ISSN:0003-2654    

DOI:10.1039/AN98409FX021

出版商:RSC    

年代:1984

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98409BX023

出版商:RSC    

年代:1984

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98409BP041

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST. JUNE 14’83 VOI I09 677 Recent Developments in Detection Techniques for Hig h-performance Liquid Chromatography Part I. Spectroscopic and Electrochemical Detectors A Review Peter C. White Metropolitan Police Forensic Science Laboratory 109 Lambeth Road London SE I 7LP UK Summary of Contents Introduction Classification of detectors Requirements of a detector Detector characteristics Noise Sensitivity Response Spectroscopic detectors Ultraviolet Visible wavelength FI uorescence lnfra red Atomic absorption Atomic emission Nuclear magnetic resonance Electron spin resonance E I ect roc hem ica I detect0 rs Permittivity or dielectric constant Conductivity Potentiometric Vol ta m metr ic References Keywords Review; high-performance liquid chromatography; spectroscopic detectors; electrochemical detectors Introduction It has been accepted that high-performance liquid chromato-graphy (HPLC) appeared as a separation technique in 1964-65.The detection systems employed in this analytical method are based on instrumentation designed to measure a particular physical or chemical property of a solute in an eluate that flows from a chromatographic column. As a large number of these properties can be used to detect a solute and because there is no universal detector available numerous detection systems have been devised. This review is an attempt to bring together detector developments that have been reported during the past 7 years, but as a result of the tremendous amount of coverage that has been devoted to this topic it has been necessary to divide it into two parts.In Part I spectroscopic and electrochemical detection methods will be covered and Part I1 will survey other types of HPLC monitoring systems. A brief conclusion to the whole review will appear at the end of Part 11. Following a brief discussion on the classification of detec-tion systems and general requirements and characteristics of a typical HPLC detector this paper reviews the development of monitoring systems to date. In addition to describing the mode of operation limitations and cost of these systems the potential of newly devised techniques has also been examined. More extensive information can also be gained from the many references that have been cited. Classification of Detectors In most textbooks HPLC detectors have been broadly divided into two categories the bulk property detector which measures some bulk physical property of the eluate and the solute property detector which measures a physical/chemical characteristic property of the solute.These have also been referred to as non-selective and selective detectors respec-tively. This type of classification is misleading as the ultraviolet (UV) detector for example which is classified as a solute property detector will produce a background signal if the eluent has a UV absorption. Another classification method that has been used is based on whether the detector measures the concentration or the mass of the eluted solute but only a limited number of detection systems are of the mass-dependent type.Clearly these divisions are not ideal and in this review four simple classifications have been used (i) spectroscopic detec-tors; (ii) electrochemical detectors; (iii) detectors involving no phase change of the eluate; and (iv) detectors involving a phase change of the eluate. Detectors of type (i) and (ii) are studied in Part I and types (iii) and (iv) in Part 11. Requirements of a Typical Detector The function of any detector employed in HPLC is to monitor accurately the concentration or amount of the sample com-ponents eluted from the column and generally the following requirements are necessary (a) capability to detect 1 part of solute or less in 106 parts of eluent as larger concentrations will overload a standard analytical column; (b) no re-mixing of components as they pass through the detector; (c) wide linear dynamic range to ensure that quantitative analysis can be accomplished in a straightforward manner; (d) low drift and noise level so that small amounts of solute can be observed; (e) fast response time to record accurately rapidly eluting peaks 67 8 ANALYST.J U N E 1984 VOL. 109 (f) insensitive to flow-rate changes pulsation and tempera-ture; (g) insensitive to changes in eluent composition so that gradient elution can be performed; and (h) ease of operation and reliability. As can be seen from the detector systems described here, very few of them actually satisfy all of these requirements. More recently the introduction of small particle size packings (4 ptn) and microbore columns these requirements become even more important as smaller amounts of solute are eluted in very low volumes of eluent.1 Detector Characteristics To be able to select a detector for a specific application it is preferable to compare performances on a uniform set of criteria. Scott’ has covered this subject in detail and therefore only the most important parameters are covered below. Noise Any disturbance of a detector output that is unrelated to an eluted solute can be defined as noise and three common forms exist. The first is termed short-term noise and produces a “fuzzy” base line as a result of signal fluctuations occurring at a frequency of greater than 10 cycles min-1. This form of noise can arise from the electronics associated with the detector or recorder.and pump pulsations. The measurement of short-term noise is an important parameter as it is used in defining the detection limit of a system. The second form of noise is referred to as long-term noise and is due to random and low-frequency variations of the output signal. The base line produced is erratic and the signals formed by this type of noise cannot be differentiated from a component peak of similar amplitude. Impurities temperature and pressure fluctuations will give rise to this type of signal. Both the long- and short-term noise should be expressed in terms of the physical property being measured by the detector. The last type of noise is known as drift. This is recognised by a continuing increase or decrease in the detector signal and the effect is produced by changes in temperature or mobile phase composition.The amount of drift is quoted as the variation in the physical property over a fixed period of time, usually 1 h. Sensitivity The minimum amount of solute that can be detected is referred to as sensitivity. The limit of detection (LOD) and noise equivalent of concentration (C,) are expressions com-monly used to define this quantity and both are measured with respect to the short-term noise level. The limit of detection is the amount of solute that will produce a signal (S) equivalent to twice the short-term noise level (N). The noise equivalent concentration differs slightly in that it is the amount of solute that produces a signal equal to the short-term noise level.The units are g ml-1 for concentra-tion detectors and g s-1 for mass-dependent detectors. As the sensitivity will vary for different solutes the solute that is used should always be specified. Response The chromatographer should be aware of two very different response ranges that are quoted for a detector. The dynamic range (R,) is the solute concentration range over which the detector will produce a concentration-dependent output. The limit of detection as described above will be the minimum Concentration. and the maximum is where the detector becomes saturated. Usually the response is expressed in orders of magnitude of solute concentration. The other form of response is the linear dynamic range (Rid) which is the range of concentration over which the detector response is linear.This range is typically one tenth of the dynamic range and is similarly expressed in orders of magnitude of solute concentration. The linear dynamic range is determined by plotting the logarithm of the response of a particular solute against the logarithm of concentration. For a perfectly linear response the gradient of the slope of the graph which is referred to as the response index (Y) would be exactly unity but in practice the value generally lies between 0.98 and 1.02. Great care must be taken when comparing detector characteristics because apart from a recently published ASTM procedure for fixed-wavelength photometric detec-tors,’ there are no universally accepted definitions for these criteria and their units are rarely standardised.This is particularly important to remember if one is considering the purchase of a detector as instrument manufacturers tend to use obscure definitions or units in their specifications in order to promote their instrument. Spectroscopic Detectors Ultraviolet The ultraviolet (UV) detector was one of the earliest HPLC detectors and is still the most popular. Any solute with a UV absorption can be monitored with this system and for a highly absorbing species a detection limit of 1 ng is feasible. In all UV detection systems the solutes are detected by measuring the amount of light that is absorbed as it passes through the flow cell. The concentration of the solute is determined from Beer’s law which states that for monochro-matic light the fraction of radiation absorbed is proportional to the number of absorbing molecules [equation (l)] where A = absorbance Z = incident light intensity I = transmitted light intensity E = molar absorptivity I = path length of flow cell and c = solute concentration.It follows that the sensitivity is proportional to the path length of the flow cell. The design of the flow cell is very important as turbulence, band dispersion and cell volume all affect the over-all performance of the detector.4.5 In order to reduce turbulence, which is responsible for changes in refractive index and hence flow sensitivity three cell designs are commonly used in commercial instruments. They are generally known as the Z pattern,6 the H pattern7 and the tapered cell.8 These cells usually have an optical path length of 10 mm and a volume of 7.5-10 pl.With micro-HPLC the cell volumes have to be reduced to 2-3 pl to ensure that the separations achieved through the column are not degraded in passing through the detector. Most UV detectors use photocells to measure the light intensities. Fixed-wavelength detectors were one of the first type to be developed commercially and are one of the cheapest forms of UV detector currently available. Low-pressure mercury lamps give a range of discrete wavelengths and the 254-nm line is the most predominant. This serves therefore as an excellent source for an HPLC detector because a very large number of organic molecules absorb at this particular wavelength. The range of solutes that can be analysed by a fixed-wavelength detector can be extended with the use of suitable phosphors.Irradiation of these phosphors with light of a wavelength of 254 nm will produce emissions at other fixed wavelengths. More recently with the introduction of zinc and cadmium lamps solutes with low wavelength absorptivities can be analysed because these lamp5 produce a very strong radiation at 214 and 229 nm respectively. Some new fixed-wavelength detectors use deuterium lamps which provide a continuum of radiation rather than a line source. When used in conjunction A = log (ZJZ) = ~ c l (1 ANALYST. JUNE 1984. VOL. 109 679 with good quality filters any solute absorbing in the wavelength range 150-400 nm can be detected. A logical progression in UV detector design was the introduction of variable-wavelength instruments.With these instruments the desired wavelength is obtained by manual operation of a diffraction grating and a deuterium lamp generally provides the source of radiation. A scanning UV spectrophotometer can be modified for use as a variable-wavelength detector by incorporating a flow cell.9 The advantage of a continuously variable-wavelength detector is that the selectivity can be enhanced by choosing the wavelength at the which the solute exhibits maximum absorp-tion. It has recently been shown. however that fixed-wavelength detectors even when operated at a wavelength that does not coincide with the absorption maximum of a solute. will give greater sensitivity than a variable-wavelength detector because they produce less background noise.10 Most fixed- and variable-wavelength detectors have a fixed band width of 2-5 nm. as for too wide a band width Beer's law is not obeyed and too narrow a band width will increase the noise level. The commercial instruments that do provide for a selection of band widths can offer some advantages. For instance minor components can be detected if wider band widths are employed and more precise quantitation can be obtained by selecting the narrowest spectral band. In practice single-beam instruments are satisfactory pro-vided that the eluents used give better than 70% transmission at the selected wavelength. With lower transmission levels or when gradient elution is being used the dual-beam instruments do offer some degree of compensation for the background absorbance.To minimise temperature variations and hence reduce the flow sensitivity of the UV detector a number of manufactur-ers fit a heat exchanger (i.e a length of narrow-bore steel tubing) to the flow cell. One important point to note is that as the heat exchanger is fitted on the inlet side of a flow cell this effectively increases the cell volume and therefore a decrease in detector efficiency will occur. One very useful property of UV detection is that it is a non-destructive technique and therefore a UV detector can be coupled together in series with any other monitoring device. Solute identification and discrimination can be achieved via this dual detection approach by simply calculating the response ratio of the two detectors.For example a fixed-wavelength and variable-wavelength detector have been coupled together to help in the discrimination of barbiturates by comparing the absorbance ratios at several wavelengths. 