Molecular Size-Selective Electrodialytic Desalters for Electrospray Ionization-Mass Spectrometry.

A membrane-based electrodialytic desalter has been developed for the selective removal of buffer/salt constituents from a fluid stream while retaining larger charged molecules such as proteins prior to mass spectrometric (MS) detection. The salts are deleterious to MS causing signal suppression, formation of adducts, and eventual contamination of the inlet. The new device uses dialysis membranes (DMs) paired with ion exchange membranes (IEMs) flanking the carrier flow channel in a planar configuration. The DMs contact the carrier channel preventing adsorptive losses of large, charged molecules to the IEMs. Ions are removed under an applied electric field using four pairs of electrodes along the flow channel. Removal of both anions and cations is more energy intensive than conversion of a suitable MS friendly salt into its respective acid or base, for example, ammonium acetate into acetic acid. The energetics and optimal voltage profiles for both scenarios have been thoroughly investigated. The DMs resulted in nonlinear increases in energy required for desalting over standard IEM devices due to electroosmotic flow of water into the interstitial space between the membranes. For a device channel with nominal volume of 15.2 µL, a maximum concentration of 200 mM ammonium acetate flowing at 0.25 mL/min was converted into acetic acid. Recovery of bovine serum albumin measured at 280 nm was 67%-96% at tested salt concentrations, and dispersion volumes were less than 200 µL2 and may be suitable for coupling to liquid chromatography.

Indirect Potentiometric pH Detection of Weak Acids with Absolute Quantitation by a Theoretical Approach.

A fast response potentiometric flow-through pH sensor was applied for organic acid determination. The analyte response with the pH sensor was obtained by eluent pH modification following ion exclusion chromatography with HClO4 as an eluent. The response characteristics depend on the adjusted baseline pH. The baseline pH adjustment was successfully done with an ammonia permeation device without solution mixing, which may cause analyte dilution, dispersion, and mixing noise. After pH adjustment, the pH response was universal to the equivalent of introduced analyte acids because the pH response was obtained by the titration of the permeant ammonia by the analytes. The average of limit of detections (S/N = 3) was 0.06 mM for seven target organic acids. Furthermore, the pH response follows the theoretical pH calculation with the concentrations of the eluent, pH modifier, and analyte. Thus, the analyte concentration in the sample can be theoretically calculated from the pH response without calibration by the analyte standard. Predicted concentrations of injected standards were within 5% of the actual standard concentration. Additionally, analysis of real samples was performed and compared with the conventional post-column reaction with a bromothymol blue (BTB) method. The results obtained with the present system (absolute quantification with theoretical calculation) and conventional BTB method agreed within 10% of errors.

Automated Programmable Generation of Broad pH Range Volatile Ionic Eluents for Liquid Chromatography.

Many of the universal detectors in liquid chromatography, including mass spectrometry, must completely volatilize the chromatographic eluent first before further processing and detection of the analytes. A basic requirement is that the eluent does not contain a nonvolatile dissolved component. However, separation of biomolecules must be conducted in mostly aqueous media of compatible pH and ionic strength if their biological activity must survive the separation process. Combinations of ammonia with acetic and formic acids are commonly used as eluent for this purpose but generally maximum concentrations that can be tolerated are relatively low. Further, buffering is good only over a limited pH range. We describe a system where the eluent is generated in an automated pressure-programmed manner from high-purity gaseous NH3 and CO2 through gas-permeable membrane devices. This can be aided by the prior presence of formic/acetic acids in the mobile phase to extend the attainable low pH limit. We outline the fundamental pH, ionic strength, and buffer intensity considerations and demonstrate the application of such eluents in the separation of amino acids, proteins, and monoclonal antibodies. We also demonstrate the use of dissolved CO2 as an ion-pairing agent in the separation of chiral amines.

Optimum Cell Pathlength or Volume for Absorbance Detection in Liquid Chromatography: Transforming Longer Cell Results to Virtual Shorter Cells.

The cell volume permissible for a specified degree of loss of efficiency can be computed from response volume considerations. For an open tubular column, the permissible illuminated length can be computed for the unretained peak. Similar estimations can be made for the maximum permissible cell volume with a packed column for a given column efficiency and flow rate. The packed column case does require an assumption on the degree to which the cell behaves as a mixer. An altogether different question is if the data from a long cell, with its considerable advantage in S/N, can be used and the associated dispersion mathematically removed. Experimental data for a variable path length (0-60 mm) HPLC detection cell indicate that an exponential model fits the observed dispersion. Once fit parameters are determined, the same can be applied to a peak, not part of the original training set, obtained in a longer path cell and the effects of dispersion mathematically reversed without major loss of S/N. Results with shorter path cell dispersion characteristics are then obtained with much higher S/N. A comparison is made with Fourier transform-inverse Fourier transform based deconvolution that can be used to achieve the same ends.

