Simple and Rapid Determination of Denaturants in ... - Springer Link

Simple and Rapid Determination of Denaturants in Alcohol Formulations by Hydrophilic Interaction Chromatography 2006, 63, 373–377

ˇ . Daunoravicˇius1, I. Juknaite_ 2, E. Naujalis3, A. Padarauskas1,& Z 1

2 3

Department of Analytical and Environmental Chemistry, Vilnius University, Naugarduko 24, 03225, Vilnius Lithuania; E-Mail: [email protected] Department of Chemistry, Vilnius Pedagogical University, Studentu St. 39, 08106, Vilnius Lithuania Laboratory for Metrology in Chemistry, Semiconductor Physics Institute, A. Gostauto 11, 01108, Vilnius Lithuania

Received: 16 February 2006 / Revised: 3 March 2006 / Accepted: 6 March 2006 Online publication: 10 April 2006

ously must exhibit exceedingly unpleasant taste or odour. Denatonium benzoate is among the most bitter substances known to man and as such, has been used for at least 25 years as a denaturant for ethyl alcohol and vegetable oils and as aversive agent when added to toxic substances to prevent accidental ingestion, particularly by children and animals [1]. It is marked under trade name Bitrex. The Lithuanian Government Regulations specify the level of Bitrex in specially

denaturated alcohol to be 20 mg L)1. In most Lithuanian alcohol formulations containing Bitrex, other denaturants such as methylene blue or crystal violet can also be present. The concentration levels of these components are usually 5 times lower than that of Bitrex. The rate of duty on the denaturated alcohols and its products depends on whether the specifications are met. Both companies adding denaturants to their products and governmental agencies monitoring compliance need a simple analytical technique for the rapid and reliable analysis of the denaturated alcohol and its products. Methods available for the determination of Bitrex involve high-performance liquid chromatography (HPLC) [2–8], gas chromatography [9] and capillary electrophoresis (CE) [10]. The determination of methylene blue or crystal violet in the denaturated alcohol formulations usually is performed spectrophotometrically. Therefore an approach which enables the determination of the denaturants in a single analysis is highly desirable. To our knowledge, no reports concerning the simultaneous determination of Bitrex, methylene blue and crystal violet were found in the literature. The current report describes the use of polar stationary phases for the simultaneous determination of three denaturants in the hydrophilic interaction chromatography (HILIC) mode. The method was validated and successfully applied to the analysis of several denaturated alcohol formulations.

Chromatographia 2006, 63, April (No. 7/8)

373

Abstract A hydrophilic interaction chromatography (HILIC) technique has been developed and validated for determination of common denaturants (denatonium benzoate, crystal violet and methylene blue) in denaturated alcohol formulations. Among the three different polar stationary phases (i.e., aminopropyl, cyanoethyl and silica) studied the cyanoethyl phase provided much stronger retention for the organic cations. It was shown that high efficiencies were reached only with anionic ion-pairing reagent that reduces the interactions with the silanol groups. The anion ionpairing strength under HILIC conditions was: acetate < formate < trifluoroacetate < perchlorate. This study also investigated the effect of various experimental factors on the retention of the cyanoethyl stationary phase, such as acetonitrile content, pH and ion-pairing anion concentration in the mobile phase. The separation of three denaturants was achieved in about 8 min with a mobile phase containing 60% (v/v) acetonitrile and 10 mmol L)1 HClO4. The proposed method was validated and applied to the determination of danaturating agents in various Lithuanian denaturated alcohol formulations.

Keywords Hydrophilic interaction chromatography Denaturants Alcohol formulations

Introduction The ethyl alcohol is second only to water as a universal solvent. It is widely used in making commercial products such as lotions and perfumes, antifreeze and cleaning fluids, anticeptics, insect repellents and many others. In order to make ethyl alcohol unfit for use as a beverage it must be denaturated before it leaves the distillery or petrochemical plant where it is produced. The denaturants used obvi-

Original DOI: 10.1365/s10337-006-0768-z 0009-5893/06/04

2006 Friedr. Vieweg & Sohn/GWV Fachverlage GmbH

Inc.) was used for data acquisition. Direct UV detection was employed at 214 nm.