11 Compounds can also be identified by further analysis of trapped fractions or by stop-flow UV scanning using the same variable-wavelength detector. 12 When compounds are difficult to resolve. UV detectors can still be used for the quantitation of the individual components. l 3 A more recent approach to this problem utilises the small spectral differences that exist between compounds. and employs two selected wavelengths that are sent through the same cell and along the same light path in a time-sharing manner.I4 A commercial instrument based on this technique appeared in 1982 and its design is unique in that the monochromator is servo-driven and the grating is suspended in a magnetic field.Very fast changes in the magnetic field allow monitoring at any two chosen wavelengths. and also by rapid scanning of the LJV -visible wavelength region spectra o f eluted components can be obtained without having to stop the flow. Traditionally UV detectors have been used for the analysis of absorbing species but recently two techniques have been developed that permit the detection of ionic non-absorbing species. As these methods are suitable for either organic and/or inorganic ions a whole new range of important applications can be accomplished with the UV detector. The first o f these methods was reported by Denkert et al.) 15 and was based on the detection of non-absorbing ions by forming ion pairs with a UV-absorbing component dissolved in the eluent.A variety of organic ions were detected via this approach and more recently a number of inorganic ions have been detected. 16 The other method for detecting non-absorbing species was introduced by Small and Miller17 in 1952. and has been named indirect photometric chromatography (IPC). With this tech-nique a small amount of a UV-absorbing compound e.g. a phthalate is dissolved in the eluent and any non-absorbing ion eluted from an ion-exchange column will produce a negative peak owing to the displacement of the TJV absorbing ions from the eluent. Several other groups have now devel-oped applications based on this technique for the detection of a wide variety of organic and inorganic anions and cations.18-20 In the above-mentioned applications of IPC the eluent composition and UV wavelength monitoring regions were chosen so that the ions gave a negative response. Recently, Wheals21 has shown that the discriminative power of the method can be improved by using citric acid as the UV-absorbing compound. This permits monitoring at 220 nm, where some common inorganic anions display sufficient absorption to be detected as positive peaks and those which do not absorb produce negative peaks as illustrated in Fig. 1. The improvements that have taken place in microcomputer technology over the past few years have led to a new breed of rapid-scanning UV spectrometers that permit simultaneous multi-wavelength detection.These detectors are based on optical multi-channel analysers22 that use either a silicon-intensified vidicon tube (SIT) or a linear diode array (LDA) for the rapid capture of UV light. The SIT detector evaluated by McDowell and Pardue23>24 uses a grating polychromator to disperse the eluate transmission spectrum across the surface of several hundred photosensitive diode junctions each of which corresponds to a specific wavelength. Each of these can be monitored by a continuously scanning electron beam.25 The incident radiation falling on these diodes leads to a loss of charge and the recharging current from the electron beam is proportional to the radiation intensity. 0 4 8 12 16 Ti me/mi n Fig.1. Indirect photometric detection o f some UV-absorbing and non-absorbing ions using an eluent containing citric acid. 1 = Phosphate 2 = chloride; 3 = nitrite 4 = nitrate 5 = sulphate; and 6 = iodide. Courtesy of B. B. Wheals. Metropolitan Police Forensic Science Laboratory Londo 680 ANALYST JUNE 1984 VOL. 109 The LDA analysers employ Reticon 256 element diode arrays2630 and operate on the same principle as the SIT detectors. By comparison with the latter type better operat-ing characteristics in the UV region can be achieved with the LDA system and therefore these have been used in the design of HPLC multi-wavelength instruments. A UV - visible spectrophotometer was the first commercial analytical instrument to use this new technology and although not intended for use as an HPLC detector two groups of workers have achieved this by replacing the cuvette with a flow ceIl.31-36 Within a short period of time a less sophisticated (and considerably cheaper) version of this instrument was manufactured specifically as an HPLC detector,37 and the basic components of this detection system are shown in Fig.2. A very important feature is that there are no moving parts in the detector. This optical configuration (reverse geometry optics) is a departure from traditional UV detectors in that instead of passing light of a selected wavelength through the flow cell, the sample is subjected to light of all wavelengths emitted by the source lamp. As this new configuration reduces light losses samples can be monitored at 190 nm.Some analysts have expressed concern over this design as it is feared that samples can either fluoresce or undergo photolytic decom-position although the instrument manufacturers have not noticed these effects with the compounds that they have studied to date. Most companies who manufacture HPLC monitoring systems have started to market multi-wavelength detectors and without exception use an LDA analyser.38 With some of these instruments up to eight wavelengths can be monitored simultaneously. The wealth of information gained in a single chromatographic run from these detectors can be used to obtain absorbance and time plots together with UV spectra of the separated components. The information can be stored either in a computer and then manipulated as required or presented in real-time format by algorithms devised by Klatt.39 Deconvolution of unresolved components can also be carried out very easily with these detectors and this particular topic has been reviewed by Fell.25 One of the major advantages of multi-wavelength monitor-ing is that absorbance ratioing can be performed and therefore solute identification and discrimination can be achieved.A group of workers who were particularly interested in this technique developed their own relatively inexpensive multi-wavelength detector. As illustrated in Fig. 3 this detection system does not use a photodiode array but exploits a very novel rotating filter disc method for monitoring up to four wavelengths simultaneously.4()~41 Typical noise levels of 10-4 absorbance units and detection limits of 31 ng of injected material have been recorded with this instrument and further, -u Grating Fig.2. Optical layout of an LDA multi-wavelength detector. Reprinted with permission of Hewlett-Packard in a recent evaluation study excellent long-term reproducibil-ity of the absorbance ratios was reported.42 Another feature of this instrument is that absorbance ratio plots can be obtained and used to check peak homogeneity. To summarise UV detection fixed- and variable-wavelength detectors will without doubt continue to be the most popular of all HPLC monitoring devices. Increased usage of these instruments can also be expected owing to the development of the indirect photometric detection technique for the analysis of non-absorbing ionic species.As some useful multi-wavelength techniques can be performed with the rotating filter disc detector it will be interesting to see if any commercial development of this instrument occurs in the future especially as its costs and performance are similar to those of single- and variable-wavelength detectors. Analysts are advised to look very closely at the LDA-based rnulti-wavelength detection systems because they can offer many factilities but their sensitivity is poorer than that of a single-wavelength detector. Further as the cost of these detectors is at least five times that of a good variable-wavelength detector, it is difficult to envisage their usage in laboratories where only routine analyses are being performed. Visible Wavelength Detectors that operate solely in the visible wavelength range of 400-800 nm are available but generally manufacturers tend to produce instruments that cover the UV and visible range.The visible region on these detectors can be used by simply switching to a tungsten source. This detector is suitable only for compounds that produce a visible wavelength spectrum. Greater exploitation of this detector can be achieved by post-column derivatisation to produce coloured species and the ninhydrin reaction for the detection of amines and amino acids is a classic e~ample."?~4 With developments in both pre- and post-column derivatisa-tion techniques the visible wavelength detector can now be used to offer good selectivity with sensitivities approaching 50 ng of injected material.Some interesting applications that have been reported are given in Table 1. Fluorescence Fluorescence is a luminescence phenomenon that occurs when a compound absorbs radiation and then emits it at a longer wavelength. This form of radiation can be generated in all regions of the electromagnetic spectrum but only that produced by the absorption of UV or visible light is of interest in HPLC detection. Flow fluorimeters have been available since the earliest days of commercial liquid chromatography, and besides offering excellent selectivity they are amongst the most sensitive detectors that are available for HPLC. Before discussing some of the recent advantages of fluorescence detection it is worth considering some of the more important principles upon which fluorescence is based and how they can be exploited to improve the design of flow fluorimeters.For Deuterium lamp 0 Lens 17 Flow cell motor K Filter Fig. 3. Optical layout of the rotating disc multi-wavelength detector. Courtesy of T. C. Catterick Metropolitan Police Forensic Science Laboratory Londo ANALE'S'I'. JLllVE 1984 VOL. I09 68 1 Table 1. Some applications o f the \. isible wavelength detector when used in conjunction with pre-and post-column derivatisation Cotnpounda analysed Creatine . . . . . Pol y am ine s-spe rnii n e . putrescine spermidene Phospholipids . . . 0 x 0 acids of phosphorus Lanthanides . . . . . Alkalineearths . . ~ Cannabinoids . . . Cations-Mn Ni. Co, Cu.ZnandPb . . . Complexity reagent Reference Picric acid 35 N i n h y d r 111 Mol)bdenum dfter 47 46 co n v e r t i n & organ I c phosphorus to in o r gd n I c form Molybdenum 18 Xylenol ormge 39 Sulphosallcyllc d c ~ d - 51 Arvxazo I11 5 0 chlorophosphonnzo I11 Fdst Blue B 52 ~-(2-Pyridylazo)resorcinol 53 monosodium salt Fluorophor concentration -+ Fig.4. fluorescence intensity Graph illustrating the effect of fluophor concentration on readers who are interested in a more detailed account of the theory of fluorescence the discussions by Guilbauit'j and Udenfriend55 are highly recommended. Only molecules that are highly conjugated will display fluorescence and with compounds of this type excellent selectivity can be obtained. Fluorescence emission will always occur at a longer wavelength than that of the exciting radiation wavelength and hence it is possible to irradiate a fluorophor (a solute that will produce fluorescence) that absorbs strongly in the UV region and observe the fluorescence in the visible region.As the irradiating light can easily be removed the measurement of fluorescence is obtained in theory against a zero background level. This is different to UV monitors, which measure small differences between the incident and transmitted light intensities and therefore fluorescence detec-tors are two to three orders of magnitude more sensitive. With the newer detectors that are available a good fluorophor will give a noise equivalent concentration (C,) as low as 10-12 g mi-1 Fluorescence will emerge randomly from an irradiated sample and therefore the emitted light can be viewed in any direction.In most fluorimeters the emitted light is detected at right angles to the exciting beam but a 2n steradian configuration can also be used provided that the exciting radiation is filtered out. The most basic type of fluorimeter produced uses filters to select the desired excitation and emission wavelengths more versa t i 1 e flu o ri me t ri c d e t e c t or s are now fitted with a monochromator to select the emission wavelength. The right-angled optical configuration offers further advantages as the absorption of the radiation by a solute can be measured independently of the fluorescence, thus permitting dual simultaneous fluorescence - absorption detection.56 A number of manufacturers offer this facility with detectors that are fitted with monochromators to select the emission and excitation wavelengths.Greater selectivity can also be achieved with these instruments because fluorophors possess different absorption and fluorescence spectra. The multi-channel systems that have recently been devel-oped for UV detection can also be used in fluorimeters. Jadamec et al.57 characterised petroleum fractions with one of these systems and more recently details of instrumentation58 and applieationsj9 of this form of detector have been published. Compounds that fluoresce will emit only a fraction of the absorbed light as fluorescence and this is referred to as the quantum efficiency of the solute (&). Although the value of gf can be as high as unity it is generally in the range 0.1-0.9.Fluorescence intensity (F) is related to the quantum efficiency as shown in the equation where I, = incident light intensity and I = transmitted light intensity and by substituting the exponential form of Beer's F = (I" - It)& . . . . * . (2) law in this equation a relationship between the fluorescence intensity and solute concentration (c) can be obtained [equation ( 3 ) ] . where K = efficiency of collection of fluorescence E = molar absorptivity and 1 = path length of the flow cell. A plot of fluorescence intensity L J ~ Y S U S concentration is shown in Fig. 4 and it can be seen that at low concentrations the curve is almost ,linear and without losing too much precision equation (3) can be approximated to equation (4).F = @fKZ,(I - 10-Eq . . . . ( 3 ) F = 2 . 