Metals analysis of botanical products in various matrices using a single microwave digestion and inductively coupled plasma optical emission spectrometry (ICP-OES) method

Abstract

Presented here is the development and validation of a simple and singular sample preparation method for the detection of heavy metals (i.e. arsenic, cadmium, lead, and mercury) and other elements (i.e. iron, sodium, calcium, phosphorus, and zinc) in botanical products by microwave digestion and inductively coupled plasma optical emission spectroscopy (ICP-OES). This one methodology is successfully applied to botanical samples existing in three very different forms; powdered raw material, liquid phyto-caps (i.e.glycerin-based matrices) and ethanolic tinctures. Method validation results were obtained by the generation of calibration curves with multi-element aqueous standard solutions and a standard addition method. NIST standard reference material (SRM) 3241 Ephedra sinica Stapf Native Extract, SRM 3243 Ephedra-Containing Solid Oral Dosage Form and SRM 3246 Ginkgo biloba (leaves) were used for the method validation. Limits of detection obtained from the calibration functions range from 4 to 100 ng/mL. Concentration ranges determined for the various botanical extracts were as follows: iron (1–434 µg/g), sodium (3–6742 µg/g), calcium (4–46537 µg/g), phosphorus (24–14710 µg/g), and zinc (7–7247 µg/g). For the majority of the heavy metals (As, Cd, Hg and Pb) the concentration results were not detectable (ND), i.e. below the detection limits of the method.

---

Introduction

The marketing, sale, and consumption of botanical products (aka, dietary supplements or nutraceuticals) has been on the upsurge over the last 20 years because of their perceived health benefits towards heart disease, cancer and other conditions. In 2007, the US nutritional product market was responsible for $94 billion in consumer sales, approximately an 11% increase from 2006.¹ In the past, the overall assurance of product safety and the subsequent health effects claimed on the labels required no substantiation. With the increase in sales, the safety and efficacy of these products has become a very important issue. For decades there have been efforts toward the establishment of rules and regulations on the manufacturing and testing of the botanical products. Because of the variety of products, compositions, and manufacturer processes available, the creation of these regulations is a very arduous and time-consuming process. In addition, the wide diversity of product sources and analytical capacities makes the development of unified standards quite difficult.

There are two distinct aspects to the regulation of botanical product commerce: truth in labeling and quality/safety assurance. In 1994, the Dietary Supplement Health and Education Act (DSHEA) introduced new regulations for dietary supplements.² This act defined the specific criteria that dietary supplements should meet and began to address several quality/safety concerns of supplements in the market place. In 2003 the Food and Drug Administration (FDA) proposed regulations that would make dietary supplement manufacturing, packaging, and storage be in compliance with current good manufacturing practices (cGMPs). Overall, the cGMPs address the safety concerns with regards to the claims made on the products label.^2–4 In addition to these federal regulations, the state of California has enacted Proposition 65, an amendment to the Safe Drinking Water and Toxic Enforcement Act of 1986 which establishes “Safe Harbor Levels” for many substances and compounds that are known or suspected to cause cancer or adverse reproductive effects.^5,6 Although this California law does not target botanical products, it provides specific guidelines for the daily maximum exposure to toxic species (e.g. heavy metals), some of which can potentially be found in botanical extracts. Specifically, the maximum allowable dose levels in Proposition 65 for arsenic, cadmium, lead and mercury are 0.1, 4.1, 0.5 and 0.3 µg/day, respectively.

Botanical products can be found in a wide variety of forms/matrices; including ethanolic tinctures, softgels, tea bags, powders, capsules and tablets. Ideally, the monitoring of the elemental components in different types of sample matrices could be carried out by a single sample preparation and analytical determination method. However, due to the nature of the various matrices, the development of such a methodology is very challenging. Several laboratories have reported digestion and analysis procedures for dietary and botanical supplements, as well as for food and other biological samples.^4,7–9 The sample preparation and detection techniques used for these matrices generally consist of either wet and dry ashing or microwave digestion with atomic absorption spectrometry (AAS), inductively coupled plasma optical emission spectrometry (ICP-OES), or inductively coupled plasma mass spectrometry (ICP-MS). To this point, the concept of broad-ranging matrix capabilities has not been demonstrated.

The present work describes the development and validation of a single botanical product preparation and analysis method using a microwave digestion procedure that is applied to three diverse matrices (powdered dried raw material, liquid-phyto caps, and ethanol-based tinctures) analyzed by ICP-OES for As, Cd, Hg, Pb, Fe, Na, Ca, P and Zn. Once the optimization of the digestion parameters was achieved, NIST standard reference material (SRM) 3241 Ephedra sinica Stapf Native Extract, SRM 3243 Ephedra-Containing Solid Oral Dosage Form and SRM 3246 Ginkgo biloba (leaves) were employed for the validation of this method by generating calibration curves with aqueous standard solutions and by the standard addition method. Special emphasis during the course of this study is given to the heavy metal content in the commercial botanical products. It is believed that this straightforward, unified approach provides a cost-effective alternative to the use of multiple, matrix-specific approaches to dietary supplement analysis.

