Physicochemical and Phytochemical

Properties of Cold and Hot Water Extraction

from Hibiscus sabdariffa
C: Food Chemistry

Milena M. Ramirez-Rodrigues, Maria L. Plaza, Alberto Azeredo, Murat O. Balaban, and Maurice R. Marshall

Abstract: Hibiscus cold (25 C) and hot (90 C) water extracts were prepared in various timetemperature combinations

to determine equivalent extraction conditions regarding their physicochemical and phytochemical properties. Equivalent
anthocyanins concentration was obtained at 25 C for 240 min and 90 C for 16 min. Total phenolics were better
extracted with hot water that also resulted in a higher antioxidant capacity in these extracts. Similar polyphenolic
profiles were observed between fresh and dried hibiscus extracts. Hibiscus acid and 2 derivatives were found in all
extracts. Hydroxybenzoic acids, caffeoylquinic acids, flavonols, and anthocyanins constituted the polyphenolic compounds
identified in hibiscus extracts. Two major anthocyanins were found in both cold and hot extracts: delphynidin-3sambubioside and cyanidin-3-sambubioside. In general, both cold and hot extractions yielded similar phytochemical
properties; however, under cold extraction, color degradation was significantly lower and extraction times were 15-fold
longer.
Keywords: anthocyanins, extraction, Hibiscus sabdariffa, phytochemicals, water

Practical Application: Hibiscus beverages are prepared from fresh or dried calyces by a hot extraction and pasteurized,

which can change organoleptic, nutritional, and color attributes. Nonthermal technologies such as dense phase carbon
dioxide may maintain their fresh-like color, flavor, and nutrients. This research compares the physicochemical and
phytochemical changes resulting from a cold and hot extraction of fresh and dried hibiscus calyces and adds to the
knowledge of work done on color, quality attributes, and antioxidant capacity of unique tropical products. In addition,
the research shows how these changes could lead to alternative nonthermal processes for hibiscus.

Introduction
Hibiscus sabdariffa L (family Malvaceae) is a tropical annual shrub.
China, Thailand, Mexico, Egypt, Senegal, and Tanzania are among
the main producing countries. In Mexico, this plant is known as
flor de jamaica or simply jamaica. The red calyces are the part
of the plant with commercial interest and are rich in organic acids,
minerals, anthocyanins, and other phenolic compounds (Morton
1987).
Hibiscus extracts contain 2 major anthocyanins; delphinidin-3sambubioside (D3S) and cyanidin-3-sambubioside (C3S). Their
spectral characteristics (Degenhardt and others 2000), mass spectrometry (MS) fragmentation patterns (Giusti and others 1999),
and potential antioxidant (Wang and others 2000) and anticancer
(Chang and others 2005; Hou and others 2005) activities have
been studied. Similarly, other polyphenolic compounds, including protocatechuic acid (Lee and others 2002; Olvera-Garca and
others 2008), hibiscus acid, and its 6-methyl ester (Hansawasdi

MS 20101165 Submitted 10/14/2010, Accepted 1/19/2011. Authors RamirezRodrigues, Plaza, Azeredo, and Marshall are with Food Science and Human Nutrition
Dept., Univ. of Florida, P.O. Box 110370, Gainesville, FL 32611, U.S.A. Author
Balaban is with Fishery Industrial Technology Center, Univ. of Alaska Fairbanks,
118 Trident Way, Kodiak, AK 99615, U.S.A. Direct inquiries to author Marshall
(E-mail: martym@ufl.edu).

C428

Journal of Food Science r Vol. 76, Nr. 3, 2011

and others 2000), have also been found to be present in hibiscus extracts and have been associated with pharmacological
activities.
Different extractions methods can be applied to obtain hibiscus extracts. The variety and extraction conditions used (type
of solvent, concentration, time, and temperature) can potentially
affect the polyphenolic profile of the extracts and thus makes
comparisons between studies difficult (Prenesti and others 2007:
Segura-Carretero and others 2008).
Traditionally, fresh hibiscus is either frozen or dried in the sun for
preservation and used in the production of natural color, flavor extracts, and/or beverages (Morton 1987). Preparation of a hibiscus
beverage includes an extraction step followed by pasteurization.
The use of nonthermal technologies such as dense phase carbon
dioxide (DPCD), pulsed ultraviolet (UV) light, high hydrostatic
pressure, and pulsed electric fields as a preservation method does
not justify an extraction step that involves high temperature, and
an alternative extraction at a lower temperature should be considered for these nonthermal processes (Del Pozo-Insfran and others
2006).
The objectives of this study were (1) to compare the effects of
cold (25 C) and hot (90 C) water extraction on the physicochemical and phytochemical properties of hibiscus extracts and (2)
to identify and quantify the anthocyanins and major polyphenolics present in extracts obtained from fresh and dried hibiscus by
equivalent cold and hot water extraction conditions.
R

C 2011 Institute of Food Technologists
doi: 10.1111/j.1750-3841.2011.02091.x

Further reproduction without permission is prohibited

Properties of Hibiscus sabdariffa extracts . . .

Chemicals and standards

Commercial standards of chlorogenic acid, gallic acid, protocatechuic acid, and quercetin were purchased from SigmaAldrich (St. Louis, Mo., U.S.A.). Delphinidin-3-glucoside
and cyanidin-3-glucoside were purchased from Polyphenols Laboratories AS (Sandnes, Norway). AAPH (2,2 azobis(2-methylpropionamidine) dihydrochloride), fluorescein
(free acid), Trolox (6-hydroxy-2,5,7,8-tetramethylchroman-2carboxylic acid), and FolinCiocalteus reagent were purchased
from Sigma-Aldrich.

