Microwave-Assisted Subcritical Water Extraction of Steviol Glycosides From Stevia rebaudiana Leaves

Abstract

A rapid method that combines microwave heating and subcritical water extraction to extract steviol glycosides from Stevia rebaudiana (Bertoni) or stevia leaves was developed and compared to conventional thermal solvent extraction methods. Within only 1 minute after subcritical water condition at 140°C being reached, the yields of major steviol glycoside, rebaudioside A and stevioside, and less polar steviol glucoside, rebaudioside C, from stevia leaves were comparable to conventional extraction method that used 70% ethanol under sonication for 45 minutes. This new hybrid extraction technique may provide a cost-effective and green approach for commercial production of high-purity steviol glycoside sweeteners and their “organic” versions without the use of highly expensive “organic” ethanol.

The increasing popularity of stevia-based high-intensity zero calorie sweeteners is partially attributed to advanced processing techniques that greatly reduces production cost and enables the commercial production of highly potent minor steviol glucosides. The sweet component, a mixture of over 30 steviol glucosides or diterpene glucosides, ¹ is cumulated in the leaves of Stevia rebaudiana (Bertoni) plant, a perennial shrub originating from the Amambay region of Paraguay and now being popularly cultivated in China. In commercial production, dried stevia leaves are subject to a series of extraction, separation/purification, crystallization, and drying processes to yield high-purity individual steviol glucosides, such as rebaudioside A (reb A), or high-purity steviol glucoside mixtures . Extraction of bioactive components from natural products has been intensively studied, and many techniques have been developed and applied commercially to meet different purposes. ² For the extraction of steviol glucosides, the techniques reported in the literature cover from time-consuming conventional maceration and thermal solvent extraction, ³ to more advanced techniques such as pressurized fluid extraction, ⁴ supercritical fluid extraction, ⁵ microwave assisted extraction, ⁶ and electro-technologies ⁷ that significantly reduce extraction time while maintaining or improving the recovery of steviol glucosides and other bioactive compounds such as polyphenols. In most cases, however, an organic solvent, preferably methanol or ethanol, is involved in the extraction process, due to the relatively low water solubility of reb A (8 g/L) and stevioside (stv) (1.3 g/L), ⁸ the two major steviol glucosides in stevia leaves. Because of safety and environmental concerns, many stevia suppliers have moved away from methanol to ethanol or to solvent-free green processes.

A few green or water-only extraction methods, including subcritical water extraction (SWE), ^9,10 and enzyme assisted extraction ¹¹ were also reported. Subcritical water is used to describe the stage of water with temperature between 100°C and 374°C and pressure high enough to maintain its liquid state. It can also be referred to as pressurized hot water extraction (PHWE), superheated water extraction, or accelerated solvent extraction with water as a solvent. In its subcritical state, liquid water experiences property changes that result in an increase in solvating power and diffusivity as well as a decrease in polarity, viscosity, and surface tension, ¹² enabling water to behave similarly to less polar solvents like ethanol. Subcritical water extraction is getting more attention recently, because it has demonstrated that it can offer shorter production duration compared to conventional methods, and milder operating conditions and lower operational costs than other green technologies such as supercritical fluid extraction.

Here, we report a hybrid extraction technique that uses rapid microwave heating to create subcritical water condition to speed up water-only extraction of steviol glycosides from dry stevia leaves. Within only 1 minute after the preset temperature at 140°C being reached, the yield of reb A and stv, and other less polar steviol glucosides, such as rebaudioside C, from stevia leaves was comparable to conventional extraction method that uses 70% ethanol under sonication for 45 minutes. The availability of those green techniques eliminates the use of organic solvents in the process, not only provides environmentally friendly alternatives but also enables the production of organic steviol glucosides without the use of highly expensive organic ethanol.