Experimental

Instrumentation

Digestion of the samples was performed with a MARS Xpress microwave digestion system (CEM Corporation, Matthews, NC, USA). The system was equipped with a 40-place sample rotor (turret) capable of holding 75 mL PFA-Teflon sample digestion vessels operable at temperatures of up to 260 °C and 500 psi. Temperature control was achieved through feedback via an infrared sensor. Temperatures ranging from 50 °C to 80 °C in combination with hold times of 10, 15 and 20 minutes were evaluated for the pre-digestion step with the power set at 300 W. In the case of the digestion step (power at 1200 W), temperatures ranging from 150 °C to 210 °C with ramp and hold time variations of 10, 15, and 20 minutes were evaluated. Caution must be taken to allow pressurized vessels to come to room temperature before opening to atmosphere. Table 1 presents the optimal microwave digestion system operating conditions employed in the quantitative evaluation of the method.

Table 1 Optimized microwave digestion program

Stage	Power (W)	Ramp time (min.)	Temperature (°C)	Hold time (min.)	Cool down time (min.)
Pre-digestion	300	0	80	15	15
Digestion	1200	10	180	15	15

---

The quantitative elemental analysis of the botanical extracts was performed by inductively coupled plasma optical emission spectrometry (ICP-OES) using a Jobin-Yvon Ultima 2 (Longjumeau, France) equipped with a radial-view plasma, a Meinhard concentric glass nebulizer and a cyclonic spray chamber. The Ultima 2 spectrometer consists of a 1.0 m Czerny-Turner monochromator equipped with 2400 grooves mm⁻¹ holographic grating, controlled by JY Analyst v5.2 data acquisition software. In order to obtain the optimal ICP-OES performance, the experimental conditions (i.e. power, sample introduction rate, nebulizer gas flow rates and the emission wavelengths) need to be considered. For the sake of simplicity, each of the parameters, with the exception of the emission wavelength, was set to the manufacturer's default values and held constant throughout the course of the entire study. For the selection of the best emission wavelength, all or some of the transitions were selected from the software database and evaluated with a 1.0 µg mL⁻¹ multi-element standard solution containing all of the target and elements present in the botanical extracts. The wavelength responses were evaluated based on their sensitivity, absence of spectral interferences, and detection limits. Table 2 shows the ICP-OES operation parameters and wavelengths used here.

Table 2 ICP-OES operation conditions

Parameters	Condition
Power (W)	1000
Ar gas flow rate (L/min)	12.0
Nebulizer (L/min)	0.02 at 1.0 bar
Sheat gas flow rate (L/min)	0.20
Peristaltic pump speed (rpm)	20.0
Replicates	5

---

Element	Wavelength (nm)
As	193.695
Cd	214.438
Pb	220.353
Hg	194.950
Ca	211.276
Zn	213.856
Na	588.995
P	213.618
Fe	259.940

---

Materials

All samples and standards were digested in trace metal grade nitric acid (HNO₃) (Fisher Scientific, Fair Lawn, NJ, USA) and diluted in MilliQ-water (18.2 MΩ cm⁻¹, NANOpure Diamond, Barnstead International, Dubuque, IA). The samples were stored in 60 mL amber Nalgene bottles (Fisher Scientific, Fair Lawn, NJ, USA) prior to analysis. Single and multielement solutions (certified reference materials) used in the preparation of standards were obtained from High Purity Standards, Charleston, SC, USA.

NIST standard reference material (SRM) 3241 Ephedra sinica Stapf Native (hot water) Extract, SRM 3243 Ephedra-Containing Solid Oral Dosage Form and SRM 3246 Ginkgo biloba (leaves) (all in powdered form) were used for the validation of the method. Botanical extracts in the form of ethanolic tinctures (single herbs or blends) consisting of 25 to 75 percent ethanol, liquid phyto-cap samples consisting of 50 to 60 percent glycerin, and powdered raw material used for this study were provided by Gaia Herbs (Brevard, NC).

