Food Chemistry 154 (2014) 71–77

Contents lists available at ScienceDirect

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

Analytical Methods

Development of an orange juice in-house reference material and its

application to guarantee the quality of vitamin C determination in fruits,
juices and fruit pulps
A. Valente a, A. Sanches-Silva a,b, T.G. Albuquerque a,c, H.S. Costa a,c,⇑
a
Department of Food and Nutrition, National Institute of Health Dr. Ricardo Jorge, I.P. (INSA), Av. Padre Cruz, 1649-016 Lisbon, Portugal
b
Centro de Estudos de Ciência Animal (CECA), University of Oporto, R. D. Manuel II, Apartado 55142, 4051-401 Porto, Portugal
c
REQUIMTE/Faculdade de Farmácia da Universidade do Porto, R. Jorge Viterbo Ferreira 228, 4050-313 Porto, Portugal

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

Article history: Reference materials are useful for the quality control of analytical procedures and to evaluate the perfor-
Received 13 March 2012 mance of laboratories. There are few and expensive certified reference materials commercially available
Received in revised form 6 February 2013 for vitamin C or ascorbic acid analysis in food matrices. In this study, the preparation and the suitability
Accepted 11 December 2013
assessment of an orange juice in-house reference material (RM) for vitamin C analysis in fruits, juices and
Available online 18 December 2013
in fruit pulps is described. This RM was used for the development and full validation of an HPLC method.
The results showed excellent linearity (r2 = 0.9995), good accuracy (96.6–97.3%) and precision, as relative
Keywords:
standard deviation, ranged from 0.70% to 3.67%. The in-house RM was homogenous and stable at storage
Vitamin C
Ascorbic acid
conditions (80 °C) during 12 months. According to our results, this in-house RM is an excellent tool to
Reference material use in quality control and method verification purposes for vitamin C analysis of fruits, juices and fruit
Quality control pulps matrices. Furthermore, a stabilization solution with perchloric and metaphosphoric acids was
Validation developed which prevents degradation of ascorbic acid for a period of 12 months at 80 °C.
Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction
Ascorbic acid (2-(1,2-dihydroxyethyl)-4,5-dihydroxyfuran-3-
one) is one of the most important water-soluble vitamins, naturally
present in fruits and vegetables, and it is widely used as a food addi-
tive and antioxidant (Özyürek, Güçlü, Bektasßoğlu, & Apak, 2007).
Vitamin C has two forms that are biologically active, ascorbic acid
and dehydroascorbic acid (Gokmen, Kahraman, Demir, & Acar,
2000). Ascorbic acid plays a crucial role in several biochemical pro-
cesses in the human body and is reversibly converted into dehydro-
ascorbic acid, under oxidative stress. However, dehydroascorbic
acid is then recycled into ascorbic acid by dehydroascorbate reduc-
tase in different cells such as erythrocytes, smooth muscle cells,
hepatocytes, astrocytes and osteoblasts (Wilson, 2002). A dietary
daily intake of this vitamin lowers the incidence of several chronic
diseases, such as, diabetes, cardiovascular and cancer (Özyürek
et al., 2007). Vitamin C is extremely labile, especially due to the ac-
tion of high temperatures, cations like copper and iron, oxygen,
alkaline pH, light or enzymes (Odriozola-Serrano, Hernández-Jover,
& Martín-Belloso, 2007). However, with appropriate stabilization
procedures and suitable storage conditions (e.g. temperature and
light), ascorbic acid can be stable for long periods of time.
⇑ Corresponding author. Tel.: +351 217519267; fax: +351 217508153.
E-mail address: helena.costa@insa.min-saude.pt (H.S. Costa).
Many analytical methods have been reported for the determi-
nation of vitamin C in foods, including titration, amperometry, cal-
orimetry, voltammetry, chemiluminescence, enzymatic assays,
spectrophotometry, spectrofluorimetric, capillary electrophoresis
and gas chromatography. Most of these methods, are time-con-
suming, need expensive instruments, have a high limit of detection
(LOD) and may give overestimated values due to the presence of
oxidizable species other than ascorbic acid (Gokmen et al., 2000).
High-performance liquid chromatography (HPLC) is a very efficient
method in ascorbic acid analysis of fruits, vegetables or beverages
(Quirós, Fernández-Arias, & López-Hernández, 2009) and it also has
some advantages, such as the increase of selectivity, sensitivity and
the elimination of interferences.
Quality assurance is an important issue in vitamin C analysis.
Reference materials (RMs) play a key role in this field, and can be
used for method validation, estimation of measurement uncer-
tainty, as well as for the verification of a correct use of a method
and quality control (Walker & Lumley, 1999). The use of RMs is
an important step to achieve comparability and traceability be-
tween measurements at national and international level. However,
the availability of RMs in food matrices to fulfil this need is scarce.
For this reason, laboratories try to produce in-house RMs some-
times called ‘‘laboratory reference materials’’ or ‘‘quality reference
materials’’ which are not certified but fulfil the characteristics of
homogeneity and stability fit for the intended use. The use of

