE fraction, which was then
concentrated to a volume of about 200 µL at 40ºC by means of the sample concentrator. The
residue was transferred to the GC vial, rinsed with additional 200 µL toluene and mixed with
100 µL solution of injection standard 9-fluorobenzo[k]fluoranthene.

Gas chromatography—mass spectrometry conditions

A gas chromatograph with a programmable-temperature vaporization (PTV) injection port

(septumless head) was used for the analysis of the target PAHs. The GC was hyphenated to a
single quadrupole mass spectrometer operated in electron ionization (EI) mode at 70 eV.
10 μl of sample extract was injected into the PTV injector in split-vent mode at an initial
temperature of 55°C for 0.5 min at a vent flow of 100 ml/min; the temperature was then
 ramped up at 600° C/min to 400°C with a 15 min static time. Transfer line temperature and
ion source temperature were maintained at 325 °C and 300 °C, respectively. Injection was
carried out using an automated GC PAL injection system (CTC Analytics, Zwingen,
Switzerland).

Separation of PAHs was performed on a capillary column which is specified for PAHs
analysis. The oven temperature program started at 80 ºC, held for 3 min, and then continued
with a first ramp: 60 ºC/min up to 200 ºC with a static time of 0 min, second ramp: 2 ºC/min
up to 220 ºC with a static time of 0 min, third ramp: 3 ºC/min up to 285 ºC with a static time
of 0 min, fourth ramp: 5 ºC/min up to 325 ºC with a final static time of 2 min.
Downloaded by [University of Tennessee, Knoxville] at 12:50 15 March 2016

Chromatograms were recorded in Single Ion Monitoring (SIM) mode. Qualification of

recorded peaks is a prerequisite for quantification. This is frequently a limiting factor in PAH
analysis due to the poor fragmentation of the analytes. However peaks were not attributed to
the relevant PAHs unless the qualifier ions were present and the ratio between quantifier and
qualifier ions were within the accepted range. Quantification of the analytes was performed
by isotope dilution via the areas of the base peak ions (see Table 1).

Results and discussion

Method validation

The method was validated to assess its applicability to the analysis of the four EU marker
PAHs in bread and cornflakes samples applying an experimental plan which is in line with
principles specified in internationally accepted guidance documents (Eurachem 2014).
Different parameters such as selectivity, linearity, limit of detection (LOD), limit of
quantitation (LOQ), precision (repeatability and intermediate precision), recovery, trueness,
and also uncertainty were evaluated. The working range for the method was between
0.05 µg kg-1 to 12.5 µg kg-1 which corresponded to the content range for which the
instrument was calibrated.
 Selectivity and Linearity

Selectivity was evaluated from procedural blank samples (PB) and low contaminated test
samples, which were checked for interferences at the expected retention time of the analytes.
The procedural blank samples were prepared by executing the whole analysis procedure
without test sample and omitting the addition of isotope labelled PAH standards. The latter
was done in order to exclude any influence of the isotope pattern of the labelled standards on
the identification of interferences stemming from the analysis system. Selectivity was defined
acceptable in case of absence of peaks in the chromatogram of the procedural blank sample at
Downloaded by [University of Tennessee, Knoxville] at 12:50 15 March 2016

the retention time of the analytes ± 0.1 minute if peaks did not exceed 30 % of the height of
the native analyte in the chromatogram of the lowest calibration level. Spectral interference
close to the retention time of the analytes was between 2 to 16 per cent of the peak heights of
analyte peaks at the lowest calibration level. The background levels increased in
chromatograms generally with increasing total fat content of the test samples, which is due to
the limited retention of fat on the SPE cartridge. However, the focus of this study was on
food samples with low fat content. Selectivity was not compromised up to a labelled fat
content level of 3.5 %.

Linearity was evaluated from the calibration function, obtained from the ratio of the signals
of the native and the labelled PAH by visual inspection of the plot of residuals, respectively
by Mandel’s test. For the four target analytes, Mandel’s tests were passed and no trend was
observed in any of the residual plots.

