A fast and selective method for the determination of 8 carcinogenic polycyclic aromatic hydrocarbons in rubber and plastic materials

Journal of Chromatography A, 1566 (2018) 13–22

Contents lists available at ScienceDirect

Journal of Chromatography A
journal homepage: www.elsevier.com/locate/chroma

A fast and selective method for the determination of 8 carcinogenic
polycyclic aromatic hydrocarbons in rubber and plastic materials
Otmar Geiss ∗ , Chiara Senaldi, Ivana Bianchi, Ana Lucena, Salvatore Tirendi,
Josefa Barrero-Moreno
European Commission, Joint Research Centre, Directorate F Health, Consumers and Reference Materials, 21027, Ispra, VA, Italy

a r t i c l e

i n f o

Article history:
Received 22 March 2018
Received in revised form 15 June 2018
Accepted 19 June 2018
Available online 21 June 2018
Keywords:
Polycyclic aromatic hydrocarbons
Rubber and plastic materials
Randall hot extraction
Molecularly imprinted polymers

a b s t r a c t
Polycyclic Aromatic Hydrocarbons (PAHs) have been detected in rubber and plastic components of a
number of consumer products such as toys, tools for domestic use, sports equipment, and footwear, with

carbon black and extender oils having been identified as principal sources. In response to these findings,
the European Union Regulation (EU) No. 1272/2013 was adopted in December 2013, amending entry 50
in Annex XVII to the Registration, Evaluation, Authorisation, and Restriction of Chemicals (REACH) directive establishing a restriction on the content of eight individual carcinogenic PAHs in plastic and rubber
parts of products supplied to the public. This work proposes a simple, relatively fast, and cost effective
method for determining the concentrations of each of these eight carcinogenic PAHs for compliance testing. Existing methodologies were taken as a starting point, improving in particular the extraction and
the clean-up procedures. Randall hot extraction and ultrasonic extraction were compared with regard
to their extraction efficiency. Randall hot extraction proved to be more efficient (10–40%, depending
on PAH). Sample extract clean-up performance was qualitatively assessed for silica-packed columns
and molecularly imprinted polymers (MIPs) solid phase extraction (SPE) cartridges. The use of highly
selective MIP-SPE cartridges removed most of the undesired contaminants, highlighting their superiority with regard to traditional, silica-based purification methodologies. The introduction of Randall-hot
extraction for sample extraction and MIP-based solid phase extraction cartridges for selective clean-up
represents a novel advance compared with previously reported methods in this field. In combination
with gas chromatography-mass spectrometry (GC–MS) analyses in selected ion mode, the method was
found to be excellent in terms of extraction efficiency, extract purity, and speed.
© 2018 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license
( />
1. Introduction
Polycyclic aromatic hydrocarbons (PAHs) are a ubiquitous group
of several hundred chemically related, environmentally persistent
compounds, many of which have mutagenic, toxic, and/or carcinogenic properties [1,2]. Known sources of PAHs in consumer
products are extender oils and carbon black which may unintentionally contain various levels of PAHs [3,4]. Carbon black is used
as reinforcing filler in rubber formulations or as pigment in plastics [5], whereas extender oils are used as plasticiser oils/softeners.
Examples of consumer products in which PAHs were detected are:
household items, tools, clothing, footwear, toys and childcare arti-

∗ Corresponding author at: European Commission, Joint Research Centre, Directorate F – Health, Consumers and Reference, Materials, Via E. Fermi, 2749, 21027,
Ispra, VA, Italy.
E-mail address: (O. Geiss).

cles [5] as well as floor tiles used in gyms or on synthetic turf

fields made of recycled rubber [6]. Consumer exposure, in particular of children, to PAHs via oral and dermal contact with these
articles has been the subject of considerable public attention over
the past decade [7], triggering the establishment of restrictions on
PAH content. According to paragraphs 5 and 6 of entry 50 in Annex
XVII to Regulation (EC) No. 1907/2006 of the European Parliament
and Council from 18 December 2006 concerning the Registration,
Evaluation, Authorisation, and Restriction of Chemicals (REACH)
[8], only articles compliant with the conditions laid out therein,
including the limits on polycyclic aromatic hydrocarbons (PAHs),
can be sold to the general public in the European Union (EU). For
eight priority PAHs – all presumed carcinogens (group 1B according
to CLP classification) - content limits of rubber and plastic components were set to 0.5 mg kg−1 for toys and childcare articles
and to 1 mg kg−1 for all other consumer articles. Reliable methods/standards to analytically determine the concentrations of each
of these eight carcinogenic PAHs are required for compliance test-

/>0021-9673/© 2018 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license ( />

14

O. Geiss et al. / J. Chromatogr. A 1566 (2018) 13–22

ing against the established limit values. The European Commission
recently (December 2017) launched a standardisation request to
the European Committee for Standardisation [7] as no such harmonised European or international standard is available at present.
A number of methods for the determination of PAHs in a variety of
matrices, that are in most cases neither rubber nor plastics, are
available. From among these methods, that used in the German
Product Safety Commission (AfPS) GS 2014:01 PAK specification
[9], in which the AfPS sets the requirements of PAH testing in the
course of GS mark certification, comes closest to EU requirements.