H3C Cl + CH3 N

Reagents and Solutions

H3C

N

N CH3

Crystal Violet

CH3

CH3

C6H5COO-

CH3

C2H5

N

N

CH3

O

+

C2H5

Denatonium benzoate

N H3C

S

N CH3

+

Cl Methylene Blue

N

CH3

CH3

Fig. 1. Chemical structures of the compounds studied

Experimental

HPLC-grade acetonitrile was purchased from Merck (Darmstadt, Germany). Crystal violet (98%), denatonium benzoate (99%) and methylene blue (98%) were obtained from Labochema (Vilnius, Lithuania). Trifluoroacetic acid (99%), HClO4 (70%) and sodium salts (i.e., acetate, formate, perchlorate and trifluoroacetate) were of ACS grade from Sigma-Aldrich (St. Louis, MO, USA). All mobile phases and standard solutions were prepared using doubly distilled water. Stock solutions (500 mg L)1) of the analytes were prepared in ethanol. Working solutions were prepared immediately before use by serial dilution with mobile phase. All mobile phase and sample solutions were filtered through a 0.45 lm nylon 66 membrane filter (Supelco, Bellefonte, PA, USA) and degassed by ultrasonication.

Results and Discussion

Instrumentation

Comparison of Polar Stationary Phases

The HPLC instrumentation consisted of a Waters Model 501 high-pressure pump, an injection valve equipped with a sample loop of 20-lL and a Waters LambdaMax Model 481 variable wavelength UV detector set to absorb at 220 nm. The results and data were collected and plotted on a plotter/integrator SP 4290 (Spectrophysics, San Jose, CA, U.S.A.). Chromatographic separations were performed on columns (150 mm · 3 mm I.D.) with three different (5 lm aminopropyl, cyanoethyl and bare silica) Separon SGX stationary phases (TESSEK Ltd., Prague). The mobile phase flow rate was 0.5 mL min)1. Capillary electrophoretic determination of Bitrex was performed on a P/ACE 2100 apparatus (Beckman Instruments Inc., Fullerton, CA, USA) equipped with a UV detector. Fused silica capillary (Polymicro Technology, Phoenix, AZ, USA) of 75 lm I.D. and 57 cm total length (50 cm to the detector) was used. Samples were introduced by electrokinetic injection (5 kV for 10 s). System Gold software (Beckman Instruments

All three denaturants are highly polar compounds with basic (positively charged) character dominating at most pH values (Fig. 1). Hydrophilic interaction chromatography provides an attractive approach to effectively separate polar analytes on polar stationary phases with aqueous-organic mobile phases [11]. Simple acetonitrile-aqueous mobile phases can be employed to retain polar compounds while eluting nonpolar compounds relatively quickly and, consequently, to avoid problems associated with sample matrix interferences. The HILIC mechanism involves partitioning of the analyte into and out of the adsorbed water layer on the polar stationary phase surface. Additionally, depending upon the pH of the mobile phase, a positively charged analyte can undergo cation exchange with the negatively charged silanol groups. Three commercially available columns representing different polar stationary phases (aminopropyl, cyanoethyl and bare silica) were selected for this

374


study. The columns are all silica-based and have the same dimensions and similar properties in terms of particle size, pore size and surface area. A test mixture of the analytes was investigated using conventional HILIC mobile phase (90:10 v/v acetonitrile/water containing 10 mmol L)1 ammonium chloride). UV detection at 220 nm was chosen as a good to detect all three analytes. However, for all the columns tested the methylene blue was failed to elute and significant peak tailing was evident in the case of the remaining analytes. This might be the result of ion-exchange interaction between the cationic analytes and negatively charged surface silanol groups on the silica based stationary phases. Addition of anionic ion-pairing reagents such as trifluoroacetate (TFA), perchlorate and some others to mobile phase has generally been designed, for silica-based stationary phases, to neutralize the positively charged groups, thereby suppressing negatively charged silanol interactions with positively charged analytes [12–14]. Fig. 2 compares the elution profiles of the three compounds obtained in the presence of TFA. As expected, when 5 mmol L)1 sodium trifluoroacetate was added to a mobile phase of acetonitrile/water (90:10 v/v) the retention of analytes was dramatically reduced and the peak shapes were greatly improved. As shown in Fig. 2, the analytes displayed about twice stronger retention on the cyanoethyl column than on the silica column, but elution patterns were very similar with Bitrex and crystal violet co-eluting on both columns. Significantly stronger effect of the TFA on the retention was observed for the analytes on the aminopropyl column: under the conditions studied all three compounds eluted with the front. The fact that analytes were not retained on the aminopropyl phase indicated that this phenomenon might be specific to the amino phase, possibly due to the electrostatic repulsion between the cationic analytes and protonated amino groups. The results suggest that the choice of stationary phase may offer some opportunity to gain resolution but that dramatic changes in selectivity are not likely. Further work was then conducted with the cyanoethyl stationary phase and the influence of the different mobile phase parameters was studied. Original