Sample preparation

Approximately one gram of each botanical extract (ethanolic tinctures and liquid phyto-cap samples) was accurately weighed and placed in a 75 mL Teflon microwave digestion vessel. One mL of concentrated HNO₃ was carefully added to the vessel to prevent an explosive reaction. Once the initial reaction had come to completion, an additional 4 mL of HNO₃ was added to the vessel. (In the case of the glycerin-based samples, the entire 5 mL of HNO₃ was added in one step.) After the reaction between the HNO₃ and the ethanolic extract was completed, the vessels were placed in the microwave system with the caps un-torqued (not fully sealed) for the pre-digestion step. Once cool, the vessel caps were tightened and the samples were placed back in the microwave system for the final digestion step. After the conclusion of the digestion step, the vessels were allowed to cool to room temperature, vented and the samples transferred to 50 mL volumetric flasks and diluted to volume with MilliQ-H₂O. In the case of the powdered raw material, use of 1 gram of sample resulted in an undigested residue (i.e. particulate present in solution). Therefore, various amounts of the powdered raw material were investigated, with a mass of ∼0.85 g resulting in complete digestion of the various raw materials.

A 1.0 µg mL⁻¹ stock solution of the heavy metals (As, Cd, Pb, and Hg) was routinely prepared in MilliQ-H₂O from aqueous multielement standards of 20 µg mL⁻¹ and further used to prepare the aqueous calibration standards on a daily basis. For the other elements (Fe, Na, Ca, P and Zn) a 1000 µg mL⁻¹ multielement standard was used for the preparation of the standard solutions. The calibration standards were prepared to contain the same acidity (10% nitric acid) as the digested samples. For the standard addition method, a 10 µg mL⁻¹ stock solution including As, Cd, Pb and Hg was prepared and amounts of 0.050, 0.100 and 0.200 mL were added to 10 mL of the digested sample. In the case of Fe, Na, Ca, P and Zn amounts of 0.200, 0.400 and 0.600 mL from the 1000 µg mL⁻¹ multielement standard were added to 10 mL of the digested sample.

Results and discussion

Development of digestion procedure

In order to obtain correct elemental quantification, it is crucial to ensure that the prepared samples are in a suitable matrix that can be subsequently analyzed by the instrument of choice (ICP-OES in this case). To be the most practical in implementation, it was desired to develop a procedure that can be applied to multiple matrices (i.e. ethanolic tinctures, raw material, tablets and/or powder forms). The ultimate developed procedure should be simple, efficient, and easy to perform on a regular basis while providing high yields and reproducibility. Initially, open vessel hotplate methods where evaluated, wherein HNO₃ was added to the ethanolic samples for digestion and heated in open volumetric flasks.^10,11 The reaction of HNO₃ with ethanol fully digested the samples, but it should be noted that the reaction is very violent, producing nitrogen dioxide gases. While this procedure was successful for the digestion of the ethanolic tinctures, there are several disadvantages, including possible analyte (vapor) loss from the open vessels and the time-consuming (3–4 hours) nature of the reaction if done under mild conditions. Because the hotplate procedure was moderately effective for the ethanolic tinctures, the liquid phyto-cap samples were digested in the same manner, but with no success. The glycerin-based sample digestions were incomplete with undigested and oily residue material remaining. One limitation may be due to the fact that the glycerin-based samples are more concentrated with respect to botanical material than the ethanolic tinctures. In addition, each sample has different degrees of viscosity because each extract contains a different percentage of glycerin. Various nitric acid digestion procedures found in the literature for nutraceutical products⁹ and mixed-acid digestion procedures of plant materials¹² were applied for the hotplate digestion of the glycerin-based samples. It was hoped the procedures from the literature would be applicable to the different sample-types (i.e. ethanolic tinctures and liquid phyto-cap samples), but they were attempted with no success. Hydrochloric acid and sulfuric acid were also unsuccessfully explored for the use in sample digestion,^4,13–15 therefore the application of microwave digestion was considered.

Microwave digestion has been widely applied to the analysis of numerous types of samples, including the botanical product and dietary supplement fields.^8,9,16 The application of microwave enhanced chemistry for sample preparation allows for shorter reaction times (i.e. digestion), reduction in the number of discrete sample preparation steps, greater sample homogeneity after digestion, increased sample throughput and better precision.^10,17,18 The processes are also very well suited for standardization and automation during method development.¹⁰

During the development of the digestion procedure, the microwave operation parameters (e.g., run time and temperature) were evaluated for the different botanical matrices. Given the diversity of materials, a particular digestion was deemed successful when an optically homogeneous, and temporally stable, solution was produced. The same amount of concentrated nitric acid (5 mL) used while evaluating the hotplate digestion was used for each sample-type during the microwave digestion. After the addition of the nitric acid to the samples, the initial reaction time was varied from matrix-to-matrix (0 to 30 min). Therefore, a pre-digestion step (first stage) was added to the microwave program to initiate the reaction between the matrix and acid, followed by the second stage of digestion (Table 1) where the temperature is ramped from 80 °C to 180 °C over the course of ten minutes at a power of 1200 W, following a hold time at 180 °C for 15 minutes and cool time for another 15 minutes.