The yields of reb A and stv under different extraction conditions are summarized in Table 1. Both reb A and stv showed similar trends across time-temperature combinations from this simple design of experiments. Although further optimization work is warranted, highest yields were obtained under each temperature setting that is 5 minutes at 100°C, 30 minutes at 120°C, and 1 minute at 140°C. Overall, holding at 140°C for 1 minute extracted most reb A and stv from the leaves among all of the 12 conditions tested. As steviol glycosides are antioxidants and may be subject to thermal oxidation when oxygen is present, sparging the reaction with nitrogen to reduce oxygen-induced oxidation was tested. However, removing oxygen had little effect, as extraction yields were within the range of analytical uncertainty compared to those without nitrogen sparging.

Table 1

Rebaudioside A and Stevioside (% Dry Leaf by Mass) Extraction Yields From Microwave-Assisted Subcritical Water Extraction.

	1 min		5 min		15 min		30 min
	reb A	stv	reb A	stv	reb A	stv	reb A	stv
100°C	8.4	3.6	9.1	3.9	8.3	3.6	7.4	3.2
120°C	9.1	3.9	8.8	3.8	9.2	3.9	9.5	4.1
140°C	9.7	4.3	8.1	3.5	8.3	3.6	8.2	3.5

reb A, rebaudioside A; stv, stevioside.

Interestingly, the decrease in reb A yields at higher temperature and longer extraction time is correlated to the increase in rebaudioside B (reb B) concentration (Table 2). This is likely due to hydrolysis of reb A to reb B. The same trend was observed for reb B and steviobioside (Table 2). This may provide a green alternative for the production of reb B from the hydrolysis of reb A, which in general is catalyzed by acids or bases. Based on the data, it was believed that thermal degradation may also take place during the extraction process that takes place with extended extraction time.

Table 2

Rebaudioside B and Steviobioside (% Dry Leaf by Mass) Extraction Yields From Microwave-Assisted Subcritical Water Extraction.

	1 min		5 min		15 min		30 min
	reb B	stb	reb B	stb	reb B	stb	reb B	stb
100°C	0.38	0.18	0.39	0.21	0.35	0.20	<0.1	0.57
120°C	0.42	0.20	0.39	0.20	0.43	0.25	0.51	0.25
140°C	0.55	0.23	0.52	0.25	0.97	0.38	1.42	0.57

reb B, rebaudioside B; stb, steviobioside.

Since SWE at 140°C already yielded similar results as conventional 70% ethanolic extraction method that takes 45 minutes (Table 3), higher temperatures or pressures were not assessed. For less polar rebaudioside C (reb C), the yield is also comparable. Therefore, this was not needed to further reduce the hold time, because the total cycle time, which included heat up and cool down, was about 8 to 10 minutes. The much faster extraction speed than standalone microwave-assisted extraction ⁶ or PHWE ^9,10 may be due to a synergistic effect of the unique microwave heating mechanism and the beneficial water property under subcritical condition. With only about 2.6 bar (0.26 MPa) of operating pressure, the mild condition can be easily and cost effectively adopted to large commercial scale, providing another advantage over techniques that require much higher operating pressure.

Table 3

Comparison Between the Developed Subcritical Water Extraction and Conventional Method.

Method	Conventional method	SWE
			Solvent	70% Ethanol	Water
Hold time	45 min	1 min
Temperature	Ambient	140°C
Pressure (bar)	-	2.6
reb A (% dry leaf by mass)	9.9	9.7
stv (% dry leaf by mass)	4.3	4.2
reb B (% dry leaf by mass)	0.4	0.5
reb C (% dry leaf by mass)	1.3	1.3
stb (% dry leaf by mass)	0.29	0.23

reb A, rebaudioside A; reb B, rebaudioside B; reb C, rebaudioside C; stb, steviobioside; stv, stevioside; SWE, subcritical water extraction.

We also tested water-only extraction with conventional heating we reported before, ¹³ with varying leaf to water ratios. Without further cleanup, the conventional water extract (60°C for 2 hours) showed lower recovery and strong interference around reb A and stv compared to 70% ethanol extraction (Figure 1). Conventional water extraction also extracted nearly no reb C from the same leaves.

