Asymmetric flow field‐flow fractionation of liposomes: 2. Concentration detection and adsorptive loss phenomena

The applicability of different concentration detection methods for online quantification of liposomes upon asymmetric flow field‐flow fractionation was investigated. Filter‐extruded egg phosphatidylcholine liposomes of different size were used. Online quantification using a differential refractive index (dRI) detector was found feasible for relatively high sample loads in the magnitude of 100 μg lipid (under the chosen fractionation conditions). UV–Vis detection of the turbidity of liposomes was ruled out as online detection method because turbidity increases with particle size and the signal is not only concentration but also particle‐size dependent. Staining of liposomes by Rhodamine phosphatidylethanolamine or Sudan Red and subsequent online UV–Vis detection at the absorption maximum of the dye enabled quantification with much higher sensitivity than dRI detection. Furthermore analyte loss and carry‐over phenomena upon repeated injection of varying liposome sample loads were studied using regenerated cellulose (RC) membranes as accumulation wall. It could be shown that RC membranes are prone to adsorption in case of very small sample loads (0.5 μg). This effect may be overcome by pre‐saturation of the membrane with sample loads of at least 2 μg. For higher sample loads adsorptive losses play a minor role. Recovery from pre‐saturated membranes reached approximately 100% and carry‐over was found negligible.     

Impact of non‐ideal analyte behavior on the separation of protein aggregates by asymmetric flow field‐flow fractionation

Asymmetric flow field‐flow fractionation is a valuable tool for the characterization of protein aggregates in biotechnology owing to its broad size range and unique separation principle. However, in practice asymmetric flow field‐flow fractionation is non‐trivial to use due to the major deviations from theory and the influence on separation by various factors that are not fully understood. Here, we report methods to assess the non‐ideal effects that influence asymmetric flow field‐flow fractionation separation and for the first time identify experimentally the main factors that impact it. Furthermore, we propose new approaches to minimize such non‐ideal behavior, showing that by adjusting the mobile phase composition (pH and ionic strength) the resolution of asymmetric flow field‐flow fractionation separation can be drastically improved. Additionally, we propose a best practice method for new proteins.     

Numerical analysis of a dielectrophoresis field‐flow fractionation device for the separation of multiple cell types

In this study, a dielectrophoresis field‐flow fractionation device was analyzed using a numerical simulation method and the behaviors of a set of different cells were investigated. By reducing the alternating current frequency of the electrodes from the value used in the original setup configuration and increasing the number of exit channels, total discrimination in cell trajectories and subsequent separation of four cell types were achieved. Cells were differentiated based on their size and dielectric response that are represented in their real part of Clausius–Mossotti factor at different frequencies. A number of novel designs were also proposed based on the original setup configuration. It was seen that by reducing the length of the main channel and the number of electrodes at low frequencies and not changing the inlet flow velocities, cell separation was still achieved successfully, although with a slightly larger electrode voltage. The shorter main channel decreased the residence time for the cells on the chip and also reduced the overall size of the device—these were improvements over the original design. The obtained results can be used to analyze other cell types by knowing their size and dielectric properties to design geometries that can ensure separation.     

Analysis of liposomes using asymmetrical flow field-flow fractionation: Separation conditions and drug/lipid recovery

Liposomes composed of dipalmitoylphosphatidylcholine and dipalmitoylphosphatidylglycerol were analyzed by asymmetrical flow field-flow fractionation coupled with multi-angle laser light scattering. In addition to evaluation of fractionation conditions (flow conditions, sample mass, carrier liquid), radiolabeled drug-loaded liposomes were used to determine the liposome recovery and a potential loss of incorporated drug during fractionation. Neither sample concentration nor the cross-flow gradient distinctly affected the size results but at very low sample concentration (injected mass 5 μg) the fraction of larger vesicles was underestimated. Imbalance in the osmolality between the inner and outer aqueous phase resulted in liposome swelling after dilution in hypoosmotic carrier liquids. In contrast, liposome shrinking under hyperosmotic conditions was barely visible. The liposomes themselves eluted completely (lipid recoveries were close to 100%) but there was a loss of incorporated drugs during separation with a strong dependence on the octanol-water partition coefficient of the drug. Whereas corticosterone (partition coefficient ～2) was washed out more or less completely (recovery about 2%), loss of temoporfin (partition coefficient ～9) was only minor (recovery about 80%). All fractionations were well repeatable under the experimental conditions applied in the present study.     