3 0 3 @ f K f ~ ~ l . . . . . . (4) As c)~ K I, and E can be held constant it can be seen that the fluorescence intensity is proportional to the concentration, and therefore the detector is suitable for quantitative analy-tical work. Under these conditions the response index (Y) of the detector is unity. but will fall to 0.98 at 0.02 absorbance units (a.u.). The main reason for non-linearity is self-absorption by the solute but the linear range (RL) of a fluorimeter is typically five orders of magnitude. The choice of solvents used in the separation process is very important when using fluorescence detection. Fluorescence is very sensitive to certain deactivating species or quenchers.Eluents containing highly polar solvents buffers or halide ions should all be avoided as they promote fluorescence quench-ing. Further as molecular collisions of the solute also quench fluorescence high temperatures should not be used and eluents with high viscosities are preferred in order to reduce such collisions. The flow cells designed for fluorimeters tend to differ from those employed in a UV detector because only low levels of light are emitted. Although the best commercial detector systems offer cell volumes of about 5 1.11 these cells produce considerable background noise levels because of the stray light that is formed through scattering at the liquid - cell interface. To improve the fluorescence collection efficiency of detectors, parabolic mirrors have been used to collect the emitted light and focus it on the photomultiplier but this also increases the noise level.Recent developments in fluorimeter design have been concerned with the production of flow cells to reduce noise levels and the use of high-power sources. It was shown earlier that the fluorescence intensity is proportional to the incident light intensity [equation (411 and in fact the conventional sources that have been used in commercial detectors produce relatively weak radiation intensities. High-intensity pulsed xenon sources with fibre optics can improve the situation but lasers appear to be most favourable. Lasers have opened up new frontiers in spectroscopy and their use in chromatographic detection systems has generated considerable interest particularly in fluorescence detection schemes.Excellent papers covering the principles of lasers6 682 ANALYST. JUNE 1984 VOL. 109 and their applications in spectroscopy61 and fluorometric detection in HPLC62,63 have appeared in the literature. Lasers produce at least 103 times more power than a conventional source and commercial lasers that are currently available can provide about 10 W of power in either the visible region (argon ion laser) or the UV region (Kr - F laser). Laser radiation is highly monochromatic and because the beam diameters are small reductions in stray light can be achieved. Some types of lasers are able to produce this intense radiation in pulses and as these pulses are short in comparison with fluorescence lifetimes greater sensitivity can be obtained owing to a reduction in the background noise level.Conventional fluorimetric detectors are designed to operate with excitation band widths of about 10 nm and undesirable “bands” are produced owing to Rayleigh scattering (due to the eluent and optical components) and Raman scattering (due to the eluent). The use of monochromatic laser radiation permits a reduction of these band widths to about 1 nm and therefore larger spectral “windows” are available for fluorescence detection. Rayleigh and Raman scattering generally produce polarised light and although conventional detectors in con-junction with polarisers can produce plane polarised light and therefore reject these bands half of the source intensity will be lost.However laser light is inherently plane polarised because of the optical design and therefore by suitable placement of the detector these bands can be removed. In fluorescence detection systems Raman and Rayleigh scattering and fluorescence from optical components and con-taminants increase background noise levels. The sensitivity of any detection system is limited by this noise level and for this reason the flow cell design is critical. With the development of laser systems alternatives to conventional flow cells have been examined as smaller cell volumes can be used with the highly collimated laser radiation. An interesting concept for the elimination of stray light has been based on suspending the eluate in a droplet from a capillary tube attached to the column exit.64 Distortions of the droplet shape can cause problems with the focusing of the emitted radiation.These distortions occur if air is present or if the viscosities and surface tensions of the eluent change e.g., in gradient elution. Further if substantial laser absorption takes place a phenomenon called “thermal lensing” can occur in the droplet which will alter the refractive index of the medium and cause further focusing problems. This type of flow cell cannot be used in series with any other detector unless it is last in the sequence. Recently detection limits of 0.03 to 5.0 ng ml-1 have been reported for a variety of polynuclear aromatic hydrocarbons with this type of flow In order to overcome air bubble problems a capillary tube has been used in conjunction with fibre optics.66 In this arrangement the eluate flows up the capillary to a fibre optic and the flow cell volume can be varied by altering the length of the capillary.It was reported that with a cell volume of 20 1.11 detection limits for adriamycin and daunorubicin of 10 and 15 pg respectively can be achieved at a signal to noise ratio of 3. The same authors have also shown the value of this detection system for studying coal liquids.67 The sheath flow principle has also been investigated as a possibility for a flow cell.68 The chromatographic effluent is injected into the centre of an ensheathing solvent stream. but does not mix with it because of laminar flow conditions. Although the liquids are retained by quartz windows the sheath fluid forms a “wall-less” flow cell and by varying the relative flow-rates of the sheath and sample liquids a cell volume of 6-150 nl can be obtained.This system was tested with porphyrins and with mesoporphyrin IX dimethyl ester, and excellent linearity over the range of 53 pg-53 ng of injected material was obtained. Calculations for a 53-pg injection showed that with a cell volume of 53 nl there were 2.5 X 107 molecules in the cuvette and only 5 x 105 molecules cell .65 in the laser beam thus indicating the potential sensitivity of such a system. A very interesting process known as two-photon excited fluorescence can be generated with a laser because of its high power. Increased selectivity can be obtained if used as an HPLC detection system and this has been demonstrated in the separation of oxadiazoles@ and studies on refined coal liquid mixtures .67 Lasers have improved sensitivity and selectivity in fluor-escence detection but as with any system there are some disadvantages.Visible lasers (e.g. argon ion) are relatively inexpensive but can only produce plasma lines between 458 and 515 nm which restricts its applications. UV lasers (e.g., Kr - F) although available are more expensive but only produce a narrow operating wavelength range. To be of any great value in fluorescence detection inexpensive tunable iasers are required so that a large wavelength range can be covered. Hopefully with the steady advancement in laser technology suitable lasers should appear within this decade.The discussions so far have concentrated on detectors that promote fluorescence by the excitation of fluorophors with electromagnetic radiation. Fluorescence can in fact be gener-ated by any system that can provide sufficient excitation energy. For instance certain chemical reactions will produce fluorescence (chemiluminescence) . Recently there has been considerable interest shown in using chemiluminescence detection systems with HPLC because the chemical reactions offer a high degree of selectivity and the detector need consist only of a photomultiplier. As this type of system does not require an external light source no stray radiation will reach the photomultiplier as in a conventional fluorimeter and therefore greater sensitivity is feasible. The production of chemiluminescence takes place after the components have been eluted from the column and this is the main area where problems have been encountered with these systems.To obtain luminescence several chemicals have to be mixed together and careful control of the pumping rate of reactants. and their mixing is essential to the success of this approach. As with fluorescence chemiluminescence can be quenched and aqueous eluents buffer ions and pH can all affect the intensity and half-life of chemiluminescence. With almost no exceptions chemiluminescence involves an oxidation process and many systems exploiting this phenom-enon have been described in several reviews.70-7s Several applications of chemiluminescence detection in HPLC have been reported in the literature over the last 4 years and these are summarised in Table 2.With some of these applications detection limits in the low picogram range have been quoted and Japanese workers have shown that in comparison with conventional fluorescence the sensitivity of the oxalate - peroxide chemiluminescence system was about 20 times higher for fluorescarnine-labelled catechol-amines and the detection limit was about 25 fmo1.82 Without doubt this method of producing fluorescence has considerable potential for the future. Table 2. Some applications of chemiluminescence detection in HPLC Compounds detected Dansyl amino acids . . Creatine kinase enzymes Ascorbicacid . . . . Polyaromatics nitrogen heterocycles nitrogen -sulphur heterocycles, hydrazines azides and sulphur compounds .. Chlorophyll . . . . Polycyclic aromatics . . Chemiluminescence system . Oxalate - peroxide . Firefly luciferase . Lucigenin . Oxygen - ozone . Sodium hypochlorite -peroxide . Oxalate - peroxide Reference 76 77 78 79 80 8 ANALYST JUNE 1984. VOL. 109 683 One commercial chemiluminescence detection system the thermal energy analyser (TEA analyser) is available for the monitoring of HPLC eluates. This detector was originally developed for gas chromatography and is based on the nitric oxide - ozone chemiluminescence reaction. This system has subsequently been used as an HPLC monitor but its applications are limited mainly because the detector output is unstable if large amounts of aqueous eluents inorganic buffers or ion-pair reagents are used.In order to detect solutes the column eluate is catalytically decomposed and then mixed with ozone. Only those compounds which contain NO groups will produce chemiluminescence and therefore the detector can offer excellent selectivity. Non-volatile N-nitroso com-pounds,s*-85 trinitroglycerine and its metabolites in blood86 and vasodilators and metabolites87 have been monitored with this detector. Recently because a micro-HPLC system was employed ionic nitrosamines in a 70 + 30 methanol - water eluent containing heptanesulphonic acid were detected.88 Compounds that break down to give NO2 under normal operating conditions can be detected if the catalytic process is modified to generate NO. For example low nanogram levels of some explosives have been identified with this modified catalytic process.89 The TEA analyser is relatively expensive (ca.E l 5 000) but, provided that some of the problems of linking it up to an HPLC system can be solved then this detector has consider-able potential in certain areas because of its excellent specificity. The natural decay of a radionuclide can also be used as a source of excitation energy for the production of fluorescence. In 1979 some preliminary experiments based on P-induced fluorescence (BIF) using a 63Ni source were reported.90 The results were encouraging and in the following year the same authors reported on a prototype detector for use under normal-phase HPLC conditions.91 Mainly as a result of loss of radioactivity from the 63Ni source into the eluent and the cost of obtaining adequately protected high-activity small surface area 63Ni sources,~~7Pm was used as the radionuclide in this prototype detector.Although 147Pm has a half-life of 2.6 years which is short for a commercial instrument a greater fluorescent photon output per Pm @-decay should result because it has a higher 13-decay energy (225 keV) than 6”Ni (67 keV). The 147Pm that was used in this detector was protected by silver foil and the surface area was only 1 mm2. This resulted in the production of an easily constructed variable volume (1.5-10 plj flow cell. For the normal-phase separations hexane or hexane -toluene was used as the eluting solvent but to overcome losses in detector response deoxygenation by purging with argon was essential.It was found that the fluorescence could be enhanced if toluene was present in the eluent. This did produce a large increase in the background signal but when the short-wavelength eluent fluorescence was removed with a 370-nm cut-off filter the signal to background noise level was improved by more than two orders of magnitude. With the compounds studied detection limits of 0.1-1 ng were achieved, thus making it comparable to a conventional fluorimeter. In extending the applications of fluorescence detection even further a very large number of post- and pre-column reaction and derivatisation methods have been developed to induce fluorescence in a non-fluorescent molecule. Many of these procedures are based on functional group reactions and therefore offer increased specificity.This subject is outside the context of this review but it is important to realise that even with the fluorimeters that are currently available a large number of analytical problems can be solved by incorporating a small amount of practical chemistry into the procedure. For those interested in derivatisation methods the publication by Lawrence and Freig2 is highly recommended. Photochemical derivatisation by irradiation with a high flux of UV light was not covered in the book cited,g2 but is a novel method of converting a solute in the eluate into a fluorescent by-product. A design for a photochemical reaction detector that consisted of a water-cooled unit containing a mercury lamp and a quartz capillary reaction coil was patented in 1978.93 In a subsequent paper94 the fluorescent photoproduct of cannibinol was detected with a sensitivity of less than 1 ng of injected material.A reaction time of less than 5 s was required and negligible effects on resolution were reported. More recently this photochemical reactor was used by another group of workers to determine diethylstilbestrol in biological sam-ples and detection limits of 1 ng ml-1 were rep0rted.