0308-8146/$ - see front matter Ó 2014 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.foodchem.2013.12.053
72 A. Valente et al. / Food Chemistry 154 (2014) 71–77
RMs and the regular participation in proficiency testing (PT)
schemes have become fundamental pillars for assuring and con-
trolling the quality of analytical data in terms of precision and
accuracy, thus providing the competence of analytical laboratories
(Emons, Linsinger, & Gawlik, 2004). Commercially available certi-
fied reference materials for vitamin C determination in foods
include ‘‘Brussels sprouts powder’’, ‘‘Whole milk powder’’, and
for ascorbic acid analysis ‘‘Infant/Adult nutritional formula’’ and
‘‘Low calorie cranberry juice cocktail’’. The ‘‘Brussels sprouts pow-
der’’ and ‘‘Whole milk powder’’ materials are certified by Commu-
nity Bureau of Reference and the ‘‘Infant/Adult nutritional formula’’
and ‘‘Low calorie cranberry juice cocktail’’ materials by National
Institute of Standards and Technology.
The aim of this study was to produce and assess the suitability
of a new in-house RM for quality control of the vitamin C analysis
in fruits, juices and fruit pulps. To achieve this goal an HPLC
method was developed and full validated.
2. Materials and methods
2.1. Standards and chemicals
L(+)-ascorbic acid extra pure crystals was from Riedel-de Haën
(Seelze, Germany). Metaphosphoric acid (HPO3) was obtained from
Prolabo (Mollet del Vallés, Barcelona, Spain), Tris-[2-carboxyethyl]
phosphine hydrochloride (TCEPHCl) was acquired from Fluka
(Buchs, Switzerland); while perchloric acid (HClO4), orthophos-
phoric acid (H3PO4) and ammonium dihydrogenphosphate (NH4-
H2PO4) were purchased from Merck (Darmstadt, Germany). All
solutions were prepared with water from a Milli-Q Ultrapure
Water Purification System (Millipore Corp., USA).
2.2. Solutions and calibration
The developed acid solution used to stabilize and precipitate
proteins in the samples was prepared as follows: 10% (v/v) per-
chloric acid and 1% (w/v) metaphosphoric acid in ultrapure water.
With respect to the in-house RM (orange juice), a TCEPHCl solu-
tion (250 lg/mL) was prepared in ultrapure water and used to re-
duce the oxidized forms of vitamin C. A standard stock solution of
ascorbic acid (1 mg/mL) was freshly made on each analytical day.
The working standard solutions were prepared from the stock
solution by appropriate dilution with mobile phase to obtain final
concentrations of ascorbic acid (1, 20, 40, 60, 80 and 100 lg/mL).
2.3. Selection of the candidate to in-house RM
The orange juice matrix was collected from major supermarket
chains in the region of Lisbon. The selected product was, according
to the label, 100% orange juice (juice and pulp) and packages were
randomly chosen and corresponded to different batches. The can-
didate for RM was first used to perform the HPLC method valida-
tion and then was characterized by homogeneity and stability
studies. For method validation, forty packages of orange juice,
330 mL each, were mixed to obtain a sample pool. Homogenization
was carried out under yellow light, protected from daylight. More-
over, the glass material used during sample preparation was
always amber in order to avoid oxidation.
2.4. Sample extraction
Four grams of sample were weighed into a 50 mL polypropylene
tube and 12 mL of the stabilization solution and 32 mL of the
TCEPHCl solution (see Section 2.2) were added. The mixture was
thoroughly vortexed for 1 min. This solution was diluted to a
50 mL volumetric flask with mobile phase. The samples were fil-
tered twice, first with a 150 mm diameter filter paper from Mache-
rey–Nalgel (Macherey–Nalgel GmbH & Co. KG, Germany), and then
filtered through a 0.45 lm PVDF Millipore filter (Millipore Corpo-
ration, Bedford, MA, USA).
2.5. Instrumentation and chromatographic conditions
Separation and quantification of ascorbic acid was performed
on an Alliance 2695 HPLC system (Waters, Milford, MA, USA),
equipped with a Waters 2996 photodiode array detector (PDA),
using Synergi™ Hydro-RP (150  4.6 mm I.D., 4.0 lm particle size)
analytical column from Phenomenex (Torrance, California, USA)
protected with a SecurityGuard Cartridge AQ C18 (40  2.0 mm
I.D., 5 lm particle size) from Phenomenex (Torrance, California,
USA). The detection was achieved at 246 nm and the peak areas
were quantified and processed with an Empower™ version 2.0
software (Waters, Milford, MA, USA). The mobile phase consisted
of 20 mM NH4H2PO4, pH 3.5 (adjusted with H3PO4 85%), and con-
taining 0.015% of m-H3PO4 (w/v). The mobile phase was filtered
through a 0.45 lm GH Polypro membrane Pall filter (Gelman Lab-
oratory, Canada) and then degassed for 30 min. The total run time
of analysis was 5 min at a flow rate of 0.8 mL/min. Column temper-
ature was kept at 30 °C and autosampler at 4 °C. A volume of 30 lL
was injected into the HPLC system.
2.6. Validation
The analytical method was validated according to Food and
Drug Administration (FDA) (Center for Drug Evaluation & Research,
1994) and International Conference on Harmonization guidelines
(ICH, 1997) and the following parameters were determined: selec-
tivity, linearity, LOD, limit of quantification (LOQ), precision, recov-
ery, matrix effect and robustness.
Specificity of the method was assessed by confirming the ab-
sence of any interference at the retention time of the ascorbic acid
peak and by comparing the spectra of ascorbic acid in orange juice
samples with ascorbic acid in pure standard.
Six calibration curves in the range 1–100 lg/mL were per-
formed using six standard concentrations and injecting six times
each. Linearity of calibrations curves were evaluated by linear
regression analysis, plotting peak areas versus standard concentra-
tions. Parameters like regression, coefficient of determination (r2),
slope and intercept were calculated.
According to international guidelines from Center for Drug
Evaluation and Research (1994) and International Conference
Harmonization (1997), the LOD was considered to be the lowest
analyte concentration, which gives a signal-to-noise (S/N) ratio of
at least 3:1. The LOQ was determined as the lowest concentration
which gives a S/N ratio of at least 10:1. The achieved limits were
confirmed by analyses of five independent standard solutions at
LOD and LOQ concentrations.
The intra-assay precision was investigated by injecting in dupli-
cate, six samples from an orange juice pool on the same day. The
inter-assay precision was evaluated on three different days by pre-
paring six samples from the same pool on each day. Mean, stan-
dard deviation (SD) and relative standard deviation (RSD) were
calculated.
The recovery studies were performed by analyzing a sample
pool before and after addition of a known ascorbic acid concentra-
tion. The recoveries were measured after spiking samples with
three independent standard solutions of ascorbic acid (20, 60 and
100 lg/mL) on three different days. The measurements were per-
formed by using five determinations per concentration. The matrix
effect of the method, i.e. the presence of parallelism (r = [0.9–1.0])
was evaluated by the absolute recovery of ascorbic acid in the
 A. Valente et al. / Food Chemistry 154 (2014) 71–77
sample matrix and in the pure standard. The percentage of total
recovery was calculated as the ratio between the slopes of calibra-
tion curves prepared in sample and pure standard:
 
