Separation and Purification Technology 255 (2021) 117755

Contents lists available at ScienceDirect

Separation and Purification Technology

journal homepage: www.elsevier.com/locate/seppur

Effects of recycling on the aqueous two-phase extraction of bioactives from

haskap leaves
Kar Yeen Chong , Marianne Su-Ling Brooks *
Department of Process Engineering and Applied Science, Dalhousie University, Halifax, Nova Scotia B3H 4R2, Canada

A R T I C L E I N F O A B S T R A C T

Keywords: Aqueous two-phase extraction (ATPE) is an emerging technique that integrates separation, concentration, and
Aqueous biphasic system partial purification of target compounds from natural resources. One potential issue which may prevent indus­
Lonicera caerulea trial application of ATPE, is the large consumption of phase-forming components such as alcohol, polymers, salt,
Recycling
and sugar. On top of that, post-handling and disposal of these components will increase associated costs. This
Salt recovery
Alcohol recovery
research evaluated the effects of recycling ATPE components on the extraction of bioactive compounds from
haskap (Lonicera caerulea) leaves. In this work, the ATPE systems were combinations of ammonium sulphate/
ethanol, sodium dihydrogen phosphate/ethanol, glucose/1-propanol, and maltose/1-propanol. Total phenolic
content (TPC), flavonoids, and chlorogenic acid were assessed in terms of their partitioning behavior, extraction
efficiencies, and yields. The results showed that with two recycling stages, the extraction performance for all four
ATPE systems was maintained. The average TPC extraction efficiencies across the first extraction stage and two
recycling stages were 91.4% for ammonium sulphate/ethanol, 99.6% for sodium dihydrogen phosphate/ethanol,
85.1% for glucose/1-propanol, and 85.0% for maltose/1-propanol systems. This suggests that recycling could
make ATPE a more sustainable technique in recovering high-value compounds from natural resources.

1. Introduction methods. Emerging technologies such as ultrasound [14], microwave

[15], and high-pressure [16] have been used to intensify and assist
The haskap plant (Lonicera caerulea) is a rich source of bioactive conventional extraction processes, however, these methods often
compounds, and is mainly cultivated for commercial production of require sophisticated equipment and parts such as ultrasound trans­
haskap berries, which have high levels of ascorbic acid and anthocyanin ducer, magnetron, and pressure vessel.
content in comparison to other berries [1–3]. In addition, haskap leaves Aqueous two-phase systems (ATPSs) have been explored as a simple
have been reported to contain various phenolic compounds and tri­ green liquid-liquid or solid-liquid extraction technique. They are formed
terpenoids [4–7], however they are an underutilized resource and by mixing two immiscible aqueous phases such as combinations of
comparatively few studies have been reported on the extraction of alcohol/salt, polymer/salt, polymer/polymer, or ionic liquid/salt solu­
bioactives from haskap leaves [8]. Functional compounds in plants have tions. Separation occurs after some time and two distinct phases, a top
been extracted as the main source for medicine, food, cosmetic, per­ and a bottom phase, will form. This separation is also known as an
fumery, dyes, and building materials for many years. Classical methods aqueous biphasic system as two phases are formed. Aqueous two-phase
such as Soxhlet extraction, maceration, percolation [9], reflux or extraction (ATPE) occurs when solutes from a solid or liquid material
distillation [10] are commonly used to extract plant material. However, selectively partition between the two phases. ATPE integrates separa­
these methods are time-consuming, require a lot of energy, and gener­ tion, concentration, and partial purification in one step [17]. It is an
ally use toxic organic solvents such as hexane and methanol [11,12]. adaptable process as less hazardous phase-forming components, such as
Alternatively, water is a common choice as it is safe, accessible, and alcohol, salt, and sugars, can be used with mild extraction temperatures.
economical. However, water requires a lot of energy to vaporize and has Additionally, ATPS can be designed to enhance selective partitioning of
limited solubility with non-polar compounds [13]. This has caused a compounds of interest instead of contaminants [18]. ATPE has been
shift towards other environmentally friendly or green extraction used to extract anthocyanins from red cabbage [19], chlorogenic acid

* Corresponding author.
E-mail addresses: [email protected] (K.Y. Chong), [email protected] (M.S.-L. Brooks).

https://doi.org/10.1016/j.seppur.2020.117755
Received 15 June 2020; Received in revised form 5 August 2020; Accepted 13 September 2020
Available online 17 September 2020
1383-5866/© 2020 Elsevier B.V. All rights reserved.
K.Y. Chong and M.S.-L. Brooks Separation and Purification Technology 255 (2021) 117755

