Calibration, uncertainty, and recovery in the chromatographic sciences.

This paper reviews calibration-, uncertainty-, and recovery-related documents from 10 consensus-based organizations. The main points from each treatise are summarized. Also included is a critique of the various approaches, as well as recommendations for a statistically sound protocol that is more compatible with chromatographic data.

Perchlorate in water via US Environmental Protection Agency Method 331 Determination of method uncertainties, lowest concentration minimum reporting levels, and Hubaux-Vos detection limits in reagent water and simulated drinking water.

US Environmental Protection Agency (EPA) Method 331 determines perchlorate in drinking water using non-suppressed ion chromatography with tandem mass spectrometry. This study reports the results of calibration and recovery studies in reagent water, as well as of a recovery study in simulated drinking water (i.e., total dissolved solids are 500 mg/mL each of chloride, sulfate, and bicarbonate). The perchlorate concentrations in the study ranged from 0.05 to 64 ng/mL. At 95% confidence, the Hubaux-Vos detection limit (H-V DL) was 0.04 ng/mL for the calibration study and the simulated-drinking-water recovery study, and 0.03 ng/mL for the reagent-water recovery study. The lowest concentration minimum reporting level was 0.03 ng/mL for reagent water and 0.0 7 ng/mL for simulated drinking water, again at 95% confidence.

Comparison of two cryptand separator columns for the determination of trace chloride in semiconductor-grade nitric acid.

Improved ion-chromatographic approaches for measuring trace chloride in nitric acid are presented. Two columns, the IonPac Cryptand A1 and a higher-capacity Cryptand prototype, were tested and compared. Also, the use of a Continuously Regenerated Anion Trap Column (CR-ATC) was evaluated for its ability to purify electrolytically generated eluent. Nitric acid (70%) was used as the test matrix and chloride was used as the test analyte; prior to injection, the nitric acid was diluted to 0.7% for the A1 column and to 2.8% for the prototype column. Chloride could be quantified in only 20 min on either column; detection limits computed for 70% HNO3 (at 95% confidence, alpha = beta = 2.5%) were 1.8 and 1.5 ppm for the A1 and prototype columns, respectively. Results also showed that the CR-ATC was necessary for obtaining acceptable acid blanks.

Ion-pair chromatography of bis (sodium-sulfopropyl) disulfide brightener in acidic copper plating baths.

Quantitative analysis of the brightener component bis (sodium-sulfopropyl) disulfide (SPS) in acidic copper plating baths poses a real challenge due to the complex chemical matrix containing large amounts of Cu(II) ion and sulfuric acid together with other organic additives and additive decomposition products. We developed a new ion-pair chromatography method to analyze micro-molar amounts of SPS directly in plating bath samples without the need for sample pre-treatment. Addition of tetra-N-methylammonium cation as ion-pairing agent to a methanol-sulfuric acid-water eluent increases the retention time of the anionic SPS2- on a C18 column sufficiently to separate this compound from Cu(II) ion and additive by-products.

Determination of chloride and sulfate in semiconductor-grade etchants comprised of acetic acid, nitric acid and phosphoric acid.

The variable-capacity Dionex Cryptand A1 column was used for the determination of low-ppm levels of chloride and sulfate in etchants comprised of acetic acid, nitric acid and phosphoric acid. All possible ratios of the three acids could be analyzed for chloride and sulfate, if the samples were first diluted 1:100. However, a suitable eluent program was found to be needed for each mixture. A proprietary formulation was chosen to undergo this suitability determination. The resulting gradient was 10 mM KOH with a step to 30 mM NaOH at 15 min, flow-rate=0.5 ml/min; column temperature=29 degrees C; sample loop=7.5 microl. Under these conditions, a low-ppm calibration study (using the proprietary mix as the matrix) was performed and the associated prediction intervals were determined. At 50 ppm (in the original etchant), the +/- prediction interval was +/- 7 ppm for chloride and +/- 20 ppm for sulfate, both at the 95% confidence level. This step gradient was found to be a good starting place for separating the five components in all other ratios of these three acids.

