ISSN:0003-2654    

DOI:10.1039/AN98409FX029

出版商:RSC    

年代:1984

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98409BX031

出版商:RSC    

年代:1984

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98409BP065

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST AUGUST 1984 VOL. 109 973 Recent Developments in Detection Techniques for Hig h-performance Liquid Chromatography Part 11.* Other Detectors A Review P. C. White Metropolitan Police Forensic Science Laboratory 109 Lambeth Road London SE 1 7LP UK Summary of Contents Introduction Detectors involving no phase change of the eluate Refractive index detectors Optical activity Circular dichroism Photoionisation detectors Piezoelectric crystal detection systems Low-angle laser light scattering detectors Radioactivity detectors Mass detector Gas-ch romatog rap hic detectors Mass-spectrometric detectors Detectors involving a phase change of the eluate Concluding remarks References Keywords Review; high-performance liquid chromatography; detectors Introduction In Part I of this review,274 recent developments in spectro-scopic and electrochemical high-performance liquid chromat-ographic (HPLC) detection techniques were reviewed.In this final part other types of HPLC monitoring systems are similarly appraised. These detectors have been classified into two main groups the division being based on whether any phase change of the eluate occurs before the solute is detected. The numbering of the Figures Tables and references continues from Part I. Detectors Involving No Phase Change of the Eluate Apart from the spectroscopic and electrochemical detection methods that can be used to monitor HPLC eluates there are a number of other detection techniques that can be used to detect a solute directly in an eluate.As will be shown some of these systems are well established whereas others are relatively new developments. Refractive Index Detectors The refractive index (RI) detector was one of the first on-line detection systems to be employed in HPLC. It is one of the most versatile detection devices to be used in liquid chromato-graphy and its mode of operation is based on the measurement of a change in either velocity or direction of light as it passes from one medium to another. As the refractive index is essentially an additive property it follows that the detector response depends on the difference between the refractive indices of the pure mobile phase and the eluate. For this reason all detector cell designs are based on a differential ~ ~ ~ ~ _ _ _ _ _ ~ * For Part I see reference 274.mode of operation and analyses must be carried out under isocratic conditions. Four methods have been used in commercial detectors to measure changes in refractive index viz. the Christiansefi effect the deflection principle Fresnel's method and the interference principle. Excellent discussions on these well established techniques have been p~blished2~5.275 and the last three methods appear to be the most widely used with typical cell volumes of about 10 pl being used. Fresnel's method requires a change of prism to encompass the whole refractive index range of 1-1.75 refractive index units (RIU). All of these RI detector designs suffer from a number of disadvantages which are related to the sensitivity of refractive index to changes in eluate composition temperature and pressure.As small differences in refractive index are being measured (ca. 0.1 RTU) the detection limits of these systems vary between 10-6 and 10-8 g ml-1. Refractive index is very temperature dependent and decreases rapidly with small increases in temperature. To be able to detect 10-6-10-8 g of solute a noise equivalent concentration of 10-8 RIU is essential and to maintain this level the temperature of the detection system must be thermostated to within + l o - 3 "C. All commercial detectors incorporate some degree of thermo-stating to reduce this problem. Pressure fluctuations due to the pumping of the eluent must be reduced and usually pulse dampers are fitted to the pumping system when work is to be carried out at high sensitivity.For optimum results the pressure variations should be held below 2 x 10-4 bar (0.3 mmHg). One particular instrument manufacturer has developed an RI detection system that incorporates facilities to control temperature and pressure fluctuations and the results are very impressive. It has been reported recently that with some sugar samples detection limits of 10 ng of injected material could be achieved with this system.27 974 ANALYST AUGUST 1984 VOL. 109 It was shown in Part I (Fig. 8) that the RI detector does not have a uniform response to a particular group of compounds, e.g. alkanes and therefore for quantitative analyses indivi-dual calibration graphs for each of the compounds have to be prepared. One particular advantage of the RI detector is that its linear range is greater than that of almost any other detector.Column overloading effects will often occur before any deviation in linearity is noticed. For this reason the detector is ideal for preparative HPLC. The RI detector is also very useful for monitoring size exclusion separations. Pro-vided that the polymer has more than ten monomer units the RI response is directly proportional to the concentration of the polymer and independent of the relative molecular mass. Despite some of the problems associated with RI detection, this method will continue to be employed for many more years. It is still one of the most “universal” techniques and hopefully manufacturers will improve the thermostating of their instruments.Increased sensitivity can also be expected if these are used in conjunction with the pulse-free pumps that are now available. Optical Activity An area in HPLC detection that has been exploited recently is that of polarimetry i.e. the measurement of optical activity. Only compounds with an asymmetric carbon atom will rotate a plane of polarised light and show optical activity and therefore polarimetry is a highly selective detection tech-nique. Many of these compounds exist as optical isomers and although they cannot usually be separated by conventional packing materials it is possible to differentiate between them by polarimetry. This differentiation is possible because polarimetry measures both the direction and degree of rotation of polarised light produced by the isomers.For optically active compounds the extent of rotation depends on the number of molecules in the light path of the radiation and for a solution this will vary according to the concentration of the solute and the cell length in which it is contained. It is also dependent on the wavelength of the radiation and temperature. In conventional polarimeters, optical activity is measured under static conditions in cells with long path lengths and the sodium D line is used as the radiation wavelength. One of the major problems associated with using a polarimeter as an HPLC detector is that of sensitivity owing to the large flow cells that are normally used. Recently a modified commercial polarimeter that had been fitted with a low volume flow cell was used to monitor HPLC eluates.277 For some sugars the detection limits ranged from 5 to 20 pg, i.e.5-10 times less sensitive than with UV or RI detection. It was also shown in this study that the optical rotation and hence the detector response is very dependent on the wavelength of the radiant beam in that the rotation increases with a decrease in wavelength. The advantages of this detection technique can be seen in Fig. 15 for the analysis of an orange juice. As solutes can rotate the plane of polarised light in either direction the polarimeter will give both positive and negative peaks which increases the selectivity of the detection system. A number of similar applications have been described by Di Cesare and Ettre.278 An excellent example of the value of polarimetry as a detection method was also given namely the quantitation of D- and L-isomers in a penicillamine prepara-tion.These isomers were not separated by the column. Another group used a more complex polarimeter with a laser as the radiation source.279 As the laser produced a highly collimated beam the internal diameter of the flow cell was reduced substantially. The flow cell described in this paper had an internal diameter of 1.58 mm and a path length of 10 cm which gave a volume of 200 pl. Alignment of the laser was critical otherwise scattering and reflection off the cell walls 3 4 4 0 4 8 12 16 Time/mi n Fig. 15. Analysis of an orange juice using (a) a polarimeter and ( b ) a refractive index detector in series (the peaks are labelled for comparative purposes).From reference 278 with permission of Elsevier Science Publishers occurred. When tested with a fructose - raffinose mixture (5 pg of each injected in 100 p1) a detection limit of 0.5 pg at a signal to noise level of 3 was reported for both sugars. Untreated urine samples were also analysed with this system. An HPLC detector based on optical activity monitoring can offer several advantages. For example most chromatographic solvents are not optically active and therefore either isocratic or gradient elution can be performed with a wide range of eluents. The major advantage of this type of detector is that it is extremely selective and analysts can differentiate between optically active isomers. This is important when applications involving chemical environmental and energy problems are being studied as only one isomer is usually responsible for producing biological activity in a sample.For these reasons this mode of detection will probably receive considerable attention in the future. Circular Dichroism A technique related to optical activity namely circular dichroism has been used for HPLC detection. A plane polarised beam of light after passing through a sample, consists of two beams that can be circularly polarised i ANALYST AUGUST 1984 VOL. 109 975 opposite directions. The absorbance related to each of these beams will vary from compound to compound and the difference in absorbance values (6A) between the two beams is called circular dichroism. Circular dichroism (CD) can give information about stereo-chemical features of optically active species that contain a suitable chromophore.The technique is extremely selective and as CD is also dependent on wavelength it can be utilised in a similar manner to that of variable-wavelength UV detection. Conventional dual-beam spectrophotometers are designed to determine the difference in absorbance between a sample and a standard and they can therefore be easily adapted to measure CD. Although no cell dimensions were given this type of detector was used for the analysis of pavine enantio-mers.280 Commercial CD spectrophotometers are available and one of these has been used as an HPLC detector.2*1 The spectrophotometer was fitted with a micro-flow cell (volume 10 1.11) and with L-tryptophan as the test solute the detection limit was 3 pg.Although this system was also applied to the analysis of several complex alkaloid mixtures CD would appear to have limited usage as an HPLC detector. Photoionisation Detectors When molecules are excited with sufficient energy ionisation can occur and the resultant currents associated with this effect can be amplified and used to detect a compound. Photoionisa-tion detection has been used for several years in gas chromatography and with recent developments the technique can now be used to monitor solutes in HPLC eluates. In 1975 Schmermund and Locke produced a photoionisa-tion detector (PID) that used the resonance lines from a variety of discharge lamps to irradiate the flash-evaporated eluate from an HPLC column.282 This system unfortunately suffered from problems related to solvent evaporation and recently Locke et ~ 1 .2 8 3 developed a direct liquid-phase photoionisation detector. To avoid evaporation energy from a microwave-excited continuum xenon source was used to irradiate the eluate in a 10-1.11 flow cell. As this is a low-energy source it is suitable for ionising molecules with ionisation potentials of less than 7.81 eV. As water is easily ionised (6.31 eV) separations were limited to substances that could be eluted with non-aqueous solvent systems. With hexane - propan-2-01 as the eluent this system was used to analyse some polynuclear hydrocarbons and some oxygen- and nitrogen-containing compounds. Minimum detection levels at the picogram level were reported for the polynuclear hydrocarbons and at the nanogram level for the other compounds.Some degree of selectivity was obtained as phenols and chlorinated compounds could not be detected. With pyrene as a test solute the amplified photocurrents were reported to be linear over a range of 2 x 10-11 to 2 x 10-5 g of injected material. This appears to be a very sensitive detection system and worthy of further development. Although the technique is restricted to chromatographic separations that involve non-polar non-aqueous solvents Locke et al. are currently investigating other detector designs to overcome these prob-lems so that solutes separated by reversed-phase techniques can be identified. Piezoelectric Crystal Detection Systems A voltage can be obtained from some materials e.g.quartz crystals and certain ceramics by compressing them. Con-versely if a voltage is applied to these materials the crystals expand or contract in certain directions. This is known as the piezoelectric effect and it has been used in a wide range of electrical devices for many years. Recently several groups of workers have attempted to use these properties to detect solutes in an HPLC eluate. Two twes of detector have been developed namely the mass detector and the photoacoustic detector. Recently some preliminary work has been reported on the application of piezoelectric crystals as mass detectors in liquid chromatography.284 The principle behind the operation of this detection system is that when a material is surface-adsorbed on to an oscillating piezoelectric crystal the resonance frequency changes linearly with the mass adsorbed.It was found that two crystals were required to obtain any results. One of these crystals (reference) was treated with trimethyl-(dimethy1amino)silane to overcome problems related to density effects. The other crystal (indicator) was used to detect the solute and to increase the surface adsorption this was treated with a long-chain hydrocarbon (CIS or C22). The sensitivity of this detection system was poor owing to the low adsorption efficiency of the indicator crystal. Further the crystals became saturated with solute after several injections. It would appear that if this technique is to be used for monitoring HPLC eluates considerable work will have to be done to enhance and control the surface adsorptivity of the crystal.Better results have been obtained with piezoelectric crystals when they are used in photoacoustic detectors. The principle of photoacoustic detection is that when a solute is dissolved in a solvent and irradiated absorption of light occurs and the solution expands. The expansion against the face of an oscillating piezoelectric crystal produces a change in fre-quency. Photoacoustic detection of compounds in static solutions is an extremely sensitive technique especially when a laser is used to irradiate the sample.285 For this reason the use of this type of detection system for the trace analysis of solutes in HPLC eluates has been investigated. Oda and Sawada286 designed a laser-activated photoacous-tic detector for HPLC.The flow cell developed for this detector had an internal volume of 20 pl and its construction is shown in Fig. 16. m rl -BNC connector P Effluent entry tube Effluent exit tube 2 Fig. 16. Schematic diagram of a flow cell designed for photoacoustic detection. From reference 286 with permission of the American ,I ~ ~~ __. Chemical Societ 976 ANALYST AUGUST 1984. VOL. 109 Pressure fluctuations caused by the pulsating flow from the pump seriously affected the measurement of the photoacous-tic signal but the authors reduced this effect by increasing the modulation frequency of the laser. The system was evaluated with three isomeric forms of chloro-4-(dimethyl-amino)azobenzene. The detector response was linear over a range from 0.31 ng to 2.6 kg and the detection limit was 0.21 ng of injected material.The photoacoustic system was reported to be about 25 times more sensitive than a UV system. The “windowless” flow cell that was developed for flu-orimetry has also been used with a piezoelectric crystal in some photoacoustic studies.65 Detection levels of some polynuclear hydrocarbons ranged from 20 to 400 ng ml-1 but the system has not yet been coupled up to an HPLC column outlet. Photoacoustic detection offers little selectivity but as this Uchnique gives information that is complementary to fluor-escence simultaneous detection with these two methods would be useful. Further applications involving the photo-acoustic detection methods are sure to appear in the future because of the sensitivity of the system.The development of inexpensive tunable lasers should also extend the range of applications. Low-angle Laser Light Scattering Detectors The development of lasers has led to the production of another HPLC detection technique namely low-angle laser light scattering (LALLS). The principle of operation of a LALLS detector and the basic theory of light scattering have been described recently,287 and therefore only the most important points have been covered in this review. Light scattering arises from the interaction of visible electromagnetic radiation with a non-uniform medium. When a solute is completely dissolved in a solvent the resultant solution is generally considered to be homogeneous but if the solute has a high relative molecular mass light scattering can be produced by the molecules.The only method suitable for the separation of high relative molecular mass solutes (i.e., polymers and proteins) is size-exclusion chromatography (SEC) and therefore LALLS detectors have been used exclusively with this separation technique. The basic parameter measured in light scattering is the excess Rayleigh factor (RH) which corresponds to the product of the relative molecular mass ( M ) and concentration (c) i . ~ . , Re = cM. In many applications a concentration-dependent detector (e.g. R I or UV) is coupled together in a series with the LALLS detector. This combined detection system then allows the analyst to obtain the relative molecular mass of the solute and calibration graphs do not have to be prepared.The scattered light intensity is highest for small scattering angles and hence scattering is usually measured at very small angles to the incident beam with a photomultiplier. One advantage of using a laser is that the scattered light can be measured at these very small angles typically about 5”. Another advantage of using a laser is that very small scattering volumes (cO.1 pl) are analysed and this reduces problems arising from dust particles. Signals due to dust particles can be differentiated from those of a solute because they appear only as spikes in the chromatogram.288 The intensity of the scattered light can be reduced consider-ably by solute interactions and chromatographic separations should be performed on very dilute samples.Samples must neither absorb the incident beam nor fluoresce and to reduce the probability of such phenomena sources that provide light of a long wavelength are generally used. It has also been shown that Re is dependent on the wavelength of the incident beam and scattered blue light is more intense than scattered light from any other part of the spectrum. There are numerous problems related to using another Table 10. Applications of LALLS detection in HPLC Application Reference MD of dextran polyacrylamide poly(ethy1ene oxide) and proteins . . . . . . . . . . . . . . . 290 Shear degradation of polyolefins . . . . . . . . 291 Characterisation of linear and branched-chain homopolymers . . . . . . . . . . . . .. 292 Branching in polymers . . . . . . . . . . . 293 MD of poly(methy1 methacrylate) . . . . . . . . 294 MD of eye lens proteins . . . . . . . . . . . . 295 Detection of proteins . . . . . . . . . . . . 