U.S.A.) was used to control temperature. All the obtained extracts

were filtered under vacuum (Whatman nr 4 filter paper), and their
physicochemical and phytochemical properties were measured.

pH, total solids, and titratable acidity

pH was measured using an Orion EA920 pH meter (Orion
Research, Boston, Mass., U.S.A.) as described by Ferrentino and
others (2009). For total solids (TS) determination, a 3-g sample
was placed in an aluminum plate and dried at 105 C for 24 h
in an oven (Precision Scientific) at atmospheric pressure. Sample weights were recorded before and after drying, and the TS
content was calculated by difference in weight. A Brinkmann pH
stat (Brinkmann Instruments Co., Westbury, N.Y., U.S.A.) conExtracts preparation
sisting of a Metrohm 655 Disomat, Metrohm 614 Impulsomat,
Fresh and sun dried H. sabdariffa (cv. Criollo) were obtained and Metrohm 632 pH meter was used to measure titratable acidfrom Puebla, Mexico. Hibiscus samples were stored in glass jars, ity (TA). Samples of 10 mL were used, and TA was determined
flushed with nitrogen, and kept frozen at 22 C until used.
by titration with 0.1 N NaOH until pH 8.1 and expressed as
Preliminary experiments were designed to gain more insight in percentage of malic acid (g/100 mL).
the hibiscus extraction process. In these experiments, sun dried
hibiscus was used to determine the hibiscus to water ratio (1:5 Color, color density, and hue tint
to 1:80), and it was found that a ratio of 1:40 was suitable for a
Color was measured using a ColorQuest XE colorimeter
beverage with similar color intensity as compared to commercial (HunterLab, Reston, Va., U.S.A.). Samples (40 mL) were placed in
products. Three different extraction temperatures (25, 60, and a 20-mm cell and L (lightness), a (redness), and b (yellowness)
90 C) were then applied over a wide range of extraction times parameters were recorded in total transmittance mode, illuminant
(6, 12, and 18 h for 25 C; 2, 4, and 6 h for 60 C; and 1, 2, and D65, and 10 observer angle. Color density (CD) and hue tint
3 h for 90 C). Experimental results showed that concentration (HT) were determined by measuring the absorbance (A) at 420,
of anthocyanins was almost the same between 6 and 12 h and 520, and 700 nm for samples (200 L) using a spectrophotometer
increased only 10% at 18 h at 25 C, was almost the same for 2, 4, SpectraMax 190 (Molecular Devices, Sunnyvale, Calif., U.S.A.)
and 6 h at 60 C and was around 4% lower than the 18 h extraction, and calculated as
for 90 C. The 1-h extract was 6% lower than the 18 h; while at
2 and 3 h, there was a decrease of 32% in anthocyanins content
Color density = [(A420 nm A700nm)
(1)
due to thermal degradation. From these results, it was noted that
+ (A520nm A700 nm)],
6, 12, and 18 h at 25 C; 2, 4, and 6 h at 60 C; and 1 h at 90 C
resulted in very similar anthocyanins concentration, but since our
main objective was to evaluate nonthermal extraction for DPCD
Hue tint = (A420nm A700nm)/(A520nm A700nm), (2)
processing, it was decided not to consider the 60 C treatments

and use the 25 C as an alternative method to the 90 C (which

as described by Giusti and Wrolstad (2005).
was used as a reference). Other experiments showed that stirring
applied during extraction and size reduction helped improve the
extraction yield for the 25 C, but no major improvement was Anthocyanin content, total phenolics, and antioxidant
seen at 90 C. Based on these results, experiments were designed capacity
Anthocyanin content was determined by the pH differential
to consider reducing the extraction times that would suit a DPCD
method
(A510nm and A700nm at pH 1.0 and 4.5, dilution factor (DF)
industrial process while maintaing similar equivalent extraction
of 4) and expressed in mg/L of delphinidin-3-glucoside (MW =
conditions at 25 and 90 C.
From the preliminary data above, 2 experiments were con- 465.2, = 23700) as described by Giusti and Wrolstad (2005).
ducted. First, dried hibiscus was mixed with distilled water at a Total phenolics were measured using the FolinCiocalteu assay
ratio of 1:40 (w/v) and maintained at 25 C (cold extraction [CE]) (A765nm , DF of 4) and quantified as gallic acid equivalents (mg/L)
or 90 C (hot extraction [HE]) for 4 different times (30, 60, 120, (Waterhouse 2005). Absorbance measurements for anthocyanin
and 240 min for CE and 2, 4, 8, and 16 min for HE). Eight content and total phenolics were made using a SpectraMax 190
spectrophotometer (Molecular Devices).
treatments were tested.
Antioxidant capacity was evaluated using the oxygen radical
Second, dried calyces, a common preservation step before extracting and processing, were compared with fresh calyces; 4 ex- absorbance capacity assay, and results were expressed as Trolox
tracts with equivalent anthocyanins concentration were prepared equivalents (TEs) per milliliter (mol of TE/mL) as described
using cold and hot water extraction conditions. Fresh and dried by Huang and others (2002) using a SpectraMax Gemini XPS
hibiscus were mixed with distilled water at a ratio of 1:4 and microplate sprectrofluorometer (Molecular Devices). Data were
1:40 (w/v), respectively, and extracted at both 25 C for 240 min acquired and analyzed using SoftMax Pro 5.2 software (Molecular
and 90 C for 16 min. For CE experiments, temperature was Devices).
controlled using a Constant Temperature Circulator Bath, Model
900 (Fisher Scientific, Pittsburg, Pa., U.S.A.) and stirring was ap- Characterization of major polyphenolics
Equivalent cold (25 C for 240 min) and hot (90 C for 16 min)
plied using a Corning stirrer plate Model PC-353 (Lowell, Mass.,
U.S.A.) at speed number 4. For HE, a Microprocessor Controlled water extraction conditions for anthocyanins extraction were seWater Bath, Series 280 (Precision Scientific, Winchester, Va., lected based on findings from the initial part of this study. Four
Vol. 76, Nr. 3, 2011 r Journal of Food Science C429

