Simultaneous Determination of Benzoic Acid and Caffeine Concentrations by Single Wavelength, Isocratic High-Pressure Liquid Chromatography

Nancy Karreman1* and Tom Zuzelski2
Student1, Teacher2: Bainbridge High School, 9330 High School Road, Bainbridge Island, WA, 98110.
*Corresponding author:



Sodium benzoate is a common preservative added to commercially available foods and beverages. Once dissolved in a polar solution, sodium benzoate dissociates into sodium ions and benzoic acid. It then decomposes into a federally classified human carcinogen, benzene. As such, the international limit on benzoic acid in consumer products is set at 0.1% by mass. Caffeine is also a common additive found in consumer products, which has various negative cardiovascular effects on the human body when ingested in large amounts. Both chemicals were extracted from samples of six different brands of soda by single wavelength, isocratic high-pressure liquid chromatography (HPLC). Caffeine content of all samples was determined to be within 0.016 mg/mL on average of given company standards. The mass percent of benzoic acid in all samples was determined to be below the legal limit of 0.1%. The method developed for this protocol could potentially be used to determine contents of commercial substances by single wavelength isocratic HPLC, improving efficiency and replacing the multiple wavelength method as a more cost-effective approach. 


High-pressure liquid chromatography, sometimes known as high-performance liquid chromatography (HPLC), is an analytical tool used to separate chemical compounds from a mixture for analysis or purification. Chromatography in general involves the movement of solutes over a stationary phase. For reverse-phase liquid chromatography, a non-polar column packed with carbon or silica, usually a chemically bonded octadecylsilyl coated silica, or C-18 column, is eluted with a liquid mobile phase. Samples to be separated are passed over the column’s surface by the eluting liquid. Solutes that are strongly adsorbed on the stationary phase will be retained for a longer period of time than those weakly adsorbed, allowing the separation and identification of samples based upon their retention times. To minimize the interaction between mobile and stationary phase in reverse-phase HPLC, an isocratic mixture of water and an organic polar solvent is used in combination with a non-polar column.1

Initially, HPLC columns were produced in diameters of 1 to 5 cm, however, the size of stationary phase materials has gradually decreased, allowing for the development of microparticulate columns numbering 1.5 to 10 µm in diameter. Because of the small size of the particulate and the tightly packed nature of the stationary phase, high pressures are necessary to overcome the resistance of the particles to solvent flow. A flow rate of 1-5 ml/min is typically used in analytical HPLC.2  

Columns are usually made of stainless steel and, though expensive, are reusable so their cost is dispersed across a large number of samples. Some manufacturers make HPLC systems out of plastics or glass, a more attractive option to the economically minded.2

Currently, the technique is utilized for a broad range of chemical analyses including the analysis of food, drugs, and agrochemicals. Fast, reverse-phase HPLC has been investigated and employed by the pharmaceutical industry for separations to determine methodology and to purify drug substances and products.3

Caffeine is one of the most frequently ingested pharmacologically active substances present in a variety of foods, beverages, and medications. Caffeine ingestion can have adverse effects on human health, having the most effect upon the “at risk” groups of reproductive-age women and children. Risks include general toxicity, cardiovascular effects, decreased bone status and calcium balance, behavioral changes, increased incidence of cancer, and loss of fertility in males. It is recommended that daily intake of caffeine be limited to 400 mg for healthy adults, 300 mg for pre-menopausal women, and 2.5 mg kg-1 for children.4

Benzoic acid is most commonly used as a food preservative, protecting against fungi, bacteria, and yeast in acidic environments. The least complex aromatic carboxylic acid, benzoic acid (C6H5 -COOH) is very soluble in most organic solvents. Once dissolved, benzoic acid and its salts have the capability to react with ascorbic acid to form benzene.5 Benzene is known to be a human carcinogen, which, in acute cases, causes myelogenous leukemia and increases risk of total lymphatic and hematopoietic cancers.6 A recent study suggests that exposure levels of as little as 10 ppm over the course of up to 26 years increases the number of chromosomal aberrations in peripheral lymphocytes, creating a significant number of chromosome breaks and gaps.7 Though primarily used as an industrial and chemical solvent rather than in the production of soft drinks, the unintentional ingestion of benzene has caused international regulations on allowable amounts of benzoic acid to be implemented, setting the limit at 0.1% by weight in food.8

Single wave length, isocratic (i.e. elution at constant solvent composition) HPLC presents an effective, time conscious method by which concentration of substances within a densely combined mixture can be determined. This potential was the main catalyst for the creation of this assay.