This AfPS requirement restricts 18 PAHs in articles and consumer
products and proposes a method for the analysis of PAHs in rubber
and plastic, based on 1 h ultrasound extraction in toluene at 60 ◦ C,
purification of the extract on silica gel packed columns, and subsequent analysis with GC MS. The Fraunhofer Institute for Process
Engineering and Packaging (IVV) has recently published a method
[10] to determine PAH concentrations in rubber material, based
on a more efficient and faster extraction process, that uses accelerated solvent extraction with cyclohexane instead of ultrasound
extraction with toluene. After extraction, the sample is purified
with normal-phase SPE cartridges and analysed with GC–MS.
ISO/TS 16,190 [11] describes a test method to quantitatively
determine polycyclic aromatic hydrocarbons (PAH) in footwear
materials by using n-hexane at 60 ◦ C in an ultrasonic bath for 1 h to
extract the test material. An aliquot is then analysed using either
GC–MS or HPLC without prior purification of the sample extract.
CEN EN 16,143-2013 [12] describes a method for the determination of benzo[a]pyrene and other selected PAHs in petroleum
products (e.g. extender oils). The product is dissolved in n-pentane
and submitted to a double cleaning step using silica-based column
chromatography. The final extract is then analysed by GC–MS. ISO
method 21,461 [13] allows for the selective determination of polyaromaticity of oil in vulcanized rubber compounds. This method
is not specific to individual PAHs and is based on high-cost nuclear
magnetic resonance (NMR) spectrometry. A study commissioned
by the carbon black industry [14] investigated the extraction and
migration behaviour of PAHs from rubber formulations containing carbon black. They proposed a method in which the sample
material is Soxhlet-extracted with toluene, the sample extract then
purified with silica gel packed chromatographic columns, and the
analysis carried out with GC–MS. While some of the above methods
entail high equipment costs as a main challenge for implementing them in enforcement scenarios, others lack the specificity and
capability of determining individual PAH concentrations in plastic and rubber materials which imposes severe limitations on their
applicability for these purposes. In addition, some of these methods
suffer from outdated protocols, often in the extraction or purification steps which are no longer state of the art. A comparison of

already published methods/studies in terms of sample extraction
and sample purification can be found in Table 1.
The purpose of this study was to develop a simple, relatively
fast, and cost effective method for the determination of 8 priority PAHs in rubber and plastic materials. Existing methodologies
served as starting points with the particular aim to improve the
extraction and the clean-up steps. Randall hot extraction and ultrasonic extraction were compared with regard to their extraction
efficiency. Sample extract clean-up performance was qualitatively
assessed for silica-packed columns and molecularly imprinted
polymers (MIPs) solid phase extraction (SPE) cartridges. We found
that the optimal method in terms of extraction efficiency, extract
purity, and time demands, was Randall hot extraction of the rubber or plastic material, followed by sample extract clean-up with
PAH specific solid phase extraction cartridges based on molecularly
imprinted polymers (MIPs), in combination with GC–MS analysis
in selected ion mode. The introduction of Randall-hot extraction for
sample extraction and MIP-based solid phase extraction cartridges

for selective clean-up represents a novel and significant advance
compared with previously reported methods in this field.
2. Materials and methods
2.1. Selection of test materials and sample preparation
We chose three test materials for this study: carbon black
N772 (2.5%) containing soft polyvinylchloride (PVC) and two natural/butadiene rubber (NR/BR) blends; one containing carbon black
N375 (24.1%) and distillate aromatic extract (DAE, 2.7%), and the
other containing carbon black N375 (24.1%) and treated distillate
aromatic extract (TDAE, 2.7%). These materials were chosen for
two reasons: (i) they should be representative of the two polymer groups: plastics and rubber; and (ii) they should be able to
account for the impact different PAH concentrations in the same
material could have on factors such as the recovery rates. Soft PVC
was chosen as the representative of plastic materials as its high
phthalate content poses a challenge in the clean-up step. Since singularly available, pure carbon blacks N375 and N772, as well as

pure treated and untreated distillate aromatic extracts (the same
used during the manufacturing process of the NR/BR blends and soft
PVC), were separately analysed as well. Although not described in
this study, the present protocol was also applied for the analysis of
low density polyethylene, polystyrene, ethylene propylene diene
monomer and silicone rubber and proved to be suitable. All test
materials were custom made at external laboratories. It should be
noted that the rubber blends and the PVC used in this study are
not necessarily representative of common commercial consumer
products.
2.2. Sample preparation – size of sample material
Using scissors, the sample materials were cut into pieces with
an edge length of < 2 mm.
2.3. Extraction
While the main extraction method used in this study is Randall
hot extraction, we also employed ultrasonic extraction for purposes
of comparison. The weighed-in quantities of sample material were
adjusted so their PAH content would fall within the linear calibration range of the GC detector. The weighings given below are
therefore only valid for the specific materials investigated in this
study.
2.3.1. Randall hot extraction
Around 40 mg of the DAE containing NR/BR, or 100 mg of either
the TDAE containing NR/BR or the PVC were weighed exactly into
cellulose extraction thimbles (Whatman, Maidstone, UK, single
thickness, 33 mm x 80 mm, Product Code 2,800,338), which were
transferred to the respective extraction cups, adding 20 ␮L of isotope labelled internal standard (2500 ng mL−1 in toluene) to the
interior base of the extraction thimbles, next to the sample material. Then, 95 mL of toluene were added to the extraction cups.
The extraction was done with a Velp SER 158 solvent autoextractor (Velp Scientifica, Usmate, Italy) set to the highest heating level,
using 120 min. for immersion, 20 min. for removal, 30 min. to wash,
7 min. for recovery, and 15 min. for cooling, which resulted in a total

extraction time of just over 3 h and a final sample extract volume of
approximately 20 mL. Gaskets made of Vaflon were used between
the connection funnel solvent and the extraction cup.
2.3.2. Ultrasonic extraction
Around 40 mg of the DAE containing NR/BR or around 100 mg of
either the TDAE containing NR/BR or the PVC were weighed exactly


Table 1
Comparison of methods/studies for the extraction and determination of PAHs from rubber and/or plastics with regard to sample extraction and sample purification.
Sample Extraction

Sample extract purification

Remarks

Technique

Timed

Coste

Extraction
efficiency

Technique

Time consuming?

Coste

Clean-up
efficiency

Study commissioned 2009
by carbon black
industry [14]

Toluene

Soxhlet

16 h
(320 cycles)

3-5 kD
(6 pos.)

Complete
extractiondc

Silica gel packed
column

Yes. Packing and
long conditioning
procedure.

Variable

Unselective.
Only polar
compounds are
retained.

ISO/TS 16,190 [11]

2013

n-Hexane

Ultrasounds
extractions

1h

1-2 kD

Incomplete
extractionb

No clean-up

n/a

n/a

No clean-up

CEN EN 16,143 [12]

2013

n-Pentan

n/aa

n/aa

n/aa

n/aa

Silica gel packed
column

Yes. Packing and
long conditioning
procedure.

