Food Chemistry 210 (2016) 613–622

Contents lists available at ScienceDirect

Food Chemistry
journal homepage: www.elsevier.com/locate/foodchem

New insights in the chemical composition of benzoin balsams

Pauline Burger a, Alexandre Casale b, Audrey Kerdudo a, Thomas Michel a, Rémi Laville a,1,
Francis Chagnaud c, Xavier Fernandez a,⇑
a
Institut de Chimie de Nice (ICN), UMR CNRS 7272, Parc Valrose, 06108 Nice Cedex 2, France
b
Faculté des Sciences, Master 2 FOQUAL, Parc Valrose, 06108 Nice Cedex 2, France
c
Agroforex Company, 292 Kaoliao Road – Sibounheuang, 01000 Vientiane, Laos

a r t i c l e i n f o a b s t r a c t

Article history: Benzoin balsam is an anthropic exudate obtained from the bark of several species of Styrax trees that is
Received 8 January 2016 mainly used as a perfume fixative as well as a flavouring agent. Benzoe tonkinensis Laos (also commer-
Received in revised form 29 April 2016 cialized under the denomination Siam benzoin balsam) displaying characteristic vanilla notes and
Accepted 2 May 2016
already being largely used to flavour all kinds of edible goods, was intended to be proposed by
Available online 3 May 2016
Agroforex Company to the Codex Committee on Food Additives for evaluation as a food additive. For this
purpose, the present paper reports the phytochemical characterisation of both the volatile and non-
Keywords:
volatile fractions of benzoin balsams and the quantitation of some of the major components by gas
Styracaceae
Siam benzoin (Benzoe tonkinensis Laos)
and liquid chromatography techniques. Four coniferyl and two morinol derivatives were characterised
Sumatra benzoin (Benzoe sumatranus for the first time in Benzoe tonkinensis Laos. Finally, two liquid chromatographic methods used to easily
Indonesia) discriminate Siam from Sumatra balsam (also known as Benzoe sumatranus Indonesia) were developed.
Chemical composition Ó 2016 Elsevier Ltd. All rights reserved.
Food additive

1. Introduction paralleloneurum Perkins on the other hand. Styrax tonkinensis was

tapped for centuries mainly in Laos and probably also in Thailand
Benzoin – not to be confused with ‘benjamin gum’ or Benzoïne to obtain the so called Siam benzoin balsam. S. benzoin and S. par-
– is a balsamic resin or balsam (AFNOR NF T 75-006, 1998) exuding alleloneurum are essentially native to Indonesia (Sumatra and Java)
from Styrax species (Styracaceae) that was employed since ancient and produce Sumatra benzoin balsam (Fernández, 2004; Supple-
time. First evidences confirmed that benzoin balsam was traded to mentary material 1). The generic term ‘benzoin’ either consists in
China about A.D. 800 to be used notably as a fixative for perfumes the exudate of one or the other of the two sources or in their mix-
(Arctander, 1960; Langenheim, 2003). In the Mediterranean basin, ture (WHO technical report 966, 2011).
it was valued for its therapeutic, pharmacological and odoriferous Benzoin balsam does not exude naturally from the trunk, but is
properties: it was notably used in the Middle-East as a substitute rather a pathological product resulting from tapping Styrax trees,
to incense in religious ceremonies, often in combination with e.g. cutting multiple superficial incisions through the cambium of
frankincense (Hovaneissian, Archier, Mathe, Culioli, & these trees (Langenheim, 2003; Pauletti, Teles, Silva, Araújo, &
Vieillescazes, 2008; Hovaneissian, Archier, Mathe, & Vieillescazes, Bolzani, 2006). As excessive tapping often induce the death of
2006; Modugno, Ribechini, & Colombini 2006). the tree, this traditional practice is nowadays well-defined: S. tonk-
The Styrax trees and shrubs originated from regions displaying inensis tree may be tapped once a year for about ten years once it is
warm and temperate climates mainly from Southeast Asia 7 years old, whereas S. benzoin and S. paralleloneurum may be
(Hovaneissian et al., 2006). Nowadays, the main benzoin producers tapped twice a year for about twenty years once they are 8–
are the deciduous Styrax tonkinensis (Pierre) Craib ex Hartwich. on 10 years old (Chagnaud, pers. comm.). After the yellowish exudate
one hand, and the evergreen Styrax benzoin Dryander and Styrax hardens upon exposition to air, it is collected by scraping the cut in
the trunk (79th JECFA – CTA, 2014; Langenheim, 2003). Once col-
⇑ Corresponding author. Tel.: +33 4 92 07 64 69; fax: +33 4 92 07 61 89. lected, benzoin balsam is cleaned, sorted into four different grades
E-mail addresses: Pauline.BURGER@unice.fr (P. Burger), Alexandre.CASALE@ according to the size and the homogeneity of the pieces, and then
unice.fr (A. Casale), audreykerdudo@gmail.com (A. Kerdudo), Thomas.MICHEL@unice.fr matured. The production remains nowadays entirely manual, from
(T. Michel), Remi.Laville@cosmo-ingredients.com (R. Laville), fchagnaud@agroforex-
company.com (F. Chagnaud), Xavier.FERNANDEZ@unice.fr (X. Fernandez).
tapping to packaging (79th JECFA – CTA, 2014).
1
Present address: Cosmo International Ingredients, 855 avenue du Dr Maurice These three Styrax species are the most significant producers of
Daunat, 06250 Mougins, France. balsam in terms of tonnage per year (Kashio & Johnson, 2001).

http://dx.doi.org/10.1016/j.foodchem.2016.05.015
0308-8146/Ó 2016 Elsevier Ltd. All rights reserved.
614 P. Burger et al. / Food Chemistry 210 (2016) 613–622