Model‐based analysis of a dielectrophoretic microfluidic device for field‐flow fractionation

We present the development of a dynamic model for predicting the trajectory of microparticles in microfluidic devices, employing dielectrophoresis, for Hyperlayer field‐flow fractionation. The electrode configuration is such that multiple finite‐sized electrodes are located on the top and bottom walls of the microchannel; the electrodes on the walls are aligned with each other. The electric potential inside the microchannel is described using the Laplace equation while the microparticles' trajectory is described using equations based on Newton's second law. All equations are solved using finite difference method. The equations of motion account for forces including inertia, buoyancy, drag, gravity, virtual mass, and dielectrophoresis. The model is used for parametric study; the geometric parameters analyzed include microparticle radius, microchannel depth, and electrode/spacing lengths while volumetric flow rate and actuation voltage are the two operating parameters considered in the study. The trajectory of microparticles is composed of transient and steady state phases; the trajectory is influenced by all parameters. Microparticle radius and volumetric flow rate, above the threshold, do not influence the steady state levitation height; microparticle levitation is not possible below the threshold of the volumetric flow rate. Microchannel depth, electrode/spacing lengths, and actuation voltage influence the steady‐state levitation height.     

Molecular characterization of branched polysaccharides from Tremella fuciformis by asymmetrical flow field‐flow fractionation and size exclusion chromatography

To accurately characterize branched polysaccharides with high molecular weights from medicinal and edible mushrooms and identify the limitations of size exclusion chromatography, molecular characteristics of polysaccharides from Tremella fuciformis were determined and compared by asymmetrical flow field‐flow fractionation coupled with multiangle laser light scattering and refractive index detection, and size exclusion chromatography coupled with multiangle laser light scattering and refractive index detection, respectively. Results showed that molecular weights of three batches of T. fuciformis polysaccharides were determined as 2.167 × 10⁶ (TF1), 2.334 × 10⁶ (TF2), and 2.435 × 10⁶ Da (TF3) by size exclusion chromatography, and 3.432 × 10⁶ (TF1), 3.739 × 10⁶ (TF2), and 3.742 × 10⁶ Da (TF3) by asymmetrical flow field‐flow fractionation, as well as 3.469 × 10⁶ Da (TF1) by off‐line multiangle laser light scattering, respectively. Results suggested that size exclusion chromatography was unable to accurately characterize T. fuciformis polysaccharides, which may be due to its limitations such as shear degradation and abnormal coelution. Compared to size exclusion chromatography, asymmetrical flow field‐flow fractionation could be a better technique for the molecular characterization of branched polysaccharides with high molecular weights from medicinal and edible mushrooms, as well as from other natural resources.     

A one-dimensional transient model of electrical field flow fractionation

A well-developed classical theory is available for constant-voltage electrical field flow fractionation (EFFF). Recent experimental research, however, has demonstrated that pulsed fields may enhance retention in some cases. A generalized mathematical approach is presented for the prediction of retention ratios for any field type, pulsed or constant. The methodology is applied and demonstrated for a square wave protocol. Complex concentration profiles arise wherein particles are focused either towards the walls or into the channel center. The computational results indicate that pulsation can either increase retention time or decrease retention time by manipulating the effective electric field and suggest that separation resolution may also be improved.     

Dielectrophoretic field‐flow fractionation of electroporated cells

We describe the development and testing of a setup that allows for DEP field‐flow fractionation (DEP‐FFF) of irreversibly electroporated, reversibly electroporated, and nonelectroporated cells based on their different polarizabilities. We first optimized the channel and electrode dimensions, flow rate, and electric field parameters for efficient DEP‐FFF separation of moderately heat‐treated CHO cells (50°C for 15 min) from untreated ones, with the former used as a uniform and stable model of electroporated cells. We then used CHO cells exposed to electric field pulses with amplitudes from 1200 to 2800 V/cm, yielding six groups containing various fractions of nonporated, reversibly porated, and irreversibly porated cells, testing their fractionation in the chamber. DEP‐FFF at 65 kHz resulted in distinctive flow rates for nonporated and each of the porated cell groups. At lower frequencies, the efficiency of fractionation deteriorated, while at higher frequencies the separation of individual elution profiles was further improved, but at the cost of cell flow rate slowdown in all the cell groups, implying undesired transition from negative into positive DEP, where the cells are pulled toward the electrodes. Our results demonstrate that fractionation of irreversibly electroporated, reversibly electroporated, and nonelectroporated cells is feasible at a properly selected frequency.     