Limit of detection (LOD) and Limit of quantification (LOQ)

The LOD for the analysis of bread was determined from the measurement of eight different
bread samples applying the paired observation approach (Fajgelj & Ambrus 2000). This
approach was chosen because firstly all tested bread samples contained low amounts of the
analytes, and secondly it allowed for the estimation of LOD based on different bread samples
(with low, but different content levels of native PAHs), which is advantageous for covering
matrix effects. The paired observation approach consists of the analysis of both native low
contaminated bread samples and the same samples spiked with the analytes to a low content
 level. The difference of the signal values of the spiked sample and of the native sample
represents the signal which is attributable to the spiked amount. Consequently, the variability
of this signal is used for the estimation of LOD and LOQ according to DIN 32645:2008-11
(DIN 2008). LOD and LOQ values for cornflakes were derived analogously.

LOD was calculated according to Equation 1. The constant term takes into account equal
probabilities (p=0.05) for false positive and false negative decisions and the respective
degrees of freedom.

𝒔
Downloaded by [University of Tennessee, Knoxville] at 12:50 15 March 2016

Equation 1: 𝒙𝑳𝑶𝑫 = 𝟓. 𝟒 ∗
𝒃

x LOD : content value for LOD, b = slope of the calibration curve, s = standard deviation of the
signal difference

In analogy to LOD, LOQ was estimated via the paired observation approach. It was
approximated for the individual analytes by Equation 2.

𝐬
Equation 2: 𝒙𝑳𝑶𝑸 ≈ 𝟏𝟎. 𝟎 ∗
𝒃

b = slope of the calibration curve, s = standard deviation of the signal difference

As shown in Table 2, the LOD values were below the pursued target level of 0.05 µg kg-1 and
the LOQ values were within the target interval required by Commission Regulation (EU) No
836/2011 (EC 2011b).

Precision (repeatability and intermediate precision)

Precision was assessed at the lowest limit of the working range level from composite low
contaminated cornflakes and composite low contaminated bread samples, which were spiked
 with a PAH standard solution resulting in an increase in the content level of 0.1 µg kg-1.
Triplicate analyses on three different days were performed on each spiked sample. The results
were evaluated for homogeneity of variances using the Cochran test, which was passed for all
data sets. Repeatability and intermediate precision estimates were calculated by ANOVA and
expressed as relative standard deviation. The obtained repeatability relative standard
deviations (RSDr) were for all four analytes below 15 %.

Intermediate precision was evaluated from triplicate analysis in three independent analysis
sequences distributed over a period of one month. The obtained intermediate precision
relative standard deviations (RSD IP ) were for all four analytes below 20 %. The results for
Downloaded by [University of Tennessee, Knoxville] at 12:50 15 March 2016

repeatability and intermediate precision are shown in Table 2.

Additionally, in accordance with Commission Regulation (EU) No 836/2011 both HORRAT r

and HORRAT R values were calculated. HORRAT r is defined there as value of observed
repeatability relative standard deviation (RSD r ) divided by the corresponding precision
estimates derived from the modified Horwitz equation, whereas relative intermediate
precision values (RSD IP ) are applied for calculation of HORRAT R values (see Table 2).
Analytical methods may be regarded fit-for-purpose, if the respective HORRAT values are
below a value of two (EC 2011b). These criteria were fulfilled for all four analytes.

Recovery

Recovery values (see Table 3) were calculated from the measurement results of 13C- labelled
PAHs, assuming that their behavior during all steps of the analytical procedure does not
significantly differ from that of the native PAHs. The average recoveries and their relative
standard deviations were calculated from all results for each labelled PAH via response
factors, which were determined from calibration solutions, between labelled PAHs and the
injection standard (9-FBkF). It has to be noted that recovery values were not used for
correction of analysis results. They were only used to monitor the extraction yield. Recovery
values of individual PAHs were in the range between 86 % and 91 %, with RSDs between
5 % and 9 %. Table 3 contains a summary of the results for cornflakes samples and bread
samples.
 Trueness