Separation Optimization Although TFA showed very strong ionpairing properties, it exhibits some UV absorbance and therefore causes an increase in the baseline noise at 220 nm. For this reason, in addition to TFA), other anions, namely perchlorate, formate, and acetate were investigated to determine the appropriate anion strength for eliminating ion exchange interactions. The mobile phases used in this study contained 80% acetonitrile and 10 mmol L)1 of appropriate salt. All four salts were sodium salts to assure any effects of the salts on analyte retention behaviour may be assigned solely to the anionic counterion. The pHs of the aqueous salt solutions were not adjusted before mixing with acetonitrile. Table 1 shows the retention times of the analytes on the cyanoethyl column in the mobile phases containing four ion-pairing reagents. The addition of acetate or formate allowed the elution of all compounds in about 20–30 min but did not eliminate tailing. In contrast, TFA) and ClO4) were strong enough to eliminate tailing and sequester the apparent ion exchange interactions. The anion ion-pairing strength under HILIC conditions on cyanoethyl stationary phase was: acetate < formate < TFA) < perchlorate, where ClO4) was the strongest ion-pairing reagent. An explanation of the fact that less hydrophobic perchlorate is more effective in reducing of electrostatic interactions than TFA) may lie in the stronger chaotropic properties of the ClO4) anion which is therefore dehydrated much more readily than TFA). Strong ion-pairing reactions require the exclusion of water molecules from the interaction between oppositely charged species, i.e., the ion-pairing anions must be dehydrated to form the ion-pair with the cationic analytes [14]. Perchlorate containing mobile phases also have the advantage of good transparency for lowwavelength UV detection. Since no improvement in selectivity was apparent over the pH range 2–6 (results not shown), subsequent work was done under acidic conditions using perchloric acid as ion-pairing reagent. Furthermore, at low pH values the suppression of free silanol ionization additionally decreases undesirable ionic interactions. The effect of perchloric acid concentration on the retention was investigated

Original

Table 1. Retention time of the analytes with different sodium salts Sodium salt ClO4TFACHCOOCH3COOa

a

Retention time, min Bitrex Crystal violet

Methylene blue

5.65 7.81 11.8 16.7

9.12 13.2 20.6 29.4

5.80 7.95 11.2 15.5

Mobile phase: 80:20 v/v acetonitrile/water containing 10 mmol L)1 sodium salt

by varying HClO4 concentration from 1 to 15 mmol L)1 (Fig. 3). It is clearly seen that all three analytes show a similar and significant decrease in retention as the amount of perchloric acid in the mobile phase increases. As expected, higher HClO4 concentrations weakened the ionic interactions between cationic analyte groups and negatively charged free silanols, thus leading to decreasing retention. In addition, the formed less hydrophilic ion-pairs have higher affinity to the mobile phase which also results in shorter retention times. However, the concentration of HClO4 has no significant influence on the resolution of Bitrex and crystal violet: both analytes partially or completely coelute in the perchloric acid concentration range studied. Another important parameter affecting the separation performance is the amount of organic solvent in the mobile phase. The acetonitrile concentration was varied from 50 to 95%. The plots of retention time versus acetonitrile concentration are presented in Fig. 4. As can be observed, the retention decreases initially as the acetonitrile content decreases and at a level of 60% acetonitrile, there is a minimum in retention. A further decrease of the less polar component of the mobile phase led to a gradual increase in retention indicating that in addition to hydrophilic interactions, non-polar (hydrophobic) interactions with the cyanoethyl stationary phase become important as the polarity of the mobile phase increases. Based on these results, an organic/aqueous solvent ratio of 60:40 (v/v) was selected for further separations because it provided the shortest separation time with acceptable resolution. The chromatogram obtained under optimum conditions for a standard solution is shown in Fig. 5. As can be seen, an excellent separation of three compounds was obtained in about 8 min.