Reproducible digestion method for ion-chromatographic analysis of anions in 30% hydrogen peroxide.

Semiconductor-grade hydrogen peroxide (30%) is analyzed for anion contamination to be certain the levels of each analyte do not exceed 30 ppb (w/w). This paper presents a reproducible, platinum-decomposition approach that uses ion chromatography to quantify the various analytes (fluoride, chloride, sulfate, bromide, nitrate, and phosphate). Important to the success of the method are: (1) use of disposable HDPE bottles for the digestion, (2) immersion of the bottles in a water bath, (3) careful re-mixing of the digesting peroxide after two (of six total) hours, and (4) careful clean-up and sample-handling procedures (to avoid contamination). Except for fluoride and nitrate, all analytes exhibited recoveries from 89.6 to 98.3%, with +/- prediction intervals (at the 95% confidence level) from 1.5 to 3.0 ppb. Fluoride's recovery was low (74.9%), but reproducible (+/- prediction interval at 95% confidence=2.0 ppb). Nitrate reeovery was 99.1%, but noisy (+/- prediction interval at 95% confidence=8.7 ppb); this imprecision was thought to be due to contamination from atmospheric nitrite. A Dionex DX 500 microbore ion chromatograph with AS15 column and 1-ml sample loop were used for all determinations; detection was by conductivity. Statistical analyses were performed using JMP software.

Ion-chromatographic analysis of common anions, acetate, and formate in 30% hydrogen peroxide statistical evaluation of two automated microbore systems.

Two hydroxide-selective microbore analytical columns (the Dionex AS11 and AS15) were tested and compared for the quantitation of anionic species in 30% hydrogen peroxide. The ions of interest were fluoride, acetate, formate, chloride, bromide, nitrate, sulfate, and phosphate. Statistically sound calibration and spiking studies were carried out, investigating the range of a blank to 60 ppb. Prior to injection onto the separators, peroxides were loaded without pretreatment onto a concentrator column, which was then washed with deionized water to remove the matrix. Although retention times gradually decreased during the spiking studies, reliable quantitation was still achievable on both columns at the target concentration of 30 ppb. However, various resolution problems meant that the AS11 should not be recommended for this application.

Ion-chromatographic assay of chromium(VI)-containing semiconductor etchants statistical comparison of two columns.

Semiconductor etchants are concentrated-acid mixtures that are prepared under tight specifications. Assay procedures are needed to ensure that the proportion of each component is within a small percentage (usually 10 relative percent or less) of the target concentration. One such etchant class contains chromium trioxide, usually in combination with HF and HNO3, While several ion-chromatographic columns can be used to analyze most mineral acids, chromium(VI) presents a problem. This latter species is highly retained by many separators and may also degrade the resins. This paper compares two columns that showed potential for success with these assays: the AS 11 and the AS16 separators (both from Dionex). These columns permit elution of chromium(VI) as chromate in 25 min or less. A representative mixture of HF, HNO3 and CrO3 was used in the research. Simultaneous calibration studies were conducted and the data sets analyzed statistically. Also investigated was the effect on the columns of repeated exposure to chromate.

Lack-of-fit testing of ion chromatographic calibration curves with inexact replicates.

Calibration studies involve the preparation and analysis of replicates for multiple concentrations of standards. Curves that are fitted through the data are evaluated for their adequacy of fit. A helpful test is a lack-of-fit procedure, which is performed easily by most statistical software. When coupled with Radj2, the procedure differentiates between data that are not linear and those that are simply noisy. The test requires data from exact replicates of the various standard levels involved. However, in ppt-level ion chromatography, the above condition may be impossible to meet. With the common anions (e.g., chloride, nitrite), the working standards must be prepared by mass and all liquids must be poured; transfer pipets contaminate at these concentrations. Since it is virtually impossible to pour out the desired mass exactly, final concentrations will vary slightly. Consequently, a different approach is needed for lack-of-fit testing. This paper discusses reasonable alternatives and applies them to actual data.