mean of slopes from sample calibration curves
Matrix effect ¼
mean of slopes from pure standard calibration curves
 100:
In addition, the method robustness was investigated by the
introduction of deliberate variations in the flow rate (0.7 or
0.9 mL/min), pH value of the mobile phase (3.4 or 3.6) and column
temperature (29 or 31 °C). These method variations were evalu-
ated as one parameter at a time and simultaneously as part of a
factorial combination experiment.
2.7. Assessment of the suitability of the in-house RM
Generally, the demand for RMs exceeds supply in terms of the
range of materials and availability (ILAC-G9, 2005). It is rare to have
a choice of alternative RMs and the user must choose the most suit-
able material available. When there is a need to prepare an in-house
RM, there are key issues that need to be considered: selection of
material (appropriateness, matrix similarity, cost, availability,
etc.); homogeneity testing; stability testing; establishment of the
assigned values and uncertainty estimation (ILAC-G9, 2005).
2.7.1. Homogeneity study
The homogeneity of orange juice regarding ascorbic acid con-
tent was assessed using twenty packages from two different
batches (10 packages from each batch). The analyses were per-
formed in five random test portions from each package. Each anal-
ysis was performed in one setting under repeated conditions, on
the same day using the same set of reagents and conditions.
2.7.2. Stability study
The candidate for in-house RM was prepared from twenty pack-
ages of orange juice and stored at two different temperatures
(refrigerated at +5 ± 3 °C and frozen at 80 ± 1 °C). The stability
of the material under these conditions was evaluated. The ascorbic
acid short-term stability at +5 °C and 80 °C was evaluated at day
0, on the second day, and also after 1 and 2 weeks of storage. The
long-term stability at 80 °C was evaluated after a period of 4, 6,
9 and 12 months. At each period of time, ascorbic acid analysis in
five samples was performed by HPLC. The ascorbic acid stability
was evaluated by comparing the results obtained at each period
with the levels analyzed at the initial time (day 0). To prove the
stability of the candidate for RM, the RSD must always be lower
than the value obtained in the method inter-assay precision.
2.7.3. Uncertainty estimation
The method combined standard uncertainty (u) was calculated
for three concentrations levels (20, 60 and 100 lg/mL) and for the
assigned value, according to the method precision and accuracy.
The expanded uncertainty (U) was calculated at 95% confidence le-
vel for ascorbic acid analysis. Calculations were as follows:
jqffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffik
u¼ u2precision þ u2accuracy
jqffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffik
U¼2 u2precision þ u2accuracy ;
where, uprecision is the standard uncertainty associated to method
precision and uaccuracy the standard uncertainty from the method
accuracy.
73
2.8. Proficiency testing and applicability of the RM
The fruit juice PT scheme was conducted by the Swedish National
Food Administration (SNFA), round V-6 (Vitamins in foods) (SNFA,
2008). Food Analysis Performance Assessment Scheme (FAPAS), pro-
ficiency tests 2155 (FAPAS, 2009) and 2171 (FAPAS, 2011) were car-
ried out for orange juice and fruit purée, respectively. The laboratory
results for vitamin C were tested and the outliers were eliminated.
The individual laboratory performance was statistically evaluated
by an indicator, known as, z-score that is calculated according to:
ðz ¼ x  X Þ=s; ð1Þ
with x being the result obtained by a laboratory, X the assigned value
and s the standard deviation. Therefore, z-scores are generally inter-
preted as follows: |z| 6 2, satisfactory performance; 2 < |z| < 3, ques-
tionable performance and |z| P 3, unsatisfactory performance. In the
PT scheme conducted by SNFA, the Horrat (Horwitz ratio) value was
also calculated. This value is the ratio between the experimental RSD
from the inter-laboratory test and the one estimated from the Hor-
witz equation. The Horrat value indicates if the result of the interlab-
oratory study has a satisfactory precision. If the value exceeds two,
the precision is questionable, but not necessarily unacceptable.
Successful application of validated methods depends on its cor-
rect use. The in-house RM was used to verify the correct use of the
method and also for quality control in ascorbic acid routine analy-
sis. According to that, the method validated in this study has been
applied to the determination of ascorbic acid in a wide range of
fruits (26 exotic and 11 non-exotic fruits). Briefly, samples were
collected from major supermarket chains in the region of Lisbon.
For each fruit, at least 2 kg of samples were randomly collected.
Multiple fruits of the same variety were combined to have a com-
posite sample. The ascorbic acid analysis of 26 exotic fruits has
been previously described (Valente, Albuquerque, Sanches-Silva,
& Costa, 2011). The edible portion of fruits were homogenized in
a blender (Grindomix GM200, Retsch, Haan, Germany) at
8000 rpm during 2 min. Homogenization was carried out under
yellow light. The analyses were performed in three individual sam-
ples, n = 3, each sample analyzed in triplicate.
2.9. Statistical analysis
Statistical analyses were performed with SPSS for Windows, IBM
SPSS Statistics 18 (SPSS Inc., Chicago). Results are expressed as
mean ± SD or as percentage. For multiple comparisons of normally
distributed data, parametric one-way analysis of variance (ANOVA)
was used. A value of p < 0.05 was considered statistically significant.
3. Results and discussion
3.1. Sample preparation
Fruits are complex samples that contain large amounts of
potentially interfering compounds. Therefore, besides the chro-
matographic method, the choice of the sample preparation proce-
dure is important for the effective quantification of vitamin C
(Phillips et al., 2010).
The sample preparation procedure shall inactive degradative
enzymes which can destroy vitamin C. In line with this, a stabiliza-
tion solution containing 10% (v/v) perchloric and 1% (w/v) meta-
phosphoric acids was used. This acid solution with high ionic
strength suppresses metabolic activity after disruption of the plant
tissues and cells and precipitates proteins. Other acids, like citric,
acetic, oxalic or orthophosphoric acids may be used. For method
optimization, two acidic solutions of 10% (w/v) metaphosphoric
and 5% (w/v) trichloroacetic acids were tested. Our results showed
74 A. Valente et al. / Food Chemistry 154 (2014) 71–77