It should be noted that there is no certified reference material for PAHs in cornflakes and
bread available. Therefore, the agreement of analysis results with the nominal content of
spiked materials was evaluated by comparing for each analyte the expanded measurement
uncertainty with the deviation from the nominal analyte content. Spiking standard solutions
and calibration standard solution used in the experiments were independent from each other.
Downloaded by [University of Tennessee, Knoxville] at 12:50 15 March 2016

Trueness of results was assessed for the four analytes at the spiking level of 0.1 µg kg-1. The
absolute difference (Δ) between the measured content and the spiked content was compared
to the expanded measurement uncertainty (U). The results were considered not statistically
significantly different from the nominal spiking value if Δ was smaller than U, which was the
case for all the four PAHs spiked into bread and cornflakes.

Measurement uncertainty

Measurement uncertainty was estimated according to the law of error propagation. The
uncertainty contributions that were considered in the combined uncertainty were the
uncertainties of the preparation of both native and labelled PAH standard solutions for
instrument calibration, the uncertainty of the preparation of spiking solutions for the
preparation of test materials, the uncertainty contribution stemming from instrument
calibration, the uncertainty coming from the precision of the analyses, and the uncertainty of
bias. Measurement data at the 0.1 µg kg-1 content level were used for the calculations.

The target standard measurement uncertainties were derived via the fitness-for-purpose
approach detailed in Commission Regulation (EU) No 836/2011, which sets the tolerable
relative standard uncertainty level for analyte contents below 120 µg kg-1 to 22 % (EC
2011b). This results in a tolerable expanded relative uncertainty (k=2) of 44 %.

Uncertainty values were calculated for each analyte/matrix combination. The expanded
uncertainty was obtained by multiplying the combined standard uncertainty by a coverage
factor of two (k=2), corresponding to a confidence level of approximately 95%. However for
 reasons of simplicity the results obtained for bread and cornflakes were merged and the
higher of the two values was used as estimate of measurement uncertainty. The respective
relative expanded uncertainty values are listed in Table 2.

Applicability to commercial samples

In order to evaluate the applicability of the analysis method, ten each of commercial bread
and breakfast cereal samples with relative fat content levels below 3.5 % were analyzed.
Breakfast cereals showed higher mean levels for the sum of the four EU marker PAHs
Downloaded by [University of Tennessee, Knoxville] at 12:50 15 March 2016

(ΣPAH4) (0.23–0.87 µg kg-1) than the samples of bread (0.11–0.22 µg kg-1). Benzo[a]pyrene
(BAP) was the most prevalent PAH in these types of samples. The BaP contents of
commercial bread samples were between "below LOQ" and 0.20 µg kg-1 and those of
breakfast cereal samples between "below LOQ" and 0.30 µg kg-1. All results were below the
maximum level specified EU legislation (EC 2011a) for processed cereal-based foods for
infants.

The obtained data agree well with results of the second French total diet study. In this study,
the middle bound content level of BaP in bread and rusk was determined to 0.04 µg kg-1, with
67% of samples at levels above LOD (Veyrand et al. 2013). The PAH levels in bread
reported by Orecchio and Pupazza (2009) were much higher (e.g. BaP content levels between
0.13 µg/kg dry weight (d.w.) and 9.4 µg/kg d.w.). The high levels can be reasoned on the one
hand by expression of results on dry weight basis, and on the other hand by the application of
wood as a fuel for baking, whereas all breads investigated in the present study were baked in
electrical ovens. Wheat bread samples investigated by Rey-Salgueiro et al. (2008) did not
contain any detectable amounts of PAHs. Measurable contents were found by them only after
exposure of wheat bread slices to combustion gases of different types of fuels. Ciecierska et
al. (2013) determined low levels (0.09 to 0.17 µg/kg) of BaA and CHR in rye breads baked at
different temperatures, but infrequently detected the five-ring PAHs BbF and BaP, which
might be explained by the higher LODs of their analytical method, compared to the method
applied in this study, or used by Veyrand et al. (Veyrand et al. 2013)

The analyses of commercial samples confirmed that the developed method is suitable for
application in routine analysis. However, further sample clean-up would be necessary for
breads with higher relative total fat content levels, e.g. breads containing oilseeds. For such
 samples, a new sample preparation technique and GC/MS based analysis method is under
development.

