摘要:

ANALYST, NOVEMBER 1986, VOL. 111 1265 Liquid Chromatographic Procedure for the Separation and Characterisation of Simple Urea - Formaldehyde Reaction Products Peter R. Ludlam and James G. King Borden (UK) Ltd., North Baddesley, Southampton SO5 9ZB, UK and Robert M. Anderson Department of Engineering, Dorset Institute of Higher Education, Norwich Union House, Christchurch Road, Bournemouth BHI 3NG, UK A procedure is described for separating about 20 simple, low relative molecular mass urea - formaldehyde reaction products by liquid chromatography on an aminopropyl column using aqueous acetonitrile as eluent. The technique is shown to have considerable advantages over previously published methods. The preparation and characterisation of some reference compounds, which so far have been little studied, are given in detail.The use of reaction mixtures of urea and formaldehyde for identifying peaks is described. Almost all of the peaks observed in commercial urea - formaldehyde formulations can now be identified by this technique. Keywords: Liquid chromatography; urea - formaldehyde resins separation; characterisation; urons Resins produced by the condensation of urea and formal- dehyde occupy an important position in many branches of industry. Although numerous analytical investigations into the course of the urea - formaldehyde (UF) reaction and the nature of the finished resin have been undertaken, there are still aspects of its chemistry that are not understood. This is in part due to the complex series of reactions that occur between urea and formaldehyde, leading to the formation of methylol and methylene compounds and both linear and cyclic ethers (urons).Thus the number of compounds produced can be very large indeed. Difficulties also arise owing to both the instability of UF compounds and the very limited solubility of compounds having a relative molecular mass greater than about 200. Reliable information on the average properties of UF reaction products has been obtained (notably by NMR measurements, 1 which show the over-all chemical composi- tion and by size exclusion chromatography,2 which gives the range of relative molecular masses) but separation and quantitation of individual components in the reaction mixture has proved more difficult. Thin-layer chromatography has been used to identify and quantify a small number of the low relative molecular mass compounds that are formed, but its range is limited and the experimental procedure is long and difficult to perform.3.4 Some efforts have been made to investigate urea - formal- dehyde reaction products and other amide - formaldehyde compounds by liquid chromatography but a simple, rapid, efficient and universal method has not yet been reported.Kumlin and Simonson5.6 studied UF condensates using a cation-exchange resin in the Li+ form. The resin used was of large particle size and although a separation of about 10 compounds was achieved, the efficiency of the column was low and a full analysis run took some 120 min to complete. Murray et a1.7 used a reversed-phase column to separate biuret , triuret and methylenediurea, which occur as impurities in fertiliser-grade urea, and Davidson8 separated urea, methylenediurea and dimethylenetriurea in a similar fashion in UF fertiliser compositions.Beck et al.9 used a cation- exchange resin in the Li+ form to examine durable press finishes. Textile finishes were also studied by Kottes Andrewslo using a reversed-phase column. In this paper, the limitations of these liquid chromato- graphic procedures are discussed and a method is reported that will separate UF condensation products efficiently and quickly. Details are given for the preparation of reference materials and of the methods used for identifying the chromatographic peaks. Preliminary Investigations In order to separate the many compounds formed when urea and formaldehyde react together and to ensure that the analysis is complete in as short a time as possible, it was decided to investigate the potential of reversed-phase silica columns rather than the io exchange columns used by Kumlin and Simonson,s as, these were reported to be inefficient and slow.In the course of the investigation, several 5-ym reversed- phase packing materials from various manufacturers were examined, including Zorbax ODs, LiChrosorb ODS and Spherisorb ODs. A mixture of urea, monomethylolurea (MMU), dimethylolurea (DMU) and methylenediurea (MDU) was used as a simple test solution and it was found that the elution patterns were slightly but significantly different. With Zorbax and LiChrosorb, DMU and MDU were not separated and with Spherisorb the resolution of urea and MMU was inadequate.However, by coupling a Zorbax column to a Spherisorb column, complete resolution of the test mixture could just be achieved. Although the analysis time was long and the separation was barely adequate, this column configuration was used for several months to estimate low relative molecular mass compounds in UF compositions. In order to attempt to improve the analysis, a column packed with Hypersil CZ2 Super was used in the hope that the longer carbon chains would retain the molecules of interest to a greater degree. The results were disappointing, showing little improvement when compared with the usual ODS packings. The limitations of chromatographing UF compounds in the conventional reversed-phase mode are as follows.1, Resolu- tion of the simplest compounds is very difficult to achieve; with more complex mixtures considerable peak overlap occurs and the chromatography becomes meaningless. 2 , Many compounds such as uron and its methylol derivatives elute at the solvent front even when using the weakest solvent and it does not seem likely that this difficulty can be overcome. The analysis time is about 25 min for a typical mixture. It was thought that a better separation of the methylol compounds could be achieved by using a more polar column. Also, the increased interaction between the polar stationary phase and the UF compounds could give greater flexibility with the eluting solvent. A column was packed with a hydroxy-terminated material (LiChrosorb Diol) and the test solution of urea, MMU, DMU and MDU was chromatographed using water as the eluting solvent.The elution pattern was markedly different to that P1266 ANALYST, NOVEMBER 1986, VOL. 111 obtained with the hydrophobic ODS columns. The hydroxy compounds DMU and MMU eluted first, followed by urea and MDU. It was apparent that the -OH - - - NH- hydrogen bonding effect was more powerful than the -OH - - - OH- interaction and compounds were eluting according to their =NH content. This order of elution from the column was contrary to the ideal pattern and it was considered likely that the chromatography of UF mixtures would be complicated on such a column. The possibility of reversing this elution pattern by using an amine column seemed worth investigating.An aminopropyl-terminated phase, Techsil NH2 ( 5 pm), was examined and, using the aqueous solvent as before, all the test compounds eluted at the solvent front. However, an amino column has the characteristics of a “normal” phase and water is likely to be the strongest solvent available. Using methanol as a potentially weaker eluent, a partial separation was achieved and the elution pattern (parent compound, monomethylol compound and dimethylol compound) seemed most encouraging. Methanol was replaced by acetonitrile in the hope that the absence of solvent - solute hydrogen bonding would, firstly, accentuate differences between the zero, mono- and dimethylol compounds thus giving better separa- tion of the test mixture and secondly, increase the elution time. This proved to be so, and after a final adjustment of the solvent strength by the addition of 10% water, an almost ideal system for separating the test mixture was obtained.The flexibility of the system was demonstrated when less polar compounds such as uron and its methylol derivatives, hitherto eluting at the solvent front, were completely sepa- rated by reducing the solvent strength, i.e., the water content. Thus a system that seemed almost ideal for the chromato- graphy of simple urea - formaldehyde compounds was developed. The column had high efficiency (8000 plates), the peak shape under normal conditions was good, the analysis time was short, being approximately 10 min for the test mixture of urea, MMU, DMU and MDU, and the elution pattern was favourable. Experimental Apparatus The chromatographic equipment used consisted of a Waters 6000 A pump, a Rheodyne 70-10 injection valve fitted with a 100-pl loop and a Model 70-11 filler port, a Waters R401 differential refractometer and a Waters Model 730 Data Module.A 2-pm in-line filter was positioned between the injector and the column. The analytical columns employed were made from 250 x 4.6 mm i.d. stainless-steel tubing with zero dead volume fittings. The columns were packed using the slurry technique developed by Kirkland.11 The initial upward displacement with methanol was fol- lowed by downward displacement and “slamming” to improve the stabilisation of the bed. The slurry used for packing the 5-pm aminopropyl-bonded silica to produce the columns used for the majority of this work was prepared by dispersing 3.5 g of Techsil NH2 ( 5 pm) packing material in 70 ml of chloroform - methanol (3 + 1) by ultrasonic agitation. The amino- propyl-bonded silica columns usually have an adequate lifetime, but when used for the analysis of materials with a high proportion of free formaldehyde they can deteriorate quickly.This is possibly due to an irreversible amino - aldehyde interaction taking place and is seen by shortened retention times and a lack of resolution of methylenediurea (MDU) and asymmetric dimethylolurea (asym. DMU). In order to attempt to offset this problem, a small amount of ammonia (0.01 M) is added to the eluting solvent. Depending on the type of sample under investigation, the life of an amino column is between 50 and 200 working days.Column packing and testing take only 2-3 h and consequently column deterioration is not considered to be a serious problem. One of the preparative columns was packed using Partisil5- ym silica by the same upward - downward - “slamming” procedure previously described. Stainless-steel tubing, 300 x 7.8 mm i d . , was used for the column and 5.5 g of the packing material was slurried in 70 ml of chloroform - methanol (1 + 1). Octadecylsilane (ODs)-bonded silica (Partisil 10) was used in the other preparative column. Column packing was by the method already described after slurrying in propan-2-01. The eluting solvent used for most of the work was prepared by mixing 900 ml of acetonitrile of Hypersol grade (BDH Chemicals, Poole, UK) with 100 ml of de-ionised water and 0.5 ml of ammonia solution (sp.gr. 0.880). The solvent mixture was de-gassed with helium and the temperature allowed to rise to room temperature before use. If the mixtures under examination contained predominantly methyl ethers or urons, i.e., compounds with short elution times, then improved resolutions were obtained by weakening the eluting solvent to 2.5 or 5% water in acetonitrile. Other solvent and chemicals, except where stated, were general-purpose laboratory reagents. Preparation of Samples Where possible, reference materials (5-10 mg) were dissolved directly in 10 ml of eluting solvent. If limited solubility was a problem, a solution in 1 ml of water was prepared, with gentle heating if necessary, and then diluted to 10 ml with aceto- nitrile.With resinous samples, about 200 mg were taken and low-condensed materials that were totally soluble were dissolved directly in the eluting solvent. However, when the average molecular mass was large and a considerable propor- tion (up to 80% of the UF) was insoluble, the above techniques were unsatisfactory. The best approach was then to dissolve the sample in 1 ml of dimethylformamide and dilute to 10 ml with acetonitrile. The sample plus solvent was shaken vigorously and allowed to stand. If necessary, the sample solution was filtered through a 0.5 pm filter before injection. Reference Compounds and Mixtures Monomethylolurea (MMU), dimethylolurea (DMU) , methyl- enediurea (MDU), monomethylolurea monomethyl ether (MMU.MME), dimethylolurea dimethyl ether (DMU.DME) and dimethylurea monomethyl ether (DMU.MME) were all synthesised by methods published earlier.3 The following reaction products of urea and formaldehyde were prepared. 1.Alkaline UF condensate, molar ratio 1 : 3. Urea, 0.6 g; 50% aqueous formaldehyde solution, 1.8 g; disodium hydrogen orthophosphate (Na2HP04. 12H20), 0.2 g; water, 5 g. The buffer was dissolved in the formaldehyde solution and the urea added. The solution was stirred and allowed to stand at room temperature. Samples were taken at intervals. 2. Alkaline M D U : F condensate, molar ratio 1 1. MDU, 1.3 g; 50% aqueous formaldehyde solution, 0.6 g; disodium hydrogen orthophosphate (Na2HP04. 12H20), 0.1 g; water, 5 g. MDU was dissolved in the warmed water, Na2HP04 was added and, when dissolved, the formaldehyde was added.The mixture was allowed to stand at room temperature and samples were taken at intervals. 3. Alkaline MDU: F condensate, molar ratio 1 : 1.5. This was prepared as for No. 2, but 0.9 g of 50% formaldehyde solution was used. 4. Alkaline MDU: F condensate, molar ratio 1 : 2. This was prepared as for No. 2 but 1.3 g of 50% formaldehyde solution were used. 5. Dimethylene ether from DMU (Zigeuner and Pitter).12 DMU (10 g) was dissolved in 33 ml of 1 YO potassium carbonate solution. After 3 weeks, about 1 g of precipitate had formed. The supernatant liquor and the precipitated material were examined.ANALYST, NOVEMBER 1986, VOL. 111 1267 6. Crude dimethylenetriurea (DMTU). MDU (6.6 g, 0.05 M) was dissolved in 100 ml of water at 60 "C.Disodium hydrogen orthophosphate (0.2 g) was added and, when dissolved, 3 g (0.05 M) of 50% formaldehyde were added. The reaction mixture was allowed to stand for 1 h. The crude monomethylol MDU was filtered off, washed with cold water and dissolved in the minimum amount of water at 60-70 "c. Urea (30 g, 0.5 M) and 0.5 g of sodium dihydrogen orthophosphate were added and dissolved in the solution. After standing overnight, the crude DMTU was filtered off and washed thoroughly with water. 7. Dimethylol uron (DM uron). A 2 ml volume of 40% sodium hydroxide was added to 50% formaldehyde (72 g, 1.2 M). Urea (12 g, 0.2 M) was dissolved in this alkaline formaldehyde and the solution was heated to boiling for 1 min. The pH was adjusted to 8 with formic acid and about 40 ml of water and formaldehyde were removed by vacuum distillation at 40-50 "C in a rotary evaporator.Extraction with chloroform - acetonitrile (1 + 1) removed the dimethylol uron from the reaction mixture, presumably as the di(hemiforma1) (pure dimethylol uron is only sparingly soluble in this solvent mixture). This extraction procedure gave a good separation of the dimethylol uron from impurities such as DMU and TMU. DM uron was separated from the excess of residual formaldehyde and water by chromatography on a semi- preparative Partisil 5-pm silica column (300 x 7.8 mm i.d.) using 10% methanol in chloroform as the eluent. The structure was confirmed by infrared, 1H NMR (Figs. 1 and 2) and 13C NMR spectroscopy: ring C, 77.4 p.p.m.; chain C, 66.9 p.p.m.; carbonyl C, 152.3 p.p.m.; solvent, DMSO-d6, m.p.87-90 "C. 100 80 C 0 .- .- 4 60 5 40 I- s 20 I R I I I I I I I I 4000 3000 2000 1600 1200 800 400 Wave nu m bericm - 8. Uron and monomethyloluron. Pure dimethyl uron, 0.04 g (0.00025 M), urea, 0.03 g (0.0005 M) and sodium dihydrogen orthophosphate, 0.005 g, were dissolved in about 0.2 ml of water. The solution was heated carefully at 100 "C for 5 min, then cooled and extracted with 3 X 1 ml of acetonitrile. The combined acetonitrile extracts were evaporated to dryness at room temperature using a jet of air. The urons were separated efficiently on a 250 X 4.8 mm i.d. amine column using 2.5% water in acetonitrile as the eluting solvent. Uron itself was identified by infrared, 1H NMR (Figs. 3 and 4) and 13C NMR spectroscopy: ring C, 75.6 p.p.m.(in CDC13). Monomethyl uron was indicated by its chromato- graphic behaviour and its infrared and proton NMR spectra (Figs. 5 and 6). Insufficient sample was obtained for a 13C spectrum. I I Y , 1 I 1 20 4000 3000 2000 1600 1200 800 400 Wavenumbertcm-1 Fig. 3. Infrared spectrum of uron Fig. 1. Infrared spectrum of dimethylol uron 6.0 5.0 4.0 p.p.m. 1 I I 6.0 5.0 4.0 3.0 p.p.m. Fig. 2. a, 4.88; c, 4.65; and d, 5.72 p.p.m. 1H NMR spectrum of dimethylol uron (solvent DMSO-d6). Fig. 4. 5.35 p.p.m. 1H NMR spectrum of uron (solvent CDC13). a, 4.85; and b, 1 00 I 1600 1200 800 400 4000 3000 2000 Wavenum ber/cm - Fig. 5. Infrared spectrum of monomethylol uron1268 ANALYST, NOVEMBER 1986, VOL. 111 3 1.3 f 0 ck, ‘CH, a I I I1 HOCHzN, /NH b d c 0 I I 0 2 4 6 Ti me/rn in 7.0 6.0 5.0 4.0 p.p.m.Fig. 7. and glycerol (for peak identification, see Table 1) Chromatogram of test mixture; urea, MMU, DMU, MDU Fig. 6. d6). a, 4.70; b, 7.12; c, 4.62; and d, 5.63 p.p.m. lH NMR spectrum of monomethylol uron (solvent DMSO- I Results and Discussion Evaluation One of the major problems in the chromatography of urea - formaldehyde compounds on an aminopropyl column is the slow but inevitable change of retention characteristics. For the most part, this renders absolute retention times unreliable and it has been found to be useful to adopt a procedure of multiple relative retention times. The two compounds used for reference in this study were urea and DMU. Using this technique, the chromatograms are fairly easy to interpret.However, it is still considered to be good practice to run standard solutions at regular intervals so that the state of the column is continually monitored. The following are examples of solutions which can be used: urea, MMU, DMU, MDU and glycerol (Fig. 7); DM uron containing MM uron and uron itself if possible; methylolated MDU; and methylolated OMDU, all prepared in eluting solvent and stable for several months. Table 1 is a list, in increasing retention time, of the peaks observed in the chromatography of the various reference compounds and reaction mixtures. The retention times given are typical of those obtained with a new column . Also given are retention times relative (a) to urea for early eluting compounds and (b) to DMU for late eluting compounds. The elution times of dimethylformamide (DMF), water and formaldehyde are included. Glycerol elutes in a chromato- graphic “window” and is suitable for use as an internal standard for the determination of compounds such as urea, MMU, DMU, etc., in UF compositions.0 2 4 6 Tim e/m i n Fig. 8. and TMU (for peak identification, see Table 1) Chromatogram of reaction product 1 showing N,N-DMU recovery procedure had not significantly affected the purity of the compounds, small amounts of the collected materials were re-chromatographed. The infrared and proton NMR spectra unambiguously confirmed the nature of the peaks. It has been shown in a previous study3 that with urea - formaldehyde compounds there can exist a simple relationship between the logarithm of the retention time and the degree of substitution.This concept was again found to hold true (Fig. 9) and in this instance a plot of k’ against the number of methylol groups is a straight line. [k’ is the capacity factor, equal to (t, - t,)/t,; t, is the retention time of the compound and t, is the retention time of an unretained solute]. When log k‘ of asym. DMU and TMU are plotted to lie on the same line it can be readily seen that a second methylol substituent on a terminal nitrogen has only half the effect on the retention time as the first substituent. Peak Identification Simple compounds The chromatography of simple reference compounds and well characterised urea derivatives (see above) resulted in the easy and unambiguous identification of about half the peaks encountered in conventional urea - formaldehyde reaction mixtures.Methy lolureas Examination of reaction product 1 enabled N , N-dimethylol- urea (asym. DMU) and trimethylolurea (TMU) to be tenta- tively identified (Fig. 8). The peaks of interest were collected from the column and the solvent removed by careful evapora- tion at room temperature using a jet of air. To ensure that the Methy lolrnethy Lenediureas The reaction between MDU and formaldehyde is compli- cated, leading to 12 possible mono-, di- and trimethylol MDUs. However, many of these derivatives are unlikely and only about seven peaks are actually observed. It is possible toANALYST, NOVEMBER 1986, VOL. 111 1269 Table 1. Peaks observed in urea - formaldehyde reaction products in order of increasing elution time Retention timehin relative to No.Compound Structure 1 Dimethylformamide . . . . . . 2 Formaldehyde . . . . . . . . HCHO 3 Water . . . . . . . . . . . . 4 Dimethylolurea dimethyl ether . . CH30CH2NHCONHCH20CH3 5 Monomethylolurea monomethyl ether CH30CH2NHCONH2 n I I II . . . . . . . . . . . . HN\C/NH 6 Uron 0 0 HX' 'CH, 7 Monomethyloluron . . . . . . I I HN II 0 8 Dimethylolurea monomethyl ether . . CH30CH2NHCONHCH20H 0 HzC' 'CHZ \P/ . . . . . . . . I NCHZOH I 9 Dimethylol uron HOCHzN i 0 10 Urea . . . . . . . . . . . . H2NCONH, 11 Unknown . . . . . . . . . . 12 Monomethylolurea . . . . . . . . H2NCONHCH20H 13 Glycerol . . . . . . . . . . 14 N,N-Dimethylolurea . . . . . . H2NCON(CH20H)2 15 Methylenediurea . . . . . . . . H2NCONHCH2NHCONH2 16 sec-Monomethylolmethylenediurea . .H2NCONHCH2N(CONH2)CH20H 17 N,N'-Dimethylolurea . . . . . . HOCH2NHCONHCH20H 18 Oxymethylenediurea . . . . . . H2NCONHCH20CH2NHCONH2 19 Trimethylolurea . . . . . . . . HOCH2NHCON(CH20H), 20 Monomethylolmethylenediurea . . HOCH2NHCONHCH2NHCONH2 21 Asym. dimethylolmethylenediurea . . HOCH2NHCON(CH20H)CH2NHCONH2 or 22 Monomethyloloxymethylenediruea . , HOCH2NHCONHCH20CH2NHCONH2 23 Unknown . . . . . . . . . . 24 Dimethylenetriurea . . . . . . H2NCONHCH2NHCONHCH2NHCONH2 25 Dimethylolmethylenediurea . . . . HOCH2NHCONHCH2NHCONHCH20H 26 Dimethyloxymethylenediurea . . . . HOCH2NHCONHCH20CH2NHCONHCH20H 27 Trimethylolmethylenediurea . . . . HOCH2NHCONHCH2NHCON(CH20H)2 28 Trimethylolmethylenediurea . . . . HOCH2NHCONHCH2(CH20H)CONCH20H HOCH2NHCONHCH2N( CHZOH)CONH2 * D = definite, P = probable, T = tentative, U = unknown.Identification* Absolute Urea DMU 2.16 2.25 2.52 2.41 0.73 2.79 0.84 D D D 2.84 0.86 D 3.03 0.91 D 3.19 0.96 D 3.20 0.96 0.58 D U D D D D T D T D P T 3.32 1 0.60 3.87 1.17 0.70 4.12 1.24 0.75 4.45 1.34 0.80 4.87 1.47 0.88 5.05 0.91 5.31 0.96 5.53 1 5.60 1.01 6.53 1.18 6.75 1.22 7.35 1.33 P 7.7 1.39 U 8.6 1.56 D 9.2 1.66 P 9.85 1.78 P 10.5 . 1.90 T 11.2 2.03 T 12.2 2.21 obtain information about the nature of some of these peaks by reacting MDU and formaldehyde together at slightly alkaline pH (reaction mixtures 2, 3 and 4). With a low MDU : F ratio and a short reaction time the predominant product is monomethylolmethylenediurea (MMMDU). As the amount of formaldehyde and the reaction times are increased, di- and then trimethylol substitution is favoured.A typical chromatogram is shown in Fig. 10. As with the urea series, MDU and its mono- and dimethylol derivatives follow the simple relationship between log k' and the degree of substitution. It can be seen (Fig. 9) that two peaks occur between MMMDU and the symmetrical DMMDU. Using the behaviour of the dimethylolureas as a guide it seems likely that a second methylol substituent on the terminal nitrogen of MMMDU will have only half the effect of the first on the retention time. Plotting log k' of these two peaks on the MDU derivative line shows indeed that one peak behaves in accordance with the asym. DMMDU structure, H2NCONHCH2NHCON( CH20H),. Considering the posi- tion of the second peak, which should be one of the terminal nitrogen - chain nitrogen disubstituted structures (H2NCONHCH2N(CH20H)CONHCH20H or H,NCON- (CH20H)CH2NHCONHCH20H), it would seem that a methylol substituent on a chain nitrogen will have half the effect of a second substituent on the terminal nitrogen and only one quarter the effect of the first substituent on a terminal nitrogen (Fig.9). Two peaks occur at longer retention times, which again appear to follow the structure - retention time relationship described above. In this way these compounds have been tentatively identified as having the two most probable trimethylol MDU structures (Table 1). Methy loloxy methy lenediureas The insoluble material from reaction mixture 6 showed one major peak, which could safely be assumed to be DMOMDU.1270 ANALYST, NOVEMBER 1986, VOL.111 5 I 11 I I I I I Unsubstituted Mono NN di NN’ di NNN’ tri 1 = NH2CONHCH20CH2NHCONHz 2 = NH2CONHCH20CH2NHCONHCH2OH 3 = HOCH2NHCONHCH20CHZNHCONHCHzOH 4 = NH2CONHCH2NHCONHZ 5 = NHzCONHCHZNCONH2 I CHpOH 6 = NHzCONHCH2NHCONHCHzOH 7 = NH2CONCHzNHCONHCHzOH I CH20H /CH20H CH20H 8 = NH2CONHCHzNHCON \ 9 = HOCH2NHCONHCH2NHCONHCHzOH 10 = HOCHzNHCONCH2NHOONHCH20H /CH20H I CHZOH 11 = HOCHZNHCONHCHZNHCON 12 = NHZCONHz 13 = NHZCONHCHPOH \CH~OH / CHpOH 14 = NHZCON \CH~OH 15 = HOCHZNHCONHCHZOH /CH20H 16 = HOCHZNHCON \CH20H /O\ 17= I I 18= I I 0 HzC’ \CHZ HzC CH7 0 H2C’ \CHz 19= I 1 Fig. 9. Effect of degree of substitution on retention times. k’ = (tr - t&,, where t, is the retention time of the solute and to is the retention time of a non-retained solute 3 I 0 2 4 6 8 1 0 1 2 Time/m in Fig.10. Chromatogram of MDU and formaldehyde reacted at pH 8 for 8 h at ambient temperature (for peak identification, see Table 1) 1c 12 17 I I I I I 0 2 4 6 8 10 12 Timeimin Fig. 11. Chromatogram of UF reaction product U : F 1 : 1.4 reacted at 70 “C for 2 h at pH 9 (for peak identification, see Table 1) The chromatogram of the supernatant liquor showed another large peak at a shorter retention time, which is likely to be the mono compound. When log k’ for these compounds is plotted against the substitution pattern, a line having virtually the same gradient as the urea and MDU derivative line is obtained. Further, there is a small peak at the elution time corresponding to zero substitution, which can be tentatively attributed to the parent compound, oxymethylenediurea. Kumlin and Simonson6 identified MMOMDU and DMOMDU in the reaction mixture described by Zigeuner and PitterQ and characterised them conclusively by spectro- scopic means.Urons Chromatography of pure dimethyl uron (DM uron) from reaction mixture 7 and uron and monomethyl uron (MM uron) from reaction mixture 8 identifed the simple urons. A plot of log k’ for these compounds against the substitution pattern again gave a straight line as for the other three series of compounds previously discussed. Thus about 20 compounds formed in simple urea - formal- dehyde reactions have been identified, the nature of only two or three small peaks being as yet unknown. This study confirms the observations of Kottes Andrew@ and Kumlin and Simonsonl3 that there is no evidence for the occurrence of tetramethylolurea. Considering the formation of the uron ring, the implications are that either theANALYST, NOVEMBER 1986, VOL. 111 1271 trimethylolurea is converted into MM uron, which in the formaldehyde-rich reaction mixture is converted rapidly into DM uron, or, alternatively, tetramethylol uron is formed but is too unstable to be chromatographed. This technique has been used successfully for a year to characterise UF reaction products (Fig. 11) and to determine the more important compounds such as urea, MMU and DMU. It has proved to be a very powerful method for investigating the significance of various manufacturing parameters and will contribute in the future to a deeper understanding of urea - formaldehyde chemistry. References 1. 2. 3. Schindlbauer, H., and Schuster, J., Kunststoffe, 1983, 73, 325. Ludlam, P. R., and King, J. G., J . Appl. Polym. Sci., 1984,29, 3863. Ludlam, P. R., Analyst, 1973, 98, 107. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. Ludlam, P. R., Analyst, 1973, 98, 116. Kumlin, K., and Simonsen, R., Angew. Makromol. Chem., 1978, 68, 175. Kumlin, K., and Simonson, R., Angew. Makromol. Chem., 1981, 93, 27. Murray, T. P., Austin, E. R., Howard, R. G., Horn, R. C . , Anal. Chern., 1982, 54, 1504. Davidson, A. D., J. Assoc. Off. Anal. Chem., 1983, 66, 769. Beck, K. R., Leibowitz, B. J., Ladisch, M. R . , J. Chromatogr., 1980, 190, 226. Kottes Andrews, B. A., J. Chromatogr., 1984, 288, 101. Kirkland, J. J., J. Chromatogr. Sci., 1972, 10, 593. Zigeuner, G., and Pitter, R., Monatsh. Chem., 1955, 86, 57. Kumlin, K., and Simonson, R., Angew. Makromol. Chem., 1978, 72, 67. Paper A61132 Received May 2nd, 1986 Accepted June 26th, I986