296,297 detector in series with the LALLS detector. Band broadening is one of the most serious problems and to achieve accurate mass determinations corrections have to be applied to the detector results.289 To obtain accurate results the concentration-dependent detector must be operated within its linear range and the response must be independent of the relative molecular mass of the solute. The RI detector has been used as a concentration-dependent detector in a number of applications but it is important to remember that with this detector the response is dependent on relative molecular mass if the value of the latter is less than 10000.In polymer science it is important to have some knowledge of the relative molecular mass distribution (MD) of a polymer as variations affect its physical thermal and mechanical properties. As the MD of a polymer can be determined very easily from relative molecular mass data the LALLS/ concentration detector approach for monitoring HPLC elu-ates has received considerable attention recently as shown by the number of applications listed in Table 10. As LALLS detection is one of the few techniques that can offer relative molecular mass information further develop-ments can be expected in the future. Instrument manufac-turers are also aware of the value of this detection system and commercial systems are now available .298.299 Radioactivity Detectors Radioactive tracers in the form of radionuclide-labelled compounds are a valuable aid to analysts investigating a wide range of biological and metabolic processes.With the increas-ing desire to become more aware of the effect of chemicals on our environment and the metabolism of drugs greater use is now being made of these radiolabelied compounds. HPLC is an excellent technique for the separation of many of these compounds and until recently the off-line method of trapping fractions and subsequent analysis by liquid scintilla-tion counting had been used extensively. This method is very time consuming and furthermore samples become contami-nated with the scintillation solution. Through development work in recent years these problems have been minimised and on-line monitoring systems are now available.Weak P-emitters such as I T 3H and 32P and the y-emitter 1251 are used extensively for labelling compounds and two major approaches to the problem of monitoring these radionuclides in HPLC eluates have been studied. These have been classified as either homogeneous or heterogeneous systems depending on the manner in which the eluate is mixed with the scintillant. In a homogeneous system the eluate is mixed with a liquid scintillation cocktail before passing through a flow cell positioned in a scintillation counter. In the heterogeneous system the eluate passes through a flow cell packed with a solid scintillator. The essential components for both systems are shown in Fig.17. Before covering some of the developments that have taken place in this field with the homogeneous and heterogeneous systems it is worth looking at the analyser unit of the instrumentation which is common to both techniques. Radio ANALYST AUGUST 1984 VOL. 109 977 Sample injector Column Other detectors Splitter Fraction collector Spectrometer or waste _IL Digital output Chart recorder Solvent -+Pump - Waste Fig. 17. HPLC radioactivity monitoring systems. The components within the broken lines are not required with the heterogeneous system Table 11. Some common solutes and solvents that produce scintillant quenching Strong quenchers Mild quenchers -0COCOR -CH=CWR -I > -Br -C1 -NR -NH, -NHR -OH -NO2 -COOH -SH= -SR -Cl2 -C13 -CHO active compounds if present in a flow cell will disintegrate, and the particle or ray emitted will strike the scintillant and produce a tiny flash of visible light.These flashes are detected in a scintillation counter that incorporates a photomultiplier tube and the number of incident particles or photons are counted over a period of time with a ratemeter. Studies have shown that spurious pulses which arise from stray and cosmic radiation increase background noise levels when working with low levels of radiolabelled compounds. Most modern radioactivity analysers reduce this problem by using two identical photomultipliers that form a sandwich around the cell. Thc photomultipliers are connected to a “coincidence circuit” and the electronics will transmit a signal to the ratemeter only when it receives simultaneous signals from both photomultipliers.The circuitry for a coincidence counter developed for monitoring HPLC eluates has recently been reported .300 The energy windows of some radionuclides are different, and now most analyser systems that are commercially avail-able enable monitoring of two radionuclides to be performed simultaneously. Examples of these radionuclide combinations include 32P and 14P with 3H and 14C with 3H. It is impossible to monitor ‘2~I-labelled compounds in combination with other radionuclides. Homogeneous detection systems Homogeneous detection systems were among the first to be developed for monitoring radioactivity in HPLC eluates. From Fig.17 it can be seen that much more equipment is required for this type of system and involves post-column reaction instrumentation. For scintillation to occur the labelled compound must be brought into contact and mixed with a solution containing a scintillant (scintillation cocktail). The selection of chromatographic eluents and scintillant solvent is important as they must be miscible and produce minimum radioactivity quenching. Some solutes and solvents that cause radioactivity quenching are shown in Table 11. Scintillation cocktails containing naphthalene - 2,5-diphenyloxazole (PPO) - 1,4-bis-2-(5-phenyloxazolyl)benzene (POPOP) in dioxane have been recommended and it has been shown that the counting efficiency improves as the proportion of mobile phase to dioxane decreases.301 This can lead to undesirably high scintillant flow-rates and in general these authors quoted optimum scintillant to eluent ratios of 1 5 for chromatographic solvents such as ethyl acetate and 10 1 for aqueous mobile phases.Because addition of the scintillant to the column eluate contaminates the sample a splitter should be fitted prior to the mixing tee in any homogeneous system if pure samples are required. The design of a flow cell intended for use in any radioac-tivity system is important as the sensitivity of these detectors is dependent on the total amount of activity in the cell. The outcome of this is that relatively high flow cell volumes (200-1000 p1) have been used in homogeneous detector systems and these have been constructed from long lengths (1.5-2 m) of narrow-bore transparent tubing such as glass, PTFE or polypropylene.The choice of material is governed by such factors as background noise level and transparency. Although glass is generally preferred these cells are more difficult to construct. Radioactive potassium in glass can increase background noise levels; this problem can be minimised by using tubing with low potassium levels. Pre-treatment of glass with 10% orthophosphoric acid has also been recommended to prevent interaction of the surface with the scintillant. Further phosphorescence from cell materials can produce problems if the cell is not maintained in complete darkness. Any increase in flow-rate in a homogeneous system will reduce the over-all sensitivity of the detector as the decreased transit time of the solute will not allow a sufficient number of counts to be obtained.This is an important factor in quantitative analysis as the precision of any estimate of radioactivity depends statistically on the number of events contributing to that measurement. In practical terms at least 100 counts must be recorded. Factors controlling quantitative precision have been discussed by Sieswerda et d . 3 0 2 The specific activity of the labelled compound and the counting efficiency for a particular radionuclide will also govern the sensitivity of the detection system. Results published indicate that a counting efficiency of 70% can be obtained with I4C radionuclides and 40% with 3H radio-nuclides from homogeneous systems.Care must be taken in attempting to establish the lower limit of detection for these systems as both disintegrations per minute (d.p.m. a measure of activity of the labelled compound used) and counts per minute (c.p.m. measured activity in the flow cell) are quoted in results. Calculations based on recorded figures indicate that it is possible to detect solutes in the low nanogram range. Homogeneous radioactivity detection methods have been used for the analysis of gibberellin benzyl ester and indolyl-3-acetic and butyric acids.3”’ More recently a study on the metabolism of purine has been rep0rted.3~)~ The purine nucleotides nucleosides and bases were chromatographed on reversed-phase and anion-exchange packing materials using gradient elution.With the anion-exchange system it was reported that the counting efficiency decreased with increase in the pH and ionic strength of the eluent. Dugger and Orwig3O4 also used gradient elution to separate and study biotransformations of a [3H]fluproquazone and a 4,4-[ *4C]methylenebisphenol compound. An outline of some of the principles and major develop-ments with homogeneous systems have been covered in this review and more detailed work can be found in publications by Reeve and Crozier301 and Harding et al.305 Heterogeneous detection systems As stated earlier these systems involve passing the HPLC eluate through a flow cell packed with radioactive particles. Apart from requiring less instrumentation (see Fig. 17), heterogeneous methods are considered generally to have other advantages over homogeneous systems including cheaper operating costs and no sample contamination.Early studies did show however that with the heterogeneou 978 ANALYST AUGUST 1984 VOL. 109 Table 12. Typical counting efficiencies for solid scintillators in flow cells Efficiency YO Scintillator* 3H 14C Glass (Ce) . . . . . . . 0.6 30 Plastic . . . . . . . . 0.7 32 CaF,(Eu) . . . . . . . . 1.1 40 Yttriumsilicate (Ce) . . . . 1.0 45 Anthracene . . . . . . 1.0 40 * (Ce) = cerium activated; (Eu) = europium activated. systems only very poor counting efficiencies for weak p-emit-ters (14C and 3H) could be obtained. Some typical results are shown in Table 12. The sensitivity of any radioactivity monitoring technique is related to the counting efficiency and because homogeneous systems give higher efficiencies heterogeneous systems have not been so popular.This situation could change in the future as results from a recent extensive study on heterogeneous systems published by Mackey et al.300 have shown that under optimised conditions a counting efficiency for 14C of greater than 70% could be obtained which is comparable to that obtained for homogeneous systems. Cell design was found to be one of the important parameters in optimising conditions. With scintillators emitting in the 395-nm region no difference in counting efficiency was observed from quartz or Pyrex glass flow cells. An interesting observation by Mackey et al. was that the centres of the photomultiplier tube windows were approximately twice as sensitive as the peripheral areas.Consequently cells were designed to take advantage of this fact and a Z-shaped cell with a volume of 60 pl was the most efficient in terms of both radioactivity counting and chromatographic resolution. Particle size of the scintillator was the other important parameter in optimising counting efficiency. A variety of solid scintillators can be used in heterogeneous systems as shown in Table 12 and in this study several commercially available lithium glass scintillator beads of different particle sizes were evaluated. Highest efficiencies were obtained from all scintil-lators when the smallest particle size of 38-63 pm (160-180 mesh) was tested. Smaller particle sizes could not be ade-quately tested because the sieving of the friable beads introduced fines.From this work it was concluded that the increase in efficiency is due to (a) a greater area of scintillator being exposed to the radiation and (b) as the smaller particle size approximates to the average range of 14C radiation in solution (14 pm) most points in the solution are within this distance from the glass scintillator surface. These findings have also been confirmed by Everett,306who used the same type of solid scintillators and some typical results are shown in Fig. 18. As can be seen Everett confirmed that an increase in coincidence resolving time of the photomultiplier improves the counting efficiency and a considerably higher efficiency (up to 6%) can be achieved for 3H radionuclides if scintillators of smaller particle size are used.In this study it was shown that good efficiencies could be obtained for 1251 (70%) and 32P (10%) radionuclides. Other factors including refractive index pH and UV absorbance of the eluate were also investigated by Mackey et ~ 1 . 3 0 0 Changes in refractive index or pH did not affect counting efficiency which is a problem with homogeneous systems and only eluents or solutes in the eluate that absorb light at 400 nm were found to have any significant effect on efficiency. The system described in this section was used to profile some alkylethoxylate urinary metabolites and the only prob-lem encountered was irreversible adsorption of some solutes on the glass scintillator beads after extended use. Repacking I 8 6o t 1 I ' 0 60-80 1OC-120 140-160 180-200 80-100 120-140 160-180 Solid scintillator mesh range Fig.18. Effect of solid scintillator particle size and coincidence time on counting efficiencies of 1% and -iH p-emitters. Coincidence time: bold lines 100 ns broken lines 20 ns. Published with permission of Packard Instruments Ltd. HPLC column A Lead shielding It I I- -Ratemeter Sodium iodide crystal 4 Recorder 1 It- PTFE capillary tubing Fig. 19. Basic design of an HPLC y-detector the cell is suggested as adsorption or desorption of contami-nants could not be prevented by washing the beads with organic solvents or solutions of salts or dilute acids or by silanisation of the scintillation beads. Sodium iodide crystal detectors Sodium iodide crystal detectors can be used to monitor radiations from y-radionuclides and these have also been developed for detecting radioactivity in HPLC eluates.As opposed to the systems described previously the y-detector does not require any post-column reagents or instrumenta-tion. Sodium iodide crystals have been used for many years to detect radiations and as the y-emitter 1251 is used extensively as a radionuclide in biochemical studies interest in using this type of detector to monitor chromatographic separations has increased. The basic design of an HPLC y-detector is illustrated in Fig. 19. The PTFE capillary tubing is inserted in a steel tube that passes through a drilled sodium iodide crystal. Construction of these cells is therefore very simple and with the capillary tubing acting as the flow cell very low cell volumes (<5 pl) can be achieved.Two applications of an HPLC y-detector have been reported in the literature recently and as both were used to monitor preparative HPLC separations no information regarding sensitivity etc. is available. In the first of these applications a size-exclusion packing material was used to separate and purify labelled proteins.3O7 Recoveries of purified material based on initial radioactivity were reported to be in excess of 90%. In the other application a reversed-phase separation system was used to purify some 1251-labelled compounds and it was reported that even wit ANALYST AUGUST 1984 VOL. 109 979 Table 13. Summary of HPLC radioactivity monitoring systems Homogeneous Parameter systems Radionuclides detected .. 14C,3H Eluent compatability . . Some restrictions Preparative . . . . . . Yes require splitter Commercially available . . Yes Currentcosts . . . . f7000 Relative operating costs . . High-continual use of scintillation solutes and solvents Heterogeneous systems y-Detectors 14C,3H1 1251,32p 1251 Few restrictions Few restrictions Yes some adsorption Yes Yes No f 5 000 -Medium-scintillators Low problems need to be replaced occasionally the large amounts of material being injected no radioactive contamination of the column could be detected.308 Both papers reported that by comparison with older methods used to purify 125I-labelled compounds the HPLC y-detection system reduced the analysis time considerably.HPLC radioactivity monitors offer excellent selectivity and are invaluable to analysts involved with the analysis and separation of radiolabelled compounds. Many of the develop-ments discussed in this review are now incorporated into detectors that are commercially available. Two comparative studies published recently309.310 report that similar perfor-mances can be obtained from these instruments and that in comparison with the older method of static counting of trapped fractions some loss in efficiency is experienced but increased resolution and shorter analysis times can be achieved. At present it is impossible to cover all possible applications with one particular system although most manufacturers produce a system with interchangeable flow cells that permit homogeneous and heterogeneous detection.The choice of instrument will depend on a number of factors and these are summarised in Table 13. Finally with many metabolites occurring at very low levels (ng ml-I) there is a genuine need to improve the sensitivity of these detection systems. Considerable achievements have been obtained with the heterogeneous systems which make them comparable to homogeneous systems but further developments such as the introduction of micellar liquid scintillators311 can still be envisaged in the future. Detectors Involving a Phase Change of the Eluate The final section of this review includes HPLC monitoring devices that require removal of the chromatographic solvent prior to detection of the solute Le.a phase change of the eluate. Apart from one novel instrument the mass detector all the other developments have utilised a gas-chromatographic detector. The use of a mass spectrometer as an HPLC detector has been included in this section as most applications require a phase change of the eluate in order to detect the solute. Mass Detector Although the evaporative analyser (now called the mass detector) was developed and patented in 1966 by Ford and Kennard,312.313 its potential as an HPLC detector was realised only after a comprehensive study had been completed by Charlesworth314 in 1978. In his paper the construction, performance and theory of operation of the mass detector were described in considerable detail. The main features of the mass detector are shown in Fig.20 and the detector works on the following principle. The chromatographic eluate is nebulised and evaporation of the solvent produces finely divided solute particles that pass through a light beam. The scattered light produced by these particles is detected by a photomultiplier that is placed at a fixed angle to the incident light beam. Nebuliser gas Evaporator Solvent stream Nebuliser Light source r- A I Light source Light trap - Section AA Photomultiplier Cross-section AA Fig. 20. Main features of the mass detector. Published with permission of Applied Chromatography Systems Ltd. Charlesworth found that the particles deflected light pre-dominantly by reflection and refraction and by measuring this light at 1350 to the incident beam he found that results were almost independent of the refractive index of the solute.Calibration graphs obtained from this instrument were found to be linear over the range 1 x 10-4-1.5 x 10-3 g cm-3. The explanation given for the departure from linearity above and below these limits was that the nebulisation/evaporation process produces particles of varying sizes that promote other light-scattering effects viz. Mie and Rayleigh scattering. Evaporation of the solvent is maintained by the heated secondary air supply and the detector response was found to be dependent on the evaporation temperature. This effect was particularly noticeable when solutes with volatilities similar to that of the solvent were analysed and therefore the detector is best suited to applications where the solutes are less volatile than the eluent 980 Flamephotometric(FPD) .. . . Sulphurand phosphorus compounds Electron-capture (ECD) . . . . Halogens and nitro compounds Thermionic(TSD) . . . . . . Nitrogenand phosphorus compounds ANALYST AUGUST 1984 VOL. 109 All highly specific ’ and sensitive The interesting and important fact that arose from this work was that the detector response was independent of the chemical structure of the solute and thus it performs not as a concentration-dependent analyser but as a mass-dependent detector. Therefore even when pure standards of a com-pound ale not available for calibration purposes the concen-tration of this compound can be obtained with this mass detector.Also if volatile solutes are excluded the mass detector can also be considered as a universal detector and for many years there has been a requirement for this type of detector. An instrument manufacturer was aware of these advantages and by 1981 a pre-production model had been built. Trials with this detector were undertaken by Macrae and Dick who carried out some analyses of carbohydrates in food samples.315 Results from these analyses were compared with those obtained with a refractive index detector i.e. the detection system most commonly used for the analysis of these compounds in HPLC eluates. Apart from having a more limited linear range (10-200 yg on-column for glucose and maltose) the mass detector proved to have three major advantages over the RI detector increased sensitivity, improved stability and the ability to operate with gradient elution.The sensitivity was reported to be dependent on the operating conditions and for common sugars 2-3 pg on-column produced a readily quantifiable peak. This sensitivity is about a factor of ten better than most RI detectors. The ability of the mass detector to operate with gradient elution was found to be a very beneficial feature. With gradients containing varying proportions of water and aceto-nitrile improved separations of monosaccharides were obtained whilst verbascose (a pentasaccharide) was still eluted within 25 min. This separation could not be obtained in one analysis under isocratic conditions. In analyses that were performed with these aqueous systems it was reported that the mass detector noise level was dependent on both the evaporating temperature and the amount of water in the eluent and that the response was temperature dependent.These were considered not to be serious problems as acceptable noise levels and good sensitivi-ties for specific applications could be achieved by choosing suitable detector operating conditions. Another recent publication316 has illustrated that the mass detector is particularly useful in applications where solutes cannot be analysed by techniques such as gas chromato-graphy because of their instability at elevated temperatures. Triglycerides are susceptible to thermal decomposition but no evidence of this was detected when some of these compounds were analysed with the mass detector.The reason put forward for this was that the separations could be achieved with a volatile eluent (acetone - acetonitrile) which permitted a very low detector evaporation temperature. A mass detector suitable for microbore HPLC has been reported recently.317 Carbon dioxide was used to vaporise the solvent and a helium - neon laser was employed to illuminate the solute particles. Some triglycerides n-alkylbenzenes and fatty acid methyl esters were analysed with this instrument and the over-all performance was similar to that of the mass detector previously described. The commercial version of the mass detector has been available for a relatively short time and the work published so far has shown that this technique is suitable for a variety of applications.It is certainly a useful addition to the range of available HPLC detectors and can provide a much improved alternative to the well established RI detector. The instrument currently costs just under &3 000. Gas-chromatographic Detectors HPLC was introduced at a time when gas chromatography (GC) was already a well established separation technique. Some of the GC detectors then available could offer good sensitivity and/or selectivity (see Table 14) and it was not surprising therefore to see several groups of workers attempting to modify these detectors for the purpose of monitoring HPLC eluates. All GC detectors require the solute to be in the vapour phase and before they can be used as an HPLC monitoring device the eluates must undergo a phase change to permit detection of the solute.In attempts to overcome this incom-patibility problem two approaches have been studied (a) transport systems and (b) direct volatilisation of the eluate. With the transport system the HPLC eluate is either dropped or sprayed on to some moving carrier i.e. wire, chain belt disc etc. and the eluent is removed in an evaporation oven. The transporter then coated with the solute passes to the detector. Some systems were developed whereby a pyrolysis or combustion unit was incorporated prior to the detector. FIDs ECDs and TSDs have been used with these transport methods and a wire transport system was commer-cially available for several years. A detailed account of the development of these systems has been published by Scott.2 Apart from a recent patent for a system that incorporated a porous belt spun from a quartz thread,318 interest in transport detectors has declined mainly because all the systems so far devised have failed to justify their original promise.Mechanical instability i. e.) breaking wires selective solvent removal poor collection efficiency and coating of the solute on the carrier and increased band spreading are some of the major reasons why transport systems have not been successful. The direct volatilisation technique is the other alternative approach that has been developed for interfacing GC detec-tors with HPLC systems. With this method the components of the HPLC eluate viz. eluent and solute are volatilised and the atomised solute species are monitored directly by the detector.Detection of volatile solutes containing halogens, nitrogen phosphorus or sulphur are eminently suited to this process and therefore ECDs FPDs and TSDs detectors have been employed. The principle of this technique was developed by Nota and Palombari,319 and a nebuliser was used to produce the atomised species which were subsequently detected with an electron-capture detector. Willmott and Dolphin320 improved the design of this ECD system by replacing the nebuliser with a heated transfer line which was situated in an oven. By this method the volume increase involved in vaporisation forces the vapour into the ECD which is enclosed in a separate oven. Table 14. GC detectors used as HPLC monitoring device ANALYST AUGUST 1984 VOL.109 981 A nitrogen purge was used to prevent back-diffusion of the condensed vapour into the ECD. This detector was taken up commercially and marketed in 1976. The sensitivity of this system towards electron-capturing compounds was reported to be extremely high 1.1 x 10-1° g ml-1 but the linear dynamic range was poor? Many of the subsequent developments have been reviewed by Brinkman322 and readers will appreciate that the HPLC -ECD systems have failed when used for the analysis of high-boiling compounds. A more recent paper323 has shown, however that with an improved interface which creates more efficient heat transfer in the volatilisation process high-boiling halogenated aromatics could be detected. For a variety of organochlorine pesticides detection limits of 8-100 pg were reported.Systems involving the use of FPDs for monitoring phos-phorus and sulphur have been covered in Part I when atomic-emission based detector systems were discussed (see Spectroscopic Detectors-Atomic Emission). Recently a dual flame TSD detector with a micro-HPLC system has been de~cribed.