C: Food Chemistry

Materials and Methods

Properties of Hibiscus sabdariffa extracts . . .

hibiscus extracts were prepared: dried hibiscus cold water extract,
dried hibiscus hot water extract, fresh hibiscus cold water extract,
and fresh hibiscus hot water extract. LC-MS and HPLC analysis were performed in order to identify the major polyphenolic
compounds including anthocyanins present in these extracts.

C: Food Chemistry

LC-MS identification
Chromatographic analyses were performed on an Agilent 1200
series HPLC (Agilent, Palo Alto, Calif., U.S.A.) equipped with
an autosampler/injector and diode array detector. A Dionex C18
5 m 120A column (250 4.6 mm) was used for compound separation (Dionex, Sunnyvale, Calif., U.S.A.). Mobile phases consisted of water (phase A) and 60% methanol in water (phase B),
both adjusted to pH 2.4 with formic acid. A gradient solvent program ran phase B from 0% to 60% in 20 min, 60% to 100% in 20
min, 100% for 7 min, 100% to 0% in 3 min and final conditions
were held for 2 min (Pacheco-Palencia and others 2007). The flow
rate was 0.8 mL/min, and detection was done at 260, 280, 320,
360, and 520 nm.
Electrospray ionization MS was performed with a high capacity
series ion trap (HCT) mass spectrometer (Bruker Daltonics, Billerica, Mass., U.S.A.). Column effluent was monitored in positive
and negative ion mode of the MS in an alternative manner during
the same run. Other experimental conditions on the mass spectrometer were as follows: nebulizer = 0.31 MPa, dry gas (nitrogen)
= 11.0 L/min, dry temperature 350 C, ion trap = scan from m/z
90 to 1000, smart parameter setting = compound stability = 50%,
trap drive level = 60%. The mass spectrometer was operated in
Auto MS2 mode. MS2 was used to capture and fragment the most
abundant ion in full scan mass spectra.
Polyphenolics were identified by comparison of UV/vis (190
to 660 nm) spectral interpretation, retention time, comparison to
standards, and MS-fragmentation patterns.

HPLC quantification
Anthocaynins and polyphenolics were quantified using a
Dionex HPLC system equipped with an autosampler/injector and
diode array (PDA 100) detector (Dionex). Compounds were separated on a 250 4.6 mm Dionex C18 5 m 120A column
(Dionex). Mobile phases consisted of water (phase A) and 60%
methanol in water (phase B), both adjusted to pH 2.4 with ophosphoric acid. A gradient solvent program ran phase B from 0%
to 60% in 20 min, 60% to 100% in 20 min, 100% for 7 min, 100%
to 0% in 3 min and final conditions were held for 2 min. The flow
rate was 0.8 mL/min, and detection was done at 260, 280, 320,
360, and 520 nm.

Statistical analysis
Each extraction condition (time + temperature) was repeated
in triplicate. Analysis of variance (ANOVA) and mean separation
using Tukeys test ( = 0.05) were performed to evaluate the
differences between extraction times, temperatures, and treatments
using SAS 9.0 Statistical software (SAS Inst. Inc., Cary, N.C.,
U.S.A.).

Results and Discussion

Effect of extraction conditions
CD, anthocyanins content, total phenolics, and antioxidant capacity increased with increasing time for both extraction temperatures (25 and 90 C), while L values decreased with time in both
cases (Table 1). There were no significant differences between the
first 2 (30 and 60 min, and 2 and 4 min) and last 2 (120 and
240 min, and 8 and 16 min) extraction times at both temperatures for HT values; although the latter time values were higher
(Table 1).
For cold water extraction (25 C), time had a significant effect
(p < 0.0001) in all the parameters measured but pH. TS and b
increased from 30 to 60 min and then remained constant (measurements at times 60, 120, and 240 min were not significantly
different). TA increased from 30 to 120 min and remained constant
at 240 min (measurements at 120 and 240 min were not significantly different), while a values increased from 30 to 60 min,
remained constant from 60 to 120 min, and decreased at 240 min
(Table 1). For hot water extraction (90 C), time had a significant
effect (p < 0.0335) in all the measured parameters. pH increased
from 2 to 4 min and remained constant until 16 min. TS and
TA increased until 8 min and remained constant until 16 min. a
values were constant for times 2 and 4 min and decreased at 8
and 16 min, while b values were constant at 2, 4 and 8 min and
decreased at 16 min (Table 1).
There was a significant effect (p < 0.0001) of treatment conditions (temperature + time) in all the measured parameters but
pH. Treatments CE for 30 min and HE for 2 min were equivalent
in TS and TA (Table 1), while treatments CE for 60 min and HE
for 4 min were equivalent in TS, TA b , CD, and anthocyanins
content (Figure 2). Treatments CE for 120 min and HE for 8 min
were equivalent in TS, TA, and anthocyanins content, while treatments CE for 240 min and HE for 16 min were equivalent in TS,
TA, CD, and anthocyanins content.
L values were significantly lower (darker color) in hot water
extracts as compared to cold water ones, while a values were
slightly higher in the cold water extracts (Table 1). A change of
1 L unit will result in a change of 1 E (difference in color)
unit, at least. Around 3 E units is an easily perceptible change

