HPLC: Separation And Quantification Of Components In Diet Soft Drinks


Separation And Quantification

Of Components In Diet Soft Drinks


1.      Principles of Instrumental Analysis, 5th Edition, Douglas Skoog, F. James Holler, Timothy Nieman, Saunders College Publishing, Philadelphia, 1998.

2.      “The Analysis of Artificial Sweeteners and Additives in Beverages by HPLC,” Journal of Chemical Education, vol. 68(8), August 1991, p A195-A200.


The purpose of this experiment is to quantify the caffeine content of a diet cola sample using high performance liquid chromatography (HPLC).  In order to quantify the caffeine, it must be isolated from the other components in the mixture.  In this experiment, you will determine a set of HPLC conditions suitable for separating caffeine, benzoic acid, and aspartame and then quantify the caffeine content of your cola sample using a standard calibration curve.  At the end of this experiment you should understand the mechanisms by which components in a mixture are separated, identified, and quantified using HPLC and understand how to vary experimental parameters to optimize a separation.


The fundamentals of chromatographic separations and a detailed discussion of the application of HPLC are covered in reference 1 (chapters 26 and 28).  The following discussion summarizes important concepts from these chapters, but the student is encouraged to read the full text.

Introduction to HPLC and Instrument Components.

High performance liquid chromatography (HPLC) is an important analytical tool for separating and quantifying components in complex liquid mixtures.  By choosing the appropriate equipment (i.e. column and detector), this method is applicable to samples with components ranging from small organic and inorganic molecules and ions to polymers and proteins with high molecular weights.  The various types of HPLC and their characteristics are summarized in the table below.  In this experiment, we will use reversed-phase partition chromatography.

Table 1.  Various Types and Applications of HPLC 







non-polar to somewhat polar

100 – 104

silica or alumina

non-polar to polar

Partition (reversed-phase)

non-polar to somewhat polar

100 – 104

non-polar liquid adsorbed or chemically bonded to the packing material

relatively polar

Partition (normal-phase)

somewhat polar to highly polar

100 – 104

highly polar liquid adsorbed or chemically bonded to the packing material

relatively non-polar

Ion Exchange

highly polar to ionic

100 – 104

ion-exchange resins made of insoluble, high-molecular weight solids functionalized typically with sulfonic acid (cationic exchange) or amine (anionic exchange) groups

aqueous buffers with added organic solvents to moderate solvent strength


non-polar to ionic

103 – 106

small, porous, silica or polymeric particles

polar to non-polar

Figure 1.  shows the components of our Hewlett-Packard Model 1100 HPLC.  The system consists of:

·         reservoirs to hold the solvents used to make up the mobile phase

·         a solvent degasser to prevent bubbles in the mobile phase

·         a programmable quaternary pump that mixes the solvents in the prescribed ratios and pumps them through the column and past the detector

·         a column compartment that houses and thermostats the HPLC column

(in our case a ZORBAX, reversed-phase C18 column; dimensions 4.6mm x 15 cm)

·         an autosampler that draws prescribed volumes from sample vials and injects them onto the column

·         a diode array detector that monitors the entire UV-vis spectrum of the column effluent at regular intervals

Control of the above components and data acquisition and analysis are performed on a personal computer. 

Figure 1.  HPLC 1100 Instrument Components

Optimization of Resolution and Column Performance.

The goal of any HPLC experiment is to achieve the desired separation in the shortest possible time.  Time is critical because “time is money” and because as we’ll see, the more time the sample spends on the column, the more the bands containing the components spread, resulting in reduced resolution.  Optimization of the experiment thus usually involves manipulation of column and mobile phase parameters to alter the relative migration rates of the components in the mixture and to reduce zone broadening.  These can generally be optimized fairly independently.