ISSN:0003-2654    

DOI:10.1039/AN9861101265

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST NOVEMBER 1986 VOL. 111 1273 Selectivity and Efficiency Measurements in High-performance Liquid Chromatography Using Micellar Hexadecyltrimethylammonium Bromide in the Mobile Phase Frank G. P. Mullins and (the late) Gordon F. Kirkbright Department of Instrumentation and Analytical Science University of Manchester Institute of Science and Technology P.O. Box 88 Manchester M60 IQD UK Good chromatographic efficiency in high-performance liquid chromatography (HPLC) was obtained using micellar hexadecyltrimethylammonium bromide (CTAB) in the mobile phase. Dithiocarbamate salts of varying hydrophobicity were used as test solutes. The efficiency measurements obtained for the hydrophobic solutes (phenol and benzene) were compared with those obtained for the ionic dithiocarbamate salts.The influence of CTAB concentration above the critical micellar concentration on retention efficiency and selectivity and the effects of varying the concentration of organic modifier on the efficiency and retention with a micellar CTAB mobile phase were investigated. The chromatographic efficiency remained high when methanol was used as the mobile phase modifier. When acetonitrile was used as the modifier the efficiency obtained was poor. Conductance measurements present a possible explanation for the difference in efficiencies obtained for these widely used mobile phase modifiers. The measurements obtained with a micellar mobile phase were compared with those obtained with a non-m ice1 I e form i ng quaterna ry am mon i u m sa I t tetra bu ty la m mon i u m bromide.Keywords Efficiency measurements; high-performance liquid chromatography; micellar mobile phase; dith ioca rbama tes The chromatography of ionisable substances such as organic acids and amines often presents problems with respect to retention plate efficiency and peak symmetry. Ion-pair chromatographyl-4 has been shown to be valuable for the high-performance liquid chromatography (HPLC) of ionis-able compounds. Ion-pair chromatography has great flexibil-ity but the complexities of the ion-pair equilibria may result in peak tailing although these problems can usually be over-come by control of the mobile phase parameters. Previously we reported the retention characteristics of the dithiocarba-mate salts using tetramethyl- tetrabutyl- and tetrahexylam-monium cations5 but were unable to solve the problem of peak tailing even by controlling the mobile phase parameters.Other workers6 have noted the inadequacy of the ion-pair partition method for the analysis of dithiocarbamate salts and no alternative method was available that allowed the rapid separation and determination of these salts commonly used as fungicides. We investigated and reported the separation of dithiocarba-mate salts using a mobile phase containing hexadecyltrimethyl-ammonium bromide (CTAB) (above its critical micellar concentration)7. The optimisation of chromatographic separa-tions using micellar mobile phases requires an understanding of the effects of commonly used organic modifiers on micelles. The aim of this study was to investigate the effect of organic modifiers on the separation efficiency of a CTAB micellar mobile phase.Dithiocarbamate salts were chosen as test solutes because of their toxicological significance. 8 Conduc-tance measurements of micellar solutions were investigated in conjunction with chromatographic studies in an attempt to elucidate the effect of organic modifiers on micellar inter-actions. Experimental The HPLC system consisted of a Water Associates M6000A pump and a Waters Associates Lambda-max Variable UV detector used in conjunction with a Model 7125 sample injector with a 2O-pl loop (Rheodyne Berkeley CA USA). The column (250 x 5 mm i.d.) was packed with 5-ym Spherisorb ODS (HPLC Technology Macclesfield UK). A Hewlett-Packard 3390A recording integrator was used to obtain retention times.A chart recorder (J.J. Instruments, Southampton UK) was used to record the chromatograms. Both the column and mobile phase temperatures were controlled by immersion in a thermostated water-bath (Gallenkamp London UK). Reagents Dithiocarbamate salts. Supplied by J. D. Campbell & Sons, Warrington; Fluorochem Glossop; and Robinson Bros., West Bromwich UK. Hexadecyltrimethylammonium bromide (CTA B) . AnalaR grade supplied by BDH Chemicals Poole UK. Methanol and acetonitrile. HPLC grade supplied by Rath-burn Chemicals Walkerburn UK. Potassium dihydrogen orthophosphate. AnalaR grade, obtained from BDH Chemicals. Tetrabutylammonium bromide. Puriss grade obtained from Fluorochem. Procedure The CTAB adsorbed was determined using the elution method described by Hung and Taylorg; this involves desorb-ing the surfactant with methanol from a previously equilib-rated column and weighing the total amount recovered.The breakthrough methodlo was also used. Chromatographic measurements were carried out in a standard phosphate buffer of pH 6.8. The flow-rate was as indicated and the wavelength of absorption used for measure-ment of the dithiocarbamate salts was 286 nm and for phenol and benzene 254 nm. The micellar mobile phases were prepared by dissolving the appropriate amount of surfactant in buffered water - organic modifier and filtering through a 0.22 ym filter (Duropore Millipore). Stock solutions of the test solutes were prepared freshly each day in water.The peak shapes obtained in ion-pair and micellar chromatography are often non-Gaussian which may lead to overestimation of the plate count by as much as 100% if the more common Gaussian based equations are employed. For this study it was decide 1274 ANALYST NOVEMBER 1986 VOL. 111 to use an equation that would include an asymmetry factor. The equation described by Foley and Dorseyll was used throughout this study for theoretical plate count calculations. Peak widths were measured manually. Plate counts were determined using the equation11 (tR/W0.1)2 N = 41.7 X (HA) + 1.25 where tR is the retention time BIA is the asymmetry factor and Wo,l is the peak width at 10% of the peak height. Conductance measurements were performed with a digital conductimeter (Orion Research Model 101) using platinum electrodes.Solutions were thermostated at 25 “C in a water-bath and degassed ultrasonically before conductance measurements were made. Results and Discussion Charged surfactants have been widely used as mobile phase modifiers to improve the partitioning characteristics of charged solutes in reversed-phase HPLC. Several retention mechanisms have been proposed and discussed previously.7 Quaternary ammonium salts containing one long hydrophobic alkyl chain are called amphiphiles. These amphiphiles e.g., hexadecyltrimethylammonium bromide (CTAB) can form micelles in polar solutions i.e. the ions interact to form discrete aggregates possessing a hydrophobic core and a polar surface. Quaternary ammonium salts such as tetrabutylammo-nium bromide are not amphiphiles and cannot form micelles.A micellar mobile phase differs from a conventional ion-pairing mobile phase in two important respects. Firstly, micellar solutions can be regarded as microscopically hetero-geneous being composed of the micellar aggregate and the “bulk” surrounding medium. An ion-pairing mobile phase is homogeneous. Secondly the concentration of surfactant in micellar chromatography is above its critical micellar concen-tration (CMC) i.e. the concentration above which micelle formation becomes appreciable. Below its CMC CTAB can be used as an ion-pairing reagent. HPLC separations performed with a micellar mobile phase have been reported previously.12-15 The effect of variation of the organic modifier concentration on the efficiency of separations obtained with a micellar mobile phase has been briefly discussed by a number of workers.Dorsey et al.I4 advised a low concentration of organic modifier to “maintain the integrity” of the micelle. Yarmchuk et al.15 concluded that the small gains in efficiency were not worth the incorporation of organic solvents in micellar eluents. Most workersl4J5 have used neutral hydrophobic test solutes which are known to interact with the hydrophobic core of the micelle in their efficiency studies. Yarmchuk et ~ 1 . 1 5 discussed the restricted mass transfer of hydrophobic solutes in micellar chromato-graphy in terms of the effect of entrance - exit rate constants of phenol and benzene with micelles. Almgren et ~1.16 discussed the dynamic and static aspects of the solubilisation of neutral arenes in ionic micellar system.They deduced that the exit rates of solutes from micelles approximately parallel the solubility of the solute in water i.e. the greater the solubility of the solute in water the faster is the exit rate from a particular micelle. We propose from the deductions of Almgren et al. 16 that the interaction of the dithiocarbamate salts with the CTAB micelles must be very rapid because of their high water solubility e.g. sodium N-methyldithiocarbamate has a water solubility of 722 g 1-1 at 20 OC.17 Tagashiralg discussed the interaction of dithiocarbamate salts with micelles and pro-posed that they interacted with the polar “mantle” of the micelle. Benzene and phenol are expected to interact with the hydrophobic core of the micelle.15 With this in mind we decided to study both the hydrophobic solutes (benzene and phenol) and the hydrophilic dithiocarbamates in our investiga-tion of the effects of organic modifiers on CTAB micellar separation efficiency.Fig. 1 shows the variation in CTAB adsorbed as the concentration of methanol and acetonitrile organic modifiers employed was increased. The results show that as the concentration of organic modifier is increased the concentra-tion of CTAB adsorbed decreases. The loading achieved with methanol was greater than that achieved with acetonitrile. These plots may be used to predict the modifier concentration necessary to achieve a particular surface adsorption. For a 1 x 10-2 M CTAB solution containing methanol -water (90 + 10 V/V) buffered to pH 6.8 (10 mM phosphate), 0.29 mM of CTAB is adsorbed per gram of adsorbent.The support surface area per molecule is only 0.57 Biz. As the molecular volume of the CTAB cation is about 500 A3 the degree of surface coverage is very high. This high surface coverage indicates that the CTAB chain is perpendicular to the surface and that the surface coating by methanol must be low (Fig. 1). As the concentration of methanol in the eluent is increased the methanol loading on the surface is increased, I I 1 10 30 50 Organic modifier YO Fig. 1. Variation of CTAB adsorbed on the Spherisorb ODS stationary phase with percentage of organic modifier in the mobile phase. A Methanol; and B acetonitrile 7.0 -5.0 F > c m a m 2 5 3.0 1 .o .\ 0 10 30 50 Acetonitrile “10 Fig.2. Variation of square root of capacity ratio with percentage of acetonitrile. Flow-rate 1.0 mlmin-1; UVdetectionat 254and286 nm; 25 “C. A Sodium N,N-diethyldithiocarbamate; B ammonium tetra-rnethylenedithiocarbamate; C sodium N,N-dimethyldithiocarba-mate;D,sodiumN-methyldithiocarbamate;E,phenol; andF benzen ANALYST NOVEMBER 1986 VOL. 111 resulting in a lower degree of surface coverage by CTAB. The use of acetonitrile as the organic modifier results in lower surface coverage by CTAB than with the corresponding percentage of methanol. Hung and Taylor9 discussed the surface loading of CTAB on Hypersil ODs. They reported a reduction in surface loading on addition of acetonitrile as the organic modifier.Unfortunately they did not discuss the possible effects on surface adsorption of using methanol as the mobile phase modifier. The effect of variation of the acetonitrile concentration in the mobile phase is illustrated in Fig. 2. A decrease in the capacity ratio is observed with a corresponding increase in organic modifier (acetonitrile) concentration for both the hydrophobic and the hydrophilic solutes. The elution order for the dithiocarbamate salts remains constant as does the elution order for phenol and benzene. However the over-all elution order of the charged and uncharged solutes changes. This change is probably due to a lower surface coverage of CTAB by a mobile phase containing high concentrations of acetonitrile (Fig.1) and to an alteration of the micelle structure in the presence of relatively high concentrations of acetonitrile. This will change the retention and efficiency characteristics of the dithiocarbamates. If micelles are disrup-ted and no longer involved in the chromatographic separation, hydrophobic solutes such as phenol and benzene return to the more conventional reversed-phase mode of separation. The effect of variation of the methanol concentration in the mobile 2.0, 20 40 60 Methanol O/" Fig. 3. Variation of log (capacity ratio) with percentage of methanol. Flow-rate 1.0 ml min-1; UV detection at 286 nm; 25 "C. A Sodium N,N-diethyldithiocarbamate; B ammonium tetramethylenedithio-carbamate; C sodium N,N-dimethyldithiocarbamate; D sodium N-rnethyldithiocarbamate; E phenol; and F benzene 1275 phase is illustrated in Fig.3. The over-all elution order for the five solutes both hydrophobic phenol and benzene and hydrophilic dithiocarbamate salts remains constant. This indicates that there is no change in the mechanism of interaction as the methanol concentration increases. The variation of the capacity ratio ( k ' ) with log[CTAB] is shown in Fig. 4. A buffered (10 mM phosphate) aqueous solution containing methanol - water (55 + 45 V/V) was chosen for this study as it separated the solutes in a short time. With no CTAB present all the dithiocarbamate salts are unretained. The retention of the dithiocarbamate anions increases with a constant order of elution in each instance up to about 1 X 10-2 M CTAB and then decreases.Benzene appears to be unaffected by the increase in CTAB concentra-tion but the capacity ratio of the more polar phenol also increases slightly to approximately 1 X 10-2 M and then decreases. Emerson and Holtzer'g reported a method for determining CMCs using conductivity measurements in methanol - water mixtures containing surfactant. The CMC for CTAB was calculated using the break in the graph of specific conductivity versus concentration of CTAB in 55 + 45 V/V methanol - water. The break occurred at 1 X 10-2.20 We propose that in 55 + 45 V/V methanol - water mixture benzene does not interact with the micelles and the capacity ratio remains constant over the range of CTAB concentrations chosen. Phenol does appear to interact with the micelles but to a lesser extent than the dithiocarbamate anions.These anions show increased retention from the more hydrophilic 9 7 s E 0 .- 4-4- ,5 .-0 Q crr u 3 1 -3.3 -2.7 -2.0 -1.4 Log ([CTABIh) Fig. 4. Variation of capacity ratio ( k ' ) with log([cTAB]/~). Flow-rate 1 ml min-*; UV detection at 286 nm; 25 "C. A Sodium N,N-diethyldithiocarbamate; B ammonium tetramethylenedithio-carbamate; C sodium N,N-dimethyldithiocarbamate; D sodium N-methyldithiocarbamate; E phenol; and F benzene Table 1. Variation of plate height (mm) with concentration of CTAB. Conditions as in Fig. 4 Concentration of CTAB/M Analyte 1 x 10-3 5 x 10-3 1 x 10-2* 3.16 x 10-2 Sodium N-methyldithio-carbamate . . . . . . . . 0.40 0.23 0.07 0.19 Sodium-N N-dimethyldi-thiocarbamate .. . . . . 0.31 0.27 0.15 0.20 Sodium N N-diethyldithio-carbamate . . . . . . . . 0.25 0.30 0.15 0.21 Ammonium tetrarnethylene-dithiocarbamate . . . . . . 0.21 0.18 0.11 0.15 Phenol . . . . . . . . . . 0.22 0.23 0.27 0.17 Benzene . . . . . . . . . . 0.14 0.13 0.11 0.10 * CMC of CTAB in 55 + 45 WV methanol - wate 1276 ANALYST NOVEMBER 1986 VOL. 111 monomethyl derivative to the more hydrophobic diethyl derivative. Table 1 shows the variation in the plate height (mm) with concentration of CTAB. The dithiocarbamates all show minimum plate heights at the CMC whereas phenol and benzene do not seem to show the same dependence on surfactant concentration. The existence of micelles in a solution containing a substantial concentration of organic modifier e .g . methanol is still a debatable topic in micellar chromatography. We believe that the results shown in Fig. 4 indicate the existence of some type of “micelle” with which the anions are interacting. A more quantitative approach to the question of the existence of micelles in the 55 + 45 V/V methanol - water solution may be found in the work of Knox and Laird.10 From their deductions we can assume that with dilute solutions of CTAB the CTAB and the dithiocarbamate anions will exist in the ionised form in the mobile phase, whereas CTA+ - (dithiocarbamate)- ion pairs are adsorbed on the hydrophobic stationary phase surface. It may also be expected that the surface of the reversed-phase packing will have a monolayer of adsorbed CTAB cations at the mobile phase - stationary phase interface.The over-all situation for the dithiocarbamate salts may be represented by the interlink-ing equilibria shown in Scheme 1. The subscripts ads. and aq. CTA+*, + dTC-*,,. ’ Kaq [CTA+dTC-I,, Kads dTC-,ds ’- ICTA’dTC - ] a d s Scheme 1 refer to the interface and eluent phases respectively. The species with the asterisks are thought to be present in high relative concentrations at low CTAB concentrations and dTC- refers to the dithiocarbamate anion. Fig. 5 shows that CTAB is indeed adsorbed by the hydrophobic stationary phase from a buffered mobile phase containing 55 + 45 V/V methanol - water with CTAB. The adsorption was measured using the breakthrough method. 10 The adsorption isotherm is significantly curved and obeys a simple Freundlich-type equation: It is assumed that the sorption of the dTC- ion on the surface is linear with respect to the concentration in the mobile phase.This is reasonable as the concentration of dTC- is low. In addition to equation (l) the following equilibria must be considered P dTC-ads (dTC-)aq 0.5 0.1 c I 0.01 0.03 0.05 [CTABIIM Fig. 5. Variation of CTAB adsorbed on the stationary phase with concentration of CTAB in the mobile phase. Flow-rate 1 ml min-l; 25 “C where P is the equilibrium constant for adsorption of dithiocarbamate from the mobile phase on to the cationic surface Kaq is the equilibrium constant for ion-pair formation in the mobile phase and Kads. is the equilibrium constant for ion-pair formation on the stationary phase surface.If the concentration of dTC-,ds is assumed to be negligible compared with that of the dTC pair it has been shown10 that the distribution coefficient D for the dithiocarbamate anion between the stationary phase and the mobile phase is . . . . (2) aPKacis.[CTA+aq.lo.75 1 + Kaq.[CTA+aq.l D = This expression indicates that the capacity ratio ( k ’ ) which is proportional to D will increase as [CTA+aq,] increases. From Fig. 5 the exponent of [CTA+,,.] was found to be 0.75 which is in good agreement with the gradients found for sodium N N-dimethyldithiocarbamate (0.64) and sodium N-methyldi-thiocarbamate (0.68). The gradients for the more hydro-phobic dithiocarbamates ammonium tetramethylenedithio-carbamate and sodium N,N-dimethyldithiocarbamate were 0.65 and 0.64 respectively.These gradients agree fairly well with the exponent of 0.75 expected for ion-pair formation at this low concentration of CTAB. Equation (2) indicates that if Kaq or (CTAt,,.] is very large (virtually all the dithiocarba-mate ions in the mobile phase are in the form of ion pairs) then the largest negative gradient should be 0.25. The negative gradients for sodium N-methyldithiocarbamate and sodium N,N-dimethyldithiocarbamate were 0.18 and 0.26 respec-tively. These are close to the negative gradient of 0.25 expected for ion-pair formation. The negative gradients obtained for both ammonium tetramethylenedithiocarbamate and sodium N,N-diethyldithiocarbamate were 0.35 and 0.50, respectively. A likely explanation for this large discrepancy in negative gradient is the formation of micelles as discussed by Knox and Laird.10 The equilibrium in the micellar mobile phase must be revaluated as K’aq.rnCTA+,,. + dTC- (rnCTA+ dTC-),,. where rnCTA+ form. Equationa?2) then becomes is the number of CTA+ ions in micellar aPKads.[CTA+]0.75 D = . . . . 1 + K’,,,[CTA+]m (3) where Klaq is the equilibrium constant for the micellar interaction. The negative gradients obtained for sodium N-methyldithiocarbamate and sodium N,N-dimethyl-dithiocarbamate are explained if rn = 1. This means that each dithiocarbamate group is associated with one CTA+ cation in the expected ion-pairing type interaction. With the more hydrophilic dithiocarbamates the interaction is purely elec-trostatic.The negative gradients for ammonium tetramethyl-r 0.6 0 cz v) 10 30 50 70 Methanol o/o Fig. 6. Variation of specific conductivity with percentage of methanol for (A) TBAB (9.3 x 10-3 M) and (B) CTAB (1 x 10-2 M ) . 25 “ ANALYST NOVEMBER 1986 VOL. 111 c I 6 0.8 cn E > .? 0.6 -. c .-c 3 U S 8 0.4 .-'c 0 R 1277 ---ene dithiocarbamate and sodium diethyldithiocarbamate are explained if rn = 1.25. This agrees fairly well with the results obtained by Knox and Laird10 for the interaction of monosul-phonic acids with micellar aggregates. From the present data it appears that each of the more hydrophobic dithiocarbamate salts i. e. ammonium tetramethylenedithiocarbamate and sodium diethyldithiocarbamate associates with about 1.2 CTAB cations when the concentration of CTAB is greater than its CMC.This greater extent of interaction may be attributed to the partitioning of the more hydrophobic dithiocarbamate into a micelle with the hydrophobic portion solubilised in the micellar core and the hydrophilic portion interacting with the polar positively charged mantle. 286 nm 254 nm I.' a.u.f.s. 6 i 2 1 3 I I ' I 16 8 0 Tirneimin Fig. 7. CTAB micellar chromatogram of dithiocarbamate salts, benzene and phenol with methanol in the mobile phase as the organic modifier. Packing Spherisorb ODS modified silica; column 250 x 5 mm i d . ; particle size 5 ym; solvent methanol - water (55 + 45 V/V); 0.01 M phosphate buffer pH 6.8; UV detection at 286 nm; flow-rate 1 ml min-I.Solutes 1 = benzene; 2 = phenol; 3 = sodium N-methyldithiocarbamate; 4 = sodium N N-dimethyldithio-carbamate; 5 = ammonium tetramethylenedithiocarbamate; and 6 = sodium N N-diethyldithiocarbamate Table 2. Variation of efficiency and asymmetry with concentration of methanol. Spherisorb ODS column (250 x 5 mm i.d. 5 pm particle size) 1 x M CTAB buffered to pH 6.8,lO mM phosphate buffer, 25 "C; flow-rate 1.5 ml min- l . The solute used for these experiments was ammonium tetramethylenedithiocarbamate Methanol % K' N* BIA 1-30 31.6 5459 2.1 40 22.4 4212 1.93 50 10 521 1 1.75 60 3.5 3468 1.87 70 2.5 4723 1.70 * N = Number of theoretical plates. t BIA = Asymmetry ratio of eluted peaks. The results obtained from this alternative approach evidently suggest that the dithiocarbamates are indeed inter-acting with micelles in solution.Finally we decided to investigate the effects of methanol and acetonitrile on micellar properties using conductance measurements. These measure-ments indicate the degree of ionisation of the micelles and hence their surface charge density. The effects of primary straight-chain alcohols (ethanol to hexanol) on micellar properties were studied by Zana et al.21 using conductance measurements. Conductance measurements have been used effectively by various worker+-26 to study the effects of additives on surfactants. In order to study the effect of variation in organic modifier concentration on micelle stabil-ity specific conductivity measurements were obtained at four B a.u.f.s.LJ 8 6 4 2 0 Timeimin Fig. 8. TBAB ion-pair chromatogram of dithiocarbamate salts using methanol as the mobile phase modifier. Conditions as in Fig. 7. Solutes as in Fig. 3. Flow-rate 1 ml min-I; 25 "C; UV detection at 286 nm; 0.01 M phosphate buffer pH 6.8; solvent methanol - water (55 + 45 v/V) I Y v) I I 1 I 10 30 50 70 Aceto n it ri I e "10 Fig. 9. Variation of s ecific conductivity with percentage of aceto-nitrile for (A) TBAB (5.3 X 10-3 M) and (B) CTAB (1 X lop2 M). 25 "C Table 3. Variation of theoretical plate number N and resolution R, with variation in methanol concentration. Conditions as in Table 2 Sodium N-methyldithio- Sodium N N-dimethyldithio- Ammonium tetramethylenedi-Methanol % carbamate carbamate thiocarbamate k' N R S k' N R k' N Rs 30 13.2 2128 6.45 19.9 3025 9.34 31.6 5459 -50 5.4 3449 4.0 6.7 5352 8.20 10.0 5211 -70 0.03 3528 1.16 0.5 3595 2.20 2.5 4723 1278 different modifier concentrations between 10 and 70% V/V.Fig. 6 illustrates the variation in specific conductivity with change in methanol concentration for CTAB. Increasing the methanol concentration results in a corresponding increase in the specific conductivity. This increase may be attributed to partial solubilisation of methanol by the micelles. The first effect of solubilisation of alcohol molecules is steric the alcohol molecules become intercalated between the surfactant ions and increase the average distance between ionic head groups. This results in a decreased micelle surface charge density which in turn results in increased ionisation.It is this increased ionisation that appears as the increase in specific conductivity with increasing methanol concentration in Fig. 6. Tetrabutylammonium bromide (TBAB) was also studied because although it is a cationic ion-pairing reagent it cannot form micelles. Fig. 6. shows an initial decrease in specific conductivity for TBAB with increasing concentration of methanol. This effect may be attributed to the high free energies of transfer of both the tetrabutylammonium cation and the bromide anion from water to methanol.27 As the methanol concentration increases both ions become strongly solvated causing the specific conductivity to plateau. We believe that the micelle structure is not altered substantially by the addition of methanol.Fig. 7 illustrates the Separation of the dithiocarbamates in a buffered 55 + 45 V/V methanol - water solution. Table 2 shows that the efficiency of the mobile phase does not decrease significantly with increase in the concentration of methanol as illustrated by ammonium tetramethylenedithiocarbamate. Table 3 shows the variation of the theoretical plate number N and resolution R, for three dithiocarbamate salts with variation in methanol con-centration. The resolution remains fairly high and the theoretical plate number is both high and fairly constant over 3 4 0.01 a.u.f.s. I 1 I _ 1 2 8 4 0 Timeimin Fig. 10. CTAB micellar chromatogram of dithiocarbamate salts using acetonitrile as the mobile phase modifier.Conditions and solutes as in Fig. 7 ANALYST NOVEMBER 1986 VOL. 111 the range of methanol concentrations studied. Fig. 8 shows the separation of four dithiocarbamate salts obtained with TBAB under the same conditions used with the CTAB. Both CTAB and TBAB contain the same number of carbon atoms so if the micellar properties associated with CTAB were destroyed by a high concentration of methanol the separation achieved using TBAB should be similar to that achieved using CTAB. The fact that this is not so points to the survival of the micelles. The results obtained for the variation in specific conductiv-ity versus concentration of acetonitrile shown in Fig. 9 for a micellar solution indicate a large initial increase in specific conductivity for acetonitrile concentrations in the range 10-30%.This increase may be attributed to the initial penetration of the micelles causing increased ionisation , similar to the effect of methanol. The greater increase in specific conductivity for acetonitrile almost double that for methanol could be attributed to total micelle disruption at higher concentrations of acetonitrile. The free energies of transfer for tetraalkylammoniurn cations are greater in aceto-nitrile than in methanol.27 This effect may be sufficient to cause micelle breakdown accounting for the dramatic increase in specific conductivity. Fig. 9 also shows that there is a slight increase in specific conductivity for TBAB with increased acetonitrile concentration again attributable to solvation of both the anion and cation by acetonitrile.Fig. 10 shows a separation of dithiocarbamate salts with CTAB in a buffered (50 + 50 V/V acetonitrile - water mobile phase. The separation achieved is very poor and the peaks are badly tailed. Table 4 shows the variation in capacity ratio ( k ’ ) and efficiency with increase in acetonitrile concentration. As the acetonitrile concentration increases the efficiency decreases until at 50 + 50 V/V acetonitrile - water the efficiency obtained is very poor compared with that obtained for methanol at the same concentration. We conclude that if the special effects attributed to a micellar mobile phase are required methanol should be selected as the mobile phase modifier. If these effects are not desired acetonitrile should be chosen.The CMC should be evaluated to ensure that the concentration of the surfactant is sufficiently high to result in the formation of micelles. Our results indicate that a separation should be carried out at the CMC rather than above it or below it for optimum efficiency. We advise using a micellar mobile phase to separate charged organics; there seems to be no advantage in using micellar mobile phases to separate hydrophobic organics. The separation efficiency is high for dithiocarbamate salts in a CTAB micellar solution with methanol as the modifier. Our results indicate that the separation mechanism depends on the interaction with the micelles and not an “ion-pairing” inter-action. We are grateful to Thames Water Authority for full support and partial funding for this project.Table 4. Variation ODS column (250 phosphate) 25 “C, of capacity ratio ( k ’ ) and efficiency (number of theoretical plates N) with concentration of acetonitrile. Spherisorb X 5 mm i.d.) 5 pm particle size water - acetonitrile as indicated 1 x M CTAB buffered to pH 6.8 (10 mM flow-rate 1.5 ml min-1 UV detection at 286 nm Ammonium Sodium N-methyldi- Sodium N N-dimethyl- tetramethylene-Acetonitrile ,% Benzene Phenol thiocarbamate dithiocarbamate dithiocarbamate k’ N k’ N k‘ N k‘ N k’ N - - 10 15.0 5436 16.6 3903 23.8 793 41.2 3281 30 4.0 2740 5.1 2539 5.1 321 6.9 107 11.9 464 50 0.8 228 1.06 124 0.7 90 0.9 23 1.3 8 ANALYST NOVEMBER 1986 VOL. 111 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. References Eksborg S.and Schill G. Anal. Chem. 1973 45,2092. Persson B.-A, Acta Pharm. Suecica 1968 5 343. Schill G. Modin R. and Persson B.-A. Acta Pharm. Suecica 1965 2 119. Karger B. L. Su S. C. Marchese S . and Persson B.-A., J. Chromatogr. Sci. 1974 12 678. Kirkbright G. F . and Mullins F. G. P. Anal. Chim. Acta, 1984 156 279. Smith R. M. Morarji R. L. Salt W. G. andstretton R. J., Analyst 1980 105 184. Kirkbright G. F. and Mullins F. G. P. Analyst 1984 109, 493. Newsome W. H. in Zweig G. and Sherma J. Editors, “Analytical Methods for Pesticides and Plant Growth Regu-lators,” Volume 3 Academic Press New York 1964 p. 197. Hung C. T. and Taylor R. B. J . Chromatogr. 1981 209, 175. Knox J. H. and Laird G. R. J . Chromatogr. 1976 122 17. Foley J.P. and Dorsey J. G. Anal. Chem. 1983 55 730. Yarmchuk P. Weinberger R. Hirsch R. F. and Cline Love, L. J. Anal. Chem. 1982,54,2233. Armstrong D. W. and Nome F. Anal. Chem. 1981 53, 1662. Dorsey J. G. DeEchegaray M. T. and Landy J. S . Anal. Chem. 1983 55 924. Yarmchuk P. Weinberger R. Hirsch R. F. and Cline Love, L. J. J . Chromatogr. 1984 283 47. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 1279 Almgren M. Grieser F. and Thomas J. K. J. Am. Chem. SOC. 1979 101 279. Ottnad M. Jenny N. A. and Roder C. H. in Zweig G. and Sherma J. Editors “Analytical Methods for Pesticides and Plant Growth Regulators,” Volume 10 Academic Press New York 1978 p. 563. Tagashira S . Anal. Chem. 1983 55 1918. Emerson M. F. and Holtzer A. J. Phys. Chem. 1967 71, 3320. Mullins F. G. P. and Kirkbright G. F. unpublished work. Zana R. Yiv S. Strazielle C. and Llanos P. J. Colloid Interface Sci. 1981 80 208. Singh H. N. Singh S . and Mahalwar D. S . J. Colloid Interface Sci. 1977 59 386. Tominaga T. Stem T. B. and Evans I. F. Bull. Chem. SOC. J p n . 1980 53 795. Robins D. C. and Thomas I. L. J . Colloid Interface Sci., 1968 26,407. Manabe M. and Koda M. Bull. Chem. SOC. J p n . 1978 51, 1599. Abu-Hamdiyyah M. and El-Danab C. M. J. Phys. Chem., 1983 87 5443. Karger B. L. LePage J. N. and Tanaka N. in Horvath C., Editor “High Performance Liquid Chromotography ,” Volume 1 Academic Press New York 1980 p. 126. Paper A6188 Received March 17th 1986 Accepted June 19th 198