~Z~ This detector was designed so that it spatially separates the fundamental flame process such as nebulisation and desolvation from the analytical measure-ment process. It was shown that with this detector high concentrations of organic modifiers in the mobile phase could be tolerated without increasing matrix interferences. For some phosphorus compounds analysed with this HPLC - TSD system the detector was reported to be linear over a range of 2-750 ng of injected phosphorus and detection limits were nearly two orders of magnitude less than for a similar micro HPLC - FPD system.An injected amount of 2 ng of phosphorus could be detected clearly above the background noise level. The authors hope that through optimisation of the detector operating parameters selective detection of nitrogen-containing compounds can be achieved in the future. From the facts that have emerged in this study a critical assessment of the situation is that apart from a possible saving in sample preparation time there is little to be gained by using GC detectors to monitor solutes in HPLC eluates. Virtually all the applications that can be performed with HPLC - GC detector systems can be accomplished more easily by gas -liquid chromatography.Further non-volatile and thermally labile compounds are not suitable for analysis by gas - liquid chromatography and therefore no useful purpose can be served by interfacing HPLC systems with GC detectors. It is generally recognised that HPLC is an ideal separation technique for these particular types of solute and therefore other detection techniques should be appraised. Mass-spectrometric Detectors Over the past decade a considerable amount of research time and money has been expended in attempting to produce HPLC - mass spectrometer (LC - MS) systems. The combina-tion of interfacing problems and the use of a variety of MS ionisation techniques has generated many papers and reviews in the literature. A number of recent reviews have covered LC - MS developments and applications in depth325-328 and this survey will therefore be an over-all summary of the situation.As the principles theory and instrumentation of mass spectrometry are well documented these topics have also been omitted here. The interest in LC - MS stems from the fact that both chromatographers and spectroscopists could achieve consider-able analytical information from this technique. As shown in Table 15 the former group of scientists regard the mass spectrometer as an “ideal” detector whereas the latter consider HPLC to be a “perfect” separator. The mass spectrometer will function as an analyser only if the solutes being examined exist in the gaseous phase and this technique is therefore ideally suited to monitoring gas-chromatographic (GC) effluents.With LC the solutes are in the liquid phase and LC and MS systems are not compatible unless the solvent is removed and the solute is vaporised. This is exactly the same situation experienced with other HPLC -GC detection systems (see Gas-chromatographic Detectors) and with the numerous problems that existed GC was considered to offer few advantages. Development of LC - MS systems has continued for two reasons. Firstly considerably more important analytical information can be obtained and secondly these systems are ideally suited to the analysis of non-volatiles. MS detection of these compounds is possible because they have sufficiently high vapour pressure~.3~~,~3~j To interface HPLC with MS the following features are essential (a) high enrichment of solute to solvent; (b) high collection and transference efficiency; (c) produce minimum band spreading; and (d) permit analysis of samples with low volatility.In attempting to produce an LC - MS system that will match up to these requirements two methods of sample introduction have been studied namely transport devices and direct liquid injection systems. Tramp o rt devices Much of the work in developing this type of interface was based on knowledge previously gained from other HPLC - GC detector transport devices. One of the first LC - MS systems was produced by Scott et al.331 and incorporated a moving wire. McFadden et ~1.332 decided to replace the wire with a polyimide belt to achieve better collection efficiency of the solute and hence improve the sensitivity of the system.The first commercially available LC - MS interface was based on this technique and a diagram of the system is shown in Fig. 21. With this system sufficient solute ions were available in the MS source to permit either electron impact (EI) or chemical ionisation (CI) detection techniques to be used. The main disadvantages of this technique are that column flow-rates are restricted to a maximum of about 1 ml min-l, and eluates containing polar modifiers or buffers produce a drastic reduction in sensitivity. The latter problem means that many applications based on reversed-phase separations are not suited to this technique. Table 15. Potential advantages of LC - MS Advantages of using MS as an HPLC detector: (a) Structural identification.(b) Excellent sensitivity and selectivity particularly if multiple or (c) A variety of ionisation processes can be used to obtain different (d) Total ion current traces give a very close approach to universal single ion monitoring is performed. spectra and hence better discrimination. detection. Advantages of using LC as a separator: (a) Offers an alternative method for introducing samples into the MS source especially as HPLC is a useful separation technique for non-volatile and thermally labile compounds. Ion source LC effluent Infrared reflcjctor spliL vaporiser Vac-’ocks / \ I option Flash / l-nLddv@ I I I A A ““I “VYUl Clean-up wheels heater Pump Pump 1 solvent Waste Fig.21. Schematic diagram of the moving belt LC - MS interface. A, Scrubbing solution enters; B extraction of salt deposits; and C waste. Published with permission of Finnigan-MAT Ltd 982 ANALYST AUGUST 1984 VOL. 109 Yorke and co-worker~333~334 modified this type of transport system by positioning the moving belt interface so that it entered directly into the source of the mass spectrometer. In comparison with the above-mentioned technique better handling of aqueous solvents was reported and the sensitivity appeared to be improved. This interface can be used with both magnetic and quadrupole mass spectrometers and two com-mercial systems using this concept are currently available. To extend the range of HPLC separations performed with these transport systems microbore columns could be used as they operate with very small amounts of eluent.Recently Games et a1.335 studied the effects of band broadening with this type of system and reported that the mass spectrometer is an excellent low dead volume detector with a time constant of 0.08 s. Although this type of LC - MS system has been used for some applications,332J36 it is generally agreed that microbore columns should be used in interfacing techniques that do not use a transport device i.e. direct liquid injection systems. Continuous flow extraction schemes using air337 and solvent segmentation338 techniques have also been studied in attempts to overcome some of the problems experienced using reversed-phase eluents and buffers with belt transport systems. With solvent segmentation techniques solutes are extracted from the column eluate into a volatile solvent e.g., methylene chloride and then deposited on to the moving belt interface.Some success has been achieved with this type of system but incorporating an extractor can introduce band broadening and affect the chromatographic resolution. Direct liquid injection systems The alternative approach to belt transport interfacing is to deliver the column eluate directly into the MS source and use chemical ionisation techniques as devised by McLafferty and co-workers.339J40 Methods based on this principle have been called direct liquid injection (DLI) systems. Apart from one recent method of trying to obtain electron impact (EI) spectra with an ultra-micro HPLC - MS system,341 chemical ionisation (CI) techniques have to be used as the amount of liquid entering the mass spectrometer increases the source pressure to a level that precludes the use of other ionisation techniques.Some analysts consider this to be a disadvantage because CI spectra in general provide limited structural information. However an advantage of CI LC - MS systems is that they are ideally suited to the use of polar solvents as these promote protonation of the solute and production of the pseudo molecular ion (MH+). This ion is prominent in most CI spectra and in other MS applications it is usually generated by the addition of a reagent gas but with LC - MS the solvent vapours can be used successfully for this purpose. Apart from using pure polar solvents polar eluents containing up to 70% of water and some buffers can be tolerated and therefore CI DLI interfaces are suited to applications requiring reversed-phase separations.Another advantage of these systems is that the technique is amenable to the analysis of non-volatile and thermally labile compounds. The explanation for this is that in the solvent evaporation process adiabatic expansion occurs which pro-motes the formation of droplets and a decrease in tempera-ture. Under these conditions and with the lifetime of the droplets being 0.5-4 ms solute molecules can be transported to the ion source without degradation. A paper published recently has discussed some of the theoretical aspects of this process.342 A major problem with DLI systems is that most mass spectrometers can only accept liquid flow-rates of less than 50 pl min-l which is considerably less than those associated with conventional 4 mm i.d.HPLC columns (i.e. 1-2 ml min-1). This reduced flow-rate can be achieved by using a splitter at the end of the column but then only about 10% of the solute originally injected will be available for analysis. To increase the sensitivity of this type of system whilst still retaining the use of conventional HPLC columns eluate flow switching can be performed. Kenyon et a1.343 have shown that with this technique positive and negative chemical ionisation could be used and detection limits for the scanning mode and the multiple ion detection mode were reported to be 10 ng and 100 pg of injected material respectively.Although an automatic flow switching method can be used this system is not ideally suited for the analysis of non-routine samples unless the analyst devotes 100% of his time to running the instrument. Several groups of workers believe that the best method of performing direct liquid injection LC - MS is to use microbore columns as they operate at flow-rates of 5-50 p.1 min-1. This type of system allows 100% transfer of injected material to the mass spectrometer and although only sub-microgram amounts of solute can be handled by these columns without generating overloading problems detection limits are repor-ted to be about 100 times greater than can be achieved by any other HPLC - MS interface. In early studies microbore column eluates were introduced into the mass spectrometer via the glass capillary of the probe inlet.With current methods the glass capillary has been omitted and the column eluate passes through a pinhole (1-3 pm) in a diaphragm and directly into the ion source. This latter method is favoured as dead volumes are reduced considerably and the eluate enters the source as a fine spray. Many chromatographers are not in favour of this technique as the microbore columns can generate a number of problems. For example the columns are difficult to pack and their efficiency deteriorates rapidly with extended use. Admittedly this is a new area of separation technology and these problems could be resolved in the future. The use of microbore columns also leads to extremely long separation times and although a flow-rate splitter has been devised to reduce this problem,344 this is another factor that has to be considered.Vestal and co-workers345J46 introduced another type of DLI interface that has sometimes been referred to as the thermospray detector. One primary advantage of the system is that it is capable of handling total eluate volumes from conventional HPLC columns. Vaporisation of the eluate is achieved with rapid heating by four oxy-hydrogen torches. This process produces a jet of vapour and an aerosol and therefore chemical ionisation spectra can be obtained from non-volatiles. It is evident that users of all types of LC - MS interfaces, including those which are commercially available experience problems with reproducibility and maintaining continuous operation of the mass spectrometer.This problem has been attributed to the deposition of high-boiling residues in the source. A recent paper suggests however that silica is a major contribution to source contamination.347 It was shown that with a DLI interface modification of the pinhole in the diaphragm changed the eluate jet and the mass fragmentation pattern. Silica particles originating from the column and through dissolution of the column packing material were detected on the diaphragm which altered the shape of the orifice. As source contamination requires several hours to remove strip clean and replace a source this is a serious problem and certainly requires considerable attention in future developments. Sadly it is fair to say that although the mass spectrometer is the chromatographer's ideal detector innumerable problems still exist in attempting to interface the two techniques.The LC - MS systems that are currently available can perform only certain types of analysis and their continuous use imposes serious source contamination problems. Incompatibility between the requirements of HPLC and MS is the major problem and although the use of off-line HPLC - MS techniques might appear to be a defeatist attitude it is the only method available that allows the analyst to exploit all the advantages of the mass spectrometer as an HPLC detector ANALYST AUGUST 1984. VOL. 109 983 Concluding Remarks HPLC is an extremely valuable analytical technique but its full potential can only be realised if suitable detectors are available.The extent of this review shows how much research effort has already been devoted to the development of HPLC monitoring devices and yet the “ideal” detector remains an elusive goal. This review has shown that there have been three distinct approaches to detector development. Firstly new technology such as lasers diode arrays and computers has been incorpor-ated into instruments exploiting well established detection techniques. From this type of work improvements in sensitiv-ity selectivity and versatility have been achieved. ,4n example of the latter is the introduction of multi-wavelength UV detectors that permit the simultaneous monitoring of several wavelengths absorbance ratioing and profiling and scan spectra of eluted components to be obtained.The second approach has been the development of detec-tion systems that monitor physical or chemical properties of the solute that have not been previoulsy studied. Some of these systems have been o f only academic interest whereas others are valuable alternatives or additions to existing techniques. The most useful of these new techniques include radioactivity electrochemical low-angle laser light scattering and mass detectors and these are all now commercially available. The final type of investigative work that has been carried out involves the use of well established detection techniques with modifications of other parts of the HPLC system or the sample being studied. Chemiluminescence indirect photo-metric detection and sample derivatisation are examples of this type of approach although the last was not included in this review.Without doubt chromatographers have benefited from all of these approaches in devising more superior detection methods but there will be demands from analysts for continued improvements. For this reason more interesting developments of HPLC detection systems will certainly be seen over the next few years. This review is based on a dissertation initially submitted to the Council of the Royal Society of Chemistry in fulfilment of conditions required for admission to CChem MRSC via their counselled experience route. I thank Mr. Brian Wheals and Dr. Anthony Bowd who acted as my two counsellors for their guidance and the time that they have given to me.Finally I am indebted to my wife Ruth and Sandra Tye for typing the manuscripts. 274. 275. 276. 277. 278. 279. 280. 281. 282. 283. 284. 285. References White P. C. Analyst 1984 109 667. Simpson C. F. Editor ‘.Practical High Performance Liquid Chromatography,” Heyden London 1976. Thente K. and Silverbage S . Focus 1982 5 12. Boehme W. Chromatogr. Newsl. 1980 8 31. DiCesare J. L. andEttre L. S . J . Chromatogr 1982,251 1. Yeung E. S. Steenhoe L. E. Woodruff S. D. and Kuo, J. C. Anal. Chem. 1980 52 1399. Drake A. F. Could J. M and Mason S . F. J . Chrornatogr., 1980 202 239. Westwood S. A. Games D. E and Sheen L. J . Chromat-ogr. 1981 204 103. Schmermund J. T. and Locke D. C. Anal. Lett. 1975 8, 611. Locke D.C. Dhingra B. S . and Baker A. D.,Anal. Chern., 1982 54 447. Konash P. L. and Bastiaans G. J Anal. Chem. 1980 52, 1929. Voigtman E. Jurgensen A and Winefordner J . Anal. Chem. 1981 53 1442. 286. 287. 288. 289. 290. 291. 292. 293. 294. 295. 296. 297. 298. 299. 300. 301. 302. 303. 304. 305. 306. 307. 308. 309. 310. 311. 312. 313. 314. 315. 316. 317. 318. 319. 320. 321. 322. 323. 324. 325. 326. 327. 328. 329. 330. 331. 332. Oda S . and Sawada T Anal. Chem. 1981 53 471. Martin M Chromatographia 1982 15 426. Kaye W. Havelik D. J. McDaniel J. B. J . Polym. Sci. Polym. Lett. 1971 9 695. Hamielec A. D. Ederer H. J. and Ebert K. H. J. Liq. Chrornatogr. 1981 4 1697. Fukutomi M.Fukada M. and Hashimoto T. Toyo Soda Kogyo K. K. Japan 1980 24 33. Rooney J. G. and Ver Strate G. in Cazes J. Editor “Liquid Chromatography of Polymers and Related Materials 111,” Marcel Dekker New York 1981 p. 207. Gallot Z. zn Cazes J. Editor. “Liquid Chromatography of Polymers and Related Materials 111,” Marcel Dekker New York 1980 p. 113. Jordan R. C. and McConnell M. L. ACS Symp. Ser. 1980, No. 138 107. Jenkins R. and Porter R. S J . Polym. Sci. Polym. Lert., 1980 18 743. Bindels J . G. DeMan B. M. and Hoenders H. J . J . Chromatogr. 1982 252 255. ‘Takagi T. J . Biochem. 1981 89 363. Maezawa S . and Takagi T. J . Chromatogr. 1983 280 124. Jordan R. C. J . Liq. Chromatogr. 1981 3 439. McConnell M. L. Am. Lab. 1978 10 63. Mackey L. N. Rodriguez P.A and Schroeder F. B. J . Chromatogr. 1981 208 1. Reeve D. A. and Crozier A. J . Chrornatogr. 1977,137,271. Sieswerda G. B. Poppe H. and Huber J. F. K. Anal. Chirn. Acta 1975 78 343. Webster H.K. and Whaun J. M J . Chromatogr. 1981,209, 283, Dugger H. A. and Orwig B. A . Drug. Metah. Rev. 1979, 10 247. Harding N. G. L. Farid Y. Stewart M. J. Shepherd J . and Nicoll D. Chromatographia 1982 15 468. Everett L. J. Chromatographia 1982 15 445. Von Stetten O. and Schlett R. J . Chromatogr. 1981 218, 591. Von Stetten O. and Schlett R. J . Chromatogr. 1983 254, 229. Frey B. M. and Frey F. J. Clin. Chem. 1982 28 689. Kessler M. J. J . Chromatogr. Sci. 1982 20 523. Ewer M. J. and Harding N. G. L. in Crook M. A . . and Johnston P. Editors “Liquid Scintillation Counting,” Volume 3 Heyden London 1974.Ford D. L. and Kennard W. J . Oil Colour Chern. Assoc., 1966 49 299. Ford D. L. and Kennard W. Aust. Pat. Appl. No. 33406. Charlesworth J. M. Anal. Chem. 1978 50 1414. Macrae R. and Dick J. J . Chrornatogr. 1981. 210 138. Macrae R. Trugo L. C. and Dick J. Chromatographia, 1982 15 476. Stolyhwo A. Colin H. and Guiochon G. J. Chrornatogr., 1983 265 1. Dixon J. B. and Hall. R. C. U.S. Pat. 4 271 022. 1981. Nota G. and Palombari R. J . Chromatogr. 1971 62 153. Willmott F. W. andDolphin R. J J. Chromatogr. Scz. 1974, 12 695. “LC - EC Chromatograph,” Pye Unicam Cambridge 1976. Brinkman U. A. T. in Zlatkis A. and Poole C. F. Editors, “Electron Capture-Theory and Practice ,” Elsevier Amster-dam 1981 p.407. DeKok A . Geerdink R. B. Brinkman U. A. T. J. Chromatogr. 1982 252 101. McGuffin V. L. and Novotny M. J . Chromatogr. 1981,218, 179. Arpino P. J. and Giuochon G. Anal. Chem. 1979 51, 682A. McFadden W. H. J . Chromatogr. Sci. 1980 18 97. J. Chromatogr. Chromatogr. Rev. 1982 251. “Spectra,” Finnigan-MAT Publication CA USA 1983 No. 9. Blakley C. R. McAdams M. J. and Vestal M. L. J . Chromatogr. 1978 158 261. Arpino P. J. and Guiochon G. J . Chromatogr. 1982 251, 153. Scott R. P. W. Scott C. G. Munroe M. and Hess J. J . Chrornatogr. 1974. 99 395. McFadden W. H. Schwartz H. L. and Evans S . J. Chromatogr. 1976 122. 389 ANALYST AUGUST 1984 VOL. 109 984 333. 334. 335. 336. 337. 338. 339. 340. Yorke D. A. in “HPLC Interface for LC - MS,” VG Publication VGOLC V. G. Organic Ltd. Altrincham Che-shire 1979 No. 9. Yorke D. A. Millington B. S . and Burns P. in Quayle A., Editor “Advances in Mass Spectrometry and Allied Topics,” Volume 8 Heyden London 1981 p. 1819. Games D. E. Hewlins M. J. Westwood S. A. and Morgan, D. J. J . Chrornatogr. 1982 250 62. Games D. E. J. Chrornatogr. 1982 251 165. Karger. B. L. Kirby D. P. and Vouros P. J . Chrornatogr. Sci. 1980 18 111. Kirby D. P. Vouros P. Karger B. L. Hidy B. and Peterson B . J. Chrornatogr. 1981 203 139. Baldwin M. A. and McLafferty F. W. Org. Mass. Spec-trorn. 1973 7 1353. Arpino P. J. Dawkins B. G and McLafferty F. W. J. Chrornatogr. Sci. 1974 12 574. 341. 342. 343. 344. 345. 346. 347. Takeochi T. Ishii D. Saito A. and Ohki T. J . High Resolut. Chrornatogr. Chromatogr. Cornmun. 1982 5 91. Deiden M. Juin C Arpino P. J. Bounine J . P. and Guiochon G. J . Chrornatogr. 1982 251 203. Kenyon C. N. Melera A. and Erni F. J . Anal. Toxicol., 1981 5 216. Krein P. Devant G. and Hardy M. J. Chrornatogr. 1982, 251 129. Blakley C. R. Carmody J. J. and Vestal M. L. Anal. Chern. 1980 52 1636. Blakley C. R. and Vestal M. L. Anal. Chern. 1983,55,750. Mauchamp B. and Krein P. J. Chrornatogr. 1982 236 17. Paper A4142 Received January 27th 1984 Accepted February 6th 198