Variable

Unselective.
Only polar
compounds are
retained.

AFPS GS2014:01
PAK [9]

2014

Toluene

Ultrasounds
extractions

1h

Incomplete
extractionb

Silica gel packed
column

Yes. Packing and
long conditioning
procedure.

Variable

Unselective.
Only polar
compounds are
retained.

Fraunhofer IVV [10]

2017

Cyclo-hexane

Accelerated solvent
extraction (ASE)

45 min

45-50 kD

Complete
extractionc

Normal phase SPE

No. Commercially
available.

4 D /cartr.

Unselective.
Only polar
compounds are
retained.

Toluene

Randall-hot
extraction

3h

10-14 kD (6
pos.)

Complete
extractionc

MIP-SPE

No. Commercially
available and
easy/fast protocol.

5 D /cartr.

Highly selective.

Method proposed in
this study
a
b
c
d
e

Year of
publication

Extraction of PAHs
from cured rubber

formulations
containing carbon
black. Not tested for
plastic materials.
Method for
determination of
PAHs in footwear
materials. Not
specific for rubber
and plastic materials
Method for
determination of
benzo[a]pyrene and
other PAHs in
petroleum products.
Used in the course of
German GS mark
certification.
Method which
comes closest to the
method described in
this study.
Method for
extraction of PAHs
from recycled
rubber. Not tested
for plastic materials.

O. Geiss et al. / J. Chromatogr. A 1566 (2018) 13–22

Extraction
Solvent

Method/Study

Petroleum product directly solved in n-pentane. No extraction required.
The current study suggests that ultrasound extraction may be less efficient compared to Randall hot extraction.
Extraction efficiency verified in referenced study.
Total extraction time.
Indicative acquisition cost (Italy).

15


16

O. Geiss et al. / J. Chromatogr. A 1566 (2018) 13–22

into 100 mL Erlenmeyer flasks, to which 95 mL toluene and 20 ␮L of
isotope labelled internal standard (2500 ng mL−1 in toluene) were
added. The capped flasks were then placed into an ultrasonic bath
(800 W, 59 KHz, bath area 900 cm2 ) for 1 h at 60 ◦ C.
2.4. Clean-Up/Purification
Toluene sample extracts were evaporated to dryness with a
rotary evaporator, in a 60 ◦ C water bath and adjusting the vacuum
to 90 mbar. The clean-up procedure was done with solid phase
extraction cartridges (SPE) filled with a sorbent based on highly
PAH-selective molecularly imprinted polymers (MIPs). Results
from this approach were compared against results obtained applying the German AfPS method [9], which is based on adsorption
chromatography on silica gel.

TM

2.4.1. Solid phase extraction with SupelMIP cartridges
The dry extract was reconstituted in 1 mL hexane and cleanedup using solid phase extraction (SPE) cartridges filled with MIPs
TM
(Supelco, SupelMIP PAHs, 50 mg/3 mL, product code 52773-U).
The following procedure was used: conditioning 1 mL cyclohexane, loading of sample, washing with 2 mL cyclohexane, elution
with 3 x 1 mL ethylacetate, evaporating the ethylacetate extract to
dryness with a nitrogen evaporator (heating block set to 40 ◦ C), and
reconstituting in 1 mL toluene for GC–MS analysis.
2.4.2. Clean-up with silica gel packed columns
A glass chromatography column (20 cm × 2 cm) was packed
with 4 g of previously deactivated silica (Supelco, Washed silica,
product code 21342-U), employing the wet packing method, to
which 1 cm of anhydrous sodium sulphate was added. Deactivation was achieved by adding 10% in weight of ultrapure water to
the silica and subsequent homogenisation for 1 h. The column was
conditioned with 10 mL petroleum ether. The dry extract was then
reconstituted in 1 mL toluene and loaded onto the column. Elution of PAH was achieved with 50 mL petroleum ether. In the next
step, the petroleum ether extract was evaporated to dryness with a
nitrogen evaporator (heating block set to 40 ◦ C) and reconstituted
in 1 mL toluene for GC–MS analysis.
TM

2.4.3. Recovery determination using SupelMIP SPE cartridges
To determine the recovery rates, 100 mg of distillate aromatic
extract, for which the PAH content had previously been determined, were initially dissolved in 100 mL hexane. Then 1 mL of this
extract was loaded onto the SPE columns and treated as described
above (Section 2.4.1). A further 1 mL of hexane extract was evaporated to dryness with a nitrogen evaporator and reconstituted
in 1 mL toluene. Both solutions were analysed with GC–MS to
determine the ratio of the respective peak areas. Absolute masses

loaded onto the cartridges were 18 ng of benzo[a]anthracene (BaA),
90 ng of chrysene (Chr), 40 ng of benzo[b]fluoranthene (BbF), 6 ng
of benzo[k]fluoranthen (BkF), 8 ng of benzo[j]fluoranthene (BjF),
100 ng of benzo[e]pyrene (BeP), 25 ng of benzo[a]pyrene (BaP) and
2 ng of dibenzo[a,h]anthracene (DbahA). Recovery determinations
were made in quintuplicate.
2.5. Analysis
2.5.1. Preparation of isotope labelled internal and native PAH
standard solutions
Native and isotope labelled internal PAH standard calibration kits were purchased from Lab Service Analytica
Srl (Anzola Dell’Emilia, Italy). The native standard contained benzo[a]anthracene, chrysene, benzo[b]fluoranthene,
benzo[k]fluoranthene, benzo[j]fluoranthene, benzo[e]pyrene,
benzo[a]pyrene, and dibenzo[a,h]anthracene, each at a

concentration of 25 ␮g mL−1 in toluene. The isotope
labelled
standard
contained
benzo[a]anthracene-d12,
chrysene-d12, benzo[b]fluoranthene-d12, benzo[k]fluoranthened12,
benzo[e]pyrene-d12,
benzo[a]pyrene-d12
and
dibenzo[a,h]anthracene-d14, each at a concentration of 25 ␮g
mL−1 in toluene. Spiking solutions of the isotope labelled internal
standards were prepared in toluene at a concentration of 2500 ng
mL−1 .
2.5.2. Analytical determination of PAHs with GC–MS
The concentrations of each of the eight PAHs were determined
through gas chromatography coupled to a mass spectrometer (Agilent 7890 A/5975C, Santa Clara CA, USA) in selected ion mode (SIM)