Theoretical analysis of the local orientation effect and the lift‐hyperlayer mode of rodlike particles in field‐flow fractionation

We investigated theoretically the effects of the cross‐stream migration and the local average orientation of rodlike particles on the shape‐based separation using field‐flow fractionation. The separation behavior was analyzed by comparing the retention ratios of spheres and rods. The retention ratio of a rod was evaluated through the derivation of its cross‐sectional concentration profile by considering the rod migration and the local average orientation. Our study in various flow conditions showed that the rod migration, caused by the hydrodynamic interaction with a wall, can affect the separation behavior as a lift‐hyperlayer mode. We also demonstrated that the local average orientation, which is a function of a local shear rate and a rotational diffusivity, results in the transverse diffusivity that is different from its perpendicular diffusivity. These results suggest that the experimental separation behaviors of rods in field‐flow fractionation may not be fully explained by the current theory based on the normal mode and the steric mode. We also characterized each condition where one of the normal mode, the steric mode of spheres, and the lift‐hyperlayer mode of rods is dominant.     

Optimization of cyclical electrical field flow fractionation

Cyclical electrical field flow fractionation (CyElFFF) is a variation on electrical field flow fractionation (ElFFF) where cyclical electrical fields are used instead of steady DC fields to increase the effective field experienced by particles in the flow channel. Even though the effective field increases more than 20-fold compared to normal ElFFF, the retention and resolution in CyElFFF has not been shown to be better than in ElFFF. In this paper we report how one can optimize operational parameters in CyElFFF to obtain good retention and resolution in CyElFFF. The effects of offset voltage, frequency, flowrate, concentration of particles and sample size on retention, resolution and retained peak/void peak ratio have been observed. The results obtained from these experiments were analyzed and suggestions have been made to improve both retention and resolution. A 4-fold improvement in retention without a significant increase in band broadening is reported.     

Asymmetric flow field flow fractionation of aqueous C60 nanoparticles with size determination by dynamic light scattering and quantification by liquid chromatography atmospheric pressure photo-ionization mass spectrometry

A size separation method was developed for aqueous C₆₀ fullerene aggregates (aqu/C₆₀) using asymmetric flow field flow fractionation (AF4) coupled to a dynamic light scattering detector in flow through mode. Surfactants, which are commonly used in AF4, were avoided as they may alter suspension characteristics. Aqu/C₆₀ aggregates generated by sonication in deionized water ranged in size from 80 to 260 nm in hydrodynamic diameter (D_h) as determined by DLS in flow through mode, which was corroborated by analysis of fractions by DLS in batch mode and by TEM. The mass of C₆₀ in each fraction was determined by LC–APPI–MS. Only 5.2 ± 6.7% of the total aqu/C₆₀ mass had D_h less than 80 nm, while 58 ± 32% of the total aqu/C₆₀ mass had D_h between 80 and 150 nm and 14 ± 9.2% of the total aqu/C₆₀ were between 150 and 260 nm in D_h. With the optimal fractionation parameters, 77 ± 5.8% of the aqu/C₆₀ mass eluted from the AF4 channel, indicating deposition on the AF4 membrane had occurred during fractionation; use of alternative membranes did not reduce deposition. Channel flow splitting increased detector response although channel split ratios greater than 80% of the channel flow led to decreased detector response. This is the first report on the use of AF4 for fractionating a colloidal suspension of aqu/C₆₀.     

Assessment of the diffusional behavior of polystyrene sulfonates in the dilute regime by hollow‐fiber flow field flow fractionation

We applied flow field flow fractionation in a ceramic tubular membrane to study the diffusional behavior of sodium polystyrene sulfonate (NaPSS) as a function of salt and polyelectrolyte concentrations in the dilute concentration regime. NaPSS standards were fractionated in an aqueous carrier liquid with salt concentrations of 1–100 mmol L^?1 and injected amounts of 0.2–20 μg. A sharp increase in the apparent diffusion coefficients, due to peak overloading, occurred at a critical polyelectrolyte concentration at the accumulation wall. A model has been developed that compares repulsive electrostatic interactions, calculated from the Derjaguin–Landau–Verweij–Overbeek theory, to the thermal energy. It was found that the observed critical concentration approximated the concentration of a closely packed lattice of hard spheres within an order of magnitude. The critical concentration appeared to be dependent on the salt concentration, but the influence of the molecular mass was less clear. Expressions for the migration velocity in overloaded channels were derived and confirmed experimentally for the two lowest molecular masses. Deviations occurred for high molecular mass NaPSS, possibly because of mutual attraction. These experiments confirmed the observations made by small‐angle neutron scattering and dynamic light scattering that the ordering of polyelectrolytes already occurred in the dilute regime. © 2001 John Wiley & Sons, Inc. J Polym Sci Part B: Polym Phys 39: 1756–1765, 2001     