ISSN:0003-2654    

DOI:10.1039/AN9840900973

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST AUGUST 1984 VOL. 109 985 Gas - Liquid Chromatographic Method for the Determination of Peracids in the Presence of a Large Excess of Hydrogen Peroxide Fulvio Di Furia Maurizio Prato Ugo Quintily Susanna Salvagno and Gianfranco Scorrano Centro CNR Meccanismi di Reazioni Organiche lstituto di Chimica Organica Via Marzolo 7 35737 Padova ltaly The oxidation of organic sulphides is very fast with peracids and very slow with hydrogen peroxide. Thus, addition of an excess of methyl p-tolyl sulphide to a mixture of an organic peracid and hydrogen peroxide results in the formation of methyl p-tolyl sulphoxide equimolar with the peracid. The gas-chromatographic determination of the residual sulphide or of the produced sulphoxide affords a quantitative evaluation of the peracid present in the mixture.The method has been applied to the determination of m-chloroperbenzoic acid and peracetic and perpropionic acids. Keywords Gas chromatography; peracid determination; methyl p-tolyl sulphide Several methods are available for the quantitative determina-tion of both peroxocarboxylic acids and hydrogen peroxide. Procedures for determining the concentration of peracid in solutions containing also hydrogen peroxide have been reported.2 Amongst these the method based on the oxidation of hydrogen peroxide by Ce(1V) is probably the most frequently used.3 Although this procedure gives excellent results when the concentrations of the two peroxo species are comparable we found that the Ce(1V) method fails to give accurate results when very small amounts of peroxocarboxylic acid need to be determined in solution in the presence of a large excess of hydrogen peroxide.Thus we propose here as an alternative a procedure based on the fact that under the appropriate experimental condi-tions the oxidation of an organic sulphide to the correspond-ing sulphoxide may be very fast with peracids and very slow with hydrogen peroxide.4 Indeed when the latter is employed in synthetic processes the presence of a catalyst either the proton or a transition metal derivative is necessary.4-5 In the proposed method a known excess of sulphide is added to the solution containing both peracid and hydrogen peroxide. The reduction of the peracid is complete within a few seconds whereas hydrogen peroxide is not consumed.The concentration of the peracid is then determined by gas - liquid chromatographic measurement of the remaining sulphide concentration and the amount of sulphoxide formed simul-taneously. The two measurements are in very good agree-ment. The total concentration of peroxidic material and hence by difference the concentration of hydrogen peroxide, may be obtained in parallel experiments by a standard iodimetric titration .6 The method proved suitable for solu-tions where the hydrogen peroxide to peracid ratio was as high as 100 whereas the Ce(1V) procedure lacked reproducibility and led to large experimental errors. As an example we have successfully used the procedure to monitor the slow equilibrium formation of two peroxocar-boxylic acids ,7 peracetic and perpropionic acids respectively, from hydrogen peroxide and the parent carboxylic acid.Experimental Apparatus A Varian 3700 gas chromatograph equipped with a flame-ionisation detector (FID) and a 0.5 m x 2 mm i.d. glass column packed with 3% FFAP on Chromosorb W AW DMCS (80-100 mesh) was used at an oven temperature of 70 "C for 2 min with a temperature gradient of 20 "C min-1 to 180 "C, remaining at 180 "C for 3 min an injector temperature of 190 "C a detector temperature of 240 "C and nitrogen as the carrier gas at a pressure of 12 p.s.i.g. at 70 "C. A Perkin-Elmer Sigma 10 computing integrator was also used. Reagents All reagents were of analytical-reagent grade and tested for interfering impurities. Chloroform . Ethanol. Hydrogen peroxide 35"/b.Obtained from Carlo Erba. Sodium hydrogen carbonate. Sodium chloride. Sodium metabisulphite. Sulphuric acid 96%. m-Chloroperbenzoic acid 85%. Octadecane. Gas-chromatographic grade obtained from Methyl p-tolyl sulphide and methyl p-tolyl sulphoxide. Carlo Erba. Prepared and purified by reported procedures.6.8 Gas Chromatography Under the conditions described under Apparatus the reten-tion time of methyl p-tolyl sulphide is 4.5 min that of the corresponding sulphoxide is 8.1 min and hence the total GLC analysis can be completed in 11 min. The calibration is obtained with 6-pl injections of standard solutions in chloroform of sulphide and sulphoxide (and octadecane as an internal standard) using peak areas as the measured parameters. The results are the means of at least three determinations.Analysis Solution A Commercial m-chloroperbenzoic acid is weighed into a 100-ml calibrated flask diluted to the mark with ethanol and analysed by the 12 - Na2S203 method.6The solution is stored in a refrigerator. Solution B A solution of hydrogen peroxide is obtained by diluting with ethanol the commercial product (35 mlV solution in water) and standardised by iodimetric titration using ammonium molybdate as a catalyst.6 Solution C Aliquots of solutions A and B were mixed in a 50-ml calibrated flask and diluted to the mark with ethanol. Solution C can be analysed both titrimetrically and by gas chromatography. Titration analysis The total peroxide content (H202 + MCPBA) is measured by iodimetric titration using ammonium molybdate as a catalyst.986 ANALYST AUGUST 1984 VOL. 109 Table I . Determination of rn-chloroperbenzoic acid (MCPBA) in the presence of hydrogen peroxide by gas-chromatographic and titrimetric methods Gas-chromatographic determination Titrimetric determination MCPBA MCPBA MCPBA determined/ determinedl determined/ X N x10 * N >; N Hz02 MCPBAii (remaining Standard (sulphoxide Standard H,02 MCPBAI/ [Ce(W) Standard MCPBA* x 10-2 N sulphide) deviationh formed) deviationh MCPBA* X 10k2 N titration] deviationh 102.4 101.9 101.9 50.9 10.2 10.3 1.02 1.03 2.08 5.30 2.10 2.10 2.10 5.25 2.10 5.25 2.09 5.29 2.09 2.11 2.08 5.25 2.09 5.23 0.08 0.01 0.04 0.04 0.06 0.01 0.01 0.02 2.02 5.31 2.14 2.16 2.12 5.12 2.05 5.10 0.04 114.0 1.93 2.13 0.22 0.02 85.2 1.93 1.97 0.18 0.09 53.1 1.93 1.93 0.08 0.06 11.3 1.93 1.88 0.07 0.02 1.1 1.93 1.93 0.01 0.06 0.03 0.01 * Prepared by mixing known volumes of titrated solutions of the two species.t Calculated from dilution of solution A (see text). The peracid and the hydrogen peroxide are titrated stepwise by the Ce(IV) method.3 Gas-chromatographic Analysis Methyl p-tolyl sulphide (0.4608 g) and the internal standard octadecane (0.0892 g) are weighed into a 100-ml flask and diluted to the mark with ethanol. A 10-ml volume of solution C is mixed with 10 ml of the sulphide solution and left at room temperature with stirring for 5-10 min. The reaction is then quenched by addition in order of crushed ice (10 g) sodium hydrogen carbonate (20 ml of a saturated aqueous solution), sodium metabisulphite (in excess) sodium chloride (10 g) and chloroform (60 ml).The resulting mixture is well shaken and the organic layer separated and analysed by GLC for the residual sulphide and the sulphoxide formed. Both analyses give comparable results. We suggest however that the determination of the sulphide and of the sulphoxide is carried out as a check of internal consistency. In fact some sulphoxide might be lost in the aqueous phase if the above procedure is not followed carefully. Peracetic and Perpropionic Acid Determination The carboxylic acid (10 g) is weighed into a 50-ml calibrated flask 0.1 ml of 96% sulphuric acid is added diluted to the mark with 35% aqueous hydrogen peroxide and left in a thermostated bath at 25 "C.Aliquots are withdrawn at time intervals and diluted 100-fold with water (to reach a peracid concentration suitable for analysis ca. 10-2 N) and substituted for solution C in the method described above. The H202 to peracids ratio was in the range 50 1-10 1. Results and Discussion Known volumes of stock solutions of m-chloroperbenzoic acid (MCPBA) and H202 were mixed in order to obtain solutions of the two peroxidic materials of different proportions. These samples were analysed for the peracid content by both methods i.e. the Ce(IV) and the gas-chromatographic procedures. The results together with the standard devia-tions are given in Table 1. A typical chromatogram is shown in Fig.1. This comparative study indicates a reasonable agreement between the results obtained by the two methods for solutions where the H202 to MCPBA ratio is less than 50 whereas at higher ratios the Ce(IV) method leads to an overestimation of the peracid content. In contrast the gas-chromatographic procedure gives reliable results over the entire range explored. Table 2. Peracid equilibrium formation from carboxylic acids and hydrogen peroxide at 25 "C in aqueous medium as monitored by the GLC method Peracid YO Sulphide Timeimin analysis Peracetic acid-217 6.8 446 14.1 707 18.5 1615 24.2 1927 25.7 2995 29.2 Perpropionic acid-214 7.3 435 16.5 699 19.7 1596 26.4 1924 27.6 2991 30.9 ~ Standard deviation 0.5 0.1 0.1 0.1 0.1 0.3 0.1 0.2 0.1 0.2 0.4 0.4 Sulphoxide analysis 6.1 11.8 18.0 23.3 25.6 28.8 6.6 13.7 17.1 25.7 28.1 29.3 Standard deviation 0.1 0.2 0.1 0.2 0.1 0.1 0.1 0.3 0.3 0.3 0.2 0.4 0 2 4 6 8 Ti me/mi n 3 Fig.1. Typical chromatogram containing A methyl p-tolyl sul-phide; B octadecane; and C methyl p-tolyl sulphoxid ANALYST AUGUST 1984. VOL. 109 These results are confirmed by the data shown in Table 2, where the kinetic of formation of peracetic and perpropionic acid already studied by Swern,’ is followed by using the gas-chromatographic procedure. It may be noticed that even at the lowest peracid concentration corresponding to the initial part of the reaction very accurate determinations are possible.This might allow inter alia the evaluation of the initial rates. Conclusion This work was aimed at finding an alternative procedure for the determination of peroxocarboxylic acids when the method based on Ce(IV) owing to the contemporary presence of a large excess of hydrogen peroxide leads to unreliable results and generally to an overestimation of the reagent. On the basis of the results reported the gas-chromatographic pro-cedure proves to be both convenient and accurate. It should be noticed that the general feature behind this procedure e.g., the different reactivity of the various oxidants towards organic sulphides might be used in developing other analytical methods for the selective determination of a given peroxidic species in the presence of peroxides of different nature e.g., dialkyl peroxides versus peresters.1 987 versus alkyl hydroperoxides or peracids 1. 2. 3. 4. 5. 6. 7. 8. References Mair R. D. and Hall R. T. in Swern D. Editor “Organic Peroxides,” Volume Two Wiley-Interscience New York, 1971 Chapter 6. House H. O. “Modern Synthetic Reactions,” Second Edition, Benjamin Cummings Publishing Co. Menlo Park CA 1972. Greenspan F. P. and Mackellar D. G. Anal. Chem. 1948, 20 1061. Curci R. Di Prete R. Edwards J. O. and Modena G. in Covington A. K. and Jones P. Editors “Hydrogen-bonded Solvent Systems,” Taylor and Francis London 1968 p. 303. Di Furia F. and Modena G. in Tsutsui M. Editor, “Fundamental Research in Homogeneous Catalysis,” Volume 2 Plenum Press New York 1978 p. 255. Modena G. and Maioli L. Gazz. Chim. Ital. 1957,87 1306. Swern D. Chem. Rev. 1949 45 1. Leandri G. Mangini A. and Passerini R. Gazz. Chim. Ital., 1954,84,3. Paper A3142 7 Received December 5th 1983 Accepted February 20th 198

ISSN:0003-2654    

DOI:10.1039/AN9840900985

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST AUGUST 1984 VOL. 109 989 Gas-chromatographic Method for the Determination of Low Relative Molecular Mass Alcohols and Methyl tea-Butyl Ether in Gasoline Leonard A. Luke and John E. Ray The British Petroleum Company plc BP Research Centre Sunbury-on-Thames Middlesex TW16 7LN, UK Increasing use is being made of oxygenated compounds such as methanol 2-methyl propano and methyl tert-butyl ether (MTBE) as octane improvers in gasoline. As a consequence reliable methods are required for the determination of these oxygenated compounds when they are present in commercial fuels. A rapid gas-chromatographic method has been developed for the simultaneous determination of MTBE and C1-C4 alcohols in gasoline that shows excellent precision over a wide concentration range (1-20% VlV) of any of the component oxygenates.The sample with an internal standard added is injected on to a column containing ethylene glycol succinate on Chromosorb P and the lower boiling hydrocarbons (up to heptane) are allowed to elute and are vented. The flow of the carrier gas is then switched and the higher boiling hydrocarbons as well as the oxygenated compounds which are retained on this polar phase are back-flushed on to a second column containing Porapak P. The oxygenates which are less volatile than the remaining hydrocarbons are separated individually from the latter and then quantified. The flow is again switched in order to back-flush the heavy hydrocarbons as soon as possible. Both columns can be operated at 150 "C and can be fitted into virtually any standard single oven gas chromatograph.The analysis takes 25 min to complete. Keywords Gas chromatography; gasoline; methanol; methyl tert-butyl ether; oxygen compounds Many oxygen compounds enhance the octane rating of gasolines and are already being used as fuel blending components in some parts of the world. As legislation comes into force reducing lead levels in gasoline the incorporation of alcohols ethers and other oxygen compounds in this fuel will undoubtedly increase. In Europe the most frequently used oxygenates are likely to be methanol 2-methylpropan-l-01 2-methylpropan-2-01 and methyl tert-butyl ether (MTBE). Ethanol is likely to become a significant fuel component in the USA and parts of the world where this alcohol can be produced cheaply from agricultural products.A number of workers have published methods for the determination of one or more oxygenates when present in gasoline; Fry et a1.2 used infrared spectrometry to determine MTBE in gasoline and Pauls and McCoy3 used a water extraction procedure to isolate the alcohols from the gasoline followed by GC analysis of the aqueous phase. The latter method worked well in our laboratories provided the same gasoline sample that was used in the blend was available for calibration purposes. However we found it less reliable for completely unknown samples. Further a secondary tech-nique employing conventional liquid chromatography is required if MTBE is to be analysed. Lockwood and Caddock4 have recently published a GC method using a fused-silica capillary column coated with dimethylsilicone gum.This is aimed at providing quantitative data for both the oxygenates and the gasoline hydrocarbons in one analysis. In this method separations are illustrated employing dual ramp temperature programming with initial sub-ambient operation. This method requires a relatively sophisticated GC instrument. Further as a capillary column giving a specified performance must be purchased it is doubtful whether this method will gain widespread acceptance as a standard procedure for oxygenates determination unless simultaneous determinations of the hydrocarbons are required. A number of multiple-column methods have been described utilising column switching valves and/or back-flush systems for firstly separating the oxygenates from potentially interfering hydrocarbons and then resolving the individual oxygenates for identification and measurement.Thus the Atlantic Richfield Companyj.6 has described methods for determining either MTBE 2-methylpropan-2-01 and ethanol or 2-methylpropan-2-01 and methanol in gasoline. A multi-dimensional switching system was used by SevCik7 to determine methanol, 2-methylpropan- 1-01 and propan-2-01. Universal Oil Pro-ducts8 has described a method capable of analysing a wider range of oxygenated compounds. In this method the sample was injected on to a polar column [25% mlm tetracyanoethyl-pentaerythritol (TCEPE) on Chromosorb PI. The oxygenates and the higher boiling hydrocarbons were retained whilst the lower boiling hydrocarbons were eluted.At a specified time the column flow was reversed and the retained components were flushed on to a less polar column (Porapak P) for analysis. The main problem with this system is that the TCEPE column is run at 115 "C and the Porapak P at 170 "C. Because the TCEPE column is unstable at the higher temperature the chromatograph must be- fitted with an auxiliary oven. We carried out an investigation to find a dual-column system where the phases were temperature compatible and would give the required separation under isothermal condi-tions. This would clearly have the advantage that the column system could be installed in virtually any standard GC instrument with the minimum of modification and difficulty. This objective has been achieved and the method has been applied successfully to the analysis of a number of gasolines containing oxygenates in a wide concentration range.Experimental Apparatus and Materials Gas chromatograph. Equipped with a single flame-ionisa-tion detector and with provision for installing an eight-port rotary valve in the column oven. Computing integrator and 1-mV recorder. Column 1. Made with stainless steel 4.6 m length 3.2 mrn 0.d. and packed with 30% mlm ethylene glycol succinate on Chromosorb P (85-100 mesh). (The buffer column is identical with Column 1.) Column 2. Made of stainless steel 2.7 m length 3.2 mm 0.d. and packed with Porapak P (80-100 mesh). Syringe 10 pl. Fitted with a Chaney adaptor. Carrier gas helium. Standards alcohols and ethers of purity greater than 99%.Used for the calibration of the system 990 ANALYST AUGUST 1984 VOL. 109 Operation of the Instrument The valve and columns are installed as shown schematically in Fig. 1. The buffer column is used to reduce base line disruption when the valve position is changed. Insertion of a 1-m length of 0.25 mm i.d. stainless-steel tubing between the valve and the FID provides a flow restriction that prevents flame extinction upon operation of the valve. The operating conditions are as follows carrier gas flow-rate 30 ml min-1; detector and injector port temperatures 200 "C; oven temperature 150 "C isothermal; and sample size 1 pl. Calibration Prepare a blend containing 5.0% V/V each of MTBE and 2-methylpropan-2-01 in pentane. Initially connect column 1 to the detector and inject 1 pl of the sample.Measure the time ( t l ) at which the MTBE just starts to elute. Next set up columns 1 and 2 as shown in Fig. 1. With the valve in position 1 inject 1 pl of the blend and then switch to column 2 after (tl -10) s. This will ensure that all the 2-methylpropan-2-01 and MTBE pass to column 2. Determine the ratio of the peak areas of the two oxygen compounds. In successive experi-ments increase the time at which the valve is switched. Comparison of the peak areas for the two compounds will establish the maximum time ( t 2 ) at which the ratio remains constant and beyond which the MTBE concentration starts to fall (the 2-methylpropan-2-01 is more strongly retained on column 1). This is the optimum switching time for the individual system.In our apparatus t2 is 98 s and for samples containing MTBE but no 2-methylpropan-1-01 this time is adjusted to 95 s. If no MTBE is present the switch time can safely be increased to 100 s. Prepare a blend containing all of the oxygenates of interest, including the desired internal standard each at 5.0% V/V in heptane. Carry out an analysis of this blend and measure the areas of the peaks. Calculate the relative response correction factors (RRCF) which are required to compensate for the difference in detector response of the various compounds. Analysis Add 5% V/V of internal standard (usually propan-2-01 if it is not already present) to the unknown sample and mix thormghly. Inject an aliquot into the GC with the valve in position 1. At time f 2 turn the valve to position 2.After the last oxygenated compound of interest has eluted (15 min if 2-methylpropan-1-01 is present) turn the valve to position 1 to back-flush the higher boiling hydrocarbons. As soon as a steady base line is regained a second sample can be injected. With a computing integrator the RRCF factors can be applied automatically to the peak areas. Gasolines containing any or all of the following oxygenates can be analysed methanol, ethanol propan-2-01 2-methylpropan-1-01 2-methylpropan-2-01 and methyl tert-butyl ether. A typical chromatogram is shown in Fig. 2. Results A large number of runs were carried out using known blends of oxygenates in gasoline in order to check the linearity of response of each oxygenate component across a wide concen-tration range.The data are presented in Fig. 3. Duplicate determinations were carried out on indepen-dently blended samples whose compositions were unknown to the operator at the time of the analysis in order to assess the typical levels of accuracy that could be expected. The data are presented in Table 1. Six consecutive analyses were made of a blend containing a number of the oxygenates to assess the method repeatability. The data are given in Table 2. Buffer column Column 2 Column 1 VALVE IN POSITION 1 I k v w w v w 4 VALVE IN POSITION 2 Fig. 1. Schematic diagram of the system ~~ __ 25 20 15 10 5 0 Time/m i n Fig. 2. Chromatogram for the separation of individual oxygenates. Peaks A gasoline heavy ends; B 2-methylpropan-1-01; C MTBE; D 2-methylpropan-2-01; E propan-2-01; F ethanol; and G methanol I5O0 I m 1000 0 X m 7 2 Y m 2 500 0 5 10 15 2c O/O VIV Fig.3. Graph showing the linearity of the oxygenates FID response. A 2-Methylpropan-2-01; B 2-methylpropan-1-01; C MTBE; D, ethanol; and E methano ANALYST AUGUST 1984 VOL. 109 99 1 Discussion A key aspect of the development of the method was the search for a liquid phase that would have a selectivity of separation for the lower relative molecular mass hydrocarbons and oxygenates similar to that of TCEPE but which had better temperature stability. Examination of the literature (McRey-nold's constants Table 3) indicated a number of possible alternatives. Experimental work showed that ethylene glycol succinate was the most suitable material when used at the concentration and column length as indicated under Experimental.Varying the porous polymers used in the second column revealed some interesting data. It was found that MTBE had different relative retentions on two Porapaks QS and R, compared with the lower alcohols. This is demonstrated in Fig. 4. These data can be used to tailor single columns containing mixed lengths of both adsorbents for optimising the Table 1. Results of duplicate determinations of blends of oxygenates in gasoline. The actual blended value is given in parentheses. Propan-2-01 is blended as an internal standard at 5.0°/0 V/V Content of samples 1-6,"/0 ViV -Component 1 Methanol. . . . 14.8 15.0 (15 .O) Ethanol . . .. -2-Methylpropan-- 2-01 . . . . 2-Methylpropan-1-01 . . . . 2.1 2.1 (2.0) Methyl tert-butyl ether . . . . -2 3.0 2.9 (3.0) -3.0 2.9 (3.0) --3 1.8 2.0 (2.0) -3.0 3.0 (3.0) -10.4 10.3 (10.0) 4 9.8 9.7 3.3 3.2 (10.0) (3.0) 3.0 3.0 (3.0) --6 12.0 11.9 2.2 2.2 (12.0) (2.0) 1 .0 1 .o (1.0) 2.1 2.1 (2.0) 4.1 4.1 (4.0) separation of an ether such as MTBE from gasoline hydrocar-bons. This might be of value if a method was required solely for the determination of MTBE in gasoline. However for our multi-oxygenate determination there were no practical advan-tages in using a mixed QS and R Porapak over that of Porapak P and the latter was adopted because it gave a shorter analysis time.The chromatogram shown in Fig. 2 using the method finally developed indicates that generally good resolution is obtained between the oxygenates of interest. Base line separation is not quite achieved between methanol and ethanol owing to some tailing of the methanol peak. The significance of this will tend to be enhanced if the former component is in much higher concentration than the latter. This is also illustrated in Table 1 where the determined levels of ethanol are about 10% absolute higher than the blended values at a 2-370 V/Vconcentration of ethanol. In practice it is unlikely that many gasolines will be marketed with both I I 1 0 20 40 60 80 100 Proportion of Porapak R O/O Fig. 4. Lo arithmic plot of the relative retention (RRT) data for Porapaks Q! and R.A Benzene; B 2-methylpropan-1-01; C MTBE; D 2-methylpropan-2-01; E propan-2-01; F ethanol; and G methanol Table 2. Repeatability of analysis. Propan-2-01 is blended as an internal standard at 5.0% ViV Amount Analysis number blended, Component O/O VIV 1 2 3 4 5 6 Mean Methanol . . . . . . 10.0 10.14 10.05 10.04 9.86 10.05 9.85 9.99 2-Methylpropan-2-01 . . 5.0 5.00 4.99 4.99 5.06 4.97 5.05 5.01 Methyl tert-butyl ether . . 5.0 4.97 4.95 4.97 5.07 4.95 5.09 5.00 2-Methylpropan-1-01 . . 5.0 4.97 4.93 5.03 5.02 4.95 5.11 5.00 Relative standard deviation, O/O 1.2 0.7 1.3 1.3 Table 3. Characteristics of selected gas chromatographic phases McReynold's constants Benzene Butanol c; Diethyleneglycolsuccinate . . . . 499 75 1 3543 Ethyleneglycolsuccinate .. . . 537 787 3759 Tetracyanoethylpentaerythritol . . 526 782 3742 Carbowax 20M . . . . . . . . 322 536 2308 Silar 10C . . . . . . . . . 523 757 3682 OV275 . . . . . . . . . 781 1001 4938 Upper temperature 1imiti"C 150 255 200 275 200 25 992 ANALYST AUGUST 1984 VOL. 109 methanol and ethanol present. The data in Table 1 indicate that in most instances the determined values of the oxygenates are within +5% absolute of the blended value. Linearity of response for all oxygenates considered over a 1-20% VIV concentration range is excellent (Fig. 3) and repeatability is seen to be entirely satisfactory (Table 2). As the technique uses standard GC equipment requiring only isothermal operation in a single oven it should be suitable for use in service laboratories refineries blending plants and depots as well as in the research situation.A thermal conductivity detector could be used instead of an FID if safety considerations required this without any need for changes in the technique. Response factors would of course be different to those obtained on the FID. Because of its simplicity and wide applicability the technique has the potential to become a standard reference method for use in the petroleum industry. Conclusions A rapid dual packed column technique has been developed for the determination of oxygenates in gasolines with good precision. It is simple to operate runs isothermally and requires only a standard single-oven chromatograph. References 1. 2. 3. 4. 5 . 6. Parkinson G. S. Pet. R e v . February 1984 39. Fry S. E. Fuller M. P . . White F. T. and Battiste D. R., Anal. Chem. 1983 55 407. Pauls R. E. andMcCoy R. W. J. Chromafogr. Sci. 1981 19. 558. Lockwood A. F. and Caddock B . D. Chromatographia. 1983 17 65. ARCO Analytical Method 9149 Atlantic Richfield Company, Harvey IL. ARCO Analytical Method 9157 “Oxygenated Fuels Technical Bulletin,” Atlantic Richfield Company Harvey IL Novem-ber 1982. SevEik J. J. High Resolut. Chromatogr. Chromatogr. Com-mun. 1980 3 166. Universal Oil Products “Proposed Standard Test Method for Oxygenates in Gasohols by Gas Chromatography,” for ASTM Committee D-2 R-D Division IV-L Correlation Programme, June 1982 Toronto. 7. 8. Pup e r A 41 9 7 Received March 13th 1984 Accepted April 17th 198