Low-level calibration study for a new ion chromatographic column to determine borate in deionized water.

In the semiconductor industry, there is interest in determining borate at sub-ppb levels in ultrapure water, since borate is an early breakthrough ion from ion-exchange resin beds. Although dissolved silica is the most common species currently used to monitor the breakdown of the deionization systems, it is thought that borate probably breaks through earlier than silicate. To be of use as an early-warning indicator, borate must be determined at ppt levels. This paper discusses benchtop results with several new column products designed to deliver low-ppt detection limits for boron as borate. The system uses a prototype borate-specific concentrator column that is coupled to an ion-exclusion separator and suppressed-conductivity detection. The acidic eluent, containing mannitol, quantitatively elutes the borate from the concentrator. The analytical separation is performed using a specially designed ion-exclusion column. Data presented are from two multilevel calibration studies. Included is a discussion of detection-limit calculations and recommended formats for reporting results.

Structure of Gialpha1.GppNHp, autoinhibition in a galpha protein-substrate complex.

The structure of the G protein Gialpha1 complexed with the nonhydrolyzable GTP analog guanosine-5'-(betagamma-imino)triphosphate (GppNHp) has been determined at a resolution of 1.5 A. In the active site of Gialpha1. GppNHp, a water molecule is hydrogen bonded to the side chain of Glu43 and to an oxygen atom of the gamma-phosphate group. The side chain of the essential catalytic residue Gln204 assumes a conformation which is distinctly different from that observed in complexes with either guanosine 5'-O-3-thiotriphosphate or the transition state analog GDP.AlF4-. Hydrogen bonding and steric interactions position Gln204 such that it interacts with a presumptive nucleophilic water molecule, but cannot interact with the pentacoordinate transition state. Gln204 must be released from this auto-inhibited state to participate in catalysis. RGS proteins may accelerate the rate of GTP hydrolysis by G protein alpha subunits, in part, by inserting an amino acid side chain into the site occupied by Gln204, thereby destabilizing the auto-inhibited state of Galpha.

Olanzapine-related fatality.

A 43-year-old male psychiatric outpatient died within hours of ingesting as much as 600 mg of olanzapine, a newer antipsychotic agent related to clozapine. Analysis of postmortem blood and urine by gas chromatography with nitrogen-selective detection yielded olanzapine concentrations of 1238 and 6987 micrograms/L, respectively, greatly in excess of levels expected following therapeutic administration of the drug. Based on the toxicology findings, the decedent's known history of suicide attempts, and the circumstances surrounding the death, this case was ruled a suicide by olanzapine overdosage.

Crystal structures of the G protein Gi alpha 1 complexed with GDP and Mg2+: a crystallographic titration experiment.

The effect of Mg2+ binding on the conformation of the inactive GDP-bound complex of the heterotrimeric G protein alpha subunit Gi alpha 1 has been investigated by X-ray crystallography. Crystal structures of the Gi alpha 1.GDP complex were determined after titration with 5, 10, 100, and 200 mM Mg2+. Comparison of these structures with that of the Mg2+-free complex revealed Mg2+ bound at the same site as observed in the structure of the active, Gi alpha 1. GTP gamma S.Mg2+-bound complex of Gi alpha 1, with a similar coordination scheme except for the substitution of a water molecule for an oxygen ligand of the gamma-phosphate of Gi alpha 1.GTP gamma S. Mg2+. In contrast to the GDP.Mg2+ complex of Gt alpha and of other G proteins, switch I residues of Gi alpha 1 participate in Mg2+ binding and undergo conformational changes as a consequence of Mg2+ binding. Partial order is induced in switch II, which is disordered in the Mg2+-free complex, but no order is observed in the switch III region. This contrasts with the GDP.Mg2+ complex of Gt alpha in which both switch II and III switch are ordered. Mg2+ binding also induces binding of an SO42- molecule to the active site in a manner which may mimic a Gi alpha 1.GDP.PO42-.Mg2+ product complex. Implications of these findings are discussed.