ISSN:0003-2654    

DOI:10.1039/AN9861101273

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, NOVEMBER 1986, VOL. 111 1281 High-performance Liquid Chromatography with Anodic Amperometric Detection for the Determination of Cefotaxime and its Metabolites Huguette Fabre and Marie Dominique Blanchin Laboratoire de Chimie Analytique, Faculte de Pharmacie, 34060 Montpellier Cedex, France and Ubbo Tjaden Division of Analytical Chemistry, Gorlaeus Laboratories, Centre for Bio-Pharmaceutical Sciences, PO Box 9502,2300 R A Leiden, The Netherlands The possibility of using liquid chromatography with anodic amperometric detection has been investigated for the determination of cefotaxime and its two main decomposition products. The electrochemical behaviour of these compounds was monitored using a glassy carbon electrode. An optimum signal to noise ratio was obtained at a +0.95 V potential (vs.SCE). At this potential, the detection limit was found to be in the picomole range and was equivalent to or better than UV detection. Linearity and repeatability results showed that anodic amperometric detection could be used as an alternative detection mode for stability studies. Keywords: Cefotaxime determination; cefotaxime metabolites; high-performance liquid chromatography; anodic amperometric detection The study reported here is part of an investigation to develop sensitive determination methods for cephalosporins and their decomposition products in pharmaceutical formulations and biological fluids. 1-3 As high-performance liquid chromato- graphy (HPLC) with electrochemical detection has, in most instances, proved to be more sensitive4 and selective than UV absorption detection, the detection of cephalosporins by amperometry was investigated.Cathodic reduction of ceph- alosporins has been widely reported (references 5-7 and references cited therein) but the application of an HPLC method using reductive amperometric detection cannot be carried out easily because of oxygen disturbances.8 There are no references in the literature dealing with the anodic oxidation of these compounds (except for cefadroxil,5 a compound with a phenolic hydroxy substituent on the side-chain in position 7 of the A3-cephem ring) although such electroactivity can be assumed for most cephalosporins. Preliminary studies showed that most cephalosporins are electrochemically active at a glassy carbon electrode using a +1.10 V potential (vs.SCE). In this paper, the results obtained for cefotaxime (I) and two major metabolites, deacetylcefotaxime (11) and deacetylcefotaxime lactone (111) are given. Cefotaxime was chosen because it is the prototype of the third generation cephalosporins and is used as a reference in testing microbiological activity. COO -Na+ Cefotaxime (I) COO-Na4 Deacetylcefotaxime (11) Deacetylcefotaxime lactone (111) Experimental Apparatus The liquid chromatograph consisted of a Constametric pump (LDC, Riviera Beach, FL, USA) equipped with an U6K injector (Waters Associates, Milford, MA, USA), a fixed wavelength (254 nm) UV absorbance detector (Model 440, Waters Associates) coupled in series with a laboratory-made amperometric cell unit and a potentiostat (E 230, Bruker, Karlsruhe, FRG), which was connected to an a.c.voltage stabiliser (Sorrensen, Zurich, Switzerland). The ampero- metric detector was a wall-jet type detector with a glassy carbon working electrode, a stainless-steel auxiliary electrode and a saturated calomel reference electrode. For UV detec- tion, a sensitivity of 0.02 a.u.f.s. was used, and for the amperometric detector the sensitivity range was varied from 20 to 200 nA full-scale. The time constant of the potentiostat was set at 2.2 s with a gain of 1. The analytical column was a slurry-packed MOS-Hypersil column (5 pm, 10 cm x 0.3 cm i.d.). The flow-rate was 0.8 ml min-1 and the pressure was about 110 bar. The column and the mobile phase were thermostated at 25 _+ 0.1 "C and the amperometric cell was used at ambient temperature. The chromatograms were recorded on two separate BD 8 iecorders (Kipp and Zonen, Delft, the Netherlands) at a speed of 5 mm min-1.Reagents and Materials Cefotaxime sodium, deacetylcefotaxime sodium and deacetyl- cefotaxime lactone were standards from Roussel UCLAF Laboratories (Romainville, France) and were used as received. All other chemicals were of analytical-reagent grade. Water was purified by a Milli-Q water purification system (Millipore, Bedford, MA, USA).1282 Solutions Stock solutions of cefotaxime (50 mg 1-1) and deacetylcefo- taxime (50 mg 1-1) were prepared in water. A stock solution of deacetylcefotaxime lactone (50 mg 1-1) was prepared in the mobile phase. Because of its instability, this solution was prepared twice daily and kept at 4 "C before dilution.The three stock solutions were suitably diluted in water just before use to give a mixed standard solution of the desired concentration. Aliquots of 10 1.11 of this mixed standard solution were injected in triplicate on to the chromatograph. ANALYST, NOVEMBER 1986, VOL. 111 Mobile phase The mobile phase was McIlvaine buffer, pH 7.65, diluted in water (1 + 3) - methanol (93 + 7). The buffer solution was 0.0011 M citric acid and 0.031 M disodium phosphate with an ionic strength of 0.083. The mobile phase was de-gassed by sonication for 10 min and was magnetically stirred during the analysis. No special precautions were taken in order to decrease the background signal, except that during the night the mobile phase was recycled.Electrochemical Detection Procedure Before the experiments, the glassy carbon electrode was polished to a mirror finish with a 1-vm diamond paste. The working electrode was then electrochemically pre-treated (eight cycles of 5 min each from -1 to +1 V) in the mobile phase. For the construction of the voltammograms, the appro- priate electrode potentials were applied and the samples were injected after base-line stability was achieved at a low detector sensitivity. This operation required about 2 h when highly positive potentials (over + 1.1 V) were applied. The noise was evaluated after stabilisation over a period of at least 5 min and is given as a top-to-top value. Results and Discussion Representative chromatograms of a mixed standard solution are shown in Fig.l(a) (UV detection) and Fig. l(b) (amperometric detection). The capacity ratios determined from the retention time of each compound and the retention time of a 20 mM potassium iodide solution were k' = 3.13 for 11, 17.6 for I11 and 30.6 for I. Electrochemical Characteristics The voltammetric characteristics of each compound were determined using the described chromatographic conditions in the potential range 0.5-1.2 V. At each potential, the peak height of both the UV and the amperometric detector signal was measured. The peak heights were converted to the response as obtained using a 0.005 a.u.f.s. deflection (for UV detection) and a 5 nA full-scale current range (for amperometric detection), respectively. The ratio of the amperometric response to the UV absorption response was plotted for each compound (Fig.2). For deacetylcefotaxime, no characteristic voltammogram was obtained, whereas for cefotaxime a limiting current was reached at about +1 V. For deacetylcefotaxime lactone, two limiting currents (+0.95 and +1.05 V) can be noted. The decrease in the voltammogram observed for I and I11 at potential values over +1.10 V may be related to the electroactivity of the mobile phase, which can compete with the compounds of interest. Preliminary studies carried out on the series of cephalo- sporins and on other decomposition products of cefotaxime (thiazoximic acid, deacetoxycefotaxime and 7-aminocephalo- sporanic acid) seem to indicate that the aminothiazole substituent on the side chain in position 7 of the A3-cephem 0.002 I AU c a, S .- 3 c a, C .- - J.T Fig. 1. Specimen chromatograms obtained using (a) UV detection and (b) amperometric detection. Cefotaxime, 112 ng; deacetylcefo- taxime, 108 ng; deacetylcefotaxime lactone, 100 ng. Chart paper speed, 5 mm min-1. (a) h = 254 nm, 0.02 a.u.f.s; and ( b ) + 0.95 V, 200 nA full-scale C f B q0.5 +1.0 PotentialiV Fig. 2. Voltammetric characteristics of cefotaxime and its meta- bolites. A, Cefotaxime, 112 ng; B, deacetylcefotaxime, 84 ng; and C, deacetylcefotaxime lactone, 108 ng ring is the electroactive group that undergoes anodic oxida- tion. The ease of oxidation is compatible with the oxidation of an aromatic amine. Signal to Noise Ratio Fig. 3 shows that up to +0.95 V the noise is relatively low and constant. The low noise observed up to this potential may be related to the low percentage of methanol in the mobile phase as methanol is electrochemically oxidised at higher rates than water.8 Moreover, recycling the mobile phase during the night contributes to lower noise levels.A dramatic increase in the noise level is observed at potentials higher than +0.95 V owing to the electrochemical oxidation of the mobile phase. In Fig. 4 the signal to noise ratio vs. potential is given for compounds I and 11; for the lactone compound the plot was scattered (because of the presence of a plateau in the voltammogram) and is not reported.ANALYST, NOVEMBER 1986, VOL. 111 10 8 - 6 - P . v) 0 Z .- 4 - 2 - 1283 - For all three compounds, an optimum signal to noise ratio was obtained at +0.95 V.Therefore, this potential value was used in the subsequent experiments. The stabilisation of the detector at this potential value can be achieved within 1 h. +0.6 +0.8 +1.0 +1.2 PotentialiV Fig. 3. Noise vs. potential at the glassy carbon electrode. Mobile phase, McIlvaine buffer (pH 7.65) - CH,OH (93 + 7). Flow-rate = 0.8 ml min-1 Q) 0 .- ; 200 Y - m C 0, i7, 100 +1.0 PotentialiV 0 Fig. 4. and B, deacetylcefotaxime (84 ng) Signal to noise ratio vs. potential for A, cefotaxime (112 ng) Linearity The linearity of the response was checked for each compound within the range 10-100 ng. The response was measured by peak height and peak area. The parameters of the linear regression analysis and the correlation coefficients were calculated and the results were compared with those obtained with UV detection (Table 1).The correlation is in all instances highly significant and allows the use of peak height and peak area measurements for the determination of these com- pounds. The confidence limits ( P = 95%) for the intercept showed that all the calibration graphs passed through the origin. Sensitivity The sensitivity, defined as the change in the measured area (AU s for UV detection and nA s for amperometric detection) or in the peak height (AU for UV detection and nA for amperometric detection) resulting from a change of one unit (ng) of an injected compound was calculated for each compound (Table 1). As can be seen from peak area measurements, the response of the compounds using ampero- metric detection follows the sequence I < I1 < 111.Detection Limit The detection limit, defined as the mass of compound that yields a signal to noise ratio of two, was calculated for each compound (Table 1). The detection limit using amperometric detection is equivalent to (for I) or better than (for I1 and 111) that with UV detection. The limit of determination can be evaluated to about three times the detection limit. The detection limit in electrochemical detection (in the picomole range) shows that cefotaxime and its two main metabolites can easily be detected at a glassy carbon elec- trode. This detection limit could be lowered by taking special precautions such as the electrical shielding of the equipment and the use of a damping system. The detection limit in UV is, however, of the same order for several reasons: firstly, cefotaxime and its metabolites are strongly UV absorbing compounds and their absorbance at 254 nm is about 90% of the maximum absorbance (238 nm under the chromatographic conditions used); and secondly, a fixed wavelength detector is particularly favourable for evaluating the UV detection limit.Repeatability The repeatability was determined by injecting 20 times on to the chromatograph a mixed standard solution (112 ng of I, 78.4 ng of I1 and 100 ng of 111). The repeatability of the Table 1. Comparison of linearity, sensitivity and detection limits in the determination of cefotaxime and its main metabolites using UV absorption detection and amperometric detection at +0.95 V Amperometric UV detection detection Parameter I I1 I11 I I1 I11 Linearity rangehg .. 10-100 10-100 10-100 10-100 10-100 10-100 Correlation coefficient: Peakarea . . . . 0.999 0.998 0.981 0.998 0.999 0.983 Peakheight . . . . 0.999 0.997 0.984 0.998 0.999 0.989 Peakarea . . . . 16 x 10-4 19 x 10-4 20 x 16.20 21.26 27.28 Peak height . . . . 0.37 x 2.1 x 0.65 x 0.37 1.77 1.26 Detectionlimithg . . -2.3 -0.4 -1.3 -1.8 -0.4 -0.6 * AU s ng-l for UV detection by area measurements; AU ng-1 for UV detection by height measurements; nA s ng-l for amperometric Sensitivity* : . . . . detection by area measurements; nA ng-1 for amperometric detection by height measurements.1284 ANALYST, NOVEMBER 1986, VOL. 111 electrochemical response to UV response ratio was calculated using peak height measurements for each compound.The repeatabilities expressed by the coefficient of variation were 2.06% (I), 1.68% (11) and 2.00% (111). The repeatability and the linearity tests show that no particular adsorption problem occurred at the electrode during the experiments. More than 80 injections of a mixed standard solution (three compounds) were made without any noticeable change in the response. Moreover, the sensitivity of the detection allows a reduction of the amount injected to about 20 ng of each compound for routine analysis. Conclusion The data show that anodic amperometric detection can be used as an alternative detection mode in HPLC for cefotaxime and its two main metabolites. The method is very sensitive and its detection limit in the picomole range suggests its use as a stability indicating assay of cefotaxime. Further investigations will be devoted to the elucidation of reaction mechanisms, to the extension of this detection mode to other cephalosporins and to its application to biological fluids. References 1. 2. 3. 4. 5. 6. 7. 8. Fabre, H., and Hussam-Eddine, N., J. Pharm. Pharmacol., 1982, 34, 423. Fabre, H., Hussam-Eddine, N., and Berge, G., J. Pharm. Sci., 1984, 73, 611. Fabre, H., Blanchin, M. D., Lerner, D., and Mandrou, B., Analyst, 1985, 110, 775. Musch, G., De Smet, M., and Massart, D. L., J . Chromatogr., 1985, 348, 97. Ivaska, A., and Nordstrom, F., Anal. Chim. Acta, 1983, 146, 87. Sengun, F. I., Ulas, K., and Fedai, I., J. Pharm. Biomed. Anal., 1985, 3, 191. Sengun, F. I . , Gurkan, T., Fedai, I., and Sungur, S . , Analyst, 1985, 110, 1111. van der Lee, J. J., van der Lee-Rijsbergen, H. B. J . , Tjaden, U. R., and van Bennekom, W. P., Anal. Chim. Acta, 1983, 149, 29. King, W. P., Kuriakose, T. J., and Kissinger, P. T., J. Assoc. Off. Anal. Chem., 1980, 63, 137. 9. Paper A61164 Received May 23rd, 1986 Accepted June 25th, 1986