Apart from them, several other minor species that yield benzoin to also means that supplies of natural vanilla must be carefully and
lesser extent and quality are also common in these regions (Burkill, sustainably managed over the long-term. The necessity of alterna-
1966). Another minor source of benzoin is Styrax officinalis from tive vanilla sources has long been explored: vanillin was first syn-
the Mediterranean region that was already known by the Greeks thesized in 1874 from coniferin (Tiemann & Haarmann, 1874), and
Herodotus (5th c. A.D.) and Theophrastus (4th-3rd c. A.D.) and rec- ‘‘vanilla extract” was invented as a cost effective substitute for the
ommended by Hippocrates (5th-4th c. A.D.) as a remedy real vanilla. Furthermore, over the last decades, food and drink
(Langenheim, 2003). It is worthwhile to note that the term Styrax manufacturers were seeking for cheaper alternatives to meet the
deriving from ‘astirax’, an Arabic locution meaning ‘sweet- pressing consumer demand for more natural products: benzoin
smelling exudation’ (Langenheim, 2003), has generated a nomen- balsam being a natural source of vanillin was a perfect candidate.
clatural confusion quite frequent in the resin terminology: gums Siam benzoin balsam was intended to be proposed by Agroforex
from Liquidambar spp. (Hamamelidaceae), so-called storax, are Company to the Codex Committee on Food Additives (CCFA, Codex
often marketed as benzoin but are not to be considered as such Alimentarius) for evaluation as an authorized food additive. Hence,
from botanical and chemical points of view (Hovaneissian et al., the necessity to meet the JECFA (Joint FAO/WHO Expert Committee
2006; Langenheim, 1996). on Food Additives) regulations implied to get a better knowledge of
In the present paper, we focus on Siam (also known as Benzoe the composition of benzoin balsam. In fact, despite the number of
tonkinensis Laos, BTL; Eur. Ph. 2016) and Sumatra (also known as studies reporting chemical investigations of benzoin balsam, ana-
Benzoe sumatranus Indonesia, BSI; Eur. Ph. 2008) benzoin balsams. lytical techniques frequently used do not take into account the
Their chemical compositions have been widely studied over the whole material and issues with botanical and geographical origins
last decades (Supplementary material 2). A hundred years ago, are often reported.
Reinitzer found that the major constituent of Siam balsam was Agroforex Company already submitted a file summarizing the
coniferyl benzoate 1 (75–80%) (Reinitzer, 1914). Later, Schroeder’s requested information about characterisation and toxicological
work shows that Siam benzoin balsam was composed of coniferyl data on Benzoe tonkinensis Laos (September 2010), an Activity
benzoate 1 (65–75%), p-coumaryl benzoate 2 (10–15%), cinnamyl Report (May 2011), Complementary Data (December 2012) and a
cinnamate 3 (0.5–6%), benzoic acid 10 (12%), and siaresinolic acid Report (January 2014) to JECFA (79th JECFA – CTA, 2014). Finally,
17 (6%) (Schroeder, 1968). On the other hand, cinnamic acid 11 BTL was concluded not to be of safety concern regarding the ‘no
and its esters, mainly cinnamyl cinnamate 3 and coniferyl cinna- observed adverse effect level’ (NOAEL) criteria (WHO technical
mate 16, and pinoresinol have been characterised as components report 990, 2014). However, the previous evaluations of the ques-
of Sumatra benzoin balsam (Bruneton, 2009; Pastorova, de tion of the BTL registration as a food flavour evidenced some miss-
Koster, & Boon, 1997). Vanillin, prevalent in Siam balsam, is never- ing information and particularly additional data about its overall
theless present in Sumatra balsam and gives rise to their familiar chemical composition, including quantitation of benzoin balsam
vanilla odour (Kashio & Johnson, 2001). constituents, were requested (FAO JECFA monograph, 2011; WHO
Our research group recently studied more particularly the vola- technical report 983, 2013). The development of relevant analytical
tile fraction of both benzoin types. For this, volatile extracts methodology to distinguish between benzoin of both botanical ori-
obtained by hydrodistillation of 50 kg of each benzoin gum were gins was also claimed.
analysed using High Temperature-Gas Chromatography (HT-GC), For this purpose, analytical and semi-preparative HPLC-DAD-
Gas Chromatography coupled to Mass Spectrometry (GC/MS) and ELSD (High-Performance Liquid Chromatography coupled to Diode
electronic nose using different sample pre-treatment strategies Array and Evaporative Light Scattering Detectors), GC-FID
(Castel et al., 2006; Fernandez, Castel, Lizzani-Cuvelier, (Gas Chromatography with a Flame Ionization Detector), GC/MS,
Delbecque, & Puech Venzal, 2006; Fernandez, Lizzani-Cuvelier, NMR (Nuclear Magnetic Resonance), and UPLC-QToF (Ultra-
Loiseau, Perichet, & Delbecque, 2003). The GC/MS study of these Performance Liquid Chromatography–Quadrupole Time of Flight
extracts led respectively to the identification of 20 and 29 com- mass spectrometry) were used to characterise these balsamic exu-
pounds. The major compound identified in benzoin balsams of dates via both the identification of new aromatic compounds of the
both botanical origins was benzyl benzoate 12 (76.1–80.1%). Yet, phenylpropanoids family, and the quantitation of some of their
while Siam benzoin balsam was characterised by benzoic acid 10 main constituents.
(12.5%), methyl benzoate (1.5%) and allyl benzoate (1.5%), Sumatra
balsam was defined by its content of cinnamic acid 11 (3.5%) and
2. Materials and methods
benzyl cinnamate 13 (3.3%). This work also led to the detection
and identification of an unusual acenaphthylene-type sesquiter-
The complete experimental procedure employed for this study
penic hydrocarbon in these volatile extracts (Filippi et al., 2009).
is summarized in Fig. 1.
Besides aromatic compounds, Siam balsam also contains several
triterpenoids (siaresinolic and sumaresinolic acids, etc.).
Despite their traditional uses, benzoin balsams are well known 2.1. Material and sample preparation
for their characteristic scents. Siam benzoin balsam has a pleasant
sweet-balsamic odour with distinct notes of vanilla, amber, choco- Benzoin samples of both botanical origins were provided by
late, caramel and nuts while Sumatra benzoin balsam presents a Agroforex Company in 2011: Siam benzoin balsam (BTL Grade 3)
strong styrax-like odour (Castel et al., 2006). Hence, benzoin bal- was collected in Laos and Sumatra balsam (BSI Grade B) in Indone-
sam of both botanical origins is widely used in the fragrance indus- sia. Siam and Sumatra samples are respectively referred to as
try as perfume fixatives and preservatives with a slight prevalence BSiam and BSum in the current article (Supplementary material 1).
for the Siam product (Langenheim, 2003). All the samples were solubilized in ethanol, filtered and evapo-
Benzoin balsam is also largely priced in the flavour industry: the rated to obtain ethanol extracts at a final concentration of 30 mg/
Siam one was notably widely used to flavour all kinds of beverages, mL (extraction yield: 98%).
confectioneries, etc. to add characteristic vanilla notes (Burdock,
1995; Burdock, 1997; Leung & Foster, 1996; Moyler, 1998). In fact, 2.2. Chemicals
the warm, soft and sweet notes of natural vanilla are among the
world’s favourite scents and vanillin is a compound of high interest Acetonitrile (ACN, ChromasolvÒ for HPLC), water (H2O, Chroma-
for the flavour industry. However, its popularity among consumers solvÒ for HPLC), dichloromethane (DCM, ChromasolvÒ for HPLC),
 P. Burger et al. / Food Chemistry 210 (2016) 613–622 615

Fig. 1. Experimental procedure employed for the identification and quantitation of the main volatile and non-volatile constituents of benzoin balsam.

