Fuel 188 (2017) 61–68

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Full Length Article

Evaluation of the oxidation stability of biodiesel stabilized with

antioxidants using the Rancimat and PDSC methods
Jian Zhou ⇑, Yun Xiong, Xiao Liu
Department of Fuel Chemistry, Logistical Engineering University, Chongqing 401311, China

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

Article history: In this study, both the standardized Rancimat and thermoanalytical pressure differential scanning
Received 7 August 2016 calorimetry (PDSC) methods were employed to investigate the effects of antioxidants on the oxidation
Received in revised form 28 September stability of biodiesel. The differences in the antioxidant effectiveness between these phenolic compounds
2016
could be clearly distinguished by both methods. Results revealed that the efficiency of antioxidant was
Accepted 2 October 2016
Available online 7 October 2016
highly dependent on the test conditions used in each method. Among the five antioxidants investigated,
tert-butylhydroquinone (TBHQ) showed the highest sensitivity to different conditions between two
methods: the highest oxidation stability in Rancimat vs. the lowest efficiency in PDSC. Butylated hydrox-
Keywords:
Biodiesel
yanisole (BHA) and butylated hydroxytoluene (BHT) acted more efficiently in PDSC than in Rancimat due
Antioxidants to lower volatilization losses, but not as well as pyrogallol (PY) and propyl gallate (PG) in increasing the
Oxidation stability oxidation stability of biodiesel. Results from isothermal and nonisothermal PDSC methods were basically
Rancimat consistent with each other with respect to the antioxidant efficiency: PY > PG > BHA, BHT > TBHQ. A pos-
PDSC itive linear relationship between results obtained from both methods could be observed within each type
TBHQ of antioxidant, but the overall correlation for the entire data set between Rancimat and PDSC was not as
strong as that between isothermal and nonisothermal PDSC methods.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction by free radicals. Synthetic antioxidants are generally more pre-

Over the past few decades, biodiesel has received increasing
attention as a renewable alternative biofuel. Biodiesel offers many
advantages over conventional petrodiesel: inherent lubricity, non-
toxic and biodegradable, free of sulfur and aromatics, higher cetane
number and flash point, and lower exhaust emissions such as CO2,
CO, PM, and HC, excepting higher NOx emissions [1,2]. However,
biodiesel is highly susceptible to oxidative degradation because
of its high content of unsaturated methyl esters. The resistance
of biodiesel against oxidation has long been a critical concern to
fuel manufacturers, distributors and consumers, which depends
not only on the fatty acid composition profile, but also on the pro-
duction process and storage conditions [3,4]. During the long-term
storage of biodiesel, the presence of light, oxygen, metal ions, high
temperature, etc. should be avoided as much as possible [5].
Most of the biodiesel produced requires the use of antioxidants
to meet the oxidation stability requirements set forth in ASTM
D6751 (3 h) or EN 14214 (8 h) [6,7]. Addition of antioxidants to
biodiesel could help slow down the process of oxidation caused
⇑ Corresponding author.
E-mail address: zhoujian2005@yahoo.com (J. Zhou).
ferred than natural antioxidants due to higher efficiency. The most
commonly used in biodiesel are the following synthetic phenolic
compounds: pyrogallol (PY), propyl gallate (PG), butylated hydrox-
yanisole (BHA), butylated hydroxytoluene (BHT), and tert-
butylhydroquinone (TBHQ). These antioxidants, more easily oxi-
dized than biodiesel itself, could interrupt the chain propagation
reactions by donating hydrogen atoms to peroxyl radicals from
their phenolic hydroxyl groups, with the formation of less reactive
antioxidant radicals that do not initiate another free radical or
propagate the oxidation process [8].
To evaluate the antioxidant efficiency on the oxidative stability
of biodiesel, the following accelerated oxidation methods are most
often employed: Rancimat, EN 14112, the determination of oxida-
tion products; PetroOXY, ASTM D7545, the determination of oxy-
gen uptake; PDSC, ASTM D6186 and E2009, monitoring of the
heat evolved during the oxidation reactions [9]. These methods
provide elevated temperature and/or pressure to accelerate the
oxidation reaction and shorten the experimental time. Although
actual storage conditions are not truly reflected by these acceler-
ated oxidation methods, the study of oxidation stability would take
too long and be unfeasible without them.
Rancimat is the only accepted method by EN 14214 standard for
assessing the oxidation stability of biodiesel. The induction period

http://dx.doi.org/10.1016/j.fuel.2016.10.026
0016-2361/Ó 2016 Elsevier Ltd. All rights reserved.
62 J. Zhou et al. / Fuel 188 (2017) 61–68
is determined by monitoring the continual increase of conductiv-
ity, which is mainly caused by the volatile secondary oxidation
products, such as formic and acetic acids [10–12]. Usually, the
longer the induction period, the better the oxidation stability.
The biggest disadvantage to this method is the test duration being
too long: a single test will require much more time than before
given that the oxidation stability specification in the EN 14214
standard has increased from the earliest 6 h to the present 8 h.