The cytotoxic action of 1,3,5-triazabicyclo[3.1.0]hexane-2,4-diones and 1,3,5-triazine-4,6-(1 H,5 H)-diones in murine and human tumor cells.

1,3,5-Triazabicyclo[3.1.0]hexane-2,4-diones proved to be potent antineoplastic and cytotoxic agents in murine and human cancer cells. In L1210 lymphoid leukemia cells DNA synthesis was significantly suppressed over 60 min by the agents from 25 to 100 microM. DNA synthesis was blocked at multiple sites including DNA polymerase alpha, ribonucleoside reductase, dihydrofolate reductase, PRPP-amido transferase, and nucleoside kinases which would be additive overall in suppressing DNA synthesis. The DNA molecule itself did not appear to be at target of the agents since no alkylation of nucleotide bases, intercalation between base-pairs or cross-linking of strands occurred after 24 h incubation at 100 microM. Nevertheless, L1210 DNA fragmentation did occur after 24 h incubation at 100 microM which is usually associated with tumor cell apoptosis.

Efficacy of two commercial products for altering urine drug test results.

OBJECTIVE: We have become aware of several commercial products that, when orally ingested, will purportedly not only eliminate "toxins" from a person's system, but will also correct any urinary imbalances caused by excessive water consumption. METHOD: Unblinded study of one volunteer subject, tested weekly x 4 for 24-hour urine elimination of test drug under conditions of control, control plus 1200 mL water, Quick Flush', and Eliminator. RESULTS: Each of the treatment protocols studied caused reductions of drug or metabolite concentrations as measured by gas chromatography-mass spectrometry in urine specimens collected up to 24 hours after ingestion of amphetamine, 9-carboxy-11-nor-delta-9-THC, benzoylecgonine, or codeine, yet the radioimmunoassay screening results demonstrated very little effect. Water alone was approximately as effective as the two commercial products in reducing the metabolite level. None of the treatment protocols employed in this study altered urinary pH, specific gravity, or creatinine concentration outside the normally accepted physiological range. CONCLUSIONS: Attempts to conceal drug abuse by water dilution are most likely to play a substantial role when concentrations are at or near the detection threshold for a particular assay such as the terminal stages of drug eliminations.

Structural and biochemical characterization of the GTPgammaS-, GDP.Pi-, and GDP-bound forms of a GTPase-deficient Gly42 --> Val mutant of Gialpha1.

The Gly42 --> Val mutant of Gialpha1 was characterized structurally and biochemically to elucidate two important features of Gialpha1-catalyzed GTP hydrolysis. The crystal structure of the GTPgammaS-bound G42VGialpha1 protein demonstrates that the steric bulk of Val42 pushes the Gln204 residue into a catalytically incompetent conformation, providing a rationale for the diminished GTPase activity of this mutant. The same phenomenon may also account for the diminished GTPase activity of the homologous transforming Gly42 --> Val mutation in p21(ras). Similarly, the steric bulk of the unique Ser42 residue in Gzalpha may account for the comparatively slower rate of GTP hydrolysis by this Galpha subunit. The G42VGialpha1 subunit was also characterized structurally in its GDP.Pi- and GDP-bound states, providing a unique opportunity to view three "snapshots" of GTP hydrolysis. Hydrolysis of GTP to a transient GDP.Pi-bound intermediate is associated with substantial conformational changes in the switch II segment of the protein. Eventual release of Pi results in further removal of switch I from the active site and a highly mobile switch II segment. Despite their disparate biochemical properties, the structural similarity of G42VGialpha1 to the G203AGialpha1 mutant in the GDP.Pi-bound form suggests that both mutations stabilize a conformation of the GDP. Pi-bound protein that occurs only transiently in the wild-type protein. The structures of the GDP-bound forms of the wild-type and mutant proteins are similar.