from blueberry leaves [20], flavonoids from lotus leaves [21], poly­ Table 1
phenols from wheat chaff [22], and saponins from ginseng root [23]. Aqueous two-phase extraction optimized parameters [30].
Despite the advantages, the relatively high salt concentrations and ATPE systems Composition Extraction time Sample loading
cost of polymers can be an issue for ATPE. For example, the salt may
(NH4)2SO4/ethanol 20 wt% (NH4)2SO4, 87.5 min 0.190%(w/w)
corrode metal pumps, lines, or other equipment [24]. Additionally, post 27.5 wt% ethanol,
handling and disposal of phase-forming components will increase 52.5 wt% water
associated costs and this may be a barrier for large scale applications NaH2PO4/ethanol 10 wt% NaH2PO4, 5 min 0.500%(w/w)
[17]. Recycling is a potential method that can be used to reduce the 37.0 wt% ethanol
53 wt% water
amount of phase-forming components required. Past studies have Glucose/1-propanol 20 wt% glucose, 5 min 0.100%(w/w)
included recycling attempts for ATPE combinations of potassium 42.0 wt% 1-propanol,
phosphate/2-propanol [25], sodium dihydrogen phosphate/ethanol 38 wt% water
[26], ammonium sulphate/ethanol [27,28], and ammonium sulphate/1- Maltose/1-propanol 25 wt% maltose, 120 min 0.427%(w/w)
30.0 wt% 1-propanol,
propanol [29]. To the best of our knowledge, this is the first study that
45 wt% water
investigates the effects of recycling on the recovery of bioactive com­
pounds from haskap leaves.
The aim of this research was to evaluate recycling on the perfor­ A flowchart of the experimental work for the Extraction as well as
mance of food compatible ATPE systems (i.e. salt/ethanol and sugar/1- Recycling Stages 1 and 2 is shown in Fig. 1.
propanol ATPE systems), as a means of increasing the economic and For the initial extraction of haskap leaves, freshly prepared compo­
environmental sustainability of this technology and potential applica­ nents were used. Here, the 20 g ATPE systems were prepared in 50 mL
bility to the functional food industry. Ammonium sulphate ((NH4)2SO4) centrifuge tubes with internal diameter of 2.8 cm and occupied height of
and sodium dihydrogen phosphate (NaH2PO4) were used in salt systems 4.5 cm. The ground leaves were then mixed thoroughly for 15 s and
while glucose and maltose were used in sugar systems. This work builds incubated at 25 ◦ C for the required extraction time. The tubes were
upon the investigation by our previous study [30] on the recovery of centrifuged at 1500g (Sorvall T1 centrifuge, Thermo Scientific, USA) for
bioactive compounds from haskap leaves, as it uses the optimized ATPE 1 min to ensure complete phase separation. The volumes of top and
parameters that were obtained for these salt and sugar-based systems as bottom phases were recorded, and 5 mL top phase was reserved for UV
a basis for recycling studies. The performance of the recycling stages was spectroscopy analysis. The remaining top phase was kept for recycling.
assessed by examining the partition coefficient, extraction efficiency, The bottom phase required pre-treatment before analysis due to
and yield of chlorogenic acid (CGA), flavonoids, and total phenolic interference from the salt and sugar [31]. About 1–1.5 mL bottom phase
content (TPC) obtained from haskap leaves. This study is novel as it is was purified with solid phase extraction (SPE) using Hypersep SPE 500
the first time that the recycling of food compatible ATPE systems of salt/ mg/2.8 mL C18 solid phase extraction cartridges (Thermo Scientific,
ethanol and sugar/1-propanol have been investigated for the recovery of Rockwood, Tennessee, USA). Briefly, the bottom phase was loaded onto
bioactive compounds from haskap leaves. the cartridges which were preconditioned and equilibrated with water.
The cartridges were then washed with acidified water, 0.1% (v/v) for­
2. Materials and method mic acid, to remove the salt and sugar components. The organic frac­
tions were then eluted with 75% (v/v) ethanol and 100% (v/v) ethanol.
2.1. Materials The fractions were dried using rotary evaporator (HiTEC RE-51, Yamato
Scientific America, California, USA) set at 50 ◦ C at 200 rpm. They were
Ethyl alcohol (95% volume) was obtained from Greenfield Specialty dissolved in 80% methanol prior to analysis using UV spectroscopy.
Alcohols Inc. (Toronto, Ontario, Canada) while sodium dihydrogen The remaining bottom phase was mixed with 3–5 mL alcohol to
phosphate 96% was purchased from Alfa Aesar (Ward Hill, Massachu­ precipitate salt according to the concept of dilution crystallization
setts, USA). Aluminum nitrate nonahydrate (ACS, ≥97%), ammonium [26,27,29]. The solution was then filtered with Grade 415 qualitative
sulphate (ACS, ≥99%), chlorogenic acid (≥95%, titration), Folin & filter paper (VWR International, Pennsylvania, USA) and the salt residue
Ciocalteu’s phenol reagent (2 N), gallic acid monohydrate (ACS, ≥98%), was dried in a vacuum oven (Lindberg/Blue M, Thermo Scientific, USA)
D-(+)-maltose monohydrate (95%, grade II), 1-propanol (≥99%, food at 70 ◦ C to remove any remaining solvent. The recycled salt was added
grade), rutin hydrate (≥94%, HPLC grade), sodium carbonate (ACS, with an appropriate amount of fresh salt to maintain the concentration
99.95–100.05% dry basis), and sodium nitrite (≥97%, ACS) were pur­ for two-phase formation [29].
chased from Sigma-Aldrich (Oakville, Ontario, Canada). D-(+)-glucose The alcohol filtrate was combined with the remaining top phase and
monohydrate (extra pure) was purchased from Acros Organics (Morris it was dried using the rotary evaporator at 50 ◦ C and rotation speed of
Plains, New Jersey, USA) and deionized water was obtained from a Milli- 200 rpm. The alcohol recovered in the receiving flask was then tested for
Q water purification system (Millipore, Bedford, Massachusetts, USA). its alcohol content and density using an EasyDens meter (Anton-Paar
Fresh haskap leaves of Aurora variety were harvested in July 2018 GmbH, Graz, Austria). The recycled alcohol was topped up with fresh
from a haskap farm in Nova Scotia, Canada. The plants were in their first alcohol until the desired volume was achieved [32], where the amount
growth year. After the berries were harvested, fully developed leaves of fresh alcohol required was calculated from Eqs. (1) and (2), and the
were picked. At this stage of natural senescence, leaf browning was combined solution designated as recycled alcohol.
visible. The samples were frozen a − 20 ◦ C and subsequently freeze-dried ( 3)
kg m
in a Labconco FreeZone 2.5 Plus freeze dryer (Labconco, Kansas City, Correction factor( ) = Alcohol content of recycled ethanol% 3
MO, USA). The dried samples were sealed in vacuum pouches and stored kg m
at − 20 ◦ C in the dark. The samples were freshly ground, sieved through a Theoretical density of alcohol mkg3
× (1)
500 µm mesh opening, stored at − 20 ◦ C, and used within 17 days. Density of recycled alcohol mkg3