methanol (MeOH, ChromasolvÒ for HPLC), diethyl ether (Et2O, Pur- gas was helium in constant flow mode at 1 mL/min. Split ratio
iss), absolute ethanol (Puriss), formic acid (Puriss), petroleum ether was 1/10 with an injection volume of 1 lL and the injector and
(PE, Laboratory Reagent), vanillin (purity 99%), benzyl benzoate transfer line were set at temperatures of 250 °C and 270 °C respec-
(purity P99%), cinnamic acid (purity P99%), benzoic acid (purity tively. Acquisition was performed in scan mode (35–400 a.m.u./s)
P99.5%), deuterated chloroform (CDCl3) for NMR experiments, n- and mass spectra were generated at 70 eV.
hexane, acetic acid, dioxane, p-methoxybenzaldehyde and sulfuric GC-FID analyses were carried out under similar conditions
acid were all supplied by Sigma-Aldrich (Saint-Quentin-Fallavier, using the same instrument equipped with the same column. The
France). carrier gas was hydrogen and samples were injected in split mode
(split ratio 1/10). Following a 5 min isothermal hold at 80 °C the
2.3. Analysis of the benzoin volatile fraction oven was temperature programmed to 320 °C at a rate of 8 °C/min-
ute with the final temperature held for 40 min.
2.3.1. Gas chromatography Data acquisition and the modules mentioned previously were
GC/MS analyses were carried out using an Agilent piloted by a computer equipped with the Agilent Chemstation soft-
6890N/5973N system equipped with an Rxi-5HT column ware. Peak assignations were based on computer matching of mass
(30 m
0.25 mm; film thickness, 0.25 lm; Restek). The carrier spectra against a commercial library (Wiley, MassFinder 2.1
616 P. Burger et al. / Food Chemistry 210 (2016) 613–622

Library) and a laboratory-made mass spectra library built up from The same instrument operated in the same conditions but
pure substances and MS literature data. Some structures assigna- equipped with a Luna C18 column (10
250 mm, 5 lm; Phenom-
tions were confirmed by co-injection with the authentic standards enex; injection volume: 80 lL; flow rate: 2.5 mL/min) was used for
available at the laboratory (vanillin 9, benzoic acid 10, cinnamic semi-preparative HPLC, in order to isolate and purify compounds
acid 11 and benzyl benzoate 12; Supplementary material 2). of interest. The elution gradient was adapted to each extract/frac-
tion injected on the column (see Section 2.4.4).
2.3.2. Quantitation of major volatile compounds
2.4.3. Quantitation of major non-volatile compounds
The quantitation of the major volatile compounds was carried
The quantitation of coniferyl benzoate 1, p-coumaryl benzoate 2
out by external calibration in HTGC-FID using calibration curves
and cinnamyl cinnamate 3 was carried out by external calibration
constructed with 5 concentrations ranging respectively from
on the same HPLC instrument equipped with the same column
0.012 mg/mL to 2 mg/mL for vanillin 9, from 0.25 mg/mL to
using the following solvent gradients: A) BSiam: the initial system
20 mg/mL for benzoic and cinnamic acids 10 and 11, and from
consisted in H2O/ACN 65:35 (v/v) and reached 100% ACN within
0.01 mg/mL to 1 mg/mL for benzyl benzoate 12. Commercial mole-
20 min; 2) BSum: the initial system consisted in H2O/ACN 50:50
cules were used as standards and each solution was injected in
(v/v) and reached H2O/ACN 20:80 (v/v) within 20 min. The elution
triplicate. The resulting calibration relations are described as
solvents were acidified with 0.1% formic acid.
follows:
The quantitation was performed using calibration curves con-
%i ¼ ða
Ai Þ þ bðR2 Þ structed with 5 concentrations ranging respectively from 0.5 mg/
mL to 8 mg/mL for coniferyl benzoate 1, from 0.7 mg/mL to
with %i representing the percentage of the compound ‘i’ in the 6 mg/mL for p-coumaryl cinnamate 2 and from 0.4 mg/mL to
ethanolic extract, and Ai, the area of the corresponding HTGC-FID 4 mg/mL for cinnamyl cinnamate 3. The previously isolated com-
peak. The R2 coefficients were all acceptable above 0.995 (for a pounds were used as standards and each solution was injected in
and b values, see Supplementary material 3). triplicate.
The calibration curves were characterised by an equation repre-
senting the logarithm of the concentration of each compound in
2.4. Analysis of the benzoin non-volatile fraction
BSiam/BSum samples as a function of the logarithm of their respec-
tive peak area in ELSD for an injection of 30 lL:
2.4.1. High performance thin-layer chromatography
High performance thin-layer chromatography (HPTLC) was per- logðCiÞ ¼ a:logðAiÞ þ b;
formed on 10 cm
20 cm HPTLC silica gel 60F254 pre-coated plates
with Ci representing the concentration of the compound ‘i’ in the
(Merck). Standards and samples were applied as 7.4 mm bands,
sample, and Ai, the area of the corresponding ELSD signal. The R2
20 mm from the left edge and 8 mm apart by means of automated
coefficients were all acceptable above 0.995 (for a and b values,
ATS4 sampler (Camag). Two microliters of each standard and sam-
see Supplementary material 3).
ples were applied on the plate. The mobile phase consisted in the
The proportion of each compound in samples was calculated
supernatant of a solution of n-hexane/MeOH/acetic acid 40/10/5
with the following formula:
to get a general triterpenic survey in benzoin balsams. The separa-
tion on the plate was performed in an automatic ADC2 developing Ci
BSiamorBSumð%Þi ¼ :
chamber (Camag; developing distance: 7 cm) with the tank previ- C BSiamorBSum
ously saturated for 20 min with the developing solvent. After
developing and drying (5 min), plates were dipped for 1 s in the
2.4.4. Benzoin fractionation and compounds isolation
sulfuric anisaldehyde detection reagent using Camag immersion
Siam benzoin 10 g of BSiam ethanolic extract (Supplementary
device, dried in a stream of warm air and heated at 105 °C with a
material 1) were fractionated by column chromatography over sil-
TLC plate heater (Camag) for 5 min.
ica gel (150 g) and eluted with a solvent system of increasing
Sulfuric anisaldehyde was prepared as described by Reich and
polarity (PE, Et2O, MeOH; 500 mL of each solvent; Supplementary
Schibli (2006): 10 mL of sulfuric acid were carefully added to an
material 4). The resulting fractions were analysed using the stan-
ice-cold solution of 170 mL MeOH and 20 mL acetic acid, followed
dard HPLC method.
by the addition of 1 mL of p-methoxybenzaldehyde. Documenta-
The ether fraction (i.e. fraction 5; Supplementary material 4)
tion of chromatographic plates (visualization and images capture)
was then purified by semi-preparative HPLC. The elution was per-
was performed under white light and at 366 nm by the TLC Visual-
formed using the following solvent gradient (solvents acidified
izer (Camag) operated with winCATS Version 1.4.6 software, to
with 0.1% formic acid): the initial system consisted in H2O/ACN
detect mainly terpenoids.
40:60 (v/v) in isocratic mode for 15 min, then reached H2O/ACN
30:70 (v/v) within 5 min, remained isocratic for another 10 min
2.4.2. High performance liquid chromatography and finally reached 100% ACN in 10 min. A total of 9 major peaks
HPLC analyses were performed on a HPLC Agilent 1200 series were observed and 5–10 mg of each of the corresponding com-
equipped with a double detection system consisting in a DAD pounds were isolated and purified. Along with 5 well-known ben-
and an ELSD. Separations were performed on a Luna C18 column zoin balsam components, coniferyl benzoate derivatives 4, 5, 6 and
(4.6
150 mm, 5 lm; Phenomenex). A standard HPLC method 7, and coeluting morinol derivatives 8 and 80 (Tables 1 and 2) were
was developed (referred to as ‘standard HPLC method’ in the fol- characterised for the first time in benzoin.
lowing discussion). The elution gradient consisting initially in Sumatra benzoin 10 g of BSum ethanolic extract were fraction-
H2O/ACN 95:5 (v/v) reached 100% ACN within 35 min and ated using the same procedure. Each fraction was analysed using
remained isocratic for 5 min. The elution solvents were all acidified the standard HPLC method.
with 0.1% formic acid. Injection volume was 30 lL and compounds Purification of the fraction PE/Et2O 7:3 (i.e. fraction 2; Supple-
were eluted at 1 mL/min at room temperature. The DAD was set at mentary material 4) was performed by semi-preparative HPLC
254, 280 and 366 nm, and ELSD conditions were set as follows: using an elution gradient consisting initially in H2O/ACN 30:70
nebulizer gas pressure 3.5 bars, evaporative tube temperature (v/v) and reaching H2O/ACN 25:75 (v/v) within 35 min (the elution
40 °C and gain 3. solvents were all acidified with 0.1% formic acid). Four major peaks
 P. Burger et al. / Food Chemistry 210 (2016) 613–622 617