To shorten the time of analysis, raising the test temperature from
110 °C to 130 °C was suggested. However, higher temperatures
could yield misleading results that do not accurately reflect antiox-
idant performances as some antioxidants, like BHA and BHT, could
easily volatilize and decompose at higher temperatures [13,14].
PetroOXY provides a fast, safe and promising alternative for
measuring the oxidation stability of biodiesel. The high pressure
in this method could decrease the volatilization loss of additive
and saturate the sample with oxygen. The induction period is
determined by monitoring the pressure drop resulting from the
consumption of oxygen in oxidation reactions, which corresponds
to the time required to achieve a fixed (10%) pressure drop. The
fuel sample can be subjected to more severe test conditions in Pet-
roOXY than in Rancimat: there is a maximum pressure of approx.
1000 kPa when the temperature reaches 140 °C as per ASTM
D7545. This method has shown good repeatability, reproducibility,
and judgement between different replicates [15–18]. However, a
minimum value of induction period using the PetroOXY method
has not been accepted yet, and the correlation between PetroOXY
and Rancimat is still under investigation [19–21]. Note that, this
method is not the focus of this article; the correlation between
Rancimat and PetroOXY will be explored in detail in a separate
paper.
The thermoanalytical technique, PDSC, that directly measures
the heat flow associated with chemical reactions, has been widely
used over the past 25 years to investigate the thermooxidative sta-
bility of lubricating oils and greases, aviation turbine oils, veg-
etable oils and polymers [22]. This method can be run in an
isothermal mode to give an oxidation induction time (ASTM
D6186) or in a programmed temperature (nonisothermal) mode
to give an oxidation onset temperature (ASTM E2009) [23]. The
acceleration of autoxidation could be easily detected by this
method: the heat flow remains stable prior to oxidation but rises
sharply to the maximum exothermic peak when oxidation reaction
begins. The high temperature and pressure recommended by ASTM
E2009 and D6186 (130–210 °C, 500 psi) could effectively suppress
the volatilization of sample materials and shorten the experimen-
tal time to minutes instead of hours. But these recommended
parameters are not suitable for evaluating the oxidation stability
of biodiesel due to the fundamental differences in the chemical
structures between biodiesel and lubricants. Therefore, a prior
selection of appropriate temperature and pressure is required to
give a measurable induction time (isothermal mode) or onset tem-
perature (nonisothermal mode): the induction time will either be
impractically too short or too long when chosen at too high or
too low temperature, respectively. Lower temperatures and pres-
sures (110 °C, 200 or 80 psi) were most often employed in previous
studies, in which the PDSC method demonstrated good indication
of the oxidation stability of biodiesel [2,24–28].
A pilot study using the above mentioned three methods to
investigate the oxidation stability of biodiesel was carried out by
Murta Valle et al. in 2014 [27]. Results showed that the sensitivity
of the oxidation stability of soybean oil biodiesel to the antioxidant
concentration, bisphenol and phenol, was highly dependent on the
method applied; PDSC presented better sensitivity and shorter
analysis time as compared to Rancimat and PetroOXY. However,
most research done to date have been limited to a single test
method for assessing antioxidant efficiency [4,6,7,16]. There’s not
much attention paid to investigating the possible differences in
the antioxidant efficiency between two or more methods, and
few systematic research has been carried out to verify the depen-
dence of the sensitivity of antioxidant efficiency on the applied
method. Based on this background, the standardized Rancimat
and promising PDSC methods were employed in this study to
investigate the efficiency of the antioxidants on the oxidation sta-
bility of biodiesel, and attempts were also made to establish the
correlation between these two methods with different test
conditions.
2. Materials and methods
2.1. Materials
The waste cooking oil biodiesel was used in this study, which
was provided by Chengdu Hengrun Hi-tech Co., Ltd., Sichuan,
China. This biodiesel did not contain any antioxidant additives,
and fulfilled the EN 14214 quality specifications with the exception
of oxidation stability. The main chemical and physical properties of
the fresh biodiesel were given as follows: viscosity at 40 °C,
4.20 mm2/s; density at 15 °C, 0.878 g/cm3; CFPP, 5 °C; flash point,
159 °C; cetane number, 54.0; sulfur content, <1 ppm; water con-
tent, 300 ppm; acid number, 0.37 mg KOH/g; copper strip corro-
sion (3 h at 100 °C), 1.0; oxidation stability at 110 °C, 3.01 h.
The following antioxidants, purchased from Aladdin Industrial Cor-
poration (analytical grade), were investigated for their impacts on
improving the oxidation stability of this biodiesel: PY, PG, BHA,
BHT, and TBHQ. These antioxidants were added into the biodiesel
at the concentration of 100, 250, 500, 750, and 1000 ppm (by
weight), after which the treated biodiesel was kept in the dark at
room temperature.