For Eq. (1), the theoretical densities of ethanol and 1-propanol used
2.2. Method
were 789 kg/m3 and 803 kg/m3, respectively. For Eq. (2), the weight of
the 100% alcohol required was calculated based on the ATPE compo­
2.2.1. Aqueous two-phase extraction (ATPE) and recycling
sition in Table 1.
ATPE systems of 20 g were prepared according to Table 1, where our
previous study [30] had determined the optimal parameters for the
extraction of phenolic compounds and flavonoids from haskap leaves.

2
K.Y. Chong and M.S.-L. Brooks Separation and Purification Technology 255 (2021) 117755

Fig. 1. Flowchart of experimental work.

Weight of 100% alcohol required 1 min at 1500g and absorbance of the supernatant was read at 510 nm.
Weight of alcohol required = (2)
Correction factor Rutin hydrate was used as the standard and results were expressed in
terms of rutin equivalents (RE).
Sugars were not recovered for ATPE recycling as there were diffi­
Chlorogenic acid. Absorbance of the extract was read at 328 nm [26].
culties in the recovery process. From our observations, glucose and
A calibration curve was plotted using chlorogenic acid as the standard.
maltose were unable to precipitate and formed a slurry when the dilu­
tion crystallization technique was applied. Rotary evaporation was then
2.2.3. Performance indicators and statistical analysis
tested as an alternative method. However, the sugars in the evaporating
Partition coefficient, k, and extraction efficiency, EE, were deter­
flasks remained as slurry due to their high degree of dissolution [33] and
mined by Eqs. (3)–(5).
only aqueous 1-propanol was evaporated. Although Flood and Srisanga
[34] have described an improved crystallization process with seed CT
Partition coefficient, k = (3)
crystals for the recovery of glucose monohydrate from aqueous solu­ CB
tions, this was not investigated in the present study.
The recycled ethanol and dried salt were used to form the ATPE Volume ratio, RV =
Volume of top phase
(4)
system for Recycling Stage 1. The same procedure was repeated and the Volume of bottom phase
next ATPE system was used for Recycling Stage 2. Ground leaves were ( )
kRV
retrieved from the freezer for every stage and the spent leaves were Extraction efficiency, EE(%) = × 100 (5)
discarded at the end of each extraction. For sugar-based ATPE, only 1- 1 + kRV
propanol was recycled. There were two independent experiments for
where CT and CB are the concentration of compounds in the top phase
each ATPE.
and bottom phase, respectively.
Yield was determined using Eq. (6).
2.2.2. UV–Vis spectrophotometric analysis of bioactive compounds
The following assays used a UV/Visible spectrophotometer (Genesys CT × VT
Yield = (6)
10S, Thermo Scientific, USA) to determine the concentration of bioac­ Dry weight of leaves
tive compounds in the extract. The corresponding top phases without
One-way analysis of variance (ANOVA) and post hoc Tukey’s test
samples and methanol were used as blanks for the top phase and bottom
were performed using Minitab statistical software (Version 18, Minitab
phase, respectively.
Inc. USA) at P < 0.05. Values in figures and tables were expressed in
Total phenolic content (TPC). Following the Folin-Ciocalteu method
average ± standard error of mean.
[35], 100 µL extract was mixed with 200 µL Folin-Ciocalteu reagent
(diluted 1:10 by volume with water). After incubation for 5 min, 800 µL
3. Results and discussion
of 7.5% (w/v) aqueous sodium carbonate solution was added. The so­
lution was mixed thoroughly and incubated for 2 h in the dark at room
3.1. Comparison of partition coefficients with previous optimization study
temperature. The solution was centrifuged for 1 min at 1500g and
absorbance was read at 665 nm for ATPE top phase. For bottom phase
The partition coefficients from this study can be compared to those
extract which was diluted in 80% methanol, absorbance was read at 765
from a previous study, which focused on the optimization of ATPE pa­
nm. Gallic acid monohydrate was used as the standard and results were
rameters for the extraction of bioactive compounds from haskap leaves
expressed in terms of gallic acid equivalents (GAE).
[30], as shown in Table 2.
Total flavonoid content. According to the concept of aluminium
Corresponding partition coefficients were assessed in Table 2 as
complex formation [21] extract of 450 µL was combined with 75 µL of
partition coefficients from the present study refer to the values from the
5% (w/v) sodium nitrite solution. The solution was mixed and incubated
first extraction using fresh phase-forming components. The values from
for 6 min. Then, 75 µL of 10% (w/w) aluminium nitrate solution was
this study are mostly much higher than that of the previous work.
added, followed with thorough mixing and 6 min incubation. Subse­
Although some had statistically insignificant difference, the increase in
quently, 750 µL of 4% (w/w) sodium hydroxide solution was added,
partitioning for the present study may be mostly due to the difference in
mixed well, and incubated for 15 min. The solution was centrifuged for