against internal standards. Five calibration levels were prepared in
the range 0–200 ng mL−1 (5, 10, 50, 100, and 200 ng mL−1 ; internal
standard 50 ng mL−1 ). The separation column used was a Rxi-PAH
column (Restek, Bellefonte PA, USA, 30 m, 0.25 mm ID, 0.10 ␮m df,
product code 49,318) with an injection volume of 1 ␮L in splitless
mode (1 min valve time). The injector and transfer line temperatures were set at 300 ◦ C and 320 ◦ C respectively, the carrier gas was
helium at a constant flow of 1.75 mL min−1 . Additional details can
be found in the supplementary data (SD1).
2.6. Quality control and quality assurance
To test for possible contaminations during work-off (extraction, clean-up and analysis), blank (95 mL toluene) solutions spiked
with 20 ␮L internal standard solution (2500 ng mL−1 ) were analysed for each batch of analyses. Pure toluene was injected prior to
each sample sequence to exclude solvent contamination. Solvents
were always found free of PAH contamination. A reference chromatogram with all native and deuterated standards can be found
in the supplementary data (SD2).
2.7. Method performance parameters
2.7.1. Recovery, precision, analytical range, linearity and control
sample
The recovery was determined for concentrations at the lower
and upper end of the calibration curve. Elftex® TP carbon black
containing polyvinylchloride (PVC), which was previously shown
not to contain any PAHs, was used as an interference-free blank
matrix. Approximately 50 mg of the blank matrix was spiked with
10 ␮L and 80 ␮L of native standard solution (2500 ng mL−1 ) corresponding to 25 ng and 200 ng of each of the 8 PAHs respectively.
Sample extraction, sample concentration and clean-up were carried out as described earlier. The dry extract obtained after the
evaporation step (2.4.1) was reconstituted with 50 ng mL−1 internal standard containing 1 mL toluene for GC–MS analysis. Recovery
was calculated by dividing the analytically determined concentration with the theoretical spiked concentration. Five replicates and
one non-spiked blank sample were analysed for each spiking level.
For the determination of the precision, samples were analysed in
quintuplicate. Native and internal standards were used to calibrate
the GC–MS instrument over a concentration range of 0–200 ng

mL−1 . Each batch of analyses included a control sample, consisting of 95 mL of toluene spiked with 20 ␮L internal standard and
40 ␮L native standard solutions (both at 2500 ng mL−1 ). All subsequent steps (sample concentration and clean-up) were carried out
as described above for the sample extracts.
2.7.2. Trueness
In the absence of a reference material, trueness - defined as the
closeness of agreement between a test result and a reference value could not be directly determined. However, the content of distilled
aromatic extracts (2.7%) and carbon black (24.1%) were known for


O. Geiss et al. / J. Chromatogr. A 1566 (2018) 13–22

the NR/BR blends and each of these two ingredients were separately
available and could be analysed with regard to their PAH content.
Assuming that carbon black and the distilled aromatic extracts were
the only sources of PAHs, the theoretical final mass-concentration
for the test-material could be determined and compared against
the effectively measured concentrations.
2.7.3. Limit of Quantification (LoQ)
The limit of quantification was determined following the
approach specified in the European Pharmacopoeia [15]. The peakto-peak noise around the analytes retention times was measured,
and subsequently, the concentration of the respective analytes that
would yield a signal-to-noise ratio equal to 10:1 was estimated.
3. Results and Discussion
3.1. Extraction
3.1.1. Considerations concerning the selection of an appropriate
extraction solvent
Selecting a suitable solvent is crucial for the extraction process
as its effectiveness depends on the solvent’s polarity and boiling
point. The non-specific interaction between carbon surfaces with
organic molecules is dominated by dispersion forces [16]. Due to

the large surface area of carbon black, it has a relatively strong
adsorption affinity for PAHs, with extractability becoming optimal
at high temperatures. The two most commonly reported [9–11,14]
solvents for the extraction of PAHs in various matrices are toluene
(bp 101.6 ◦ C) and hexane/cyclohexane (bp 68.7/80.7 ◦ C). Hexane
has a lower polarity index (close to the polarity index of the highly
lipophilic PAHs). However, its boiling point is around 40 ◦ C lower
than for toluene. For this reason, toluene was given preference over
hexane as extraction solvent.
3.1.2. Randall hot extraction compared to ultrasound extraction
Randall hot extraction and ultrasonic extraction were compared
with regard to their extraction efficiency for the materials under
investigation and the conditions used. With the exception of the
extraction technique, all other steps were identical (clean-up with
SPE and analysis).
Mass concentrations achieved when extracting with Randall
hot extraction were always higher compared to the concentrations obtained with ultrasound extraction (Fig. 1). This difference
appears to be independent of the PAH content (Fig. 1A and B) and
the type of material. The contents achieved with ultrasound extraction were between 10–40 % lower, depending on the specific PAH,
suggesting that ultrasound extraction under these conditions may
be less efficient compared to Randall hot extraction. Our extraction procedure deviated from the AfPS protocol [9] in that the AfPS
protocol extracts higher weightings of sample material in lower
amounts of solvent. This deviation was necessary to ensure identical extraction conditions in terms of loading factor for the Randall
and ultrasound extractions and should have no effect on extraction efficiency. The Randall hot extraction process represents an
improvement over the classical Soxhlet extraction technique in
that it considerably shortens the extraction time. Compared to
the classical Soxhlet method where the condensed solvent is at
a temperature below the boiling point, the Randall method has
the sample material completely immersed in boiling solvent which
provides great time savings as analytes are more soluble in boiling solvent. Other benefits of the hot extraction process include

short process paths, low solvent requirements, and a process that
is gentler on the extract (due to the shorter extraction period). A
general drawback of ultrasound extraction is linked to the underlying principle behind the effects of ultrasound sonication in liquids:
cavitation. Cavitation refers to the formation, growth, and collapse