ISSN:0003-2654    

DOI:10.1039/AN9861101281

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, NOVEMBER 1986, VOL. 111 1285 Tetracyanoethylene in Pharmaceutical Analysis Part 1. A Spectrophotometric Method for the Determination of Some Pharmaceutically Important Hydrazine and Pyrazolone Derivatives F. A. Ibrahim, M. S. Rizk and F. Belal Department of Analytical Chemistry, Faculty of Pharmacy, Mansoura University, Mansoura, Egypt A simple and sensitive spectrophotometric method is described for the determination of some hydrazine and pyrazolone derivatives. The determination is based on the formation of charge-transfer complexes between tetracyanoethylene as a n-acceptor and the studied drugs as n-donors in acetonitrile solvent. The spectra, various experimental parameters and the stoicheiometry and stability of the reaction products were studied. The complexes formed were found to absorb at 395,380,410,420 and 440 nm for the complexes formed with phenelzine sulphate, isonicotinic acid hydrazide, hydralazine hydrochloride, amidopyrine and antipyrine, respectively.Beer’s law is obeyed in the concentration ranges 20-80, 1-10, 2-40, 10-60 and 30-80 pg ml-1, respectively, for the studied drugs. The proposed method is simple and can be applied to the determination of the studied drugs in their pharmaceutical dosage forms. Keywords: Hydrazine and pyrazolone derivatives; tetracyanoeth ylene; charge-transfer complexes; phar- maceutical analysis; spectrophotometry Tetracyanoethylene (TCNE) is known to yield charge-transfer complexes and radical ions with a variety of electron donors, including amines. 1-3 Merrifield and Phillips4 studied the interaction of pyridine, which may act either as an n- or n-donor, with TCNE and reported that the product of the interaction showed a doublet at 400 and 421.5 nm, which was attributed to the formation of a complex.Middleton et al.5 observed that the pentacyanopropenide ion is formed during the interaction of TCNE with aqueous pyridine and absorbs at 393 and 412 nm. TCNE has also been used for the detection by thin-layer chromatography and subsequent identification of a number of biologically important indole derivatives.610 The interaction between TCNE and indole derivatives consisted in the formation of a charge-transfer complex followed by trans- formation to the tricyanovinyl derivative. 11 A recent review of the charge-transfer complexes3 did not mention the quantitative determination of hydrazine or pyrazolone derivatives by charge-transfer complex formation, except for the interaction of pyrazolone derivatives with 2,5-dichlorobenzoquinone.12 In this study, the charge-transfer complex formed between TCNE and hydrazine derivatives [phenelzine sulphate, iso- nicotinic acid hydrazide (INH) and hydralazine hydro- chloride] and with the pyrazolone derivatives [ amidopyrine (aminophenazone) and antipyrine (phenazone)] was adopted in a spectrophotometric method for the assay of these drugs, either in pure form or in their pharmaceutical dosage forms. A trial for the detection of the studied drugs on TLC plates with TCNE as a chromogenic agent was also performed. The interaction products are stable and have large molar absorp- tivities.Experimental Apparatus Spectra were recorded on a Pye Unicam SP 1800 spectropho- tometer with 1-cm cells. Materials Amidopyrine, antipyrine, phenelzine sulphate, isonicotinic acid hydrazide (INH) and hydralazine hydrochloride were obtained from commercial sources and their purities were determined by the USP method.13 Reagents All chemicals were of analytical-reagent grade and the solvents were of spectroscopic grade. Tetracyanoethylene (Merck) solutions, 0.1% mlV and 8 X low4 M, were prepared in the required organic solvent. Silica gel G plates were prepared and activated according to the method described by Stahl,l4 and the solvent (mobile phase) was butanol - ammonia (6 + 1). Preparation of Sample Solutions Amidopyrine, antipyrine and INH solutions Solutions, 1 X l O - 3 ~ , of the drugs were prepared in the required solvent; 8 x 1 0 - 4 ~ solutions were prepared in acetonitrile by successive dilution from the 1 X l o - 3 ~ solutions.Phenelzine sulphate or hydralazine hydrochloride solution A 100-mg portion of the drug salt was transferred into a 250-ml separating funnel with a small volume of distilled water and made alkaline with 6~ ammonia solution. The liberated hydrazine derivative base was then extracted with five 15-ml portions of chloroform and the combined extract was dehy- drated by shaking with anhydrous sodium sulphate for 5 min. The extract was filtered into a 100-ml measuring flask, the sodium sulphate and the filter were rinsed with chloroform and the washings added to the filtrate and the solution was diluted to the mark with chloroform.For working solutions other than chloroform, a volume corresponding to the required concentration was evaporated to dryness at room temperature with a stream of nitrogen and the residue was dissolved in the desired solvent. An 8 X l o - 4 ~ solution was prepared as described above. General Procedure Accurately measured volumes of drug in acetonitrile, equi- valent to 10-60, 30-80, 20-80, 1-10 and 2-40 pg ml-1 final dilution of amidopyrine , antipyrine , phenelzine sulphate, INH and hydralazine hydrochloride, respectively, were trans- ferred into a series of 10-ml flasks. A 2-ml volume of TCNE solution was added to each flask and the solutions were mixed and diluted to volume with acetonitrile.The solutions were then either treated or left at room temperature for the1286 ANALYST, NOVEMBER 1986, VOL. 111 appropriate time (Table l), and the absorbances were measured at A,,,, against a reagent blank. The concentration of the pure drug, either alone or in its pharmaceutical dosage form, is calculated from the regression equation or from a calibration graph. Procedure for Tablets Twenty tablets were weighed and finely powdered. An accurately weighed amount of the powder, equivalent to 50 ml of the drug or drug salt, was treated in the same way as the pure drugs in the calibration procedure. Accurately measured volumes of acetonitrile solution were prepared, containing about 30,50,50,5 or 20 yg ml-1 final dilution of amidopyrine, antipyrine, phenelzine sulphate, INH or hydralazine hydro- chloride, respectively, and were transferred into 10-ml measuring flasks and treated in the same way as standards.Procedure for Injection or Syrup An accurately measured volume of the mixed contents of ten ampoules or mixed syrup, equivalent to 50 mg of the drug, was treated as described for tablets. Results and Discussion Tetracyanoethylene solution in acetonitrile gives an absorp- tion spectrum with an absorption maximum at 285 nm. On the Table 1. Data for the reaction of TCNE with some pyrazolone and hydrazine derivatives Molar Working L l I a x . 1 absorptivity1 range/ Temperature/ Time/ Molar Compound nm 1 mol-1 cm-1 pg ml-1 "C min ratio Amidopyrine . . . . . . . . 420 1.9 x 103 10-60 40 5 2 : 3 Antipyrine . . .. . . . . 440 1.55 x 103 30-80 80 5 1:1 Phenelzinesulphate . . . . . . 395 2.57 x 103 20-80 25 20 1:1 Isonicotinic acid hydrazide . . 380 6.04 x 103 1-10 25 10 2 : 3 Hydralazine hydrochloride . . 410 9.9 x 103 2-40 25 5 2 : 3 Table 2. Assay of pyrazolone and hydrazine derivatives alone and in pharmaceutical preparations using tetracyanoethylene in an organic solvent. Excipients used: talc, 8.1 mg; starch, 50 mg; magnesium stearate, 0.9 mg; and lactose, 57 mg Recovery* 5 c.v., % Compound Amidopyrine . . . . . . . . . . . . Amidopyrine tablets$ . . . . . . . . . . Antipyrine . . . . . . . . . . . . . . Antipyrine tablets$ . . . . . . . . . . Phenelzine sulphate . . . . . . . . . . Nardil tablets,§ 0.015 g per tablet . . . . . . INH . . . . . . . . . . . . . . . . Isocid tablets,lO.O5 g per tablet .. . . . . Hydralazine hydrochloride . . . . . . . . Apresoline tables,(( 0.05 g per tablet . . . . * Average of three determinations. t Spectrophotometric assay.17 $ Laboratory-made tablets. 4 Warner Company, Michigan, USA. fi Chemical Industries Development Co., Egypt. (1 Ciba Giegy, Basle, Switzerland. Tetracyanoethylene method 100.03 k 1.15 101.01 k 1.9 100.00 5 1.41 101.10 & 1.6 100.00 k 1.85 99.88 f 2.1 99.997 k 2.2 99.68 5 2.4 100.00 5 0.46 101.00 5 1.03 Official method13 101.12 & l.08t 99.87 f 1.8 99.84 2 1.11 101.73 k 1.32 99.94 k 1.2 98.99 2 1.29 98.96 f 1.65 101.46 f 2.1 100.81 2 0.8 100.38 5 1.15 Table 3. Statistical analysis of tetracyanoethylene method compared with official USP 1980 method13 Tetracyanoethylene method Official method Sample Mean recovery, % N V Mean recovery, YO N V Amidopyrine .. . . . . . . 100.03 k 1.15 6 1.32 101.125 1.08* 8 1.19 t = 1.499 (1.86) F = 1.109 (9.3) t = 0.188 (1.86) F = 1.146 (9.3) t = 1.154(1.78) F = 1.21 (3.9) t = 0.057 (1.83) F = 2.375 (8.9) t = 0.899 (1.796) F = 3.09 (4.1) Antipyrine . . . . . . . . 100.00 k 1.41 6 1.988 99.842 1.11 4 1.23 INH . . . . . . . . , . 99.997 f 2.20 10 4.84 98.96 k 1.65 4 2.67 Phenelzine sulphate . . . . . . 100.00 5 1.85 7 3.42 99.94 k 1.2 4 1.44 Hydralazine hydrochloride . . 100.00 k 0.46 9 0.21 100.31 f 0 . 8 4 0.65 * According to spectroscopic method.18ANALYST, NOVEMBER 1986, VOL. 111 1287 0.8 - 330 350 370 390 410 430 450 470 490 I I 1 .--’ I I I I Wavelengthlnm Fig. 1. Absorption spectra for the product of the reaction between tetracyanoethylene and the studied groups. A, Phenelzine sulphate, 80 c1.g ml-1; B, hydrolazine hydrochloride, 10 pg ml-I; C, isonicotinic acid hydrazide, 10 pg ml-1; D, antipyrine, 60 yg m1-I; and E, amidopyrine, 40 pg ml-1 addition of pyrazolone or hydrazine derivatives to TCNE solution, a bathochromic shift to longer wavelengths is obtained, for a limited time, at the optimum temperature.The reaction product formed is stable for at least 24 h, thus permitting quantitative analysis to be carried out with good reproducibility. The new absorption band (Fig. 1) formed is the result of the formation of a charge-transfer complex through the interac- tion of TCNE as a n-acceptor and the studied drugs as n-donors: D + TCNE [D + TCNE] n-n, complex D*+ + TCNE*- .ir Radical ions The effect of the solvent on the formation of the charge- transfer complex was studied in acetonitrile and chloroform; in other solvents such as dioxane or cyclohexane, no absorp- tion band was observed in the electronic spectra at the concentration tested.Acetonitrile is preferred as a solvent owing to the high molar absorptivities of the complexes formed in it. The effects of the solvent on the formation of several donor - acceptor systems have been reported,’ and the formation of contact charge-transfer spectra by the interaction of amines with halomethanes15 has also been reported. The molar absorptivities are 1.9 x lO3,1.55 x lO3,2.57 x lO3,6.04 X l o 3 and 4.93 x l o 3 1 mol-1 cm-1 for amidopyrine, antipyrine, phenelzine sulphate, INH and hydralazine hydrochloride, respectively.This illustrates the higher sensi- tivity of this method over the 2,5-dichlorobenzoquinone method12 for the determination of amidopyrine and anti- pyrine, which reported molar absorptivities of 666 and 212 1 mol-1 cm-1 for amidopyrine and antipyrine, respectively. The stoicheiometry of the reaction between TCNE and the studied drugs was studied by Job’s method of continuous variation16 and it was found that the complexes are formed in the ratios 2 : 3, 1 : 1, 1 : 1, 2 : 3 and 2 : 3 (donor : acceptor) for amidopyrine, antipyrine, phenelzine sulphate, INH and hydralazine hydrochloride, respectively. Calibration graphs were constructed by plotting the absorbance versus the concentration of the drug in pg ml-1 final dilution.The most effective TCNE concentration was found to be given by the use of 2 ml of 0.1% solution. A concentration of up to about 1 yg could be detected as a yellow spot on TLC plates using tetracyanoethylene as a spray reagent. Regression analysis indicated that the value of the intercept is small; the values were found to be 0.028, -0.0082, 0.0071, 0.0007 and 0.0006 for amidopyrine, antipyrine, phenelzine sulphate, isonicotinic acid hydrazide and hydralazine HCl, respectively. The method was applied to the determination of the studied drugs in their pharmaceutical dosage forms (Table 2). The concentration of the drug in the preparation is calculated from the corresponding calibration graph or regression equation. Statistical analysis of the results obtained, compared with those of the official method, showed no significant difference in the accuracy and precision of the two methods (Table 3).The method is reproducible, accurate and precise. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. Rao, C. N. R., Bhat, S. N., and Dwivedi, P. C., in Brame, E. G., Editor, “Applied Spectroscopy Review,’’ Volume 5 , Marcel Dekker, New York, 1972, pp. 1-170. Melby, L. R., in Patai, S . , Editor, “The Chemistry of the Cyano Group,” Interscience, New York, 1970, pp. 639-670. Ibrahim, F. A., PhD Thesis, Faculty of Pharmacy, University of Mansoura, Egypt, 1984. Merrifield, R. E., and Phillips, W . P., J. Am. Chem. SOC., 1958, 80, 2778. Middleton, W. J., Little, E. L., Coffmard, D. D., and Engelhardt, V. A., J. Am. Chem. SOC., 1958, 80, 2795. Hutzinger, O., Anal. Chem., 1969, 41, 1662. Macke, G. F., J. Chromatogr., 1968,36, 537. Hutzinger, O., J. Chromatogr., 1969, 40, 117. Hutzinger, O., and Jamieson, W. D., Anal. Biochem., 1970, 35, 351. Hutzinger, O., Heacock, R. A., MacNeil, J. D., and Frei, R. W . , J. Chromatogr. 1972, 68, 173. Heacock, R. A., Forrest, J. E., and Hutzinger, O., J . Chromatogr., 1972, 72, 343. Abdine, H., Elsayed, M. A., Chaaban, I., and Abdel-Hamid, M . E., Analyst, 1978, 103, 1227. “United States Pharmacopoeia, XXth Revision, National Formulary, XVth Edition,” USP Convention, Rockville, MD, 1980. Stahl, E., “Thin-layer Chromatography,” Springer-Verlag, Berlin, 1969, p. 17. Stevenson, D. P., and Koppinger, G. M., J. Am. Chem. SOC., 1962, 84, 149. Rose, J., “Advanced Physico-Chemical Experiments,” Pitman, London, 1964, p. 54. Clarke, E. G. C., “Isolation and Identification of Drugs,” Volume 1, Pharmaceutical Press, London, 1978, p. 185. Paper A6140 Received February IZth, 1986 Accepted April 18th, 1986

ISSN:0003-2654    

DOI:10.1039/AN9861101285

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, NOVEMBER 1986, VOL. 111 1289 Allylthiourea as a Reagent for the Spectrophotometric Determination of Osmium Basilio Morelli Dipartimento di Chimica, Universita di Bari, Via Amendola 173, 70126 Bari, Italy Allylthiourea is proposed as a reagent for the spectrophotometric determination of osmium(Vlll). Osmium(VIII) forms a 1 : 1 complex with allylthiourea. Conformity to Beer's law was observed for up to 20 Fg ml-1 of osmium in acidic medium (molar absorptivity 2.17 x lo4 I mol-1 cm-1, at 298 nm); the Sandell's sensitivity of the reaction was 0.0087 pg cm-2 per 0.001 absorbance unit. The effect of foreign ions was also studied. A statistical evaluation of the experimental results was undertaken and a comparison with the most important recent spectrophotometric reagents for osmium is presented.The proposed method has the advantages of simplicity, reasonable sensitivity and rapid determination without the need for extraction or heating. Keywords: Osmium determination; allylthiourea; spectrophotometry A wide variety of chromogenic reagents have been employed for the spectrophotometric determination of osmium, but a large number of these lack adequate sensitivity or selectivity and necessitate the separation of osmium by extraction or distillation before the determination. Most of the proposed methods have not found wide application because of their complexity . In a recent paper,l a sensitive and rapid spectrophotometric method was reported for the determination of palladium(I1) using allylthiourea (ATU) as a reagent. It was of interest to see if ATU could also be used for the determination of osmium.In this paper a detailed study of the osmium(VII1) - ATU system is described. The method for the determination of Os(VII1) involves a simple technique and the sensitivity is promising and compares well with that of other known met hods. Experimental and Results Reagents Osmium( VIII) standard solution. Working solutions of about 7.8 X 10-4 M osmium were prepared as follows. A 0.1-g ampoule of osmium tetroxide (Riedel-de-Haen, Seelze, Hannover) was scratched with a file, weighed and broken beneath the surface of about 20 ml of 0.2 M sodium hydroxide solution. The red - orange solution was washed into a 500-ml calibrated flask and mixed with 50 ml of concentrated nitric acid. After stirring for about 30 min, the solution was diluted to volume with distilled water.The recovery and weighing of the broken glass gave the mass of osmium tetroxide by difference. Solutions were stored in a refrigerator. Allylthiourea solution. Stock solutions, 0.03 M, were pre- pared by dissolving the requisite amount of allylthiourea (Carlo Erba, Milan) in distilled water. Foreign ion solutions. Analytical-reagent grade salts of various elements were used. The solutions contained 2-10 pg ml-l of the ions. Apparatus The spectrophotometric measurements were made in the double-beam mode with a Perkin-Elmer 555 spectrophoto- meter using 1-cm quartz cells. The automatic base-line corrector was employed, the base line being determined with both sample and reference cuvettes filled with reagent blank solution.Procedure Into a 5-ml calibrated flask, transfer 1.5 ml of 0.03 M ATU solution and 200 pl of 6 M hydrochloric acid. Add a known volume of sample containing up to 100 pg of osmium, then dilute to volume with distilled water and measure the absorbance at 298 nm against a reagent blank. Effect of Heating In a preliminary investigation, the absorbance of solutions prepared as described under Procedure was measured at 298 nm, at different temperatures as a function of time, by varying the standing time after the addition of the reagent. Fig. 1 shows the results of this study. The graphs were obtained with samples containing 25 pg of osmium per 5 ml. The samples measured at 20-90°C showed no change in the spectral curve. Maximum absorbance was reached im- mediately after the preparation of the samples.At room temperature (line A) the absorbance was stable for at least 1 h (further measurements were not made). At 50 "C (line B) and 70°C (line C) the absorbance was maximum and constant up to about 10 and 5 min, respectively, and then it gradually decreased by about 2 and 6%, respectively, within 1 h. At 90 "C (line D) an irregular behaviour was observed. For these reasons, all further measurements were made at room temperature, at about 10 min after sample preparation. 0.60 I 0 10 20 30 40 50 60 Ti me/mi n 0.45 Fig. 1. Effect of temperature and heating time on absorbance of Os(VII1) - ATU complex. Conditions: 25 pg of 0 s per 5 ml; h = 298 nm; reference = reagent blank. A, 20 "C; B, 50 "C; C, 70 "C; and D, 90°C1290 ANALYST, NOVEMBER 1986, VOL.111 0 0.2 0.4 0.6 0.8 1 .o 6 M HCli ml per 5 ml 0.70 Fig. 2. Effect of hydrochloric acid concentration on the absorbance of Os(VII1) - ATU complex. Conditions: 32 pg of 0 s per 5 ml; h = 298 nm; reference = reagent blank 1 I I I I J 0 0.5 1 .o 1.5 2.0 2.5 0.55 ' 0.03 M ATUiml per 5 ml Fig. 3. Effect of allylthiourea concentration on the absorbance of Os(VII1) - ATU complex. Conditions: 27 pg of 0 s per 5 ml; h = 298 nm; reference = reagent blank 1 .o 0.8 0.6 0 m e rn a 0.4 0.2 0 I . - - - - - - - 280 320 360 Wavelengthinm Fig. 4. Absorption spectra of: A, Os(VII1) - ATU complex; and B, ATU (reagent blank). Conditions: A, 40 pg of 0 s per 5 ml; reference, reagent blank. B, 1.5 ml of 0.03 M ATU per 5 ml; reference = water Effect of Acidity To test the effect of acidity, the absorbance of solutions as a function of hydrochloric acid concentration was measured at 298 nm against a reagent blank.A typical graph of absorbance vs. HC1 concentration is shown in Fig. 2. The graph was obtained with samples containing 32 pg of osmium per 5 ml of solution. The absorbance was found to be constant and maximum in solutions containing 150-350 pl of 6 M HC1 per 5 ml and then decreased outside this range. Hence, a volume of 1 .o a, C m e s: 2 0.5 0 I I I 0.5 1 .o 1.5 Os(VIII) : mol of ATU Fig. 5. Molar ratio of the Os(VII1) - ATU complex by the molar-ratio method. Conditions: ATU concentration, 6.7 X M; h = 298 nm; reference = water Table 1. Tolerance of the osmium(VII1) - allylthiourea system to foreign ions.All solutions contained 26 pg of 0 s per 5 ml. The tolerance to a foreign ion was taken as the largest amount that gives an absorbance not more than 1% different from that of osmium alone Amount tolerated relative Foreign ion to Os, % Na(1) . . . . . . 220 Li(1) . . . . . . 350 Rb(1) . . . . . . 250 Cs(1) . . . . . . 1650 Mg(I1) . . . . 200 Ca(I1) . . . . . . 650 Sr(1I) . . . . . . 33 Ba(I1) . . . . . . 660 Cr(II1) . . . . 10 Mn(I1) . . . . 35 Fe(II1) . . . . 3 Ni(I1) . . . . . . 80 Cu(I1) . . . . . . 6 Tl(1) . . . . . . 800 La(II1) . . . . 30 Co(I1) . . . . . . 120 Foreign ion Ag(1) . . . . Zn(I1) . . . . Cd(I1) . . . . Al(II1) . . Sn(I1) . . . . Pb(I1) . . . . As(II1) . . Sb(II1) . . Bi(II1) . . . . Ru(II1) . . Rh(II1) . . Pd(I1) .. . . Ir(II1) . . . . Hg(I1) . . . . Pt(I1) . . . . Amount tolerated relative to os, Yo . . 100 . . 3000 . . 450 . . 10 . . 350 . . 450 * . 100 . . 70 . . 30 . . 15 . . 15 . . 15 . . 13.5 . . 10 . , 40 200 pl of 6 M HC1 was chosen for all subsequent measure- ments, Effect of ATU Concentration To a series of samples containing a fixed amount of osmium, increasing volumes (from 0.2 to 2.5 ml) of 0.03 M ATU solution were added. Some experimental results, obtained with samples of 27 pg of osmium per 5 ml, are illustrated in Fig. 3. The graph was plotted at the wavelength of maximum absorption, 298 nm. The most pronounced effect was obtained by adding from 1.4 to 2.5 ml of the 9 0 3 M ATU solution in the standard procedure. The amount of ATU solution was therefore chosen as 1.5 ml.Absorption Spectra Fig. 4 shows the absorption spectrum of the Os(VII1) - ATU complex (40 pg of 0 s per 5 ml) against a reagent blank, obtained under optimum conditions by the proposed method (line A), along with that of the reagent blank against water (line B). The Os(VII1) - ATU complex shows an absorption maximum at 298 nm. The ligand, however, exhibits someANALYST, NOVEMBER 1986, VOL. 111 1291 absorption around this wavelength (which sharply decreases and becomes insignificant from 320 nm onwards), emphasis- ing the need for measurements at 298 nm to be performed against the reagent blank. Calibration Graphs The system was found to obey Beer’s law up to 100 pg of osmium in the final 5 ml. The molar absorptivity at 298 nm, calculated from the linear regression coefficient, is 2.17 X 10-4 1 mol-1 cm-1.The Sandell’s sensitivity for the determina- tion of osmium is 0.0087 pg cm-2 per 0.001 absorbance unit. Molar Composition of the Complex The stoicheiometry of the complex was determined by the molar-ratio method. Fig. 5 shows a typical molar-ratio plot obtained with samples containing 6.7 x 10-5 M ATU and increasing amounts of osmium as required. The graph shows a break at a molar ratio of 1.0, corresponding to a ratio of osmium to ATU of 1 : 1. Effect of Foreign Ions In order to study the effect of other platinum metals and common cations, solutions were prepared containing 26 pg of Os(VII1) per 5 ml and various amounts of foreign ions. The tolerance of the Os(VII1) - ATU system to an interfering substance was defined in usual way,l-5 i.e., as the largest amount of that substance that would give an absorbance not more than 1% different from that of osmium alone.The substances tested and the tolerance, as defined above, are reported in Table 1. The tolerance to foreign ions is satisfactory. Table 2. Statistical analysis of the spectrophotometric determination of osmium(VII1) with allylthiourea. Level of significance, p = 0.01; number of standard specimens, n = 15 Angular Detection Correlation coefficient/ limit/ y-Intercept coefficient cm-1 yg-1 ml Variance(s,*) pg ml-1 -0.001 0.9995 0.1140 4.75 x 10-4 0.5 Statistical Analysis of Results A critical evaluation of the proposed method follows, devel- oped by assessing the results of the statistical analysis of the experimental data.These results are shown in Table 2. The linearity of calibration graphs and the conformity to Beer’s law of the system is proved by the high value of the correlation coefficient of the equation of regression and the intercept on the y-axis close to zero. The value of the detection limits and variance are evidence for the sensitivity of the method and the negligible scatter of the points with respect to the line of regression. Statistical analysis of the calibration graph allows the calculation of the error, S,, in the determination of a given concentration, c, by using the following equation‘? where So = VX(A - ACalc,)2/n-2; A = experimental value of absorbance; Acalc. = absorbance value calculated from the regression equation; b = angular coefficient of the regression line; and c’ and A’ = average concentration and absorbance values, respectively, for n standard specimens.The graph of S, against 0 s concentration is shown in Fig. 6, in form of histograms. The error reaches a minimum for the valueA = A’. 0.190 0.185 d 0.180 0 5 10 15 20 IOsl, p.p.m. Fig. 6. Histograms of the error in the determination of the concentration of 0 s Table 3. Comparison of spectrophotometric reagents for osmium(VII1) Reagent Allylthiourea . . . . . . . . . . . . . . Salicylaldehyde hydrazone . . . . . . . . . . Phenanthraquinone monoxime* . . . . . . . . Detn. as osmate . . . . . . . . . . . . . . Thiocyanate and hexamethylphosphoramide* . . 2-Methyl-l,4-naphthoquinone thiosemicarbazone * 1,5-Diphenylcarbazide * .. . . . . . . . . HglI and 4,5-diamino-6-hydroxypyrimidine sulphate 3-Hydroxy-2-methyl- 1,4-naphthoquinone monoxime Phenanthraquinone monosemicarbazone . . . . N-(4-Methoxyphenyl)-a-thiopicolinamide* . . . . N-(4-Methylphenyl)-ar-thiopicolinamide* . . . . Pyrimidenethiol* . . . . . . . . . . . . . . Phenanthraquinone monoxime* . . . . . . . . 1 -Phenyl-4,4,6-trime t hyl-( 1 H,4H)-2-pyrimidine t hi01 * Perphenazine dihydrochloride . . . . . . . . Tetramethylthiurammonosulphide* . . . . . . Cyclohexane- 1,3-dione bisthiosemicarbazone monohydrochloride . . . . . . . . . . . . * By solvent extraction. . . . . . . . . . . . . . . * . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Range1 pg ml-l 0.5-20 S19.7 S17.78 69.51 G30 63.96 613 1.58-1 1.22 1-20 1-18 8-64 1-20.9 per 10 ml 4-77 per 10 ml 0.1-5.2 5-45 0.5-16 5-150 0.57-9.13 Molar absorptivity1 Wavelength1 1 mol-1 cm-l nm 2.17 x 104 3.1 x 103 1.85 x 104 1.9 x 104 2.74 x 103 2.09 x 104 5.76 x 104 2.7 x 104 1.08 x 104 1.6 x 104 8.3 x 103 6.5 x 103 4.6 x 103 1.33 x 104 3.43 x 104 3.53 x 103 8.8 X lo4 298 430 500 470 340 595 470 560 500 430 440 430 500 475 520 528 580 1.87 x 104 375 Molar ratio 1 : l 1 : l 1 : 2 1 : 4 1 : l : l 1 : 2 1 : l 1 : 2 : 2 1 : 2 1 : l 1 : l 2 : 3 1 : l 1 : l 1 : l 1 : l - 1 : 2 Reference This work 9 10 11 12 13 14 15 16 17 18 18 19 20 21 22 23 241292 40 30 8 20 10 0 B ;I^ 5 10 15 20 10~1, p.p.m. Fi .7. Variation of confidence limits at p = (A) 0.01, (B) 0.05 and <4 0.1 levels of significance, presented as percentage uncertainty on the concentration The quantity S, also allows the determination of confidence limits.6 These results are shown graphically in Fig.7 in the form of percentage uncertainty on the concentration, TpS&901~3~7~8 (Tp is the Student’s quantity) at levels of significance p = 0.01, 0.05 and 0.1 (99, 95 and 90% probability, respectively). This is a useful way of representing confidence limits because it allows a direct calculation of the relative uncertainty on concentration over the full range of concentrations tested, and hence it is a guide to the level of precision that may be expected from the application of the analytical procedure proposed. To test the accuracy and precision of the proposed method, ten successive determinations on the same sample solution (12.5 pg ml-1 of 0s) were carried out at one time.The results were in the range 12.4-12.7 pg ml-1 with a mean of 12.54 pg ml-1, a standard deviation, SD, of 0.084 pg ml-1 and a coefficient of variation, CV, of 0.67%. Conclusions The proposed method is very simple, using a minimum number of reagents and reaction sequences, and allows the determination of osmium with good accuracy and precision. From a review of the main spectrophotometric methods for osmium developed in recent years (Table 3), ATU is seen to ANALYST, NOVEMBER 1986, VOL. 111 rank among the more sensitive reagents. The proposed method is also faster and simpler than other sensitive methods, which often require prior solvent extraction of the complex and/or are only utilisable in narrower concentration ranges.The author thanks Ms. Marina Mariani and Mr. Pasquale Peluso of the Dipartimento di Chimica of Universita di Bari for carrying out part of the experimental work. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21 * 22. 23. 24. Morelli, B., Anal. Lett., 1984, 17, 2267. Morelli, B., Analyst, 1983, 108, 386. Morelli, B., Analyst, 1983, 108, 870. Morelli, B., Analyst, 1983, 108, 959. Morelli, B. , Analyst, 1984, 109, 47. Nallimov, V. V., “The Application of Mathematical Statistics to Chemical Analysis,” Pergamon Press, Oxford, 1963, p. 189. Morelli, B., Analyst, 1983, 108, 1506. Morelli, B., and Peluso, P., Anal. Lett., 1985, 18, 1113. Ray, H. L., Garg, B. S., and Singh, R. P., J . Indian Chem. Soc., 1979, 56, 975. Kamil, F., Sindhwani, S. K., and Singh, R. P., Mikrochim. Acta, 1980, 1, 345. Kamil, F., Sindhwani, S. K., and Singh, R. P., Ann. Chim. (Rome), 1980, 70, 417. Marczenko, Z . , Balcerzak, M., and Kus, S. , Talanta, 1980,27, 1087. Pal, B. K., Chowdhury, R. P., and Mitra, B. K. , Talanta, 1981, 28, 62. Kamini, M., Sindhwani, S. K., Singh, R. P. , Fresenius 2. Anal. Chem., 1982,311, 521. Jaya, S., and Ramakrishna, T. V., Talanta, 1982, 29, 619. Kudra, S. K., Katyal, R. P. , and Katyal, M. , Acta Cienc. Zndica Ser. Chem., 1982, 8, 6. Kamini, M., Sindhwani, S. K., and Singh, R. P., Ann. Chim. (Rome), 1983, 73, 223. Bag, S. P., and Bhattacharya, B., J . Indian Chem. SOC. 1983, 60, 283. Anuse, M. A., Mote, N. A. , and Chavan, M. B. , Talanta, 1983, 30, 323. Wasey, A., Bansal, R. K., and Puri, B. K., Mikrochim. Acta, 1984, 1, 211. Wasey, A., Bansad, R. K., Puri, B. K., and Rao, A. L. J., Talanta, 1984, 31, 205. Gowda, A. T., Gowda, H. S . , and Gowda, N. M. M., Analyst, 1984, 109, 651. Ilyas, S. Q. R., and Joshi, A. P., Chem. Anal. (Warsaw), 1984, 29, 231. Reddy, K. H., and Reddy, D. V., Anal. Lett., 1984, 17, 1275. Paper A6198 Received March 24th, 1986 Accepted April 30th, 1986