3
K.Y. Chong and M.S.-L. Brooks Separation and Purification Technology 255 (2021) 117755

Table 2
Comparison of partition coefficient obtained from previous work and this study.
Previous work1 First extraction from this study Previous work1 First extraction from this study

(NH4)2SO4/ethanol NaH2PO4/ethanol
k TPC 16.05 ± 7.58a 11.15 ± 2.79a 6.59 ± 3.28a 28.85 ± 15.47a
k flavonoids 16.44 ± 5.02a 16.99 ± 1.69a 3.50 ± 0.74a 28.17 ± 7.14a
k CGA 8.16 ± 1.44a 8.72 ± 0.17a 1.73 ± 0.23a 10.99 ± 2.87a
Glucose/1-propanol Maltose/1-propanol
k TPC 2.41 ± 0.52a 3.91 ± 0.27a 4.14 ± 1.37a 11.03 ± 1.75a
k flavonoids 2.88 ± 0.63a 4.81 ± 0.46a 5.52 ± 1.62a 17.56 ± 2.66a
k CGA 1.84 ± 0.56b 5.08 ± 0.14a 4.76 ± 1.04b 14.94 ± 1.60a

Different superscripts indicate statistically significant difference between the previous study and this study at P < 0.05.
1
Partition coefficient from previous work [30].

the height-to-diameter (H/D) ratio of the extraction vessel. In the pre­ For CGA, the partition coefficients were 8.72, 10.06, and 10.32. There
vious study, 15 mL centrifuge tubes were used for the 10 g ATPE and the was an increase in partition coefficient from the first extraction. High
H/D ratio was 5.3. The height refers to the height of the solution in the extraction efficiencies for flavonoids were recorded at about 95% while
tube. In the present study, 50 mL centrifuge tubes were used for the 20 g for TPC and flavonoids, they averaged at about 91%. Wu et al. [27]
ATPE and the H/D ratio was 1.6. The time needed for phases to separate investigated multiple extractions of (NH4)2SO4/ethanol ATPE for mul­
has been reported to be much faster in batch settlers with large cross- berry anthocyanins. The authors’ methodology was different such that
sectional area [36] and a H/D ratio of less than one has been recom­ most of the top phase from the first extraction was used in the second
mended. The results in Table 2 show that the appropriate selection of the extraction with fresh bottom phase. The partition coefficient increased
extraction vessel dimensions is critical for partitioning performance. from 3.53, 3.36, and lastly 4.50 for the third partitioning step. These
partition coefficients from Wu et al. [27] for mulberry anthocyanins
were lower than the partition coefficients for the TPC of haskap leaves
3.2. Effects of recycling on ATPE partition coefficient and extraction reported in this study. It should be noted that as anthocyanins are a
efficiency subset of compounds belonging to the diverse group of polyphenolic
compounds, the partition coefficients for TPC may be higher than for the
In Fig. 2(a), (NH4)2SO4/ethanol ATPE had partition coefficient for anthocyanins, as a broader group of compounds would be detected. The
TPC of 11.15, 10.51, and 9.37 for the first extraction, first and second ethanol and salt from Wu’s work were recycled although it was unclear
recycling stage, respectively. For flavonoids the partition coefficients how those substances were used after recycling [27]. In a different
were 16.99, 19.40, and 18.61. The values were not statistically different.

Fig. 2. Partition coefficient and extraction efficiency of (a) (NH4)2SO4/ethanol, (b) NaH2PO4/ethanol, (c) glucose/1-propanol, and (d) maltose/1-propanol ATPE for
the first extraction and subsequent two recycling stages. Different letters indicate significant difference between the stages at P < 0.05. Subscripts 1, 2, and 3
distinguish bars for TPC, flavonoids, and CGA. An asterisk indicates significant extraction efficiency decrease from the rest of the stages. Unmarked extraction ef­
ficiencies indicate no statistical difference.

