ANALYST. JUNE 1984. VOL. 109 Determination of Cobalt Chromium and Vanadium as 8-Hydroxyquinoline Complexes by High-performance Liquid Chromatography Lauri H. J. Lajunen," Erkki Eijarvi and Tim0 Kenakkala Department of Chemistry University of Oulu SF-90570 Oulu 57 Finland The HPLC behaviour of Co(ll) Cr(lll) and V(V) chelates of 8-hydroxyquinoline was studied on silica-gel, reversed-phase and size-exclusion columns using tetrahydrofuran - chloroform methanol - water or aceto-nitrile and tetrahydrofuran as mobile phases respectively. The calibration graphs of peak area or height versus the amount of metal injected over the low nanograms to milligrams range were linear. The relative standard deviations were between 0.5 and 5%. Keywords Cobalt chromium and vanadium determination; h ig h-perfo rmance liquid chroma tog rap h y; 8-h ydroxyquinoline During the past few years increased attention has been paid to the use of high-performance liquid chromatography (HPLC) for the separation identification and determination of metal complexes and organometallic compounds.The complexation reagents studied most with regard to the HPLC of metal complexes are various dithiocarbamates,4-8 but there are also some papers dealing with the HPLC of 8-hydroxy-quinoline metal chelates by means of silica-gel columns.9-13 In this work we studied the separation of cobalt chromium and vanadium as 8-hydroxyquinoline complexes by thin-layer chromatography (TLC) and HPLC techniques. Silica-gel, reversed-phase and size-exclusion columns were used. The main aim of the study was to investigate the simultaneous determination of Co Cr and V by HPLC using different techniques.Experimental Instrumentation HPLC studies were performed on a Perkin-Elmer 1220 liquid chromatograph equipped with a UV detector operating at 254 nm. Columns The silica column was 250 x 4 mm i.d. packed with 5 pm LiChrosorb SI 60 (E. Merck) the reversed-phase column was 250 x 4 mm i.d. packed with 10 pm LiChrosorb RP-8 or RP-18 (E. Merck) and the size-exclusion column was 250 X 8 mm i.d. packed with 10-15 pm Shodex 80115 (Showa Denko). TLC Plates TLC studies were performed using 20 x 10 cm plates coated with silica gel G (E. Merck). Reagents All reagents were of analytical-reagent grade. 8-Hydroxy-quinoline (Fluka) was used without purification.All metal salts were used without purification. All solvents were distilled and passed through a column packed with silica gel and aluminium oxide before use. * To whom correspondence should be addressed. HPLC Procedure The pH of 0 - 6 ml of the metal salt solution was adjusted to the desired value (4.5) with an acetate buffer solution the solution was diluted to 10 ml then 300 mg of solid 8-hydroxy-quinoline were added. The mixture was kept in a rotating retort at 90 "C for 30 min. When the mixture had cooled it was extracted with 20 ml of chloroform. The excess of ligand present in the chloroform layer was destroyed with 0.1 M HCl or NaOH solution (for 5 ml of chloroform solution 20 ml of HCl or NaOH solution were added). Both the HC1 and NaOH treatments were effective.After the acid or base treatment, the chloroform phase (which contained the metal chelates) was evaporated to dryness at room temperature and dissolved in an eluent solution. Before each HPLC run the air was removed from the solvents by using a vacuum ultrasonic mixer or a helium flow. The columns were cleaned before use with methanol (flow-rate 0.2 ml min-1 24 h). TLC Procedure Spots of sample solution containing metal chelates in chloro-form were placed on the plates the solvent was evaporated at room temperature then the TLC runs were executed. Results and Discussion TLC The experiments in which an