Fig. 2. Separation of denaturants on (a) cyanoethyl, (b) silica, and (c) aminopropyl columns. Conditions: mobile phase 5 mmol L)1 NaTFA in acetonitrile/water (90:10 v/v); flow rate 0.5 mL min)1; UV detection at 220 nm. Peaks: 1=Bitrex; 2=crystal violet; 3=methylene blue

Method Validation Once the optimized conditions were selected, the method was validated with

375

Table 2. Recovery study of denaturants (mg L)1) spiked to synthetic samples (n=3) Sample

Analyte

Spiked

Found

Recovery (%)

Ethanol

Bitrex

5.00 20.0 2.00 5.00 2.00 5.00 5.00 20.0 2.00 5.00 2.00 5.00

4.93 20.4 1.96 4.94 2.03 5.10 5.00 19,7 2.05 5.06 1.94 4.86

98.6 102.0 98.0 98.8 101.5 98.5 100.0 98.5 102.5 101.2 97.0 97.2

Crystal violet Methylene blue Glass cleaner

Bitrex Crystal violet Methylene blue

30 Bitrex

Methylene blue

Crystal violet

Retention time, min

25 20 15 10 5 0

0

10 15 Perchloric acid, mmol/L

20

Fig. 3. Effect of HClO4 concentration on the retention of Bitrex, crystal violet, and methylene blue. Conditions: mobile phase HClO4 in acetonitrile/water (80:20 v/v); other conditions as in Fig. 2

14 Bitrex

Methylene blue

Crystal violet

Retention time, min

12 10 8 6 4 2

40

50

60 70 80 Acetonitrile, %

90

100

Fig. 4. Effect of acetonitrile concentration on the retention of Bitrex, crystal violet, and methylene blue. Conditions: mobile phase 10 mmol L)1 HClO4 in acetonitrile/water; other conditions as in Fig. 2

respect to the following parameters: stability of the solutions, linearity, method limit of quantification, accuracy and precision.

376

Fig. 5. Chromatogram of a standard solution under optimized conditions. Conditions: mobile phase 10 mmol L)1 HClO4 in acetonitrile/ water (60:40 v/v); flow rate 0.5 mL min)1; UV detection at 220 nm. Peaks: 1=Bitrex; 2=crystal violet; 3=methylene blue

Although the stability test is often considered as a part of the ruggedness of the procedure, it should be carried out at the beginning of the procedure validation because it conditions the validity of the data of the other tests. The response factors of standard and sample solutions were found to be unchanged for at least up to three months. Less than 1.5% concentration difference was found between the solution freshly prepared and those aged for three months. The solutions can therefore be used within this period without the results being affected. The linearity of the method was tested by preparing a calibration curve for each analyte with six points. The tested concentration range was from 1 to 50 mg L)1 and each concentration level was injected three times. The assay showed linearity with a relative standard deviation (RSD) £ 1.8% for the relative responses (peak area divided by concentration) obtained in the tested concentration range and correlation coefficients >0.999 found for all three analytes. In addition, Chromatographia 2006, 63, April (No. 7/8)

it was checked that at the lower limit of the analytical range, defined as method limit of quantification, analytical performances were satisfactory. Consequently, for the first concentration (1 mg L)1) of the analytical range, the signal to noise ratio was measured and the RSD of the peak area was calculated with five replicates. Signal to noise ratio ranged from 12 for methylene blue to 46 for Bitrex and repeatability RSD from 1.4% for Bitex to 3.8% for methylene blue. Thus, for all three compounds the signal to noise ratio was higher than 10 and the RSD less than 5%. The detection limits (three times the baseline noise) were 0.08, 0.15 and 0.30 mg L)1 for Bitrex, crystal violet and methylene blue, respectively. These data support the suitability of the proposed method for its application to real samples. To evaluate the accuracy of the method, two synthetic samples free from the denaturating agents were prepared. Both samples were then spiked with the analytes at two different concentrations and analysed using the calibration curves obtained with the external standard method. The results obtained are summarized in Table 2. It can be seen in the table that the concentrations found were generally in good agreement with the added concentrations, with recoveries between 97 and 102%. These results suggest that interferences by the other matrix components are not significant and the HILIC conditions are suitable to obtain adequate method accuracy. Method precision was determined by measuring repeatability and intermediate precision (between-day precision). In order to determine the repeatability of the method, six replicate injections of two synthetic samples spiked with denaturants at three concentration levels (2, 10 and 25 mg L)1) were carried out. The intermediate precision was evaluated over 3 days by performing six successive injections daily. Relatively little dispersion was observed for retention times because both repeatability and intermediate precision were less than 0.7%. For peak areas, repeatability was in the range 1.2–2.6%, whereas intermediate precision ranged from 1.5 to 3.3%, depending on the analytes. Irreversible adsorption of sample matrix components is generally a bigger concern with polar stationary phases than with reversed phase systems. In order to determine if such phenomenon Original

