Department of Chemistry

Stage 2 Chemistry Social Relevance Projects.

Home | Contents | Cross Referenced Index | Experimental Procedures | Data Analysis

Determination of caffeine in soft drinks by High Performance Liquid Chromatography.


The following experiment illustrates the utility of high performance liquid chromatography (HPLC) as an analytical tool.

In HPLC, narrow columns with internal diameters 2-80 mm are used. These columns are packed with particles having an average diameter of less than 50 microns (50 x 10-6m). Such columns require high pressures (1000-6000 psi) to maintain a convenient flow of the eluting solvent, usually in the range 0.1-10 mL/min, but occasionally higher. Resolution is considerably superior to that achieved with an ordinary column, in part because of the tight packing of the stationary phase, which reduces lateral diffusion, and because of the large surface area of the packing.

Compared with classical column chromatography, where the columns are gravity fed and a separation can take hours or even days, HPLC can offer analysis times of 5-30 min. Such times are comparable to that needed for GLC analyses.

HPLC is especially suited to the analysis of compounds not readily assayed by GLC. For example, thermally labile compounds can be analysed by HPLC at ambient temperatures, and highly polar or involatile compounds can be analysed. Sample treatment is often minimal since aqueous solutions can be used in HPLC.

Since its inception in the late 1960's, HPLC has made significant practical impact on the areas of pharmaceutical, clinical, forensic, environmental and industrial research and development analyses, and is an ideal complementary technique to GLC. Preparative HPLC has also found an important use in the isolation and purification of various compounds.



There are two main classes of column: "normal" and "reversed" phase.

Normal phase columns are most usually packed with silica gel; they work in the partition/adsorption mode in the same manner as a normal silica gel column in conventional chromatography.

Reversed phase chromatography, which is the most common form of HPLC, is a type of partition chromatography. Frequently, reversed phase columns are packed with a chemically bonded octadecylsilyl coated silica; such columns are referred to as C-18 and are very non-polar. Other popular bonded phase columns have octasilyl, cyanopropyl, or phenylsilyl packings.

The eluent used with reversed phase columns is relatively polar, e.g. MeOH/H2O. Unlike normal phase chromatography, the more polar components of a mixture elute first, since these partition relatively unfavourably on the highly non-polar packing. Increasing the polarity of the solvent increases the retention time of a particular component. The situation is the reverse of normal adsorption chromatography:

Normal vs. Reversed Phase

 Normal Reverse
Packing polarity High Low
Solvent polarity Low High
Elution order Non-polar first, then polar Polar first, then non-polar
Effect of increasing solvent polarity Decreases retention time Increases retention time


In this experiment, the amounts of benzoic acid and caffeine in various commercially available soft drinks will be quantitatively determined. These components can be selectively detected in the presence of substantial quantities of other materials (e.g. sugar) by choosing u.v. absorption (254 nm) as a method of detection. An accurately known amount of a reference compound will be added to a known volume of the soft drink, and the peak areas will be compared. If the detector response for each compound, compared to the reference compound, is known (or can be determined), it follows that the absolute amounts of the component can be determined. This method of quantitative determination is known as the internal standard method, and gives results of high accuracy, which do not depend on the volume of the injection. The internal standard method is also widely employed in GLC analyses.

Operation of the HPLC - General Procedure

Cartridge: 8 x 100 mm m-Bondapak C18 (a reversed phase column).

Solvent: 0.5% phosphoric acid in 40% aqueous methanol.

The greatest enemy of HPLC is fine particulate matter, which can damage the pumping system and irreversibly block the column. Therefore, all solvents have been filtered through fine membranes (0.4-0.5 micron) and all solutions to be injected MUST be prepared either with filtered solvent, or filtered as specified later in these notes.

The Waters instrument comprises three main components: the injector (located on the top right-hand corner of the console), the solvent delivery system (right-hand side) and the u.v. detector (left-hand side).

Start-up Procedure

First, ensure that there is sufficient filtered solvent in the reservoir for your run.