Improved theory of cyclical electrical field flow fractionation

Previously reported theories for cyclical electrical field flow fractionation (CyElFFF) are severely limited in that they do not account for diffusion, steric, or electric double layer effects. Experiments have shown that these theories overpredict the retention of particles in CyElFFF. In this work, we present a model for prediction of steric, diffusion, and electrical effects. The electrical double layer effects are treated using a lumped electrical circuit model that accounts for the field shielding by the electrical double layer formed at the electrode-carrier interface. The electrical effects are shown to dominate retention times and outweigh the contributions of diffusion and particle size. Detailed results from the simulations are presented in this work, and a comparison between the theoretical and experimental results obtained from the retentions of polystyrene particle standards is presented in this paper. The models are shown to correctly predict the retention of the polystyrene standards in CyElFFF with a reasonable error, while existing models are shown to have significant failings.     

Miniaturized asymmetrical flow field-flow fractionation: application to biological vesicles

Asymmetrical flow field-flow fractionation (AFlFFF) has been carried out in a miniaturized channel by reducing the channel dimensions. Performance of the miniaturized AFlFFF (mAFlFFF) channel was evaluated with standard proteins and polystyrene latex spheres from nanometer to micrometer size. By reducing the channel dimension, proteins or particulate materials can be separated within a few minutes without a significant loss in resolution. The mAFlFFF channel was applied for the separation of exosomes harvested from immortalized human mesenchymal stem cell line. It shows a potential to fractionate exosome vesicles according to sizes which can be useful for proteomic studies in relation to immunotherapeutic applications.     

CE and asymmetrical flow‐field flow fractionation studies of polymer interactions with surfaces and solutes reveal conformation changes of polymers

Amphiphilic diblock copolymers consisting of a hydrophobic core containing a polymerized ionic liquid and an outer shell composed of poly(N‐isoprolylacrylamide) were investigated by capillary electrophoresis and asymmetrical flow‐field flow fractionation. The polymerized ionic liquid comprised poly(2‐(1‐butylimidazolium‐3‐yl)ethyl methacrylate tetrafluoroborate) with a constant block length (n = 24), while the length of the poly(N‐isoprolylacrylamide) block varied (n = 14; 26; 59; 88). Possible adsorption of the block copolymer on the fused silica capillary, due to alterations in the polymeric conformation upon a change in the temperature (25 and 45 °C), was initially studied. For comparison, the effect of temperature on the copolymer conformation/hydrodynamic size was determined with the aid of asymmetrical flow‐field flow fractionation and light scattering. To get more information about the hydrophilic/hydrophobic properties of the synthesized block copolymers, they were used as a pseudostationary phase in electrokinetic chromatography for the separation of some model compounds, that is, benzoates and steroids. Of particular interest was to find out whether a change in the length or concentration of the poly(N‐isoprolylacrylamide) block would affect the separation of the model compounds. Overall, our results show that capillary electrophoresis and asymmetrical flow‐field flow fractionation are suitable methods for characterizing conformational changes of such diblock copolymers.     

Hollow-fiber flow/hyperlayer field-flow fractionation for the size characterization of airborne particle fractions obtained by SPLITT fractionation

Hollow-fiber flow field-flow fractionation (HF FlFFF) was applied for the separation and size characterization of airborne particles which were collected in a municipal area and prefractionated into four different-diameter intervals >5.0, 2.5-5.0, 1.5-2.5, <1.5 microm) by continuous split-flow thin (SPLIIT) fractionation. Experiments demonstrated the possibility of utilizing a hollow-fiber module for the high-performance separation of supramicron-sized airborne particles at steric/hyperlayer operating mode of HF FlFFF. Eluting particles during HF FlFFF separation were collected at short time intervals (approximately 10 s) for the microscopic examination. It showed that particle size and size distributions of all SPLITT fractions of airborne particles can be readily obtained using a calibration and that HF FlFFF can be utilized for the size confirmation of the sorted particle fraction during SPLITT fractionation.     