Analytical response characteristics

Once the optimization of the operation parameters for the primary dissolution was achieved, the analytical response characteristics were determined for each of the elements of interest using aqueous multielement standard solutions. The calibration curves were generated for each of the elements through the acquisition of five intensity measurements across a concentration range from 0 (i.e. analytical blank) to 300 ng mL⁻¹ for the heavy metals and 0 to 50 µg mL⁻¹ for Fe, Na, Ca, P and Zn. Good linearity and satisfactory coefficients of correlation (R² values) were observed for each of the elemental response functions. The limits of detection (LOD = 3σ_blank/m) were also calculated from each calibration response. Table 3 shows the analytical response characteristics obtained by ICP-OES for each of the elements of interest based on the use of aqueous calibration standards.

Table 3 ICP-OES analytical response characteristics

Pneumatic nebulization				CMA spray chamber
Element	Response function	R ^2	LODs (ng mL^−1)	Response function	R ^2	LODs (ng mL^−1)
Fe	y = 9E^+4x + 3E^+5	0.9995	5.0
Na	y = 7E^+3x + 2E^+4	0.9998	15.0
P	y = 7E^+2x + 2E^+2	0.9999	100.0
Zn	y = 5E^+4x + 5E^+5	0.9935	19.0
Ca	y = 5E^+2x + 2E^+3	0.9994	15.0
As	y = 6E^+4x − 6E^+2	0.9996	6.0	y = 1E^+5x − 8E^+2	0.9989	3.0
Cd	y = 1E^+5x − 4E^+2	0.9999	4.0	y = 2E^+5x − 2E^+2	0.9999	4.0
Pb	y = 1E^+5x − 1E^+3	0.9989	6.0	y = 9E^+4x − 9E^+1	0.9994	15.0
Hg	y = 1E^+5x − 4E^+2	0.9982	5.0	y = 1E^+6x − 6E^+3	0.9995	5.0

---

The determination of some metals/metalloids can be achieved with better sensitivity through the use of hydride generation sample introduction. In addition to the conventional solution nebulization described above, determination of the heavy metal concentrations was also performed following hydride generation using the concomitant metal analyzer (CMA) spray chamber (Jobin-Yvon, Longjumeau, France). When using the CMA spray chamber, the reaction of sodium borohydride and an acidic solution (i.e.hydride formation) takes place in the chamber after being delivered by a peristaltic pump. One gram of sodium borohydride was dissolved in 100 mL of water, with three different hydrochloric acid (HCl) concentrations (0.5, 1.0 and 3.0 M) evaluated to determine the best acid composition. Calibration curves were obtained for As, Cd, Hg, and Pb over the concentration range of 0 to 300 ng mL⁻¹, at each HCl concentration. The best elemental responses during the hydride generation experiments were observed with a 1 M HCl concentration. As in the case of using conventional nebulization, good linearity and satisfactory correlation coefficients were observed for each response function, as shown in Table 3. Overall, the limits of detection for As were improved by a factor of 2, but at the expense of a ∼3× increase in Pb LOD. In the case of Hg and Cd, no changes in the LODs were observed. Due to the fact that the LODs obtained without the CMA chamber are in the low ng mL⁻¹ levels, and fall below the Prop 65 guidelines, the quantitative elemental analysis of the various botanical extracts was carried out without hydride generation.

Method validation

Upon development of the singular digestion procedure, it is necessary to ensure the procedure's efficiency to digest the samples in such a way that an accurate representation of elemental concentrations is obtained. The ultimate goal of this study was to validate the developed digestion procedure by analysis of standard reference materials (SRMs). The selected SRMs for the validation experiments need to be in a suitable matrix that is representative of the botanical extracts. However, because commercial botanical products have only recently come under scrutiny, very few SRMs targeting botanical products exist. The National Institute of Standards and Technology (NIST) had initiated the development of dietary supplement SRM suites, encompassing materials from the different preparation/production steps (e.g., harvest to final manufactured product).¹⁹ The first two suites of NIST botanical SRMs available in the market were Ephedra sinica and Ginkgo biloba. Three reference materials; SRM 3241 Ephedra sinica Stapf Native (hot water) Extract, SRM 3243 Ephedra-Containing Solid Oral Dosage Form and SRM 3246 Ginkgo biloba (ground leaves) were employed during the validation experiments. The primary material for SRM 3241 was prepared by hot water extraction of the plant material under pressure, followed by filtration and concentration to produce the native product.²⁰ The materials making up SRM 3243 and SRM 3246 were prepared from various commercially available sources, ground and sieved for production of the packaged SRM.^19,20 The certified values for these SRMs include reports of the active organic components, as well as the trace levels of toxic metals and nutrients. Due to the very low concentrations of the heavy metals in these SRMs (where certified at all) it was necessary to validate the method by inclusion of nutrient elements (Fe, Na, P, Zn, Ca) to the analyte list. Thus, the method can be validated over a very wide range of elemental concentrations, as well as physical and chemical characteristics.