ISSN:0003-2654    

DOI:10.1039/AN9840900989

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST AUGUST 1984 VOL. 109 993 Separation of Some Transition Metals by Reversed-phase Paper Chromatography C. G. Yeole and V. M. Shinde" Analytical Laboratory Department of Chemistry Shivaji University Kolhapur 4 76 004 India A reversed-phase paper chromatographic method has been developed for the mutual separation of transition metals such as Fe(lll) Cu(ll) Pb(ll) Ni(ll) Ag(l) and Au(ll1). The separations were performed on Whatman No. 1 filter-paper with liquid ion exchangers such as trioctylamine triisooctylamine or Aliquat 336 as stationary phases and sodium acetate sodium malonate or sodium succinate as mobile phases. Keywords Transition metal separation; reversed-phase paper chromatography Selective separation of transition metals by reversed-phase paper chromatography is a recent problem but few methods have been investigated.Papers impregnated with N N-bis-[ (2-methylheptyl)acetamide] 1 bis[ (2-ethylhexyl) hydrogen phosphate] ,* N-phenylbenzohydroxamic acid ,3 sulphoxide sol-vents for petroleum,4 trioctylamines and 4-benzoyl-3-methyl-1-phenylpyrazolin-5-one6 have been used for chromato-graphic studies of iron cobalt nickel silver and gold but a systematic separation study is lacking. In addition these methods are time consuming,h require a high concentration of the mobile phase1.5 and for petroleum sulphoxides strict temperature control to prevent their oxidation.4 In this work we have attempted to exploit the anion-exchange properties of liquid ion exchangers such as trioctylamine (TOA) triisoctyl-amine (TIOA) and Aliquat 336 which were used as stationary phases on Whatman No.1 filter-paper as des-cribed earlier? The metal ions were developed with organic complexing agents such as sodium acetate sodium malonate and sodium succinate as the mobile phases. The optimum conditions for the separation of the transition metals were evaluated from a critical study of the effects of pH concentra-tion of mobile phase and stationary phase on the R values of the metals and the method was applied to the analysis of synthetic mixtures and samples. Experimental Preparation of Stationary Phase A 2 or 5% benzene solution (Caution-Benzene is highly toxic and appropriate precautions should be taken) of a liquid ion exchanger such as TOA TIOA or Aliquat 336 (obtained from Koch-Light or Fluka) was pre-equilibrated (for 1 h) with equal volumes of 2 M sodium acetate sodium malonate or sodium succinate solutions (to convert the amine into the acetate, malonate or succinate) and sprayed on to Whatman No.1 filter-paper (15 x 4 cm). The paper was dried with warm air and used for ascending reversed-phase chromatography, performed in glass jars. The pH of the mobile phase was adjusted after measurement with a Philips precision pH meter. The stock solutions of copper(II) nickel(II) lead(II), iron(III) silver( I) and gold(II1) were prepared by dissolving copper(I1) sulphate (4.91 g) nickel chloride (5.06 g) lead nitrate (1.99 g) ammonium iron(II1) sulphate (10.73 g), silver nitrate (1.96 g) and chloroauric acid (1.01 g) in 250 ml of water.The solutions were standardised by established methods and diluted as required. Working Procedure Take Whatman No. 1 filter-paper strips (15 X 4 cm) impreg-nated with liquid ion exchangers spot the metal ion or mixture * Present address Department of Chemistry University of Bombay. Kalina Santacruz (E). Bombay 400 098 India. of metal ions with the help of microcapillaries dry and keep in a glass jar (21.5 x 5.2 cm) containing the aqueous solution of either sodium acetate sodium malonate or sodium succinate adjusted to the required pH. Allow the solvent to run using the ascending technique for 10 cm and after drying the paper detect the metal ions by spraying with colour-forming reagents. Calculate the R value for the individual cation by the usual procedure.Detection Nickel8 and copper9 are detected as blue or olive green spots with an alcoholic solution of 0.1% rubeanic acid; silver is detected as a pink spot with 0.1% dithizonelo solution in carbon tetrachloride exposed to ammonia; gold is detected with a tin(I1) chloride - potassium iodide mixture," which gives a brown spot; iron is detected with thiocyanatelz; and lead is detected with an aqueous solution of 0.1% 4-(2-pyridylazo)resorcinol,13 giving a red - brown spot. Results and Discussion Various experiments were carried out to study the effect of the pH of the mobile phase concentration of the mobile phase and concentration of the stationary phase on the R values of individual cations (Table 1). The R values were found to be almost constant between pH 4 and 7 in an organic complexing solution such as sodium acetate sodium malonate and sodium succinate (it is difficult to adjust the pH to below 4) whereas at a pH higher than 7 irreproducible results were observed.The variation in concentration of mobile phase (0.01-0.1 M) showed that for good results either a 0.05 or a 0.01 M aqueous solution of mobile phase could be employed. For concentrations between 0.05 and 0.1 M the R values show an insignificant change. Similarly variation in the concentration of liquid ion exchangers (1-5% using benzene as diluent) show insignificant changes on R values between 2 and 5%. A 1% solution of amine gave erratic results and so for detailed studies a 2-5%solution of amine (as the stationary phase on papers) and a 0.05-0.1 M solution of the mobile phase (sodium acetate sodium malonate or sodium succinate) adjusted to either pH 4 or 7 were used.Temperature had no effect on the R values and so all the experiments were carried out at room temperature. The method is suitable for the separation of ternary and quaternary mixtures (Fig. I). The time required for a 10-cm development ranges from 25 to 50 min. The standard deviation and coefficient of variation were found to be 0.009-0.014 and 1.05-1.5% respectively. The proposed method is applicable to the separation and detection of iron(III) copper(II) nickel lead and silver in pharmaceutical samples brass ball bearings and nickel - silver alloy. The separation conditions are reported in Table 2 994 ANALYST AUGUST 1984 VOL.109 ~ Table 1. R values of Ni(II) Cu(II) Fe(II1) and Pb(I1) on paper impregnated with high relative molecular mass amines as the stationary phase using acetate malonate and succinate solutions (pH 4) as the mobile phase. Ag(1) and Au(II1) do not move and remain at the bottom Mobile phase Stationary phase Sodium acetate Sodium malonate Sodium succinate Metal ion Aniine tration O/O 0.05 M 0.01 M 0.05 M 0.1 M 0.05 M 0.01 M Ni(I1) . . . . TOA 2 0.89 0.90 0.96 0.96 0.97 0.95 5 0.92 0.90 0.98 0.96 0.95 0.97 TIOA 2 0.92 0.94 0.97 0.9s 0.98 0.96 5 0.94 0.94 0.92 0.90 0.95 0.94 Aliquat 336 2 0.9.5 0.94 0.82 0.80 0.93 0.95 5 0.92 0.94 0.73 0.76 0.94 0.92 Cu(I1) . . . . TOA 2 0.84 0.85 0.87 0.85 0.91 0.90 5 0.84 0.85 0.90 0.93 0.90 0.92 TIOA 2 0.88 0.90 0.85 0.83 0.86 0.85 5 0.92 0.92 0.85 0.90 0.85 0.85 Aliquat 336 2 0.95 0.96 0.74 0.72 0.85 0.87 5 0.94 0.93 0.46 0.60 0.72 0.75 5 0.88 0.95 0.99 0.99 0.93 0.9.5 TIOA 2 0.93 0.97 0.99 0.99 0.89 0.90 5 0.90 0.90 0.99 0.99 0.89 0.90 Aliquat 336 2 0.85 0.86 0.01 0.01 0.95 0.90 5 0.90 0.85 0.01 0.01 0.90 0.95 Pb(I1) .. . . TOA 2 0.82 0.84 0.86 0.88 0.83 0.85 5 0.80 0.82 0.88 0.90 0.85 0.88 TIOA 2 0.90 0.91 0.70 0.72 0.8.5 0.87 5 0.87 0.88 0.72 0.74 0.82 0.85 Aliquat 336 2 0.83 0.80 0.40 0.60 0.28 0.30 5 0.85 0.83 0.27 0.36 0.25 0.28 Fe(II1) . . . TOA 2 0.93 0.98 0.99 0.99 0.94 0.99 Table 2. Optimum conditions for the separation of mixtures of transition metal ions present in drugs and alloys Ion detected by the proposed Sample Composition method Separation conditions Fersolate (Glaxo) .. . . FeS04 (IP* 0.195 g) Fe 2% TIOA 0.1 M sodium malonate (pH 4); CuS04 (IP 2.6 mg), MnSO (BPt 2.6 mg) c u R values Fe(0.99) Cu(0.83); time of development 25 min. Or 5% TIOA 0.05 M sodium malonate (pH 4.0); R,values Fe (0.99) Cu(0.85); time of development 2.5 min Supradin . . . . . . . . FeS04 (IP 32.0 mg), MnS04 (BP 2.0 mg), CuSO,(IP 3.4 mg) Ca,(PO4)2(0. 13 g) 9 Copper - nickel alloy (Indian Govt. Mint) . . Cu(74.6%) Ni(25.2%), Mn(O.l%) Fe(0.03%) Nickel - silver alloy (Kamini Industries) . . Cu(S4.6%) Ni(17%), Pb (0.1 % ) Sn (0.0.5 YO), Mn( 0.2 1 Yo) Fe c u 2% TIOA 0.1 M sodium malonate (pH 4); R,values Fe(0.99) Cu(0.82); time of development 25 min. Or 5% TIOA 0.05 M sodium malonate (pH 4.0); R,values Fe (0.99) Cu(0.84); time of development 25 rnin c u Ni R,values Fe(0.99) Ni(0.92) Cu(0.85); Fe 5Y0 TIOA 0.05 M sodium malonate (pH 4); time of development 25 min Ni c u 5% Aliquat 336,0.05 M sodium succinate (pH 4); R,values Ni(0.94) Cu(0.72); time of development 50 min Brass .. . . . . . . Pb(2.35%). Zn(40.6%) Ni 5% Aliquat 336 0.05 M sodium succinate (pH 4); Cu(56.9%) Fe(0.009%) c u R values Ni(0.93) Cu(0.72) Pb(0.24); Pb time of development 50 min Steelball-bearing . . . . Fe Ni R,values Fe(0.99) Ni(0.92) Cu(0.85); c u 5% TIOA 0.05 M sodium malonate (pH 4); time of development 25 rnin Indian 10 Ps coin . . . . Cu(60%) Ni(40%) c u Ni Synthetic mixture of silver and copper metal . . Ag(20%) Cu(80%) Ag c u Synthetic mixture of silver, nickel and copper metal Ag(20%) Ni(30%) Ag CU(50%) Ni c u * IP Indian Pharmacopoeia.t BP British Pharmacopoeia. 5% Aliquat 336,0.05 M sodium succinate (pH 4); &values Ni(0.94) Cu(0.72); time of development 50 min 5% Aliquat 336,0.05 M sodium succinate (pH4); R values Cu(0.72) Ag(O.O1); time of development 50 min 5% Aliquat 336,O.O.S M sodium succinate (pH 4); R,values Ni(0.94) Cu(0.72), Ag(0.01); time of development 50 rni ANALYST AUGUST 1984 VOL. 109 995 Fig. 1. Reversed-phase paper chromatograms for the se aration of transition metal mixtures. (1) Mobile phase 0.05 M sodium malonate (pH 4); stationary phase Aliquat 5%; time of development 5 f m i n ; R values Ni (0.73) Cu (0.46) and Ag (0.01). (2) Mobile phase 0.05 M sodium malonate (pH 4); stationary phase Aliquat 5%; time of development 50 min; R values Ni (0.78) Cu (0.48) and Fe (0.01).(3) Mobile phase, 0.05 M sodium malonate (pH 4); stationary phase Aliquat 5%; time of development 50 min; R values Ni (0.73) Cu (0.46) Pb (0.27) and Ag (0.01). (4) Mobile phase 0.05 M sodium malonate (pH 4); stationary phase Aliquat 5 % ; time of development 50 min; R values Ni (0.77) Cu (0.48) Pb (0.24) and Fe (0.01). (5) Mobile phase 0.05 M sodium succinate (pH 4); stationary phase Aliquat 5%; time of development, 50 min; R values Ni (0.92) Cu (0.71) and Au (0.01). (6) Mobile phase 0.05 M sodium succinate (pH 4); stationary phase Aliquat 5 % ; time of development 50 min; R values Ni (0.92) Cu (0.71) Pb (0.24) and Au (0.01). 7) Mobile hase 0.05 M sodium malonate (pH 4); stationary hase TIOA 2%; time of development 25 min; R values Fe (0.99) Pb (0.76 and Au fO.01).(8) Mobile phase 0.05 M sodium malonate 6 H 4); stationary phase TIOA 2%; time of development 25 min; R values Fe (0.99) Pb (0.70) and Ag (0.01) 1. 2. 3. 4. 5 . 6. 7. 8. References Lu Z . Fen Hsi Hua Hsueh 1981 9 152; Anal. Abstr. 1981, 41 4B22. Cerrai E. and Ghersini G . J. Chromatogr. 1965 18 124. Fritz J . S . and Sherma J. J. Chromatogr. 1966 25 153. Murinov Yu. I. Kartavtseva A . G. and Nikitin Yu. E. Zh. Anal. Khim. 1977 32 904. Przeszlakowski S . and Soczewinski E. Chem. Anal. (Warsaw) 1966 11 895. De A . K. and Sarkar S. K. Sep. Sci. 1974 9 431. Yeole C. G. and Shinde V. M. Analyst 1983 108 1102. Bhatnagar R. P. and Bhattacharya K. M. J. Indian Chem. SOC. 1978 55 105. 9. 10. 11. 12. 13. Lemoine A. Anal. Chim. Acta 1952 6 528. Krasiejko M. Marczenko Z. and Kowalski T . Chem. Anal. (Warsaw) 1979 24 1037. Wagif Husain S. and Rasheedzad Sh. Mikrochim. Acta, 1978 1 11. Gruzins I. Leiniece I. Alksnis A Silis U. and Surna J., Latv. PSR Zinat. Akad. Vestis Kim. Ser. 1978 6 705. Malakhova N. M. Olenovich N. L . and Kotel’naya N. I., Zh. Anal. Khim. 1980 35 475. Paper A4132 Received January 23rd 1984 Accepted February 29th 198