ISSN:0003-2654    

DOI:10.1039/AN9861101289

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, NOVEMBER 1986, VOL. 11 1 1293 Leucoquinizarin as an Analytical Spectrophotometric and Fluorimetric Reagent Part 2.* Determination of Beryllium Miguel Angel Bello Lopez, Manuel Callejon Mochon,t Jose L. Gomez Ariza and Alfonso Guiraum Perez Department of Analytical Chemistry, Faculty of Chemistry, University of Seville, 4 10 12 Seville, Spain The reaction between beryllium and leucoquinizarin was studied both spectrophotometrically and spectrofluorimetrically, and methods for the determination of beryllium are proposed on the basis of the 1 : 1 chelate formed. The yellow chelate is formed a t pH 5.3-5.5 in a medium containing 20% of ethanol, and the absorption is measured at 438 nm. The molar absorptivity (1.2 x 104 I mol-1 cm-1) allows the determination of 0.1-0.8 p.p.m.of beryllium. The fluorimetric method allows the determination of 5-100 p.p.b. of Be2+ in a medium containing acetate buffer (pH 4.6) and 70% of ethanol (A,,, = 450 nm, he, = 476 nm). The fluorimetric method has been applied t o the determination of beryllium in an aluminium alloy. Keywords: Leucoquinizarin; beryllium determination; spectrophotometry; fluorimetry; aluminium - magnesium alloy Leucoquinizarin is a hydroxy anthracene compound and shows a close relationship with hydroxy anthraquinone compounds such as alizarin. The analytical applications of leucoquinizarin have been considered in an earlier paper1 and the spectrophotometric determination of magnesium in phar- maceutical preparations proposed. This paper considers the spectrophotometric and fluorimetric determination of beryl- lium and the evaluation of the beryllium content in an aluminium - magnesium alloy.Experimental Reagents All solvents and reagents were of analytical-reagent grade and de-ionised water was used throughout. Leucoquinizarin solution in ethanol, 2 x 10-3 M. Prepared from the commercial product (Ega-Chemie). Beryllium standard solution, 1.014 g 1-1. Prepared from Be(N0&.4H20 by dissolving in 0.01 M HC1. The beryllium content was determined by precipitation with ammonia solution .2 Working solutions were prepared by dilution. Acetylacetone solution, 5% VIV. Buffer solutions, p H 5.4 and 4.6. 1 M acetic acid - sodium acetate. Apparatus Two spectrophotometers were used: a Coleman 55 (a digital instrument used for measuring absorbances at fixed wavelengths) and a Perkin-Elmer 554 recording spectropho- tometer for absorbance scanning.Matched glass and quartz cells of 1.0-cm optical path length were used. A Beckman 70 pH meter with a combined saturated calomel - glass electrode was used for the pH measurements. All fluorescence measure- ments were made with a Perkin-Elmer LS-5 spectroflu- orimeter, equipped with Ultrathermostat Colora K5 and 1.0-cm cells. * For Part 1, see reference 1. t To whom correspondence should be addressed. Procedures Spectrophotometric determination of beryllium Samples are prepared at room temperature in 25-ml calibrated flasks by mixing a volume (up to 5 ml) containing 2.5-20 pg of beryllium with 5 ml of 2 x 10-3 M leucoquinizarin solution in ethanol and 5 ml of acetic acid - sodium acetate buffer (pH 5.4), diluting to the mark with water.The absorbance is measured at 438 nm against a blank. Unknown concentrations are determined from a calibration graph established from known concentrations of beryllium. Fluorimetric determination of beryllium Into a 25-ml calibrated flask place an aliquot (up to 5 ml) of sample containing 0.125-2.50 pg of beryllium. Add 1 ml of 2 X 10-3 M leucoquinizarin solution in ethanol and 2.5 ml of acetic acid - sodium acetate buffer (pH 4.6) and dilute to the mark with ethanol. After 10 min, measure the fluorescence (25 rl: 1 "C) at 476 nm, using an excitation wavelength of 450 nm. Determine the concentration of beryllium in the sample from a calibration graph prepared under identical conditions.Determination of beryllium in aluminium Weigh accurately a suitable amount (1 .0-1.5 g) of sample and dissolve in 1 : 1 nitric acid. Filter the solution into a 50-ml calibrated flask and dilute to the mark with 1 : 1 nitric acid. Transfer 5.0 ml of the solution, with a pipette, into a vessel, add 20 ml of distilled water and adjust the pH to about 2 with NaOH solution. Add 50 ml of 10% mlV EDTA solution and adjust the pH to about 7-8 with NaOH solution. Transfer into a separating funnel and add 5 ml of 5% V/V acetylacetone solution, mix well and wait for 5 min. Finally extract three times with 7.0-ml aliquots of chloroform for 2 min. Transfer the extracts into a platinum crucible, add 2 ml of 16 M nitric acid and 2 ml of 60% perchloric acid and evaporate to dryness.Repeat the evaporation with the same volumes of nitric acid and perchloric acid. Dissolve this final residue in 15 ml of 0.1 M nitric acid and transfer again into a separating funnel. Repeat twice the treatment with EDTA and the extraction with acetylacetone. Dissolve the final residue in water and transfer into a 10-ml calibrated flask, adding de-ionised water to the mark. Beryllium was determined in 3 ml of this solution as described under Fluorimetric determination of beryllium.1294 ANALYST, NOVEMBER 1986, VOL. 111 Results and Discussion Study of the Beryllium - Leucoquinizarin System In an excess of leucoquinizarin the beryllium solutions give a yellow colour, the absorption spectrum of which is shown in Fig. 1. This chelate also has fluorescence properties and the excitation and emission spectra at pH 4.6 are shown in Fig.2 (A,,, = 450 nm, A,, = 476 nm). 1 .o 0.8 al C m 0.6 2 a D 0.4 0.2 0 Wavelengt hin m Fig. 1. Absor tion spectrum of beryllium - leucoquinizarin in ethanol - water & + 3) solution. Conditions: [beryllium] = 1.1 X M, [leucoquinizarin] = 2.0 X M; pH = 5.4. Broken line = reagent blank Wavelengt hinm Fig. 2. (a) Excitation spectra: A, beryllium - leucoquinizarin; B, reagent only. (b) Emission spectra (hex, = 450 nm): A, beryllium - leucoquinizarin; B, reagent only. Conditions: [leucoquinizarin] = 8.0 x 10-5 M; [beryllium] = 1.1 x 10-5 M; 60% V/V ethanol; pH = 4.5 0.6 (I) m 0.5 2 2 0.4 2 0.3 PH 3.5 4.0 4.5 5.0 5.5 PH Fig. 3. Influence of pH on the beryllium - leucoquinizarin complex.(a) Graph of absorbance versus pH at 438 nm. ( b ) Graph of fluorescence versus pH (hex, = 450 nm and he, = 476 nm). [Leucoquinizarin] = 4 x 10-4 M; [beryllium] = 8.8 x M Influence of Experimental Variables The pH of the medium had a major effect on the absorbance and fluorescence intensity. Fig. 3 shows a narrow pH interval (between 5.3 and 5.5) on the absorbance versus pH graph, at 438 nm, in which the absorbance is independent of the pH. For more alkaline media the complex is precipitated and the absorbance decreases. For this reason, an acetic acid - acetate buffer solution (pH 5.4) was used and a 5-ml volume was selected as the most suitable. An analogous study was carried out on the fluorescence properties of the complex (Fig. 3), which shows a maximum fluorescence intensity at pH 4.54.7.Acetate buffer (pH 4.6) was used in all subsequent fluorescence work, 2.5 ml being the most suitable volume. A study of the influence of the percentage of ethanol in the medium revealed that the ethanol concentration is critical and must be precisely controlled. A high percentage of ethanol is necessary for high sensitivity but limits the volume of aqueous solutions (buffer and cation) that can be added. For these reasons the percentage of ethanol in the samples was fixed at 70%. The absorbance of the complex was studied as a function of the molar ratio of leucoquinizarin to beryllium. At least a 5-fold molar excess of the reagent over beryllium was necessary in order to obtain the maximum absorbance value. The order of addition of the reagents was not found to be important.A sequence of metal ion, reagent and then buffer was adhered to during the preparation of all measured solutions. Under these conditions, the colour is developed immediately and remains for at least 3.5 h. The effect of leucoquinizarin concentration on the intensity of the fluorescence for a 1.1 x 10-5 M beryllium solution was also tested under analogous conditions to those given under Fluorimetric determination of beryllium. The intensity of the fluorescence increased when the concentration of leuco- quinizarin increased, reaching a maximum and constant value for 9.9 x 10-5 M reagent. The fluorescence intensity dimin- ished by 4% on increasing the temperature from 25 to 30°C. Further increases to 70°C caused a further decrease of 18%.The work reported here was carried out at 25 k 1 "C. Stoicheiometry of the Complex Because beryllium forms a highly dissociated chelate with leucoquinizarin, classical methods such as those of Yoe and Jones3 and Job4 do not yield reliable results for the determination of the stoicheiometric ratio. For this reason the Asmus5 and modified6 Holme and Langmyhr methods, which are more suitable for this type of chelate, were investigated. Fig. 4 shows the results obtained, which indicate a 1 : 1 molar ratio. 1 .o 0.8 $ 0.6 7 - 0.4 0.2 4.4 - 4.2 =' a 4.0 3.8 7 - -1 3.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1 iA Log (X-1) Fig. 4. Determination of the composition of the beryllium - leucoquinizarin complex. (a) Asmus method, and ( b ) modified Holme and Langmyhr method.Conditions: pH 5.3 and h = 438 nm. A = absorbance; L = free ligand concentration; V = volume of reagent stock solution taken; X = Amax,/A; n = ratio of ligand to metal ion in the complexANALYST, NOVEMBER 1986, VOL. 111 1295 The Asmus method5 was also applied to the chelate in the excited state, measuring the fluorescence intensity at 476 nm with excitation at 450 nm. The results were in agreement with those of the chelate in the ground state. Spectrophotometric Determination of Beryllium with Leuco- quinizarin The relationship between the concentration and the colour intensity obeyed Beer’s law in the range 0.1-0.8 p.p.m. of beryllium at 438 nm. The molar absorptivity is 1.2 x 104 Table 1. Effect of foreign ions on the spectrophotometric determination of 0.4 p.p.m.of beryllium Tolerance, p.p.m. Foreign ion 400 Alkalimetals, S042-, NO3-, C1-, Br-, I-, AsO2-, C104-, SCN-, C103-, NO2-, S032-, 103-, S2032-, Br03- 40 Mg(II), Ca(II), Ba(II), La(III), Zn(II), Cd(II), TI(I), Pb(II), Mo(VI), C032-, tartrate 20 phthalate, As043- 10 Sr(II), Co(II), Ni(I1) 200 S20,2-, B4072- 120 104- Ag(I), Ce(IV), U(VI), In(III), Bi(III), 4 Zr ( W , W V ) , Hg(II), Sn(II), V(V), Cr(VI), W(VI), EDTA 2 Au(III), 0.4 Pd(IV),Sb(III), C2042-,HP042- <0.4 Ti(IV), Cr(III), Fe(III), Mn(II), Cu(II), AI(III), F-, citrate Table 2. Effect of foreign ions on the fluorirnetric determination of 60 p.p.b. of beryllium Tolerance, p.p.b. 30 000 6000 3000 ’ 1500 600 60 <60 Foreign ion Alkali metals, S042-, NO3-, C1- Mg(II), Ca(II), Sr(II), Ba(II), Zr(IV), Mn(II), U(VI), Cd(II), Ga(III), T1(1), V(V), Clod-, Br-, I-, 104-, 103-, As043-, S 2 0 8 2 - , Br03-, Y(III), Co(II), Ni(II), Ag(I), Zn(II), In(III), Mo(VI), C103-, S 2 0 3 2 - Sn(II), Pb(I1) La(III), Pd(IV), Cu(II), Os(IV), Ce(IV), Sb(III), Cr(VI), C032-, SCN-, HP042-, phthalate Au(III),H~(II),P~O,~-,SO~~-,EDTA, C ~ 0 4 ~ - , NO2-, Se032-, As0,- , tartrate Ti(IV), Cr(III), Fe(III), AI(III), Bi(III), W(VI), F-, citrate B4072- 1 mol-1 cm-1.The sensitivity of the reaction as defined by Sandell is 0.00075 pg cm-2.7 The optimum concentration range evaluated by a Ringbom plot is 0.15-0.65 p.p.m. The relative error of the method (P = 0.05) for 11 measurements was f0.23%. Effect of foreign ions For the determination of 0.4 p.p.m. of beryllium by this method, foreign ions (added before the leucoquinizarin reagent) can be tolerated at the levels given in Table 1.The criterion for interference was an absorbance value varying by more than +3% from the expected value of beryllium alone. Fluorimetric Determination of Beryllium with Leucoquinizarin Using the method described earlier, beryllium can be deter- mined fluorimetrically with leucoquinizarin. The calibration graph was linear in the range 5-100 p.p.b. of beryllium. The relative error (P = 0.05) from 11 samples (60 p.p.b.) was +O.ST%o. Effect of foreign ions The effect of several ions on the fluorescence intensity of beryllium (at a level of 60 p.p.b.) was studied. The criterion for interference was fixed at f 4 % variation of the average fluorescence intensity, calculated for the established level of beryllium.The results are given in Table 2. Application The fluorimetric procedure was satisfactorily applied to the determination of beryllium in an A1 - Mg alloy (MBH, GO5 H2: Cu, 0.21%; Mg, 3.09%; Si, 0.31%; Fe, 0.61%; Mn, 0.22%; Ni, 0.16%; Zn, 0.16%; Pb, 0.05%; Sn, 0.15%; Ti, 0.16%; Cr, 0.20%; and Be, 0.002%). The sample was dissolved as described under Determination of bery ZZium in aluminium and beryllium was separated following the method proposed by Merrill et aZ.8 Owing to the low concentration of beryllium in the sample, the results obtained with the fluorimetric procedure were compared with those given by atomic absorption spectrometry, using hydrofluoric acid and 8-hydroxyquinoline to avoid interferences from aluminium and magnesium, respectively. The measurements were made at 234.9 nm using a dinitrogen oxide - acetylene flame. The percentage of beryllium calculated as the average of three determinations was 0.0026%.The result from AAS was 0.0022% , the average of six determinations. Table 3. Characteristics of other reagents for the fluorimetric determination of beryllium Reagent Excitation 2-Hydroxy-3-naphthoicacid . . . . . . 380 Tetracycline . . . . . . . . . . . . 406 2-(Salicylidenearnino)benzoic acid . . . . 395 Arsenazo I . . . . . . . . . . . . 366 Chlorphosphonazo I . . . . . . . . 366 3-Hydroxy-2-naphthoicacid . . . . . . 369 3-Amino-5-sulphosalicylic acid . . . . 370 Chlorphosphonazo I11 . . . . . . . . 570 2-Quinizarinsulphonicacid . . . . . . 475 Arsenazo I11 .. . . . . . . . . 580 2-Ethyl-5-hydroxy-7-methoxyisoflavone . . 366 Leucoquinizarin 450 . . . . . . . . . . Emission 460 506 485 565 565 459 465 650 650 575 430 476 Apparent optimum pH 3-4 6.2-7.5 6.6 7.3-7.8 6.5-7.3 6.2-6.3 7.0 - AcOH, 5% 9.0 4.5-4.7 Linear fluorescence range, p.p.b. 150-220 100-300 2-200 4-32 4-32 10-100 5- 100 36-320 36-320 1-7 0.4-12 5-100 Reference 9 10 11 12 12 13 14 15 15 16 171296 ANALYST, NOVEMBER 1986, VOL. 111 Laslo Erdey, L., “Gravimetric Analysis (Part II),” Pergamon Press, Oxford, 1965. Yoe, J. H., and Jones, A. L., Ind. Eng. Chem., Anal. Ed., 1944, 16, 111. Job, P., Ann. Chim. (Paris), 1928, 9, 113. Asmus, E. A , , Anal. Chem., 1960, 178, 104. JimCnez, J. C., Muiioz Leyva, J. A., and Roman Ceba, M., Anal. Chim. Acta, 1977, 90, 223.Sandell, E. B., “Colorimetric Determination of Traces of Metals,’’ Third Edition, Interscience, New York, 1959, p. 443. Merrill, J. R., Honda, M., and Arnold, J. R., Anal. Chem., 1960, 11, 1420. Cherkesov, A. I., and Zhigalkina, T. S., Zavod. Lab., 1961, 27, 658. Alykova, T. V., and Cherkesov, A. I., Izv. Vyssh. Uchebn. Zaved. Khim. Khim. Tekhnol., 1972, 15, 1107. Morishige, K., Kinki Daigaku Rik. Kenkyu Hokoku, 1982,17, 411. Savvin, S. B., and Danilin, E. S . , Zh. Anal. Khim., 1980, 35, 457. Gladilivich, D., and Grigor’ev, N., Zh. Anal. Khim., 1978,38, 2113. Alykov, N. M., and Cherkesov, A. I., Zh. Anal. Khim., 1976, 31, 1104. Savvin, S. B., and Chernova, R. K., Zh. Anal. Khim., 1976,31, 269. Guiradm, A., and Vilchez, J. L., Quim. Anal., 1975, 29, 265.Murata, A., and Tominaga, M., Bunseki Kagaku, 1974, 23, 1349. Comparison of Leucoquinizarin with Other Reagents for Beryllium Determination Although there are other reagents for the spectrophotometric and fluorimetric determination of beryllium, they are not numerous and the proposed method is of interest for the resolution of analytical problems in which the classical reagents for beryllium are not suitable. The applicability of leucoquinizarin as a spectrophotometric reagent is restricted by the numerous interferences, and the advantages arising from the use of the acid pH buffer and aqueous medium are also present in other reagents. However, leucoquinizarin is a good fluorimetric reagent (Table 3) and compares favourably with Arsenazo I, Chlorophosphonazo I, 2-ethyl-5- hydroxy-7-methoxyisoflavone and others, especially as an acidic medium is used; its sensitivity is similar to that of the former reagents. The selectivity of leucoquinizarin is good, the reagent having analogous interferences to those of Arsenazo I, Chlorophosphonazo I and 2-ethyl-5-hydroxy- 7-methoxyisoflavone with Al, Cr(II1) and Fe(II1) and surpass- ing 2-(salicylideneamino)benzoic acid, 3-amino-5-sulphosal- icylic acid, Chlorophosphonazo 111 and Arsenazo 111, which exhibit numerous interferences. Leucoquinizarin therefore increases the number of fluorimetric reagents available for beryllium determination and has an appreciable reliability and versatility. References 1. Bello Lbpez, M. A., Callejbn Mochbn, M., Gomez Ariza, J. L., and Guiraum PCrez, A., Analyst, 1986, 111, 429. 2. 3. 4. 5 . 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. NOTE-ReferenCe 1 is to Part 1 of this Series. Paper A6150 Received February 17th, 1986 Accepted April 25th, 1986