4
K.Y. Chong and M.S.-L. Brooks Separation and Purification Technology 255 (2021) 117755

study, (NH4)2SO4/ethanol ATPE was used to recover lignans [28]. corresponding number of moles in parentheses) were 4 g glucose (0.022
Although the partition coefficient values were not reported, the authors mol) and 5 g maltose (0.015 mol) [38]. As glucose can form five
stated that the partition coefficient hardly decreased and showed no hydrogen bonds while maltose can form eight hydrogen bonds, the
difference after recycling the ethanol and salt for five times. This indi­ number of hydrogen bonds that can be formed for maltose is higher than
cated that the ethanol and salt were not contaminated by impurities and that of glucose. Since maltose has a greater capacity to bond with water,
target compounds. These findings aligned with our results which re­ more 1-propanol will be pushed to the top phase, resulting in higher
ported stable partition coefficient and extraction efficiencies between partitioning and extraction efficiency. Fig. 2 shows that the partition
the three stages. coefficient and extraction efficiency were fairly constant with the two
Fig. 2(b) shows the performance indicators for NaH2PO4/ethanol recycling stages, indicating that it is feasible to recycle the phase-
ATPE. The partition coefficients for TPC were 28.85, 19.18, and 53.07. forming components.
Although it seemed that the partition coefficient increased with recy­
cling, the large standard error of mean made the difference statistically 3.3. Effects of recycling on ATPE extraction yields
insignificant. For flavonoids, the partition coefficients were 28.17,
23.75, and 29.70 while for CGA they were 10.99, 8.21, and 11.20. TPC Table 3 shows the extraction yields obtained from fresh phase-
and flavonoids had more partitioning to the top phase than CGA in forming components, first, and second recycling stage.
NaH2PO4/ethanol ATPE. Extraction efficiencies were nearing 100%. In Table 3, the yields had slight variation between the stages and
Similarly, Tan et al. [26] reported that the extraction efficiency of only a few statistically different values were observed, which also sup­
NaH2PO4/ethanol ATPE to recover chlorogenic acid from ramie leaf ports the feasibility of recycling ATPE components. For the NaH2PO4/
hardly decreased after recycling the salt for 3 times. This phenomenon ethanol ATPE, the first recycling stage had a TPC yield that was 12%
was attributed to impurities being transferred to the alcohol used for salt greater than that obtained from the original extraction. However, the
precipitation. flavonoids yield from the (NH4)2SO4/ethanol ATPE was the highest at
For glucose/1-propanol ATPE in Fig. 2(c), TPC had a significant in­ the first extraction and then decreased with the rest of the stages. It was
crease in partition coefficient of 7.81 in the second recycling stage. This also observed that salt-based ATPE produced higher yields in compari­
may be due to the accumulation of phenolic content in the 1-propanol son to sugar-based ATPE, which is consistent with the results from our
from previous cycles [29]. Flavonoids had partition coefficient of previous work [30]. Among the salt-based and sugar-based ATPE sys­
4.81, 3.31, and 4.53 with the first extraction, first and second recycling tems, NaH2PO4/ethanol ATPE and glucose/1-propanol ATPE had higher
stage, respectively. CGA had corresponding partition coefficient of 5.08, extraction yields, respectively.
3.90, and 5.44. The extraction efficiencies for flavonoids and CGA Like the present work, other studies in the literature have shown that
averaged at 86.6% and 88.03%. They had no statistically significant the recycling of ATPE components can result in a fairly consistent
difference between the stages. TPC had a decreased extraction efficiency extraction yield across the recycling stages. For example, Wu et al. [27]
of 76.70% at the first recycling stage. reported that the average anthocyanin yield in the top phase was 84.7%
In Fig. 2(d) for maltose/1-propanol ATPE, TPC had partition coeffi­ and there was no statistically significant difference across three extrac­
cient of 11.03, 15.19, and 10.11, flavonoids had partition coefficient of tions. It is worth noting that the authors used a different approach from
17.56, 17.87, and 15.24, and CGA had partition coefficient of 14.94, our study, where they reused the top phase instead of recycling so that
15.64, and 15.27, respectively. The extraction efficiencies for TPC, fla­ their second partitioning step comprised of the partitioned top phase
vonoids, and CGA averaged at 85.04%, 89.02%, and 88.17%, respec­ from the first step with fresh bottom phase. Similarly, Cheng et al. [28]
tively. All the values had no statistical difference. In general, there are reported that the extraction yields of lignans hardly decreased and
very few studies on the effect of recycling on sugar/1-propanol ATPE. showed no difference after recycling five times. However, Show et al.
However, Show et al. [25] used ATPE comprising of potassium phos­ [25] demonstrated that when the top phase was recycled and the bottom
phate and 2-propanol to recover lipase. One of the performance in­ aqueous phase reused for the subsequent stages, lipase extraction yields
dicators was concentration coefficient and it was defined as the ratio decreased from 99.2%, to 89.9%, then 73.2% for the first extraction and
between concentration of lipase at the top phase at time t and original two consecutive recycling stages, respectively. This decrease in yield
concentration of lipase in bottom phase. The concentration coefficient was attributed to the bottom phase being saturated with contaminants.
decreased from 16.1, 13.4, and to 5.4. Fresh components were used in When both top and bottom phases were recycled, the yields improved to
the first extraction while successive first and second recycling stages 96.7% and 97.3% for two consecutive stages, indicating better perfor­
used recycled 2-propanol and reused bottom phases. They also observed mance when both phases are recycled. In another recycling study by
that separation efficiencies decreased across the stages. However, when Sankaran et al. [29], lipase extraction yields were 71.69%, 58.87%, and
both top and bottom phases were recycled, the separation efficiencies 79.39% for the first extraction and two consecutive recycling stages
were 88.6% and 88.5% for the fourth and fifth stages respectively. This where both the top and bottom phases were recycled.
is similar to our findings that the extraction efficiency remained fairly When determining the feasibility of ATPE for industrial scale pro­
constant without any statistically significant difference. cessing, it is important to consider the savings that would result from
Among salt-based systems, NaH2PO4/ethanol ATPE had higher par­ recycling and the reduction in fresh ATPE components needed for the
titioning and extraction efficiency of bioactive compounds to the top process. In this study, the average amount of salt and ethanol recovered
phase. In contrast, our previous study showed that (NH4)2SO4/ethanol from the (NH4)2SO4/ethanol ATPE over the three extraction stages were
ATPE performed better [30]. Results from the present study also did not 36.1% and 73.1%, respectively, where the salt was recovered after
adhere to the Hofmeister series’ salting-out ability where the H2PO−4 ion drying in the vacuum oven, while the ethanol was recovered after rotary
is expected to have lower salting-out ability. This may be due to the evaporation. In comparison, for the NaH2PO4/ethanol ATPE the average
container geometry used in this study where a larger surface area will salt recovery was much lower at 8.9%, while the ethanol recovery was
facilitate rapid partitioning to the top phase. Taking NaH2PO4/ethanol 69%. Experiments with glucose/1-propanol ATPE were able to recover
ATPE as an example, a container that provides larger interfacial area 47.7% propanol, while 54% propanol was recovered with the maltose/
will reduce the time needed for phase separation as compared to a 1-propanol ATPE. In comparison to the average salt recovery of 60.8%
narrow container [37] in the span of 5 min. The extraction time of 87.5 and alcohol recovery of 69.9% reported by Show et al. [25] for potas­
min for (NH4)2SO4/ethanol ATPE allowed ample time for biomolecules sium phosphate/2-propanol ATPE, lower salt recoveries have been ob­
to partition. In sugar-based systems, maltose/1-propanol ATPE had tained the present work. It is possible that the low salt recoveries from
better extraction performance than glucose. This is consistent with our our study could be due to the SPE cartridges used to purify the bottom
previous findings [30]. The masses of sugar in a 20 g system (and their phase for analysis. In any case, these results indicate that economic