Table 1Measured pH, total solids (TS), titratable acidity (TA), and color (L , a , b values, color density [CD], and hue tint [HT])
for the extracts.
Extract
CE30
CE60
CE120
CE240
HE2
HE4
HE8
HE16

T ( C)
25
25
25
25
90
90
90
90

Time (min)
30
60
120
240
2
4
8
16

pH
a

2.37
2.32a
2.32a
2.31a
2.37a
2.37a
2.36a
2.33a

TS
d

0.68
0.92bc
0.97ab
1.00ab
0.79cd
0.90bc
0.95ab
1.08a

TA
d

0.28
0.38abc
0.40ab
0.44a
0.33cd
0.37bc
0.39ab
0.43a

a
a

54.18
46.29b
43.78cd
40.79ef
44.82bc
42.13de
39.34f
35.26g

b
d

65.85
67.65a
67.50a
67.16ab
66.73bc
66.39c
65.65d
63.93e

CD
e

45.39
66.92abc
68.76a
68.22ab
65.19c
67.03abc
65.70bc
60.29d

1.04
1.82de
2.05c
2.55ab
1.73e
2.00cd
2.34b
2.70a

HT
0.35cd
0.35d
0.36c
0.36c
0.38b
0.38b
0.39a
0.39a

CE = cold extraction, HE = hot extraction. Data represent the mean of n = 9. Values with similar letters within columns are not significantly different (Tukeys test, p > 0.05).

Expressed as g of solids/100 mL of extract.

Expressed as g of malic acid/100 mL of extract.

C430 Journal of Food Science r Vol. 76, Nr. 3, 2011

for anthocyanins were found. As can be seen from Figure 2, total

phenolics were better extracted with hot water (90 C) than with
cold water (25 C). Prenesti and others (2007) also found that
hot water (100 C for 3 min) extracted a higher phenolic content
compared to cold water hibiscus extracts. The higher concentration of polyphenolic compounds other than anthocyanins in
hot water extracts may have contributed to a higher antioxidant
activity in these extracts as compared to cold water extracts (Figure 2). Tsai and others (2002) found that hibiscus anthocyanins
contributed to 51% of total antioxidant capacity and that other
phenolic compounds were responsible for the remaining activity.
Qualitative differences were observed between the cold and hot
water hibiscus extracts. Cold extracts had a clear appearance and
bright red color, whereas hot extracts presented a more opaque red

in color by the human eye (Kim and others 2002). CD values

were significantly higher in hot water extracts, which is associated
with a darker color. HT is a measurement of color degradation in
anthocyanin containing products. From Table 1, it can be observed
that the extracts obtained with cold water have lower HT values
than those obtained with hot water. This indicates that temperature
had an effect on hibiscus extracts color and, thus, anthocyanins.
A higher HT value is associated with an increase in absorbance at
420 nm (yellow tones) in relation to that at 520 nm (red tones); this
is undesirable because it is an indication of anthocyanins degrading.
Anthocyanin content was not significantly different between
treatments CE for 60 min and HE for 4 min, CE for 120 min
and HE for 8 min, and CE for 240 min and HE for 16 min
(Figure 2), so equivalent cold and hot water extraction conditions

H
COOH
COOR1
R2

Compound name
hibiscus acid
hibiscus acid glucoside

Figure 1Structures of the phenolic compounds

identified in this study.

R1
H
H

hibiscus acid 6-methyl ester

CH3

COOH

R3

Compound name
gallic acid
protocatechuic acid glucoside

R1

R1
OH
O-glu

R2

COOH

HO

R1

R3

R2

Compound name
OH quinic acid
3-caffeoylquinic acid
4-caffeoylquinic acid
OH 5-caffeoylquinic acid

-O
caffeoyl

R1
OH
caffeoyl
OH
OH

OH
OH
O

HO

R1

Compound name
dpd-3-sambubioside
cyd-3-sambubioside

R1
OH
H

O-xyl-glu
OH

OH
O

HO

OH
O-rha-glu

OH

O
quercetin-3-rutinoside

Vol. 76, Nr. 3, 2011 r Journal of Food Science C431

C: Food Chemistry

Properties of Hibiscus sabdariffa extracts . . .

Properties of Hibiscus sabdariffa extracts . . .