ISSN:0003-2654    

DOI:10.1039/AN9861101293

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, NOVEMBER 1986, VOL. 111 1297 Spectrophotometric Determination of Trace Amounts of Silver with 5-[4-(2-Methyl-3-hydroxy-5-hydroxymethyl)pyridylene]rhodanine R. Escobar Godoy and Alfonso Guiraum Perez Department of Analytical Chemistry, Faculty of Chemistry, University of Seville, 4 I0 12 Seville, Spain The synthesis of 5-[4-(2-methyl-3-hydroxy-5-hydroxymethyl)pyridylene]rhodanine (PYR), its characteristics, its reactions with various noble metal ions and its application to the selective determination of silver are described. PYR reacts with silver to form a red complex (metal to ligand ratio 1 : 2 ) having an absorption maximum at 530 nm in aqueous solution containing 8% VNof dimethylformamide at pH 8.2 (bicine - NaOH buffer solution). The molar absorptivity is 1.5 x lo4 I mol-1 cm-1 and the relative error for 3 pg ml-1 of silver (1 1 replicates) is 0.85%.The application of the proposed method to the determination of silver in drugs, silver pre-concentrate and galena is also described. Keywords: Silver determination; 5-[4-(2-methyl-3-hydroxy-5-hydroxymethyl)pyridylene]rhodanine; spectrophotometry; galena ore analysis Rhodanine and its derivatives are used for the quantitative analysis of silver(I),1-7 gold(III), palladium(II), copper(I), copper(I1) and other ions. As a continuation of our studies on the analytical applications of rhodanine derivatives,&lO the derivative of rhodanine with pyridoxal has been synthesised and characterised. The introduction of the pyridoxal group produces an increase in the dipole moment and hence the chromophoric nature of the molecule, which contributes to the reaction with silver instead of exerting only a resonance effect.A selective and sensitive method involving 5-[4-(2-methyl- 3-hydroxy-5-hydroxymethyl)pyridylene]rhodanine (PYR) is proposed for the determination of silver. The method was applied successfully to the determination of silver in drugs, in silver pre-concentrate and in galena ores. Experimental Apparatus Absorption spectra were recorded with a Perkin-Elmer Model 554 spectrophotometer. A Coleman 55 (digital) spectropho- tometer was also used for simple measurements. In all instances 1-cm glass or quartz cells were used. A Philips PW 9408 pH meter with glass - calomel electrodes was used for pH measurements. Reagents All reagents were of analytical-reagent grade unless stated otherwise.All solutions were prepared with distilled, de- ionised water. PYR solution. Dissolve 0.025 g of PYR in 100 ml of dimethylformamide. The solution should be prepared fresh daily and must be kept in a amber-glass bottle at 4°C. Standard silver(Z) solution. Dissolve silver nitrate (1.070 g) in water and standardise gravimetrically with chloride. Bicine - NaOH buffer (PH 8.2). Mix 800 ml of 0.1 M bicine [ N , N-bis(2-hydroxyethyl)glycine] solution and 11 1 ml of 0.1 M NaOH solution and dilute to 1 1 with water. Synthesis of PYR Equimolar amounts of pyridoxal hydrochloride and rhodanine (0.02 mol) in ethanol containing 2 g of sodium acetate and 5 ml of glacial acetic acid were heated at 60 "C in a water-bath for about 15 min. The orange precipitate was filtered off and washed with ethanol - water (1 + 1 V/V).The results of the elemental analysis were as follows: found, C 46.4, H 3.8, N 9.3, S 22.0%; calculated for CllH10N203S2: C 46.7, H 3.5, N 9.9, S 22.7%. Recommended Procedure Silver determination Place a standard solution containing from 0.25 to 4 pg ml-1 of silver in a calibrated flask and add 2 ml of 0.025% m/V PYR solution in dimethylformamide and 5 ml of bicine - NaOH buffer. Dilute to the mark with water and 5 min after preparation measure the absorbance of the solution at 530 nm against a blank prepared in the same way but without silver. Preparation of samples Collyrium (eye lotion). A 1-ml sample is evaporated to dryness in the presence of 5 ml of concentrated nitric acid and the residue is dissolved in 5 ml of 0.05 M nitric acid.The solution obtained is neutralised to pH =: 7 with 0.1 M ammonia solution and diluted to 100 or 250 ml with water, depending on the concentration of silver in the pharmaceutical preparation. Silver pre-concentrate. About 1 g of sample is dissolved in concentrated nitric acid and evaporated to dryness. The residue is dissolved in 10 ml of concentrated nitric plus 5 ml of perchloric acid and the solution is evaporated to dryness. The residue is dissolved in 15 ml of 0.05 M nitric acid and a slight excess of ammonia solution (1 + 1 V / V ) is added, followed by boiling for 5 min. After filtration to remove iron hydroxide the filtrate is diluted to 100 ml with water. Galena ore. About 3 g of sample, containing 84.5% of Pb, is evaporated to dryness twice in the presence of 30 ml of concentrated nitric acid.The solution is made alkaline with sodium hydroxide solution, then 15 ml of saturated sodium carbonate solution and 25 ml of 30% m/V sodium hydroxide solution are added, followed by boiling for 20 min. The black precipitate of Ag20 formed is cooled and filtered off using a paper filter of small pore size. This precipitate is dissolved in 15 ml of concentrated nitric acid and the solution is evaporated nearly to dryness. The residue is extracted with 10 ml of 0.05 M nitric acid, neutralised with 0.1 M ammonia solution to pH = 7 and diluted with water to 25 ml. Analysis of sample solutions Volumes of 1 ml of the above sample solutions are taken for the determination of silver using the PYR method.1298 ANALYST, NOVEMBER 1986, VOL.111 Results and Discussion Properties and Characteristics of PYR The infrared spectrum of PYR was measured with a potassium bromide disc in order to confirm its structure. The spectrum had absorption peaks at 3240 and 3135 cm-1, which were assigned to ~ ( 0 - H ) and v(N-H) stretching vibrations, at 1690 cm-l, assigned to v(C=O), at 1610-1570 cm-1 (three peaks), corresponding to the pyridine ring, at 1415 cm-1, assigned to v(C=S), and at 1250-1190 cm-1, bands characteristic of the rhodanine ring. From these results and those of elemental analysis, the synthesised PYR was presumed to have the structure shown. CHZOH OH CH3 The dissociation constants were determined spectropho- tometrically.The pK values of the first, second and third ionisation steps were calculated from the variation of the absorbance with pH at different wavelengths by the Phillips and Merritll and Sommerl2 methods (99% water, 0.1% dimethylformamide, y = 0.1). The mean values found were 2.95, 5.95 and 8.75, respectively. The first pK value can be attributed to the protonated pyridine ring of pyridoxal, the second value corresponds to the equilibrium protonation of the rhodanine ring and the third value may be assigned to the deprotonation of the phenolic group of the pyridoxal. These values are in accord with those in the literature for rhodanine and pyridoxal derivatives. 13-15 We conclude that the dissociation equilibrium proceeds as shown. LHS’ H+ LH2 CH CH II I ‘c=s C-NH/ c-s II 0 II C-NH’ II c-s I ‘C=S 0 L2- LH - Absorption Spectra and Effect of pH The absorption spectrum of the red complex formed between silver(1) and PYR in aqueous solution containing 8% V/V of dimethylformamide has an absorption maximum at 530 nm, as can been seen in Fig.1. The complex is stable for at least 2 h. A study of the effect of pH on the complexation of PYR with silver(1) shows that the complex is formed over a wide range of pH values, but the maximum absorbance was obtained in the pH range 7.30-9.10 (Fig. 2). The complex cannot be measured in the pH range 4-6.5, possibly owing to the insolubility of the reagent as a result of the formation of zwitterions in the pyridoxal group of the molecule. Effect of Reagent Concentration A concentration of PYR three times greater than the silver concentration (molar ratio) was necessary for complete complexation.A greater excess of reagent decreases the absorbance. Composition of the Complex Job’s method of continuous variation16 showed that the complex is formed with metal to ligand ratios of 1 : 1 and 1 : 2. The former complex is formed with a deficit of reagent and we observed that this complex absorbs at the highest wavelength. The 1 : 2 complex is formed with an excess of the reagent. This is in agreement with the observations of Borissova et a1.6 Calibration Graph, Sensitivity and Precision Straight-line calibration graphs passing through the origin were obtained using the recommended procedure. The equation of the line obtained by a least-squares treatment was Ag (p.p.m.) = 7.46 A (at 530 nm) where A is the absorbance. The optimum range for the determination of silver is between 0.25 and 4 yg ml-l and the molar absorptivity is 1.5 x l o 4 1 mol-1 cm-1.Eleven standard solutions containing 3 yg ml-1 of silver were analysed by the recommended procedure. The precision was estimated from 11 results for 25-ml aliquots containing 3.0 pg ml-1 of silver(1). The results give a mean absorbance of 0.402 with a standard deviation of the mean of 0.0015 absorbance unit and a relative error of 0.85%. Effect of Foreign Ions Solutions containing various amounts of foreign ions and a fixed amount of silver (2 pg ml-1) were subjected to the described procedure. The tolerable amounts of each ion giving a maximum error of k3Y0 in the determination are sum- marked in Table 2.It can be seen that the noble metals, bromide, iodide and thiosulphate anions are the most impor- tant interferents. I CH c-s C-N’ 0 I K3 II c-s\ I ,c-s- II II II 7 CH I ‘c-s- C-N 0 Reactivity of PYR with Metal Ions The reactions of 40 cations with PYR were tested at different pH values. The most sensitive were silver(I), palladium(II), mercury(1) and -(II) and copper(1) and -(II). The most interesting results are summarised in Table 1. Application to Real Samples In order to confirm the usefulness of the proposed method, it was applied to the determination of silver in drugs, in a silver pre-concentrate and in a galena ore. The results obtained are summarised in Table 3, in comparison with those given by a standard procedure that uses p-dimethylaminobenzylidene- rhodanine.The proposed method has the advantage over other methods that use rhodanine derivatives of forming true solutions. We were able to carry out a systematic optimisation of the experimental conditions, which was not possible with other rhodanines proposed as spectrophotometric reagents forANALYST, NOVEMBER 1986, VOL. 111 1299 Table 1. Reactivity of PYR with some noble metal ions pH 2.2 Metal ion A,,,. &/lo4 1 mol-1 cm-1 AgI . . . . 520 0.86 - - CU" . . . , c u r . . . . HgII . . . . HgI . . . . Pd" . . . . 510 0.74 Au"' . . . . o~v"' - - - - - - - - - - . . . . pH 4.9 pH 8.2 Amax. &/lo41 mol-1 cm-1 A,,,. &/lo41 mol-1 cm-1 510 0.25 - - 510 0.69 - - 470 3.40 - - 490 1.10 - 520 2.3 5 10 1.2 5 10 0.60 490 0.60 - 530 1.50 - - - - - - Wavelengthhm Fig.1. reagent blank; and B, reagent vs. water blank A, Absorption spectra of PYR and its silver complex vs. Table 2. Tolerance limits of ions in the determination of 2 p.p.m. of silver Ions C104-, NO3-, S042-, F-, C032-, PO4'-, EDTA . . . . . . . . . . . . . . Alkali metals, alkaline earths, AllI' . . . . . . . . Mn", Nirr, CdrI,* Pb",* tartrate, and oxalate . . CoII, Zn", VV, TiIV, FexI, MoVI, U072+, - . so32-, c1- . . . . . . . . . . HgII, Au"1 . . . . . . . . . . . . . . OsvIIr, BPI, CuII, Br- . . . , . . . . . , Pd", I-, s 2 0 3 * - . . . . . . . . . . . . * With 1 ml of 0.01 M EDTA solution. Tolerance limit, p.p.m. 1 .ooo 100 50 10 5 2 silver17-'9, as they only form chelates in colloida] solutions with low stability and under very strict conditions (pH, nature and composition of the solvent, addition of protective colloid, reagent concentration).Further advantages of the PYR method are that it is direct, simple, rapid and reproducible and is subject to few interferences. The authors thank Prof. Dr. D. Francisco Pino for his valuable collaboration and Dr. J. L. Puente Hernaiz, Metallurgical Chief, "La Cruz", Linares ( J a h ) , for providing the silver pre-concentrate and the galena ore samples. References 1. Feigl, F., 2. Anal. Chern., 1928, 74, 380. 2. Feigl, F., "Chemistry of Specific, Selective and Sensitive Reactions," Academic Press, New York, 1949. 0.4 e, C $ 0.3 0 s 0.2 , Fig. 2. Effect of pH on absorbance of the complex. Silver, 2.7 X 10-5 M; PYR, 1.0 x 10-4 M; wavelength, 530 nm; reference, reagent blank.- - - Indicates the pH range in which the absorbance could not be measured owing to the formation of a precipitate Table 3. Determination of silver in real samples Silver found, % PYR pDimethylaminobenzy1- Sample method idenerhodanine method Collyrium . . . . . . . . 1.77 1.75 Silverpre-concentrate . . . . 0.94 0.95 Galena ore . . . . . . . . 0.002 0.0019 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. Sheppard, S. E., and Brigham, H. R., J. Am. Chem. SOC., 1936, 58, 1046. Pedley, E., Anal. Chirn. Acta, 1952, 7, 387. Stephen, W. I., and Townshend, A., Anal. Chirn. Acta, 1965, 33, 257. Borissova, R., Koeva, M., and Topalova, E . , Talanta, 1975, 22, 797. Snell, F. D., "Photometric and Fluorometric Methods of Analysis," Wiley-Interscience, New York, 1978. Galan Alfonso, G., Thesis Doctoral, University of Seville, 1983. Galan Alfonso, F., Gomez Ariza, J. L., and Guiraum Perez, A., Anal. Lett., 1984, 17 (A13), 1447. Gonzalez Gonzalez, G., Gomez Ariza, J. L., and Pino Perez, F., An. SOC. ESP. Fis. Quirn., 1985, 81,208. Phillips, J. P., and Merrit, L. L., Jr., J. Am. Chem. Soc., 1948, 70, 410. Sommer, L., Folia Fac. Sci. Nut. Univ. Purk., Brno, 1964,5,1. Galan Alfonso, G., and Gomez Ariza, J. L., Microchern. J., 1981,26, 574. Lunn, A. K., and Morton, R. A., Analyst, 1952, 77,718. Umbreit, W. W., O'Kane, D. J., and Gunsalus, I. C., J. Biol. Chern., 1948, 176,629. Job, P., Ann. Chirn. (Paris), 1928, 8, 113. Sandell, E. B., and Neumayer, J. J., Anal. Chim. Acta, 1951,5, 445. Cave, G. C., and Hume, D. M., Anal. Chem., 1952,24,1503. Sandell, E. B . , "Colorimetric Determination of Traces of Metals," Interscience, New York, 1959. Papers A51261 Received July 17th, 1985 Accepted May 16th, 1985

ISSN:0003-2654    

DOI:10.1039/AN9861101297

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, NOVEMBER 1986, VOL. 111 1301 Anisaldehyde-4-phenyI-3-thiosemicarbazone as an Analytical Reagent for the Extractive Spectrophotometric Determination of Gold Kinthada M. M. S. Prakash", L. D. Prabhakar and D. Venkata Reddy Department of Chemistry, Sri Krishnadevara ya University, Anantapur 515 003, India The synthesis, spectral characteristics and analytical applications of anisaldehyde-4-pheny1-3- thiosemicarbazone (APT) are described. A simple, rapid, selective and sensitive spectrophotometric method for the determination of microgram amounts of gold, alone or in the presence of associated metals, is developed, based on the colour reaction between the metal ion and anisaldehyde-4-phenyI-3-thiosemicarbaz- one. The yellow - brown complex (Amax. = 365 nm) is extracted into ethyl acetate.Gold(ll1) reacts with the reagent in the ratio 1 : 1 (metal to ligand) over the pH range 4.0-7.0. Beer's law is obeyed over the concentration range 0.1-12.3 yg ml-1 of gold. The molar absorptivity and Sandell's sensitivity of the method are 2.12 x lo4 I mol-1 cm-1 and 0.0092 pg cm-2, respectively. The relative standard deviation for ten replicate determinations of 45 yg of gold(ll1) was 1.71%. The interference of various ions has been studied and conditions were developed for the determination of gold in some synthetic samples. Keywords: Gold determination; anisaldeh yde-4-phen yl-3-thiosemicarbazone; spectrophotometry; extraction Only a few phenyl thiosemicarbazones have been synthesised and used as chromogenic reagents for the spectrophotometric determination of metal ions.Renewed interest in the synthesis of phenyl thiosemicarbazones~-7 is due to their sensitive reactions with metal ions and the easy extractability of metal - phenyl thiosemicarbazone complexes into organic solvents. Gold is generally separated by the extraction of chlorauric and bromauric acids using oxygen-containing solvents such as isobutyl methyl ketone (IBMK), ethyl acetate, amyl acetate, mesityl oxide , diisopropyl , diethyl or dichloroethyl ether, tri-n-butyl phosphate (TBP) or tri-n-octyl phosphate oxide (TOPO) solutions in cyclohexanone. Gold may also be determined using various complexing agents such as diethyl- dithiocarbamate,s 8-rnercaptoquinoline79 2-quinolylaldox- ime,lO thioamides,ll aniline,l* ethyl xanthate,l3 2-pyridine aldoxime,14 anthranilic acid,ls di-2-thienyl ketoximel6 or chromophryozole.17 In most of the methods, the sensitivity is very poor and the colour fades after a few minutes.In some instances the complex is formed only after heating for a long period, whereas others suffer from interferences from other metal ions. A thorough survey of the literature showed that no previous attempt has been made to employ thiosemicarb- azones for the spectrophotometric determination of gold. Hence , in this paper, anisaldehyde 4-phenyl-3-thiosemicar- bazone (APT) is proposed as a reagent for the rapid extraction and direct spectrophotometric determination of gold(II1). Various parameters such as pH, reagent concentration, equilibration time and interference of foreign ions have been studied. This method has been applied to the determination of gold(II1) in certain synthetic samples.The reagent has been found to be sensitive and selective compared with other complexing agents. Experimental Reagents All chemicals were of analytical-reagent grade (BDH Chemi- cals). Synthesis of APT. A solution prepared by dissolving 0.835 g * To whom correspondence should be addressed. of 4-phenyl-3-thiosemicarbazide in 2 ml of glacial acetic acid was added dropwise to a solution of anisaldehyde (0.68 g) in 50 ml of methanol, with stirring. A white product quickly separated and this was recrystallised from methanol or carbon tetrachloride (m.p. 182-184 "C, yield 75%). Colour reagent. A 0.285 g amount of APT was dissolved in 50 ml of dimethylformamide (DMF) (2 X M).Solutions of lower concentrations were obtained by dilution with DMF. M. A stock solution of gold(II1) was prepared by dissolving the requisite amount of gold(II1) chloride in distilled water and adding hydrochloric acid in order to make the solution 1 M with respect to the acid. The solution was standardised gravimetrically by the hydro- quinone method.18 Buffer solutions. Buffer solutions of various pH values were prepared using 1 M HC1- 1 M sodium acetate (pH 1-3), 0.2 M acetic acid - 0.2 M sodium acetate (pH 4-7) and 0.2 M ammonium chloride - ammonia solution (pH 8-12). Standard gold(ZZ1) solution, Apparatus The pH was measured with an ELICO Model LI-120 pH meter. IR and NMR spectra were recorded with Perkin-Elmer 257 and EM-390 MHz spectrophotometers, respectively.C, H analyses were carried out with a Coleman CH Analyser. A Beckman DU 2 spectrophotometer was used for absorbance measurements. Procedure Complexes of APT with metal ions To 5 ml of buffer solution of the desired pH (1-12), taken in a 25-ml separating funnel, 1 ml of solution containing 3-8 p.p.m. of the metal ion and 1 ml of 1.5 X 10-3 M APT solution were added. The aqueous phase was shaken for 1 min with 10 ml of suitable solvent (solvents were selected in which the metal complexes showed the maximum absorbance). After shaking, the phases were allowed to settle and the aqueous layer was separated. The organic layer was dried with anhydrous sodium sulphate and the absorbance spectra measured in the wavelength region 300-500 nm against the corresponding reagent blanks prepared under identical con- ditions,1302 ANALYST, NOVEMBER 1986, VOL.111 Recommended procedure for the determination of gold To an aliquot of the sample solution [containing 1.25 x M, 10-100 pg Au(III)] in a 25-ml separating funnel, add 5 ml of buffer (sodium acetate - acetic acid) solution of pH 5 and 1 ml of 1.25 X 10-3 M reagent solution. Shake the aqueous phase for 1 min with 10 ml of ethyl acetate. Allow the layers to separate and discard the aqueous phase. Dry the organic layer with anhydrous sodium sulphate and measure the absorbance of the Au - APT complex at 365 nm against a reagent blank prepared in an identical manner. The calibration graph is prepared as described below. To 5 ml volumes of buffer solution (pH 5.0), taken in different 25-mi separating funnels, are added known aliquots of standard gold(II1) (1.25 x 10-4 M) solution and 1 ml of APT (1.25 X 10-3 M) solution.The total volume of the aqueous phase in each instance is adjusted to 10 ml with distilled water and shaken with 10 ml of ethyl acetate for 1 min. The organic layer is separated, dried over anhydrous sodium sulphate and the absorbance measured at 365 nm against a corresponding reagent blank. A plot drawn of the volume of the metal ion versus absorbance gives a straight line passing through the origin. The gold concentration of the sample is deduced from the calibration graph. Preparation of solid Au - APT complex To 30 ml of gold(II1) solution (5 x 10-2 M) were added 30 ml of 2 X 10-4 M APT in methanol and the mixture was refluxed for about 1 h in methanolic medium.The resulting solution was concentrated by distilling the excess of solvent under reduced pressure and the yellow solid that separated was filtered and washed with water. The precipitate was dried under vacuum. Results and Discussion Properties and Characteristics of Anisaldehyde-4-phenyl-3- thiosemicarbazone Identification of the synthesised APT was carried out by taking the results of its elemental analysis and the IR and NMR spectra into consideration. The results of the elemental analysis are given in Table 1. The infrared spectra, measured in potassium bromide pellets, showed absorption peaks that were assigned to the stretching vibrations of an azomethine bond (:C=N-) at 1610 and 1540 cm-l.The peaks at 1070, 1025, 835 and 740 cm-1 are assigned to the C=S bond, and a band corresponding to -NH is observed at 3330 cm-1. In Table 1. Structure and physical properties of APT H C=N-NH- C -NH-C& It CH30 S addition the nuclear magnetic resonance spectra, measured in dimethyl sulphoxide, also showed a peak assigned to an azomethine bond at 8.1 p.p.m. A peak at 9.8 p.p.m., which is attributed to the -NH group, further indicates that the ligand is in the thione form. On the basis of these results, the APT synthesised is presumed to have the structure shown in Table 1. The solubility of APT in various solvents is shown in Table 2. Of the solvents tested, DMF was the best. The method of Phillips and Merrittlg was used for the determination of the ionisation constant in DMF medium (2 ml of 1 X M APT solution in DMF, diluted to 25 ml with distilled water).The pK value (Fig. 1) found was 8.6. The pK value shown is an arithmetic mean of values obtained from measurements at four different wavelengths. Reactivity of APT The reactivity of APT with various metal ions at different pHs are summarised in Table 3. APT is found to react with copper(II), nickel(II), cobalt(II), platinum(IV), palla- dium(II), osmium(VIII), gold(III), ruthenium(III), silver(1) and mercury(I1) to give coloured complexes with high molar absorptivities. The table suggested the usefulness of the synthesised APT for the determination of gold, and therefore the Au - APT complex was examined in subsequent studies. Selection of the Extractant The Au - APT complex can be extracted into organic solvents such as chloroform, amyl alcohol, butanol, ethyl acetate, benzene and IBMK.As the complex showed a maximum absorbance in ethyl acetate, this solvent was selected for the extraction spectrophotometric studies. Absorption Spectra The absorption spectrum of Au - APT shows a maximum absorption at 365 nm, where the reagent shows comparatively low absorbance (Fig. 2). Hence in all instances the absorption was measured at 365 nm against a corresponding reagent blank. Effect of pH The determination of gold(II1) was studied over the pH range 1.g12.0. The metal content of the organic phase was determined from the absorbance of the coloured complex using the previously constructed calibration graph at pH 5. The optimum pH range for constant, maximum percentage Property Value Molecular formula c 15 N3 H15 0s Melting-poin t 182-184 "C Yield 75 yo pK value 8.6 Elemental analysis, % m/m: C 63.38 (63.08)* H 5.316 (5.257)* S 11.360 (11.250)* * Figures in parentheses indicate calculated values.Table 2. Solubility of APT Solvent Solubility of APT/g 1-1 Dimethylformamide . . Nitrobenzene . . . . . . Butanol . . . . . . . . Methanol . . . . . . Benzene . . . . . . Acetone . . . . . . . . Water . . . . . . . . Carbon tetrachloride . . >165 > 20 8.70 1.75 2.90 3.00 Insoluble InsolubleANALYST, NOVEMBER 1986, VOL. 111 1303 10.0 1 A ' - \ A 2 4.0 ':I I I , , , I , , I I , 1 .o 1 2 3 4 5 6 7 8 9 1 0 1 1 PH Fig. 1. 400 nm Determination of pK value. A, 325; B, 340; C, 320; and D, Table 3.Characteristics of anisaldehyde-4-phenyl-3-thiosemicarb- azone complexes of metal ions ~ r n a x . J Ion pH nm ~ / 1 mol- cm-1 Solvent Co(I1) . . . . 5 370 6000 Butanol Co(I1) . . . . 10 370 15000 Butanol Ni(I1) . . . . 10 375 16000 Benzene Cu(I1) . . . . 5 365 45000 IBMK Cu(I1) . . . . 9 370 44000 IBMK Au(II1) . . 5 365 21200 Ethyl acetate Pt(1V) . . . . 1 370 15000 Ethyl acetate Pt(1V) . . . . 7 360 7000 Ethyl acetate Pd(I1) . . . . 4 375 19000 Ethyl acetate Os(VII1) . . 5 360 9000 IBMK Ru(II1) . . . . 2 365 5000 Ethyl acetate Ag(1) . . . . 6 365 10000 Ethyl acetate Hg(I1) . . . . 4 367 12000 Benzene Hg(I1) . . . . 8 370 12000 Benzene 1 .o 0.8 0, 5 0.6 + s n a 0.4 0.2 350 360 370 380 390 0 Wavelengthhm Fig. 2. Absor tion spectra of: A, APT; and B, Au - APT.Conditions: AuKII), 5 x 10-4 M; and APT, 5 X M ~ ~~ Table 5. Determination of gold in synthetic gold - copper - silver samples Composition Au-Cu-Ag(50 + 20 + 30) Au-Cu-Ag(60+ 35 + 5 ) . . AU - CU - Ag (20 + 70 + 10) Amount of gold taken/ ygml-l . , 2.46 3.68 4.92 5.16 . . 2.96 4.42 5.90 6.48 . . 2.44 3.06 3.68 4.90 * Average of five determinations. Amount of gold found*/ yg m1-1 2.44 3.64 4.94 5.12 2.94 4.40 5.92 6.52 2.46 3.08 3.64 4.92 Table 4. Effect of foreign ions. Conditions: [Au(III)] = 50.0 yg; [APT] = 5 X M; pH = 5 Ions EDTA, C103-, P043-, Sr2+ . . . . . . . . . . Tartrate2-, Pb2+, Ca2+ Br0,-, Mn2+, AP+, Cr6+', Hg2;; K+', *W6+ ,'Mgi', SO>-, so32- 7 B4072-, Pt4+, c1- . . . , . . . * . . NO3-, CH3COO-, Cd2+, Zn2+ . . . . . , . . Citrate2-, F-, V5+ . .. . . . . . . . . . Ce4+, Sn2+, Br-, malonate . . . . , , . . . . Ni2+ * . . . . . . . . . . . . . . . . , . CU2+,t co3+ . . . . . . . . . . . . . . . . Pd2+$ . . . . . . . . . . . . . . . . . . Os*+,SCN- . . . . . . . . . . . . . . u6+ Zr4+ , s 2 0 3 2 - 9 & 0 d 2 - . . . . . . . . . . * Masked with 1000 p.p.m. of tartrate. t Masked with 2500 p.p.m. of EDTA. $ Masked with 1160 p.p.m. of DMG. . . . . . . . . . . . . . . . . Mo6+ . . . . . . . . . . . . . . . . . . . . . . . . . . . . * . . . . . . . . . . . . , . . * . . . . . . . . . . . . , , . . . . . Tolerance limit/ L% 50 000 20 000 10 000 7 500 5 500 3 500 2 300 1000 400 320 150 Interfere seriously extraction of the complex was found to be 4.0-7.0. Hence sodium acetate - acetic acid buffer solution of pH 5 was selected for further studies. Effect of Reagent At optimum pH, absorbance measurements were carried out by varying the reagent concentrations.It was observed that the addition of a 10-fold excess of reagent was sufficient for maximum and constant absorbance. Effect of Order of Addition of Constituents After fixing all the other parameters, a few further experi- ments were carried out to ascertain the influence of the order of addition of reagents on the determination of gold. The order of addition of constituents of the reaction mixture has no effect on the absorbance. Equilibration Time and Stability of the Complex A single extraction for 60 s with 10 ml of APT was sufficient for quantitative extraction and prolonged shaking had no1304 ANALYST, NOVEMBER 1986, VOL.11 1 Table 6. Determination of gold in synthetic mixtures corresponding to alloys Alloy Au-Pdalloy . . . Alloy for electrical contacts . . . Low melting-point dental magnetic alloy . . . . Amount of Amount of Synthetic gold taken/ gold found*/ composition, % pg ml-1 pg ml-1 Au 50.0 2.46 2.42 Pd 50.0 3.68 3.72 4.92 4.96 . Pd35.0 1.48 1.50 Ag 30.0 1.96 1.98 Pt 10.0 2.46 2.48 Cu 14.0 Au 10.0 Zn 1.0 Pd 34.0 1.48 1.50 Ni 24.0 1.96 1.98 c o 22.0 2.46 2.48 Au 10.0 * Average of five determinations. Table 7. IR spectral and conductance data of the solid Au - APT complex Spectral bands in the Infrared spectra Ligand Complex region(v/cm-l) . . . . . . 3300 3300 NH 1610(~) 1600(~) C=N2'J' 1510(s) 1510(s) v(CN) 1075(s) 1065(ms) (C=S)22 - 430(~) Au-S23 - 370(~) Au-CP Molar conductivity (in DMF) (ohm-' cm-2 mol-1) 40 Non-ionic nature of complex 5 A B Fig.3. Structures of: A, APT; and B, Au - APT complex further effect on the extraction. The absorbance of the Au - APT complex was stable for 4 h. Composition of the Complex The composition of the Au - APT complex was established by Job's method of continuous variation and molar ratio meth- od. In both instances the gold to reagent ratio was found to be 1 : 1. Beer's Law and Sensitivity A calibration graph was constructed under the optimum conditions described above. The system obeys Beer's law over the concentration range 0.11-12.3 pg ml-1 of gold. The molar absorptivity and Sandell's sensitivity of the method are 2.12 X l o 4 1 mol-1 cm-1 and 0.0092 pg cm-2, respectively.The standard deviation for ten replicate determinations of 50 pg of gold was 0.020 pg. Effect of Foreign Ions Various amounts of foreign ions were added to a fixed amount of gold (5 pg ml-1) and the recommended procedure for the extraction and spectrophotometric determination was fol- lowed. A maximum error of 2% in the absorbance reading was considered tolerable. The tolerance limit of foreign ions is given in Table 4. Amongst the anions examined, only thiocyanate interfered seriously. Among the cations exam- ined, Pd4+, Os8+, Cu2+ and Ni2+ interfered. The interference from Ni2+, Cu2+ and Pd4+ can be eliminated by masking with tartrate, EDTA and dimethylglyoxime, respectively. Thus the data in Table 4 indicate the reasonable selectivity of the method in the presence of associated ions.Application to Synthetic Samples In order to confirm the usefulness of the proposed extraction spectrophotometric method, it was applied to the determina- tion of gold(II1) in certain synthetic samples corresponding to alloys (Tables 5 and 6). Tartrate (1000 p.p.m.), EDTA (2500 p.p.m.) and DMG (1160 p.p.m.) were added to mask Ni(II), Cu(I1) and Pd(II), respectively. Ag(1) was separated by precipitation as AgCl. IR Spectral Investigation of Solid Au - APT Complex The IR spectral data shown in Table 7 allow some conclusions to be drawn about the bonding sites of Au(II1) with APT. The band at 1610 cm-1 is shifted towards lower frequency (1600 cm-I), indicating the participation of the azomethine nitrogen atom in coordination.20221 The band at 1075 cm-1 due to the C=S bond in the ligand is lowered to 1065 cm-1 in the complex, suggesting the sulphur coordination of the ligand.22 The absence of a band due to the S-H stretching mode near 2570 cm-l suggests that in the solid state the molecule remains in the thioketo form.The far-IR spectrum shows bands at 430 and 370 cm-1 corresponding to Au-S and Au-Cl bands, respectively.23 The molar conductivity (40 ohm-1 cm-2 mol-1) in DMF indicates that the complex behaves as a non-electrolyte. Based on the infrared spectral investigation, the structure of the Au - APT solid complex is assumed to be as shown in Fig. 3. Conclusions The reagent can be easily synthesised by the direct condensa- tion of anisaldehyde and 4-phenyl-3-thiosemicarbazone in acetic acid medium.The reagent APT has been compared with some well known reagents for the spectrophotometric determination of gold (Table 8). Rhodamine B37 is a highly sensitive reagent for gold determination but involves many difficulties. The colour of the complex is stable for only 30 min and as the efficiency of extraction depends on the concentration of hydrochloric acid in solution, the concentration in both sample and standard solutions must be kept constant (at ca. 0.5 M HC1). The initial isolation of trace amounts of gold by coprecipitation with tellurium is required to overcome the interferences due to associated ions. Rhodamine spectrophotometric methods for gold currently in use suffer from the drawbacks of the limited solubility of the reagent, sensitivity to the time of colour development and to the acidity of reaction medium and also the instability of the reaction product.The proposed method for the spectrophotometric determi- nation of gold is sensitive, simple and rapid. It does not require heating and is virtually free from interferences from many associated foreign ions. The colour development is instantaneous at room temperature (27 "C) and the whole process can be completed within 15 min.ANALYST, NOVEMBER 1986, VOL. 111 1305 Table 8. Comparison of reagents for the spectrophotometric determination of gold L a x . / €1 Reagent nm Trifluoroethyl xant hate . . . . . . . . . . 452 5-(p-Ethoxyanilino)-5,6-dihydrouracil Azide . . . . . . . . . . . . Benzoic acid hydrazide . . . . . . Thiocaprolact am . . . . .. . . Isonicotinic acid hydrazide . . . . Nitron . . . . . . . . . . . . Diantipyrylmethane . Triisooct ylamine . . . . . . . . Aniline . . , , . . . . . . . . Triflupromazine hydrochloride . . 4-(2-Thiazolylazo)resorcinol . . . . Promethazine hydrochloride . . . . . . . . 323 . . . . 330 . . . . 520 . . . . 400 . . . . 520 . . . . 405 . . . . 336 . . . . 325 . . . . 560 . . . . 503 . . . . 520 . . . . 517 2,2‘-Dipyridyl ketoxime . . . . . . . . . . 450 Prochlorpromazine maleate . . . . . . . . 503 Mepazine hydrochloride . . . . . . . . . . 514 Tetraphenylarsonium chloride . . . . . . . . 323 Di t hizone . . . . . . . . . . . . . . 420 Rhodanine . . . . . . . . . . . . . . 562 Rhodanine B . . . . . . . . . . . . . . 555 4,4’-Bis(dimethy1amino)thiobenzophenone . .540 Anisaldehyde-4-phenyl-3-thiosemicarbazone . . 365 1 mol-1 cm-1 Medium Comments Interferences 1.09 x 103 1.30 x 103 1.32 x 103 3.065 x 103 3.70 x 103 3.88 x 103 4.0 x 103 5.5 x 103 5.8 x 103 6.2 x 103 1.303 x 104 1.470 X 104 1.609 X lo4 2.0 x 104 2.05 x 104 2.18 x 104 2.42 x 104 2.80 x 104 3.20 x 104 9.70 x 104 1.20 x 105 2.12 x 104 Aqueous 15 rnin required for colour development Butanol Colour stable Aqueous - - for 15 rnin - Aqueous - Cu(I1) CHC13 - - Aqueous - - Dichloroethane - Ga(III), Fe(III), CHC13 cc1, - Aqueous Aqueous Aqueous Aqueous Aqueous Aqueous CHC13 CC14 Aqueous Diisoprop yl ether Aqueous Ethyl acetate Sb(1IIj ’ ’ Shaking time 5 min - - - - Colour stable Ce(IV), Cr(VI), for 30 rnin Shaking for 5 rnin - 8 rnin required Ce(VI), Cr(VI), development 10 min required for colour development - 10 rnin required for colour development and colour stable for 30 rnin - Colour stable for 20 min - pH and chloride content of solution critical - Colour stable for 30 min.Heat for 15 rnin at 50-60°C - Colour stable for 30 min - 30 rnin required for colour development - Mn(II) 7 V(V> for colour M a ) , V(V> - - Reference 24 25 26 27 28 29 30 31 32 12 33 34 33 35 36 33 33 37 37 37 38 This work Thanks are due to the Council of Scientific and Industrial Research, New Delhi, for awarding Senior Research Fellow- ships to two of the authors (K. M. M. s. P. and L. D. P.). 1. 2. 3. 4. 5. 6. 7. 8. References Gomez Ariza, J. L., Can0 Pavon, J. M., and Pino, F., Talanta, 1976, 23, 460. Martinez Aguilar, M. T., Cano Pavon, J.M., and Pino, F., Anal. Chim. Acta, 1977, 90, 335. Can0 Pavon, J. M., Sanchez, J. C. J . , and Pino, F., Anal. Chim. Acta, 1975, 75, 335. Bhatt, Y. N., Patel, K. K., Shah, K. J., and Patel, R. S . , J. Indian Chem. SOC., 1975, 52, 1214. Balairon Gonzalez, M., Cano Pavon, J. M., and Pino, F., Talanta, 1976, 26, 71. Bautista, J. M., and Can0 Pavon, J. M., Talanta, 1980,27,923. Rodriquez, J . , Garcia De Torres, A., and Can0 Pavon, J. M., Talanta, 1981, 28, 131. De, A. K., Khopkar, S. M., and Chalmers, R. A., “Solvent Extraction of Metals,” Van Nostrand Reinhold, New York, 1970, p. 143. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20 * 21. 22. Suprunovish, V. I., and Shevchenko, Yu. I., Zh. Anal. Khim., 1979, 34, 1738. Dutta, N. K., and Dhar, S. N., J. Znst. Chem. Calcutta, 1978, 50, 83.Radhusev, A. V., and Golomolzin, B. V., Zh. Anal. Khim., 1979,34,742. Rzeszutko, W., and Kopec, T., Fresenius 2. Anal. Chem., 1977, 285, 125. Donaldson, E. M., Talanta, 1976, 23, 411. Gagliardi, E., and Presinga, P., Mikrochim. Ichnoanal. Acta, 1965, 791. Makovsch, M. E., Talanta, 1969, 16, 443. Holland, W. J., and Gerard, J.,Anal. Chim. Acta, 1968,43,71. Busev, A. I., Simonova, L. N., Misharina, T. A., and Zayukova, N. D., Zh. Anal. Khim., 1972,27,298. Vogel, A. I . , “A Text Book of Quantitative Inorganic Analysis,” Third Edition, Longmans, London, 1969. Phillips, J. P., and Merritt, L., Jr., J. Am. Chem. SOC., 1948, 70, 410. Sharma, B. D., and Bailer, J. C., J. Am. Chem. SOC., 1955,77, 5476. Dutt, N. K., and Chakdar, N. C., J. Inorg. Nucl. Chem., 1970, 32,2303. Jain, M. C., and Jain, P. C., J. Znorg. Nucl. Chem., 1977,39, 2183.1306 ANALYST, NOVEMBER 1986, VOL. 111 23. 24. 25. 26. 27. 28. 29. 30. 31. Borissova, R., Talanta, 1975, 22, 797. Hussain, M. F., Bansal, R. K., Puri, B. K., and Satake, M., Analyst, 1984, 109, 1151. Marshall, D. A., and Meloan, C. E., Anal. Lett., 1969,2,595. Clem, R. G . , and Huffman, E. H., Anal. Chem., 1965, 37, 1155. Krishna Rao, P. V., and Sambasiva Rao, R., Curr. Sci., 1975, 44,551. Sikorska Tornicka, H., Mikrochim. Acta, 1970, 1006. Adil, A. S., Ayazi, A. A., Amjal, A. I . , and Hashmi, M . H., Mikrochirn. Acta, 1970, 606. Dimitrova, A., and Bogdanova, V., Natura (Plovdiv, Bulg.), 1973, 6 , 85. Petrov, B. I., Moskvitinova, T. B., and Galinova, K. G., Zh. Anal. Khim., 1983, 38, 1000. 32. 33. 34. 35. 36. 37. 38. Mirza, M. Y., Talanta, 1980, 27, 101. Sanke Gowda, H., and Thimmaiah, K. N., Indian J . Chem., Sect. A, 1976, 14, 632. Subrahmanyarn, B . , and Eshwar, M. C . , Anal. Chim. Acta, 1976, 82, 435. Holland, W. J . , and Bozic, J., Anal. Chem., 1967, 39, 109. Sanke Gowda, H., and Ramappa, P. G., Fresenius 2. Anal. Chem., 1976, 280, 221. Marczenko, Z., “Spectrophotometric Determination of Ele- ments,” Wiley, New York, p. 282. Sukhara, T., Talanta, 1977, 24, 633. Paper A6183 Received March 12th, 1986 Accepted May 19th, 1986