~5 Other designs have been proposed that incorporate xenon and xenon - mercury sources with different types of air cooling,9698 PTFE reaction coils99 and solvent segmentation. 100 Clearly, photochemical reaction detectors can be used to good advantage in HPLC and one interesting application uses the idea of identifying ergot alkaloids by observing the decrease in their natural fluorescence after samples have been irrad-iated.101 From this review of fluorescence detection techniques it can be appreciated that considerable benefits can be gained in both sensitivity and selectivity provided that the solute can be made to fluoresce. Inexpensive conventional fluorimeters are suitable for many applications and with the development of new chemical derivatisation reagents and reaction methods it should be possible to extend the scope of fluorimetric analysis. Chemiluminescence in principle can offer excellent specificity and sophisticated instrumentation for its detection is not required. If some of the problems that were previously mentioned can be overcome then this could become a very useful technique.Laser sources are still very much in their infancy but they do appear to offer a further increase in sensitivity over conventional sources. If inexpensive tunable lasers can be produced in the future then a new generation of fluorimeters could appear on the market. Infrared Most compounds can absorb energy in the infrared (IR) region of the electromagnetic spectrum and this phenomenon can be utilised for HPLC detection. In comparison with absorption methods that have been previously described IR absorption processes are considerably weaker and microgram amounts of a sample are required to obtain acceptable IR spectra with the conventional dispersive detector. Apart from this relatively poor sensitivity many solvents absorb strongly in the IR region and this method of detection was considered to be unsuitable for monitoring solutes in HPLC eluents until about 10 years ago.This change was brought about mainly as a result of the development of size exclusion chromatography (SEC) packing materials which permitted polymers and other high relative molecular mass compounds to be separated. One other important factor was that these chromatographic separations could be achieved with eluents that do not absorb strongly in the IR region as the eluent serves only to dissolve the sample and plays no part in the separation process. Following these developments commercial detectors based on dispersion IR spectrophotometers were produced for monitoring HPLC eluates. These instruments are fitted with low-volume flow cells that have sodium or calcium fluoride windows and allow monitoring in the IR wavelength range of 2.5-14.5 pm.Interference filters are used to select the desired wavelength and by tuning the IR detector it can act as a universal detector or as a highly selective detector. For example by monitoring the C-H stretch region (ca. 3.5 pmj the first condition is satisfied whereas operation at 5.8 pm can be used to detect compounds that contain the C=O functional group. These detectors can be used to scan a wavelength range but as they are not high-resolution instruments they are no 683 ANALYS'I. JUXE 1984. L ' O L 1041 particularly suitable for obtaining a complete IR spectrum of a compound as it is eluted. Most applications employing the dispersive IR detector have been associated with the analysis of polymers but Table 3 shows that there have also been other application5.In comparison with the refractive index detector which offers similar limits of detection these IR systems give a constant detector response for any class of compound c . g . , alkanes and therefore have the advantage of being true mass detectors. In one particular application an IR instrument was used as a universal detector to monitor organic carbon by measuring the C=O absorption after catalytic oxidation. lo4 This method is of course restricted to eluents that do not contain organic solvents. Many of the limitations associated with dispersive instru-ments have been overcome with the development of Fourier transform infrared (FT-IR) techniques.110 Following the initial studies carried out by Vidrine and Mattson several groups of workers have subsequently monitored HPLC eluates using this relatively new technique.Amongst its advantages are that sub-microgram amounts can be detected, several IR bands can be monitored simultaneously and complete spectra corresponding to each point on the chro-matogram can be obtained."' FT-IR monitoring of eluates has been accomplished in two ways firstly by direct detection via a flow cell and secondly by detection after solvent removal; the direct method offers advantages of simplicity of operation and quantitation and has a universal application. IR absorptions of a compound are generally measured in transmission cells but attenuated total reflection (ATR) of a thin film over a reflective surface can be used.The latter technique is subject to changes in refractive index and therefore pressure variations. IR band-shape distortions are also caused by refractive index changes. It is interesting to note that in a comprehensive review covering flow cell detection because of difficulties in removing solvent background spectra solvent programming cannot be used but flow programming can be used as the FT-IR detector is insensitive to changes in flow-rate. Direct detection of components separated on microbore columns has also been studied.*l"l15 One advantage of this system is that it allows solvents with large IR windows to be used that were considered previously too expensive for use with conventional columns.FT-TR analysis of compounds after the removal of the eluent has been investigated chiefly where the process is other than SEC. Solvent removal techniques do offer the theoretical possibility of obtaining 1R spectra in regions of complete solvent opacity although subtractive techniques to remove solvent bands from FT-IR spectra have been studied.I'h The technique of transferring the eluate on to a heated metal ribbon and measuring the reflection - absorption spectra has been attempted but very poor results were recorded because the area over which the solute was deposited was much larger than the beam diameter. Evaporation of eluates in light pipes suffers from similar problems of uneven solute deposition on the surface.The only solvent removal system that has produced an acceptable result involves dropping the eluate on Table 3. Applications of HPLC with dispersive IR detection techniques Compound detected Reference Silicones . . . . . . . . . . 102 Poly(styrene - rert-butyl methacrylate) 104 Polystyrene - acrylonitrile . . . . . . 105 Coal-derivedproducts . . . . . . 106 Triglycerides . . . . . . . . . . 107 Organometallics . . . . . . . . 108 Poly(styrene -vinyl stearate) . . . . 103 a rotating carousel that contains cups of powdered KBr. After evaporation of the solvent diffuse reflectance infrared Fourier transform (DRIFT) spectroscopy was used to detect the ~ o l u t e . ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ A level of 100 ng of indophenol blue was detected with this system.but it would not appear to be suitable for routine analyses because of the complexity of its operation. A similar but simpler system whereby the eluate was evaporated on to a KHr plate has also been reported,l19 but the results are not very impressive. In summary IR monitors are useful for the detection of solutes after separation by SEC particularly as good selectiv-ity can be gained by using a specific wavelength. The simpler and certainly less expensive dispersive instruments are satis-factory for most applications despite their relatively poor sensitivity. Although the FT-IR spectrometer can offer an improvement in sensitivity record spectra and monitor at several wavelengths at a current cost of around f20000 it is difficult to justify the purchase of such an instrument for the purpose of monitoring HPLC eluates.Atomic Absorption Atoms are just as capable as molecules of absorbing energy, and this absorption and its quantitative correlation with the concentration of metal ions originally present in a sample solution serve as the basis of analytical atomic-absorption spectroscopy (AAS). The use of AA spectrophotometers as detectors is well established and these can offer excellent sensitivity and selectivity for a large number of elements. HPLC can be used to separate a wide variety of organometal-lic species and hence the use of AAS for monitoring HPLC eluates has generated considerable interest especially in environmental and biological studies. Standard commercially available AA instrumentation has been used for these studies and no specialised equipment is required for monitoring HPLC eluates.Textbooks covering all aspects of atomic-absorption theory and instrumentation are available12(L1** and therefore only points pertinent to the coupling up of a spectrophotometer to an HPLC system will be covered in this review. Atomic absorption obeys the same general rules as molecular absorp-tion but for the former process to take place free atoms of the element in the gaseous phase must be produced and so sample preparation requires volatilisation followed by dis-sociation of the molecules. The conversion of metal ions in a liquid sample to their atomic form in the vapour phase is usually accomplished by heat energy and both flame and electrothermal methods have been used in HPLC - AAS systems.Flame systems will be considered initially but many of the points made are also relevant to the electrothermal systems. In classical AAS the sample solution is aspirated into the flame via a pneumatic nebuliser but for most nebulisers the uptake rate of 2-10 ml min-l is greater than the HPLC flow-rate which is typically 0.5-2.0 ml min-1. After the first work carried out in 1973 by Manahan and Jones,l23 several groups of workers have tried to optirnise these conditions. Jones et af.124 studied the effects of column flow-rate on nebulisation efficiency and found that for several transition metals a decrease in the flow-rate increased the determined AAS efficiency. Under these conditions the nebuliser was being starved of liquid L e .the flow-rate was less than the nebuliser rate. Starving a burner results in a reduced pressure region post-column which leads to bubble formation. Attempts to adjust the flow-rate to match the aspiration rate have been carried out by Yoza and Ohashi.125 but balancing the flow-rates was found to be very difficult. Their solution to this problem was to tee in a make-up flow of eluent but this condition is undesirable for trace analyses as the sample is diluted even further. In a more recent study,J26 nebuliser design was shown to be an important parameter and for som 685 commercially available spectrophotometers a flooded burner (i.e. flow-rate greater than the aspiration rate) improved the signal to noise ratio. Under this condition a back-pressure is produced and it was reasoned that the improvements were due to the formation of small droplets which result in a better transport rate and hence a greater proportion of the analyte is available in the atomic state.The atomic-absorption injection method127 was developed for the analysis of small samples and this technique has been used in HPLC - AAS to overcome some of the above problems. As the eluate from the HPLC column collects in a drop former the droplet produced falls into a PTFE funnel and is then sucked into the flame. It has been shown127 that one drop (ca. 100 pl) is sufficient to permit an AAS instrument and a recorder to reach a steady-state signal and hence any LC flow-rate greater than the nebuliser flow uptake rate can be used.With HPLC - flame AAS systems there is no restriction on solvents and in some instances improved sensitivity can be obtained by changing the organic part of the solvent. Gradient elution can also be used and therefore a wide range of samples can be analysed by this method. The sensitivity of the technique is dependent on the species that is being deter-mined but is generally in the nanogram region. The applica-tions listed in Table 4 illustrate the selectivity of this system and cover the analysis of inorganics organometallics and compounds that can be determined by metal labelling in a variety of matrices. Electrothermal heating methods have been used in HPLC -AAS studies to provide gains in sensitivity for certain elements. This increase in sensitivity arises from the ability to select and control the atomisation temperature more pre-cisely.and the graphite furnace has been used in all electrothermal AAS work with HPLC so far reported. The major problem with the graphite furnace AAS (GFAAS) detection system is that it cannot be used as an on-line monitor. Unlike flame methods several time-consuming stages are required to produce free atoms of an element in a furnace and hence samples must be collected and stored prior to analysis. 