Optimizing the NIST Standard Sequential Extraction Protocol Using a Full-Factorial Experimental Design

Michael K. Schultz¹, Kenneth G. W. Inn², and William C. Burnett³.

1. PerkinElmer Instruments, 801 S. Illinois Avenue, Oak Ridge, TN 37831-0895, michael.schultz@perkinelmer.com , Phone: (843) 841-2446.

2. National Institute of Standards and Technology, Radioactivity Group, 100 Bureau Drive, Mailstop 8461, Gaithersburg, MD 93001, kenneth.inn@nist.gov, Phone: (301) 975-5541.

3. Florida State University, Department of Oceanography, Tallahassee, FL 32306-4230, burnett@ocean.fsu.edu, Phone: (850) 644-6703.

Extended Abstract

The NIST Standard Sequential Extraction Protocol

We are developing a Standard Sequential Extraction Protocol, termed the NIST Standard Sequential Extraction Protocol (NSSEP), for determining the geochemical fractionation of uranium and plutonium in soils and sediments. Sequential extraction methods are used to investigate geochemical relationships of actinides and trace metals in soils and sediments (Tessier, et al. 1979). These geochemical relationships are used to understand the conditions under which contaminants may be released to the surrounding environment.

The sequential extraction approach employs a series of operationally-defined chemical extractions in an attempt to dissolve selectively discrete geochemical components of soils and sediments. For example, analyte metals that may become environmentally available through changes in ionic strength can be identified by the application of an ion-exchange solution such as MgCl₂. As a further example, analyte metals that are occluded in or irreversibly adsorbed to the lattice structure of oxides of Fe and Mn can be identified by the application of a reducing solution (e.g. hydroxylamine hydrochloride NH₂OH·HCl).

Unfortunately, interpretation of sequential extraction results can be complicated by several factors. Although each extraction in the sequence is designed to attack a single geochemical phase, complete specificity is not likely. We developed an experimental plan to evaluate systematically the experimental conditions under which extraction selectivity could be optimized for uranium and plutonium in a NIST Standard Reference Material (SRM). The NSSEP includes five chemical extractant solutions (Table 1) ¾ arranged in a sequence designed to maximize the dissolution of the target phase, while minimizing the destruction of non-targeted phases. Stable metal analyses were combined with radiometric analyses to determine the specificity of reactions (Schultz, et al., 1996, Schultz et al., 1998a, Schultz et al., 1998b, Schultz et al., 1999).

The NSSEP is being developed by NIST to establish a standard method for comparison of the results of sequential extraction experiments. The method is being developed using NIST Standard Reference Materials SRM’s . The experiments described here were conducted using NIST SRM 4357 Ocean Sediment. The Ocean Sediment material is a mixture of contaminated material obtained from the Irish Sea and relatively uncontaminated material collected from the Chesapeake Bay. The resulting standard consists of isotopic plutonium concentrations well in excess of fallout and naturally-occurring isotopic uranium concentrations that are easily measured by alpha-spectrometry.

Table 1 Summary of the extraction sequence being tested for the development of the NIST Standard Sequential Extraction Protocol.

Desired Fraction	Extractive Reagent	Reagent/Sample Ratio (mL/g)
Water Soluble/ Exchangeables	H₂O/MgCl₂ pH 4.5	15:1
Carbonates	NH₄Ac in 25% Hac pH 4	15:1
Oxides (Fe/Mn)	NH₂OH·HCl in 25% Hac pH 2 (HNO₃)	15:1
Organic matter	30% H₂O₂ in 0.02 M HNO₃ pH 2	15:1
Acid Soluble	8M HNO₃	15:1
Residue	Total dissolution	as needed

Design of Experiment ¾ The Full-Factorial Approach

In this paper, we present some of the fundamentals of so-called Design of Experiment DEX ¾ as they relate to the development of the NSSEP. The results of the optimization experiments conducted are used as examples of the DEX approach.

We define DEX as a rigorous-systematic approach for solving scientific and engineering problems so that unambiguous results are obtained at minimum cost. The main elements of our approach include:

· Establishing criteria for response success.

· Establishing a level of statistical significance and eliminating “unimportant” variables.

· Establishing high and low settings for variables.

· Conducting experiments that simultaneously include three variables.

· Conducting auxiliary experiments.