that the addition of individual solutions of metaphosphoric or tri-

chloroacetic acids originated serious analytical interactions with
C18 silica-based column. In both cases, these interactions resulted
in drifts in the baseline and retention time. The combination of
perchloric and metaphosphoric acids provided an excellent
extraction, stabilization of ascorbic acid and no interference be-
tween metaphosphoric acid and C18 silica-based column was
observed.

3.2. Chromatographic optimization

In order to develop an isocratic HPLC method to quantify

ascorbic acid, a hydrophilic molecule, several columns were tested
and the best results were achieved with Synergi™ Hydro-RP,
which allows the retention of ascorbic acid away from the solvent
front but with reduced time of analysis. After testing different mo-
bile phases with other acids, such has, acetic acid containing
0.015–0.030% (w/v) or trichloroacetic acid containing 0.015–
0.025% (w/v) and other phosphate concentrations (10, 20 and
30 mM), the optimal mobile phase was found to be 20 mM ammo-
nium dihydrogen phosphate, pH 3.5 (adjusted with H3PO4 85%),
containing 0.015% (w/v) of HPO3. Fig. 1 (a) shows a chromatogram
of ascorbic acid standard and (b) shows a typical chromatogram of
an orange juice sample. The retention time of ascorbic acid was
4.2 min.