C: Food Chemistry

color and some haze. These differences were possibly associated identified as 5-CQA by comparison with an authentic standard.
with higher concentrations of phenolic compounds other than According to Clifford and others (2003), 5-CQA is characterized
by an intense base peak at m/z 191 and a weak secondary ion at
anthocyanins in the hot water extracts.
m/z 179. Peak 6 was identified as 3-CQA, since it is characterized
Polyphenolics identification
by a base peak at m/z 191 and a relatively intense secondary ion
Anthocyanins and other polyphenolics present in hibiscus ex- at m/z 179, while peak 8 was identified as 4-caffeoylquinic (4tracts were identified on the basis of their retention time, absorp- CQA) acid with a characteristic base peak at m/z 173 (Clifford
tion spectrum, MS-fragmentation pattern, and where possible by and others 2003). Peak 10 was tentatively identified as a CQA
comparison to an authentic standard. The chemical structures of isomer from its absorption spectrum and fragmentation patterns
(Table 3). The presence of 5-CQA has been previously reported in
the compounds identified in this study are shown in Figure 1.
Peaks 1 (tR = 4.7 min, max = 265 nm), 2 (tR = 7.4 min, hibiscus extracts (Mourtzinos and others 2008; Segura-Carretero
max = 262 nm), and 3 (tR = 10.2 min, max = 263 nm) were and others 2008).
Peaks 9 (tR = 24.2 min, max = 529 nm) and 11 (tR = 26.2 min,
tentatively identified as hibiscus acid, hibiscus acid glucoside, and
hibiscus acid 6-methyl ester (Figure 3) by their MS-fragmentation max = 521 nm) were identified as delphynidin-3-sambubioside
patterns. Hibiscus acid showed a negatively charged molecule ion
([M-H] ) at m/z 189 that fragmented to produce a secondary
fragment ion (MS2 ) at m/z 127 (Table 3). The same fragmentation
patterns for hibiscus acid were reported by Rodrguez-Medina and
others (2009). Hibiscus acid glucoside showed a negatively charged
molecule ion ([M-H] ) at m/z 351 that fragmented to produce 2
secondary fragment ions (MS2 ) at m/z 189 and 127 (Table 3). The
difference between m/z 351 and 189 is 162 and corresponds to
glucose. Hibiscus acid 6-methyl ester showed a negatively charged
ion ([M-H] ) at m/z 203 that fragmented to produce 2 secondary
fragment ions (MS2 ) at m/z 185 and 127 (Table 3). Hibiscus acid
is a lactone form (2S, 3R)-(+)-2-hydroxycitric acid and along
with its 6-methyl ester was shown to be an -amylase inhibitor
that could result in reduced blood glucose levels (Hansawasdi and
others 2000).
Peak 4 (tR = 13.4 min, max = 271 nm) was identified as gallic
acid (Figure 3) by comparison of the absorption spectrum with
a standard. This was confirmed by MSMS analysis that showed
the presence of a negatively charged molecule ion ([M-H] ) at
m/z 169 that fragmented to produce a secondary fragment ion
(MS2 ) at m/z 125 (see Table 3). The presence of gallic acid in
hibiscus extract was measured previously by GC-MS (Mourtzinos
and others 2008).
Peak 5 (tR = 17.1 min, max = 259 nm) was identified as protocatechuic acid glucoside (Figure 3). The absorption spectrum
was compared to a protocatechuic acid standard; the presence of
the glucose molecule slightly shifted the retention time. MS analysis of the peak revealed a [M-H] at m/z 315 that fragmented to
yield the ion m/z 153 that corresponds to protocatechuic acid (Table 3). The difference between ions 315 and 153 gave an ion with
m/z 162 that corresponds to glucose. The same MS-fragmentation
patterns for protocatechuic acid glucoside were reported in dried
plum (Fang and others 2002). Protocatechuic acid isolated from
hibiscus extracts was demonstrated to have antiatherosclerosis (Lee
and others 2002), antitumor (Tseng and others 1998; OlveraGarcia and others 2008), antioxidant (Lin and others 2003), and
anti-inflammatory (Liu and others 2002) activities.
Peaks 6 (tR = 18.7 min; max = 326 nm), 7 (tR = 23.0 min,
max = 327 nm), 8 (tR = 23.6 min, max = 327 nm), and 10 (tR =
24.3 min, max = 331 nm) were identified as caffeoylquinic acids
(CQA), which are esters formed between caffeic and quinic acids
(Figure 3). Their identification was based on previously developed
structure-diagnostic hierarchical keys (Clifford and others 2003), Figure 2Total anthocyanins content expressed as delphinidin-3-glucoside
UV-vis spectrum, and retention time was compared relative to a (mg/L). (A) Total phenolics content expressed as gallic acid equivalents
(mol of TE/mL) (C) for the extracts.
commercial 5-CQA (chlorogenic acid) standard. Peaks 6, 7, 8, (mg/L) (B) and antioxidant capacity
The upper time scale is for the 90 C (- -) curve and the lower time scale
and 10 produced a [M-H] at m/z 353, and MS2 ions at m/z 191 is for the 25 C (--) curve. Data represent the mean of n = 9. Values
(corresponds to quinic acid), 179 (corresponds to caffeic acid), 173, with similar letters within the figure indicate that treatments are not
and 135 (peak 7 only had MS2 ions at m/z 191 and 173). Peak 7 was significantly different (Tukeys HSD, p > 0.05).
C432 Journal of Food Science r Vol. 76, Nr. 3, 2011

black and green tea (Del Rio and others 2004) and in pear skins
(Lin and Harnly 2008).
It was not possible to identify several other peaks labeled with
letters (a, b, c, d, e, f, and g) in Figure 3. These could correspond
to hibiscus acid derivates or other hydroxybenzoic acids, since
they were only observed at 260 nm. There were some differences
in these compounds between the dried and the fresh hibiscus
extracts. Compound a concentration decreased in the dried
extracts while compounds f and g were only present in the
fresh extract. On the other hand, compound e was present in
higher concentration and compound d was only detected in the

(D3S) and C3S, which are the 2 major anthocyanins present in

hibiscus (Figure 3). Identification was based on their absorption
spectrum and MS-fragmentation patterns that have been previously reported (Giusti and others 1999; Degenhardt and others
2000; Juliani and others 2009). The difference between the MS
of the molecule (597) and the aglycone (303) for D3S gave a m/z
of 294 that corresponds to xylose-glucose (132 + 162) known as
sambubiose. Similarly, the MS for the C3S molecule (581) and the
aglycone (287) yields the sambubiose disaccharide (Table 2).
Peaks 12 (tR = 29.0 min, max = 359 nm), 13 (tR = 30.9
min, max = 348 nm), and 14 (tR = 32.0 min, max = 356
nm) were tentatively identified as flavonols by their characteristic
absorption spectrum with max approximately 360 nm. Peak 15
(tR = 35.7 min, max = 355 nm) was also tentatively identified
as quercetin-3-rutinoside by its absorption spectrum and MSfragmentation patterns that revealed a base peak at m/z 609 and
MS2 at m/z 301 (Figure 3, Table 3). The difference between
m/z 609 and 301 gave a m/z of 308 that corresponds to the
disaccharide rutinose formed between rhamnose (m/z 146) and
glucose (m/z 162). The presence of rutinose has been previously
reported in hibiscus extract as part of an anthocyanin (cyanidin-3rutinoside) by Segura-Carretero and others (2008). Quercetin-3rutinose with the same MS-fragmentation patterns was found in