Migration Rates

The length of time it takes for a given component/solute to travel through the column and be detected is determined by the flow rate of the mobile phase, m, and the partitioning of the solute between the mobile and stationary phases.  Since the solute molecules can only travel when they are dissolved in the mobile phase, the greater their concentration in the mobile phase, the faster they will elute.  The partition coefficient, K, is defined in equation 1

where CS is the concentration of the solute dissolved in or adsorbed to the stationary phase, and CM is the concentration of the solute in the mobile phase. 

The quantities CS and CM, however, are rarely determined in chromatographic experiments.  Instead, a quantity called the retention factor, k’, is determined.  The retention factor for a component A is defined as

where tR is the retention time of component A and tM is the retention time of an unretained species (also called the dead time).  The average rate of linear migration of component A is related to both the flow rate of the mobile phase and the retention factor. 

The retention factors should normally lie in a range of 2-5, but for complex mixtures a larger range may be required to separate all the components.  The value of the retention factor for a given component depends on the chemical identity of the component and the following experimental variables:

·         mobile phase flow rate

·         mobile phase composition

·         column temperature

·         column composition

Zone Broadening

The extent to which the component bands spread as they travel down the column affects the efficiency of the separation.  The theoretical plate height, H, is defined in equation 4 and is based on a Gaussian analysis of the peak width, s, as it exits the column at point L.


and W is the width of the peak at the base.  The data analysis program on our HPLC actually reports the width at half maximum, W1/2, for each peak rather than the width at the base.  Assuming a Gaussian peak shape,


The number of theoretical plates in the column, N, is

Efficient columns have small H and large N for a given component.  The theoretical plate height is affected by the following experimental parameters:

·         mobile phase flow rate

·         diffusion coefficient of the solute in the mobile phase

·         diffusion coefficient in the stationary phase (depends on temperature and viscosity)

·         retention factor

·         diameter of the particles packing the column

·         thickness of the liquid coating on the stationary phase


The resolution of two adjacent peaks, RS, is determined by their separation and their widths.

In other words, RS depends on both migration rates and zone broadening.  A resolution of 1.5 means that the overlap of the peaks is about 0.3%, so conditions should be optimized to achieve at least this resolution if possible. 

In this experiment, you will adjust only the composition of the mobile phase to optimize the retention factors and resolution.  We will not attempt to optimize the zone broadening independently by changing the column or the flow rate.

Components in Diet Soft Drinks

            The ingredient list for most diet soft-drinks includes caffeine, benzoic acid, and aspartame (Nutrasweet®).  The structures of these compounds are shown below along with their UV-vis spectra. 



MW = 194.19

pKa = 10.4

Benzoic Acid




Provided by instructor:

mobile phase  –           HPLC-grade methanol

                                   20 mM phosphate buffer, pH 3

standard samples  –     mixture containing caffeine, aspartame, benzoic acid and uracil

                              –     individual samples of each of the above components        

Student prepared:

diet cola sample –        degassed for 20 minutes with air stream followed by filtration with 0.22mm syringe filter

caffeine standards –    100% = 0.045g/250 mL water

                                    85% dilution




                                    (and additional dilutions as necessary to bracket cola sample)


Model 1100 Hewlett-Packard HPLC system

ZORBAX reversed-phase bonded C18 HPLC column (4.6 mm x 15 cm)

250mL volumetric flask

40-200 mL automatic pipet

200-1000 mL automatic pipet

1mL HPLC vials with caps

vial crimper

analytical balance

Notes on Use of HPLC:

·         Your instructor will show you how to use the equipment and the software.  A primer guide/refresher is also found in the appendix.

·         At the beginning of the day, first open the valve with the black knob on the front of the pump manifold.  This sends the mobile phase to waste instead of the column.  Allow the pump to run for about 10 minutes at 5 mL/min to purge the lines of any air bubbles.

·         At the end of the day, run an 80% water/20% methanol mixture through the column at 1 mL/min for 10 minutes followed by 100% methanol for 10 minutes.  This should flush the column of any potential salt forming materials and stores the column in a compatible solvent.