3.3. Validation

Linear curve fitting was applied to calculate the calibration

curve over the concentration range (1–100 lg/mL). The mean
(±SD) regression equation from six replicate calibration curves,
on different days was: y ¼ 9:96  104 ð1:41  103 Þx  2:65 104
ð1:52  104 Þ. Coefficients of determination (r2) were always equal
or greater than 0.9995, indicating suitability for quantification.
Slope RSD (n = 6) was 1.4%. The ANOVA one-way test was
performed to prove variance homogeneity in the working range.
Results indicate that no statistical differences (p = 0.05) were
achieved between all calibration curves.
The achieved LOD and LOQ were 0.035 and 0.09 lg/mL, respec-
tively. An interesting review on methods to determine ascorbic
Fig. 1. HPLC chromatograms corresponding to (a) standard of ascorbic acid with
acid and dehydroascorbic acid allows the comparison of our results 60 lg/mL, (b) orange juice sample showing total ascorbic acid with 53 lg/mL.
with others (Nováková, Solich, & Solichová, 2008). The LOD ob-
tained in this study was better than the one achieved by other Table 1
authors: 8.29 lg/mL (Pérez, Olias, Espada, Olias, & Sanz, 1997) or Validation data for method precision.
0.05 lg/mL (Nojavan, Khalilian, Kiaie, Rahimi, Arabanian, Chalavi, Precision Ascorbic acid
2008).
Mean (lg/mL) ± SD RSD (%)
Within-assay and between-assay results for ascorbic acid are
presented in Table 1. The intra-assay RSD (n = 6) ranged from Repeatabilitya Day 1 63.83 ± 0.28 0.43
Day 2 68.99 ± 0.41 0.60
0.43% (63.83 ± 0.28 lg/mL) to 0.70% (68.71 ± 0.48 lg/mL). The in- Day 3 68.71 ± 0.48 0.70
ter-assay RSD (n = 18) was 3.67% (67.18 ± 2.47 lg/mL). Intermediate precisionb Between-day 67.18 ± 2.47 3.67
The results of analytical recoveries spiked with 20, 60 and a
Mean values of six individual samples, n = 6, each sample analyzed in duplicate.
100 lg/mL of ascorbic acid are shown in Table 2. The method b
Mean values of eighteen individual samples, n = 18, each sample analyzed in
recoveries for ascorbic acid varied from 96.6% (58.0 ± 1.34 lg/mL) duplicate.
to 97.3% (97.3 ± 1.53 lg/mL). In relation to the matrix effect study,
the method proved to have a total recovery of R = 0.95, showing
that there was a parallelism between calibration curves prepared 3.4. Assessment of the suitability of the in-house RM
in sample matrix and in pure standard.
The robustness is a measure of the capability of the method to 3.4.1. Homogeneity study
remain unaffected by small, but deliberate variations in method The procedure of the IUPAC Technical report (Thompson,
operational parameters. The results of the method robustness are Ellison, & Wood, 2006) was used to perform the homogeneity
presented in Table 3. The effect of the individual and combined study. The orange juice samples were analyzed and the results of
interactions of analytical parameters is not statistically significant the analysis of one-way ANOVA for the homogeneity study showed
(p > 0.05). The method proved to be robust for small variations in that there are no significant differences between samples from
each individual analytical parameter and for the combination of the same batch and from different batches (F-value = 4.72 < critical
the three parameters under study. F-value = 9.55). Therefore, the tested samples of orange juice can
 A. Valente et al. / Food Chemistry 154 (2014) 71–77 75

Table 2
Validation data for method accuracy.
80 a
Ascorbic acid

Ascorbic acid (mg/100 g)

L-Ascorbic acid added (lg/
mL) Mean (lg/ Recovery RSD 60
mL) ± SD (%)a (%)
20 19.3 ± 0.26 96.7 1.32
60 58.0 ± 1.34 96.6 2.31 40
100 97.3 ± 1.53 97.3 1.57
a
Overall mean recovery of ascorbic acid added to fifteen samples = 96.9%.
20

be considered suitable, in terms of homogeneity, as a quality con-

trol material for analysis of ascorbic acid.
0
0 2 4 6 8 10 12 14
3.4.2. Stability study Storage time (days)
Previous stability studies reported in the literature (Urbany &
80
Horti, 1992; Sahari, Boostani, & Hamidi, 2004; Vanderslice, Higgs, b
& Block, 1990) describe that vitamin C content in fruits, juices

Ascorbic acid (mg/100 g)

and vegetables decreases markedly after short periods of time at 75
different storage temperatures (+4, 12, 18, 20, 24, 30,
40 °C). Some of these studies have been performed with and 70
without addition of citric or metaphosphoric acid (Vanderslice
et al., 1990). As far as we know, the stability study described in this
65
paper is the first with addition of perchloric and metaphosphoric
acids at two different storage temperatures.
After verifying that orange juice samples were homogeneous, 60
ascorbic acid was determined after storing samples at different
periods of time and at two different temperatures (+5 and 550
80 °C). Results are shown in Fig. 2. The levels of ascorbic acid were 0 2 4 6 8 10 12 14
uniform over 14 days at 80 °C but not stable at +5 °C, where a deg- Storage time (days)
radation of 80.6% was found. In the long-term stability study, ascor-
80
bic acid was found to be stable over 12 months of storage at 80 °C.
c
Ascorbic acid (mg/100 g)

75
3.4.3. Uncertainty estimation
The expanded uncertainty (with a coverage factor of k = 2) for
ascorbic acid measurement of the in-house RM was 0.23 lg/mL 70
for the 20 lg/mL level; 1.20 lg/mL for the 60 lg/mL level and
1.37 lg/mL for the 100 lg/mL level. The relative expanded uncer- 65
tainty for the assigned value (55.1 lg/mL) of ascorbic acid in the
in-house RM was 2.2%. 60

3.5. Proficiency testing and applicability of the in-house RM 55

0
0 2 4 6 8 10 12
Besides the use of RMs, the participation in PT schemes can also Storage time (months)
improve the quality of analytical data and laboratories are able to
Fig. 2. Stability study during samples storage: (a) and (b) short-term stability study
Table 3 of ascorbic acid in orange juice at +5 °C and 80 °C during 14 days, respectively; (c)
Combination of the analytical parameters variation for robustness by multifactor test. long-term stability of ascorbic acid in orange juice at 80 °C during 12 months.
Uncertainty bars represent the standard deviation of five independent replicate
Test Parametersa L-ascorbic acid
analyses. The dotted lines show the tolerance limits established during the method
(lg/mL)
A B C AB AC BC ABC validation.