2.2. GC/MS analysis
The fatty acid composition of the waste cooking oil biodiesel
was determined by gas chromatography/mass spectroscopy (GC/
MS). The chromatographic analysis was performed on Agilent
7890A GC system (Agilent Technologies) equipped with a 5975C
MSD and a HP-5MS UI column (30 m  250 lm  0.25 lm) using
Helium as the carrier gas. The column temperature was held con-
stant at 180 °C for 3 min, ramped to 200 °C at 1 °C/min, and then
held at 200 °C for 2 min, and lastly at 20 °C/min to 250 °C, where
it was kept for 2 min. The transfer line between GC and MS was
kept at 260 °C. The EI mass spectra (70 eV) was acquired in the
m/z range 10–400. The injection volume was 1 ll and a split ratio
of 100:1 was employed. The biodiesel was dissolved in tetrahydro-
furan and quantification analysis was performed by internal stan-
dard calibration using methyl heptadecanoate (Sigma-Aldrich).
The fatty acid profile of this biodiesel was presented as follows:
myristic acid (C14:0), 0.59%; palmitic acid (C16:0), 15.53%; palmi-
toleic acid (C16:1), 1.27%; stearic acid (C18:0), 6.80%; oleic acid
(C18:1), 41.53%; linoleic acid (C18:2), 27.78%; linolenic acid
(C18:3), 1.10%; arachidic acid (C20:0), 0.41%; eicosenoic acid
(C20:1), 1.44%; behenic acid (C22:0), 0.24%; erucic acid (C22:1),
2.27%.
2.3. PDSC method
The PDSC measurements were carried out using a DSC Q2000
TA Instruments, coupled to a pressure cell. The instrument was cal-
ibrated with high-purity indium before test. Approximately
3 ± 0.3 mg sample was carefully weighed in an open aluminum
pan and placed in the equipment’s sample chamber, and an identi-
cal empty pan was placed on the rear platform as a reference. The
 J. Zhou et al. / Fuel 188 (2017) 61–68
experiments were performed at 1400 kPa (200 psi) under static
oxygen atmosphere (zero purge gas flow). In the isothermal mode,
the cell was heated from 40 °C to the specified temperature 110 °C
at a heating rate of 40 °C/min, then maintained at the regulated
temperature and pressure until an exothermic reaction occurred.
The nonisothermal measurements were performed under the same
conditions of atmosphere, pressure, and sample quantity, but at a
heating rate of 10 °C/min in the temperature range of 40–200 °C.
Before the start of heating, an equilibrium time of about 2 min
was recommended to enhance the baseline. Experimental data
were analyzed using TA Universal Analysis software. The oxidation
induction time and oxidation onset temperature were determined
by the intersection of the extrapolated baseline and tangent to the
maximum exothermic peak.
2.4. Rancimat method
The Rancimat test was conducted as per EN 14112 using an 873
Biodiesel Rancimat instrument from Metrohm, Herisau, Switzer-
land. Approximately 3 ± 0.2 g sample was carefully weighed in a
glass test tube and analyzed under constant airflow of 10 L/h
within the thermostat-controlled block heater at 110 °C. The
stream of purified air swept the volatile oxidation products formed
during the accelerated aging out, which were collected in an elec-
trical conductivity measuring vessel containing 50 mL of deionized
water where the conductivity was continuously measured. Induc-
tion period was defined as the time taken from the start of the test
to the onset of oxidation, and determined from the second deriva-
tive of the conductivity curve automatically. The experiments were
carried out in at least duplicate to evaluate the reliability of both
methods employed. The variability of our determinations was
within ±5%, which did not exceed the repeatability limit of their
respective standard method.
3. Results and discussion
3.1. Rancimat: induction period
Typical conductivity curves for biodiesel treated with antioxi-
dants, as determined by the Rancimat method, were plotted
against time as shown in Fig. 1. As the oxidation reaction pro-
gressed, the conductivity increased rather slowly and almost lin-
early with time until accelerated oxidation stage, where it
Fig. 1. Rancimat conductivity curves for biodiesel treated with 250 ppm of
antioxidants measured at 110 °C. The navy dashed line indicates the 2nd derivative
of conductivity curve for biodiesel stabilized with TBHQ.
63
showed a sudden sharp increase in conductivity. The induction
period, defined as the time corresponding to the inflection point
of the conductivity vs. time curve, was used for quantitative deter-
mination of the oxidative stability of biodiesel. The longer the
induction period, the more stable is the biodiesel. The untreated
biodiesel showed an induction period of 3.01 h and failed to meet
the EN 14214 specification (8 h), which could be mainly attributed
to the structural composition and oxidative reactivity of the unsat-
urated fatty acid methyl esters. For the waste cooking oil based
biodiesel used in this study, 23.57% were saturated, 46.51% were
monounsaturated, and 28.88% were polyunsaturated: methyl ole-
ate was the predominant fatty acid methyl ester (41.53%), followed
by methyl linoleate (27.78%), and methyl palmitate (15.53%). The
composition profile of this biodiesel was in good agreement with
previous reports that also choosing used cooking oil as raw mate-
rials [4,14]. As can be seen from Fig. 1, addition of antioxidants to
biodiesel could enhance the oxidation stability by extending the
induction period: values of 6.27, 7.24, 15.10, 20.07, and 22.30 h
were found for biodiesel treated with 250 ppm of BHT, BHA, PG,
PY, and TBHQ, respectively.