Table 1
Chemical structures and 1H and 13
C NMR chemical shifts of the newly identified coniferyl benzoate derivatives 4–7 (siamyl mono- and di-benzoates).

Siamyl benzoate 4 Siamyl-1,3-dibenzoate 5 Siamyl-1,2-dibenzoate 6 Siamyl-2,3-dibenzoate 7

OMe 5'' 5'
OMe
4' 4'' 4'
4' OH
OH 5' 3'
OH 2 3
5' 3' 2 3 4
OH 4 3'' 2'' 3' 2'
2'
O 8 OMe OMe
2' 8 1' 5
1' O 5 1 4'
1 7 OH
7 9 6
5' 3' 2
OH
9 6 O 1'' O 3 1' 2 3
O O O 4 O O 4
1''
O OH 2' 8
O 8
1' 5 HO 5
1 1
2'' 7 7
9 6 9 6
3''
O OH O O
4''
5''

13 1 13 1 13 1 13 1
C H C H C H C H
1 132.01 128.6 131.27 128.53
2 114.6 6.93 d (1.5) 109.92 6.95 d (1.9) 109.03 6.98 d (1.6) 110.23 7.02 d (1.8)
3 146.9 146.79 146.99 146.78
4 145.9 146.25 146.1 146.24
5 120 6.88 dd (8.1, 1.6) 114.88 6.91 d (8.2) 114.83 6.93 d (8.0) 114.63 6.88 d (7.9)
6 109.12 6.9 d (8.2) 120.26 7.01 dd (8.1, 1.9) 120.15 6.92 m 120.49 7.04 dd (7.9, 2.0)
7 74.8 4.69 d (7.0) 77.36 6.1 d (6.7) 73.64 5.04 d (6.8) 76.42 6.04 d (6.0)
8 74.6 4.06 (br m) 72.93 4.4 m 75.73 5.66 m 72.41 4.46 m
9 65.9 4.38 dd (11.8, 3.7) 65.69 4.44 dd (11.2, 3.6) 63.47 4.58 m 65.61 4.61 m
4.23 dd (11.8, 5.8) 4.26 dd (11.4, 5.0) 4.3 dd (12.0, 6.0) 4.48 m
10 167 166.67 166.33 166.92
20 129.78 129.7
30 129.83 8.02 dt (7.1, 1.5) 129.83 8.03 d (7.8) 129.98/129.8 8.07/7.98 d (7.9) 129.89 8.02 d (7.9)
40 128.62 7.45 tt (7.7, 1.5) 128.64 7.45 m 128.63 7.44 m 128.63 7.44 m
50 133.47 7.58 tt (7.3, 1.3) 133.46 7.58 m 133.45 7.57 m 133.45 7.57 m
100 166.07 166.33 165.63
200 130.0
300 129.89 8.1 d (8.0) 129.98/129.8 8.07/7.98 d (7.9) 129.98 8.07 d (7.9)
400 128.64 7.45 m 128.63 7.44 m 128.63 7.44 m
500 133.46 7.58 m 133.45 7.57 m 133.45 7.57 m
OCH3-3 56.1 3.87 s 56.05 3.83 s 56.02 3.81 s 56.13 3.89 s
OH-5 5.6 5.6
OH-7 2.78
OH-8 2.92

Notes: Arrows: HMBC correlations H?C; Thick bonds: COSY correlations.

were observed and the corresponding compounds already identi- tion mode (ESI-) to obtain the molecular masses of the unknown
fied in benzoin were isolated (5–10 mg of each of the compounds). compounds [MH+].
The semi-preparative HPLC of the fraction PE/Et2O 1:1 (i.e. frac- Infusion was performed in resolution mode. The capillary volt-
tion 3; Supplementary material 4) was performed using an elution age was set at 2 kV and the cone voltage at 45 V. Source and des-
gradient consisting initially in H2O/ACN 40:60 (v/v) and reaching olvation temperatures were respectively set at 100 °C and 200 °C.
H2O/ACN 10:90 (v/v) within 35 min (solvents acidified with 0.1% The nebulization and the cone gas were adjusted to 600 L/Hr and
formic acid) and leading to the detection of four major peaks and 10 L/Hr.
the subsequent isolation of four compounds (5–10 mg of each)
well-known to enter the composition of benzoin. 3. Results and discussion

2.5. Structural elucidation of isolated compounds The intended use of Siam benzoin balsam being as a food
flavouring agent, it has to be registered as a food additive in the
2.5.1. 1H and 13C NMR analysis Codex Alimentarius which aims at protecting consumers health
1
H and 13C NMR spectra were recorded in CDCl3 at 25 °C on and ensuring fair practices in the food trade; it has hence to meet
either 200 MHz or 500 MHz Bruker Avance NMR spectrometers. the JECFA regulations. For this purpose, some additional informa-
Chemical shifts were expressed in Hz relative to CDCl3 or d6- tion about benzoin balsam overall chemical composition appeared
DMSO. 1H and 13C NMR data for compounds already known to necessary. The development of a relevant analytical methodology
enter the composition of benzoin are presented in the Supplemen- to distinguish between benzoin of both botanical origins was also
tary material 5 section. Structural elucidations of unknown com- claimed.
pounds 4–8 and 80 are displayed respectively in Tables 1 and 2.
3.1. Chemical composition of benzoin balsams
2.5.2. Molecular mass determination
Elemental analyses were performed by UPLC-high resolution As already stated in the introduction, the analysis of the volatile
electrospray ionization MS (HRESIMS) on an ACQUITY UPLCÒ/ fraction of both benzoin types has previously been dealt with
XevoTM G2 QTof (Waters) operated in negative electrospray ioniza- HTGC/MS, GC/MS and electronic nose using different sample
618 P. Burger et al. / Food Chemistry 210 (2016) 613–622

Table 2
Chemical structures and 1H and 13
C NMR chemical shifts of the morinol derivatives 8 and 80 (Siam morinol and siam hydroxymorinol) newly identified in benzoin.