5
K.Y. Chong and M.S.-L. Brooks Separation and Purification Technology 255 (2021) 117755

Table 3
Effect of extraction and recycling on TPC yield (mg GAE/mg), flavonoids yield (mg RE/mg), and CGA yield (µg CGA/mg).
Extraction Recycling 1 Recycling 2 Extraction Recycling 1 Recycling 2

(NH4)2SO4/ethanol NaH2PO4/ethanol
TPC yield 0.050 ± 0.00a 0.049 ± 0.00a 0.051 ± 0.00a 0.067 ± 0.00b 0.075 ± 0.00a 0.067 ± 0.00b
Flavonoids yield 0.146 ± 0.00a 0.141 ± 0.00b 0.142 ± 0.00b 0.248 ± 0.00a 0.247 ± 0.00a 0.247 ± 0.00a
CGA yield 31.75 ± 0.27a 31.85 ± 0.78a 29.60 ± 0.09a 39.75 ± 0.23a 40.67 ± 0.99a 39.10 ± 0.38a
Glucose/1-propanol Maltose/1-propanol
TPC yield 0.033 ± 0.00a 0.036 ± 0.00a 0.036 ± 0.00a 0.025 ± 0.00a 0.022 ± 0.00a 0.022 ± 0.00a
Flavonoids yield 0.062 ± 0.00a 0.075 ± 0.01a 0.080 ± 0.00a 0.055 ± 0.00a 0.054 ± 0.00a 0.054 ± 0.00a
CGA yield 19.62 ± 1.16a 24.54 ± 3.09a 24.52 ± 2.70a 17.37 ± 0.22a 17.13 ± 1.10a 16.76 ± 0.16a

Different superscripts indicate statistically significant difference between extraction, first, and second recycling stages at P < 0.05.