Fig. 2 shows the dependence of the Rancimat induction period
of the biodiesel on the antioxidant concentration measured at
110 °C. We can see clearly that the higher the concentration, the
longer the induction period. The addition of different antioxidants
could improve the oxidation stability of the biodiesel to different
degrees. As can be observed from Fig. 2, PY, PG, and TBHQ per-
formed much better than BHA and BHT. The induction periods of
PY and PG were at least two times as large as those of BHA and
BHT at all concentrations. PY demonstrated better oxidative stabil-
ity than PG, especially at concentrations below 500 ppm, while the
difference between them got smaller as the concentration
increased further. As can be seen from these results, PY and TBHQ
could significantly improve the stability of biodiesel at the lowest
concentrations, while no noticeable improvements could be
observed for BHA and BHT even at the highest concentrations.
Among the five antioxidants studied, almost 500 ppm of BHA and
BHT was needed to meet the EN 14214 oxidation stability require-
ment (8 h); however, 100 ppm was sufficient to achieve this in the
case of PY, PG, and TBHQ: the antioxidants PY and TBHQ at
100 ppm were far more effective than BHA and BHT at all concen-
trations tested in this study. Overall, the effectiveness of these
antioxidants could be ranked as TBHQ > PY > PG > BHA > BHT, in
agreement with a previous study [4]. Although PY showed the
Fig. 2. Rancimat induction period vs. antioxidant concentration measured at
110 °C.
64 J. Zhou et al. / Fuel 188 (2017) 61–68
longest induction period (16.11 h) at the lowest concentration
(100 ppm), TBHQ demonstrated far better antioxidant efficiency
than the other four antioxidants at higher concentrations
(>100 ppm). The induction period of TBHQ at 1000 ppm was about
4.6, 4.7, 2.0, and 2.1 times as long as those of to BHA, BHT, PY, and
PG, respectively.
The differences in the antioxidant efficiency could to a large
extent be explained by the molecular structures of these phenolic
compounds [8]. The hydrogen atom abstracted from the hydroxyl
group of the antioxidant could be transferred to the oxidized free
radicals, which serves to interrupt the chain propagation reactions
and inhibit the rate of oxidation. The stable antioxidant radical
formed can further react with other free radicals to contribute to
the inhibition of oxidation. PY and PG have three hydroxyl groups
attached to their aromatic rings, and could offer more sites than
BHA and BHT for the formation of complex between antioxidant
radical and free radical. TBHQ bears two hydroxyl groups in the
aromatic ring, but its derivative products formed during the
antioxidative process may also retain antioxidant properties, and
some of them have even been reported to have higher activity than
TBHQ [29,30]. BHA and BHT possess only one hydroxyl group in
their aromatic rings and are proved to be less effective than PY,
PG, and TBHQ in most cases. Another contributing factor for the
poor antioxidant efficiency of BHT and BHA is the possible
volatilization losses during high temperature measurements
because of their relatively high volatility [13].
3.2. Isothermal PDSC: oxidation induction time
The striking advantage of the PDSC method is that the use of
high pressure helps to decrease the sample volatility by elevating
the boiling point. Fig. 3 shows isothermal PDSC curves for biodiesel
stabilized with 250 ppm antioxidants. The onset of oxidation at the
specified temperature is signaled by an abrupt increase in the sam-
ple’s evolved heat, which results in an exothermic peak in the PDSC
thermal curve. The time interval from the start of the isothermal
experiment to the appearance of the exothermic peak is defined
as oxidation induction time. During the induction phase of biodie-
sel oxidation, peroxyl radicals can be efficiently trapped by the
antioxidant present in the sample. Once the antioxidant is con-
sumed to a critical level, it is no longer effective to retard oxidative
deterioration, at which point the biodiesel sample catastrophically
oxidizes liberating large amounts of heat (exotherm). The length of
the induction time is often considered as a measurement of the
Fig. 3. Isothermal PDSC curves for biodiesel treated with 250 ppm of antioxidants
measured at 110 °C.
oxidation stability. The longer the induction time, the better the
oxidation stability. The untreated biodiesel exhibited poor oxida-
tion stability with an induction time of 23.54 min at 110 °C, which
could be raised to 62.04, 96.38, 101.44, 119.44, and 157.34 min for
biodiesel stabilized with 250 ppm of TBHQ, BHA, BHT, PG, and PY,
respectively. It should be noted that the type of antioxidant used
cannot be distinguished directly from the shape of endothermic
peak, but the induction time obtained can be used to rank the
effectiveness of various antioxidants.