Siam morinol 8 Siam hydroxymorinol 80

OMe OMe
HO O HO O
2 2
1'' 1''
1 1
HO O HO O
7 7
1' 8' 1' 8'
O O
MeO 8 MeO 8
2' 7' 9' 9 1''' 2' 7' 9' 9 1'''

O O
+ -OH*
13 1 13 1
C H C H
1 130–131 130–131
2 109.9 6.92 109.6 6.89 d (1.8)
3 146.65 146.65
4 145.7 145.7
5 114.69 6.91 d (8.0) 114.69 6.89 d (8.0)
6 120.17 6.99 dd (8.0, 1.6) 119.87 6.96 dd (8.2, 1.8)
7 76.52 6.12 d (4.4) 77.16 6.13 d (4.5)
8 44.02 2.67 m 44.21 2.63 m
9 64.05 4.56 dd (11.2, 5.0) 64.37 4.38 dd (11.4, 4.9)
4.52 dd (11.2, 5.4) 4.21 dd (11.3, 4.7)
10 130–131 130–131
20 108.16 6.81 s 108.1 6.78 s
30 146.65 146.65
40 145.3 145.3
50 114.63 6.89 d (8.1)
60 119.9 6.78 m 119.87 6.78 m
70 132.58 6.27 d (15.8) 132.49 6.35 d (15.8)
80 124.51 6.01 ddd (15.8, 7.4, 6.8) 124.84 6.08 m
90 32.01 2.42 m 31.81 2.65 m
2.31 ddd (14.2, 8.0, 7.4) 2.54 m
100 165.8 165.66
200 130–131 130–131
300 129.8 8.08 d (8.2) 129.76 8.05 d (8.2)
400 128.6 7.4 m 128.6 7.4 m
500 133.3 7.55 m 133.3 7.55 m
1000 166.62 166.45
2000 130–131 130–131
3000 129.71 8.0 d (8.2) 129.71 8.0 d (8.2)
4000 128.6 7.4 m 128.6 7.4 m
5000 133.3 7.55 m 133.3 7.55 m
OCH3-Ar 55.98 (56.1) 3.85 (3.78) 55.98 (56.1) 3.85 (3.78)
OH-Ar 5.59–5.56 5.59–5.56
⁄
Notes: Position of the additional -OH group not defined; Arrows: HMBC correlations H?C; Thick bonds: COSY correlations.

Norm. 2 1

2500
3
2000

1500

1000

500
Siam
BSiam

0 Sumatra
BSum
0 10 20 30 40 50 min

Fig. 2. HPLC chromatograms obtained at 280 nm of BSiam and BSum, presenting the major compounds identified as coniferyl benzoate 1, p-coumaryl cinnamate 2 and
cinnamyl cinnamate 3. Column Luna C18 (4.6
150 mm, 5 lm). Flow rate of 1 mL/min. Elution gradient using H2O and ACN both acidified at 0.1% formic acid.

pre-treatment strategies (Castel et al., 2006; Fernandez et al., 2003; The non-volatile fraction of benzoin extracts, seldom analysed
Fernandez et al., 2006). These studies led to the conclusion that, until nowadays, has been rightfully suggested to hold further inter-
apart from benzyl benzoate 12 characterised as the major esting information to help the distinction of benzoin balsam of
compound in both types, numerous differences exist between both both botanical origins. Liquid chromatography appears to be the
balsams and their distinction based on the acute analysis of their best technique to unravel such fraction, e.g. to gain further access
volatile fraction appears to be quite evident. to its chemical composition.
 P. Burger et al. / Food Chemistry 210 (2016) 613–622 619