Figure 1

HPLC chromatogram of stevia leaf extract. Top trace (blue): 70% ethanol extraction; bottom trace (red): water-only extraction.

Experimental

Materials and Chemicals

Dried stevia leaves with high reb A content were obtained as retain samples from Cargill’s past agronomy program. The dried leaves were grounded into powder by using a coffee grinder before extraction. Karl Fischer titrator (Metrohm, Herisau, Switzerland) was used to determine the water content of the leaves. Food grade 200 proof ethanol that was used for extraction was purchased from Brenntag Great Lakes, Inc. (Saint Paul, MN, United States). High Performance Liquid Chromatography-grade (HPLC-grade) acetonitrile was purchased from Fischer Scientific (Hampton, NH, United States). Water was purified using a Sartorius Arium Pro water purification system (Goettingen, Germany).

Conventional Solvent Extraction Method

This is a standard method used in our lab for stevia leaf evaluation. Add 1 g of the ground dry leaf powder to a 100 mL volumetric flask and dilute to 1 to 2 cm below the mark with 70% ethanol. Sonicate the solution for 45 minutes and let it equilibrate to room temperature before diluting it to the mark with 70% ethanol. The supernatant was collected, filtered through a 0.2 µm polytetrafluoroethylene (PTFE) filter, and diluted 10 times with 30% acetonitrile prior to HPLC analysis.

Conventional Water Extraction Method

The ground dry leaf powder was steeped in water at a 1:100 weight to volume ratio at 60°C for 2 hours. The supernatant was collected, filtered through a 0.2 µm PTFE filter and diluted 10 times with 30% acetonitrile prior to HPLC analysis

Microwave-Assisted Subcritical Water Extraction

Stevia leaf was extracted with subcritical water in a Anton Parr (Graz, Austria) Monowave 300 reactor. The reactor has operation limits of 300°C and 30 bar. Fifty milligrams of ground dry leaf powder was combined with deionized (DI) water at a 1:100 weight to volume ratio in a 10 mL standard borosilicate glass vial and then it was sealed with a polyether ether ketone (PEEK) snap cap and a PTFE-coated silicone septum. The samples were heated to temperatures between 100°C and 140°C (to ensure subcritical water effect) as fast as possible (1-2 minutes depending on final temperature), held at that constant temperature for 1 to 30 minutes, cooled to 70°C by forced air in the reactor (over 2-5 minutes), then transferred to a closed vial in ice bath, and stored at 4°C until analysis. The effect of minimizing oxygen concentration in an attempt to minimize oxidative degradation was tested for 15 minutes at 120°C condition, by first sparging the vial for 10 minutes with nitrogen to reduce oxygen concentration to 0.05% of normal concentration in air. The supernatant was collected and filtered through a 0.2 µm PTFE filter prior to HPLC-UV analysis.

HPLC Analysis

HPLC-UV analysis was carried out on a Waters Acquity ultra-high performance liquid chromatographic system (Milford, MA, United States) equipped with a photodiode-array detector operating at 210 nm. An Agilent Poroshell EC-C18 column (150 × 4.6 mm, 2.7 µm) held at 50°C was used for the separation with 0.01% trifluoroacetic acid in water (solvent A) and acetonitrile (solvent B) as mobile phases. The flow rate was 0.6 mL/min with an elution gradient as follows: 0 to 2 minutes 75% A, 2 to 5 minutes 75% to 68% A, 5 to 14 minutes 68% A, 14 to 16 minutes 68% to 50% A, 16 to 18 minutes 50% A, 18 to 18.5 minutes 50% to 10% A, 18.5 to 22 minutes 10% A, 22 to 22.5 minutes 10% to 75% A, and 22.5 to 28 minutes 75% A. The injection volume was 10 µL. The method has report levels of 0.1% (w/w) in stevia leaves for all of the 9 steviol glycosides listed in 2010 JECFA monograph. ¹⁴ The method precision is less than 4% for both reb A and stv.