benefits would be possible when scaling up to the industrial scale, as [5] R. Becker, C. Pączkowski, A. Szakiel, Triterpenoid profile of fruit and leaf cuticular
waxes of edible honeysuckle Lonicera caerulea var. kamtschatica, Acta Soc. Bot.
small percentage recoveries would be significant at the larger scale.
Pol. 86 (2017), https://doi.org/10.5586/asbp.3539.
[6] K. Machida, J. Asano, M. Kikuchi, Caeruleosides A and B, bis-iridoid glucosides
4. Conclusion and future work from Lonicera caerulea, Phytochemistry. 39 (1995) 111–114, https://doi.org/
10.1016/0031-9422(94)00853-L.
[7] J.K. Dawson, Concentration and Content of Secondary Metabolites in Fruit and
This study presented the effects of recycling ATPE phase-forming Leaves of Haskap (Lonicera caerulea L.), 2017.
components. Using the recovery of TPC, flavonoids, and CGA from [8] D. Bonarska-Kujawa, H. Pruchnik, S. Cyboran, R. Żyłka, J. Oszmiański, H.
haskap leaves as an applied model, two recycling stages for ammonium Kleszczyńska, Biophysical Mechanism of the Protective Effect of Blue Honeysuckle
(Lonicera caerulea L.1. Bonarska-Kujawa D, Pruchnik H, Cyboran S, et al (2014)
sulphate/ethanol, sodium phosphate/ethanol, glucose/1-propanol, and Biophysical Mechanism of the Protective Effect of Blue Honeysuckle (Lonicera
maltose/1-propanol ATPE produced consistent partitioning, extraction caerulea L. var. kamtschatica Se, J. Membr. Biol. 247 (2014) 611–625. https://doi.
efficiencies, and extraction yields. We concluded that recycling has the org/10.1007/s00232-014-9677-5.
[9] B. Vongsak, P. Sithisarn, S. Mangmool, S. Thongpraditchote, Y. Wongkrajang,
potential to reduce the amount of materials and post handling costs W. Gritsanapan, Maximizing total phenolics, total flavonoids contents and
required for ATPE while maintaining its performance. Further studies antioxidant activity of Moringa oleifera leaf extract by the appropriate extraction
may include integration of sugar crystallization with ATPE to enhance method, Ind. Crops Prod. 44 (2013) 566–571, https://doi.org/10.1016/j.
indcrop.2012.09.021.
its sustainability. [10] T. Nakamura, Y. Murata, Y. Nakamura, Characterization of benzyl isothiocyanate
extracted from mashed green papaya by distillation, Food Chem. 299 (2019),
CRediT authorship contribution statement 125118, https://doi.org/10.1016/j.foodchem.2019.125118.
[11] E.S. Ong, Extraction methods and chemical standardization of botanicals and
herbal preparations, J. Chromatogr. B 812 (2004) 23–33, https://doi.org/10.1016/
Kar Yeen Chong: Methodology, Software, Validation, Formal anal­ j.jchromb.2004.07.041.
ysis, Investigation, Writing - original draft, Visualization. Marianne Su- [12] M.D. Luque de Castro, F. Priego-Capote, Soxhlet extraction: past and present
panacea, J. Chromatogr. A 1217 (2010) 2383–2389, https://doi.org/10.1016/j.
Ling Brooks: Conceptualization, Resources, Supervision, Project
chroma.2009.11.027.
administration, Funding acquisition, Writing - review & editing. [13] F. Chemat, M. Abert-Vian, A.S. Fabiano-Tixier, J. Strube, L. Uhlenbrock,
V. Gunjevic, G. Cravotto, Green extraction of natural products. Origins, current
status, and future challenges, TrAC Trends Anal. Chem. 118 (2019) 248–263,
Declaration of Competing Interest
https://doi.org/10.1016/j.trac.2019.05.037.
[14] A.L.-S. Eh, S.-G. Teoh, Novel modified ultrasonication technique for the extraction
The authors declare that they have no known competing financial of lycopene from tomatoes, Ultrason. Sonochem. 19 (2012) 151–159, https://doi.
interests or personal relationships that could have appeared to influence org/10.1016/j.ultsonch.2011.05.019.
[15] W. Routray, V. Orsat, Y. Gariepy, Effect of different drying methods on the
the work reported in this paper. microwave extraction of phenolic components and antioxidant activity of highbush
blueberry leaves, Dry. Technol. 32 (2014) 1888–1904, https://doi.org/10.1080/
Acknowledgements 07373937.2014.919002.
[16] P. Kraujalis, P.R. Venskutonis, E. Ibáñez, M. Herrero, Optimization of rutin
isolation from Amaranthus paniculatus leaves by high pressure extraction and
The authors would like to thank Cynthia Swinimer of Lone Tree Farm fractionation techniques, J. Supercrit. Fluids 104 (2015) 234–242, https://doi.org/
and Dr. Evie Kemp of Haskapa for assisting in the leaf collection process. 10.1016/j.supflu.2015.06.022.
[17] R.R.G. Soares, A.M. Azevedo, J.M. Van Alstine, M.R. Aires-Barros, Partitioning in
aqueous two-phase systems: analysis of strengths, weaknesses, opportunities and
Funding threats, Biotechnol. J. 10 (2015) 1158–1169, https://doi.org/10.1002/
biot.201400532.
[18] R. Hatti-Kaul, in: Aqueous Two-Phase Systems: Methods and Protocols, Humana
This work was supported by funding from the Natural Sciences and Press, Totowa, NJ, 2000, https://doi.org/10.1385/1-59259-028-4:1.
Engineering Research Council of Canada (NSERC). [19] C. Jampani, K.S.M.S. Raghavarao, Process integration for purification and
concentration of red cabbage (Brassica oleracea L.) anthocyanins, Sep. Purif.
Technol. 141 (2015) 10–16, https://doi.org/10.1016/j.seppur.2014.11.024.
References [20] T. Wang, W.-J. Xu, S.-X. Wang, P. Kou, P. Wang, X.-Q. Wang, Y.-J. Fu, Integrated
and sustainable separation of chlorogenic acid from blueberry leaves by deep
[1] J. Harb, B. Khraiwesh, J. Streif, R. Reski, W. Frank, Characterization of blueberry eutectic solvents coupled with aqueous two-phase system, Food Bioprod. Process.
monodehydroascorbate reductase gene and changes in levels of ascorbic acid and 105 (2017) 205–214, https://doi.org/10.1016/j.fbp.2017.07.010.
the antioxidative capacity of water soluble antioxidants upon storage of fruits [21] B.J. Hou, Y.Q. Wei, F. Ma, X.N. Wang, S.Z. Yang, Chelatometric salting-out
under various conditions, Sci. Hortic. (Amsterdam) 125 (2010) 390–395, https:// extraction and characteristics of flavonoids from Folium nelumbinis based on an
doi.org/10.1016/j.scienta.2010.04.031. ethanol/K2HPO4system, Sep. Sci. Technol. 53 (2018) 717–724, https://doi.org/
[2] L. Chen, X. Xin, H. Zhang, Q. Yuan, Phytochemical properties and antioxidant 10.1080/01496395.2017.1397023.
capacities of commercial raspberry varieties, J. Funct. Foods. 5 (2013) 508–515, [22] T. Đorđević, M. Antov, Ultrasound assisted extraction in aqueous two-phase system
https://doi.org/10.1016/j.jff.2012.10.009. for the integrated extraction and separation of antioxidants from wheat chaff, Sep.
[3] G.E. Pantelidis, M. Vasilakakis, G.A. Manganaris, G. Diamantidis, Antioxidant Purif. Technol. 182 (2017) 52–58, https://doi.org/10.1016/j.seppur.2017.03.025.
capacity, phenol, anthocyanin and ascorbic acid contents in raspberries, [23] A. He, B. Dong, X. Feng, S. Yao, Extraction of bioactive ginseng saponins using
blackberries, red currants, gooseberries and Cornelian cherries, Food Chem. 102 aqueous two-phase systems of ionic liquids and salts, Sep. Purif. Technol. 196
(2007) 777–783, https://doi.org/10.1016/j.foodchem.2006.06.021. (2018) 270–280, https://doi.org/10.1016/j.seppur.2017.05.041.
[4] J. Oszmiański, A. Wojdyło, J. Gorzelany, I. Kapusta, Identification and [24] P.A.J. Rosa, A.M. Azevedo, S. Sommerfeld, M. Mutter, M.R. Aires-Barros,
characterization of low molecular weight polyphenols in berry leaf extracts by W. Bäcker, Application of aqueous two-phase systems to antibody purification: a
HPLC-DAD and LC-ESI/MS, J. Agric. Food Chem. 59 (2011) 12830–12835, https://
doi.org/10.1021/jf203052j.