ISSN:0003-2654    

DOI:10.1039/AN9861101301

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST NOVEMBER 1986 VOL. 111 Pre-concentration and Determination Spectrophotometry Dennis P. DeSalvo” Department of Chemistry Lo yola University Chicago IL Kenneth W. Street Jr.t 1307 of Aquatic Sulphide by Visible 60626 USA Department of Chemistry Kent State University Kent OH 44240 USA The concentration of sulphide in aqueous environmental samples was determined by first pre-concentrating the samples on a cadmium(l1)-exchanged zeolite sorbent followed by conversion to methylene blue. Sulphide was finally quantified by visible spectrophotometry. The various parameters affecting the sulphide recovery of this method such as flow-rate zeolite column bed height sample size sample volume and aspiration time were investigated. A comparison of this method with the currently accepted sulphide spectrophotometric method indicated that the pre-concentration technique yields a lower limit of detection a greater certainty of results and the highest sensitivity.The application of the standard additions technique as a means to circumvent potential sample matrix problems in aqueous samples is also discussed. Keywords Sulphide determination; cadmium(l1)-exchanged zeolite; meth ylene blue; visible spectro-photometry; aquatic sulphide Until recently little was known about the concentration of sulphide in and its subsequent effects on the aquatic environment. The presence of sulphide in the aquatic environ-ment can be traced back to a variety of biological and industrial sources ranging from the anaerobic action of microorganisms on domestic sewage to oil refinery tannery and paper mill effluents.1-4 A wide variety of effects have been observed in the presence of sulphides even at low concentra-tions.Included among these are the corrosion of metal surfaces the degradation of concrete by oxidation to sulphate, reduced fish spawning rates and odour and taste nuisance effects.5-7 Various methods have been employed to determine sul-phide ion concentrations. Several “standard methods” as studied evaluated and presented by the American Public Health Association the American Water Works Association and the Water Pollution Control Federation have been developed to determine the total sulphide ion concentration in aqueous samples. The detection limits of the current titri-metric iodine (No. 427 D) and colorimetric methylene blue (No.427 C) methods are only at the p.p.m. level.8 The Ag2S solid-state ion-selective electrode approach is prone to irre-producible results from high noise levels signal drift and chemical interferences in the sample matrix.gJ0 These methods are inadequate to determine reliably the p.p.b. levels present in aqueous environmental samples. The method presented here uses a pre-concentration technique. Sulphide is pre-concentrated on zeolite ion-exchange molecular sieves. 1 1 ~ 2 Earlier work has shown that the cadmium(1I) species readily exchanges with the sodium ions on zeolite AAA and this Cd-exchanged zeolite sorbent was used as the pre-concentration reagent in the determina-tion of atmospheric sulphide.13 The aim of this work was to develop a spectrophotometric method employing a pre-concentration technique to deter-mine quantitatively total sulphide in aqueous environmental samples.The method is based on a Cd-exchanged zeolite pre-concentration sorbent as previously reported by Vasireddy et a1.13 * Present address Stepan Co. 22 North Frontage Road, Northfield IL 60093 USA. t To whom correspondence should be addressed. Experimental Instrumentation and Apparatus All the absorbance measurements were made on a Beckman ACTA I11 UV - visible spectrophotometer. Standard and sample solutions were scanned from 700 to 640 nm using 5-cm path length glass cells (Hellma) and the absorbances were read from a strip-chart recorder. The sulphide aspiration collection apparatus consisted of a reservoir filter tube 160 mm X 32 mm i.d.and a 75 mm long stem tube connected to a zeolite sorbent glass tube. This was in turn connected to a calibrated stainless-steel ball-glass flow meter (Brooks Instrument Divi-sion) and a 1000-ml vacuum flask. The individual components were connected with Y4 in i.d. x 3% in 0.d. Tygon tubing as detailed in Fig. 1. Reagents The methylene blue chromophore-forming reagents N N-dimethyl-p-phenylenediamine monohydrochloride (DPA) -sulphuric acid and iron(II1) chloride were prepared as previously described. 13 A stock sulphide solution was pre-pared by dissolving 0.71 g of analytical-reagent grade Na2S. 9H20 (Mallinckrodt) in de-ionised water and then diluting this to 1 1 (100 p.p.m. as H2S). A pellet of analytical-reagent grade NaOH (Mallinckrodt) was added to the stock solution (yielding a pH >lo) to minimise any H2S evolution.The sulphide solution was prepared by the dilution of a 10-ml aliquot of the sulphide stock solution to 11 with de-ionised water. A pellet of NaOH was also added to the standard sulphide solution (1000 p.p.b.). Other standards were pre-pared by serial dilution of the 1000 p.p.b. standard with de-ionised water. Zeolite Sorbent Tube Two of the three zeolite sorbent tubes were prepared from the top portion of 5-ml Pyrex glass (40 mm bed height) and 1-ml Corning disposable glass (60 mm bed height) volumetric pipettes. These sample tubes served as a convenient inexpen-sive disposable and uniform diameter source of glass tubing because of the pre-formed constriction at the outlet end.A full description of the construction of the sorbent tube was given by Vasireddy et a1.13 The stem of the reservoir filter tub 1308 ANALYST NOVEMBER 1986 VOL. 111 (labelled Reservoir in Fig. 1) served as the 20 mm bed column. The zeolite AAA molecular sieve [Aridzone AAA, Na(A102)(Si02)3,0.3.9 H20] was provided by Amco Arizona Minerals. 14 The cadmium(I1)-exchanged zeolite was prepared using the same procedure as for the determination of H2S in air samples.13 The zeolite pre-concentration tubes were prepared by weighing 200 k 1 mg of the cadmium-exchanged zeolite. Prior to the addition of the zeolite a plug of silanised glass-wool7 2-3 mm in height was placed in each of the glass tubes with the aid of a small diameter glass rod.The glass-wool plug served as the base of the zeolite column allowing an unrestricted flow of the aqueous solution yet retaining the zeolite bed. The prepared zeolite sorbent tubes were stored in a desiccator over an anhydrous phosphorus pentoxide desic-cant prior to use. Sample Collection Environmental samples were collected in 1-1 brown glass, round-bottomed bottles sealed tightly with a PTFE-lined plastic cap. Each sample bottle was filled to the top and the pH adjusted on site to a minimum pH of 10 with 50% sodium hydroxide solution. Each sample bottle was gently inverted several times to ensure the homogeneity of the solution. Sample bottles were transported from the collection site in an insulated Styrofoam cooler at 15.6 "C and kept cool until the actual aspiration procedure was performed.Analytical Procedure All standard and sample solutions were aspirated through the zeolite sorbent collection system in Fig. 1 at the appropriate flow-rate using the vacuum pump. Prior to the removal of the sorbent tube from the apparatus five drops of anhydrous methanol were added to remove any water. It has been Reservoir (50 ml) Calibrated flow meter Collection flask -+ To aspirator Tygon tubing Fig. 1. Sulphide pre-concentration - collection apparatus previously observed that wet zeolite has a tendency to adhere to the walls of the glass sorbent tubes hence impeding complete transfer to the volumetric flask. Further the sulphide may be oxidised during field storage under wet conditions; methanol facilitates air drying thereby simulating field sampling conditions.Air was then drawn through the tube for an additional 5 min to dry the sorbent. The zeolite was transferred into a closed 25-ml volumetric flask and cooled for 2 min in a water - ice bath. The chromophore-forming reagents (0.6 ml of DPA reagent and 1 drop of FeCI3 reagent respectively) were added followed by the immediate addition of 5 ml of water to the flask. The flask was swirled vigorously for a few seconds and then the solution was decanted into another 25-ml volumetric flask. The zeolite was quickly washed twice with 3-ml portions of water followed by a final wash with 5 ml of methanol all of which were decanted into the second 25-ml volumetric flask. The resulting solution was diluted to the mark with de-ionised water and allowed to stand for 30 min to ensure the complete formation of the methylene blue chromophore.Based on the information from earlier studies,l3 one extraction of the cadmium-exchanged zeolite bed was considered adequate for sulphide levels of 2.2 pg and below. The methylene blue solution was transferred into the 5-cm path length sample cell and scanned from 700 to 640 nm on the spectrophotometer. All absorbance readings were taken at the peak maximum of 665 nm versus a blank solution in the reference cell. The absorbance of the blank solution did not exceed 0.03 units at this wavelength. Results and Discussion Flow-rate versus Percentage Recovery as a Function of Bed Length Based on the earlier work13 on determining H2S in air samples 200 mg of cadmium-exchanged zeolite were found to be acceptable for collection of up to 2 pg of sulphide.A single extraction was all that was required for the complete recovery of the sulphide. The sorbent collection tubes were chosen to increase the sorbent bed length by 20 mm intervals ranging from 20 to 60 mm. Similarly sample solutions were run at several flow-rates through the various sorbent bed lengths. As expected the larger the sorbent length and the slower the flow-rate the higher was the recovery of sulphide as seen in Table 1. However significant differences in recovery were only observed between the 20 and 40 mm bed length i e . an average 15.8% increase. A 2.6% increase in recovery was observed using the 60 versus the 40 mm bed length at 20 ml min-1.These results suggested the use of the 60 mm tube at a flow-rate of 6 ml min-1 in order to yield the highest recovery i.e. optimum conditions. However the 40 mm sorbent tube run at 20 ml min-1 was actually chosen for all experiments in this study. Justification for this took into account several additional factors that influenced the sulphide recovery. For example some evolution of sulphide as H2S gas, Table 1. Percentage recovery as a function of bed height and flow-rate. The standard deviation was calculated as on-l for quadruplicate analysis Column bed Flow-rate/ height/mm ml min-1 20 6 20 28 40 6 20 60 20 Average recovery,% * 30.1 27.6 22.4 47.1 42.2 44.8 Standard Relative standard deviation YO deviation % 3.4 11.4 7.7 27.9 3.3 14.6 7.4 15.7 7.2 17.0 9.6 21 .o Average 17.9 * As compared directly with aqueous standards that have not been subjected to pre-concentration on zeolite sorbent beds ANALYST NOVEMBER 1986 VOL.111 1309 even at a pH greater than 10 is inevitable. Consequently a longer aspiration time i.e. a lower flow-rate increased the loss of sulphide from the sample solution to the environment. Only 100-ml volumes of sulphide solutions were used in this study. At a flow-rate of 6 ml min-1 the aspiration time was 16.7 min. The application of this method to a 1000-ml sample would result in an unreasonable aspiration time of 2.8 h for a single trial. This is considered excessive especially for on-site stream monitoring.The 20 ml min-1 flow-rate yielded a slightly lower recovery but gave a more reasonable aspiration time i. e. 5 min for 100-ml and up to 50 min for 1000-ml samples. A 40 mm sorbent bed length was chosen in prefer-ence to the 60 mm length for two reasons firstly the irregular packing of the tube and secondly the unconfirmed suspicion of channelling through the sorbent bed. Both are interrelated through the relationship between the zeolite particle size and the i.d. of the sorbent tube. The 40 mm bed length and 20 ml min-1 flow-rate parameters represent the best over-all compromise between sulphide ion recovery and collection time. These parameters could be applied to a variety of sample solution volumes and are within the range of sulphide concentrations to be determined.The results from an independent test in which Cd-exchanged zeolite and standard sulphide solution were con-tacted directly indicated not greater than 50% recovery compared with the aqueous sulphide standard that had not been pre-concentrated. This may be due in part to losses to the atmosphere or some other unknown mechanism. Although the technique appears to be analyst dependent it is repro-ducible. Relationship of Sample Size to Absorbance Standard sulphide solutions ranging from 0.5 to 2.1 pg in a 100-ml total volume were aspirated through the 40 mm cadmium-exchange sorbent tubes at a flow-rate of 20 ml min-1. The analysis was performed at least in triplicate on each solution. The resulting calibration graph of sulphide (in micrograms) versus methylene blue absorbance was generated using a linear least-squares fit equation on all the data.The slope of the calibration graph was 7.67 X 10-2 with a y-intercept of -5.0 X 10-4 absorbance units (i.e. essentially zero). A correlation coefficient of 0.97 indicated good agreement among the data. The relative error in general decreased with an increase in the sulphide concentration although the standard deviation remained consistent implying a constant error. This error is probably due to the loss of sulphide as H2S gas from the aqueous solution awaiting aspiration incomplete removal of the sulphide by the sorbent and/or during the addition of DPA - sulphuric acid reagent in the chromophore-forming procedure. In earlier work13 on analysis of HZS gas 20% was lost during the chromophore-forming step.which was iden-tical with that used in this work implying that further losses result from incomplete immobilisation or gas loss during pre-concentration. Effect of Sample Volume on Percentage Recovery An examination of Fig. 2 indicates that the effect of increasing the volume of sample solutions from 100 to 500 ml is a reduction in sulphide recovery by only 3%. Based on the "best fit" on the data the sulphide versus sample volume plot followed a quadratic rather than a linear function. The data were best represented by the quadratic equation ax2 + bx + c = y where a and b = -1.38 x 10-5 and c = 33.4%. The variables x and y correspond to the sample volume and percentage recovery respectively.The computer program used was limited by the fact that a must equal b. Comparison of the correlation coefficients (0.90 for the linear and 0.94 for the quadratic function) indicated that the non-linear graph 40 1 I I I I 1 1 0 250 500 7 50 1000 Sample volumeiml I I I I I I I I ' ' I 0 5 10 15 20 25 30 35 40 45 50 Aspiration tirneimin Fig. 2. recovery. Flow-rate = 20 ml min-1 Plot of sample volume and aspiration time versus percentage gave a better fit of the data. Visual inspection of Fig. 2 also tends to confirm the non-linear trend of the data. The reduction in sulphide recovery above 500 ml may be attributed to several factors Firstly the increase in the volume of the sample solution to be analysed increased the number of transfers of the solution to the funnel resulting in losses of gaseous H2S.Secondly the remaining sulphide solution awaiting aspiration was susceptible to oxidation by dissolved oxygen and/or the release of H2S fumes that escaped during the repeated removal of the glass stopper from the volumetric flask. At present it is recommended that the sample size be limited to below 500 ml for this procedure. Determination of Sulphide in Well Water Source The Day 1 Mundelein Illinois well water sample was initially analvsed by pre-concentrating the sulphide followed by the methylene blue absorbance spectrophotometric (Cd zeolite MBAS) method to ensure that the sulphide concentration of this source was within the working range of this method. The Day 2 well water samples were submitted to two State certified and approved laboratories that specialise in water and wastewater analysis and were also analysed in-house by the Cd zeolite MBAS method.Each of the three laboratories analysed the samples within the designated hour. Laboratory B which ran the standard methylene blue method (No. 427 C) reported less than 100 p.p.b. This is the method's limit of detection for a 1-cm cell. Laboratory A reported an average of 12 p.p.b. using the same method but with a 1.9-cm cell. An average of 3.6 p.p.b. was determined by the Cd zeolite MBAS method. The examination of the sample absorbance results (Table 2) between laboratories for their corresponding sulphide levels indicated a difference of a factor of 11.1 (i. e. 0.122 divided by 0.011) between the Cd zeolite MBAS and the Laboratory A methods.Even scaling the results up to a 5-cm cell against the 1.9-cm cell of Laboratory A followed by multiplying the Laboratory A result by the ratio of the path lengths of the two cells ( i . e . 0.011 X 5.0h.9) yielded an absorbance of only 0.029. This result is still a factor of 4.2 times below the 0.122 absorbance for the Cd zeolite MBAS method. The absorbance difference is due solely to the use of the pre-concentration technique. Additionally any variability of the blank solution at an absorbance level as low as 0.011 would have a greater impact on the reported results than that of the 0.122 absorbance value which is obtained at the expense of the 20% loss of sulphide discussed previously. The limit of detection for a 500-ml sample volume for Laboratories A and B differed by a factor of 25 and 500, respectively compared with the Cd zeolite MBAS method.This sizeable difference is again due to the fact that Labora-tories A and B do not concentrate the sulphide in each sample prior to their analysis and hence no sensitivity gain is achieyed by using larger sample volumes. On the other hand lower detection limits for the Cd zeolite MBAS method can be achieved with a larger sample volume. The limitation here is the mass of sulphide recovered 1310 ANALYST NOVEMBER 1986 VOL. 111 Table 2. Interlaboratory comparison of methods for aquatic sulphide determination on Day 2 sample. Laboratory A is Suburban Laboratories, Inc. Laboratory B is Enviro-Test Inc. and Laboratory C is Stepan Co.Absorbance Sulphide p.p.b. Pre- Cell path Limit of Limit of Laboratory concentration lengthkm Sample detection Sample detection . . . . . . A No 1.9 0.01 1 0.004 12 5 B . . . . . . No 1 .o - - <loo 100 c Yes 5.0 0.122 0.007 3.2 0.2 . . . . . . A standard additions technique was chosen as an accuracy “check” on the Cd zeolite MBAS single data point approach. The slope of the line was 3.10 x 10-2 with a y-intercept of 0.172 absorbance unit. A correlation coefficient of 0.99 was generated from three sulphide solutions. The data from the standard additions linear plot yielded a 5.5 p.p.b. sulphide result versus a 4.4 p.p.b. result by the single data point approach. This 1.1 p.p.b. difference represents a 25% relative error. However this is considered to be reasonable agreement for this dilute concentration indicating that the single point -calibration graph approach discussed earlier is adequate for sulphide determination in this concentration range.This reported error was still considerably lower than the 62% error of Laboratory A using the “standard” methylene blue absorption spectroscopy method. As a comparison the accuracy of any of the “standard” MBAS procedures for a 100 p.p.b. detection limit was of the order of 10%. The use of the standard additions technique would also eliminate the effect of certain sample matrix problems. In examples where doubt as to the nature of the sample matrix exists the standard additions technique should be employed in lieu of multiple measurements by the direct method.Conclusion The cadmium(I1)-exchanged zelite MBAS method which has been proved effective for gaseous samples,l3 has been extended to the determination of sulphide in aqueous environ-mental samples. By application of the pre-concentration technique this method gives a lower limit of detection and higher sensitivity than the standard methylene blue absor-bance spectroscopy method. In addition it is potentially a way to collect the sample at the site and preserve it for subsequent transport to the analytical facility. The potential interference of dissolved oxygen in the sample during transport would be eliminated by the field immobilisation of sulphide on the Cd zeolite once the sorbent is dried. Also the impractical act of transporting numerous large volumes of liquid samples back to the laboratory will be replaced by transporting the same number of small sorbent collection tubes if this method is used.The authors express their grateful appreciation to Stepan Co. for use of their facilities during the course of this project. They also thank the Society of Analytical Chemists of Pittsburgh for financial support in the form of an SACP starter grant and to Arizona Minerals for the donation of the Aridzone AAA zeolite sorbent used throughout this study. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. References Lawrence A. W. and McCarty P. L. Air Wat. Pollut. 1966, 10 207. Camp T. R. and Meserve R. L. “Water and its Impurities,” Second Edition Dowden Hutchinson and Ross Stroudsburg, PA 1974 p.257. Kirk R. E. and Othmer D. F. “Encyclopedia of Chemical Technology,” Third Edition Volume 14 Wiley New York, 1981 pp. 208-209. Kirk R. E. and Othmer D. F. “Encyclopedia of Chemical Technology,” Third Edition Volume 17 Wiley New York, 1981 p. 125. Kirk R. E. and Othmer D. F. “Encyclopedia of Chemical Technology,” Third Edition Volume 19 Wiley New York, 1981 pp. 395-416. Boon A. G. and Lister A. R. in Jenkins S. H. Editor, “Progress in Water Technology,” Volume 7 Pergamon Press, Oxford 1975 pp. 289-300. Boon A. G. and Lister A. R. in Jenkins S. H. Editor, “Progress in Water Technology,” Volume 7 Pergamon Press, Oxford 1975 pp. 599-605. Greenberg A. E. Connors J. J. and Jenkins D. Editors, “Standard Methods for the Examination of Water and Waste-water,” Fifteenth Edition American Public Health Associa-tion Washington DC 1980 pp. 442-450. Kirk R. E. and Othmer D. F. “Encyclopedia of Chemical Technology,” Third Edition Volume 13 Wiley New York, 1981 pp. 720-721. Frant M. S. and Ross J. W. assigned to Orion Research Inc. US Pat. 3 672 962 1972. Stocky G. D. and Dwyer F. G. Editors “Intrazeolite Chemistry,” American Chemical Society Washington DC, 1983 pp. 3-11. Kirk R. E. and Othmer D. F. “Encylopedia of Chemical Technology,” Third Edition Volume 24 Wiley New York, 1981 p. 322. Vasireddy S . Street K. W. Jr. and Mark H. B. Jr. Anal. Chem. 1981 53 868. Ataman 0. Y. and Mark H. B. Jr. Anal. Chem. 1977,49, 1331. Street K. W. Jr. Mark H. B. Jr. Vasireddy S . LaRue-Filio R. C. Anderson C. W. Fuller M. P . and Simon S. J., Appl. Spectrosc. 1985 39 68. Paper A61113 Received April l l t h 1986 Accepted June 16th 198