Prior to validating the entire digestion and ICP-OES analysis method for botanical products, it was first necessary to do so for the test elements present in neat aqueous (standard) solutions. To do so removes the chemical digestion efficiency aspect of the process, but includes aspects of solution preparation, transfer among the various vessels, and performing the ICP-OES quantification procedure. The first column of Table 4 shows the recovery values obtained for a mixture of the aqueous standards (100 ng mL⁻¹ each) taken through the complete sample preparation (microwave digestion) and ICP-OES analysis. Recoveries of 94% and higher were obtained for each of the elements, with sample-to-sample variabilities of ≤4% RSD, demonstrating that there was minimal elemental loss during the sample preparation procedures. There is a question as to why the recoveries of the nutrient elements are all above 100%, albeit not by much. These elements are the most likely to be present in the de-ionized water used in the solution preparations, thus leading to somewhat elevated blank levels.

Table 4 Elemental recoveries for aqueous standard solutions and the three commercial botanical product matrix types taken through microwave digestion process and ICP-OES analysis

Element	Recovery (%)
	Aqueous standards (n = 5)	Ethanolics (n = 16)	Phyto-caps (n = 14)	Raw (n = 10)	Cumulative (n = 40)
Fe	103	102	105	96	102
Na	100	106	107	100	105
P	109	110	104	100	105
Zn	103	108	106	103	106
Ca	103	110	106	105	107
As	96	97	97	96	97
Cd	97	99	97	101	99
Pb	94	92	91	92	92
Hg	95	60	62	61	61

---

The validation of the microwave digestion procedure developed for the three different matrices was accomplished using both the external calibration and standard addition methods, which are the most common approaches for ICP-OES measurements. Table 5 shows the validation results obtained for the nutrient elements in SRMs 3241 and 3243, using the external calibration and standard addition procedures. (Values are not certified for these elements in SRM 3246.) Overall, good recoveries were obtained for these elements, with values of 86% and higher, as well as having variabilities of ≤15% RSD. The precision here is in fact better than provided on the SRM certificates of analysis (overall variability of ≤21% RSD). Table 6 presents the validation results obtained for As, Cd, Hg, and Pb using external calibration and standard addition. For the detectable elements, As and Pb (in most cases), the determined values were comparable to the certified values provided by NIST, with recoveries of ≥95% obtained by external calibration and standard addition. The precision is not as good with the heavy metals here in comparison to the NIST values, presumably due to the use of less sensitive ICP-OES than ICP-MS used for NIST quantification. The goal behind using both calibration techniques was to determine the potential effects of the different botanical matrices on the ICP analysis; i.e., are there potential matrix effects that make calibrations curves unsuitable, and only standard addition is a viable means of quantification? Because both validation procedures provided good results and the fact the number of botanical samples to be analyzed is high, the analysis of the botanical extracts was performed by using the external calibration method.

Table 5 Validation results for nutrient elements in NIST SRMs 3241 and 3243

Element	Reference values (mg kg^−1)	Calculated values (mg kg^−1)	% Recovery	Calculated values (mg kg^−1)	% Recovery
		Aqueous calibration standards		Standard addition method
(X%) = relative standard deviation.a Reference value in percentage.
SRM 3241
Fe	900 ± 100 (11%)	803 ± 58 (7%)	90	955 ± 89 (9%)	106
Na	2480 ± 280 (11%)	2446 ± 103 (4%)	99	2849 ± 351 (12%)	115
P	not certified	3169 ± 173 (5%)	—	3173 ± 177 (6%)	—
Zn	not certified	ND	—	ND	—
Ca	8450 ± 500 (6%)	9103 ± 241 (3%)	108	9063 ± 1098 (12%)	107
SRM 3243
Fe	760 ± 160 (21%)	787 ± 27 (3%)	104	650 ± 20 (3%)	86
Na	1960 ± 140 (7%)	1915 ± 98 (5%)	98	2283 ± 48 (2%)	116
P	6800 ± 1000 (15%)	6863 ± 120 (2%)	101	7070 ± 464 (7%)	104
Zn	3250 ± 310 (10%)	3851 ± 102 (3%)	118	3993 ± 350 (9%)	123
Ca	1.03 ± 0.05^a (5%)	1.06 ± 0.14 (13%)	103	1.44 ± 0.21 (15%)	140