132-113 The GFAAS system also suffers Table 4. Applications of HPLC - flame AAS Species Application Reference Silicon . . . . . . SiliconlevelsinHPLCeluates 128 Tin . . . . . . Organometallics 130 Lead . . . . . . Tetralkylleadingasoline 131 Copper . . . . . . Detertninationofhistidine 120 from a number of other problems.For example sample cuvettes deteriorate with use ‘35 difficulties exist in obtaining the same absorbance from various species of the same metall36 and more interference problems are observed than with flame methods. 137 Despite these problems a variety of metallic species in HPLC eluates have been detected including phosphorus in lubricating oils,IJs arsenic compound~,1”~139 tin in water,’“) tin and silicon in polymers and silicates141 and lead in organolead substances.13bh’37 Trace analysis of metals in environmental and biological samples is becoming an important topic and HPLC - AAS can provide valuable analytical data. For the reasons given flame methods are preferable despite the fact that electrothermal techniques may offer increased sensitivity.Finally one import-ant point that should not be overlooked is that if AAS is used to monitor trace metals a number of HPLC solvents do contain appreciable levels of metallic species. Atomic Emission Ever since the work of Bunsen and Kirchoff (1860) it has been known that many elements when subjected to suitable excitation emit radiation of characteristic wavelengths. For a number of years atomic-emission (AE) spectrophotometers have been used to measure these emissions. In comparison with AAS AES can be operated as a simultaneous multi-element technique which lends itself to on-line monitoring of HPLC eluates. For atomic emission to occur the excitation source tem-perature must be substantially higher than that required for the same element in AAS analysis because as large a fraction as possible of the vapourised atoms must be energetically excited rather than simply dissociated.Flames and plasmas are two sources that are commonly used in spectrophotomet-ers and both types have been used in the monitoring of HPLC eluates. The detection limits for a number of elements by a variety of AAS and AES methods are given in Table 5 . It can be seen that plasma AES offers excellent sensitivity. The burners used for static sampling in AES analyses are of the type known as “total-consumption’’ burners [Fig. 5 ( u ) ] . With the total-consumption burner the sample is aspirated into the flame by either the fuel or the oxidant but the gases are not pre-mixed. Julin et ul.142 found that the burners could not handle satisfactorily the volume flow-rate used in HPLC, and produced a modified burner [Fig. 5 ( b ) ] . By changing the diameter of the aspiration needle eluate flow-rates of 0.8-5.4 ml min-I could be accommodated. With a pre-mixed gas supply phosphorus and su!phur could be detected at 0.02 and Table 5. Detection limits for some elements using AAS and AES Detection Atomic absorption limiti Clgml ‘ Flame GFAAS 10-7 Be Cd. Cr Fe, Mg Mn Zn 10-6 Al Cd Co c‘u Na. Pb 10-5 Ba Ni V lop3 Mg Si Ti 10 Ag,Al,Be,Ca. Cd Co Cr Cu, Fe Mn. Na Ni, Zn 1 0 ~ ’ Ba Pb. Sn Sr P Atomic emission Flame Ca Na Ag Ba. Cr. Mg, Mn Sr Al Co Cu Fe, Ni Pb T1 B. Sn Plasma Ca Sr B Ba. Be Cd, Co. Cu Fe Mg, Mn.Na Ni Pb Sn. Th Ag Al Cr Hg, P Si TI 10 Si,Tl,V >1 B Th Hg Si Zn 1 FIE! V C 686 ANALYST. JUNE 1984. VOL. 109 Plasma box (b) Plasma box 1 r I q F - Fuel Sample solution t (b) m- Light-tight vent Finned heat exchanger Burner chimney ’ t \yDriin for excess sample Column Nitrogen effh~ent Nebuliser section Fig. 5. Atomic-emission burners. ( u ) Total consumption burner; ( b ) modified burner for monitoring HPLC eluates. Reprinted from reference 142 with permission of Elsevier Science Publishers 0.2 pg ml-1 respectively. With this system organic components in the eluent were found to affect the detector response. For example on changing from water to water plus 1% methanol the phosphorus response dropped to SO% and if 100% methanol was employed no response was detected.Acetone and acetone - water eluents were acceptable. This system was also susceptible to interferences from other elements. A total-consumption burner configuration used to monitor phosphorus species separated on a microcapillary column has been reported. 143 With aqueous methanol solutions (&SO%), ethanol and acetone no decrease in the emission was observed. The detector response was linear up to 100 ng and the detection limit was 2 ng. A reduction in interference levels by organic and inorganic matrix components has been observed by using an “inverted flame” burner.133.145 With this type of burner the gas inlet configuration is reversed i.e air is used to nebulise the sample in a hydrogen atmosphere but the system was more suitable for reversed-phase HPLC.Detection levels of 1 ng for ionic phosphorus and 50 ng for non-ionic phosphorus were quoted. A dual hydrogen flame total-consumption burner as pro-posed by Gilbert136 has been developed as a selective halide detector.’37.138 Any organic halide (RX) present in the eluate is combusted in the hydrogen - air flame and in the hydrogen atmosphere after passing over a platinum net the halide is converted into HX. This species passes over a net of indium and the InX vapour produced is burnt in an upper cool hydrogen flame. The indium(II1) halide is then reduced to the excited indium(1) halide which gives a very strong emission. Total column flow-rates of up to 1.5 ml min-1 could be tolerated and a detection limit of 300 pg s-1 for chloroform in water was obtained with this detector.Alternative chemical reactions and the presence of other elements can produce severe interference problems when using low-temperature flames in AES studies. More recently high-temperature inert gas plasma (non-flame) methods have been used and interference problems have been reduced. thereby increasing the sensitivity of AES (see Table 5 for comparison of detection levels of flame versus plasma). Both direct current (d.c.)139 and inductively coupled plasma (ICP) chamber Fig. 6. HPLC - ICP interface configurations. (a) Internal spray chamber; ( b ) external spray chamber. Reprinted from reference 152 with permission of the American Chemical Society Table 6. Some recent applications of HPLC - ICP Application Reference Copper in chelates .. . . . . . . . . . . . . 153 Metals in organometallics . . . . . . . . . . . . 154 Phosphorus in phosphates . . . . . . . . . . . . 155 Arsenic in biological samples . . . . . . . . . . 156 Protein analysis by determination of carbon, phosphorus and several metals . . . . . . . . 157 Nucleotides by determination of phosphorus . . . . 158 Arsenic and selenium . . . . . . . . . . . . 159 Amino acid analysis by determination of carbon and sulphur . . . . . . . . . . . . . . . . 160 methods have been investigated but the latter are preferred as no electrodes are required and therefore interferences from the source are minimised. In ICPs the plasma is generated by inducing a magnetic field around the top of an assembly of coaxial silica tubes through which a conducting gas (commonly ionised argon) is flowing.A solution or an HPLC eluate is introduced into the plasma via a nebuliser. usually of the pneumatic type. In one of the earliest papers that described the monitoring of HPLC eluates with an ICP Gast et a1.’50 found that this type of detector placed few restrictions on the choice of mobile phase or flow-rate. However plasma ignition problems appeared to arise if concentrations of greater than 40% methanol or 10% tetrahydrofuran (THF) were used. This problem can be overcome by ignition of the plasma with initially ethanol - water as eluent and then switching to the methanol - or THF - water mixtures. The ICP detector was reported to have excellent linearity which permitted analyses over a concentration range of four orders of magnitude.A major drawback was that only aromatic hydrocarbons could be used as a mobile phase for normal-phase absorption chromatography. Spray chamber designs have been evaluated for the on-line detection of organometallics separated by SEC using non-aqueous eluents.151 A tapered conical spray chamber gave the most favourable response with toluene. and greater plasma stability was obtained if a cooling jacket was fitted to the spray chamber. It has been reported that the location of the ICP spray chamber with respect to the column and plasma can influence profoundly the ICP response.15’ In this study two HPLC - ICP interface configurations were used to monitor aqueous sol-vents eluting from an ion-exchange column and the two systems are shown in Fig.6. The arrangement in Fig. 6 ( u ) minimises sample transport as an aerosol whilst that in Fig. 6(b) decreases the amount of sample transport in the liquid phase. Placement of the spray chamber external to the plasma box appeared to be the better configuration as peak heights were largely independent of the mobile phase flow-rate. Data obtained in this study cannot of course. be directly related to organic eluents because solvent evaporation plays a much greater role i n aerosol transport processes ANALYST JUNE 1984 VOL. 109 687 At present ICPs have been used to monitor HPLC eluates because they can offer excellent selectivity as shown by some recent applications given in Table 6.Atomic emission is a detection technique that can be employed in HPLC for monitoring trace elements in biological and environmental samples. Many of the problems inherent in flame emission systems can be overcome by using plasma-based detectors as they can offer excellent sensitivity and selectivity. The main restriction in the use of these systems is that of cost. A plasma detector system with full multi-wavelength monitoring capabilities can be purchased cur-rently for about f40000. A good quality flame emission spectrophotometer costs only 23 000. Nuclear Magnetic Resonance Absorption of radiation can be observed for certain nuclei e.g. 1H. 19F 13P and lSN. This phenomenon of nuclear magnetic resonance (NMR) is an extremely powerful tech-nique for the structural elucidation of organic compounds, especially if the resonance of protons is studied and in recent years NMR spectrometers have been applied to the monitor-ing of HPLC eluates.Sensitivity is one of the problems associated with this detection technique and therefore Fourier transform spec-trometers have been used to try and reduce the sample size. Another important factor is that as proton magnetic reso-nances are being studied deuteriated solvents (which are very costly) must be used and this limits the choice of HPLC solvent. Attempts to overcome this problem with an external lock arrangement by placing an ampoule of D20 near to the cell have been considered.16' When HPLC - NMR systems have been used it has been found that the resolution of the spectra is dependent on the eluate flow-rate and cell volume of the systems.lhl-162 In a study carried out with a 90-MHz NMR spectrometer fitted with a 0.4-ml flow cell Bayer et al.161 found that chemical shifts and integration could be obtained from the recorded spectra, but coupling constants could be measured only if the resolution was >3 Hz. In addition they found that the chemical shift references (e.g. tetramethylsilane) could be added to the HPLC eluent with negligible effects on the separation of components. NMR (100 MHz) - preparative HPLC studies have been performed on jet fuels and from the spectra obtained differentiation between straight- and branched-chain alkanes was reportedl6'. Proton integrals could also be used to reproduce the chromatogram.In a follow-up study Haw er a1.164 used a 200-MHz spectrometer with a superconducting magnet. With this system much smaller sample volumes were injected (10-15 pl) and the separation required only 15-50 ml of solvent. These improvements yielded improved chromatographic and spec-troscopic resolution. This system has been used also for the analysis of coal conversion recycle solvents16' and aviation fuels. 166 In summary NMR is one of the few techniques that can offer structural information but as microgram amounts of material are required in order to obtain satisfactory spectra this detection technique is not ideal for the monitoring of HPLC eluates. Further if for no other reason the cost of Fourier transform NMR instrumentation precludes its use as a routine HPLC detector.Electron Spin Resonance As electrons always possess a spin compounds with unpaired electrons e.g. transition metal ions and free radicals yield electron spin resonance (ESR) spectra. Most ESR analyses are performed with spectrometers that operate at a constant frequency (low GHz region) and vary the magnetic field to scan the resonance signals. The efficiency of transfer of energy to a sample is usually very low (ca. 1-30%) and in order to increase the sensitivity of this detection system in HPLC studies the magnetic field is generally fixed at values where the ESR signals are expected. To date all applications of HPLC - ESR have been associated with the analysis of transient free radicals and paramagnetic intermediates.To produce these pre- and post-column thermal photochemical or radiochemical reac-tions have been performed. In one study,167 the sample loop on a valve injector was modified so that thermally and photochemically generated free radicals of phenoxy hydrazyl, nitroxide and semiquinones could be detected after separation on a silica column eluted with benzene. A similar chromato-graphic separation system has also been used to demonstrate the formation of a rhenium complex.168 As many of these free radicals are short-lived intermediates, "spin-trapping" compounds are often used to produce more stable radicals. 