Using the screw on the right hand side of the RCM, pressurise the column to 800 psi (beginning of the yellow region). Switch the solvent selector on the inlet manifold at the front of the pump to the running solvent. Switch on the power to the pump and slowly increase the flow rate to 3 mL/min. Switch on the U.V. detector and once the absorbance reading has settled (ª10 min), set the zero to 0.01 AU. Switch on the Data Module and enter the date and time (eg. 1999^ 03^ 16^ 09^ 32^) and answer 0^ to "Is this a new file?"

To inject a sample: Switch the injector lever (top) to "load". Switch the lower lever to vertical and remove the plug (store in hole in injector switch). Wipe the needle with a clean tissue and insert into the injector. Inject the sample into the loop with even pressure (excess of solvent in the loop will be pushed out of the vent tube on the right of the injector). Replace the injector plug and move the low lever to the horizontal position. Smoothly switch the injector lever to "inject", and at the same time press "inject" on the data module to start the data collection. The data module plots a real time chromatogram, and at the end of the run time (15 min) replots the chromatogram with details of retention time (RT), peak area (A or H) and relative areas of the peaks (conc). Although the integration is not affected if a plotted peak is off scale, the chromatogram can be replotted at a different attenuation by resetting the ATN (powers of 2, the bigger the attenuation setting the smaller the peaks will look, normally set at 30) and then recalling the plot (Recal). The plot is stored until the next injection. "Feed" moves the paper forward for the next plot.

Shut down procedure

Stop the flow slowly and change the solvent selector to flushing solvent (methanol). Increase the flow to 3mL/min slowly and flush the column with solvent for 10 min. Stop the flow and switch off the pump. Remove the plots from the Data module and switch off at the front of the unit. Switch off the U.V. detector. Depressurise the column by unscrewing the pressure screw on the right of the RCM until 4 threads are showing.

Notes: Normal running pressure for this experiment is between 1500-2200 psi. If a high pressure shutdown occurs (>2500 psi), consult a demonstrator. Look for leaks at connections through the system; if there are any, consult a demonstrator.

Listen to the pump during the experiment; if there are any unusual noises, consult a demonstrator. Look at the outlet flow; it should be a thin stream. If there is no flow, consult a demonstrator.

Determining detector response ratios

Pipette 5.00 mL of each of 1.00 mM aqueous m-methoxybenzoic acid (the internal standard) (provided), 0.200 mM caffeine and 1.00 mM benzoic acid into a 25 mL conical flask equipped with a ground glass joint. Mix well. Inject 50 microlitres onto the column. All components should elute within 15 min.

Identify each peak by injecting solutions (20 microlitres) of the pure components individually.

From the trace of the mixture, determine the ratio of the areas relative to that of the internal standard, and hence determine the relative molar response factors for caffeine and benzoic acid relative to the internal standard.

Standard solutions

Run A

Peak Area

Relative Area

Molar Response Ratio
Benzoic acid      
p-Methoxybenzoic acid      

Run B - duplicate as above

Determining the amount of caffeine and benzoic acid in soft drinks

Obtain about 10 mL of an appropriate soft drink from the refrigerator. Do not leave the lid off the container for longer than is necessary to transfer your sample.

Pipette accurately 5.00 mL of the filtered drink sample and 5.00 mL of the internal standard into a 25 mL conical flask and mix thoroughly. Inject 25 microlitres onto the column and record the trace. Determine the relative peak areas and correct for the molar detection responses of the individual components. Using the known concentration of the internal standard, calculate the concentration of those components present in the soft drink. Express this concentration also in grams/litre.



(a) Data should be presented in the form of a Table which is easily read and which clearly shows how the values were obtained.

(b) Calculations should be clearly shown with an explanation for each step.



Harris, D.C. "Quantitative Chemical Analysis; 5th Edition"; W.H. Freeman and Company: New York..

Home | Contents | Cross Referenced Index | Experimental Procedures | Data Analysis