17

of vapour or gas bubbles due to ultrasound [17]. Depending on the
frequency and intensity of the ultrasound waves, the cavitation
bubbles produce very high temperatures and pressures, which can
cause degradation and transformation of organic molecules [18].
This can result in lower recoveries of the desired analytes and/or
increase the amount of undesired compounds extracted from the
matrix.
3.1.3. Completeness of extraction
The Randall hot extraction process can be split into 3 steps:
(i)immersion: the thimble is lowered into the boiling solvent,
(ii)rinsing/washing: the thimble is raised above the boiling solvent
for a period of time until residual extract is removed from the solid
material by the condensed solvent, and (iii) recovery: part of the
solvent is removed from the extraction cup, concentrating the analytes for further processing. Within the margin of error provided
by the standard deviations, no difference in mass concentrations
(content) was observed between immersion times of 2 h and 4 h
(Fig. 1D) which indicates that the extraction was already complete
after immersing for 2 h. Hamm and co-authors [14] investigated the
extraction time/cycles necessary to obtain complete extraction of
PAHs from carbon black with traditional Soxhlet extraction [19,20]
and concluded that 16 h (320 cycles) with toluene were required.
Randall-extraction is known to shorten extraction times by a factor of 4 to 5 when compared to classical Soxhlet extraction. This
is equivalent to a total extraction time of 3–4 hours (including all

three steps) and confirms the quantitative extraction of all PAHs in
our study.
3.2. Purification
Rubber and plastic materials are complex matrices and during the extraction process a number of undesired substances such
as monomers, aliphatic hydrocarbons, and aromatic hydrocarbons
(other than PAHs), as well as plasticizer-additives are extracted
together with the desired analytes. These contaminations lead to
higher detection/quantification limits, accelerated dirtying of the
mass-spectrometer, and - in the presence of phthalates - may
render determination of the correct analyte outright impossible.
For this reason, sample extracts undergo a clean-up before being
injected into the GC MS system. The selection of appropriate and
efficient clean-up procedures in the case of complex matrices can be
challenging [21,22]. Some of the existing methods listed in Table 1
either forego this clean-up process entirely [11] or rely on a normal phase silica clean-up using SPE-cartridges [10] or silica-gel
packed glass columns [9,12,14], which are non-selective and coextract compounds having similar physicochemical characteristics
[23]. In this work, we tested a new SPE-phase, based on molecularly
imprinted polymer (MIP) technology, on extracts obtained through
Randall hot extraction on both the NR/BR blend and the phthalate
containing soft-PVC. MIPs are highly cross-linked polymer phases
that have pre-determined selectivity for a single analyte or a group
of structurally related analytes. To the best of our knowledge, these
SPE cartridges have so far only been used for the extraction and
analysis of PAHs in olive oil [24], environmental water samples
[25], and from tea leaves [26]. Having extracted the same amount
(35 mg) of rubber for all three samples (no clean-up, clean-up with
MIPs SPE and clean-up with silica-gel packed column) allows for
direct comparisons of the areas below the different peaks in the
obtained chromatograms (Fig. 2). Clearly, the silica-gel packed column did not remove most of the undesired contaminants. Baselines
from silica gel purified and unpurified extracts show similar total

ion currents (in Scan and Selected Ion modes). Silica-gel primarily
retains polar compounds which appear to be absent in the sample extract. In particular, the aliphatic hydrocarbons eluted in the
first part of the chromatogram were not removed from the sam-


18

O. Geiss et al. / J. Chromatogr. A 1566 (2018) 13–22

Fig. 1. Comparison of extraction efficiencies. Comparison of the extraction efficiencies for Randall hot extraction and ultrasound extraction for (A) NR/BR with carbon black
N375 and TDEA (100 mg weigh-in), (B) NR/BR with carbon black N375 and DEA (40 mg weigh-in), (C) soft-PVC (100 mg weigh-in). (D) Comparing extraction efficiencies for
2-hour and 4-hour immersion times.

ple extract. The extract purified with the MIP-SPE columns shows a
much cleaner chromatogram, indicated by the lower total ion current baseline. A dominant peak elutes around 21 min. deriving from
the release of an additive from the frits used in the SPE-columns.
However, the retention time of this peak did not interfere with the
retention times of the analytes under the chromatographic conditions described in this work. A clean-up step before using the
column could however be added to remove the impurities in case
of interference. Also in selected ion monitoring mode (lower part
of Fig. 2), the MIP-SPE baseline is lower compared to the other two.
Large amounts of phthalates were extracted together with the
PAHs, which, even in selected ion monitoring mode, resulted in a
relatively large increase in the baseline signal which in turn needed
to be removed before injection (Fig. 3). Both the MIP-SPE and the
silica-gel packed columns successfully removed, almost quantitatively, the more polar phthalates from the sample extract. By
overlaying a standard chromatogram (panel 1 in Fig. 3) onto the
sample extract chromatogram (panel 2 in Fig. 3), we demonstrate
that the eight priority PAHs and their respective deuterated forms
do not co-elute with impurities released from the frits of the SPE

columns.

attributed to the design and intrinsic functioning of MIPs, which
are a class of highly cross linked polymer-based molecular recognition elements engineered to bind one specific target compound
or a class of structurally related compounds. The MIP material is
designed with cavities that are sterically and chemically complementary to the target analytes [27]. Compared to most other PAHs
investigated in this study, benzo[a]anthracene has a relatively small
molecular structure and might therefore be retained less efficiently.
Also benzo[j]fluoranthene exhibited a relatively low recovery rate
which can possibly be attributed both to the relatively low absolute
amount loaded onto the SPE cartridge and non-baseline separation from benzo[k]fluoranthene which added uncertainty to the
integration procedure (cf., RSD column in Table 2). This explanation is only partly satisfactory as it should equally apply to
benzo[k]fluoranthene, which exhibited a higher recovery rate. All
other recovery rates were found to be above 84%. The simple
usage-protocol (fast), low amount of required solvents, commercial availability (no need to pack the column), and, above all, high
selectivity resulting in lower baselines (i.e., lower detection limits),
highlight the superiority of the MIP-based solid phase extraction procedure with regard to traditional, silica-based purification
methodologies.