2-Methyl-2-nitropropane has been used as a spin-trap in HPLC - ESR studies and the spin adducts of y-irradiated L-valine169 and glycine and ~-alanine170 have been detected.Most of the papers cited contain very little information about the sensitivity or performance characteristics of this monitoring system. The results suggest however that HPLC -ESR can in principle be used for studying free radicals and therefore can be of value to either the physical or organic chemist. Electrochemical Detectors The classification of electrochemical detectors in this review encompasses all detection systems that utilise the measure-ment of any electrical property to monitor a solute in an HPLC eluate. To date capacitance resistance voltage and current have all been used to form the basis of a variety of electrochemical detection techniques and these are listed in Table 7. All of these detection systems will be discussed in detail.Permittivity or Dielectric Constant When any conducting material i.e. a dielectric is subjected to an electrical field electrons are displaced from their equilibrium position. This causes a separation of positive and negative charges and the field is said to have polarised the dielectric. As a result of polarisation if a dielectric is introduced between two plates of a capacitor the capacitance is increased by a factor called the dielectric constant (E). As all materials possess a dielectric constant this property has been used to monitor solutes in solution and therefore can form the basis of a chromatographic detection method. It must be appreciated that because the chosen eluent will always have a dielectric constant this detection system is not particularly sensitive.Further the selectivity of this type of detector is relatively poor and generally the over-all performance is similar to that of a refractive index detector. It is usually found that the dielectric constant is greater for polar substances or materials having more than one functional group although there are some exceptions. For example, dioxane has an exceptionally low dielectric constant although Table 7. Types of electrochemical detection systems currently used in HPLC Electrical property monitored Type of detector Capacitance . . . . . . . . Permittivity (dielectric constant) Resistance . . . . . . . . Conductimetric Voltage . . . . . . . . . . Potentiometric Current . . . . . . . . . . Voltammetric ( ~ .g . coulometric. polarographic and amperometric it is polar this resulting from the two dipoles which are in opposition electrically thereby nullifying the charge. Solvents that show this effect are always worth considering for chromatographic separations if sensitivity is an important factor. It has also been shown. however that better linearity can be achieved with less over-all sensitivity if the dielectric constant of the mobile phase is greater than that of the solute. 171.172 In all dielectric constant detectors the flow cell takes the form of either a cylindrical or a parallel-plate condenser. To obtain a minimum dead volume with either design the condensers have to be very small but as the sensitivity of the detection system is directly related to the capacity of the condenser.plate spacings or cylinder radii and surface areas are important parameters. Alder and Thoe1-173 have carried out an evaluation on plate spacing and shown that there is a compromise between selectivity and sensitivity because for maximum response with media of low F minimum spacing is desirable but this tends to impart more thermal and mechan-ical instability to the cell. Both variable spacing anti cylindrical have been reported. Two methods have been used to measure the changes in dielectric constant that occur as a solute passes through a cell. One method employs an a.c. bridge circuit that incorporates a flow cell in one arm of the bridge. and any out-of-balance signal indicates a change in the dielectric constant. The type of bridge used depends on the design of the cell.If the capacity of the cell is large (>lo0 pF) a Wein bridge can be used. and if small (1-10 pF) a Schering bridge is more appropriate. With both of these circuits reference cells can be accommodated in another arm of the bridge for differential operation of the detector. The modes of operation of both of these a.c. circuits have been explained in detail by Scott.2 The other more recent method that is generally preferred is the heterodyne syctem (see Fig. 7). Changes in the eluate composition followed by the variation of its dielectric constant result in changes in the capacitance of the cell which will modify the frequency cf2) of a measuring oscillator circuit. A frequency mixer is then used to compare this variable frequency against a fixed frequency if,) of another oscillator.The resultant frequency signal cfl - f2) is used to monitor any variation in the dielectric constant of a solution in the flow cell. A theoretical discussion on this detection method has been published by Haderka. 17h Pro-vided that the oscillators are very stable the detector response is reported to be linear over four orders of magnitude. It has been reported that with this heterodyne detection method two oscillators of similar frequency will tend to lock, such that any change in frequency due to a small variation in the dielectric constant cannot be detected. A larger spacing between the plates can overcome this problem but this has the undesirable effect of increasing the volume of the cell.Alder et al. 174 have recently introduced a three-oscillator system that overcomes this problem and permits the detection of com-ponents where only small dielectric constant differences exist between the solute and the mobile phase. One of the major problems associated with dielectric constant detectors is that small changes in temperature have profound effects on the dielectric constant. For operation at maximum sensitivity cell temperatures must be stable to within about 10-3 "C and a number of flow cells have therefore been constructed so that good thermal stability can be maintained. 177-179 A microprocessor-based detector that gives base-line correction for long-term drift has been reported recently. 180 The performance of the dielectric constant detector is comparable to that of the refractive index (RI) detector and is therefore suitable for monitoring eluates when sensitivity or selectivity are not important parameters.The dielectric constant detector cannot be used for gradient elution work, but will tolerate flow programming. 177 The dielectric constant detector does have some advantages over the RI detector. One major advantage is that the dielectric constant detector provides a more uniform response for a group of compounds, e.g. hydrocarbons and therefore calibration factors are not required. A comparison between the responses for some hydrocarbons from ii dielectric constant detector and an RI detector are used to illustrate this point and are shown in Fig. 8. Slightly better sensitivity can usually be expected from the dielectric constant detector because the RI detector tends to lose sensitivity faster as a function of temperature.One disadvantage of the dielectric constant detector in comparison with the RT detector is that a poorer performance is generally obtained if aqueous reversed-phase solvent eluents are used. although some applications using this separation technique have been published. 175 The dielectric constant detector has been shown to be very useful for monitoring size exclusion and hydrocarbon separations. 1x1 Die 1 e c t ric constant detect or s are commercial 1 y avai 1 a b 1 e and for many of the reasons stated above there are applica-tions where it is preferable to use this method of monitoring HPLC eluates.Conductivity Detectors When an acid base or salt is dissolved in an aqueous medium, ionisation occurs. If a voltage is applied across two electrodes placed in a solution containing these charged ions a current will flow between the electrodes and the solution is said to be conducting. As a conducting medium has an electrical resistance this property can be used to detect the presence of ions in a solution and conductivity detectors have been t- I oscillator I 1 f2 Variable-frequency oscillator 117 Recorder Fig. 7. dielectric constant Schcmatic diagram o f a heterodyne system for monitoring 50 I L 4-0 a, a 4-n Ti meim i n Fig. 8. Comparison between responses for a dielectric constant detector and a refractive index detector for some alkanes.Reprinted from reference 175 with permission of Preston Publications Inc available for this purpose for many years. It is worth noting that the specific conductance (K) of a solution is generally quoted instead of specific resistance (p) and the two are related by K = lip. As will be seen later. the circuitry required to monitor specific conductances (units ohm-' cn1-l or mho cm - 1 ) is actually based on the measurement o f resistance. In environmental and biomedical studies there is a need to separate and monitor organic anions and inorganic anions and cations. Over the years there have been developments in both separation techniques and flow cell designs that now make it feasible to perform such analyses with HPLC - conductivity systems.The conductivity flow cell generally consists of a small chamber typically less than 5 pl in volume fitted with two electrodes. These electrodes are commonly platinurn, although other inert conducting tnaterials such as stainless steel or gold can be used. In the older detector designs the resistance of the flow cell contents is measured by incorporat-ing the cell in one arm of a Wheatstone bridge type circuit.' Because electrolysis of water occurs at electrodes with d.c. voltages a.c. voltage must be used. These detectors can be operated in the differential mode by placing a reference cell in another arm of the bridge circuit. With eluents that have a relatively high conductance a sensitivity of about 10-f' g m-1 can be realised with this system. If the eluent conductivity can be reduced considerably the sensitivity can be increased b y about two orders of magnitude.The detection of ionic species in HPLC eluates is extremely difficult with these detection systems because the mobile phases that are suitable for the separation of the ions are too highly conducting. It was not until 1975 when Small et al. 1s' reported a novel technique for reducing the background conductivity. that the potential of the conductivity detector could be evaluated. In this technique the background conduc-tivity suppression is accomplished by placing a second column (suppressor) downstream of the analytical column. Both of these columns contain ion-exchange resins. and the resin used in the suppressor is capable of reducing the eluent conductivity and concerting the sample ions into a common ionic form for conductivity detection.This HPLC - conductivity system is commercially available and is known as an "ion chromato-graph. '' This system has been used in many applications that cover a range of ionic species. One of the drawbacks of this technique is that the sensitivity for certain ions particularly organic anions is poor. This occurs because the large particle size resins and the use of two columns contribute to an increase in band spreading. One group of workers have found that if very low capacity resins are used in conjunction with mobile phases that contain organic aromatic acids the suppressor column is not required and greater sensitivity can be achieved.183-186 An alternative approach is to use reversed-phase packing materials.Some organic acids have been detected in methanol -water eluentsls7 and ion-pairing reagents have also been used to aid the separation of a number of anions and cations.lgs In the latter work a specially designed conductivity cell was used to reduce the currents produced in the electrical circuits. Reducing these currents gives greater thermal stability which in turn improves the base-line sensitivity and hence increases the sensitivity of the detection system. One of the major problems produced by incorporating conductivity cells into a.c. bridge circuits is that undesirable capacitance effects are detected that restrict the conductance range of the detector. In 1970 Johnson and Enkel8" developed the bipolar pulse technique for the fast conductance measure-ments of solutions which overcame this problem.The technique consists in applying consecutive constant voltage pulses of equal magnitude but opposite polarity to a standard conductance cell. The current to voltage ratio is measured at the end of the second pulse. HPLC detectors are now being produced that use this circuitryIYO and they do show that improvements in sensitivity can be achieved. In 1977 Rodgers and Halll~l reported that the use of post-column photochemical reactions could be used to extend the applicability of the conductivity detector. This can be a very selective detection system as only certain organic compounds will undergo photolytic decomposition to produce ionic species.This photoconductivity detector which consists of a reaction coil and a conductivity flow cell was introduced comniercially in 1978. and described by Popovich et ul. 192 The column effluent is divided into two streams one of which passes through a quartz reactor coil irradiated with UV light and into a conductivity cell. The other stream passes through a delay coil and into the reference side of the conductivity cell. The photoconductive response is obtained by measuring the difference in cell conductivities. With this photoconductivity detector sensitivities for nitro-substituted compounds halo-genated aromatics and sulphonamides were reported to be in the 0.2-10 ng range and it was more selective than the UV detector for the analysis of such compounds in pharmaceutic-als and biological fluids.193 In many of these applications low flow-rates had to be employed to ensure sufficient time for the photolysis of the separated compounds to take place in the reaction cell. The over-all effect was that analysis times tended to be long. Photolysis of compounds in volatile eluents can lead to bubble formation and if quartz reaction coils are used back-pressures cannot be applied to rectify this problem. It has been reported that PTFE FEP coils can be used instead and back-pressures can be applied without altering the base-line stability over an eluent flow-rate range of 0.1-6.0 ml min-l in the analysis of some chlorinated pesticides. 