A first glimpse of the chemical composition of extracts of both never identified previously in natural products of any kind.
types of benzoin has been obtained by analytical liquid chromatog- Siamyl-1,3-dibenzoate is proposed as its common name.
raphy. The resulting HPLC chromatograms presented in Fig. 2 Compound 6 and 7 were isolated as a mixture. Their molecular
easily enable the discrimination of both benzoin balsams: BSiam formulas C24H22O7 were both established via HRESIMS analysis in
is characterised by one major peak eluting after 34 min corre- QTof (m/z 445.1255 calculated for C24H22NaO7, 445.1263, D
sponding to coniferyl benzoate 1 while the BSum profile displays 1.8 ppm). A thorough analysis of 1D and 2D NMR experiments
two major molecules corresponding to p-coumaryl cinnamate 2 revealed that compounds 6 and 7 were isomers of compound 5
(eluting after 25 min) and cinnamyl cinnamate 3 (eluting after regarding the benzoate positions: all 1H and 13C NMR signals were
37 min). doubled. The positions of benzoate moieties were assigned thanks
After fractionation of both benzoin ethanolic extracts over silica to the chemical shift of H-7 and H-8 and to HMBC correlations
gel column, semi-preparative liquid chromatography was between the quaternary carboxylic carbons C-10 or C-100 and the
employed to purify some of the non-volatile compounds observed H-7, H-8 or H-9 (Table 1). These compounds were tentatively
in both types of benzoin. Structural elucidation of the isolated named 3-hydroxy-3-(4-hydroxy-3-methoxyphenyl)propyl-1,2-dib
compounds was subsequently achieved by NMR and mass enzoate and 1-hydroxymethyl-2-(4-hydroxy-3-methoxyphenyl)et
spectrometry. hyl-1,2-dibenzoate respectively following the IUPAC rules. They
While all the compounds isolated from BSum fractions 2 and 3 were, to our knowledge, never identified previously in natural
were already well-known benzoin balsams components (i.e. cin- products of any kind. Siamyl-1,2-dibenzoate and siamyl-2,3-
namyl cinnamate 3, benzyl cinnamate 13, cinnamyl benzoate 14 dibenzoate are respectively proposed as their common names.
and sumaresinolic acid 18 in fraction 2; coniferyl benzoate 1, p- Compounds 8 and 80 were isolated as a mixture in BSiam sam-
coumaryl benzoate 2, p-coumaryl cinnamate 15 and coniferyl cin- ples that have suffered some alteration processes (maturation)
namate 16 in fraction 3; Supplementary material 5), the fractiona- over time after collection (they were not detected in the more
tion of BSiam fraction 5 is more striking. In fact, some unknown recently collected samples). Their respective molecular formulas
compounds (4, 5, 6, 7, 8 and 80 ; Tables 1 and 2) were isolated for C34H32O8 and C34H32O9 were established by HRESIMS in QTof (m/
the first time in Siam benzoin balsams, along with the following z 591.1995 calculated for C34H32NaO8, 591.1992, D 0.5 ppm; m/
well-known coniferyl benzoate 1, p-coumaryl benzoate 2, z 607.1932 calculated for C34H32NaO9). Signals and correlations
siaresinolic acid 17, sumaresinolic acid 18 and 3-oxo-siaresinolic provided by 1D and 2D NMR spectra are presented in Table 2. 1H
acid 19 (Supplementary material 5). and 13C NMR spectra showed characteristic signals of the phenyl-
Compound 4 was isolated as a yellow oil and its molecular for- propanoid family with 2 benzyl and 2 o-methoxy-phenolic moi-
mula C17H18O6 was established by HRESIMS in QTof (m/z 341.1005 eties for both isomers (for 1H NMR spectrum, see Supplementary
calculated for C17H18NaO6, 341.1001, D 1.2 ppm). 1H and 13C NMR material 6). The number of o-methoxy-phenolic moieties leds us
spectra presented interesting similarities with coniferyl benzoate. to the conclusion that both compounds consisted in dimeric coni-
Indeed the presence of both benzoic and o-methoxy-phenolic moi- feryl benzoate derivatives. Indeed, 1H–1H COSY NMR experiments
eties were assigned to the proton at dH 8.02 (2H, dt J = 7.1, 1.5 Hz, enabled the assignment of the non-aromatic core with the consec-
H-30 ), 7.45 (2H, tt J = 7.7, 1.5 Hz, H-40 ) and 7.58 (1H, tt J = 7.3, utive correlations between H-9, H-8, H-90 , H-80 and H-70 on one
1.3 Hz, H-50 ) ppm and 6.93 (1H, d J = 1.5 Hz, H-2), 6.88 (1H, dd hand and H-8 and H-7 on the other hand: such structure is similar
J = 8.1, 1.6, H-5) and 6.9 (1H, d J = 8.2 Hz, H-6) ppm respectively. to the one of the 8,90 -neolignan skeleton. A similar skeleton was
Benzoate moiety was confirmed by the quaternary carbon at dC already described in the morinols C-L for instance. HMBC correla-
167 ppm and the quaternary carbons at dC 146.9 and 145.9 ppm tions presented in Fig. 4 confirmed the respective positions of the
are in accordance with the 3-hydroxylation and 2-methoxylation aromatic moieties according to the central skeleton. Such o-
of the second aromatic ring as expected in coniferyl derivatives. methoxy-phenolic and benzoated 8,90 -neolignans were described
The proton signals corresponding to the a,b unsaturation of the here for the first time in benzoin. The coupling constant H-70 /H-
coniferyl moiety were displaced by two methines at dH 4.69 80 of 15.8 Hz confirmed the expected trans configuration of the C-
(1H, d J = 7.0 Hz, H-7) and 4.06 (1H, br m, H-8) ppm which were 70 /C-80 double bond. The coupling constants between H7 and H8
consistent with the diol derivative presented in Table 1. The (7.42 and 7.21) correspond both to the threo configuration, as
COSY key correlations between H-7, H-8 and H-9 and the HMBC determined previously by Su, Takaishi, and Kusumi (1999), in their
key correlations between H-9 and C-10 , for the benzoate connec- study of twelve novel sesquineolignans and neolignans, respec-
tion, and H-7 and C-1, C-5 and C-6, for the o-methoxy-phenol tively named morinols A-L: coupling constants of 7 Hz correspond
moiety connection further confirmed the proposed structure. No to the threo configuration whereas those of 4 Hz correspond to the
stereochemical clues were identified so far. This compound was erythro configuration. Compound 8 was tentatively named 1,5-
hence tentatively named 2,3-dihydroxy-3-(4-hydroxy-3-methoxy di(40 -hydroxy-30 -methoxyphenyl)-2-[(benzyloxy)methyl]-pent-4-en-
phenyl)propyl benzoate following the IUPAC rules: such molecule 1-yl benzoate following the IUPAC rules, while no name could be
was, to our knowledge, never identified previously in natural attributed to compound 80 as the exact position of the additional
products of any kind. Siamyl benzoate is proposed as its common -OH group could not be determined. Siam morinol and Siam
name. hydroxymorinol are respectively proposed here as common names
Compound 5 was isolated as a white powder and its molecular for these compounds that were, to our knowledge, never identified
formula C24H22O7 was established by HRESIMS in QTof (m/z previously in any natural product.
445.1275 calculated for C24H22NaO7, 445.1263, D 2.7 ppm). 1H
and 13C NMR spectra revealed similarities with those of compound 3.2. Quantitation of the major constituents of benzoin extracts
4. Nevertheless compound 5 possessed two benzoate moieties
pointed out by the two doublets at dH 8.03 (2H, d J = 7.8 Hz, H-30 ) Quantitation of the major volatile compounds was performed
and 8.1 (2H, d J = 8.0 Hz, H-300 ) ppm. Furthermore H-7 was shifted via the external standardization method in GC/FID using commer-
at dH 6.1 ppm which was in accordance with the benzoate moiety cial standards. The formulas resulting from the calibration curves
at position C-7 (Table 1). HMBC correlation between H-7 and C-100 constructed for vanillin 9, benzoic acid 10, cinnamic acid 11 and
further confirmed this observation. This compound hence tenta- benzyl benzoate 12 enabled the determination of these com-
tively named 2-hydroxy-3-(4-hydroxy-3-methoxyphenyl)propyl- pounds’ proportions (expressed in percentage in crude material)
1,3-dibenzoate following the IUPAC rules, was, to our knowledge, in various samples of BSiam and BSum (Table 3). The vanillin 9
620 P. Burger et al. / Food Chemistry 210 (2016) 613–622

Table 3
Percentages in crude material of the major volatile (dark grey cells) and non-volatile compounds (light grey cells) identified in BSiam and BSum samples determined by
quantitation in GC-FID and HPLC respectively.

Siam benzoin Sumatra benzoin

BSiam-1 BSiam-2 BSiam-3 BSiam-4 BSum-1 BSum-2 BSum-3 BSum-4
vanillin 9 <1 <1 <1 <1 <1 <1 <1 <1
benzoic acid 10 33 31 36 31 2 4 2 4
cinnamic acid 11 n.d. n.d. n.d. n.d. 16 16 18 26
benzyl benzoate 12 1 1 <1 <1 <1 <1 <1 <1
volatile n.i. ≥ 1% 1 1 1 1 2 2 2 2
coniferyl benzoate 1 37 47 56 29 n.d. n.d. n.d. n.d.
cinnamyl cinnamate 3 n.d. n.d. n.d. n.d. 8 5 5 6
coniferyl benzoate
5 5 5 5 n.d. n.d. n.d. n.d.
derivatives 4-7
p-coumaryl cinnamate 15 n.d. n.d. n.d. n.d. 19 23 19 5
non-volatile n.i. ≥ 1% 8 4 1 10 15 6 10 20
n.i. ≤ 1% 10 8 n.d. 20 32 40 32 19
n.d. not detected; n.i. not identified.