The established procedure for simultaneous determination of benzoic acid and caffeine involves lengthy and complex methods such as variable wavelength scanning and the use of a gradient mobile phase. Changing the composition of the mobile phase during the run is called gradient elution. A gradient must be established for the non-isocratic method to be effective, requiring a considerable volume of mobile phase. Both methods result in significantly elevated run times compared to fixed wavelength isocratic systems.9            

Ideally the calibration curve of benzoic acid generated by this assay will display good linearity as well as a regression factor of at least 0.99. Soft drink samples should display logical results with concentrations of the analyte within the legal range of 0 to 0.1%. The optimal result of this assay would be to establish an efficient, reliable, and replicable method for the simultaneous determination of benzoic acid and caffeine concentration that can be implemented effectively and affordably.

Materials and Methods

A mobile phase was prepared with a ratio of 60/40 methanol/water. All solutions were prepared using HPLC grade solvents. Sodium phosphate was added as a buffer and the solution adjusted to a pH of 5.0. Solid benzoic acid was dissolved in the mobile phase to a concentration of 1100 ppm and degassed to prepare a stock standard solution, then stored and chilled at 1.8oC. Concentrations of 22, 44, 66, 88, and 110 ppm were diluted from the original sample mixture. All chemicals were obtained from Flinn Scientific.

Caffeine was extracted from Vivarin tablets by the process of recrystallization.  The tablets were dissolved in hot isopropyl alcohol and gravity filtered to remove insoluble impurities. Then, the solution was recrystallized and filtered a second time. The resultant powdered caffeine was dissolved in mobile phase to a stock concentration of 1100 ppm, degassed, and chilled. Concentrations of 44, 88, 132, 176, and 220 ppm were then diluted from the 1100 ppm mixture.

Samples of six different soft drinks of varying ingredients were purchased from a local grocery store. 5 mL of each soft drink sample (Diet Coke, Coca-Cola, Mountain Dew, Sprite, Gold Peak Unsweetened Tea, and Red Bull) were diluted with the methanol/water mobile phase to 25 mL. All samples were vacuum degassed to remove effervescence and filtered (through pore size 45µm) to remove impurities.

The injector (Rheodyne 7725i) consisted of a 20 microliter loop to ensure precise and consistent injections. The ultraviolet light detector (Gilson HM Holochrome UV Monitor) was set to a wavelength of 270 nm, the optimal setting to simultaneously observe both caffeine and benzoic acid retention as determined by preparatory experimentation. The measurements of separation and determination of benzoic acid were performed using a SP880 Ternary HPLC pump and a SP4270 integrator from Spectra Physics. The column was an Altima C-18 column with a particle diameter of 5 µm.

Initial conditions were set at approximately 4,000 psi and 1.5 mL min-1 with 100% of solvent running from line A as determined by preparatory experimentation. The column was equilibrated with a non-buffered 60/40 methanol/water solution before and after each running session.


The retention time for benzoic acid was observed at 4.07 ± 0.03 minutes. A calibration curve was determined from the injection and evaluation of the prepared 22, 44, 66, 88, and 110 ppm benzoic acid solutions. The relationship between concentration and peak area was linear as expected, with a regression factor of 0.999 (Fig. 1). The chromatographic peaks of the benzoic acid samples were clear and well resolved as shown by Fig. 2.


benzoic acid calibration

Figure 1. Calibration curve for benzoic acid.



benzoic acid chromatogram

Figure 2. Example chromatogram of 6 ppm benzoic acid. (x=time in minutes, y= peak area (BC)).


The retention time of caffeine was 2.51 ± 0.01 minutes. The injection and determination of the aforementioned 44, 88, 132, 176, and 220 ppm caffeine solutions determined the calibration curve. An excellent regression factor of 0.998 was observed in the linear fit applied to the curve (Fig. 3). The peaks for caffeine were clean and isolated as illustrated in Fig. 4.


caffeine calibration

Figure 3. Calibration curve for caffeine.



  • 12 Ppm Caffeine

    Figure 4. Chromatogram of 12 ppm caffeine.



Under these optimized conditions, the HPLC method was applied to the determination of benzoic acid and caffeine content in samples of relatively unknown composition and unpredictable chromatographic visibility. The preservative was identified in the samples by comparing the known retention times of the standard pure solutions to the retention times that were displayed on the chromatogram of the samples. The same process was used to identify and determine caffeine content in the beverage samples (as shown in Figure 5).


  • Mountain Dew caffeine

    Figure 5. Chromatogram of Mountain Dew sample with caffeine (2.53) and benzoic acid (4.11) peaks. (x=time in minutes, y= peak area (BC)).