excess of ligand was present in the organic phase were unsuccessful. The ligand acid present in the organic phase was destroyed by HCI or NaOH treatment.The best separations were obtained with tetrahydrofuran - toluene or tetrahydrofuran - chloroform (Table 1). For the cobalt - 8-hydroxyquinoline system two separate spots were obtained probably owing to the oxidation of Co(I1) to Co(II1). Table 1. RF values of the 8-hydroxyquinoline complexes for various solvent mixtures Solvent Co(1) Co(2) Cr V Tetrahydrofuran - chloroform Tetrahydrofuran - toluene Methanol - chloroform ( 5 + 95) . 0.61 0.24 0.58 -(60 + 40) . . . . . . . . 0.52 0.21 0.47 -(60+ 40) . . . . . . . . 0.36 0.05 0.33 700 ANALYST. JUNE 1984 VOL. 109 HPLC Qualitative analysis For the silica-gel column the best separation was obtained by using tetrahydrofuran - chloroform (60 + 40) as the eluent.Deviations of up to about 10% in the volume ratio of the solvents did not significantly change the locations of the peaks on the chromatograms. However there are some problems with regard to the silica-gel columns it is necessary to use absolutely water-free solvents and also the conditions in the column change continuously with time. Fig. 1 shows the chromatogram for the separation of Co Cr and V chelates with the silica-gel column. The small peak before the V peak arises from the increased amount of tetrahydrofuran because of the vaporisation of chloroform. When using a reversed-phase column the eluent must be more polar than the stationary phase and the solvent mixtures used were therefore methanol - water and acetonitrile - water. The best results were obtained by using proportions of 63 + 37 and 40 + 60 respectively.Examples of runs with a reversed-phase column are shown in Fig. 2. For vanadium - 8-hydroxy-quinolinate no peaks were obtained on using a reversed-phase I co I 1 I I I 0 2 4 6 Timeimin Fig. 1. Separation of cobalt -. chromium - and vanadium - 8-hydroxy-quinolinates on silica gel. Conditions SI 60 column eluent. THF -CHCl (60 + 40); flow-rate 1 ml min-1; sensitivity of detector 0.05 absorbance unit ) Cr column. The precision obtained for the cobalt - and chromium -8-hydroxyquinolinate peaks was very satisfactory. The rela-tive standard deviation for the retention times of these peaks under the same experimental conditions was less than 1 YO. Runs with a molecular size-exclusion column (tetrahydro-furan as eluent) showed that it is not possible to separate the present chelates satisfactorily by this technique.Quantitative analysis In studying quantitative analysis by HPLC a silica-gel column and an RP-8 reversed-phase column were used (an RP-18 column gave virtually the same results as the RP-8 column). All the samples were prepared as described above. Calibra-tion graphs of peak area versus metal concentration for the metal 8-hydroxyquinolinate systems using the silica-gel col-umn are shown in Fig. 3. The slopes indicate that the sensitivity for the cobalt - 8-hydroxyquinolinate system is the highest being about 1.5 and 4.1 times high than those for the chromium and vanadium systems respectively. The linearity of all the calibration graphs is very good; the linear regression coefficient was better than 0.999 for each system.With respect to the precision of the method the relative standard devia-tions for four successive injections of the same sample were Co 2% Cr 0.4% and V 5%. The detection limits defined as the concentration where the peak height was three times the background when the sensitivity of the detector was at a maximum (0.010 absorbance unit) were 50 and 75 pg for cobalt and chromium respectively when the injection volume was 20 pl. There were difficulties in measuring the vanadium peak height with regard to the maximum sensitivity of the Table 2. Relationship between peak size and sensitivity of the detector Metal con-centration/ Sensitivity Peak height/ Peak area/ Metal mg 1 - I of detector cm cm2 co .. . . 