Nanovolume Gas-Free Hydroxide Eluent Generator for Open Tubular Ion Chromatography.

A gas-free KOH eluent generator (EG) with 210 nL of internal volume is described. It utilizes a two-membrane configuration where there is a single CEM layer on one side and a single BPM layer on the other side for use in open tubular ion chromatography systems with typical back pressures < 50 psi. At a flow rate of ∼190 nL/min, the 10-90% gradient rise time is 3.5 min. The device shows good linearity between applied current and concentration of KOH generated, which is stable over extended periods. The overall system reproducibility (that includes contributions from any changes in flow rate), as judged by the relative standard deviation (RSD) of the retention times of individual separated ions in repeat measurements (n = 6), ranged from <0.5% for isocratic to <1.2% for gradient elution schemes. Perceptible current flow and KOH production in the BPM-based EG begins at subelectrolytic applied voltages, prompting us to look more closely at exact field strength necessary for field-enhanced dissociation of water. An increase in the specific conductance of pure water is noticeable by a field strength of 105 V/m.

Attenuation Coefficients of Tubular Conduits for Liquid Phase Absorbance Measurement: Shot Noise Limited Optimum Path Length.

We trace the history of liquid core waveguides (LCWs, also called liquid core optical fibers) and the role Teflon AF (TAF) has played in their development. We show that, in any shot noise limited situation, the optimum signal-to-noise ratio (S/N) occurs at a path length of 1/αa{ln[1 + 2(αa/αb)]}, approximately 2/αb under most conditions, αa and αb being the light attenuation coefficient due to the analyte and the background, respectively. The analysis shows that LCW length should be selected depending on the applicable αb value. An overly long LCW may exhibit a lower signal-to-noise ratio. Water-filled TAF-clad fused-silica (FS) tubes show the lowest attenuation across the wavelength range. Nevertheless, except at λ ≥ 600 nm, the observed αb values far exceed those reported for pure water: it appears that both impurities in the water and waveguide losses are involved. In examining the attenuation in various water-filled tubes, we find that the transmission of air-surrounded FS tubes is second only to TAF-clad FS tubes and is better than that of TAF tubes or externally mirrored FS tubes. Surprisingly, except for a window centered at ∼250 nm, light transmission in a water-filled poly(tetrafluoroethylene) (PTFE) tube is worse than in poly(ether ether ketone) (PEEK) tubing. Light transmission in PTFE tubes improves with increasing wall thickness.

Carbonic Acid Eluent Ion Chromatography.

Alkali metals, amines and alkanolamines are separated on a poly(butadiene)-maleic acid on silica stationary phase using a carbonic acid (H2CO3*) eluent with and without a mineral acid. The H2CO3* eluent is prepared in situ by high pressure permeative introduction of gaseous CO2 through thin membranes supported upon a porous steel disk. The permeation flux and thus the eluent concentration is controlled by varying the applied CO2 pressure. This novel frit-supported membrane device tolerates much higher liquid and gas pressures than Teflon AF capillaries, permitting [H2CO3*] exceeding 0.53 M and attaining a pH of 3.3. Silicone was presently preferred over Teflon AF, both as planar membranes, as mechanical properties of the latter change as large amounts of CO2 dissolve in it. After separation, the CO2 can be efficiently removed via another gas permeable membrane device permitting detection of the eluting bicarbonate salt conductometrically in a background of nearly pure water. Most analytes are more sensitively detected after anion conversion to hydroxide using a standard suppressor, permitting 3-17 pmol LODs on 2 mm bore columns. The data, particularly comparisons with an HNO3 eluent, with or without H2CO3*, indicate that proton exchange alone does not account for the retention behavior; some reactive addition of HCO3- is involved.

Flow-Cell-Induced Dispersion in Flow-through Absorbance Detection Systems: True Column Effluent Peak Variance.

Following a brief overview of the emergence of absorbance detection in liquid chromatography, we focus on the dispersion caused by the absorbance measurement cell and its inlet. A simple experiment is proposed wherein chromatographic flow and conditions are held constant but a variable portion of the column effluent is directed into the detector. The temporal peak variance (σt,obs2), which increases as the flow rate (F) through the detector decreases, is found to be well-described as a quadratic function of 1/F. This allows the extrapolation of the results to zero residence time in the detector and thence the determination of the true variance of the peak prior to the detector (this includes contribution of all preceding components). This general approach should be equally applicable to detection systems other than absorbance. We also experiment where the inlet/outlet system remains the same but the path length is varied. This allows one to assess the individual contributions of the cell itself and the inlet/outlet system.to the total observed peak. The dispersion in the cell itself has often been modeled as a flow-independent parameter, dependent only on the cell volume. Except for very long path/large volume cells, this paradigm is simply incorrect.