Fig. 3. Gross composition of Siam and Sumatra benzoin extracts.

content, inferior to 1%, is quite similar in samples of both botanical apparently missing in Sumatra samples, which are on the contrary,
origins. A similar observation can be done for benzyl benzoate 12 characterised by the occurrence of p-coumaryl cinnamate 15 (5–
(less than 1%, except for samples BSiam-1 and BSiam-2). Quite 23%), and to a lesser extent, of cinnamyl cinnamate 3 (5–8%).
important proportions of benzoic acid 10 (31–36%) have been The gross compositions of both types of benzoin are presented
detected in BSiam extracts, compared to BSum extracts (3–5%) in Fig. 3: the distinction between both botanical origins is well evi-
which are rather characterised by significant proportions of cin- denced by the respective predominance. The part of unidentified
namic acid 11 (16–26%). These results are in accordance with pre- non-volatile compounds occurring in proportions higher than 1%
vious conclusions of studies of the volatile compositions of benzoin is significantly more important in both types of benzoin (up to
samples (Fernandez et al., 2003; Pastorova et al., 1997). 20%) compared to the volatile one, representing only up to 2%.
Quantitation of the major non-volatile compounds was per- One can also notice the higher percentage of unidentified com-
formed via the external standardization method in HPLC-ELSD pounds present as traces (61%) in Sumatra balsams (up to 40% of
using previously purified compounds. The formulas resulting from the total extract) compared to Siam samples (up to 20% of the total
the calibration curves constructed for coniferyl benzoate 1, p- extract). Future triplicate analysis and quantitation of an even lar-
coumaryl cinnamate 15 and cinnamyl cinnamate 3 allowed the ger number of samples would confirm these tendencies.
determination of the proportions of the respective compounds in
various samples of BSiam and BSum (Table 3). 3.3. Fast discrimination between both benzoin origins by HPTLC
BSiam extracts appear to be characterised by a high proportion
of coniferyl benzoate 1 (from 29 up to 56% in the four samples Previous analysis of benzoin balsams of both botanical origins
analysed), along with minor proportions of its derivatives siamyl by GC/MS evidenced the occurrence of both benzoic and cinnamic
mono- and dibenzoates 4–7 (c. 5%). Such compounds are acids 10 and 11 in Sumatra balsam, whereas only benzoic acid is
 P. Burger et al. / Food Chemistry 210 (2016) 613–622 621

the first time from Benzoe tonkinensis Laos: the coniferyl deriva-
tives were characterised in all Siam benzoin samples, whereas
the morinol ones were only isolated from ‘oldest’ samples. These
compounds probably correspond to degradation derivatives of
compounds entering the composition of fresh BTL. Experiments
of artificial ageing of BTL in an ageing chamber where various
parameters (room temperature and humidity, light intensity,
etc.) can be monitored would help to better understand the degra-
dation phenomenon taking place in Siam benzoin that lead to the
formation of these morinol derivatives.
This study constitutes a major breakthrough in the unravelling
of the overall composition of both Siam and Sumatra benzoin bal-
sams. However the part of unidentified compounds (mainly occur-
ring as traces) still represents roughly up to one third of the total
material. Systematic fractionation/sub-fractionation and isolation
A B C BSum BSiam D E of every single molecule could lead to a further move towards
the characterisation of these minor constituents of benzoin
Fig. 4. HPTLC plate developed with the supernatant of a solution of n-hexane/ balsams.
MeOH/acetic acid (40/10/5) and illuminated at 366 nm to visualize triterpenoids
present in benzoin of both botanical origins. Tracks: 1: lupenone; 2: a-amyrin; 3:
ursolic acid; 4: oleanolic acid and 5: lupeol. Conflict of interests

The authors declare that they have no conflict of interests.