Table 2Identification of anthocyanins present in hibiscus using

their spectral characteristics with HPLC-DAD and positive ions
in LC-MS and MS2 .
HPLC-DAD
data
Peak

Compounda

9
11

Dpd-3-sambubioside
Cyd-3-sambubioside

LC-MS data
(m/z)

tR
(min)

max
(nm)

(M + H)

MS2
(aglycone)

24.2
26.2

529
521

597
581

303
287

Dpd = delphinidin, Cyd = cyanidin.

mAU
900
9

750

500
11
250
0
0.0

5.0

10.0

15.0

20.0

mAU
600

30.0

Time (min)

35.0

400

7 9

200
1

0
0.0

25.0

5.0

4
3
d
ab c
10.0

10
11

8
e

15.0

20.0

mAU
450

25.0

14
12 13
30.0

15

35.0

40.0

7 9
200
1
2

a
3b c

Time (min)

14

11
10

15
12

13

0
0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

40.0

Time (min)

Figure 3HPLC chromatograms of dried hibiscus cold water extract: (A) 520 nm, (B) 260 nm, and fresh hibiscus cold water extract (C) 260 nm. AU =
absorbance unit. For peak identification see Table 2 and 3. Letters represent unidentified peaks.

Vol. 76, Nr. 3, 2011 r Journal of Food Science C433

C: Food Chemistry

Properties of Hibiscus sabdariffa extracts . . .

Properties of Hibiscus sabdariffa extracts . . .

C: Food Chemistry

dried hibiscus extracts. Further studies are needed to identify and extracts, respectively, while flavonols accounted for approximately
10% of the total polyphenolics in all 4 extracts. Anthocyanins acquantify these compounds.
counted for approximately 45% and 50% of the total polyphenolics
Polyphenolics quantification
in the dried and fresh hibiscus extracts, respectively.
Polyphenolics were quantified in the 4 hibiscus extracts studAs seen in Table 4, the dried hibiscus hot water sample had the
ied (dried hibiscus cold water extract, dried hibiscus hot water highest concentration of total polyphenols followed by dried hiextract, fresh hibiscus cold water extract, and fresh hibiscus hot biscus cold water, fresh hibiscus cold water, and fresh hibiscus hot
water extract) (Table 4). Results were expressed in milligrams per water extracts. Gallic acid was not detected in the fresh extracts
liter of extract. The concentration of Hibiscus acid derivatives and its presence in the dried hibiscus extracts could be attributed
was approximately 7 ppm for dried hibiscus extracts and around to a breakdown of another phenolic compound during the drying
5 ppm for fresh hibiscus extracts. A higher concentration of hi- process. The concentration of protocatechuic acid glucoside was
biscus acid glucoside was observed in the dried extracts while higher in fresh hibiscus extracts, and a significantly lower conhibiscus 6-methyl ester was higher in the fresh extracts. Differ- centration of CQAs was also observed compared with the dried
ences may be a result of the drying process. Hydroxybenzoic acids extracts. CQA distribution was approximately 50% 5-CQA and
accounted for approximately 2% of the total polyphenolics quan- 3% CQA for dried hibiscus extracts and approximately 47% 5tified in the dried hibiscus extracts and approximately 0.5% in the CQA and 10% CQA for the fresh hibiscus extracts. Three-CQA
fresh hibiscus extracts. CQAs accounted for approximately 45% and 4-CQA accounted for approximately 33% and 13% of total
and approximately 38% of total polyphenolics in dried and fresh CQAs in both dried and fresh extracts, respectively. This indicates
Table 3Identification of polyphenolics present in hibiscus using their spectral characteristics with HPLC-DAD and negative ions
in LC-MS and MS2 , and respective standards.
LC-MS data (m/z)
MS2

HPLC-DAD data
Peak
1
2
3
4
5
6
7
8
10
15
a

Compound
c

Hibiscus acid
Hibiscus acid glucosidec
Hibiscus acid 6-methyl esterc
Gallic acida
Protocatechuic acid glucosideb
3-caffeoylquinic acid
5-CQAa
4-caffeoylquinic acid
Caffeoylquinic acid isomerc
Quercetin-3-rutinosidec

tR (min)

max (nm)

(M-H)

Base peak

4.7
7.4
10.2
13.4
17.1
18.7
23.0
23.6
26.0
35.7

265
262
263
271
216,
326
327
327
331
355

189
351
203
169
315
353
353
353
353
609

127
189
185
125
153
191
191
173
191
301

Product ions
127 (48)
127 (37)
179 (58), 173 (7), 135 (14)
179 (2), 173 (0.4), 135 (0.70)
191 (20), 179 (39), 135 (14)
173 (3)

Confirmed with authentic standards. b Confirmed with the standard of the acid. c Tentatively identified. d Values in parenthesis indicate the intensity of the peak.