ISSN:0003-2654    

DOI:10.1039/AN9861101307

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, NOVEMBER 1986, VOL. 111 1313 Automated Flow Injection Spectrophotometric Determination of Zinc Using Zincon: Applications to Analysis of Waters, Alloys and Insulin Formulations Michael A. Koupparis” and Paraskevi 1. Anagnostopoulou Laboratory of Analytical Chemistry, Department of Chemistry, University of Athens, 104 Solonos Street, Athens 10680, Greece An automated flow injection method for the determination of zinc, based on the zincon method with differential demasking of the cyanide metal complexes with cyclohexanone, is described. Zinc concentra- tions, in the range 1-10 pg ml-1 can be determined with a precision greater than I%, at a rate of 80 determinations per hour and with a detection limit of 0.05 pg ml-1. Because of the short reaction time (27 s), interferences arising from the breakdown of other metal cyanide complexes are considerably reduced.The method was evaluated for the determination of zinc in alloys and insulin formulations and the results obtained were in good agreement with those of reference methods. Recovery studies on spiked sample solutions gave excellent results. Keywords: Zinc determination; alloys; insulin formulations; automated flow injection analysis; waters The determination of zinc is of great importance in a variety of fields such as environmental chemistry, domestic and waste water control, metallurgy, galvanising and alloy manufactur- ing, agriculture (soil analysis) , pharmaceutical technology (analysis of zinc and zinc - insulin formulations) and in clinical chemistry (monitoring zinc levels in serum).In all the above fields, sensitive, selective, accurate, rapid and inexpensive methods are required. Apart from atomic absorption spectrometry, which , although the preferred analytical technique , is expensive, several spectrophotometric methods have been proposed1 with different sensitivities, selectivities and of varying complexity. Of these, the rather complicated dithizone method2.3 and the zincon method4.5 are well established and have been adopted as official methods for water ,6 soil7 and pharmaceutical analyses.8 Neither reagent is selective towards zinc, and a separation step is required. In the classical dithizone method, a preliminary extraction with the reagent is employed, whereas in the zincon method an ion-exchange procedure or differential demasking of the metal ions is used.A flow injection analysis (FIA) method for zinc determina- tion would have many advantages, e.g. , simplicity, precise time control, rapid analysis, low sample and reagent consump- tion, high precision and the possibility of automation. An FIA spectrophotometric method, using xylenol orange as the reagent and thiosulphate as a masking agent for copper and ’lead, has already been developed,g but it suffers from interference from nickel. An automated method, based on 4-(2-pyridylazo)resorcinol (PAR) and a centrifugal analyser , has also been reported.10 This has good sensitivity and detection limits but it has serious interferences and high instrumental costs. In this paper, the development of a zincon method for zinc determination using an automated flow injection analyser is described.The method is based on the differential demasking of metal cyanide complexes with cyclohexanone at pH 9.0.11 Zinc is rapidly liberated to form a 1 : 1 complex with zincon, which is then measured at 620 nm. Ascorbic acid is added to the cyanide reagent to reduce interferences from copper(II), iron(II1) and manganese.6 The automated FIA method is suitable for the analysis of copper alloys and zinc - insulin formulations. * To whom correspondence should be addressed Experimental Apparatus A laboratory-constructed automated FIA photometric ana- lyser,l* interfaced to a Rockwell AIM 65 microcomputer, was used. The optimised manifold that was developed is shown in Fig. 1. The sample solution is pumped continuously through the sample loop (300 pl) for a pre-selected load - wash time (30 s) and is then automatically injected into a stream of mixed reagent A (cyanide - ascorbate - buffer, pH 9.0) and is allowed to react in the first reaction coil.The reacted sample zone is then merged with mixed reagent B (cyclohexanone - zincon - buffer, pH 9.0) and allowed to react in the second reaction coil. The absorbance of the zinc - zincon complex is measured at 620 nm. The apparatus is cleaned with 0.10 M NaOH solution followed by distilled water. Reagents All reagents were of analytical-reagent grade and de-ionised, distilled water was used throughout. Borate buffer solution, 0.5 M, p H 9.0. This was prepared by dissolving 30.9 g of boric acid and 8.4 g of sodium hydroxide in 900 ml of water, adjusting the pH with 5 M NaOH or HCl, then diluting to 1 1.A 0.05 M borate buffer was prepared by appropriate dilution with water. Zincon stock solution, 0.1% mlV. A 200-mg amount of zincon sodium salt (2-carboxy-2’-hydroxy-5’-sulphoforma- zylbenzene) was dissolved in 200 ml of water. This solution was stable for about 1 week. ml min-’ 2.9 B A 6.4 W \ U 300 ~l Fig. 1. Schematic diagram of the automated FIA system for zinc determination. Reagents: A, cyanide - ascorbate - borate buffer (pH 9.0); B, cyclohexanone - zincon - borate buffer1312 a, 6 0.2 e 2 a 6 0.1 ANALYST, NOVEMBER 1986, VOL. 111 - - Composite reagent A (0.010 M cyanide - 0.10 M ascorbate - buffer, pH 9.0). A 0.65-g amount of KCN and 17.6 g of ascorbic acid were dissolved in 1 1 of borate buffer solution and the pH was readjusted to 9.0 with 5 M NaOH.Composite reagent B (10% VIV cyclohexanone - 0.016% mlV - zincon buffer, pH 9.0). A 50-ml volume of pure cyclohexanone and 80 ml of zincon stock solution were mixed and diluted to 500 ml with borate buffer. Zinc stock solution, 1000 yg ml-1. A 4.397-g mass of ZnS04.7H20 was dissolved in 1 1 of water. Zinc working standard solutions, 1-10 yg ml-1 in dilute 0.05 M borate buffer, pH 9.0. These were prepared daily by appropriate dilutions of the stock solution and 0.5 M borate buffer. Trichloroacetic acid, 0.61 M. A 10.0-g amount of CC1,COOH was dissolved in 100 ml of water. HNO3, 1% VIV. HCI, 1 and 5 M. NaOH, 5 M. Sample Preparation Waters For the determination of dissolved zinc, the sample was filtered through a 0.45-pm membrane filter, pre-cleaned with 1% VIV HN03 followed by distilled water. The pH was adjusted to 9.0 by mixing with the 0.5 M borate buffer in a 9 + 1 ratio.For total zinc the sample was acidified with HC1, filtered, its pH was adjusted to 9.0 as before and zinc determined during the same day. If the final sample solution contained more than 10 pg ml-1 of zinc, it was diluted with the 0.05 M borate buffer. Alloys About 0.5 g of the sample, accurately weighed, was dissolved in 10 ml of concentrated HN03 and the solution was evaporated in a water-bath to half of the solution volume. The remaining solution was filtered through a Whatman No. 42 Table 1. Effect of foreign metal ions on the determination of zinc (5 m1-l) Concentration*/ Zinc found/ Metal ion yg ml- yg ml-l Error, YO CU'I .. . . 100 5.02 0.4 NiII . . . . 100 5.13 2.6 CO" . . . . 100 5.06 1.2 A1'I1 . . . . 4 4.78 -4.4 Cd" . . . . 5 5.21 4.2 Cr"' . . . . 0.5 4.77 -4.6 Mn" . . . . 10 4.88 -2.4 Fe"' . . . . 20 5.23 4.6 Pb'I . . . . 50 5.17 3.4 * For errors less than k 5%. 0.4 a, r: ((I 2 0.3 % 2 0.2 5 7 9 Flow-rate/ml min- Fig. 2. Effect of total flow-rate of reagents on the absorbance peak. Zinc. 5 WE ml-1: samde volume. 300 ul filter-paper, pre-cleaned with 1% V/V HN03, the filter was rinsed with 1% VIV HN03 and the filtrate was then diluted to 100 ml with water. An appropriate volume of the sample solution was neutralised using 5 M NaOH and a phenolph- thalein indicator. This was further diluted with 0.05 M borate buffer to obtain zinc concentrations in the range 1-10 pg ml-1 and concentrations of the expected interferents (mainly copper) less than those given in Table 1.Zinc was determined during the same day. If any precipitation occurred, the solution was filtered while the zinc remained in solution as a soluble hydroxide complex. Ins ulin formula ti0 ns (injections and suspensions) For total zinc determination, 1-5 ml, accurately measured, of the preparation to be tested was transferred into a centrifuge tube. A 1.0 ml volume of 1 M HC1 was added for clarification and then 8.0 ml of trichloroacetic acid for deproteination. The mixture was centrifuged at 4000 g after 20 min. A 2.00-ml volume of the supernatant liquid was neutralised with 2.0 ml of 1 M NaOH, diluted to 10.0 ml with the 0.05 M borate buffer and the zinc concentration determined.For the determination of zinc in the supernatant liquid of the insulin suspension, a portion of the preparation was centrifuged at 4000g. An accurately measured volume of the clear supernatant liquid was diluted to 10.0 ml with the 0.05 M borate buffer and zinc determined during the day of preparation. Reagent blank A reagent blank solution is prepared for use in the analysis of alloys and insulin formulations. Measurement Procedure The photometer of the flow injection analyser was set at 620 nm, the reagents were pumped through and the transmittance was set at 100%. The number of standards, samples, runs per standard and sample and the injection and load - wash times (15 and 30 s, respectively) were input into the microcom- 100 300 Sample volume/pI Fig. 3.Effect of sample volume on absorbance peak. Zinc, 5 pg ml-1 0.5 a, C (D e 2 2 5.0 0 I 10 pg ml-1 I u I L 2 min Time Fie. 4. TvDical FIA absorbance Deaks obtained for zinc standards . U I 1 I . measured i% increasing and decrea'sing orderANALYST, NOVEMBER 1986, VOL. 111 1313 Table 2. Determination of zinc in standard reference alloys (alloys from Thorn Smith, USA) Other metals, %* Zn, '10 Alloy c u Pb Sn Fe Ni Certified Found? Error, '/o Brass74 . . . . . . 93 Brass78 . . . . . . 81 Brass79 . . . . . . 76 Brass86 . . . . . . 86 Brass88 . . . . . . 88 Brass90 . . . . . . 57 Zincore20 . . . . . . - Germansilver6 . . . . 64 1.00 7 6 - 4.69 9 8 - 6.13 4 4 6 3 - - 6 - - - 5.6 3.2 - - - - - 42.98 - - 12.00 1 - 0.08 18 16.30 - - - - - * Concentrations rounded.t Mean of three measurements, f S.D. 0.960 f 0.006 4.60 kO.05 6.35 5 0.04 5.50 fO.O1 3.25 kO.01 43.3 f 0 . 4 11.7 f O . l 16.2 f 0 . 2 -4.0 -1.9 +3.6 -1.8 +1.6 +0.7 -2.5 -0.6 Mean: 2.1 Table 3. Results for recovery of zinc from alloy sample solutions . . . . Zinc/pg ml-1 Alloy Measured Added Recovered Recovery, YO Brass 74 . . . . . . . . 0.86 2.00 1.93 96.5 Brass 79 . . . . . . . . 2.64 2.00 2.08 104.0 Brass 88 . . . . . . . . 2.58 2.00 2.05 102.5 Brass 90 . . . . . . . . 3.46 2.00 2.10 105.0 Zincore 20 . . . . . . . . 5.46 2.50 2.46 98.4 Germansilver6 . . . . . . 6.77 2.00 2.03 101.5 Mean: 101.3 Table 4. Determination of total and free zinc in commercial insulin formulations Total zinc/pg ml - 1 Free zinc (FIA) Formulation AAS FIA * Error, YO pg ml-1 % LenteMC .. . . . . 80 80.5 f 0.4 +0.6 43.1 53.5 SemilenteMC? . . . . 80 80.8 k 0.4 +1.0 53.6 66.3 UltralenteMC . . . . 80 78.3 f 0.2 -2.1 48.1 61.4 MonotardMC . . . . 80 79.6 f 0.5 -0.5 45.2 56.8 Mean: 1.0 * Mean of three measurements on two samples +_ S.D. t Date expired. Table 5. Results for recovery of zinc from insulin sample solutions Zinclug ml - 1 Formulation LenteMC . . LenteMC . . SemilenteMC . . SemilenteMC . . UltralenteMC . . UltralenteMC . . MonotardMC . . MonotardMC . . * Samples diluted before measurement. Measured . . 3.90 . . 4.91 . . 14.4 . . 14.1 . . 24.1 . . 25.2 . . 27.6 . . 26.3 Added 10.0 20.0 30.0 40.0 40.0 50.0 60.0 70.0 Recovered* 9.80 21.1 29.2 40.6 39.1 48.7 59.5 71.8 Recovery, YO 98.0 105.5 97.3 101.5 97.8 97.4 99.2 102.6 Mean: 99.9 puter's routine program.The determination then proceeded automatically, measuring the standards, constructing the calibration graph by linear regression analysis, measuring the samples and calculating and printing out their zinc concentra- tions. A graphical presentation of the analytical data is also available using the chart recorder. If an autosampler is not available, an interactive program can be used and the samples are manually presented to the sample probe. Results and Discussion System Op timisa tion The FIA manifold developed (Fig. 1) provides a two-step automation of the zinc determination. In the first reaction coil, zinc and other heavy metal ions present rapidly form cyanide complexes at pH 9.0.At the same time, the ascorbate reduces FeIII and CuII to FeII and CuI, respectively, the cyanide complexes of which are more stable than those of the oxidised forms. Manganese(I1) is also complexed by ascorbate. A pH of 9.0 was selected as it allows effective masking by the cyanide complexes and because the zinc - zincon complex is stable in the pH range 8.5-9.5.4 The cyanide and ascorbate concentrations were chosen to reduce effectively the interference of any other heavy metal ions if they were present at high concentrations. The length of the first reaction coil was optimised to give the maximum mixing of the sample, minimum dispersion and to allow complete cyanide complex formation and reduction by ascorbate.1314 ANALYST, NOVEMBER 1986, VOL.111 From the first reaction coil, the mixed sample zone is merged with cyclohexanone and zincon. The cyclohexanone concentration (10% VIV) and the length of the second reaction coil were chosen for rapid but selective liberation of zinc without any demasking of the other metal ions. With higher concentrations of cyclohexanone, solubility problems, base-line instability and thus poor precision were observed. The zincon concentration was sufficient for rapid complex formation during the zinc demasking step. The analytical reaction scheme can be presented as: Mn+ + xCN- [M(CN),](x - n)- Zn2+ + 4CN- C [Zn(CN),]2- [Zn(CN)& + 4C6Hio0 4H2O + Zn2+ + 4C6Hlo(OH)CN + 40H- Zn2+ + zincon blue complex The total flow-rate of the reagents was optimised for high sensitivity, high sample throughput and low reagent consump- tion.The results of this study are shown in Fig. 2. There was a linear increase in the peak height with flow-rate. The effect of the sample volume on the response is shown in Fig. 3. Validation of the Method Under the optimised conditions, the calibration graph was linear in the range 1-10 pg ml-1 with the following least- squares regression equation: A = 0.00251 + 0.0716 [Zn2+] with r = 0.9993. Typical FIA recorded peaks for the calibration graph are shown in Fig. 4 in increasing and decreasing concentration order, showing the absence of any carry-over effect. The precision of the measurements varied from 1.6 to 0.2% relative standard deviation (n = 10) for the lowest to the highest concentration of the calibration graph.The detection limit, defined as that concentration giving a peak equal to the Students t value at the 99% confidence level times the standard deviation of the repetitive determination of zinc in a 1 pg ml-1 standard, is 0.05 pg ml-1. The sensitivity of the method, calculated as the reciprocal of the slope of the calibration graph, was found to be 14 pg ml-1 and corresponds to a zinc concentration giving a peak of one absorbance unit. Under the optimised conditions, the dispersion13 of the sample, as estimated experimentally using a dye (ratio of absorbance of undiluted dye and the peak height of the dispersed dye) was 3.5. This dispersion was adequate for the effective mixing of the injected zone with the reagents and for the matching of the standards and samples with respect to the pH and borate buffer concentration, if borate buffer contami- nated with zinc was used.Measurements of a zinc standard, prepared in various borate buffer concentrations (0.01- 0.5 M), gave identical results. The proposed method is not as sensitive as the dithizone method and is thus not affected by trace amounts of zinc in reagents of analytical-reagent grade. If reagents with high zinc contamination are used, the use of a reagent blank is required. The sample throughput was 80 per hour assuming a 15-s injection time and a 30-s load - wash and data collection and manipulation time. Interference Studies The manual procedure suffers from interferences from those metal ions with unstable cyanide complexes. Special precau- tions are required to measure the absorbance in a short time (1 min) to avoid the demasking of other metal ions.14 With the automated FIA method the residence time of the sample in the reaction coils is only 27 s and this kind of interference is therefore totally eliminated.In order to study the interference of a diverse range of metal ions, synthetic solutions containing 5 pg ml-1 of zinc and various amounts of the interferents were measured. Table 1 shows the maximum allowable concentrations of the metal ions that caused errors of less than 5%. As can be seen, only CrIII, A1111 and Cd" interefere at concentrations comparable to those measured for zinc. It is noticeable that Al"1, Cr"1 and Mn", metal ions not masked by cyanide, caused negative errors. Also the threshold concentration found for Cr"' was 0.5 pg ml-1, compared with 10 pg ml-1 found with the manual method.6 These effects can be explained by the kinetic nature of the FIA procedure.In the second reaction coil, the above three metal ions are free to form complexes with zincon that do not absorb at the wavelength of measurement, and they therefore compete with the zinc ions demasked by cyclohexan- one. All three reactions, i.e., the formation of interferent - zincon complexes, the demasking of zinc and the zinc - zincon complex formation, proceed simultaneously and are not completed at the time of measurement. It seems that CrIII has the greatest effect on this kinetic procedure. Copper can be masked at high concentrations, and the method can therefore be used for the determination of zinc in copper alloys.Copper concentrations are kept at less than the allowable level (100 yg ml-1) by dilution. The interferences from the other ions are significantly reduced with respect to the manual procedure.6 Accuracy The automated FIA method was evaluated by analysing standard reference alloy samples (Thorn Smith, USA) and performing spiked recovery tests on sample solutions. As shown in Table 2, the results are in good agreement with the reported certified values, with errors varying from 0.5 to 4% (mean 2.1%). As mentioned under Sample Preparation, the alloy samples were diluted to keep the expected interferent concentrations less than those able to be masked (Table 1). The relative standard deviations ranged from 0.3 to 1.2% (n = 3). The recovery results (Table 3) ranged from 96.5 to 105.0% (mean 101.3%) The FIA method was also evaluated for the determination of total and free zinc in commercial zinc - insulin formulations.The results were found to be in good agreement with those obtained by the reference AAS method (Table 4). Recovery results (Table 5) performed on insulin sample solutions gave a mean of 99.9% (97.3-105.5%). Conclusions The proposed automated FIA method is rapid, sensitive, precise and accurate with a high sample throughput. The time required for preparation is short and the reagent consumption is relatively low. The residence time of the sample in the reaction mixture is kept precisely short and as the method is kinetically dependent all the interferences associated with the instability of the cyanide complexes in the prolonged manual procedure are therefore minimised.The automated FIA method can be used on a routine basis for the determination of zinc in copper alloys, in zinc - insulin formulations (instead of the tedious dithizone or expensive AAS method) and in water and soil analysis. The time-determining step in these determi- nations is the preparation of the samples. The proposed method is another example of how FIA can be useful in analytical problems that require the precise control of short lengths of time in analytical procedures. It also shows that FIA is a valuable technique for the safe use of toxic analytical reagents such as cyanides. References 1. 2. 3. Ackermann, G., and Kothe, J., Talanta, 1979,26, 693. Hibbard, P. L., Znd. Eng. Chem., Anal. E d . , 1937, 9, 127. Sandell, E. B . , "Colorimetric Determination of Traces of Metals," Third Edition, Interscience, New York, 1969.ANALYST, NOVEMBER 1986, VOL. 111 1315 4. 5. 6. Rush, R. M., and Yoe, J. H., Anal. Chem., 1954, 26, 1345. Platte, J. A., and Marcy, V. M., Anal. Chem., 1959, 31, 1226. American Public Health Association, American Water Works Association and Water Pollution Control Federation, “Stan- dard Methods for the Examination of Water and Wastewater,” Fifteenth Edition, American Public Health Association, Wash- ington, DC, 1980. “Methods of Soil Analysis,’’ American Society of Agronomy, Madison, WI, 1965. “United States Pharmacopeia, XXth Revision,” Mack, Easton, PA, 1980. Kuroda, R., and Mochizuki, T., Talanta, 1981, 28, 389. Goldstein, G., Madox, W. L., Kelly, M. T., Anal. Chem., 1974,46, 485. 7. 8. 9. 10. 11. 12. 13. 14. Miller, D. G., J . Water Pollut. Control Fed., 1979, 51, 2402. Koupparis, M. A., and Anagnostopoulou, P. I., 1. Autom. Chem., 1984, 6, 186. RiiiEka, J., and Hansen, E., “Flow Injection Analysis,” Wiley-Interscience, New York, 1981. Natking, R., Weiner, L. M., andZak, B., Microchem. J . , 1971, 16, 14. Paper A51451 Received December I6th, I985 Accepted May 16th, 1986

ISSN:0003-2654    

DOI:10.1039/AN9861101311

出版商:RSC    

年代:1986

数据来源: RSC