Fig. 4 shows the PDSC induction time of biodiesel treated with
antioxidants measured at 110 °C. As expected, the induction time
increased with increasing antioxidant concentration. BHA, BHT,
PY, and PG provided the greatest enhancement of oxidation stabil-
ity, whereas TBHQ presented the lowest oxidation stability, con-
trary to the highest antioxidant efficiency shown in the Rancimat
method: 1000 ppm of TBHQ could not match 250 ppm of other
four antioxidants in the PDSC method. PY and PG demonstrated
better antioxidant efficiency than BHA and BHT in the whole range
of concentrations, but the differences between them were not so
large as that observed in Rancimat, especially at higher concentra-
tions. BHT provided better protection against oxidation than BHA;
addition of BHT resulted in a longer induction time. Overall, the
order of the antioxidant efficiency of these phenolic compounds
obtained from isothermal PDSC method was as follows:
PY > PG > BHT > BHA > TBHQ.
3.3. Stabilization factor: Rancimat vs. PDSC
Although the five antioxidants displayed high effectiveness in
protecting the biodiesel from oxidative degradation in both meth-
ods, there were notable differences in the antioxidant efficiency
between Rancimat and PDSC. For the purpose of comparison
between induction periods obtained from two different methods,
the stabilization factor SF was introduced to express the antioxi-
dant efficiency: SF = IPA/IP0, where IPA and IP0 are the induction
periods in the presence and absence of antioxidant, respectively.
The larger the stabilization factor, the higher the antioxidant effi-
ciency. As can be seen in Fig. 5, these phenolic compounds per-
formed much better in PDSC than in Rancimat with an exception
of TBHQ, as the former SF was obviously larger than the latter.
The effectiveness of BHT and BHA was significantly higher in PDSC
than in Rancimat, most likely due to lower volatilization losses by
the high pressure adopted in the PDSC method, which also indi-
Fig. 4. PDSC oxidation induction time vs. antioxidant concentration measured at
110 °C.
 J. Zhou et al. / Fuel 188 (2017) 61–68
Fig. 5. Stabilization factor vs. antioxidant concentration for Ranciamt (a) and PDSC (b).
cated that the actual antioxidant efficiency could be underesti-
mated by the Rancimat method.
For the same antioxidant level, the differences in the antioxi-
dant efficiency between PDSC and Rancimat could be attributed
to the antioxidant sensitivity to the test conditions: 200 psi pure
oxygen (1400 kPa) in severe PDSC versus 1 atm air supply (10 L/
h) in mild Rancimat. The stabilization factor ratio SFR, defined as
the ratio of the larger stabilization factor to the smaller one deter-
mined by both methods (no less than 1), was introduced to express
the antioxidant sensitivity to the differences between two meth-
ods. The larger the stabilization factor ratio, the more sensitive
the antioxidant is to the different test conditions between two
methods. It can be clearly seen from Fig. 6 that TBHQ showed
the highest sensitivity, followed by BHT, BHA, PG, and PY. BHT
and BHA were far more sensitive than PG and PY to the differences
in the test conditions between Rancimat and PDSC, whereas PY and
PG proved to be more effective than BHA and BHT at inhibiting bio-
diesel oxidation. A possible explanation is that BHA and BHT could
easily volatilize during the Rancimat measurements because of
their relatively high volatility [13]. When the tests are carried
out under elevated pressure in PDSC, the volatilization of BHT
and BHA could be effectively suppressed, and they have acted more
Fig. 6. Stabilization factor ratio vs. antioxidant concentration.
65
efficiently than in the Rancimat method, although not as well as PY
and PG.
TBHQ exhibited the longest induction period in the Rancimat
method but the lowest antioxidant efficiency in PDSC; the highest
SFR value could reach 4.70 for 1000 ppm of TBHQ. It should be
mentioned that both methods were carried out at almost the same
time. Among the five antioxidants investigated, TBHQ has often
been reported to display excellent antioxidant efficiency in the
Rancimat method; the same high efficiency, however, has rarely
been reported in PDSC [25]. Pinto et al. investigated the inhibitory
effects of a-tocopherol, green tea extract, and TBHQ on peanut oil
biodiesels obtained by methyl (MPB) and ethyl route (EPB) using
both Rancimat and PDSC methods [28]. They reported that TBHQ
showed much better antioxidant efficiency in Rancimat than in
PDSC, and the SFR values obtained for 2000 ppm of TBHQ were
3.94 and 5.04 for MPB and EPB, respectively, coinciding with our
study. Similar observations about TBHQ have also been reported
by other researchers [31]. We cannot rule out the possibility that
the contradictory behavior of antioxidant TBHQ in PDSC is due to
possible participation in side reactions with excess oxygen in the
chamber or thermal decomposition losses by the high temperature
and pressure adopted in this method. The different test conditions
of PDSC from Rancimat, mainly the elevated oxygen pressure,
could alter the antioxidative mechanism of this phenolic com-
pound and increase the possible occurrence of participation in side
reactions [19]. More research is needed in the future to confirm
this assumption.