Table 4Polyphenolics content (mg/L) of hibiscus samples analyzed in this study.d

Compound (peak)e

Dried cold extract

Dried hot extract

Fresh cold extract

Fresh hot extract

Hibiscus acid derivates

Hibiscus acid (1)
Hibiscus acid glucoside (2)
Hibiscus acid 6-methyl ester (3)
Total hibiscus acid derivates
Hydroxybenzoic acids
Gallic acid (1)f
Protocatechuic acid glucoside (2)g
Total
Caffeoylquinic acidsh
3-caffeoylquinic acid (3)
5-caffeoylquinic acid (4)
4-caffeoylquinic acid 5)
Caffeoylquinic acid isomer (7)
Total
Flavonolsi
Unidentified (9)
Unidentified (10)
Unidentified (11)
Quercetin-3-rutinoside (12)
Total
Anthocyanins
Delphinidin-3-sambubioside (6)j
Cyanidin-3-sambubioside (8)k
Total
Total phenolic compounds

3.81b
4.16a
0.23d
8.19

3.81a b
2.98b
0.41c
7.19

4.07a
0.77c
1.04b
5.88

3.66b
0.46d
1.16a
5.27

0.65a
0.06b
0.71

0.58a
0.05b
0.63

ndl
0.19a
0.19

nd
0.13ab
0.13

67.53b
43.64b
17.52b
3.93b
132.62

73.01a
46.23a
18.81a
4.13b
142.18

51.97c
39.05c
14.12c
10.86a
116.00

49.70c
38.15c
13.51c
12.09a
99.94

5.56b
5.52a
10.21b
8.45a
29.74

5.70ab
5.58a
9.99b
9.20a
30.47

5.55b
5.11b
12.29a
9.38a
32.33

5.84a
4.86c
12.10a
9.21a
32.01

87.32c
41.62b
128.94
292.01

100.90a
44.88b
145.78
319.06

87.76c
50.30a
138.06
286.58

96.16b
50.89a
147.05
279.13

Data represent the mean of n = 6. d Values with similar letters within rows are not significantly different (Tukeys HSD, p > 0.05). e Peak numbers refer to the compounds identified in
Table 3 and 4. f ,g,h,i,j,k Quantified with gallic acid, protocatechuic acid, chlorogenic acid, quercetin, delphinidin-3-glucoside, and cyanidin-3-glucoside standards, respectively. l nd =
not detected.

C434 Journal of Food Science r Vol. 76, Nr. 3, 2011

that there is a difference in CQA concentration between dried

and fresh hibiscus extracts probably caused by the drying process.
Hibiscus anthocyanins distribution was approximately 68% and
64% of the total for delphinidin-3-sambubioside and 32% and
36% for C3S in dried and fresh extracts, respectively. This indicated that a significantly higher concentration of C3S was found in
fresh hibiscus extracts as compared to dried extracts. Delphinidin3-sambubioside was present in a significantly higher concentration in hot water extracts as compared to cold water ones, but
no significant differences were found in the concentration of
C3S in the cold and hot water extracts for both fresh and dried
hibiscus.

Conclusion
Equivalent cold and hot water conditions were found for anthocyanins extraction of dried hibiscus. Similar polyphenolic profiles were observed between fresh and dried hibiscus extracts,
although differences in some unidentified compounds and other
compounds concentrations were found. Hibiscus acid and its 2
derivates where found in all hibiscus extracts. Hydroxybenzoic
acids, CQAs, flavonols, and anthocyanins constituted the polyphenolic compounds identified in hibiscus extracts. CE would be
suitable for nonthermal processes, giving results comparable to
the traditional HE process, although longer extraction times are
required. Findings of this research can provide more flexibility
to hibiscus processing. Extraction process selection for industrial
applications should consider availability of raw material (fresh or
dried hibiscus), processing technology, time, and other economic
considerations.

References
Chang Y-C, Huang H-P, Hsu J-D, Yang S-F, Wang C-J. 2005. Hibiscus anthocyanins rich
extract-induced apoptotic cell death in human promyelocytic leukemia cells. Toxicol Appl
Pharmacol 205(3):20112.
Clifford MN, Johnston KL, Knight S, Kuhnert N. 2003. Hierarchical scheme for LC-MSn
identification of chlorogenic acids. J Agric Food Chem 51(10):290011.
Degenhardt A, Knapp H, Winterhalter P. 2000. Separation and purification of anthocyanins
by high-speed countercurrent chromatography and screening for antioxidant activity. J Agric
Food Chem 48(2):33843.
Del Pozo-Insfran D, Balaban MO, Talcott ST. 2006. Microbial stability, phytochemical retention,
and organoleptic attributes of dense phase CO2 processed muscadine grape juice. J Agric Food
Chem 54(15):546873.
Del Rio D, Stewart AJ, Mullen W, Burns J, Lean MEJ, Brighenti F, Crozier A. 2004. HPLCMSn analysis of phenolic compounds and purine alkaloids in green and black tea. J Agric
Food Chem 52(10):280715.
Fang N, Yu S, Prior RL. 2002. LC/MS/MS characterization of phenolic constituents in dried
plums. J Agric Food Chem 50(12):357985.
Ferrentino G, Plaza ML, Ramirez-Rodrigues M, Ferrari G, Balaban MO. 2009. Effects of dense
phase carbon dioxide pasteurization on the physical and quality attributes of a red grapefruit
juice. J Food Sci 74(6):E33341.