ISSN:0003-2654    

DOI:10.1039/AN9840900993

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST AUGUST 1984 VOL. 109 997 Simple and Sensitive Ion Chromatograph for Trace Metal Determination G. J. Schmidt and R. P. W. Scott The Perkin-Elmer Corporation Main Ave. Norwalk CT 06856 USA A simple non-suppressed ion-chromatographic system is described for the determination of cations The design of a novel detector and post-column reactor exhibiting high sensitivity and low dispersion are described in detail together with an on-column cation concentration procedure. The system has a linearity of over three orders of magnitude and is satisfactory for determining cations present in aqueous media down to a level of a few parts per billion in the presence of high concentrations of salt and organic materials. Keywords /on chromatography; trace metal determination The use of liquid chromatography for the direct determination of metals in a variety of sample matrices offers several attractive possibilities.Firstly multi-element capability is achievable owing to the inherent nature of the chromato-graphic process. This capability is not only useful for determining a particular metal in any given sample without changing the conditions of analysis but is also advantageous for the determination of a number of different constituent metals simultaneously. Secondly by suitable choice of an appropriate phase system and detector sample preparation and concentration could be achieved simultaneously and consequently high concentration sensitivity achieved. The separation and determination of metal ions by liquid chromatography has to date been accomplished with varying degrees of success.Several approaches to the determination of metals by chromatographic means have been described and these may be appropriately divided into two distinct forms: suppressed and non-suppressed ion chromatography. Suppressed ion chromatography as first described by Small et al. ,* has been widely utilised for the determination of anions but its application to metals has been limited. Moreover there are several disadvantages to the use of a suppressor column including increased band dispersion, which causes a decrease in separation efficiency and the need to regenerate the suppressor column when its exchange capacity is exhausted. In addition some metal ions may be precipitated by the hydroxide form of the suppressor resin, resulting in the formation of transition metal hydroxides.The alternative to this form of ion chromatography is non-suppressed ion chromatography. Such systems can be used with a variety of column packing including low-capacity ion exchangers and reversed-phase materials. A serious drawback to non-suppressed ion chromatography when used in conjunction with conductivity detectors is the need to employ very weakly conducting mobile phases thereby limiting the choice of the phase system. An alternative to these two chromatographic approaches has been discussed by Elchuk and Cassidy.2.i Their chromato-graphic system utilises a post-column reaction of the metal ions with an appropriate reagent to form complexes that exhibit high molar absorptivity at a specific wavelength in the visible region.This system appears promising in that the problems associated with the use of suppressor columns are removed and as the conductivity of the mobile phase is now of no consequence the choice of a wide range of phase systems becomes available. In addition the use of post-column reaction ensures greatly increased specificity relative to that obtainable with conductivity detection. In this paper a complete chromatographic system deve-loped specifically for the determination of metals using the approach of Elchuk and Cassidyz.3 is discussed. A very simple detector and post-column reaction system (designed and constructed specifically for this analysis) is described and shown to exhibit very high sensitivity while maintaining very low band dispersion characteristics.A phase system that allows the direct injection of large volumes of sample with accompanying cation concentration and that permits the determination of metals in the low parts per billion concentra-tion range is also described. Equipment Apparatus A Perkin-Elmer Series 4 solvent delivery system was used for the mobile phase supply and a Perkin-Elmer Series 10 liquid chromatography (LC) pump was used for reagent delivery. Owing to the use of isocratic development with the phase system employed a very simple pumping system can be utilised for mobile phase delivery if desired. Most of the experimental data were obtained with the use of the Series 4 pump although during system development an Eldex single-piston pump (Model A-30-S) was also successfully employed.The latter pump proved satisfactory if used with a suitable pulse damping system such as a diaphragm-type pulse damper. The particular damping device used in this work had a mobile phase volume of 600 pl at the maximum operating pressure of 6200 lb in-2 and was followed by a suitably large ballast volume situated between it and the sample valve. In the work discussed here an empty preparative LC column having dimensions of 25 x 2.2 cm i.d. was used to provide a suitable ballast volume of 97 m1. This arrangement smoothed the pressure pulses effectively and provided acceptable perfor-mance. As gradient elution development was not required, the relatively large ballast volume did not prove to be a disadvantage.Injections were made using a Rheodyne 7010 injection valve. This valve had an externally located position for installing sample loops of various volumes and in this work several different loop sizes were used. To assess the general performance of the chromatography system a 6-pl loop was used. However at the other extreme a 3 m X 1 mm i.d. loop having a volume of 2.3 ml was employed for sample preparation and concentration. The column used (25 x 0.46 cm i.d.) was packed with 10-pm CIS bonded phase. This column was relatively inefficient compared with other column types available such as those using particles 3 and 5 pm in diameter but proved adequate for the development of this particular chromatographic system. However the over-all system was designed to exhibit very low extra-column dispersion so that it could well be suited for more efficient column systems if required.The mobile phase was similar in composition to that reported earlier,3 and consisted of an aqueous solutio 998 ANALYST AUGUST 1984 VOL. 109 containing 45 mM sodium tartrate (10.35 g 1-1) and 10 mM hexanesulphonate (1.88 g 1-1 sodium salt). The solution was adjusted to pH 3.10 by the addition of phosphoric acid and was passed through a 0.45-pm filter prior to use. The flow-rate was typically set at 1 ml min-1 and no special handling precautions were required. Generally fresh mobile phase solution was prepared at the start of each working day and the column was rinsed with de-ionised water at the end of the day and stored overnight in methanol.Fig. 1 illustrates the separation of a mixture of cation standards using the phase system just described. Each peak represents approximately 1 pg of the respective cation injected on to the column. The separation of the eight different cations is complete in less than 10 min and the eluted cations were reacted with 4-(2-pyridylazo)resorcinol (PAR) prior to detec-tion .3 Under suitable conditions the reaction occurs instan-taneously and yields a very intensely coloured complex with an adsorption maximum at about 500 nm. The reagent consisted of an aqueous solution containing 2 X 10-4 M PAR (44 mg 1-I) 1 M ammonium acetate (77 g 1-1) and ammonia solution (sp. gr. 0.88) (15 ml 1-1). It was found advisable to de-gas the reagent and store it under nitrogen to prevent oxidation.Table 1 lists the absorption maxima for several cation complexes. In addition to choosing the proper chemical system for post-column reaction the appropriate apparatus for mixing the mobile phase and reagent had to be designed to ensure a complete reaction and minimum band dispersion. Perfor-mance characteristics which are important in the design of a mixer include the capability of mixing rapidly and completely and in as small a volume as possible to ensure compatibility with high-efficiency columns. A mixer was designed and built that achieves these goals and is shown schematically in Fig. 2. It consisted of two pieces of 1/16 in 0.d. X 0.007 in i.d. tube 8 and 4 cm long. The 8-cm length of tube that was used to connect the column to the detector was radially bored in the centre with a circular 60" angle conical cut just to intersect with the central conduit.The end of the 4-cm length of 0.007 in i.d. tube which was used for reagent delivery was formed into a cone with a 60" angle cut, permitting the two pieces of tube to be interlocked at right-angles. The two tubes were housed within a drilled-out lil6-in Swagelok tee. The band dispersion contribution of this mixer was deter-mined after the manner described by Scott4 and the results are given in Table 2. The dispersion (02) was calculated from the band width measured at 0.6065 of the peak height. The results were compared with the dispersion obtained when using a straight 8-cm length of 0.007 in i.d. tube. The over-all system dispersion was minimised by using a Valco injection valve with a 0 .2 4 loop and an LC-85B detector incorporating a flow cell with a volume of 1.4 pl. The data in Table 2 indicate that within experimental error, no additional contribution to band dispersion occurred when using the mixer. The dispersion of the straight length of 0.007 in i.d. tube was measured using a flow-rate of water of 197 pl min-l. A mobile phase flow-rate of 101 pl min-1 and a reagent flow-rate of 105 p1 min-1 were used for the mixer measurements and consequently the total flow-rate approxi-mated to the original flow-rate through the straight tube. The slightly reduced dispersion when using the mixer (if statistic-ally valid) could be due to increased injection valve dispersion that could have taken place during the measurements with the straight tube as a higher flow-rate through the valve was utilised (197 vs.101 pl min-1). The chromatographic performance of this mixer is shown in Fig. 3. Chromatogram A shows the injection of 1 pg of Cu. Although relatively good niixing is achieved at this sensitivity, some residual mixing noise may be observed on the base line. This remaining problem was almost completely eliminated by incorporating a 5-cm length of low-dispersion serpentine tube I 0 5 10 Ti meimi n Fig. 1. Separation of 1 yg cation standards. Mobile phase 0.045 M Na + tartrate - 0.01 M hexanesulphonate pH 3.1 1 ml min-l at room temperature. Peaks A Cu; B Pb; C Zn; D Ni; E Co; F Cd; G. Fe; and H Mn Reagent inlet / To detector 4 i - j I / 6oo cut I 1 Fig.2. Mixing tee A 4 Time --+ From column Fig. 3. Improved mixing performance using a low-dispersion serpen-tine connecting tube. A 1 yg of Cu without serpentine tube; B 0.5 pg of Cu with a 5-cm length of serpentine tube Table 1. Absorption maxima for complexes of cations with PAR Cation - PAR Cation - PAR complex Lmax inm complex i,,, /nm Cu . . . . . . 507 Co . . . . . . 506 Pb . . . . . 498 Cd . . . . . . 485 Zn . . . . . . 489 Fe . . . . . . . . 481 N i t . . . . . . . 499 Mn . . . . . . 49 ANALYST AUGUST 1984 VOL. 109 999 Table 2. Dispersion contribution of the mixer cu*lp12 Run1 Run2 Run3 Mean 8.2 cm x 0.007 in i.d. tube* . . 6.95 6.48 7.33 6.92 Mixer? . . . . . . . . . 6.22 6.32 6.22 6.25 * Mobile phase flow-rate = 197 pI min-l.i- Mobile phase flow-rate = 101 pl min-1; reagent flow-rate = 105 p1 min ~ l . Fig. 4. Photograph of serpentine tube Cell (dimensions 0.74 rnrn i.d. x 4 r n r n ) Quartz window / Quartz 0.020in i.d. outlet tube ~ 1 1 w;ndow m LED leads LED retair nut Photo -cell . Photocell ' leads P hot ocel I -retaining nut Quartz ' Cell body \ / \ ~ O . O l O i n i . ~ u a r t z window Neoprene seals (2) inlet tube window retaining nut k--- 23 r n m d retaining nut Fig. 5 . Ion chromatograph detector immediately following the mixer. This low-dispersion serpen-tine tube design introduces radial mixing and has recently been described by Katz and Scott.5 It has been shown to contribute minimally to band dispersion even in long lengths, while at the same time providing excellent radial mixing due to secondary flow.The greatly improved performance obtained with the addition of this short length of low-dispersion serpentine tubing is shown in Fig. 3B where residual mixing noise is almost completely removed. This peak represented 500 ng of copper detected at the s a z e sensitivity as in chromatogram A . A photograph of a section of the serpentine tube used in this work is shown in Fig. 4. A fundamental interest during the course of this work was the design and construction of a dedicated detection system for cation analysis. Initially several design criteria were established as being important to the over-all successful implementation of this detector.The detector had to be simple compact and exhibit small band dispersion and thus permit its use with high-efficiency columns if desired. In addition the detector was to have good stability and linearity and offer sufficiently high sensitivity to be suitable for the analysis of a wide range of sample types at low concentrations. The detector incorporating the above features and perfor-mance characteristics is shown in Fig. 5 . The system incorpor-ates an LED light source a flow cell having a very small volume and short connecting tubes and a solid-state photo-diode as the sensing device. The associated electronics are very simple in design and allow a very stable detection system to be realized. The light source utilised in this detector was a green LED (Radio Shack 276-034) having a diameter of 5 mm and as expected it has proved to be exceptionally stable and drift-free in operation.The LED operated at a current of 10-35 mA exhibited a relatively broad band spectral charac-teristic having a maximum emission intensity at approximately 550 nm. The maximum light intensity occurs at a higher wavelength than that at which the cation complexes absorb (approximately 500 nm). However owing to the broad band emission of the light source and the broad band absorption spectra of the cation complexes spectral overlap occurs. Fig. 5 shows the layout of the detector and identifies the location of the LED light source and detector photocell. The entire detection system is 23 mm in length and other dimensions are as indicated.The illuminated cell volume has dimensions of 0.74 mm i.d. x 4 mm long giving a volume of 1.7 1.11. The mobile phase inlet tube consists of a length of 0.010 in i.d. tubing intersecting the LED side of the flow cell at a 45" angle. The flow cell exit tube is prepared from 0.020 in i.d. tubing to minimise cell back-pressure and intersects the opposite end of the flow cell also at a 45" angle but rotated 180" from the inlet tube position. The angular intersection of both inlet and exit tubes is designed to minimise the unswept volume within the illuminated path of the flow cell. Fig. 5 shows that both the LED and photocell (Hamamatzu S1336-5BQ-2L) are incorporated directly into the flow cell housing. This simplifies the design and obviates the need for an additional optical arrangement within the system which would be necessary if the individual detector components were isolated with respect to one another.This simplified design improves the stability of the system and permits maximum energy from the light source to be directed into the flow cell. The LED is inserted into the flow cell housing butted directly against one side of a circular quartz window and retained by a threaded collar. The other side of the quartz window is sealed against the stainless-steel housing by means of a thin neoprene washer which is compressed when the LED is properly seated. The use of neoprene is possible owing to the restricted nature of the mobile phase (aqueous mixtures) that was to be employed with the system.The photocell side of the flow cell had a similar arrangement except that the quartz window had a convex surface facing toward the photocell. In this manner the light beam passing through the flow cell is effectively dispersed on to the total sensitive area of the photocell. The contribution of the flow cell to band dispersion was determined using procedures previously described for deter-mining the dispersion of the reagent mixing system. The dispersion of the entire flow cell including the exit tube was determined using a measured flow-rate of 106 1.11 min-1. The dispersion peak was shown to be very asymmetric. The dispersion calculated from 20 (band width) measured at 0.6065 of the peak height was 7.26 1.112. As this value also includes the dispersion contribution from the 2.8 cm X 0.020 in i.d.exit tube the dispersion of a similar length of exit tube was also measured. The peak from the exit tube alone wa 1000 ANALYST. AUGUST 1984 VOL. 109 shown to have poor symmetry with a variance 02 of 7.50 PI?. The asymmetries of the peaks for both the flow cell and the exit tube were essentially identical. Further the contribution to dispersion of the flow cell appears to be due almost exclusively to the exit tube which has no significance under normal operating conditions. Fig. 6 shows schematically the detector electronics. The circuitry is very simple in design and has proved to be both reliable and stable in operation. A voltage follower is used to convert the light source current from one of high impedance to one of lower impedance and incorporates a BIFET opera-tional amplifier (No.TLO81CP). A very simple amplifier succeeds the voltage follower utilising a 741 operational amplifier (No. UA741CP) providing a gain of 60. A 1.8 cycle second-order Butterworth filter follows the second-stage amplifier to reduce noise and utilises another 741 operational amplifier. Various filter frequencies may be obtained by simply changing the appropriate capacitors. The amplifier operates from a +12 and -12 V stabilised power supply (not shown) and the LED from an appropriately attenuated +5 V stabilised power supply (also not shown). The linearity and detection limits for several cations were determined. The minimum detectable concentrations for several cations are listed in Table 3.These concentrations are equivalent to a concentration calculated to produce a signal equal to twice the noise level. The procedure used for calculating these values is that previously recommended by Scott.4 Approximately 10-8 g ml-1 of each cation is detect-able. Nickel gives the lowest sensitivity 2.3 X 10-7 g ml-1, whereas cobalt which gives the highest sensitivity could be detected at concentrations as low as 5.4 x 10-9 g ml-1. The linearity of the ion chromatograph over the cation concentra-tion range of interest was tested and found to be acceptable. A graph is shown in Fig. 7 illustrating the linearity for two cation standards. It should be pointed out that the high molar absorptivities of the complexes and the relatively low concen-trations (e.g.zinc ca. 7.9 x 10-5 g ml-1 and cobalt ca. 5.4 x 10-6 g ml-1) restrict the linear dynamic range to about three orders of magnitude but this is adequate for general cation analysis. High cation concentrations can be easily accommo-dated by dilution. Instrument Operating Conditions The flow-rates for both the mobile phase and reagent solutions were typically set at 1 ml min-1. During preliminary experi-ments designed to determine the appropriate phase system, the effect of column temperature on the separation was evaluated. These results indicated that an increased column temperature did not provide greatly improved separations using this particular phase system so for this work all separations were performed at room temperature. The use of room temperature also simplified the instrumentation required.The operating current applied to the LED was varied between 10 and 35 mA but most experimental data were collected using an LED current of 25 mA. Chromatograms were displayed using a Model 56 recorder utilising a variable sensitivity range from 1 mV to 10 V full-scale deflection. Signals were appropriately adjusted by use of the recorder attenuator which was typically operated in the range from 100 mV to 10 V. The detector electronics were balanced by use of the fine and coarse balance control on the amplifier. During the course of this work two different injection techniques were utilised. Development of the phase system involved the use of a 6-p1 injection loop for injecting standard solutions of the various cations.In addition actual sample analysis involving high cation concentrations also utilised this injection volume. A larger volume loop (2.3 ml) was utilised for evaluating on-column cation concentration. The capability of utilising this system for determining "P-10 m 2 Y 5 aJ a Butterworth filter Voltage Amplifier G = 1.58 follower G = 60 T = 100 ms 10M 6k 27 k __t 0.05 mF t -V 500A t t -V -V Fig. 6. Detector electronics for ion chromatograph 0 200 400 600 800 0 200 400 600 800 Amount injecteding Fig. 7. Graph of amount injected versus peak area I A I I I I 0 5 10 Timeim in Fig. 8. Analysis of a brass sample. Mobile phase 0.045 M Na + tartrate - 0.01 M hexanesulphonate pH 3.4 1 rnl min-l at room temperature.A Copper (65"/,); B zinc (35%) Table 3. Minimum detectable concentrations of cations Minimum detectable Minimum detectable Cation concentration/g ml- 1 Cation concentratidg ml-I c u . . . . 3.2 x 10-7 co . . . . 5.4 x 10-9 Pb . . . . 4.7 x 10-* Cd . . . . 1.0 x Zn . . . . 7.9 X 10-8 Fe . . . . 2.4 X 10-* Ni . . . . 2.3 X 10-7 Mn . . . . 3.3 X cations at high concentrations is illustrated in Fig. 8 which shows the analysis of a brass sample. The sample was prepared by dissolving a sample of brass in 100 ml of HN03 - HCl - H20 (1 + 1 + 2). A 6-pl volume of the sample was injected. The chromatogram indicates the presence of copper (65%) and zinc (35%). As a very small sample volume of acid was employed and further the sample was injected into a buffered mobile phase special treatment of the stainless-steel surface was not necessary ANALYST AUGUST 1984.VOL. 109 1001 On-column Cation Concentration The eventual success of a liquid chromatography-based cation analyser is in part based on its ability to analyse cations in real samples at very low concentrations. Data that have already been presented indicate that very low detection limits may be achieved using a very simple detection sytem. These data were obtained using 6-p1 injections of standard cation solutions and as a result do not adequately predict the concentration sensitivity obtained when large sample volumes (>1 ml) are injected. Typically when large sample volumes are injected spreading of the chromatographic peaks occurs, resulting in poorer concentration sensitivity.A method of improving concentration sensitivity is to choose a phase system permitting the concentration of the cations on to the top of the column during injection. This is followed by the re-initiation of the elution processes and as the mobile phase again begins to pass through the column it lifts the solutes from the front of the column for subsequent development. This is commonly achieved by the use of a concentration column placed before the analytical column and an appropriate valve switching arrangement. An additional advantage can be realised in terms of both simplicity and speed if sample concentration can be achieved directly on the analytical column. The separation system described utilises hexanksulphonic acid dissolved in sodium tartrate solution as the mobile phase.The sulphonate groups act as a sorbed ion exchanger (counter ion) on the column packing with finite retention in the presence of the tartrate solution. Therefore the counter ion is distributed in equilibrium between stationary phase (packing) and the liquid phase (mobile phase). The hexanesulphonate is continuously eluted from the column and replenished by the introduction of new mobile phase. In a situation where a large volume of pure water is injected, this equilibrium is disturbed as there is no tartrate present to modify the sulphonate distribution. Consequently the hex-anesulphonate is adsorbed very strongly on to the stationary phase. The phenomena of the adsorption of sulphonates on reversed-phase materials has been described by Scott and Kucera.6 Under these conditions the cations are strongly attracted to the adsorbed ion exchanger and are collected as a thin band on the front of the column.When the mobile phase containing hexanesulphonate and tartrate again enters the column the concentrated cations behave as though injected in a smaller sample volume instead of being dispersed within the entire injection aliquot. These principles are illustrated in Fig. 9. Fig. 9(a) shows the results obtained from the injection of a 1-ml Cu sample at a concentration of 1 p.p.m. made up in the mobile phase. A very broad chromatographic peak is obtained indicating that no sample concentration at the top of the column occurs. In comparison Fig. 9(b) shows a similar Cu injection contained in a 0.1% solution of HN03 in pure water.In this instance a much more sharply defined peak is observed owing to the fact that on-column concentration occurs and thus the concentra-tion sensitivity is improved. Fig. 10 shows the effect of varying the injection volume from 0.25 to 2 ml on the relative peak height for a 100 p.p.b. solution of cobalt and zinc in 0.1% HN03. The use of peak heights is important as these measurements directly relate to the efficiency of on-column sample concentration over the sample size range and it can be seen that good linearity is obtained up to an injection volume of 2 ml. The index of determination and coefficient of variation were 0.9983,2.51% and 0.9960 3.7% for zinc and cobalt respectively. These errors of 2.5 and 3.7"/0 are associated with both the detector and the on-column concentration procedure.This level of error when determining concentrations at the parts per billion level represents a very satisfactory precision. The results shown indicate that when using these simple on-column sample concentration procedures very low con-0 4 Tirneimin 8 Fig. 9. On-column concentration of 1 p.p.m. of copper. ( u ) 1 ml injection of copper solution prepared in the mobile phase; ( h ) 1 ml injection of copper solution prepared in 0.1 '% nitric acid 0 Zinc Y a a 0.5 1.0 1.5 2.0 2.5 200 150 100 50 0 Volume injected/ml Fig. 10. containing 100 p.p.b. of zinc and cobalt Graph of peak height versus sample volume for a solution 0 5 10 Ti meim i n Fig.11. Injection of 2.3 ml of 50 p.p.b. cation solutions centrations can be determined successfully. Fig. 11 shows chromatograms from the analysis of 50 p.p.b. solutions of several others cations. It should be noted that in Fig. 11 there is a large extraneous peak due to iron. The iron was shown to be a contaminant from the chromatographic system in particular from the ferrules used in the column and tubing connections and the large sample loop. The contaminant peak decreased in size with use as the available iron was leached from the surface. It is therefore advisable to employ ferrules made from inert materials and wherever possible sample loops that are chemically passive 1002 a) ANALYST AUGUST 1984 VOL. 109 A k 0 6 12 Inj. L 0 6 A E D I I I I I I 0 6 12 0 6 12 Ti meimi n Fig.12. To analysis of 50 p.p.b. cation standards in 1% NaCl ( a ) water; ( h ) lo!NaC1. Peaks A Cu; B Pb; C. Zn; D Ni; E Co; F Fe; and G Mn. Bottom analysis of zinc in urine (c) 1 p.p.m. Zn standard; (d) urine The effect II any of matrix constituents such as sodium chloride in sea water when sampling large volumes was also briefly examined. Fig. 12 ( a ) and ( b ) shows a comparison of results obtained for a 500 p.p.b. standard test mix when injected in 0.1% HN03 and a similar solution containing 1% of NaCl. This matrix would be similar to that of sea water. but the cation concentrations are much higher than would normally be found. No observable matrix effects were noted. Fig. 12 ( c ) and ( d ) show chromatograms for a sample representing a potentially difficult biological matrix.The urine sample was filtered and diluted 3-fold and analysed directly for zinc. No matrix problems were observed and the zinc concentration was calculated to be 0.5 p.p.m. It should be pointed out that all the analyses demonstrated in this paper were carried out on columns of indifferent efficiency. The use of high-efficiency columns would provide better resolution and greatly reduced analysis times. The use of high-efficiency columns with the apparatus described here will be the subject of a future paper. Conclusions Cations can be determined by non-suppressed ion chromato-graphic procedures at very low concentrations with relatively inexpensive equipment. Utilising an absorption detector with a stable broad emission LED source coupled with the post-column reaction of the eluted cations a linear dynamic range of over three orders of magnitude can be achieved.The system permits samples to be analysed at a level of a few parts per billion. Utilising a reversed-phase column and an ab-sorbed ion exchanger the column can be also used satisfac-torily for sample concentration. The presence of high concen-trations of saline (for example in sea water) and high concentrations of organics (for example in urine) are shown to be easily tolerated and do not affect the accuracy of the analysis. It is likely that in the future ion chromatography will be more extensively used for cation analysis and be applied to a wide range of sample types offering both high sensitivity and multi-element capability. The authors thank Mr. Paul Rebmann for his help with the construction of the apparatus Ms. Laura Smith for her help with the experimental work and Prof. Charles Lochmiiller for many helpful discussions. References 1. 2. 3. 4. 5 . 6. Small H. Stevens T. S . and Bauman W. C . Anal. Chem., 1975 47 1801. Elchuk S . and Cassidy R. M. Anal. Chem. 1979 51 1434. Elchuk S . and Cassidy R. M. Anal. Chem. 1982 54 1558. Scott R. P. W. “Contemporary Liquid Chromatography,” Wiley-Interscience New York 1974 p. 139. Katz E. D. and Scott R. P. W. J. Chromatogr. 1983 268, 169. Scott R. P. W. and Kucera P. J. Chromatogr. 1977 142, 213. Paper A411 11 Received March 20th 1984 Accepted March 26th 198