1    + + +  59.0
29 3.4 0.7
2 +     + + 58.4
31 3.4 0.7 demonstrate if their performance is satisfactory for the intended
3  +   +  + 56.8 analytical purpose.
29 3.6 0.7
The results of the participation in three PT schemes are shown in
4 + +  +    56.9
31 3.6 0.7 Table 4. Successful participation was obtained for all cases and in the
5   + +   + 56.4 PT scheme conducted by SNFA, the Horrat value was lower than two.
29 3.4 0.9 The in-house RM was used in quality control of the ascorbic acid
6 +  +  +   57.4
routine analysis. Table 5 describes vitamin C content determined in
31 3.4 0.9
7  + +   +  55.6
11 fruits commercialized in Portugal. The gooseberry (44.7 ±
29 3.6 0.9 0.69 mg/100 g of edible portion) was the fruit with the highest con-
8 + + + + + + + 55.7 tent, followed by strawberry fruit with 32.6 ± 0.52 mg/100 g of edi-
31 3.6 0.9 ble portion and raspberry fruit with 12.7 ± 0.67 mg/100 g of edible
a
Parameter A: temperature 29 or 31 °C; Parameter B: pH = 3.4 or 3.6; Parameter portion. The blueberry (0.663 ± 0.06 mg/100 g of edible portion)
C: flow = 0.7 or 0.9 mL/min. and cherry (0.463 ± 0.02 mg/100 g of edible portion) fruits were
76 A. Valente et al. / Food Chemistry 154 (2014) 71–77

Table 4
Participation in proficiency testing schemes.

Sample designation Number of participants Vitamin C

Meana (mg/100 g) Assigned value (mg/100 g) SDb (mg/100 g) RSDc (%) z-score
Fruit purée 61 69.4 69.7 4.16 6.0 0.10
Fruit juice 18 81.1 80.8 6.54 8.1 0.20
Orange juice 22 43.9 45.5 8.71 22.6 0.30
a
Each sample was analyzed in triplicate.
b
SD, standard deviation.
c
RSD, relative standard deviation.

Table 5
Vitamin C content in 11 fruits commercialized in Portugal.

Common Species Family Origin Vitamin C (mg/100 g of edible portion)

name
Meana SD Literature
Mean or Reference
Range
Blueberry Vaccinium corymbosum L. Ericaceae Spain 0.663 0.06 ND Häkkinen, Kärenlampi, Heinonen, Mykkänen, and
Mykkänen (1999)
Cherry Prunus cerasus L. Rosaceae Portugal 0.463 0.02 6.00 Swiss Food Composition Database (2012), TCA
(2006)
Gooseberry Ribes uva-crispa L. Grossulariaceae Portugal 44.7 0.69 33.0 DTU (2009)
Medlar Eriobotrya japonica (Thunb.) Lindl. Rosaceae Spain 1.04 0.06 2.00 RedBEDCA (2010)
Mulberry Morus nigra L. Moraceae Spain NDb – 0.28–2.30 Koyuncu, Koyuncu, Yıldırım, and Vural (2004)
Nectarine Prunus persica var. nucipersica Rosaceae Spain 0.740 0.03 5.40 USDA National Nutrient Database (2012)
(Suckow) C. Schneider
Persimmon Diospyros kaki L. Ebenaceae Spain 2.47 0.03 3.00 TCA (2006)
Persimmon Diospyros kaki L. Ebenaceae Portugal 2.38 0.03 3.00 TCA (2006)
Pomegranate Punica granatum L. Lythraceae Chile 1.57 0.07 6.10 DTU (2009)
Raspberry Rubus idaeus L. Rosaceae Portugal 12.7 0.67 15.4–32.0 Haffner, Rosenfeld, Skrede, and Wang (2002)
Strawberry Fragaria  ananassa Duchesne ex Rosaceae Spain 32.6 0.52 32.4–66.9 Hakala, Lapveteläin, Huopalahti, Kallio and
Rozier Tahvonen (2003)
a
Mean values of three individual samples, n = 3, each sample analyzed in triplicate.
b
ND = Not detected (LOD < 0.042 mg/100 g).