Giusti MM, Wrolstad RE. 2005. Characterization and measurement of anthocyanins by UVvisible spectroscopy. In: Wrolstad RE, Acree TE, Decker EA, Penner MH, Reid DS, Schwartz
SJ, Shoemaker CF, Smith D, Sporns P, editors. Handbook of food analytical chemistry.
Hoboken, N.J.: John Wiley & Sons Inc.
Giusti MM, Rodriguez-Saona LE, Griffin D, Wrolstad RE. 1999. Electrospray and tandem mass
spectroscopy as tools for anthocyanin characterization. J Agric Food Chem 47(11):465764.
Hansawasdi C, Kawabata J, Kasai T. 2000. -Amylase inhibitors from roselle (Hibiscus sabdariffa
Linn.) tea. Biosci Biotechnol Biochem 64(5):104143.
Hou D-X, Tong X, Terahara N, Luo D, Fujii M. 2005. Delphinidin 3-sambubioside, a Hibiscus
anthocyanin, induces apoptosis in human leukemia cells through reactive oxygen speciesmediated mitochondrial pathway. Arch Biochem Biophys 440(1):1019.
Huang D, Ou B, Hampsch-Woodill M, Flanagan JA, Prior RL. 2002. High-throughput assay
of oxygen radical absorbance capacity (ORAC) using a multichannel liquid handling system coupled with a microplate fluorescence reader in 96-well format. J Agric Food Chem
50(16):443744.
Juliani HR, Welch CR, Wu Q, Diouf B, Malainy D, Simon JE. 2009. Chemistry and quality
of hibiscus (Hibiscus sabdariffa) for developing the natural-product industry in Senegal. J Food
Sci 74(2):S11321.
Kim S, Park JB, Hwang IK. 2002. Quality attributes of various varieties of Korean red pepper powders (Capsicum annuum l.) and color stability during sunlight exposure. J Food Sci
67(8):295761.
Lee MJ, Chou FP, Tseng TH, Hsieh MH, Lin MC, Wang CJ. 2002. Hibiscus protocatechuic
acid or esculetin can inhibit oxidative LDL induced by either copper ion or nitric oxide donor.
J Agric Food Chem 50(7):21306.
Lin WL, Hsieh YJ, Chou FP, Wang CJ, Cheng MT, Tseng TH. 2003. Hibiscus protocatechuic
acid inhibits lipopolysaccharide-induced rat hepatic damage. Arch Toxicol 77(1):4247.
Lin L-Z, Harnly JM. 2008. Phenolic compounds and chromatographic profiles of pear skins
(Pyrus spp.). J Agric Food Chem 56(19):9094101.
Liu C-L, Wang J-M, Chu C-Y, Cheng M-T, Tseng T-H. 2002. In vivo protective effect
of protocatechuic acid on tert-butyl hydroperoxide-induced rat hepatotoxicity. Food Chem
Toxicol 40:63541.
Morton J. 1987. Roselle. In: Morton J, editor. Fruits of warm climates. Miami, Fla.: Creative
Resources systems Inc. p 28186.
Mourtzinos I, Makris DP, Yannakopoulou K, Kalogeropoulos N, Michali I, Karathanos VT.
2008. Thermal stability of anthocyanin extract of Hibiscus sabdariffa L. in the presence of
-Cyclodextrin. 65(12):1030310.
Olvera-Garca V, Castano-Tostado E, Rezendiz-Lopez RI, Reynoso-Camacho R, Gonzalez de
Meja E, Elizondo G, Loarca-Pina G. 2008. Hibiscus sabdariffa L. extracts inhibit the mutagenicity in microsuspension assay and the proliferation of HeLa cells. J Food Sci 73(5):T7581.
Pacheco-Palencia L, Hawken P, Talcott S. 2007. Phytochemical, antioxidant and pigment stability of acai (Euterpe oleracea Mart.) as affected by clarification, ascorbic acid fortification
and storage. Food Res Int 40(5):6208.
Prenesti E, Berto S, Daniele PG, Toso S. 2007. Antioxidant power quantification of decoction
and cold infusions of Hibiscus sabdariffa flowers. Food Chem 100(2):4338.
Rodrguez-Medina CI, Beltran-Debon R, Molina VM, Alonso-Villaverde C, Joven J, Menendez
JA, Segura-Carretero A, Fernandez-Gutierrez A. 2009. Direct characterization of aqueous
extract of Hibiscus sabdariffa using HPLC with diode array detection coupled to ESI and ion
trap MS. J Sep Sci 32(20):34418.
Segura-Carretero A, Puertas-Meja MA, Cortacero-Ramrez S, Beltran R, Alonso-Villaverde
C, Joven J, Dinelli G, Fernandez-Gutierrez A. 2008. Selective extraction, separation, and
identification of anthocyanins from Hibiscus sabdariffa L. using solid phase extraction-capillary
electrophoresis-mass spectrometry (time-of-flight /ion trap). Electrophoresis 29(13):285261.
Tsai P-J, McIntosh J, Pearce P, Camden B, Jordan BR. 2002. Anthocyanin and antioxidant
capacity in Roselle (Hibiscus sabdariffa L.) extract. Food Res Int 35(4):3516.
Tseng T-H, Hsu J-D, Lo M-H, Chu C-Y, Chou F-P, Huang C-L, Wang C-J. 1998. Inhibitory
effect of Hibiscus protocatechuic acid on tumor promotion in mouse skin. Cancer Lett
126(2):199207.
Wang C-J, Wang J-M, Lin W-L, Chu C-Y, Chou F-P, Tseng T-H. 2000. Protective effect of
Hibiscus anthocyanins against tert-butyl hydroperoxide-induced hepatic toxicity in rats. Food
Chem Toxicol 38(5):4116.
Waterhouse AL. 2005. Determination of total phenolics. In: Wrolstad RE, Acree TE, Decker EA,
Penner MH, Reid DS, Schwartz SJ, Shoemaker CF, Smith D, Sporns P, editors. Handbook
of food analytical chemistry. Hoboken, N.J.: John Wiley & Sons Inc.

Vol. 76, Nr. 3, 2011 r Journal of Food Science C435

C: Food Chemistry

Properties of Hibiscus sabdariffa extracts . . .