ISSN:0003-2654    

DOI:10.1039/AN9840900997

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST, AUGUST 1984, VOL. 109 1003 Fluorescence Analysis of Thiols with Ammonium 7-Fluoro benzo-2-oxa- 1,3=diazole=4=su Ip honate Toshimasa Toyo'oka and Kazuhiro Imai* Faculty of Pharmaceutical Sciences, University of Tokyo, 7-3- I Hongo, Bunkyo-ku, Tokyo 1 13, Japan ~ ~ ~ ~ ~~~~ The fluorigenic reaction of thiols with ammonium 7-fluorobenzo-2-oxa-1,3-diazole-4-sulphonate (SBD-F) and the factors that affected the fluorescence of SBD-thiols were investigated. High pH, high temperature and organic solvents increased the reaction rates. The fluorescence intensity of the fluorophore was dependent on pH. Under suitable conditions (at pH 9.5 and 60 "C for 1 h), the detection limits for thiols such as cysteine, glutathione, captopril and bovine serum albumin were in the range 43-520 pmol ml-I.The fact that blue shifts of the fluorophore were observed depending on the hydrophobicity of the added solvents indicated the possibility of its use for the environmental analysis of reactive thiols of macromolecules with SBD-F. Keywords: Fluorescence analysis; thiol analysis; ammonium 7-fluorobenzo-2-oxa- 7,3-diazole-4-sulp honate; fluorescence probe reagent In living tissues there are many thiol-containing compounds such as cysteine or glutathione or macromolecules such as proteins or enzymes, which are known to play important roles. 1.2 For their measurement, selective and sensitive methods are required and in this respect fluorimetry appears to be the most suitable. Different types of fluorigenic reagents for the thiol groups,3 e.g., iodoacetamide derivatives such as N-(iodoacetylaminoethyl)-5-naph thylamine- l-sulphonic acid (1 ,5-I-AEDANS) ,&8 N-dansylaziridine ,9,lo N-substituted maleimides such as N-(7-dimethylamino-4-methyl-3- coumariny1)maleimide (DACM) ,11-13 bimanelG18 and halo- genated benzofurazans such as 7-fluoro-4-nitrobenzo-2-oxa- 1,3-diazole (NBD-F) ,19-22 have been reported.Among them, ammonium 7-chlorobenzo-2-oxa-l,3-diazole-4-sulphonate (SBD-Cl), recently reported by Andrews et al.,23 an analogue of 4-chloro-7-nitrobenzo-2-oxa-l,3-diazole (NBD-C1) ,19 seems to be an excellent reagent as regards its stability and solubility in water. However, its reactivity towards thiols is very low. Therefore, in previous ~0rk,24 ammonium 7-fluorobenzo-2-oxa-l,3-diazole-4-sulphonate (SBD-F) (Fig.1) was developed as an alternative to SBD-Cl in order to increase the reactivity towards thiols. It was used as a pre-column derivatisation reagent for high-performance liquid chromatography (HPLC) for the determination of glutathione and cysteine in blood and plasma25 and of captopril (an antihypertensive agent) in dog plasma.26 In this paper, we report further investigations of the reactivity of SBD-F towards thiols and discuss its applicability in the life sciences. Experimental Materials SBD-F was synthesised and purified according to the pro- cedure described previously.24 SBD-Cl (Pierce, Rockford, IL, USA) was used as received. The SBD adduct of mercapto- ethanol (SBD-mercaptoethanol) was prepared by the method F SR I S03-NH4' SBD-F SBD-SR Fig. 1. Structure of SBD-F and its reaction with thiols * To whom correspondence should be addressed.of Andrews et a1.23 (sublimed at 310 "C). Cysteine.HC1, alanine, proline, cystine, serine and tyrosine were obtained from Ajinomoto (Tokyo, Japan). Homocysteine, cysteamine.HC1, glutathione, N-acetylcysteine, P-hydroxy- phenylethylamine (HPEA) , N-methyl-p-hydroxy- phenylethylamine (MHPEA) and bovine serum a1 bumin (BSA, fraction V) were purchased from Sigma (St. Louis, MO, USA). a-Mercaptopropionylglycine (Fluka, Buchs, Switzerland), coenzyme A (Boehringer, Mannheim, FRG), mercaptoethanol (Toyko Kasei, Tokyo, Japan) and ethylene- diaminetetraacetic acid, disodium salt (EDTA.2Na) (Kanto Chemical, Tokyo, Japan) were used. l-(~-3-Mercapto-2- methyl-l-oxopropy1)-L-proline (captopril) was kindly donated by Sankyo (Tokyo, Japan).All other chemicals were of analytical-reagent grade. De-ionised, distilled water was used. Apparatus A Hitachi 65&10S fluorescence spectrophotometer was used with a l-cm quartz cell or an 18-pl flow cell and operated at ambient temperature. The fluorescence maximum wavelengths were recorded without correction. A Waters high-performance liquid chromatograph, equipped with a U6K universal injector and a Model 6000A pump was used. Ultraviolet (UV) spectra were measured with a Uvidec 505 instrument (JASCO, Tokyo, Japan). The reaction tempera- ture was controlled by a Model JB 1 water-bath (Grant Instruments, Cambridge, UK). A pBondapak CI8 column (300 X 3.9 mm i.d., 8-10 pm) connected to a guard column of Bondapak C18-Corasil (20 x 3.9 mm i.d., 37-50 pm) was used.The eluting solvent was methanol - 0.1 M phosphate buffer (pH 6.0) prepared with 0.1 M sodium dihydrogen phosphate and 0.1 M disodium hydrogen phosphate ( 5 + 95). The flow-rate of the eluent was 1.0 ml min-1. The column temperature was ambient. The eluate was monitored at 490 nm with excitation at 360 nm. Proton nuclear magnetic resonance (1H NMR) spectra were recorded on a Jeol Model FX-100 spectrometer at 100 MHz using tetramethylsilane (TMS) as an internal standard (abbreviations: s = singlet, d = doublet, q = quartet, m = multiplet). Syntheses of SBD Adducts of Cysteine (SBD-cysteine) and Homocysteine (SBD-homocysteine) SBD-F (0.47 g, 2 mmol) and an equal amount of thiol (cysteine.HCl,O.31 g, or homocysteine, 0.27 g) were dissolved in 80 ml of 0.1 M borate buffer (pH 9.5, Naf) containing 1 mM EDTA.2Na.The solution was mixed and allowed to stand at 60 "C for 1 h. After cooling with ice - water, the solution was neutralised with 2 M HC1 and then evaporated in vacuo. The1004 ANALYST, AUGUST 1084, VOL. 100 residue was dissolved in 5 ml of water and applied to a Bio-Gel P-2 column (200-400 mesh, 50 x 2.0 cm). The column was eluted with water. The fractions corresponding to the yellow fluorescent band were collected and evaporated to dryness under reduced pressure. SBD-cysteine (recrystallised from EtOH - water, yellow needles, decomposed above 180 "C): calculated for C9H7N306S2Na2, C 29.84, H 1.95, N 11.60; found, C29.66, H 2.25, N 11.90%. NMR (in DMSO-d6): 6 7.46 (1H, d , J = 7.1 bH aHQco so3 Hz, a), 7.69 (1H, d , J = 7.1 Hz, b ) , 3.17-3.92 (5H, m, c + d + e).UV: h,,, (water) = 370 nm, E (370 nm) = 3.8 X 103 1 mol-1 cm-1. Fluorescence in water: he,, 375 nm; A,, , 500 nm. SBD-homocvsteine (yellow needles, decomposed above 285 "C): calcuiated ~ O ~ ' C , ~ ~ H ~ ~ N ~ O ~ S ~ N ~ ~ . H ~ O ~ C 2.80, N 10.63; found, C 30.82, H 2.94, N 10.76%. DMSO-d6): 6,7.40 (1H, d , J = 7.1 Hz, a ) , 7.66 ( l € I , f N H2 C d I SCHZ-CHTc~- COO- I e so3- 30.38, H NMR (in d , J = 7.1 Hz, b ) , 2.06 (2H, q , J = 6.0 Hz, d ) , 3.10-3.65 (5H, m, c + e + f). UV: A,,, (water) = 380 nm, E (380 nm) = 6.9 x 103 1 mol-1 cm-l. Fluorescence in water: A,,,, 385 nm; A, , 515 nm. Synthesis of SBD Adduct of Methanol (SBD-methanol) To 40 ml of SBD-F (0.24 g, 1 mmol) in 0.1 M NaOH was added an equal volume of MeOH.After allowing the solution to stand at room temperature for 2 h, it was neutralised with 2 M HC1 and evaporated in vacuo. A 100-ml volume of methanol was added to the resultant residue and the suspension was filtered. The filtrate was evaporated and chromatographed on a Bio-Gel P-2 column (200-400 mesh, 50 X 2.0 cm) using water as the eluent. The yellow fluorescent fractions corre- sponding to SBD-methanol were collected and evaporated to dryness under reduced pressure. The SBD-methanol obtained was recrystallised from EtOH (yellow needles, decomposed above 300 "C). Calculated for C7HSN2OsSNa: C 33.34, H 2.00, N 11.11; found, C 33.54, H 1.92, N, 11.28%. NMR (in C OCH3 DMSO-d6):6,6.80(1H,d,J=7.7H~,a),7.67(1H,d,J=7.7 Hz, b ) , 4.02 (3H, s, c).UV: k,,, (water) = 348 nm, E (348 nm) = 4.9 x 103 1 mol-1 cm-1. Fluorescence in water: h,,,, 355 nm; he, , 490 nm. HPLC of SBD-methanol and a Fluorigenic Reaction Mixture A 10+1 volume of 1 PM authentic SBD-methanol in 0.1 M borate buffer (pH 9.5, Na+) was subjected to HPLC. A reagent blank was obtained after the reaction of 500 VM SBD-F in 0.1 M borate buffer (pH 9.5, Naf) - methanol (9 + 1) at 60 "C for 1 h. A 10-pl volume of the solution was subjected to HPLC (Ac, , 360 nm; A,,, , 490 nm). Investigation on the Fluorescence Reaction Homocysteine was selected as a typical thiol for the fluorigenic reaction with SBD-F. A 1-mi volume of 1 mM SBD-F in the buffer (or the buffer containing EDTA.2Na) and an equal volume of 10 VM homocysteine in the buffer (or buffer plus organic solvent) was immediately mixed in glass tubes.The tubes were capped and heated at a certain temperature (40-60 "C). At certain time intervals, the tube was taken out and cooled with ice - water. The fluorescence intensities were measured at ambient temperature with emission at 515 nm (excitation, 380 nm). The reagent blank without thiol was treated in the same manner as described above. The reaction rate constant (pseudo-first order) was calculated on the difference of the fluorescence intensity (sample fluorescence minus blank fluorescence) from that of authentic SBD-homocysteine. Fluorescence Intensities and Spectral Shifts of Authentic SBD-thiols at Various pH Values and in Different Solvents The relative fluorescence intensities of 2.5 PM SBD-thiols (cysteine, homocysteine and mercaptoethanol) at various pH values (0.05 M Britton - Robinson buffer for pH 2-12, 0.1 M HCl for pH 1 and 0.1 M NaOH for pH 13) were measured at the maximum wavelengths.SBD-homocysteine (0.38 p ~ ) was dissolved in various solvents [water, MeOH, EtOH, acetonit- rile, acetone, dimethylformamide (DMF) and dimethylsul- phoxide (DMSO)] and its maximum wavelengths and fluores- cence intensities were measured. Recommended Procedure for the Determination of Thiols To 1.0 ml of 10 VM thiol in 0.1 M borate buffer (pH 9.5, Na+) containing 2 mM EDTA.2Na was added an equal volume of 1.0 mM SBD-F in 0.1 M borate buffer (pH 9.5, Naf). The solution was thoroughly mixed and maintained at 60 "C for 1 h. After cooling with ice - water, the fluorescence intensity was measured at ambient temperature with emission at 5 15 nm (excitation, 380 nm).The reagent blank without thiol was treated in the same manner as above. For the determination of cysteine or homocysteine, the chilled solution was adjusted to pH 2 with 2 M HC1 and the fluorescence intensity was measured at certain wavelengths. Results Optimisation of the Fluorigenic Reaction As thiols are easily oxidised with dissolved oxygen in alkaline media2.23.27 and the oxidation is catalytically accelerated with various metals, EDTA.2Na was added to the reaction medium in the range 0.1-5.0 m M to investigate the effect. Even the lowest EDTA.2Na concentration (0.1 mM) of those tested was effective compared with the fluorigenic reaction without EDTA.2Na.Therefore, buffers containing 1 mM EDTA.2Na were used in subsequent work. As can be seen in Table 1, the reaction rates gradually increased with increasing pH, and the fastest rate was obtained at the highest pH tested (10.0). Enhancement of the reaction rate was also observed with increased temperature. With an increase in temperature of 10 "C the reaction rate approximately doubled (Table 1).ANALYST, AUGUST 1984, VOL. 109 1005 Table 1. Effects of pH and temperature on the fluorigenic reaction rates. 5 VM homocysteine and 500 VM SBD-F were reacted in 0.1 M borate buffer containing 1 mM EDTA.2Na. hex ,385 nm; he, ,515 nm klmin-1 PH 40 "C 50 "C 60 "C 8.0 ND* ND* 2.31 x 8.5 ND* ND* 3.02 X 10-2 9.0 1.69 x 10-2 3.18 x 6.24 X 10-2 9.5 2.44 x 4.40 x 10-2 8.01 x 10-2 10.0 NTt NTI 1.09 X 10-1 * ND = not determined. t NT = not tested.Table 2. Effects of buffer species and buffer concentration on fluorigenic reaction rates. 5 VM homocysteine and 500 VM SBD-F were reacted at 60 "C in buffer (pH 9.5) containing 1 mM EDTA.2Na he, , 385 nm; he,., 515 nm. Buffer preparations: carbonate, 0.1 M NaHC03 + 0.1 M Na2C03; ammonium, 0.1 M NH3 + 0.1 M NH,Cl; glycine, 0.1 M glycine + 0.1 M NaOH; veronal, 0.1 M sodium veronal + 0.1 M HCI; borate, disodium tetraborate(II1) + NaOH Buffer species Buffer concentrationh Carbonate . . 0.1 Ammonium . . 0.1 Verona1 . . . . 0.1 Borate . . . . 0.1 Glycine . . . . 0.1 0.05 0.02 0.01 klmin-1 8.87 x 10-3 1.28 x 10-2 2.56 x 4.97 x 10- 2 8.01 x 10-2 8.11 X lop2 7.49 x 10-2 5.68 X 10-2 0 0.053 0.111 0.176 Solvent : H20( VIW Fig.2. Effect of solvent on the fluorigenic reaction. 5 PM homocy- steine and 500 VM SBD-F were reacted with the solvents at 60 "C in 0.1 M borate buffer (pH 9.5, Na+, 1 mM EDTA.2Na). Fluorescence intensity was measured at 515 nm (hex 385 nm). Solvent: A, acetonitrile; B, DMSO; and C, acetone The type of buffer also affected the reaction rate. Of the buffers tested (ammonium, borate, carbonate, glycine or veronal), borate gave the highest yield of SBD-thiol (Table 2 ) . A concentration of borate buffer higher than 0.05 M was necessary for completion of the reaction (Table 2). As shown in Fig. 2, acetonitrile was the most effective in increasing the fluorescence yield and the reaction rate increased with increasing acetonitrile concentration, but precipitation of borate occurred with more than 20% aceto- nitrile. Acetone suppressed the reaction.With methanol, the excess reacts gradually with SBD-F to afford a high reagent blank. The reagent blank fluorescence was the same as that of authentic SBD-methanol. The resultant fluorophore obtained from the reagent blank was identified as SBD-methanol by chromatographic comparison with the authentic compound (the retention times were both 19 min). >- 800 c v) C a c 0 C a v) .- c. .- a 600 400 3 a - Lc .- e ; 200 cc 0 0.1 M 2 4 6 8 10 12 0.1 M HCI NaOH PH Fig. 3. Effect of pH on the fluorescence intensities of authentic SBD-thiols. Fluorescence measurement: A , SBD-mercaptoethanol, 520 nm (Aex, 385 nm); B, SBD-homocysteine, 515 nm (hex, 385 nm); and C, SBD-cysteine, 505 nm (hex.375 nm). Concentration of SBD-thiols: 2.5 p~ each. Buffers: 0.05 M Britton - Robinson (pH 2-12) Effect of pH on Fluorescence of SBD-thiols The fluorescence intensities of many fluorophores are affected by pH, and those of authentic SBD-cysteine, -homocysteine and -mercaptoethanol were dependent on pH. As shown in Fig. 3, high relative fluorescence intensities were observed from pH 2 to 7 with all the SBD-thiols tested, except SBD-cysteine or -homocysteine, for which the highest fluores- cence intensity was obtained at about pH 2 . The SBD derivatives of amino acids (cysteine and homocysteine) may give the higher fluorescence in the -NH3+ form because the fluorescence of SBD-mercaptoethanol was constant over a wide pH range (pH 2-11) (Fig.3). Considering the above results, an acidic medium (about pH 2) would be preferable for fluorescence measurement of thiol-containing amino acids. Detection Limits for Various Tkiols From consideration of the above results, appropriate condi- tions [in 0.1 M borate (pH 9.5, Naf), at 60 "C, for 1 h, as under Recommended Procedure for the Determination of Thiols] were selected for the determination of thiols. The detection limits (the net fluorescence intensity of twice the reagent blank fluorescence intensity) were in the range 43-2600 pmol ml-1 (Table 3). However, the detection limits of cysteine and homocysteine at pH 2 were lowered (Table 3). The reason for the higher detection limit for cysteine compared with that for homocysteine seems to be the small molar absorptivity (3.8 X l o 3 1 mol-1 cm-1 for cysteine versus 6.9 X 103 1 mol-1 cm-1 for homocysteine). In contrast, no disulphide (cystine), amine (HPEA or MHPEA), hydroxylamino acid (serine) or amino acid (alanine or proline) was detected under these conditions.The detection limit for BSA was 9.2 vg ml-1 (95 pmol ml-* SH equivalent). Correlation of Fluorescence with the Environment Increases in the fluorescence intensities of the authentic SBD-thiols were observed on changing the medium from aqueous to non-aqueous solvents, accompanied by blue shifts of the maximum florescence emission wavelength (Fig. 4, Table 4). Discussion The results presented here support the idea that SBD-F, like other halogenated benzofurazans,28J9 reacts with nucleo- philes (thiols) through an intermediate (SN2 reaction),30,31 as a1006 ANALYST, AUGUST 1984, VOL.109 Table 3. Detection limits for some thiols with SBD-F. Thiols and 500 p~ SBD-F were reacted at 60 “C for 1 h in 0.1 M borate buffer (pH 9.5, Na+) containing 1 mM EDTA.2Na Thiol Mercaptoethanol . . . . CoenzymeA . . . . . . Homocysteine . . . . Cysteine . . . . . . Cystine . . . . . . . . a-Mercaptopropionylglycine Glutathione . . . . . . Captopril . . . . . . Cystearnine . . . . . . N-Acetylcysteine . . . . Alanine . . . . . . . . Proline . . . . . . . . Tyrosine . . . . . . Serine . . . . . . . . HPEA . . . . . . . . MHPEA . . . . . . BSA . . . . . . . . Detection limit (S/B = 2)lpmol ml-1* . . 43 . . 72 . . loot . . 120t . . 150t . . 160t .. 330t; 120-1: . . 390t . . 2600;520§ . . NDlI . . ND . . ND . . ND . . ND . . ND . . ND . . 9.2pgml- lj/ (135 pmol ml- l ) * hex ,380 nm; he, ,515 nm. S = Sample fluorescence intensity; B = t Reference 24. i Adjusted to pH 2 with 2 M HCl. 9 Adjusted to pH 2 with 2 M HCl; hex , 375 nm; he, , 505 nm. 7 ND = not detected. /I hex , 380 nm; he, , 495 nm. blank fluorescence intensity. Table 4. Comparison of maximum wavelengths and relative fluorescence intensities of SBD-homocysteine in various solvents. 0.38 p . ~ SBD-homocysteine was dissolved in various solvents and fluorescence intensities were measured at maximum wavelengths. The fluorescence intensity of SBD-homocysteine in water was arbitrarily taken as 100 Solvent Water . . MeOH . . EtOH . . Acetonitrile Acetone .. DMF _ . DMSO . . . . . . . . . . . . . . . . hex 387 384 382 384 382 389 395 hem 516 499 493 488 480 487 495 Relative fluorescence intensity 100 260 550 110 230 220 170 higher pH in the presence of hydrophobic solvents such as acetonitrile was preferable for an effective reaction. When an organic solvent was added to the medium the blank value also increased. Hence, in the manual method as adopted in the present experiment, the signal to noise ratio (sample fluorescence intensity to blank fluorescence intensity) cannot be improved even in the presence of an organic solvent such as acetonitrile. However, this positive effect might be helpful for enhancing the sensitivity in post-column reactions in HPLC because the organic solvent gives high signals in spite of the high background level.The gradual reaction of methanol with SBD-F to afford SBD-methanol suggests that its use should be avoided in the measurement of thiols. On lowering the pH of the final medium, the detection limits for cysteine and homocysteine were lower than those obtained in the previous paper.24 The sensitivity of the method, comparable to that with DACM,32 might be suffi- cient for its use for the determination of thiols in biological samples. In general, the spectral shifts of the maximum fluorescence were correlated with hydrophobicity around the fluoro- ph0re.33~34 Although the blue shift of the maximum fluores- cence wavelength of SBD-thiol caused by the addition of an organic solvent is not great as that of N-( l-anilinonaphthyl)-4- maleimide(ANM)-thiol,35 the phenomenon might be applic- able to the analysis of the environment around the reactive thiols of macromolecules.As indicated in Table 3, SBD-F reacts with proteins such as BSA, and the hydrophobic region of enzymes, e.g., papain,36 blomelain36 and ficin,36 may be elucidated by applying the reagent. Major features of SBD-F are: (a) its high solubility in water and (b) more controlled reactivity towards thiols than ANM. The other features, that (c) the SBD moiety is not subjected to the S -+ N transfer known to occur for the NBD (4-nitrobenzo-2-oxa-l,3-diazole) group,37 (d) it is stable over a wide pH range, unlike the maleimide structure,3X and (e) it is not photosensitive like the I-AEDANS isomers,5 might be advantageous for its use in the life sciences. In conclusion, SBD-F might be useful for the sensitive detection of low relative molecular mass thiols using both the 60 70 80 90 Z Fig.4. Polarity dependence of fluorescence maxima of SBD- homocysteine. Concentration of SBD-homocysteine: 0.38 p ~ . Sol- vent: 0, water; ., ethanol; V, DMSO; A, acetone; A, methanol; 0, acetonitrile; and 0, DMF. kK (10-3 cm-1) is the fluorescence maximum wavelength, and 2 (kcal mol-1) is the transition energy manual and HPLC methods, and also useful for the environ- mental analysis of macromolecules such as enzymes and membraneous proteins. The authors thank Professor Z. Tamura of Keio University and Professor T. Nakajima of the University of Tokyo for valuable discussions. Thanks are also due to Mr. Y. Kawahara of Sankyo Company for the donation of captopril, Dr.Y. Watanabe of Chugai Pharmaceutical for NMR measurements and Dr. M. C. Khosla of the Cleveland Clinic for useful comments on the manuscript. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. References Friedman, M., “The Chemistry and Biochemistry of the Sulfhydry! Group in Amino Acids, Peptides and Proteins,” Pergamon Press, New York, 1973. Jocelyn, P. C., “Biochemistry of the SH Group,” Academic Press, New York, 1972. Brocklehurst, K., Znt. J. Biochem., 1979, 10, 259. Hartig, P. R., Bertrand, N. J., and Sauer, K., Biochemistry, 1977, 16, 4275. Hudson, E. N., and Weber, G . , Biochemistry, 1973, 12,4154. Wu, C. W., and Stryer, L., Proc. Natf. Acad. Sci. USA, 1972, 69, 1104. Friedman, F. K., Chang, M. Y., and Beychok, S ., J. Biol. Chem., 1978, 253, 2368. Haugeland, R. P., J. Supramol. Struct., 1975, 3, 192. Scouten, W. H . , Lubcher, R . , and Baughman, W . , Biochim. Biophys. Acra, 1974, 336, 421. Johnson, J. D . , and Schwartq, A , , J. Biol. Chem., 1978, 253, 5243. Glazer, A . N., Rev. Biochem., 1970, 39, 101. Kanaoka, Y., Yakugaku Zasshi, 1980, 100,973. Weltman, J. K., Szaro, R. P., Frackelton, A. R., Bunting, J. R., and Cathou, R. E., J. Biol. Chem., 1973, 248, 3173.ANALYST, AUGUST 1984, VOL. 109 1007 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. Machida, M., Machida, M., Sekine, T., and Kanaoka, Y., Chem. Pharm. Bull., 1977, 25, 1678. Takahashi, H., Nara, Y., and Tuzimura, K., Agric. Biof. Chem., 1978, 42, 769. Kosower, N. S., Kosower, E . M., Newton, G .L., and Ranney, H. M., Proc. Natl. Acad. Sci. USA, 1979, 76, 3382. Kosower, N. S., Newton, G . L., Kosower, E . M., and Ranney, H. M., Biochim. Biophys. Acta, 1980, 622, 201. Newton, G. L., Dorian, R., and Fahey, R. C., Anal. Biochem., 1981, 114, 383. Ghosh, P. B., and Whitehouse, M. W., Biochem. J., 1968,108, 155. Fager, R . S., Kutina, C. B., and Abrahamson, E. M., Anal. Biochem., 1973, 53, 290. Imai, K., and Watanabe, Y., Anal. Chim. Acta, 1981,130,377. Watanabe, Y., and Imai, K., J. Chromatogr., 1982, 239,723. Andrews, J. L., Ghosh, P., Ternai, B., and Whitehouse, M. W., Arch. Biochem. Biophys., 1982, 214, 386. Imai, K., Toyo’oka, T., and Watanabe, Y., Anal. Biochem., 1983, 128, 471. Toyo’oka, T., and Imai, K., J. Chrornatogr., 1983, 282, 495. Toyo’oka, T., Imai, K., and Kawahara, Y., J. Pharm. Biomed. Anal., in the press. Kawahara, Y., Hisaoka, M., Yamazaki, Y., Inage, A., and Morioka, T., Chem. Pharm. Bull., 1981, 29, 150. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. Di Nunno, L., and Florio, S . , J. Chem. SOC., Perkin Trans. 2, 1975, 1469. Di Nunno, L., and Florio, S., Tetrahedron, 1974, 30, 863. Ingold, C. K., “Structure and Mechanism in Organic Chemistry,” Cornell University Press, Ithaca, NY, 1969. Miller, J., “Aromatic Nucleophillic Substitution,” Elsevier, Amsterdam, 1969. KBgedal, B., and Kallberg, M., J. Chromatogr., 1982,229,409. Kosower, E. M., J . Am. Chem. SOC., 1959, 80, 3253. Kosower, E. M., J . Am. Chem. SOC., 1959, 80, 3261. Kanaoka, Y., Machida, M., Machida, M., and Sekine, T., Biochim. Biophys. Acta, 1973, 317, 563. Drenth, J., Jansonius, J. N., Koekoek, R., and Wolthers, B. G., “The Enzymes,” Third Edition, Volume 3, Academic Press, New York, 1971. Birkett, D. J., Price, N. C., Radda, G. K., and Salmon, A. G., FEBS Lett., 1970, 6, 346. Machida, M., Machida, M. I., and Kanaoka, Y., Chem. Pharm. Bull., 1977, 25, 2739. Paper A41 76 Received February 21st, 1984 Accepted March 15th, 1984