---

Table 6 Validation results for heavy metal elements in NIST SRMs 3241, 3243, and 3246

Element	Certified values (mg kg^−1)	Calculated values (mg kg^−1)	% Recovery	Calculated values (mg kg^−1)	% Recovery
		Aqueous calibration standards		Standard addition method
(X%) = relative standard deviation.
SRM 3241
As	1.29 ± 0.08 (6%)	1.27 ± 0.01 (1%)	99	1.332 ± 0.081 (6%)	104
Cd	0.0587 ± 0.0036 (6%)	ND	—	ND	—
Pb	0.241 ± 0.012 (5%)	ND	—	ND	—
Hg	0.00383 ± 0.00029 (8%)	ND	—	ND	—
SRM 3243
As	0.554 ± 0.018 (3%)	0.612 ± 0.035 (6%)	110	0.599 ± 0.013 (2%)	108
Cd	0.122 ± 0.003 (2%)	ND	—	ND	—
Pb	0.692 ± 0.056 (8%)	0.764 ± 0.086 (11%)	114	0.671 ± 0.112 (17%)	97
Hg	0.00900 ± 0.00044 (5%)	ND	—	ND	—
SRM 3246
Cd	20.8 ± 1.0 (5%)	ND	—	ND	—
Pb	995 ± 30 (3%)	1121.0 ± 0.2 (0.01%)	112	1063 ± 154 (14%)	107
Hg	23.1 ± 0.2 (1%)	ND	—	ND	—

---

Quantification of botanical extracts

After completion of the method validation, the three different matrices of botanical samples underwent microwave digestion and were analyzed for As, Cd, Hg, Pb, Fe, Na, Ca, P and Zn by ICP-OES. Tables 7–9 show the concentration values obtained for the elements of interest from the powdered raw material, glycerine-based and ethanolic tinctures samples, respectively. The toxic metals (As, Cd, Hg, Pb) were not detected (ND) in the glycerin-based samples and the ethanolic tinctures, indicating their safety. In the case of powdered raw materials, a few of the samples (for example; Bilberry P.E. and Burdock Root) provided detectable levels of As and Pb. Because, in many situations the powdered-raw materials are employed for the production/preparation of other consumable matrices (e.g, capsules, tablets, tinctures) the amount of the heavy metals would have to be accounted for in the final preparation.

Table 7 Elemental composition of powdered raw botanical product materials

Sample	Fe (µg g^−1)	Na (µg g^−1)	P (µg g^−1)	Zn (µg g^−1)	Ca (µg g^−1)	As (µg g^−1)	Cd (µg g^−1)	Pb (µg g^−1)	Hg (µg g^−1)
Rhodiola Rosea	8.56 ± 0.26	99.5 ± 3.0	93.2 ± 3.0	ND	727 ± 5	ND	ND	ND	ND
Milk Thistle Dry Extract	3.71 ± 0.43	4.12 ± 0.48	1354 ± 36	ND	326 ± 88	ND	ND	ND	ND
Bayberry P.E.	414 ± 13	137 ± 3	230 ± 6	ND	184 ± 3	ND	ND	ND	ND
Ashwagandha Powder	28.2 ± 1.7	285 ± 6	833 ± 20	7.48 ± 0.54	85.6 ± 3.5	ND	ND	ND	ND
Hawthorn Berry	47.5 ± 1.6	ND	618 ± 15	ND	1587 ± 15	ND	ND	0.750 ± 0.170	ND
Licorice Root	242 ± 4	522 ± 12	347 ± 27	ND	22295 ± 357	ND	ND	ND	ND
Green Tea	140 ± 2	ND	3622 ± 38	ND	2805 ± 61	ND	ND	1.10 ± 0.23	ND
Schizandra Berry	55.1 ± 2.1	12.7 ± 0.3	2036 ± 42	ND	961 ± 10	ND	ND	ND	ND
Chinese Skullcap	51.4 ± 1.0	1893 ± 29	1432 ± 18	ND	2844 ± 65	ND	ND	ND	ND
Bilberry P.E.	131 ± 2	33.5 ± 0.7	502 ± 6	ND	81.7 ± 2.5	ND	ND	24.5 ± 0.9	ND
Tribulus	43.4 ± 3.4	16.3 ± 0.5	4177 ± 94	ND	46537 ± 1616	ND	ND	ND	ND
Bitter Orange Peel	137 ± 4	47.2 ± 0.7	726 ± 26	ND	14269 ± 120	ND	ND	ND	ND
Holy Basil	433 ± 7	6548 ± 60	3632 ± 119	15.6 ± 0.6	24278 ± 280	0.966 ± 0.353	ND	ND	ND
Burdock Root	189 ± 5	3.37 ± 0.70	949 ± 18	ND	3684 ± 74	0.877 ± 0.190	ND	ND	ND