present in Siam balsam (Castel et al., 2006; Fernandez et al., 2003).
The distinction of benzoin balsam origins by GC can be achieved by
the calculation of the ratio cinnamic acid/benzoic acid, that is nil in Appendix A. Supplementary data
the case of Siam benzoin balsam. Nonetheless, GC/MS analyses are
time-consuming and sample preparation can be fastidious: hence, Supplementary data associated with this article can be found, in
there was also a need for a visual and easy-to-set-up methodology the online version, at http://dx.doi.org/10.1016/j.foodchem.2016.
to discriminate both botanical origins. HPTLC enabling the simulta- 05.015.
neous automated analysis of a great set of samples is a choice ana-
lytical method for such discrimination.
References
A survey of the terpenic content of benzoin of both botanical
origins was performed using an elution solution consisting in AFNOR NF T 75-006 (1998). Aromatic natural raw materials—Vocabulary (Indice de
n-hexane/MeOH/acetic acid (40/10/5): both balsamic resins classement: T 75-006). ICS. 01.040.71; 71.100.60.
display specific HPTLC fingerprints (Fig. 4). In fact, one can observe Arctander, S. (1960). Perfume and flavor materials of natural origin (p. 736). Carol
Stream: Allured Publishing Corporation.
the elution of 2 specific compounds at approximately RF 0.05 Bruneton, J. (2009). Pharmacognosie, phytochimie, plantes médicinales (4th ed.
(violet band) and 0.5 (beige band) in the Sumatra sample. Siam p. 1292). Paris: Tec & Doc Lavoisier.
benzoin is characterised by 2 specific compounds at approximately Burdock, G. A. (1995). Fenaroli’s handbook of flavor ingredients (3rd ed., Vol. 1)
p. 368). Boca Raton, New-York, London, Tokyo: CRC Press.
RF 0.05 and 0.1 (brown bands) and 2 others at approximately RF Burdock, G. A. (1997). Encyclopaedia of food and color additives (Vol. 1). Boca Raton,
0.2 and 0.6 (orange bands). Yellow, orange and red bands detected New-York, London, Tokyo: CRC Press, p. 3153.
under light at 366 nm confirm the already stated presence of triter- Burkill, I. H. (1966). A dictionary of the economic products of the Malay Peninsula Vols.
1–2 (2nd ed.). London: Crown Agent for the Colony, p. 2444.
penoids in both benzoin balsams (Martelanc, Vovk, & Simonovska, Castel, C., Fernandez, X., Lizzani-Cuvelier, L., Loiseau, A.-M., Perichet, C., Delbecque,
2009). C., et al. (2006). Volatile constituents of benzoin gums: Siam and Sumatra, part
These results demonstrated that HPTLC is an accurate analytical 2. Study of headspace sampling methods. Flavour and Fragrance Journal, 21(1),
59–67.
method to rapidly obtain chemical fingerprints of both benzoin European Pharmacopoeia, (2008) (7th ed.). Benzoe sumatranus. 01/2008:1814.
types allowing their distinction. European Pharmacopoeia, (2016) (8th ed.). Benzoe tonkinensis. 01/2016:2158.
Fernández, C. G. (2004). Chapter 10 Benzoin, a resin produced by Styrax trees in
North Sumatra Province, Indonesia. In K. Kusters & B. Belcher (Eds.), Forest
products, livelihoods and conservation. Case studies of non-timbers forest product
4. Conclusions systems. Volume 1 – Asia (pp. 151–168). Jakarta, Indonesia: CIFOR.
Fernandez, X., Lizzani-Cuvelier, L., Loiseau, A.-M., Perichet, C., & Delbecque, C.
This article reports for the first time an extensive characterisa- (2003). Volatile constituents of benzoin gums: Siam and Sumatra. Part 1.
Flavour and Fragrance Journal, 18(4), 328–333.
tion of the chemical composition of both Siam and Sumata benzoin Fernandez, X., Castel, C., Lizzani-Cuvelier, L., Delbecque, C., & Puech Venzal, S.
balsams. The major volatile and non-volatile compounds of ben- (2006). Volatile constituents of benzoin gums: Siam and Sumatra, part 3. Fast
zoin of both botanical origins were identified and quantified by characterization with an electronic nose. Flavour and Fragrance Journal, 21(3),
439–446.
GC-FID and HPLC-DAD-ELSD respectively. Coniferyl benzoate 1 Filippi, J.-J., Castel, C., Fernandez, X., Rouillard, M., Gaysinski, M., & Lavoine-
and benzoic acid 10 are the major compounds identified in Siam Hanneguelle, S. (2009). An unusual acenaphthylene-type sesquiterpene
balsam, whereas cinnamic acid 11 and p-coumaryl cinnamate 15 hydrocarbon from Siam and Sumatra benzoin gum. Phytochemistry Letters, 2
(4), 216–219.
are predominant in Sumatra balsam. Coniferyl benzoate 1, p- Food and Agriculture Organization (2011). Compendium of food additive
coumaryl benzoate 2, benzyl cinnamate 13, cinnamyl benzoate specifications, 74th meeting. FAO JECFA Monograph 11.
14 and coniferyl cinnamate 16, as well as the triterpenic Hovaneissian, M., Archier, P., Mathe, C., & Vieillescazes, C. (2006). Contribution de la
chimie analytique à l’étude des exsudats végétaux styrax, storax et benjoin.
sumaresinolic acid 19, were identified as minor components in
Comptes Rendus Chimie, 9(9), 1192–1202.
Sumatra balsam. Minor compounds were also identified in Siam Hovaneissian, M., Archier, P., Mathe, C., Culioli, G., & Vieillescazes, C. (2008).
benzoin: p-coumaryl benzoate 2 and some triterpenic acids such Analytical investigation of styrax and benzoin balsams by HPLC-PAD-
as siaresinolic, sumaresinolic and 3-oxosiaresinolic acids 17–19. fluorimetry and GC-MS. Phytochemical Analysis, 19(4), 301–310.
Kashio, M., & Johnson, D. V. (2001). Monograph on benzoin.Thailand: Food and
Besides, four new coniferyl mono- and di-benzoates 4–7 and Agriculture Organization of the United Nations Regional Office for Asia and the
two morinol derivatives 8 and 80 were isolated and identified for Pacific Bangkok.
622 P. Burger et al. / Food Chemistry 210 (2016) 613–622

Langenheim, J. H. (1996). Chapter 1 biology of amber-producing trees: Focus on Reinitzer, F. (1914). Untersuchungen über Siambenzoë. Archiv der Pharmazie, 252
case studies of Hymenaea and Agathis. In K. B. Anderson & J. C. Crelling (Eds.), (3–5), 341–349.
Amber, resinite, and fossil resins (pp. 1–31). Washington DC: American Chemical Schroeder, H. A. (1968). The p-hydroxycinnamyl compounds of Siam Benzoin gum.
Society, ACS Symposium Series 617. Phytochemistry, 7(1), 57–61.
Langenheim, J. H. (2003). Plant resins: Chemistry, evolution ecology and ethnobotany. Su, B.-N., Takaishi, Y., & Kusumi, T. (1999). Morinols A-L, twelve novel
Portland, Cambridge: Timber Press, p. 586. sesquineolignans and neolignans with a new carbon skeleton from Morina
Leung, A. Y., & Foster, S. (1996). Encyclopaedia of common natural ingredients used in chinensis. Tetrahedron, 55(51), 14571–14586.
food, drugs and cosmetics (2nd ed.). New-York, Chichester, Brisbane, Toronto, Tiemann, F., & Haarmann, W. (1874). Ueber das Coniferin und seine Umwandlung in
Singapore: John Wiley & Sons, p. 688. das aromatische Princip der Vanille. Berichte der deutschen chemischen
Martelanc, M., Vovk, I., & Simonovska, B. (2009). Separation and identification of Gesellschaft, 7(1), 608–623.
some common isomeric plant triterpenoids by thin-layer chromatography and World Health Organization (2011). Evaluation of certain food additives and
high-performance liquid chromatography. Journal of Chromatography A, 1216, contaminants. 74th report of the Joint FAO/WHO Expert Committee on Food
6662–6670. Additives, WHO technical report series, 966, p. 136.
Modugno, F., Ribechini, E., & Colombini, M. P. (2006). Aromatic resin World Health Organization (2013). Evaluation of certain food additives and
characterisation by gas chromatography–mass spectrometry: Raw and contaminants. 77th report of the Joint FAO/WHO Expert Committee on Food
archaeological materials. Journal of Chromatography A, 1134(1–2), 298–304. Additives, WHO technical report series, 983, p. 75.
Moyler, D. A. (1998). The flavor of gum resins, their chemistry and uses. Rivista World Health Organization (2014). Evaluation of certain food additives. 79th report of
Italiana EPPOS (Special issue), 351–360. the Joint FAO/WHO Expert Committee on Food Additives, WHO technical report
Pastorova, I., de Koster, C. G., & Boon, J. J. (1997). Analytical study of free and ester series, 990, p. 124.
bound benzoic and cinnamic acids of gum benzoin resins by GC–MS and HPLC– 79th JECFA - Chemical and Technical Assessment (2014). Benzoe tonkinensis,
frit FAB–MS. Phytochemical Analysis, 8(2), 63–73. Chemical and Technical Assessment (CTA) prepared by Susanne Rath, PhD, and
Pauletti, P. M., Teles, H. L., Silva, D. H. S., Araújo, Â. R., & Bolzani, V. S. (2006). The reviewed by Harriet Wallin, MSc.
styracaceae. Brazilian Journal of Pharmacognosy, 16(4), 576–590.
Reich, E., & Schibli, A. (2006). High-performance thin-layer chromatography for the
analysis of medicinal plants. Thieme Medical Publishers, p. 264.