3.4. Nonisothermal PDSC: oxidation onset temperature
The main advantage of nonisothermal PDSC method is that it
can be used to determine the onset temperature of oxidation and
measure the kinetic parameters of the process within a short time,
which gives another way to look at the oxidation stability of the
biodiesel. Fig. 7 depicts nonisothermal PDSC curves for biodiesel
oxidation inhibited with 1000 ppm of antioxidants. The heat flow
remains relatively constant until oxidation of the sample begins,
releasing huge amounts of energy in the process. As seen from
Fig. 7, the addition of antioxidant was manifested as an increase
of onset temperature corresponding to the first maximum of the
heat flow, which was greatly higher than the isothermal tempera-
ture. The experiment was ended soon after the first exothermal
peak appeared to save time for further experiments. The first peak
was assigned to the formation of hydroperoxides, while the second
66 J. Zhou et al. / Fuel 188 (2017) 61–68
Fig. 7. Nonisothermal PDSC curves for biodiesel treated with 1000 ppm of
antioxidants.
peak (not shown) resulted from the decomposition of the
hydroperoxides to secondary degradation products, which did
not correlate directly with the antioxidative effects [32,33]. The
onset temperature of the untreated biodiesel increased from
152.40 °C to 169.60, 176.35, 179.81, 184.96, and 191.08 °C for bio-
diesel treated with 1000 ppm of TBHQ, BHT, BHA, PG, and PY,
respectively. The higher the onset temperature, the better the ther-
mooxidative stability.
The results presented in Fig. 8 showed that the onset tempera-
ture of biodiesel treated with antioxidants could be shifted to
higher temperatures in comparison to untreated biodiesel. The
addition of antioxidants could effectively improve the thermoox-
idative stability of biodiesel. Antioxidant activity and antioxidant
concentration demonstrated a positive linear relationship: higher
concentration could contribute to better stability. The effectiveness
of the five chain-breaking antioxidants evaluated for protecting
biodiesel from thermooxidative degradation could be ranked in
the ascending order: TBHQ < BHT < BHA < PG < PY, basically in
agreement with previous observations in isothermal PDSC method.
Obviously, TBHQ was more significantly affected than other four
antioxidants by the high temperature and pressure with respect
to improving resistance to oxidation. BHA showed better ther-
Fig. 8. PDSC oxidation onset temperature vs. antioxidant concentration.
mooxidative protection than BHT, which could be attributed to
BHT’s higher rates of decomposition loss at elevated temperatures
[13]. There was no fundamental contradiction between antioxidant
efficiency determined from isothermal and non-isothermal exper-
iments, although different mechanisms were thought being
involved in inhibiting oxidative deterioration.
3.5. Correlation study: Rancimat vs. PDSC
Fig. 9 presents the correlation between PDSC oxidation induc-
tion time (min) and Rancimat induction period (h) measurements
for biodiesel stabilized with antioxidants determined at 110 °C.
The different signs correspond to different types of antioxidants,
and the lines come from regression analysis. As can be seen from
the short-dashed lines in Fig. 9, there was positive linear correla-
tions (R2 > 0.90) within each type of antioxidant between induction
period values obtained from Rancimat and PDSC, in agreement
with findings reported by Leonardo et al. [26]. All the regressions
were statistically significant with p value < 0.02. However, no over-
all correlation could be detected between Rancimat and PDSC for
the entire data set. This could be explained by the antioxidant sen-
sitivity to the different test conditions between two methods. Due
to possible volatilization, decomposition or participation in side
reactions, these antioxidants could undergo different antioxidative
mechanisms in inhibiting oxidation between Rancimat and PDSC.
Considering the contradictory behavior of TBHQ in both methods,
the overall relationship between the oxidation stability values of
Rancimat and PDSC for other four types of antioxidants, i.e. BHA,
BHT, PY, and PG was established: the red solid line in Fig. 9 showed
a coefficient of determination value of 0.69 between them. It
should be noted, however, that this correlation was dependent
on the biodiesel and antioxidants used, and could not be guaran-
teed for other kinds of biodiesel and antioxidants.
Taking into account the consistent behaviors of antioxidants in
both isothermal and nonisothermal PDSC methods, the correlation
between oxidation induction time and oxidation onset tempera-
ture was established and presented in Fig. 10, which showed a pos-
itive linear correlation (R2 > 0.93) within each type of antioxidant.
The overall correlation for the entire data set denoted a statistically
significant correlation between isothermal and nonisothermal
PDSC methods (R2 = 0.89), independent of the antioxidant used.
Fig. 9. Correlation between PDSC oxidation induction time and Rancimat induction
period measurements for biodiesel treated with antioxidants determined at 110 °C.
The red solid line indicates the overall relationship between induction period values
of Rancimat and PDSC for BHA, BHT, PY, and PG.
 J. Zhou et al. / Fuel 188 (2017) 61–68
Fig. 10. Correlation between isothermal PDSC (induction time) and nonisothermal
PDSC (onset temperature) measurements for biodiesel treated with antioxidants.