---

Table 8 Elemental composition of glycerine-based botanical product extracts

Sample	Fe (µg g^−1)	Na (µg g^−1)	P (µg g^−1)	Zn (µg g^−1)	Ca (µg g^−1)	As (µg g^−1)	Cd (µg g^−1)	Pb (µg g^−1)	Hg (µg g^−1)
Whole Body Defense	51.6 ± 1.3	628 ± 10	535 ± 16	ND	1878 ± 20	ND	ND	ND	ND
Echinacea Goldenseal	ND	153 ± 2	434 ± 21	ND	501 ± 5	ND	ND	ND	ND
Holy Basil	ND	2187 ± 28	474 ± 16	ND	42.6 ± 3.4	ND	ND	ND	ND
Saw Palmetto	ND	3.19 ± 0.26	14710 ± 223	ND	511 ± 17	ND	ND	ND	ND
Milk Thistle Seed	ND	34.8 ± 0.5	4885 ± 47	ND	170 ± 6	ND	ND	ND	ND
Antioxidant Supreme	1.68 ± 0.22	93.8 ± 1.9	1089 ± 43	ND	346 ± 11	ND	ND	ND	ND
Kava Kava Root	ND	95.8 ± 1.5	5728 ± 126	ND	239 ± 3	ND	ND	ND	ND
Cinnamon	ND	ND	712 ± 16	ND	143 ± 3	ND	ND	ND	ND
Ginkgo	ND	ND	258 ± 13	ND	42.2 ± 1.9	ND	ND	ND	ND
Thyroid Support	15.9 ± 0.9	6742 ± 68	1411 ± 45	ND	451 ± 9	ND	ND	ND	ND
Green Tea	ND	ND	660 ± 20	ND	ND	ND	ND	ND	ND
Valerian	1.23 ± 0.74	269 ± 4	719 ± 21	ND	74.3 ± 2.0	ND	ND	ND	ND
Olive Leaf	ND	30.3 ± 0.6	133 ± 5	ND	131 ± 4	ND	ND	ND	ND
Motor Oil for Men w/ Zn	2.52 ± 0.24	345 ± 3	4955 ± 41	7247 ± 105	1462 ± 15	ND	ND	ND	ND
Male Libido	3.24 ± 0.33	101 ± 4	515 ± 25	ND	405 ± 14	ND	ND	ND	ND

---

Table 9 Elemental composition of ethanolic tinctures of botanical products

Sample	Fe (µg g^−1)	Na (µg g^−1)	P (µg g^−1)	Zn (µg g^−1)	Ca (µg g^−1)	As (µg g^−1)	Cd (µg g^−1)	Pb (µg g^−1)	Hg (µg g^−1)
Maitake Gold	2.37 ± 0.19	ND	390 ± 11	ND	4.63 ± 0.72	ND	ND	ND	ND
Milk Thistle	ND	15.6 ± 0.6	365 ± 11	ND	16.7 ± 1.1	ND	ND	ND	ND
Lobelia Herb & Seed	4.70 ± 0.34	30.7 ± 0.6	96.1 ± 6.9	ND	447 ± 11	ND	ND	ND	ND
Ginkgo Leaf Extract	7.21 ± 0.29	ND	949 ± 19	ND	465 ± 8	ND	ND	ND	ND
Sangre de Drago	ND	ND	24.4 ± 1.8	ND	192 ± 4	ND	ND	ND	ND
Sanito	53.5 ± 0.9	13.3 ± 0.4	168 ± 12	ND	267 ± 5	ND	ND	ND	ND
Valerian Root	4.38 ± 0.20	10.9 ± 0.2	93.0 ± 5.6	ND	23.1 ± 1.4	ND	ND	ND	ND
Echinacea Supreme	ND	86.9 ± 0.4	172 ± 2	ND	65.8 ± 1.7	ND	ND	ND	ND
Licorice Root Extract	ND	629 ± 10	95.5 ± 7.9	ND	355 ± 8	ND	ND	ND	ND

---

In order to corroborate the fact that the ND assignments for many of the heavy metals were not the result of systematic errors, each of the botanical samples was spiked with a standard aqueous solution containing each of the test elements prior to the addition of nitric acid and the microwave digestion. Table 4 also shows the recovery values obtained for each of the elements for the three sample matrix types. Recoveries of 90% and higher were observed for each of the elements with the exception of mercury, which resulted in a 61% recovery for the different sample matrices. The uniformity of the elemental recoveries across the different matrix forms is firm validation of the efficacy and utility of the developed digestion procedure. The loss of mercury during the experiments could be due to the volatility of the element or adsorption to the digestion vessel walls or the components of the ICP sample introduction system. Based on the fact that the recovery for Hg was the same as the other elements in the case of the aqueous standard solutions (Table 4), it seems quite clear that volatile Hg species are formed in the initial nitric acid decomposition of organomercury compounds prior to the sealing of the microwave vessels. Unfortunately, processing in this manner is required as the mixture of HNO₃, with ethanol in particular, is quite rapid and exothermic. There may be some improvement in Hg recoveries by using lower concentrations of the acid, but this would occur at the expense of longer digestion times.