the ones with the lowest content of vitamin C. The analytical results
obtained for vitamin C content in gooseberry, persimmon, rasp-
berry and strawberry fruits were close to literature data, as is
shown in Table 5. The analytical results obtained for cherry, nectar-
ine, medlar, mulberry and pomegranate fruits were lower than the
values reported in literature. However, these differences can be due
to several factors, such as cultivars, seasons, country of origin (soil,
climate, and cultivation techniques), harvest, postharvest, analyti-
cal methods for quantification of vitamin C. Dietary reference
intake (DRI) for vitamin C was established by the Food and
Nutrition Board of the Institute of Medicine, National Academy of
Sciences (2002). Values given for healthy males and females,
P19 years, are 90 and 75 mg/day, respectively. From all the ana-
lyzed fruits, the best sources of vitamin C are gooseberry and straw-
berry. Considering a mean daily consumption of 100 g/day, the
gooseberry fruit contributes with 49.7% of the DRI in men and
59.6% in women. The daily consumption of strawberry fruit contrib-
utes as well to an adequate ingestion of vitamin C, representing
36.2% of the DRI in men and 43.5% in women. The method devel-
oped in this study was also successfully applied to soft drinks with
several flavours, such as, mango, orange, passion fruit, pineapple
and apple. From the five types of the analyzed soft drinks, mango
juice without sugar addition had the highest content of vitamin C
(14.1 ± 0.02 mg/100 mL), followed by the soft drink with 8% of or-
ange juice and with 11% of fruit juice (orange, passion fruit and
mango) that had an vitamin C content of 12.0 ± 0.03 mg/100 mL.
For soft drinks with 22% of apple juice and with 12% of passion fruit
the vitamin C content was 8.15 ± 0.01 and 7.62 ± 0.03 mg/100 mL,
respectively. Two types of carbonated soft drinks were also ana-
lyzed for vitamin C content. The values ranged between
7.80 ± 0.01 mg/100 mL (soft drink with 22% of apple juice) and
3.95 ± 0.04 mg/100 mL (soft drink with 1% of orange juice).
4. Conclusion
An isocratic HPLC method for the determination of vitamin C
has been validated and successfully applied to several food matri-
ces (juices, fruits and fruit pulps). This analytical method has
significant practical advantages (sensitive, rapid, easy to handle,
low cost and free of interferences) and is essential to minimize er-
ror and maximize precision of data that could be used to produce
new high quality data for food composition databases. With our
study, stabilization of ascorbic acid in food samples was achieved
by adding a newly developed solution of perchloric and metaphos-
phoric acids. With the combination of these acids, samples were
stabilized and then stored at 80 °C, assuring that the loss of ascor-
bic acid was never higher than the inter-day precision determined
during validation of the HPLC method. This is an advantage in
ascorbic acid analysis since it minimizes the degree of nutrient loss
during the frozen storage in the laboratory, especially since it facil-
itates practical storage and can be used in large scale studies.
In this study, we developed an in-house RM for vitamin C anal-
ysis that confirmed to be suitable for its intended purpose. The in-
house RM proved to be sufficiently homogeneous and stable to be
used for the assessment of a measurement method, measurement
uncertainty, quality control and for assigning values to material. In
the future, we intend to produce new in-house RMs for other food
matrices, such as, vegetables. We also intend to continue this study
 A. Valente et al. / Food Chemistry 154 (2014) 71–77
by establishing and applying a protocol to produce a standard RM
or a certified RM using this in-house RM.
Acknowledgements
The authors are grateful to the postdoctoral contract of Ana
Sanches Silva in the frame of the Program ‘‘Science 2007’’ funded
by ‘‘Fundação para a Ciência e a Tecnologia’’, Portugal. Tânia Gonç-
alves Albuquerque is grateful for research grant (BRJ/DAN-2012)
funded by National Institute of Health Dr. Ricardo Jorge, I.P.
References
Center for Drug Evaluation and Research. (1994). Reviewer guidance: validation of
chromatographic methods. Rockville: US FDA Federal Register.
DTU Food National Food Institute – Technical University of Denmark (DTU). (2009).
Danish food composition database. Available from http://www.foodcomp.dk/
v7/fcdb_default.asp. Accessed 16.01.13.
Emons, H., Linsinger, T. P. J., & Gawlik, B. M. (2004). Reference materials:
terminology and use. Can’t one see the forest for the trees? Trends in
Analytical Chemistry, 23, 442–449.
FAPAS (Food Analysis Performance Assessment Scheme). (2009). Vitamin C in
orange juice. Available from http://www.fapas.com/proficiency-testing-
schemes/. Accessed 10.01.13.
FAPAS (Food Analysis Performance Assessment Scheme). (2011). Vitamin C in fruit
purée. Available from http://www.fapas.com/proficiency-testing-schemes/.
Accessed 10.01.13.
Food and Nutrition Board, Institute of Medicine, National Academies. (2002).
Dietary reference intakes (DRIs): estimated average requirements for groups.
Available from http://iom.edu/Activities/Nutrition/SummaryDRIs/~/media/
Files/Activity%20Files/Nutrition/DRIs/
RDA%20and%20AIs_Vitamin%20and%20Elements.pdf. Accessed 15.01.13.