TM

3.2.1. Recovery experiments of SupelMIP SPE cartridges
Depending on the type of compound, PAH recoveries ranged
from 51 to 95 % (Table 2).
The lowest recoveries were found for benzo[a]anthracene and
benzo[j]fluoranthene with recoveries of 59% and 51%, respectively.
A similarly low recovery rate for benzo[a]anthracene has been
reported in an application note for the extraction and analysis of
TM
PAHs in olive oils using SupelMIP

SPE [24] and can likely be

3.3. Method performance parameters
3.3.1. Recovery, precision, analytical range and linearity
Recoveries and relative standard deviations for each of the eight
PAHs are summarised in Table 3. Values are comparable for both
spiking levels and range from 55 to 85%.


O. Geiss et al. / J. Chromatogr. A 1566 (2018) 13–22

19

Fig. 2. Comparison clean-up efficiency NR/BR extract. Overlaid chromatograms of non-purified NR/BR extract, with MIP-SPE columns cleaned NR/BR extracts and with
silica-gel packed columns purified NR/BR extracts in Scan mode (upper) and selected ion monitoring mode (lower). Sample weigh-in around 40 mg.

Table 2
TM
Recovery rates for priority PAHs purified on SupelMIP SPE cartridges.
Compound

Absolute mass loaded
on column [ng]

Average recovery [%]

RSD [%] n = 5

Recovery of spiked
olive oila [%]

Benzo[a]anthracene
Chrysene
Benzo[b]fluoranthene
Benzo[k]fluoranthene
Benzo[j]fluoranthene
Benzo[e]pyrene
Benzo[a]pyrene
Dibenzo[a,h]anthracene

18
90
40
6
8
100
25
2

59
84
94
89
51
95
93
91

4.3
3.5

2.6
10.8
9.3
1.4
3.5
3.2

65
70
82
84
n/ab
n/ab
87
82

a
b

From Supelco application note 192 [24].
Recovery rates not reported in Supelco application note 192.

Since the recovery values were determined by spiking blank PVC
matrix, values may vary for other plastic or rubber materials. Due
to its high content of phthalates, the clean-up step for this material can be considered as being particularly challenging. Therefore
it can reasonably be assumed that the recovery rates will be similar
or better for other rubber or plastic materials. For the determination of the precision, samples were analysed in quintuplicate and
the relative standard deviation was generally below 5%, except for
benzo[k]- and benzo[j]fluoranthene, where it reached 10% due to
non-baseline peak separation. The recovery of the control sample

ranged from 79 to 99 %. Native and internal standards were used to

calibrate the GC–MS instrument, yielding an R-squared >0.99 for all
compounds. The detector proved to have a linear response over the
range 0–200 ng mL−1 and all sample extract concentrations were
within this range.

3.3.2. Indirect determination of trueness
In the absence of a reference material, trueness - defined as
the closeness of agreement between a test result and a reference
value - could not be directly determined. However, the content of
treated and untreated distilled aromatic extract (2.7%) and carbon
black (24.1%) were known for the NR/BR blends and each of these


20

O. Geiss et al. / J. Chromatogr. A 1566 (2018) 13–22

Fig. 3. Comparison clean-up efficiency PVC extract. Soft-PVC extracts (100 mg sample weigh-in): (1) Standard chromatogram (scan mode) on same retention time scale as
overlaid sample extract chromatograms in panel (2); (2) overlaid chromatograms of MIP-SPE column cleaned extract and silica-gel packed column purified extract; (3) non
purified sample extract.

Table 3
Recovery and RSD [%] values for low and high spiking levels.
Spiking Level

Compound

25 ng

Recovery ± RSD [%] n = 5

200 ng
Recovery ± RSD [%] n = 5

Benzo[a]anthracene
Chrysene
Benzo[b]fluoranthene
Benzo[k]fluoranthene
Benzo[j]fluoranthene
Benzo[e]pyrene
Benzo[a]pyrene
Dibenzo[a,h]anthracene

67.0 ± 5.0
84.0 ± 5.8
72.4 ± 2.5
84.6 ± 4.4
65.6 ± 5.2
72.2 ± 3.1
72.3 ± 3.8
54.7 ± 7.4

62.2 ± 8.0
75.4 ± 5.7
74.2 ± 3.0
79.5 ± 8.2
62.6 ± 3.8
76.1 ± 3.4
65.2 ± 6.1

62.2 ± 4.3

two ingredients were separately available and could be analysed
with regard to their PAH content. Assuming that carbon black and
the distilled aromatic extracts were the only sources of PAHs, the
theoretical final mass-concentration for the test-material could be
determined from
c(PAH, mgkg−1 ) = c(PAHCarbon Black , mgkg−1
)xContent in rubber [%] + c(PAHDAE/TDAE , mgkg−1
)xContent in rubber [%],

(1)

where c(PAHCarbonBlack ) is the individual PAH concentration in
the carbon black (N375) and c(PAHDAE/TDAE ) the individual PAH
concentration in the treated/untreated distilled aromatic extract,
respectively. These concentrations are multiplied with their
respective relative contents. Results are generally in good agreement (Table 4) with only benzo[a]pyrene showing a higher
discrepancy between the theoretical and the measured values
for both NR/BR blends. Unfortunately, these results cannot be

used for a fully quantitative trueness evaluation of the method
as an insufficient number of replicate measurements were done
for the mass-concentration determination of carbon black N375,
TDAE, DAE and NR/BR (N375/TDAE). Calculation of the uncertainties under these conditions was not possible. Moreover, the values
provided by the manufacturer for the content of carbon black and
distilled aromatic extract can only be considered semi-quantitative
estimates.
3.3.3. Limit of Quantification (LoQ)
The limit of quantification is the lowest concentration of an analyte that can be determined with acceptable precision and accuracy.