Sequence of Analysis:

1.      Determine optimum conditions for separation of caffeine, benzoic acid and aspartame using the standard sample provided.  Start with a 75% buffer/25% methanol mobile phase mixture at 1 mL/min, 1mL injection volumes, and a 20 minute run time.  Uracil has been added to the standard mixture to provide a dead time marker (i.e. uracil is unretained).  The diode array detector can monitor several wavelengths at once as well as record the entire spectrum at specified intervals.  Use the UV-vis spectra of the components shown above to choose appropriate wavelength(s) to monitor the chromatograms.  Vary the ratio of buffer to methanol to achieve an acceptable separation and finally adjust the run time to end after the last component exits the column.  Use these data to calculate the retention factors for each component, the resolution between each pair of adjacent peaks, and the values for H and N for caffeine under these conditions.

2.      Using the conditions determined above and the pure samples of each component provided, determine the retention time of each component (i.e. identify the peaks).

3.      Using the column conditions determined above and 5mL injection volumes, run three samples of your most concentrated caffeine standard to determine the precision of the injections.  Then run the remaining caffeine standards and the diet cola sample to determine the concentration of caffeine in your sample using a calibration curve.  Check the full spectrum of each peak against spectra of the pure components to verify that the peaks represent isolated components.

The chemstation control software is accessed via the start menu, programs, HP Chem Stations, and then instrument 1 online.  All of the components are controlled via the run control window shown below by accessing the instrument sub-menus for each of the components.  Notice also that the status of each component is indicated on the run control screen.

The parameters are set in the various “set” menus.  The “more” menus contain control sub-menus that allow you to actually turn these components on.

You can either run individual samples one at a time using the Run Method command under the Run Control menu, or run a sequence of samples using the Sequence menu and it’s various sub-menus.  If you choose the run method option, use the Sample Info menu under the Run Control menu to tell the computer where to store your files and what to name each run.  If you use the sequence mode, you can enter this information in the sequence parameters sub-menu.

Once the data has been collected, use the View menu to see the report containing the chromatograms at each monitored wavelength and the integrated peak areas etc.  From here you can print the report as well as view and print full spectra at any peak.




NAME ________________________________                          DATE ______________


Mass caffeine   _____________

Dilution volumes:       standard    /      water

                       A:        _______  /  _________

                       B:        _______  /  _________

                       C:        _______  /  _________

                       D:        _______  /  _________

                       E:        _______  /  _________

Optimization of Conditions:

% buffer                     ____________

% methanol                ____________

flow rate                     ____________

column pressure         ____________


Retention Time (min)

Retention Factor (k’)














Benzoic Acid




Rcaffeine,aspartame  _______________

Hcaffeine  ___________________

Ncaffeine  ___________________

Quantitation of Caffeine:

wavelength chosen for quantitation  __________________


[caffeine] (mg/L)

Peak Area

diet cola    

avg. peak area for A  ____________________  95% CIm  __________________

slope   ___________________

std. dev. slope  _________________        95%CI slope   ____________________

intercept  ___________________

std. dev. intercept  _______________      95%CI intercept   __________________

R2  ________________

covariance  ___________________

[caffeine]diet cola  _____________________  95%CI  ____________________

Discussion Questions:

Discuss how the choice of monitored wavelength affects the sensitivity of the analysis.  Would you choose the same wavelength to quantify benzoic acid or aspartame?

Based on the pKa’s of the components in our samples, why do you think a mobile phase buffer of pH 3 was chosen? 

If significant zone broadening had resulted in unsatisfactory resolution in our experiment, what might we have changed to reduce its affect?  Comment on the feasibility of this.

Compare the precision obtained for the three injections of the same sample to the precision of your calibration curve.  Which limits the precision of your unknown concentration?  Would using an internal standard be justified?

Show sample calculations and attach all relevant chromatograms and spreadsheet printouts.

Bir yanıt yazın

Başa dön tuşu