194 A novel method for the selective detection of nitrogen with a conductivity detector has recently been reported.195 Organic nitrogenous compounds after separation by normal- or reversed-phase chromatography. are introduced into a fur-nace and the nitrogen is converted into ammonia by reaction with hydrogen over a nickel catalyst. The ammonia is subsequently collected in water - alcohol and detected in a conductivity cell. Submicrogram detection limits were achieved with good precision and a high nitrogen to carbon selectivity was reported. One of the major reasons for the continued development of conductivity detection systems is that very few alternatives exist for the detection of a range of ionic species. In view of the potential of some very practical and interesting UV detection work that has been developed recently for monitoring both UV-absorbing and non-UV-absorbing ions (indirect pho-tometry-see Spectroscopic Detectors Ultraviolet) it would appear that .these techniques are now viable alternatives.These techniques are based on separation on small particle size silica packing materials and sensitivities are comparable to and in some applications superior to those obtained with a conductivity detector. These newer techniques do not make the conductivity detector obsolete because in one instance?' an eluent with a low conductance was actually used so that simultaneous UV - conductivity detection could be used to add to the discrimination of certain ions. Potentiometric Although not mentioned in the previous section poten-tiometry can be used for the detection of ions in HPLC eluates.With some ion-selective electrodes excellent speci-ficity can be achieved and for this reason potentiometry has been used to monitor eluates. If an electrode is made of an element that enters into a chemical equilibrium with ions of the same element in solution a potential is produced. The potential of the electrode is related to the concentration of the ions in solutio 690 ANALYST JUNE 1984 VOL. 109 (Nernst equation). The measurement of these potentials is known as potentiometry. To obtain the concentration of an electroactive species, several conditions have to be met. The basic system requires an indicator electrode that is capable of monitoring the activity of the species of interest and a reference electrode that gives a constant known half-cell potential to which the indicator electrode potential can be related.The voltage produced by combining these electrodes must be measured so that minimal currents are drawn by the measuring circuit. Electrometer amplifiers are suitable for this purpose as they draw currents of typically less than 10-12 A. Very small changes in voltage occur with fairly large variations in concentration and there-fore the measuring system must also have considerable sensitivity. Errors can also be introduced if strict control over the temperature is not maintained. A wide variety of indicator electrodes have been developed for potentiometry,196 and HPLC eluates have been monitored with a number of ion-selective electrode systems. These electrodes show varying degrees of specificity and utilise a membrane to confine an inner reference solution and elec-trode and at the same time make electrolytic contact with the outer (sample) solution.The membrane in each of these electrodes reacts by an ion-exchange mechanism. A few typical constructions are shown in Fig. 9. With miniaturised electrodes flow-through caps can be fitted to produce cells with volumes of about 5 ~ 1 . 1 9 7 An electrode that is responsive to hydrogen ions has been used for monitoring carboxylic acids. 198 Liquid membrane cells have been employed in the detection of nanomole amounts of nitrate and nitrite. 197 Based on amounts injected the response was reported to be linear over a range of 0-10 nmol for nitrate and 0-3 nmol for nitrite. This is a versatile system because it was reported that by changing the liquid ion-exchange solution other ionic species could be detected.A micro-membrane cell has been designed for use as a differential detector for the determination of phosphate, acetate sulphate and fluoride ions.Ig9 The cell is separated into two compartments by an ion-exchange membrane. Column eluate passes through one side of the cell and pure mobile phase through the other side. The different composi-tions of the two solutions bring about a change in the membrane potential. which is used to detect the ionic species. The detector signal is reported to be independent of flow and linear over 2-3 decades. Ion-exchange or reversed-phase ion-pair chromatography can be used with this system and if dilute buffer solutions are used nanomole amounts can be detected.Solid-state membrane electrodes are reported to have some advantages over the other membrane systems in that they respond more rapidly and permit the use of aqueous alcoholic eluents.200 Copper-selective electrodes have been used indirectly for the measurement of transition and rare earth metals201 and amino acids.2M) In the former application the eluate is mixed with a copper - EDTA complex and the metal ions displace the copper. The increase in free copper ions is then measured and a detection limit of 1.5 x 10-10 mol of metal ion was quoted. The other application involves mixing the eluate with a copper ion solution and measuring the loss in copper which results from ligand formation with the amino acids.A copper tubular electrode has been designed specifically for monitoring HPLC eluates.20' The eluate flows through a tubular platinum electrode before entering the tubular copper electrode. These tubes are insulated from each other and the platinum tube is held at ground potential. The response and sensitivity of this system are quoted as being as good as those for a UV detector. The electrode system was used for the analysis of amino acids and because of its high selectivity it could be used to monitor these compounds in urine samples without having to extract them. Glass Porous disc Crystal membrane Fig. 9. Construction of some ion-selective electrodes. (a) Glass membrane; ( b ) liquid ion exchange; and (c) ionic crystal membrane 1 RE DUCT1 0 N I I -0x1 DATl ON - Olefins I Azines Esters Hydrocarbons Aminesiamides Ketones Aldehydes - Phenothiazines Olefinic esters - Catecholamines I - Quinolines I- Aromatic hydroxyls Ethers -Diazo cpds.-Nitro cpds. - F Phenols Halogens I I 1 I -2 - 1 0 + I +2 Applied voltageiV Fig. 10. Electroactivity of a range of organic functional groups. Taken from reference 205 with permission of Journal of Chromato-graphic Science I Fig. 11. Voltage - current curves (voltammograms) for three compounds. A B and C. E(+) and E ( - ) = applied potential; I = anodic (oxidation) current; I = cathodic (reduction) current; and Id = diffusion or limiting current The instrumentation required for potentiometric detection is simple and relatively inexpensive and many laboratories probably have good pH meters that can be used for recording the potentials.Potentiometric detection can be just as effective as conductivity monitoring for some ions but as was stated earlier with the latter technique the newer UV detection methods are possibly superior ANALYST. J U N E 1983. VOL. 109 69 I Voltammetric Approximately 10 years ago there was a resurgence in column chromatography when reversed-phase packing materials were introduced. The outcome of this was that many applications related to the analysis of drugs and biochemicals could be accomplished. About the same time it was realised that such eluents could be rendered conducting by the addition of a small amount of electrolyte and permit the use of voltam-metric detectors for monitoring HPLC eluates.This is currently probably one of the most active areas in detector development. Voltammetry can be used only for the detection of compounds that are electroactive. All voltammetric detectors are based on the principle that when a voltage is applied to an electrode any electroactive material will undergo electrolysis and produce an electric current.203J04 The measurements of these changes are used to monitor solutes in an HPLC eluate. These detection systems afford excellent selectivity because organic functional groups will electrolyse only at specific values of applied potential. The electroactivity of a range of functional groups is shown in Fig. 10. Two different approaches to voltammetry can be used for the detection of electroactive species (a) complete electroly-sis of a solute-coulometry; and (b) partial electrolysis of a solute-amperometry .If amperometry is performed with mercury as the electrode material this is known as polaro-graphy. The detection techniques have all been used for detecting solutes in HPLC eluates and some of these methods are extremely sensitive with detection limits of g or better. This sensitivity is possible because cell volumes of less than 1 p1 can be used and recently a cell with a volume of 1 nl was reported.206 The theory of voltammetry of quiescent solutions is well established but can be applied to flowing ~olutions.~~)~.2~)8 The basics of all voltammetric measurements can be easily understood with the help of the voltage -current relationships.Voltage - current curves (voltammograms) can be obtained for any electroactive compound by measuring the current over a range of potentials. These can be achieved fairly easily for both static or flowing solutions.209 Voltammograms are shown for three hypothetical compounds A B and C in Fig. 11. The increase in anodic current with potential indicates that compounds A and B are being oxidised whereas the increase in cathodic current indicates that compound C is being reduced. Qualitative results can be obtained because the potential at which the oxidation or reduction wave has reached half the total wave height (half-wave potential Ei) is characteristic of each solute in a given electrolyte solution and electrode system. Quantitation is based on the measurement of the current as this is proportional to the bulk concentration of the electroactive material.In most HPLC applications the applied potential is chosen so that the detector is not operated on the limiting current plateau but on the rising part of the curve. Greater sensitivity can be obtained in this manner because background currents from the electrode solvents, etc. tend to increase with an increase in applied potential. The selectivity of the system can be appreciated by using the following example. If compounds A and B (both oxidisable) are not particularly well resolved chromatographically and compound A can still be detected in the presence of B if the oxidation is performed at a low applied potential. This is possible because at this potential B is not oxidised and cannot be detected.An important feature of all modern flow-through voltam-metric flow cells is that three electrodes are required namely working. reference and auxiliary. The last is often called the counter electrode. It is possible to produce a detector with only two electrodes i. e. working and auxiliary electrodes, and monitor the current that flows when a potential is applied across them. This arrangement produces a non-linear response as a voltage drop will occur in the eluate as the current flow changes. By inclusion of a reference electrode the potential of the working electrode can be monitored. The auxiliary electrode still serves to carry the current but if there is any deviation in the potential of the working electrode from the pre-set applied potential current feedback via the auxiliary electrode can be employed to restore the balance.The types of materials used for the reference electrode (e.g., silver - silver chloride) and auxiliary electrode (e.g. platinum or stainless steel) are not critical but the choice of working electrode material is very important as it affects the detector performance. Polarography was one of the earliest forms of voltammetric detection and as was stated previously mercury was used as the working electrode material. Since then many other solid electrodes have been developed and some are very suitable for HPLC detectors. As will be shown all of these electrode materials have both advantages and disadvantages. The mercury electrode is restricted to electro-reduction work because of the anodic range limited to about +0.2 V owing to the oxidation of the mercury.It is one of the few electrode materials however that has an extensive cathodic range because of the large over-potential for the evolution of hydrogen. In suitable supporting electrolytes potentials even more negative than -2 V can be used. Polarography has always been performed with a dropping-mercury electrode (D.M.E.) and the same technique is still being employed with HPLC detector systems. The main advantage of the D.M.E. is that it provides a constantly renewed electrode surface and therefore eliminates the problem of electrode surface contam-ination. One of the major disadvantages with mercury or any other material when used in the reductive mode is that oxygen is