6
K.Y. Chong and M.S.-L. Brooks Separation and Purification Technology 255 (2021) 117755

multi-stage approach, J. Biotechnol. 139 (2009) 306–313, https://doi.org/ aqueous two-phase systems on the quantification of total proteins, Int. J. Biol.
10.1016/j.jbiotec.2009.01.001. Macromol. 83 (2016) 30–33, https://doi.org/10.1016/j.ijbiomac.2015.11.056.
[25] P.L. Show, C.W. Ooi, M.S. Anuar, A. Ariff, Y.A. Yusof, S.K. Chen, M.S.M. Annuar, T. [32] S.R. Chia, K.Y. Mak, Y.J. Khaw, N. Suhaidi, K.W. Chew, P.L. Show, An efficient and
C. Ling, Recovery of lipase derived from Burkholderia cenocepacia ST8 using rapid method to extract and purify protein – Liquid Triphasic Flotation system,
sustainable aqueous two-phase flotation composed of recycling hydrophilic organic Bioresour. Technol. 294 (2019), 122158, https://doi.org/10.1016/j.
solvent and inorganic salt, Sep. Purif. Technol. 110 (2013) 112–118, https://doi. biortech.2019.122158.
org/10.1016/j.seppur.2013.03.018. [33] L.A. Alves, J.B. Almeida e Silva, M. Giulietti, Solubility of d-glucose in water and
[26] Z. Tan, C. Wang, Y. Yi, H. Wang, M. Li, W. Zhou, S. Tan, F. Li, Extraction and ethanol/water mixtures, J. Chem. Eng. Data 52 (2007) 2166–2170. https://doi.
purification of chlorogenic acid from ramie (Boehmeria nivea L. Gaud) leaf using org/10.1021/je700177n.
an ethanol/salt aqueous two-phase system, Sep. Purif. Technol. 132 (2014) [34] A.E. Flood, S. Srisanga, An improved model of the seeded batch crystallization of
396–400, https://doi.org/10.1016/j.seppur.2014.05.048. glucose monohydrate from aqueous solutions, J. Food Eng. 109 (2012) 209–217,
[27] X. Wu, L. Liang, Y. Zou, T. Zhao, J. Zhao, F. Li, L. Yang, Aqueous two-phase https://doi.org/10.1016/j.jfoodeng.2011.09.035.
extraction, identification and antioxidant activity of anthocyanins from mulberry [35] V.L. Singleton, R. Orthofer, R.M. Lamuela-Raventos, Analysis of total phenols and
(Morus atropurpurea Roxb.), Food Chem. 129 (2011) 443–453, https://doi.org/ other oxidation substrates and antioxidants by means of Folin-Ciocalteu reagent,
10.1016/j.foodchem.2011.04.097. Methods Enzymol. (1999) 152–178.
[28] Z. Cheng, L. Cheng, H. Song, L. Yu, F. Zhong, Q. Shen, H. Hu, Aqueous two-phase [36] C. Solano-Castillo, M. Rito-Palomares, Kinetics of phase separation under different
system for preliminary purification of lignans from fruits of Schisandra chinensis process and design parameters in aqueous two-phase systems, J. Chromatogr. B
Baill, Sep. Purif. Technol. 166 (2016) 16–25, https://doi.org/10.1016/j. Biomed. Sci. Appl. 743 (2000) 195–201, https://doi.org/10.1016/S0378-4347(00)
seppur.2016.04.013. 00060-8.
[29] R. Sankaran, P.L. Show, Y.J. Yap, H.L. Lam, T.C. Ling, G.-T. Pan, T.C.-K. Yang, [37] J. Benavides, M. Rito-Palomares, Practical experiences from the development of
Sustainable approach in recycling of phase components of large scale aqueous two- aqueous two-phase processes for the recovery of high value biological products,
phase flotation for lipase recovery, J. Clean. Prod. 184 (2018) 938–948, https:// J. Chem. Technol. Biotechnol. 83 (2007) 133–142, https://doi.org/10.1002/
doi.org/10.1016/j.jclepro.2018.02.301. jctb.1844.
[30] K.Y. Chong, R. Stefanova, J. Zhang, M.S.-L. Brooks, Aqueous two-phase extraction [38] P.B. Dhamole, S. Chavan, R.G. Patil, H. Feng, M. Bule, P. Kinninge, Extraction of p-
of bioactive compounds from haskap leaves (Lonicera caerulea): comparison of coumaric acid from agricultural residues and separation using ‘sugaring out’,
salt/ethanol and sugar/propanol systems, Sep. Purif. Technol. 252 (2020), 117399, Korean J. Chem. Eng. 33 (2016) 1860–1864, https://doi.org/10.1007/s11814-
https://doi.org/10.1016/j.seppur.2020.117399. 016-0020-y.
[31] S.M. Golunski, L. Sala, M.F. Silva, R.M. Dallago, J. Mulinari, A.J. Mossi,
A. Brandelli, S.J. Kalil, M. Di Luccio, H. Treichel, Interference of salts used on