Type of Document	Dissertation
Author	Simmons, Carolyn Rebecca
Author's Email Address	csimmons@shc.edu
URN	etd-08262005-190326
Title	Deconvolution of Mobile Phase Contributions to Band Broadening in Reversed-Phase Liquid Chromatography
Degree	Doctor of Philosophy
Department	Chemistry and Biochemistry, Department of
Advisory Committee	Advisor Name Title Alan G. Marshall Committee Member Albert E. Stiegman Committee Member Bruce Locke Committee Member John G. Dorsey Committee Member
Keywords	liquid chromatography band broadening efficiency equations void volume HPLC reversed-phase
Date of Defense	2005-08-04
Availability	unrestricted
Abstract High Performance Liquid Chromatography (HPLC) is the most widely used technique for the separation and identification of components in complex liquid samples. From groundwater to biological fluids, HPLC dependably provides efficient separations. Reversed-phase liquid chromatography (RPLC) has for the last several decades been the most popular mode of LC. The nonpolar stationary phase and adjustable polarity mobile phase of RPLC provide for simple optimization schemes and the ability to separate both ionic and neutral compounds. In addition, the use of small stationary phase support particles and high pressure capabilities allows RPLC to provide the fast, efficient separations required by industrial users. More than a century has passed since the advent of liquid chromatography and since that time great advances have been made in the field. Regardless, there still remain unanswered questions to many fundamental problems associated with HPLC. One particular area of interest is band broadening. A sample band is injected onto the head of the column as a narrow pulse. Due to band broadening, the solute bands that elute are no longer narrow pulses, but rather Gaussian shaped peaks. The variance of the peaks essentially determines the peak capacity of the separation. HPLC users demand high throughput separations in a minimal amount of time, which becomes quite difficult when band broadening is excessive. What are the kinetic processes that cause the solute bands to disperse, and what can we do to prevent this detrimental occurrence? The work described in the following chapters attempts to answer these questions. Band broadening occurs due to molecular diffusion of the solutes, multiple paths available to the solutes through the packed bed, resistance to mass transfer of the solutes when partitioning into the stationary phase, and resistance to mass transfer through the mobile phase. It has been a topic of much debate as to whether the variance due to each process is independent of the others or if the effects should be coupled. The broadening taking place in the mobile phase is the topic of the following chapters. The columns in RPLC are typically packed with porous stationary phase support particles. Pressure-driven flow does not provide for convective velocity within the particle pores. As a result, the mobile phase becomes trapped and stagnant. There are conflicting views in the literature as to the magnitude of the broadening taking place in the stagnant mobile phase. It has been suggested that this source of broadening is insignificant and that the only way to reduce dispersion in the mobile phase is to redesign the interstitial region in a more ordered fashion. However, it is very unlikely that the two billion dollar a year HPLC industry will drastically change the existing internal architecture of the column without substantial experimental evidence. It was the purpose of this work to provide a quantitative assessment of how much of the broadening taking place in the mobile phase can be attributed to dispersion in the stagnant pore volume versus the interstitial regions. As a first attempt, the void volume measured by nonretained solutes was monitored over the flow rate range 0.1 to 3.0 mL/min. If the stagnant mobile phase caused a significant resistance to mass transfer, it was predicted that the measured volume would increase as the flow rate decreased. In other words, the solutes would have more time at the lower flow rates to diffuse through the stagnant mobile phase and thus measure a higher volume. The results of the study did not support this theory. The measured void volume remained constant over the flow rate range. Our next attempt to assess the stagnant mobile phase contribution to dispersion was to chromatographically isolate it from the broadening in the interstitial space. Low concentrations of ionic solutes experience electrostatic exclusion from the stationary phase pores in unbuffered mobile phases. NaNO3 was used as the probe solute for this study, because it is unretained on the reversed-phase surface. A high concentration of NaNO3 experienced dispersion as a result of both interstitial and stagnant mobile phase broadening. However, a low concentration did not sample the pore regions and was only affected by the interstitial dispersion. Using these solutes it was determined that the broadening taking place in the stagnant mobile phase was a significant source of dispersion, contributing 30-40% of the total mobile phase dispersion. These experiments were repeated on a nonporous reversed-phase column. Unfortunately, due to complications with the data analysis software, the instrument, and the column capacity this portion of the study did not further our understanding of mobile phase broadening. The final chapter of this work describes a study to determine which of the four most popular efficiency equations provides the best model for chromatographic plate height data. For more than fifty years, separations scientists have attempted to derive an equation to mathematically model band broadening contributions. Three of the most popular equations derived from theory are the equations of van Deemter, Giddings, and Horvath and Lin. The empirical Knox equation was also included in this study. Plate height data for benzene, toluene, ethylbenzene, propylbenzene, and butylbenzene were collected for four mobile phase compositions. The efficiency equations were fit to the data, and all four equations gave comparable fits. The Horvath and Lin equation consistently performed as well as or better than the others, while the Knox equation typically provided the worst fit to the data. An unexpected dependence of fit quality on retention factor k’ was observed. The plate height data for solutes with k’ values of approximately 3 provided the best fits to the four equations.
Files	Filename       Size       Approximate Download Time (Hours:Minutes:Seconds)     28.8 Modem   56K Modem   ISDN (64 Kb)   ISDN (128 Kb)   Higher-speed Access    Final_Dissertation.pdf 1.62 Mb 00:07:30 00:03:51 00:03:22 00:01:41 00:00:08