Both induction time and onset temperature parameters character-
ize the onset of oxidative degradation when the antioxidant level
dips below a critical value where it is no longer effective [23].
However, different stabilizing mechanisms may be involved in
inhibiting oxidation in these two methods, as oxidation induction
time and oxidation onset temperature data are obtained at very
different temperatures. Further studies are still necessary to verify
this relationship that has been drawn from the limited data in this
study. The strong correlation between these two parameters did
prove that the PDSC technique can be a useful tool for screening
the effectiveness of antioxidants.
4. Conclusion
In this paper, both the Rancimat and PDSC methods were
employed to evaluate the effectiveness of five synthetic antioxi-
dants on the oxidation stability of biodiesel, i.e. TBHQ, BHA, BHT,
PY, and PG. Results obtained from these two methods were basi-
cally consistent with each other, with the exception of TBHQ. TBHQ
showed the highest antioxidant efficiency in the Rancimat method
but lowest oxidation stability in PDSC; one possible explanation is
that TBHQ participates in side reactions with excess oxygen in the
test chamber or undergoes severe thermal decomposition by the
high temperature and pressure adopted in this method, and this
assumption deserves further research attention. The other four
antioxidants, especially BHT and BHA, acted more efficiently in
PDSC than in Rancimat, as the high pressure provided by the ther-
moanalytical technique could effectively reduce the volatilization
losses. Due to their molecular structural differences, PY and PG
demonstrated better antioxidant efficiency than BHA and BHT over
the concentration range studied in both methods. Results from
isothermal (induction time) and nonisothermal (onset tempera-
ture) PDSC methods basically agreed with each other with respect
to the effectiveness of the five chain-breaking antioxidants:
PY > PG > BHA, BHT > TBHQ. There was a nearly linear relationship
within each type of antioxidant between the results of the two
measurement methods, but the overall correlation for the entire
data set between Rancimat induction period and PDSC induction
time was not as strong as the positive linear trend between
isothermal and nonisothermal PDSC methods. The PDSC method
shows great potential for the discrimination of different antioxi-
dants. But it is still too early to come to the conclusion that PDSC
presents better choice than Rancimat for screening antioxidant
67
efficiency in biodiesel. Taking into consideration the inconsistent
behavior of TBHQ in both methods, more research is needed in
the future to investigate antioxidant behaviors in different test
methods.
Acknowledgements
The authors would like to thank the Scientific Instrument Cen-
ter of Logistical Engineering University for providing research facil-
ities for this work.
References
[1] Dunn RO. Antioxidants for improving storage stability of biodiesel. Biofuel
Bioprod Bior 2008;2:304–18.
[2] Dunn RO. Thermal oxidation of biodiesel by pressurized differential scanning
calorimetry: effects of heating ramp rate. Energy Fuel 2012;26:6015–24.
[3] Knothe G. Dependence of biodiesel fuel properties on the structure of fatty acid
alkyl esters. Fuel Process Technol 2005;86:1059–70.
[4] Mittelbach M, Schober S. The influence of antioxidants on the oxidation
stability of biodiesel. J Am Oil Chem Soc 2003;80:817–23.
[5] Dunn RO. Effect of temperature on the oil stability index (OSI) of biodiesel.
Energy Fuel 2007;22:657–62.
[6] Tang H, Wang A, Salley SO, Simon Ng KY. The effect of natural and synthetic
antioxidants on the oxidative stability of biodiesel. J Am Oil Chem Soc
2008;85:373–82.
[7] Sulistyo H, Almeida MF, Dias JM. Influence of synthetic antioxidants on the
oxidation stability of biodiesel produced from acid raw Jatropha curcas oil.
Fuel Process Technol 2015;132:133–8.
[8] Rizwanul Fattah IM, Masjuki HH, Kalam MA, Hazrat MA, Masum BM, Imtenan
S, et al. Effect of antioxidants on oxidation stability of biodiesel derived from
vegetable and animal based feedstocks. Renew Sustain Energy Rev
2014;30:356–70.
[9] Focke WW, van der Westhuizen I, Lofte Grobler AB, Nshoane KT, Reddy JK, Luyt
AS. The effect of synthetic antioxidants on the oxidative stability of biodiesel.
Fuel 2012;94:227–33.
[10] Focke WW, van der Westhuizen I, Oosthuysen X. Biodiesel oxidative stability
from Rancimat data. Thermochim Acta 2016;633:116–21.
[11] Flitsch S, Neu PM, Schober S, Kienzl N, Ullmann J, Mittelbach M. Quantitation
of aging products formed in biodiesel during the rancimat accelerated
oxidation test. Energy Fuel 2014;28:5849–56.
[12] de Carvalho AL, Cardoso EA, da Rocha GO, Teixeira LS, Pepe IM, Grosjean DM.
Carboxylic acid emissions from soybean biodiesel oxidation in the EN14112
(Rancimat) stability test. Fuel 2016;173:29–36.