Table 3. Determination of denaturants (mg L)1) in various denaturated alcohol formulations (n=3) Sample Denaturated ethanol Glass cleaner Windshield washer fluid a

Analyte Bitrex Bitrex Methylene blue Bitrex Crystal violet

Sample Analysis

occurred with our system, the performance of the column was re-evaluated after a period of four weeks of daily work. The number of theoretical plates obtained for each analyte was greater than 93% of the initial value and peak tailing did not increase. Retention and selectivity also indicated no sign of column deterioration. These results

Original

Alternative technique a

20.1 (1.8) 19.4 (2.2) 3.94 (3.1) 9.66 (2.5) 2.06 (3.0)

20.4 (3.4) 19.8 (4.2) 3.83 (1.5) 9.50 (3.7) 2.01 (1.8)

Values in parentheses are RSDs (%).

demonstrate that proposed HILIC system seems to be rugged enough for routine analysis.

Fig. 6. HILIC analysis of (a) windshield washer and (b) glass cleaner sample solutions. Conditions as in Fig. 5

HILIC

reversed-phase HPLC for the determination of positively charged analytes. Simple acetonitrile-aqueous mobile phases can be employed to retain polar compounds while eluting nonpolar compounds relatively quickly and, consequently, to avoid problems associated with sample matrix interferences.

To evaluate the proposed HILIC system for the real samples, it was applied to the determination of denaturating agents in several Lithuanian alcohol formulations. Fig. 6 shows the typical chromatograms for the windshield washer (a) and glass cleaner (b) sample solutions. Several samples were analyzed by the proposed method and by alternative techniques. The alternative analysis requires two separate determinations for the same sample: Bitrex was determined by CE method [10] and the determination of crystal violet and methylene blue was performed by spectrophotometric technique. The results are compared in Table 3. As can be seen, HILIC method showed good agreement with data obtained from alternative techniques. The comparison of means using a t-test has shown that there is no statistically significant difference between them at a confidence level of 0.05. In conclusion, the above results show that HILIC using polar stationary phase offers an attractive alternative to

1. Hansen SR, Janssen C, Beasley VR (1993) Vet Hum Toxicol 35:234 2. Glover MJ, Blake AJ (1997) Analyst 97: 891 3. Sugden C, Mayne TG, Loscombe CR (1978) Analyst 103:653 4. Damon CE, Pettitt Jr BC (1980) J Chromatogr 195:243 5. Kovar J, Loyer M (1984) J Liq Chromatogr 7:2163 6. Gadzala-Kopciuch RM, Buszewska T, Buszewski B (2000) J Liq Chromatogr Rel Technol 23:3133 7. Faulkner A, Demontigny P (1995) J Chromatogr A 715:189 8. Henderson MC, Neumann CM, Buhler DR (1998) Chemosphere 36:203 9. Ng LK, Hupe M, Harnois L, Lawrence AH (1998) Anal Chem 70:4389 10. Pranaityte B, Daunoravicius Z, Padarauskas A (2004) Chromatographia 60:353 11. Alpert AJ (1990) J Chromatogr 499: 177 12. Cai B, Li J (1999) Anal Chim Acta 399:249 13. Chen Y, Mehok AR, Mant CT, Hodges RS (2004) J Chromatogr A 1043:9 14. Shibue M, Mant CT, Hodges RS (2005) J Chromatogr A 1080:49


377

References