Different guidelines for the determination of the LoQ are available. By using the signal-to-noise method, the peak-to-peak noise
around the analyte retention time is measured, and then, the concentration of the analyte that yields a signal-to-noise ratio of 10:1
is estimated [15]. The absolute and relative limits of quantification
are reported in Table 5.
If required, the sensitivity could be improved by reducing the
volume of toluene used to reconstitute the dry-extract after nitrogen evaporation (Section 2.4.1).
4. Conclusions
This work proposes a simple, relatively fast, and cost effective
method for the determination of eight priority polycyclic aromatic
hydrocarbons in rubber and plastic materials. Compared with other
published methods/studies, the method proposed in this study
has two relevant advantages: In terms of extraction efficiency, the
Randall hot extraction process proved to be more efficient compared to ultrasound extraction, and represents an improvement
also over the classical Soxhlet extraction technique in that it con-


O. Geiss et al. / J. Chromatogr. A 1566 (2018) 13–22

21

Table 4
Theoretical and measured mass-concentrations of PAHs in NR/BR blends.
Mass Concentration [mg kg−1 ]

Benzo[a]anthracene
Chrysene
Benzo[b]fluoranthene
Benzo[k]fluoranthene
Benzo[j]fluoranthene
Benzo[e]pyrene

Benzo[a]pyrene
Dibenzo[a,h]anthracene
a

Carbon black N375

Treated Distillate
Aromatic Extract
(TDAE)

Distillate Aromatic
Extract (DAE)a

0.8
1.3
4.9
1.7
2.0
14.7
20.2
1.4

0.16
0.57
0.28
0.1
0.04
1.16
0.21
0

18.3
88.0
40.1
6.2
8.4
101.3
25.4
2.2

NR/BR, N375, TDAE

NR/BR, N375, DAE

Theoretical

Measured

Theoretical

Measured

0.2
0.3
1.2
0.4
0.5
3.6
4.9
0.3

0.2
0.3
0.9
0.3
0.3
3.2
3.7
0.4

0.7
2.7
2.3
0.6
0.7
6.3
5.5
0.4

0.5
2.1
2.1
0.5
0.5
6.6
4.7
0.5

Not available on the market. Used only for the purpose of this research project.

Table 5
Absolute and relative limits of quantification.
Compound

Absolute LoQ
[ng mL−1 ]

Relative LoQ (500 mg of
sample material) [mg kg−1 ]

Relative LoQ (100 mg of
sample material) [mg kg−1 ]

Relative LoQ (20 mg of
sample material) [mg kg−1 ]

Benzo[a]anthracene
Chrysene
Benzo[b]fluoranthene
Benzo[k]fluoranthene
Benzo[f]fluoranthene
Benzo[e]pyrene
Benzo[a]pyrene
Dibenzo[a,h]athrancene

0.8
0.8
3.6
0.9
0.8

0.8
1.1
0.6

0.002
0.002
0.007
0.002
0.002
0.002
0.002
0.001

0.008
0.008
0.036
0.009
0.008
0.008
0.011
0.006

0.04
0.04
0.18
0.05
0.04
0.04
0.06
0.03

siderably shortens the extraction time. Other benefits of the hot
extraction process include short process paths, low solvent requirements, and a process that is gentler on the extract (due to the
shorter extraction period). In terms of sample extract purification efficiency, the simple usage-protocol (fast), low amount of
required solvents, commercial availability (no need to pack the
column), and, above all, high selectivity resulting in lower baselines (i.e., better detection limits), highlight the superiority of the
MIP-based solid phase extraction procedure with regard to traditional, silica-based purification methodologies. The introduction of
Randall-hot extraction for sample extraction and MIP-based solid
phase extraction cartridges for selective clean-up represent a novel
and significant improvement compared with previously reported
methods in this field. The findings of this work may be beneficial to
the work conducted by the European Committee for Standardisation (CEN) which has recently (December 2017) been mandated by
the European Commission to develop a European standard for the
analytical determination of the individual concentrations of the 8
carcinogenic PAHs, listed in entry 50 of Annex XVII to the REACH
Regulation, in the plastic and rubber components of articles, in support to the enforcement of the provisions in paragraphs 5 and 6 of
the restriction.

Disclaimers
The information and views set out in this study are those of the
author(s) and do not necessarily reflect the official opinion of the
European Commission. The European Commission does not guarantee the accuracy of the data included in this study. Neither the
European Commission nor any person acting on the European Commission’s behalf may be held responsible for the use which may be
made of the information contained therein.

Appendix A. Supplementary data
Supplementary material related to this article can be found, in
the online version, at doi: />06.047.

References

[1] H.I. Abdel-Shafy, M.S.M. Mansour, A review on polycyclic aromatic
hydrocarbons: source, environmental impact, effect on human health and
remediation, Egypt. J. Petroleum 25 (2016) 107–123.
[2] N.-D. Dat, M.B. Chang, Review on characteristics of PAHs in atmosphere,
anthropogenic sources and control technologies, Sci. Total Environ. 609
(2017) 682–693.
[3] K.O. Goyak, M.H. Kung, M. Chen, K.K. Aldous, J.J. Freeman, Development of a
screening tool to prioritize testing for the carcinogenic hazard of residual
aromatic extracts and related petroleum streams, Toxicol. Lett. 264 (2016)
99–105.
[4] P. Lassen, L. Hoffmann, M. Thomasem, PAHs in toys and childcare products.
Survey of chemical substances in consumer products No. 114 2011, in: Danish
Ministry of the Environment, Environmental Protection Agency, 2012.
[5] BAuA, Annex XV Restriction Report Proposal for a Restriction of PAH,
Bundesanstalt Fuer Arbeitschutz Und Arbeitsmedizin (BAuA), 2010.
[6] M.K. Peterson, J.C. Lemay, S. Pacheco Shubin, R.L. Prueitt, Comprehensive
multipathway risk assessment of chemicals associated with recycled (ăcrumb)ă
rubber in synthetic turf fields, Environ. Res. 160 (2018) 256–268.
[7] Commission Implementing Decision of 1.12.2017 on a standardisation
request to the European Committee for StandardiZation and to the European
Committee for Electrotechnical StandardiZation as regards compliance with
maximum content criteria of Polycyclic Aromatic Hydrocarbons in rubber and
plastic components of articles placed on the market for supply to the general
public in support of Regulation (EC) No. 1907/2006 of the European
Parliament and of the Council, in: M/556 C(2017)7926 final, Brussels, 2017.
[8] Regulation (EC) No 1907/2006 of the European Parliament and of the Council
of 18 December 2006 concerning the Registration, Evaluation, Authorisation
and Restriction of Chemicals (REACH), establishing a European Chemicals
Agency, amending Directive 1999/45/EC and repealing Council Regulation
(EEC) No 793/93 and Commission Regulation (EC) No 1488/94 as well as