Conclusions

A rapid and sensitive isotope dilution GC/MS method was developed and single laboratory
validated for the analysis of the four EU marker PAHs in breads and breakfast cereals with
relative total fat content levels below 3.5 %. Sample preparation of the method is based on
pressurized liquid extraction followed by SPE on silica. The developed method here is
Downloaded by [University of Tennessee, Knoxville] at 12:50 15 March 2016

environmentally friendly due to low solvent consumption, requires short sample preparation
time, and allows many samples to be prepared in parallel without the need any time-
consuming sample preparation (e.g. gel permeation chromatography (GPC)). Recovery
values were in the range of 85 % to 95 %, and LOQs were equal or less than 0.05 µg kg-1.
The performance of the analysis method complies in terms of selectivity, precision, recovery,
limits of detection and of quantification, as well as measurement uncertainty with provisions
given in EU legislation.

Acknowledgements

Sibel Kacmaz wishes to express her thanks to TUBITAK (The Scientific and Technological
Research Council of Turkey) BIDEB-2219 international post-doctoral research fellowship
program for financial support of the project, and the European Commission's Joint Research
Centre, Institute for Reference Materials and Measurements, for hosting the project.
 References

Al-Rashdan A, Helaleh MIH, Nisar A, Ibtisam A, Al-Ballam Z. 2010. Determination of the

Levels of Polycyclic Aromatic Hydrocarbons in Toasted Bread Using Gas
ChromatographyMass Spectrometry, Int Journal of Anal Chem. Article ID 821216.

Ciecierska M, Obiedzin´ ski MW. 2013. Polycyclic aromatic hydrocarbons in the bakery
chain. Food Chem. 141: 1–9.
Downloaded by [University of Tennessee, Knoxville] at 12:50 15 March 2016

Codex Alimentarius, CCFAC. 2005. Discussion paper on polycyclic aromatic hydrocarbons

(PAHs) contamination. Thirty-seventh Sessions, The Hague, the Netherlands.

Commission of the European Communities, CEC. 2005. Commission Recommendation (EC)

No. 108/2005 of 4 February 2005 on the further investigation into the levels of polycyclic
aromatic hydrocarbons in certain foods. OJEU L34:43-45.

Commission of the European Communities, CEC. 2006. Commission Regulation (EC) No

1881/2006 of 19 December 2006 setting maximum levels for certain contaminants in
foodstuffs. OJEU L364: 5-24.

Commission of the European Communities, CEC. 2007. Commission Regulation (EC) No.
333/2007 of 28 March 2007 laying down the methods of sampling and analysis for the
official control of the levels of lead, cadmium, mercury, inorganic tin, 3- MCPD and
benzo[a]pyrene in foodstuffs. OJEU L88: 29-38.

European Committee for Standardization, CEN. 2015. Food Analysis – Determination of

benzo[a]pyrene, benz[a]anthracene, chrysene and benzo[b]fluoranthene in foodstuffs by gas
chromatography mass spectrometry (GC-MS). EN 16619:2015 E
 European Commission, EC. 2011a. Commission Regulation (EU) No. 835/2011 of 19 August
2011 amending Regulation (EC) No. 1881/2006 as regards maximum levels for polycyclic
aromatic hydrocarbons in foodstuffs. OJEU L215: 4-8.

European Commission, EC. 2011b. Commission Regulation (EU) No. 836/2011 of 19 August
2011 amending Regulation (EC) No. 333/2007 laying down the methods of sampling and
analysis for the official control of the levels of lead, cadmium, mercury, inorganic tin, 3-
Downloaded by [University of Tennessee, Knoxville] at 12:50 15 March 2016

MCPD and benzo(a)pyrene in foodstuffs. OJEU L215: 9-16.