The results for the determination of benzoic acid concentration in these samples are displayed in Fig. 6. Caffeine peak area is shown in Fig. 7.


benzoic acid in soda

Figure 6. Graph of relative peak area of benzoic acid in soda samples.



caffeine in soda

Figure 7. Graph of relative peak area of caffeine in soda samples.


The observed peak areas were then converted to mg/mL by comparing each concentration to those of the standard tested caffeine (Table 1) and benzoic acid (Table 2) peak areas.


red bull caffeine

Table 1. Caffeine concentration in soda samples.



benzoic acid coke

Table 2. Benzoic acid concentration in soda samples.



The results derived from this study confirm the original hypothesis: single wavelength HPLC is an accurate and effective method to determine concentrations of benzoic acid and caffeine in soft drink samples. Isocratic methods dramatically reduce the amount of mobile phase utilized. The pump volume can thus be scaled up to increase efficiency. This method not only allows for an efficient determination but also limits cost. Rather than using a detector that employs multiple wavelengths to determine concentration, a basic single wavelength detector can be used to achieve the same outcome.

Due to its simplicity, the method outlined by this assay is not only cost-effective but durable. Within a simple and technologically limited high school lab setting, the method proved to be fully functional and accurate. The durability of the column was also exhibited over the two month time period in which this assay was achieved. The structure of the C-18 packing within the column lends itself to reproducibility and allows it to remain relatively unaffected by chemical analytes of various acidity and astringency.

The HPLC method lends itself to a diverse range of applications, from pharmaceutical to consumer safety programs that protect the general public from elevated amounts of potentially harmful chemicals. This assay could be applied as a standard procedure for ensuring legal restrictions on content of benzoic acid and caffeine are met by beverage companies. It would not only be an efficient and cost-effective manner in which to ensure these restrictions are met, but also applicable to measurement of a vast variety of other carcinogenic and potentially harmful chemicals commonly found in consumer products such as acetaldehyde, butylparaben, lanolin, and monosodium glutamate.

The following limitations need to be considered along with the method outlined in this paper. Firstly, only single trials were performed. Secondly, the purity of the caffeine extracted from Vivarin tablets may have been compromised and potentially interfered with the measurements of concentration in the standards. Lastly, it is impossible to tell the degree to which the method outlined in this study accurately determined the amount of benzoic acid in the soda samples because the amounts of benzoic acid within the actual soft drinks were not released by their manufacturers. Therefore, while the amount of benzoic acid was determined to be within legal limits in all samples, the exact accuracy of the measurements is unknown.


  1. Kupiec, Tom. (2004). Quality-Control Analytical Methods: High Performance Liquid Chromatography. International Journal of Pharmaceutical Compounding, 8, 223-227.

  2. High Performance Liquid Chromatography, 2, John Wiley & Sons. 2-9. Lindsay, Sandie, and Kealey, D. (1987).

  3. Gerber, Frederick. (2004). Practical aspects of fast reversed-phase high-performance liquid chromatography using 3 μm particle packed columns and monolithic columns in pharmaceutical development and production working under current good manufacturing practice. Journal of Chromatography A, 1036, 127-133.

  4. Nawrot, P., Jordan, S., Eastwood, J., Rotstein, J., Hugenholtz, A. & Feeley, M. (2003). Effects of caffeine on human health. Journal of Food Additives and Contaminants, 20, 1-30.

  5. Concise International Chemical Assessment Document, No. 26. International Programme on Chemical Safety. (2000).

  6. Report on Carcinogens Research Triangle Park, NC: U.S. Department of Health and Human Services, Public Health Service, National Toxicology Program. 499. (2011).

  7. Al-Ganimi, Y., Al-Saadi, A., Zaidan, H., Ewadh, M. &Al-Ameri, Q. (2013). Cytogenetic Effects of Benzene on Human Blood Cells. Journal of Natural Sciences Research, 3, 7-12.

  8. Opinion on Benzoic Acid and Sodium Benzoate, Scientific Committee on Consumer Products, European Commission. 6-7. (2005).

  9. Ree, M. & Stoa, E. (2011). Simultaneous determination of aspartame, benzoic acid, caffeine and saccharin in sugar-free beverages using HPLC. Concordia College Journal of Analytical Chemistry, 73-77.

  10. Center for Science in the Public Interest. (2012). Caffeine content of food and drugs.


Ray Weigand at Grace Davison Discovery Sciences for donating the column, injector, tubing, fittings, and other equipment.