Performance of hollow-fiber flow field-flow fractionation in protein separation

Since hollow-fiber flow field-flow fractionation (HF FIFFF) utilizes a cylindrical channel made of a hollow-fiber membrane, which is inexpensive and simple in channel assembly and thus disposable, interests are increasing as a potential separation device in cells, proteins, and macromolecules. In this study, performance of HF FIFFF of proteins is described by examining the influence of flow rate conditions and length of fiber (polyacrylonitrile or PAN in this work) on sample recovery as well as experimental plate heights. The interfiber reproducibility in terms of separation time and recovery was also studied. Experiments showed that sample recovery was consistent regardless of the length of fiber when the effective field strength (equivalent to the mean flow velocity at the fiber wall) and the channel void time were adjusted to be equivalent for channels of various fiber lengths. This supported that the majority of sample loss in HF FIFFF separation of apoferritin and their aggregates may occur before the migration process. It is finally demonstrated that HF FIFFF can be applied for characterizing the reduction in Stokes' size of low density lipoproteins from blood plasma samples obtained from patients having coronary artery disease and from healthy donors.     

Application of an asymmetric flow field flow fractionation multi-detector approach for metallic engineered nanoparticle characterization--prospects and limitations demonstrated on Au nanoparticles

In this work we discuss about the method development, applicability and limitations of an asymmetric flow field flow fractionation (A4F) system in combination with a multi-detector setup consisting of UV/vis, light scattering, and inductively coupled plasma mass spectrometry (ICPMS). The overall aim was to obtain a size dependent-, element specific-, and quantitative method appropriate for the characterization of metallic engineered nanoparticle (ENP) dispersions. Thus, systematic investigations of crucial method parameters were performed by employing well characterized Au nanoparticles (Au-NPs) as a defined model system.For good separation performance, the A4F flow-, membrane-, and carrier conditions were optimized. To obtain reliable size information, the use of laser light scattering based detectors was evaluated, where an online dynamic light scattering (DLS) detector showed good results for the investigated Au-NP up to a size of 80 nm in hydrodynamic diameter. To adapt large sensitivity differences of the various detectors, as well as to guarantee long term stability and minimum contamination of the mass spectrometer a split-flow concept for coupling ICPMS was evaluated. To test for reliable quantification, the ICPMS signal response of ionic Au standards was compared to that of Au-NP. Using proper stabilization with surfactants, no difference for concentrations of 1–50 μg Au L⁻¹ in the size range from 5 to 80 nm for citrate stabilized dispersions was observed. However, studies using different A4F channel membranes showed unspecific particle–membrane interaction resulting in retention time shifts and unspecific loss of nanoparticles, depending on the Au-NP system as well as membrane batch and type. Thus, reliable quantification and discrimination of ionic and particular species was performed using ICPMS in combination with ultracentrifugation instead of direct quantification with the A4F multi-detector setup.Figures of merit were obtained, by comparing the results from the multi detector approach outlined above, with results from batch-DLS and transmission electron microscopy (TEM). Furthermore, validation performed with certified NIST Au-NP showed excellent agreement. The developed methods show potential for characterization of other commonly used and important metallic engineered nanoparticles.     

Microparticles manipulation and enhancement of their separation in pinched flow fractionation by insulator‐based dielectrophoresis

The separation and manipulation of microparticles in lab on a chip devices have importance in point of care diagnostic tools and analytical applications. The separation and sorting of particles from biological and clinical samples can be performed using active and passive techniques. In passive techniques, no external force is applied while in active techniques by applying external force (e.g. electrical), higher separation efficiency is obtained. In this article, passive (pinched flow fractionation) and active (insulator‐based dielectrophoresis) methods were combined to increase the separation efficiency at lower voltages. First by simulation, appropriate values of geometry and applied voltages for better focusing, separation, and lower Joule heating were obtained. Separation of 1.5 and 6 μm polystyrene microparticles was experimentally obtained at optimized geometry and low total applied voltage (25 V). Also, the trajectory of 1.5 μm microparticles was controlled by adjusting the total applied voltage.     

Poisson's effects in electrical field flow fractionation

Recent and earlier models of electrical field flow fractionation (ELFFF) have assumed that the electric field within the fluid domain is governed by Laplace's equation. This assumption results in a linear potential and a spatially constant field across the channel and is generally true for very dilute systems and relatively high effective potentials. Experimental studies show, however, that the effective potential within the channel may be less than 1% of the applied potential; this is apparently due to double layer formation and charge buildup at the poles. In such cases, local analyte concentrations can, nonetheless, be orders of magnitude higher than the bulk mean and the local potential small, both of which can lead to a nonlinear spatial distribution of the field strength. In such cases Poisson's equation must be used rather than Laplace's equation. Steady-state ELFFF simulations were performed using a Poisson's equation-based model. The domain in which Laplace's equation is valid was identified and the effects of concentration and effective field strength on device performance were explored.