Gokmen, V., Kahraman, N., Demir, N., & Acar, J. (2000). Enzymatically validated
liquid chromatographic method for the determination of ascorbic and
dehydroascorbic acids in fruit and vegetables. Journal of Chromatography A,
881, 309–316.
Haffner, K., Rosenfeld, H. J., Skrede, G., & Wang, L. (2002). Quality of red raspberry
Rubus idaeus L. cultivars after storage in controlled and normal atmospheres.
Postharvest Biology and Technology, 24, 279–289.
Hakala, M., Lapveteläinen, A., Huopalahti, R., Kallio, H., & Tahvonen, R. (2003).
Effects of varieties and cultivation conditions on the composition of
strawberries. Journal of Food Composition and Analysis, 16, 67–80.
Häkkinen, S. H., Kärenlampi, S. O., Heinonen, I. M., Mykkänen, H. M., & Mykkänen, A.
R. (1999). Content of the flavonols quercetin, myricetin, and kaempferol in 25
edible berries. Journal of Agricultural and Food Chemistry, 47, 284–290.
ILAC-G9 (ILAC Accreditation Committee in association with the ILAC Laboratory
Committee). (2005). Guidelines for the selection and use of reference materials.
Available from http://www.ilac.org/guidanceseries.html. Accessed 23.01.12.
ICH (International Conference on Harmonization). (1997). Guidance for industry Q2b:
validation of analytical procedures: methodology. Rockville: US FDA Federal
Register.
Koyuncu, F., Koyuncu, M. A., Yıldırım, F., & Vural, E. (2004). Evaluation of black
mulberry (Morus nigra L.) genotypes from Lakes Region, Turkey. European
Journal of Horticultural Science, 69, 125–131.
Nojavan, S., Khalilian, F., Kiaie, F. M., Rahimi, A., Arabanian, A., & Chalavi, S. (2008).
Extraction and quantitative determination of ascorbic acid during different
77
maturity stages of Rosa canina L. fruit. Journal of Food Composition and Analysis,
21, 300–305.
Nováková, L., Solich, P., & Solichová, D. (2008). HPLC methods for simultaneous
determination of ascorbic and dehydroascorbic acids. Trends in Analytical
Chemistry, 27, 942–958.
Odriozola-Serrano, I., Hernández-Jover, T., & Martín-Belloso, O. (2007). Comparative
evaluation of UV-HPLC methods and reducing agents to determine vitamin C in
fruits. Food Chemistry, 105, 1151–1158.
Özyürek, M., Güçlü, K., Bektasßoğlu, B., & Apak, R. (2007). Spectrophotometric
determination of ascorbic acid by the modified CUPRAC method with
extractive separation of flavonoids–La(III) complexes. Analytica Chimica Acta,
588, 88–95.
Pérez, A. G., Olias, R., Espada, J., Olias, J. M., & Sanz, C. (1997). Rapid determination of
sugars, nonvolatile acids, and ascorbic acid in strawberry and other fruits.
Journal of Agricultural and Food Chemistry, 45, 3545–3549.
Phillips, K. M., Tarragó-Trani, M. T., Gebhardt, S. E., Exler, J., Patterson, K. Y.,
Haytowitz, D. B., et al. (2010). Stability of vitamin C in frozen raw fruit and
vegetable homogenates. Journal of Food Composition and Analysis, 23,
253–259.
Quirós, A. R., Fernández-Arias, M., & López-Hernández, J. (2009). A screening
method for the determination of ascorbic acid in fruit juices and soft drinks.
Food Chemistry, 116, 509–512.
RedBEDCA, Ministerio de Ciencia e Innovación. (2010). Base de Datos Española de
Composición de Alimentos. Available from http://www.bedca.net/. Accessed
16.01.13.
Sahari, M. A., Boostani, M., & Hamidi, Z. (2004). Effect of low temperature on the
ascorbic acid content and quality characteristics of frozen strawberry. Food
Chemistry, 86, 357–363.
SNFA (Swedish National Food Administration). (2008). Proficiency testing – food
chemistry vitamins in food – round V-6. Available from http://www.slv.se/
upload/dokument/rapporter/Kvalitetssakring_lab/2008/2008_22_National_Food_
Administration_proficiency_testing_eng.pdf. Accessed 10.01.13.
Swiss Federal Institute of Technology Zurich ETH. (2012). Swiss food composition
database. Available from http://naehrwertdaten.ch/request?xml=Message
Data&xml=MetaData&xsl=Start&lan=en&pageKey=Start. Accessed 18.01.13.
TCA (2006). Tabela da Composição de Alimentos. (1st ed.). Lisboa: Centro de
Segurança Alimentar e Nutrição, Instituto Nacional de Saúde Doutor Ricardo
Jorge.
Thompson, M., Ellison, S. R. S., & Wood, R. (2006). The international harmonized
protocol for the proficiency testing of analytical chemistry laboratories. Pure
and Applied Chemistry, 78, 145–196.
Urbany, G. Y., & Horti, K. (1992). Changes of surface colour of the fruit and of the
anthocyanin content of sour cherries during frozen storage. Acta Alimentaria, 21,
3–4.
U.S. Department of Agriculture, Agricultural Research Service. (2012). USDA
national nutrient database for standard reference, release 25. Nutrient data
laboratory. Available from http://www.ars.usda.gov/ba/bhnrc/ndl. Accessed
18.01.13.
Valente, A., Albuquerque, T. G., Sanches-Silva, A., & Costa, H. S. (2011). Ascorbic acid
content in exotic fruits: A contribution to produce quality data for food
composition databases. Food Research International, 44, 2237–2242.
Vanderslice, J. T., Higgs, D. J., & Block, G. (1990). Ascorbic acid and dehydroascorbic
acid content of foods-as-eaten. Journal of Food Composition and Analysis, 3,
105–118.
Walker, R., & Lumley, I. (1999). Pitfalls in terminology and use of reference
materials. Trends in Analytical Chemistry, 18, 594–616.
Wilson, J. X. (2002). The physiological role of dehydroascorbic acid. FEBS Letters, 527,
5–9.