Deutsches Institut für Normung, DIN (2008) DIN 32645:2008-11 Nachweis-, Erfassungs-
und Bestimmungsgrenze, Ermittlung unter Wiederholbedingungen, Begriffe, Verfahren,
Auswertung. Beuth Verlag GmbH, Berlin, Germany.

Eurachem Guide. 2014. The Fitness for Purpose of Analytical Methods A, Laboratory Guide

to Method Validation and Related Topics, second edition.

https://www.eurachem.org/images/stories/Guides/pdf/MV_guide_2nd_ed_EN.pdf.

European Food Safety Authority, EFSA. 2008. Polycyclic aromatic hydrocarbons in food.
Scientific opinion of the panel on contaminants in the food chain adopted on 9 June 2008.
EFSA J 724: 1–114.

Fajgelj A, Ambrus Á. 2000. Principles and Practices of Method Validation. Royal Society of
Chemistry, Thomas graham house, Science Park, Miltun Road, Cambridge CB4 OWF, UK.
 Germán-Hernández M, Pino V, Anderson JL, Afonso AM. 2011. Use of ionic liquid
aggregates of 1-hexadecyl-3-butyl imidazolium bromide in a focused-microwave assisted
extraction method followed by high-performance liquid chromatography with ultraviolet and
fluorescence detection to determine the 15 + 1 EU priority PAHs in toasted cereals
(“gofios”). Talanta. 85: 1199– 1206.

Germán-Hernández M, Pino V, Anderson JL, Afonso AM. 2012. A novel in situ

preconcentration method with ionic liquid-based surfactants resulting in enhanced sensitivity
for the extraction of polycyclic aromatic hydrocarbons from toasted cereals, Journal of
Downloaded by [University of Tennessee, Knoxville] at 12:50 15 March 2016

Chrom A. 1227: 29– 37.

Gomez-Ruiz JA, Cordeiro F, Wenzl T. 2009. Optimisation and validation of programmed

temperature vaporization (PTV) injection in solvent vent mode for the analysis of the 15 + 1
EU-priority PAHs by GC–MS. Talanta 80, 643–650.

International Agency for Research on Cancer, IARC. 2010. Some Non-heterocyclic

Polycyclic Aromatic Hydrocarbons and Some Related Exposures. IARC Monographs on the
Evaluation of Carcinogenic Risks to Humans. Volume 92. Lyon, France, Avaliable
from: http://monographs.iarc.fr/ENG/Monographs/vol92/mono92-2.pdf. Accessed 24 August
2015.

International Agency for Research on Cancer, IARC. 2012. Chemical agents and related
occupations. A review of human carcinogens. IARC Monographs on the Evaluation of
Carcinogenic Risks to Humans..Volume 100F. Lyon, France. International Agency for
Research on Cancer. Avaliable from:
http://monographs.iarc.fr/ENG/Monographs/vol100F/mono100F.pdf. Accessed 22 February
2016.
 Joint FAO/WHO Expert Committee On Food Additives, JECFA. 2005. Summary and
conclusions of the sixty-fourth meeting, 8-17 February 2005, Rome, Italy.

Lerda D, Lopez P, Szilagyi S, Wenzl T. 2011. Method validation study on a GC-MS based
method for PAHs analysis in food matrices, Report on the inter-laboratory comparison
organized by the European Union Reference Laboratory for the validation of a method for
quantifying the four EU marker PAHs in Food. EUR 25019
EN. https://ec.europa.eu/jrc/sites/default/files/Report.pdf. Accessed 24 August 2015.
Downloaded by [University of Tennessee, Knoxville] at 12:50 15 March 2016

Orecchio S, Papuzza V. 2009. Levels, fingerprint and daily intake of polycyclic aromatic
hydrocarbons (PAHs) in bread baked using wood fuel. J Hazard Mater. 164:876-883.

Ray-Salgueiro L, García-Falcón MS, Martínez-Carballo E, Simal-Gándara J. 2008. Effects of

toasting procedures on the levels of polycyclic aromatic hydrocarbons in toasted bread. Food
Chem. 108:607-615.