Conclusions

A single microwave digestion method has been successfully applied for the elemental analysis of three different botanical matrices (powder raw material, glycerin-based samples and ethanolic tinctures) by ICP-OES. In addition, method validation was carried out by external calibration and standard addition using three NIST standard reference materials. Both calibration techniques provided good results, but due to the high number of samples, the external calibration was the technique of choice. Recovery results obtained by the addition of element standard solutions to the botanical matrices prior to addition of nitric acid and microwave digestion and carried through every step demonstrate that the presented methodology is uniform and can be applied for the elemental analysis of different botanical product matrices.

Acknowledgements

The authors would like to thank Clemson University Institute of Nutraceutical Research (INR), Gaia Herbs (Brevard, NC, USA), and the Analytical Chemistry Division of the National Institute for Standards and Technology (NIST) for support of this research.

References

1. U.S. Nutrition Industry Overview, New Hope Natural Media, Penton Media, Inc., 2008.
2. C. Noonan and W. P. Noonan, Toxicology, 2006, 221, 4–8.
3. S. Melethil, Life Sci., 2006, 78, 2049–2053.
4. P. Raman, L. C. Patino and M. G. Nair, J. Agric. Food Chem., 2004, 52, 7822–7827.
5. Proposition 65: Local Taxpayers and Public Safety Protection Act: The Summary Report from the Informational Hearing, California Senate Publications, Sacramento, CA, 2004.
6. Safe Drinking Water and Toxic Enforcement Act of 1986: Proposition 65., Office of Environmental Health Hazard Assessment, State of California, 2003.
7. N. J. Miller-Ihli, J. Agric. Food Chem., 1996, 44, 2675–2679.
8. A. Lozak, K. Soltyk, P. Ostapczuk and Z. Fijalek, Sci. Total Environ., 2002, 289, 33–40.
9. S. P. Dolan, D. A. Nortrup, P. M. Bolger and S. G. Capar, J. Agric. Food Chem., 2003, 51, 1307–1312.
10. R. C. Richter, D. Link and H. M. Kingston, Anal. Chem., 2001, 73, 30A–37A.
11. A. Krejcova, D. Kahoun, T. Cernohorsky and M. Pouzar, Food Chem., 2006, 98, 171–178.
12. J. Sucharova and I. Suchara, Anal. Chim. Acta, 2006, 576, 163–176.
13. M. Bettinelli, G. M. Beone, S. Spezia and C. Baffi, Anal. Chim. Acta, 2000, 424, 289–296.
14. C. Moor, T. Lymberopoulou and V. J. Dietrich, Mikrochim. Acta, 2001, 136, 123–128.
15. J. R. Dean, in Practical Inductively Coupled Plasma Spectroscopy, ed. D. J. Ando, John Wiley and Sons, Ltd., West Sussex, England, 1st edn., 2005, pp. 17–36.
16. L. B. Allen, P. H. Siitonen and H. C. Thompson, J. Agric. Food Chem., 1997, 45, 162–165.
17. C. G. Millward and P. D. Kluckner, J. Anal. At. Spectrom., 1989, 4, 709–713.
18. M. Tighe, P. Lockwood, S. Wilson and L. Lisle, Commun. Soil Sci. Plant Anal., 2004, 35, 1369–1385.
19. L. C. Sander, K. E. Sharpless and S. A. Wise, Life Sci., 2006, 78, 2044–2048.
20. K. E. Sharpless, D. L. Anderson, J. M. Betz, T. A. Butler, S. G. Capar, J. Cheng, C. A. Fraser, G. Gardner, M. L. Gay, D. W. Howell, T. Ihara, M. A. Khan, J. W. Lam, S. E. Long, M. McCooeye, E. A. Mackey, W. R. Mindak, S. Mitvalsky, K. E. Murphy, A. NguyenPho, K. W. Phinney, B. J. Porter, M. Roman, L. C. Sander, M. B. Satterfield, C. Scriver, R. Sturgeon, J. B. Thomas, R. D. Vocke, Jr., S. A. Wise, L. J. Wood, L. Yang, J. H. Yen and G. C. Ziobro, J. AOAC Int., 2006, 89, 1483–1495.

---