Council Directive 76/769/EEC and Commission Directives 91/155/EEC,
93/67/EEC, 93/105/EC and 2000/21/EC, in, Official Journal of the European
Union L396, 2006.


22

O. Geiss et al. / J. Chromatogr. A 1566 (2018) 13–22

[9] BAuA, GS-Spezifikation AfPS GS 2014:01 PAK. Testing and validation of
Polycyclic Aromatic Hydrocarbons (PAH) in the course of GS-Mark
Certification. Ausschuss fuer Produktsicherheit, Bundesanstalt fuer
Arbeitsschutz und Arbeitsmedizin, (2014).
[10] L. Gruber, M. Schlummer, D. Fiedler, M. Barwitz, S. Smilic, R. Franz,
Untersuchungen zur Migration von Polyzyklischen Aromatischen
Kohlenwasserstoffen (PAK) aus altreifenhaltigen Gummiprodukten. Report
PA/4453/17, in, Fraunhofer-Institut fuer Verfahrenstechnik und Verpackung,
Freising/Germany, 2017.
[11] International Organization for Standardization, ISO/TS 16190:2013 Footwear
– Critical substances potentially present in footwear and footwear
components – Test method to quantitatively determine polycyclic aromatic
hydrocarbons (PAH) in footwear materials, in, Geneva, 2013.
[12] European Committee for Standardization, EN 16143:2013 Petroleum
products - Determination of content of Benzo(a)pyrene (BaP) and selected
polycyclic aromatic hydrocarbons (PAH) in extender oils - Procedure using
double LC cleaning and GC/MS analysis, in: Brussels, 2013.
[13] International Organization for Standardization, ISO 21461:2009 Rubber –
Determination of the aromaticity of oil in vulcanized rubber compounds, in,
Geneva, 2009.
[14] S. Hamm, T. Frey, R. Weinand, G. Moninot, N. Petiniot, Investigations on the

extraction and migration behavior of polycyclic aromatic hydrocarbons
(PAHS) from cured rubber formulations containing carbon black as
reinforcing agent, Rubber Chem. Technol. 82 (2009) 214–228.
[15] A. Shrivastava, V. Gupta, Methods for the determination of limit of detection
and limit of quantitation of the analytical methods, Chronicles Young
Scientists 2 (2011) 21–25.
[16] A. Andreu, H.F. Stoeckli, R.H. Bradley, Specific and non-specific interactions on
non-porous carbon black surfaces, Carbon 45 (2007) 1854–1864.
˛
[17] M. Gagol,
A. Przyjazny, G. Boczkaj, Highly effective degradation of selected
groups of organic compounds by cavitation based AOPs under basic pH
conditions, Ultrason. Sonochem. 45 (2018) 257–266.
[18] P. Riesz, D. Berdahl, C.L. Christman, Free radical generation by ultrasound in
aqueous and nonaqueous solutions, Environ. Health Perspect. 64 (1985)
233–252.

[19] N.J. Thiex, S. Anderson, B. Gildemeister, Crude fat, diethyl ether extraction, in
feed, cereal grain, and forage (Randall/Soxtec/submersion method):
collaborative study, J AOAC Int. 86 (2003) 888–898.
[20] S. Anderson, Soxtec: Its Principles and Applications, AOCS Press, 2004.
´
[21] E. Gilgenast, G. Boczkaj, A. Przyjazny, M. Kaminski,
Sample preparation
procedure for the determination of polycyclic aromatic hydrocarbons in
petroleum vacuum residue and bitumen, Anal. Bioanal. Chem. 401 (2011)
1059–1069.
[22] M. Patrycja, F. André, B. Grzegorz, Method for the simultaneous
determination of monoaromatic and polycyclic aromatic hydrocarbons in
industrial effluents using dispersive liquid–liquid microextraction with gas

chromatography–mass spectrometry, J. Sep. Sci. (2018) 0.
[23] F. Ahmadi, H. Rezaei, R. Tahvilian, Computational-aided design of molecularly
imprinted polymer for selective extraction of methadone from plasma and
saliva and determination by gas chromatography, J. Chromatogr. A 1270
(2012) 9–19.
[24] Supelco, Extraction & Analysis of PAHs in Olive Oil Using SupelMIPTM SPE –
PAHs and GC-MS, Application Note 192 />content/dam/sigma-aldrich/docs/Supelco/Application Notes/1/t309192.pdf.
[25] M. Villar-Navarro, M.J. Martín-Valero, R.M. Fernández-Torres, M.
Callejón-Mochón, M.Á. Bello-López, Easy, fast and environmental friendly
method for the simultaneous extraction of the 16 EPA PAHs using magnetic
molecular imprinted polymers (mag-MIPs), J. Chromatogr. B 1044–1045
(2017) 63–69.
[26] L. Drabova, J. Pulkrabova, K. Kalachova, M. Tomaniova, V. Kocourek, J.
Hajslova, Rapid determination of polycyclic aromatic hydrocarbons (PAHs) in
tea using two-dimensional gas chromatography coupled with time of flight
mass spectrometry, Talanta 100 (2012) 207–216.
[27] H. Gong, S. Hajizadeh, L. Jiang, H. Ma, L. Ye, Dynamic assembly of molecularly
imprinted polymer nanoparticles, J. Colloid Interface Sci. 509 (2018) 463–471.