[13] Hamama AA, Nawar WW. Thermal decomposition of some phenolic
antioxidants. J Agric Food Chem 1991;39:1063–9.
[14] Lapuerta M, Rodríguez-Fernández J, Ramos Á, Álvarez B. Effect of the test
temperature and anti-oxidant addition on the oxidation stability of
commercial biodiesel fuels. Fuel 2012;93:391–6.
[15] Neumann A, Jebens T, Wierzbicki V. A method for determining oxidation
stability of petrodiesel, biodiesel, and blended fuels. Am Lab 2008;40.
[16] Araújo SV, Luna FMT, Rola EM, Azevedo DCS, Cavalcante Jr CL. A rapid method
for evaluation of the oxidation stability of castor oil FAME: influence of
antioxidant type and concentration. Fuel Process Technol 2009;90:1272–7.
[17] Medeiros ML, Cordeiro AM, Queiroz N, Soledade LEB, Souza AL, Souza AG.
Efficient antioxidant formulations for use in biodiesel. Energy Fuel
2014;28:1074–80.
[18] Deligiannis A, Anastasakos A, Anastopoulos G, Dodos GS, Zannikos F.
Effectiveness of various phenolic compounds (commercial and non-
commercial) on biodiesel oxidation stability. SAE technical paper; 2013.
[19] Dodos GS, Karonis D, Zannikos F, Lois E. Assessment of the oxidation stability
of biodiesel fuel using the Rancimat and the RSSOT methods. SAE technical
paper; 2014.
[20] Botella L, Bimbela F, Martín L, Arauzo J, Sánchez JL. Oxidation stability of
biodiesel fuels and blends using the Rancimat and PetroOXY methods. Effect of
4-allyl-2, 6-dimethoxyphenol and catechol as biodiesel additives on oxidation
stability. Front Chem 2014;2.
[21] Zhou J, Xiong Y, Xu S. Evaluation of the oxidation stability of biodiesel
stabilized with antioxidants using the PetroOXY method. Fuel
2016;184:808–14.
[22] Sharma BK, Stipanovic AJ. Development of a new oxidation stability test
method for lubricating oils using high-pressure differential scanning
calorimetry. Thermochim Acta 2003;402:1–18.
[23] Focke WW, van der Westhuizen I. Oxidation induction time and oxidation
onset temperature of polyethylene in air. J Therm Anal Calorim
2010;99:285–93.
[24] Dantas MB, Albuquerque AR, Soledade LEB, Queiroz N, Maia AS, Santos IMG,
et al. Biodiesel from soybean oil, castor oil and their blends: oxidative stability
by PDSC and rancimat. J Therm Anal Calorim 2011;106:607–11.
[25] Tavares MLA, Queiroz N, Santos IMG, Souza AL, Cavalcanti EHS, Barros AKD,
et al. Sunflower biodiesel: use of P-DSC in the evaluation of antioxidant
efficiency. J Therm Anal Calorim 2011;106:575–9.
68 J. Zhou et al. / Fuel 188 (2017) 61–68
[26] Leonardo RS, Murta Valle ML, Dweck J. An alternative method by pressurized
DSC to evaluate biodiesel antioxidants efficiency. J Therm Anal Calorim
2012;108:751–9.
[27] Murta Valle ML, Leonardo RS, Dweck J. Comparative study of biodiesel
oxidation stability using Rancimat, PetroOXY, and low P-DSC. J Am Oil Chem
Soc 2014;116:113–8.
[28] Pinto LM, de Souza AL, Souza AG, Santos IM, Queiroz N. Comparative
evaluation of the effect of antioxidants added into peanut (arachis hypogae
l.) oil biodiesel by P-DSC and rancimat. J Therm Anal Calorim
2015;120:277–82.
[29] Kurechi T, Aizawa M, Kunugi A. Studies on the antioxidants XVIII: oxidation
product of tertiary butyl hydroquinone (TBHQ) (I). J Am Oil Chem Soc
1983;60:1878–82.
[30] de Guzman R, Tang H, Salley S, Simon Ng KY. Synergistic effects of antioxidants
on the oxidative stability of soybean oil-and poultry fat-based biodiesel. J Am
Oil Chem Soc 2009;86:459–67.
[31] Chiriac R, Toche F, Brylinski C. Biodiesel: characterization by DSC and P-DSC.
In: Auroux A, editor. Calorimetry and thermal methods in
catalysis. Heidelberg, Berlin: Springer; 2013. p. 455–79.
[32] Litwinienko G, Kasprzycka-Guttman T. The influence of some chain-breaking
antioxidants on thermal-oxidative decomposition of linolenic acid. J Therm
Anal Calorim 1998;54:203–10.
[33] Litwinienko G. Analysis of lipid oxidation by differential scanning calorimetry.
In: Kamal-Eldin A, Pokorný J, editors. Analysis of lipid oxidation. AOCS Press;
2005. p. 152–93.