5.0 0.5 4.4 2.2 5.0 0.2 10.8 5.4 2.5 0.1 10.8 5.4 1.25 0.05 10.8 5.4 Cr . . . . 6.7 0.5 4.1 1.64 6.7 0.2 10.7 4.28 6.7 0.1 21.8 8.72 0.67 0.05 4.6 1.84 7 Cr Time/mi n co Cr 1 c Fig. 2. Separation of cobalt - and chromium 8-hydroxvquinolinates using a reversed-phase column. Conditions ( a ) RP-8 column eluent. MeOH - H,O (63 + 3 7 ) flow-rate. 1 ml min I ; sensitivity of detector. 0 . 1 absorbance unit ( h ) RP-18 d u m n eluent CH,CN - H,O (40 + 60); flow-rate. 1 ml min- 1 ; sensitivity of detector. 0.2 absorbance unit (c) RP-18 column eluent. MeGH - H 2 0 (63 + 3 7 ) flow-rate. 1 ml min-1; sensitivity of detector 0.1 absorbance uni ANALYST JUNE 1984 VOL. 109 70 1 40 N E 2 30 Y a g 20 Metal concentration/mg I-' Fig.3. Calibration graphs for (A) cobalt - (B) chromium - and (C) vanadium - 8-hydroxyquinolinates when using a silica- el column. Conditions SI 60 column; eluent THF - CHC13 (60 + 40$ flow-rate, 1 ml min-1; sensitivity of the detector 0.05 absorbance unit "E-:4 1 -Y m al a 2 -I I I I I 0 0.1 0.2 0.3 0.4 0.5 Metal concentrationhng 1-1 Fig. 4. Calibration graphs for (A) cobalt - and (B) chromium -8-hydroxyquinolinates usiyg a reversed-phase column. Conditions: RP-8 column; eluent MeOH - H 2 0 (63 + 37); flow-rate 1 ml min - l ; sensitivity of detector 0.02 absorbance unit detector because of the large solvent peak (Fig. 2). However, the detection limit for vanadium can be assumed to be less than 0.5 ng. The calibration graphs for cobalt - and chromium -8-hydroxyquinolinates using the reversed-phase column are shown in Fig.4. The sensitivity for cobalt was about 1.6 times higher than that for chromium. The linear regression coeffi-cient for the calibration graphs was about 0.999 for both systems. The relative standard deviations for four successive injections of the sample were 1 and 1.9% for Co and Cr, respectively. The detection limits were 100 pg for Co and 150 pg for Cr when the injection volume was 20 1-11. The relationship between the peak size and sensitivity of the detector is illustrated in Table 2. The size (both the height and area) of the peak remains unchanged when the sensitivity of the detector is doubled and the metal concentration is halved.For example the peak heights for solutions containing 5 and 0.5 mg 1-1 of cobalt are the same (4.4 cm) when the sensitivity of the detector is exactly ten times higher in the second run. By changing the sensitivity of the detector concentrations of 10 and 15 mg 1-1 for the upper limits of the linear range for Co and Cr respectively were obtained. 1. 2. 3. 4. 5 . 6. 7 . 8. 9. 10. 11. 12. 13. References Schwedt G. Chromatographia 1979 12 613. Tollinche C. A . and Risby T. H. 1. Chromatogr. 1978 16, 448. Eijarvi E. and Lajunen L. H. J. Kem. Kemi 1983 10,707. Liska O. Lehotay J. Brandsteterova E. Guiochon G. and Colin H. J . Chromatogr. 1979 172 384. Moriyasu M. and Hashimoto Y. Chem. Lett. 1980 117. Tande T. Pettersen J . and Torgrimsen T. Chromato-graphia 1980 13 607. Bond A . and Wallace G. Anal. Chem. 1982 54 1706. Smith R. and Yankey L. Analyst 1982 107 744. Berthod A . Kolosky M. Rocca J.-L. and Vittori O., Analusis 1979 7 395. Wenclawiak B. Fresenius Z. Anal. Chem. 1981 308 120. Wenclawiak B. Fresenius Z. Anal. Chem. 1982 310 144. Hoffmann B. and Schwedt G. J . High Resolut. Chromatogr. Chromatogr. Commun. 1982 5 439. Hambali C. and Haddad P. Chromatographia 1980 13, 633. Paper A31386 Received November I l t h 1983 Accepted December I9th 198