Scientific Committee on Food, SCF, 2002a. Polycyclic aromatic hydrocarbons – Occurrence

in foods, dietary exposure and health effects. Report No. SCF/CS/CNTM/ PAH/29 ADD1
Final, of 4 December 2002. http://ec.europa.eu/food/fs/sc/scf/out154_en.pdf. Accessed 24
August 2015.

Scientific Committee on Food, SCF, 2002b. Opinion of the Scientific Committee on Food on
the risks to human health of Polycyclic Aromatic Hydrocarbons in food, expressed on 4
December 2002. Report No. SCF/CS/CNTM/PAH/29
Final. http://ec.europa.eu/food/fs/sc/scf/out153_en.pdf. Accessed 24 August 2015.
 Veyrand B, Sirot V, Durand S, Pollono C, Marchand P, Dervilly-Pinel G, Tard A, Leblanc
J-C, Le Bizec B. 2013. Human dietary exposure to polycyclic aromatic hydrocarbons: Results
of the second French Total Diet Study. Environm Int. 54:11-17.
Downloaded by [University of Tennessee, Knoxville] at 12:50 15 March 2016
 Table 1 – Retention times, recorded ions and ion ratios of respective analyte peaks

Base peak ion Qualifier ion

Retention Ion ratio Ion ratio - lower Ion ratio - upper
PAH
Downloaded by [University of Tennessee, Knoxville] at 12:50 15 March 2016

time (Q 1 ) (Q 2 ) (Q 1 /Q 2 ) tolerance limit tolerance limit

min m/z m/z

Benz[a]anthracene (BaA) 19.01 228 114 12 6 18

Benzo[a]pyrene (BaP) 29.35 252 126 20 10 29

Benzo[b]fluoranthene (BbF) 26.65 252 126 13 6 20

Chrysene (CHR) 19.55 228 114 12 6 18

9-
25.76 270 135 15 7 22
fluorobenzo[k]fluoranthene

Benz[a]anthracene 13C 6 19.01 234 116 10 5 15

Benzo[a]pyrene 13C 4 29.35 256 128 13 10 16

Benzo[b]fluoranthene 13C 6 26.66 258 129 12 10 14

Chrysene 13C 6 19.55 234 116 11 9 13

 Table 2 – Limit of detection (LOD), limit of quantification (LOQ), repeatability relative standard deviation (RSD r ), intermediate precision

relative standard deviation (RSD IP ) and HORRAT values for repeatability (HOR r ) and reproducibility (HOR R ) and relative expanded

measurement uncertainty (U) for the determination of four EU marker PAHs in bread and cornflakes samples which were both spiked with
Downloaded by [University of Tennessee, Knoxville] at 12:50 15 March 2016

0.1 µg/kg of each of the four analytes.

Bread Cornflakes

Analyte
LOD LOQ RSD r RSD IP HOR r HOR R LOD LOQ RSD r RSD IP HOR r HOR R U* (k=2)

[μg/kg] [μg/kg] % % [μg/kg] [μg/kg] % % %

BaA 0.02 0.03 6.6 6.6 0.5 0.3 0.01 0.02 5.4 11.6 0.4 0.5 14

CHR 0.02 0.03 6.8 6.8 0.5 0.3 0.01 0.02 6.2 17.0 0.4 0.8 20

BbF 0.03 0.05 11.0 11.0 0.7 0.5 0.02 0.03 7.6 12.2 0.5 0.6 14

BaP 0.02 0.03 8.1 10.5 0.6 0.5 0.02 0.04 8.3 9.3 0.6 0.4 12

* applied for both bread and cornflakes

 Table 3 - Recovery of labelled PAHs from cornflakes and bread samples.

Cornflakes Bread

Mean Mean SD
SD %
Recovery% Recovery% %

Benz[a]anthracene 13C 6 91 6 88 5

Chrysene 13C 6 86 5 86 5
Downloaded by [University of Tennessee, Knoxville] at 12:50 15 March 2016

Benzo[b]fluoranthene 13C 6 86 9 88 4

Benzo[a]pyrene 13C 4 90 5 91 5

Graphical Abstract