ISSN:0003-2654    

DOI:10.1039/AN9840901003

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST, AUGUST 1984, VOL. 109 1009 3-Arylcoumarins as Fluorescent Indicators Lena A. DeLisser-Matthews and Joel M. Kauffman Department of Chemistry, Philadelphia College of Pharmacy and Science, Philadelphia, PA 79 704, USA A new group of fluorescent compounds, 3-arylcoumarins, were evaluated a s indicators for various types of titrations. Several of the compounds were found to have a wide spectrum of applications, not only a s adsorption indicators for precipitation titrations, but also a s indicators for acid - base titrations. Some were found t o be a s effective as dichlorofluorescein in their adsorption indicator properties in precipitation titrations. Many were also found t o compare favourably with some of the most efficient acid - base indicators such a s phenolphthalein, methyl red and phenol red.Keywords: 3-Arylcoumarins; laser and fluorescent dyes; precipitation; aqueous and non-aqueous acid - base titrations; adsorption indicators In a recent study several fluorescent compounds of the type shown in Fig. 1 were prepared primarily for evaluation as laser dyes. The compounds were found to have peak fluorescence wavelengths close to the maximum sensitivity of the human eye, and excitation peaks of sufficiently long wavelengths so that there was no need for an ultraviolet lamp for excitation. Five members of the group of compounds (Fig. 1, A-E), which exhibited very high fluorescence quantum efficiency and long photochemical half-lives, were chosen for evaluation primarily as adsorption indicators for precipitation titrations. The study was further extended to an evaluation of the dyes as acid - base indicators.Although several indicators are available for precipitation titrations, there is need for additional indicators, because many of these indicators are found to be specific for only certain types of titrations, and require very precise conditions for the attainment of good results.2 The effectiveness of an adsorption indicator depends upon the extent to which the indicator is adsorbed on to the precipitate during the course of the titration. A desirable indicator is not too strongly or too weakly adsorbed. If the indicator is too strongly adsorbed, its adsorption may actually be a primary process in which the indicator displaces the primarily adsorbed ion. This results in a colour change well before the equivalence point in the titration is reached.On the other hand, if the indicator is too weakly adsorbed, it will not be able to displace the secondary adsorbed ion, and the equivalence point will not be revealed. This is one of the disadvantages of an indicator such as eosin when used in halide titrations with silver.2 n A A r = a B A r = E A r = < n Fig. 1. Structures of 3-aryicoumarIns The effectiveness of the dyes as indicators for precipitation titrations was evaluated by comparison with dichlorofluores- cein, which appears to be one of the most widely used of the adsorption indicators. As the compounds are weak bases, it was expected that, when they were used in aqueous acid - base titrations, the colour change at the equivalence point would occur in the acid pH range.They were thus compared with such indicators as methyl orange and methyl red, which are known to change colour within this range.2 Crystal violet was used for comparison in the non-aqueous titrations. Experimental The compounds shown in Fig. 1 were synthesised by conven- tional methods.3 In all experiments except the non-aqueous titrations, the dyes were used as 0.1% mlV solutions in N, N-dimethylformamide; for the non-aqueous titrations they were prepared as 0.1% mlVsolutions in glacial acetic acid. At this concentration, they exhibited a strong green to green - yellow fluorescence. For a typical titration, 5-10 drops of the dye solution were adequate for producing a good colour change. All chemicals and reagents were of analytical-reagent grade.The hydrochloric acid, sodium hydroxide, potassium hydrogen phthalate, perchloric acid and silver nitrate solutions were made up to 0.1 N by standard methods. Sodium bromide, sodium chloride, sodium iodide and potassium thiocyanate solutions used in the precipitation titrations were all made up to between 0.05 and 0.10 N. The sodium salicylate samples used in the non-aqueous titrations consisted of 300-mg tablets manufactured by Eli Lilly and Company, Indianapolis, IN. Methyl orange and methyl red indicators were prepared as 0.1% mlvsolutions in ethanol.4 Crystal violet was made up to 1.0% mlV in glacial acetic acid.4 Dichlorofluorescein was made up to 0.1% mlVin 80% ethanol.5The pH measurements were made on a Fisher Accumet Model 140A pH meter.Results and Discussion In the evaluation of the dyes as indicators for precipitation titrations, two types of experiments were conducted. The first set involved the titration of the halides and of thiocyanate with silver ion. In the second set, the procedure was reversed. In both instances the solution being titrated was adjusted where necessary to the indicated pH by the addition of nitric acid. The results of these experiments are shown in Tables 1 and 2 in which the dyes are compared with dichlorofluorescein rep- resented as DCF. As expected, the effectiveness of the dyes as indicators for precipitation titrations is pH dependent and shows a certain selectivity for the different species being titrated. The1010 ANALYST, AUGUST 1984, VOL.109 ~~~ Table 1. Evaluation of fluorescent compounds as adsorption indicators in precipitation titrations with Ag+ as titrant Equivalence point visibility as a function of pH* Indicator DCF . . Analyte . Cl- Br- I- CNS- . Cl- Br- I- CNS - . Cl- Br- I- CNS- . c1- Br- I- CNS- . c1- Br- I- CNS- . Cl- Br- I- CNS- Observed colour change Yellow to pink Yellow to pink Yellow to pink Yellow to pink Pink to yellow Pink to pale yellow Pink to pale yellow Pink to yellow Bright orange to pale yellow Pink to pale yellow Orange to pale yellow Pink to yellow Pink to pale yellow Pink to pale yellow Bright orange to yellow Pink to yellow Pink to yellow Orange to yellow Pink to yellow Pink to pale yellow Pink to yellow - - - PH4 ++++ ++++ +++ +++ ++++ ++++ +++ +++ + + ++ + + +++ ++++ ++++ ++++ ++++ ++++ +++ +++ + ++ ++ + ++++ ++++ + ++ +++ ++++ ++ +++ ++++ + ++ - A .. . . B . . . . . + +++ - c . . . . . D . . . . . ++ - +++ - - ++ ++ - +t - E . . . ++ + - - + + * The notation + + + + to - indicates a range from excellent to poor. Table 2. Evaluation of fluorescent compounds as adsorption indicators in precipitation titrations with halide salts and CNS- as titrant Equivalence point visibility as a function of pH* Indicator DCF . . Titrant . . c1- Br- I- CNS - . . c1- Br- I- CNS-- . . c1- Br- I- CNS- . . c1- Br- I- CNS - . . Cl- Br- I- CNS- . . Cl- Br- I- CNS- Observed colour change Pink to yellow Pink to yellow Pink to yellow Pink to yellow Yellow to pink Yellow to pink Yellow to pink Yellow to pink Yellow to pink Yellow to pink Yellow to pink Yellow to pink Yellow to pink Yellow to pink Yellow to pink Yellow to pink Yellow to pink Yellow to pink Yellow to pink Yellow to pink PH2 ++ + +++ +++ +++ ++ + ++++ +++ + + ++++ ++++ +++ + +++ ++++ ++++ ++++ +++ +++ ++++ - - PH 4 ++ + +++ ++++ ++++ ++++ ++++ ++++ +++ ++++ +++ + ++ + + - - + + + +++ + + A .. . . ++++ ++++ ++++ ++++ ++++ ++++ +++ +++ +++ +++ +++ +++ ++++ +++ +++ - B . . . . c . . . . D . . . . E . . . . Pale orange to pink Pale orange to pink Pale orange to pink +++ ++++ ++++ - + + * Notation as in Table 1. behaviour of the dyes conforms with the expectation that, because they are weak bases, they would work best in titrations in which the halide salts and thiocyanate are the titrants of silver ion rather than the reverse procedure. In such instances the indicator ion, being of opposite charge to the ion being titrated, is adsorbed on to the precipitate just prior to the equivalence point in the presence of a slight excess of the titrant.Thus the colour change accompanying this adsorption occurs at the precise moment of equivalence, and is neither premature nor delayed.2 The addition of a small amount of dextrin as a “protective colloid” to prevent clumping of the precipitate and to maximise the surface area for adsorption was effective in some instances in enhancing the end-point detection. Overall, the efficiency of the indicators in many instances was as good as, or better than, that of dichloro- fluorescein.ANALYST, AUGUST 1984, VOL. 109 1011 Table 3. Evaluation of fluorescent compounds as acid - base indicators in aqueous media Equivalence pH Observed point transition Indicator colour change visibility* range Methylred .. Redtoyellow ++++ 4.2-6.2 Methyl orange . . Red to yellow ++ 3.1-4.4 fluorescence ++++ 3.7-5.7 fluorescence ++++ 3.8-5.6 fluorescence ++++ 3.4-5.2 A . . . . . . Orangetogreen B . . . . . . Yellowtogreen C . . . . . . Orangetogreen D . . . . . . Yellowtogreen + 3.9-5.2 E . . . . . . Orangetoyellow + 4.2-5.8 * Notation as in‘Table 1. Table 4. Evaluation of fluorescent compounds as acid - base indicators in non-aqueous titrations Colour Equivalence point Indicator change observed visibility* Crystalviolet . . . . Purple togreen ++++ D . . . . . . . . . . Green to yellow + E . . . . . . . . . . Yellow to pink ++++ * Notation as in Table 1.The effectiveness of the dyes as aqueous acid - base indicators was evaluated in titrations of potassium hydrogen phthalate and of hydrochloric acid with standard sodium hydroxide solution. The evaluation criteria were visibility and sharpness of the equivalence points when compared with those produced by methyl orange and methyl red. The results are shown in Table 3. Three of the compounds, namely A, B and C, were seen to be excellent as indicators in aqueous acid - base titrations, comparing favourably with phenolphthalein, methyl red and phenol red, and superior to methyl orange in signalling the equivalence point. In addition to the expected colour change, the strong fluorescence exhibited by these compounds as the equivalence point was reached was an added advantage in enhancing the end-point detection.The pH transition ranges for the various compounds tested were within the preferred limits of 2 pH units. The first set of acid - base titrations in non-aqueous media involved the titration of potassium hydrogen phthalate dis- solved in glacial acetic acid using perchloric acid as titrant. In a second set of experiments, sodium salicylate tablets were titrated in a similar way without preliminary treatment. Of the five dyes, only two, D and E, were found to be useful as indicators in non-aqueous titrations in the glacial acetic - perchloric acid system. Their effectiveness compared with that of crystal violet is shown in Table 4, where compound E appears to be as good an indicator as crystal violet. To summarise, the most significant features of this new group of fluorescent indicators are as follows.1. Versatility as indicators for a broad spectrum of titrations. 2. Usefulness over a reasonably wide pH range, i e . , pH 2-6. This is a distinct advantage over some of the commonly used dyes that are limited to a much smaller pH range. 3. Stability as shown by the absence of decomposition or fading. 4. Very high fluorescence quantum efficiency, which permits the use of ambient light and which requires no room darkening during the titrations. 5. No separate source of ultraviolet light is needed. Financial support from the Chemistry Department, Director, G. D. Chase, is gratefully acknowledged. Helen E. Plotkin and Patricia Davis aided in the preparation of the manuscript. References 1. 2. Fletcher, A. N., Bliss, D. E., and Kauffrnan, J. M., Opt. Commun., 1983, 47, 57. Day, R. A., and Underwood, A. L., “Quantitative Analysis,” Third Edition, Prentice Hall, Englewood Cliffs, NJ, 1974, Chapters 4 and 6. 3. Horning, E. C., Horning, M. G., and Dimmig, D. A . , Org. Synth., 1955, Coll. Vol. 111, 165. 4. “The United States Pharmacopeia,” Twentieth Revision, Mack, Easton, PA, 1980, pp. 1104, 1106 and 1107. 5. Dick, J . G., “Analytical Chemistry,” McGraw-Hill, New York, 1973, Chapter 15. Paper A4168 Received February 13th, 1984 Accepted March 14th, 1984

ISSN:0003-2654    

DOI:10.1039/AN9840901009

出版商:RSC    

年代:1984

数据来源: RSC