· Analyzing the results

Using the DEX approach, reaction time, reaction temperature, and extractant concentration were identified as experimental variables ¾ it was determined that changes in these variables would likely result in identifiable changes in the extraction of targeted geochemical fractions. The phase specificity of the extractions was monitored by measuring the concentration of stable elements Al, Ba, Ca, Fe, K, Mn, Pb, Sr, and Ti in extractant solutions. Extraction efficiency was determined at two extreme values for each variable and a midpoint experiment was included to identify curvature in the relationship. The resulting full-factorial design provides for an estimation of the effect (change in the % extracted U, Pu or stable metal), as the independent variables (time, temperature, and reagent concentration) are varied from a low to a high value. Once the full factorial experiment has been completed, the response “y” can be estimated mathematically by the equation:

Eq 1

where:

y = response for an individual experimental run.

m = average response over all experimental runs.

b_tm = Average effect of reaction time.

b_tp = Average effect of reaction temperature.

b_c = Average effect of reagent concentration.

c_tm = Duration of experimental run.

c_tp = Temperature of experimental run.

c_c = Reagent concentration of experimental run.

This model includes not only effects from changes in each variable, but also effects of variable interactions.

Graphically, the design becomes a three dimensional figure with the axes representing the independent variables time, temperature, and concentration (Fig. 1). The origin represents experiments conducted at the “low” value (coded “-”) for each (e.g., 1 hour, 25°C, 0.1 M MgCl₂ for the exchangeable fraction). The opposite end of each axis represents the high value (coded “+”) for each variable (e.g. 4 hours, 90°C, and 1 M MgCl₂. This paper will illustrate the determination of the optimum conditions for the sequential extraction of a soil material. This results in nine total experimental runs per fraction (including the mid-point). For the NSSEP, the experimental runs were randomized to minimize the effect of learning curve on the results.

Results Summary

Using the DEX approach we have completed the optimization of the first NIST material SRM 4357 Ocean Sediment (Table 2).

Our results indicate that over 60% of natural uranium (as U-238) is associated with a residual fraction ¾ as determined by a sodium hydroxide fusion of the residual material that remained following the extraction procedure (Fig 2a). Plutonium, on the other hand, is distributed primarily among the reducible, organic and nitric-acid extractable fractions in this sediment (Fig 2b). We present selected results from the optimization experiments. A future publication will include a complete interpretation of the results.

Table 2 Summary of the optimized parameter settings for each fraction.

Fraction	Temperature (°C)	Concentration (M)	Duration (hours)
MgCl₂	25	0.4	1
NH₄Ac (pH 5)	25	1.0	2.0
NH₂OH-HCl / 25% HAc	70	0.1	6
H₂O₂ / 0.05M HNO₃	50	pH 2	4
HNO₃	90	4	4

Figure 2. Extraction profiles of U-238 and Pu-239/240 as determined by the NSSEP. The x-axis represents the percent total extracted for four complete runs. The y-axis shows the fractions in order of extraction.

Reference

Box, G. E. P., Hunter W. G. and Hunter, J. S. (1978) Statistics for Experimenters: An Introduction to Design Analysis and Model Building. John Wiley and Sons, New York, NY.

Schultz, M. K., Burnett, W. C., Inn, K. G. W., Thomas, J. W. L. and Lin, Z. (1996) Partitioning of radioactive elements in NIST natural-matrix standards. J. Res. Nat. Inst. of Stand. and Technol. 101, 707-715.

Schultz, M. K., Inn, K. G. W., Lin, Z. C., Burnett, W. C, Smith, G. E., Biegalski, S. R. and Filliben, J. (1998a) Identification of radionuclide partitioning in soils and sediments: Determination of the optimum conditions for the exchangeable fraction of the NIST standard sequential extraction protocol. Appl. Radiat. Isot. 49, 9-11, 1289-1293.

Schultz, M. K., Burnett, W. C. and Inn, K. G. W. (1998b) Evaluation of a sequential extraction method for determining actinide fractionation in soils and sediments. J. Environ. Rad. 40, 2, 155-174.

Schultz, M. K., Biegalski, S. R., Inn, K. G. W., Yu, L., Burnett, Thomas, J. L. W., and Smith, G. E. (1999) Optimizing the removal of carbon phases in soils and sediments for sequential chemical extractions by coulometry. J. Environ. Monit., 1, 183-190.

Tessier, A., Campbell, P. and Bison, M. (1979) Sequential extraction procedure for the speciation of particulate trace metals. Anal. Chem. 51, 7, 844-850.