easily reduced and this must be completely removed to reduce interference problems.One of the major developments in HPLC voltammetric detection has been the use of solid working electrode materials. As these electrodes have a high anodic range, electro-oxidisable compounds can be detected and numerous applications have been reported. These electrodes can also be used for compounds that are easily reduced. When operated in either mode the residual currents from these electrodes are very low. Solid electrodes are generally preferred to the D.M.E. as the cell design is simpler noise levels are lower and greater sensitivity is usually achieved. One disadvantage is that of non-reproducibility of the electrode surface which is caused primarily by absorption of material and oxidation of the surface; hence care is needed in preparing these electrode surfaces.A wide variety of solid electrode materials have been used in voltammetry and many of these have been tried in flow-through detection systems as shown in Table 8. The most widely used working electrode materials are carbon pastes and glassy carbon. Carbon paste has a fair potential range which can also be adjusted by including a variety of waxes or 0ils.224 This electrode material suffers from Table 8. Solid electrode materials use in HPLC voltammetric detection systems Type of electrode material Electrode material Reference Carbon-based . . Carbon - Nujol pastes 210 Carbon -silicone oil paste 21 1,212 Metallic .. . . Glassy (vitreous) carbon Pyrolytic graphite Low-temperature isotropic carbon Wax-impregnated carbon Graphite - polymer matrices Platinum and silver Gold Cadmium Mercury - gold Mercury - platinum 213 214 215 215.2 2 17-2 220 22 1 222 22 1 223 6 Ag -reference Column Drop pi n g - rn e rcu ry Fig. 12. Design of a dropping-mercury electrode flow cell for monitoring HPLC eluates Reproduced wiith permission cf EC K G Instrument 4 I_ t d . a number of problems which limit its use ~ . g . (a) the solubility of the binder in organic solvents restricts usagc to mobile phases that contain typically less than 20% of organic modifier (b) variable results are obtained owing to the difficulty in preparing reproducibile carbon formulntions; and (c) the electrode material requires a considerable period of time (several hours) to permit operation at high sensitivity.Glassy carbon in many ways is superior because a highly polished surface can be obtained very easily and with a working range of - 1.3 to + 1 .S V it is ideal for a wide variety of applications. Glassy carbon has a lower residual current than carbon paste and operation at high sensitivity can usually be performed within 1 h of start-up. Metallic electrodes. especially those with large surface areas are not ideal for use in flow cells as oxidation of these surfaces occurs and produces high residual currents. Mercury film (amalgams) electrodes have been developed to try to increase the cathodic range of solid electrodes but problems are often encountered in trying to maintain a uniform electrode surface.As voltammetric detectors have been used to monitor HPLC eluates cell design has received considerable atten-tion. The designs of D.M.E. and solid working electrode amperometric detectors are different because of the nature of the electrode material. Coulometric detectors however are very similar to the solid electrode amperometric detectors and differ only in respect that larger electrode surfaces are required in the coulometric detector to achieve complete electrolysis of the electroactive species. Polarography (amperometry with the D.M.E.) has been performed since the 1920s but some radical modifications have been required in order to develop low-volume cells suitable for the monitoring of analytes in HPLC eluates.225 In modern systems that are commercially available the eluate impinges on a mercury drop that has been rapidly dispensed.The current produced by an electroactive species is measured only when the drop has attained a specified size. The drop time and size are important parameters and recent develop-ments have led to detzctors where these can be accurately controlled. Drop times of less than 1 s are essential if poorly retained components are to be monitored accurately because the current is only sampled once per drop. If drop times of greater than 1 s are used the electrode surface will become covered with products of the electrochemical reaction and thereby reduce sensitivity. The drop size must be reproducible as the signal (and background) are directly proportional to the size of the drop.Another reason for not using large mercury ( a ) Thin-layer cell Working electrode 1 Outiet Inlet ib) Wall-jet cell Inlet 1 1 PTFE spacers \ Working electrode .-Fig. 1.3. Cell designs de\eio ed fox u x i n d MPLC arnpt-romctrlc detectors. (0) Thin-la\er (hy v,all-let. For clarirq auuiliarj and reference electrode\ are riot \hewn drops is that with low rnohile phase flow-rates the drops tend to disturb the delivery of the eluate in the cell. The design of a commercially available D.M.E. flow cell that has a volume of about 1 pl is shown i n Fig. 12. In this design the eluate flow is perpendicular to the axis of the capillary. A cell has also been developed in which the eluate fiovi is at right-angles to the axis.”() In a theoreticai paper20x it w:\s predicted that a better signal to noise ratio shorild be obtained if the former orientation is employed.Many cell designs for solid working electrode amperometric detectors have been devised. and with very few exceptionc they are all based on d generally accepted three-electrode configuration. The reference and auxiliary electrode> are placed on the downstream side of the working electrode so that either leakage from the reference electrode or the formation of any electrochemical products at the auxiliary electrode do not interfere with the working electrode. The reference is normally placed close to the working electrode to ensure that the electrical resistance of the cell is kept to a minimum.These electrodes are fitted into the body of a flow cell which is usually constructed from a non-conducting material. Metallic cell bodies e.g, stainless steel are con-sidered to be unsuitable because high residual currents can be produced owing to metal ions being leached from the wrface of the material. Kel-F has been used in numerous cell designs as it is a good insulator easy to machine and resists attack from a wide range of solvents. Glass-filled PTFE (GFT) has been used as an alternative as it c a t s less but has very cimilar properties .227 The working electrode surface can be formed from part of a tube. wire sheet or disc of the desired material. Develop-ments with planar surfaces have been the most popular as very low cell volumes can be achieved.There are basically two designs that have been used in the development of amperometric detectorc namely thin-layer and wail-jet configurations and these are shown in Fig. 13. The wall-jet system was devised by Yamada and Mat-suda,228 and Fleet and Little119 were the first of many groups of workers to use this popular method for monitoring HPLC eluates. The thin-layer cell is easier to construct and has been found to be particularly useful for carbon paste working electrodes.2”) In numerous thin-layer and wall-jet cells the outlet tubing is stainless steel so it can be used as an auxiliary electrode and thereby simplify cell designs. In Fig. 13 it can be seen that both types of cell are made i n two halves.An insulating material. commonly PTFE is used to separate them and a slot is cut in it to allow the eluate to pass over the working electrode. By using various thicknesses of PTFE the spacing and hence the cell volume can be changed. Volumes typically less than 5 pl can be obtained with this procedure. It is worth noting that if the spacer thickness i reduced. the sensitivity of the detecticm system can be increased owing to the higher linear flow-rate of the eluate. There is however. an optimum thickness because by reducing the spacing the detection limit will not continue to improve owing to an increase in the noise level. It ic recommended that a spacing of less than SO pm 5hould not be used owing to the limited flatness of the surface o f an) electrode material.? i I 2~ Tubular platinum electrodes were investigated a< flou cells in 1963.:" Mainly as a result of surface oxidation they havc received only limited attention i n ItlPLC but in a more recent paper it wa5 shown that cell volumes of about 1 id can be obtained with this system and nanomole amounts could be detected.'?J An interesting developinent with platinurn wire electrodes using the well-jct principle has recently been published and is of value to those interested in micro-HPLC.20h By using the cross-sectional area of a 0 .1 mm diameter wire and setting the exit of a capillary HPLC column within 0.1 mm of the wire a cell volume of less than 1 nl was obtained. The response of any voltammetric detector is very suscep-tible to flow-rate fluctuations and is dependent on the mass transport of the electroactive material from the bulk solution in the flow cell to the electrode.A study ha5 been performed in an attempt to overcome these probiems by using a rotating-disc electrode (R.D.E.).ZT5 In comparison with a static disc, the R.D.E. (rotation speed 20 rev 5- 1 j gave a considerable increase in response and no change In background level was observed. As was stated earlier the design ot a coulonnetric detector is very similar to that of an ampemmetric detector. but larger surface areas of the working electrode are required in order to obtain complete electrolysis. Coulometric detectors have not experienced much popularity because their sensitivity is often below that of an amperometric detector in which usually less than 20% of the material is electrolysed.The reasons put forward for the lack of 5ensitivity are as follows (a) it is difficult to obtain 100% electrolysis efficiency; (b) back-ground currents increase with surface area of working electrode; (c) the larger cell volume; and (dj electrode surface contamination. In order to maximise efficiency coulometric cells have been based on the thin-layer design and some large surface area electrodes that have been used include carbon. silver and platinum gauze,Z2().236 large planar glassy carbon surfaces237 and tubular platinum238 and cadmium.": Coulometric voltammetry has also been carried out with optically transparent electrodes so that spectroscopic data can be obtained on the electrochemically produced species.239 These electrodes are produced by vapour depositing materials such as carbon and mercury on glass or quartz.A thin-layer optically transparent gold minigrid electrode has been evalu-ated in an HPLC system.24J" There has been considerable interest in recent years in systems employing more than one voltammetric detector to try and improve the over-all selectivity or sensitivity of the detector system. Pre-column coulometric detectors have been used to control the chromatographic selectivity by removal of possible electroactive interferences.'-",~J2 A severe limitation of these systems is that electrode reactions produce multiple products and therefore many of the more recent systems have concentrated on moniroring electrochemical conver5ions after chromatographic separation.Eggli and Asper243 used sepa-rate cmiometric and arnperwictric detectors in series both being operated in the reductive mode. Compounds containing a disulphtde bond were reduced with the coulometric cell and the electrogenerated sulphj dry' i-.ompczutids were detected at the ampercmetric detector. A similar system has been used to separate required compounds frcm interferents in a single run.z4J The selectivity of this method was also compared with that of a differential-pulse technique ;nd the results were i,,und to be comparable to or siightly superior to those obtained with the dual cell arrangement. Coulometric -amperometric detection has also been uwd to monitor conipounds that are not detectable by direct voltammetric methods.In a recent report the coulometric cell was used to liberate bromide which reacts with the compound to produce a detectable species in the arnperometric detector.2J5 Brunt and Bruins232J46 studied the concept of differential amperometric detection. In their system the sensitivity was increased because the background currents were reduced. This was accomplished by using a differential amplifier to monitor the output of two identical sample and reference amperometric cells. Further advances in dual electrode detection systems have been made based on the design and use of a single cell fitted with two working electrodes. This development overcomes problems associated with the band spreading that is created when two separate cells are employed.With the single cell -dual electrode arrangement coulometric - arnperometric and amperometric - amperometric systems have been devised, based on the thin-layer cell design. With this design three possible configurations of the electrodes with respect to the flow axis can be achieved. These ha\e been termed parallel opposed. parallel adjacent and series configurations and are shown in big. 14. Parallel opposed217 and parallel adjacent248 configurations have been used in very similar applications where a different voltage has been applied to each electrode. By monitoring the resultant output signals simultaneously considerable discrimi-nation can be obtained with this technique. The series configuration can be used to increase selectivity and with some applications greater sensitivity can be acliieved.21'),2i(' A useful detection scheme for the series dual electrode involves the detection of products from the upstream (generator) electrode at the downstream (detector) electrode.By setting the two electrodes at different potentials, and to some extent by varying the electrode materials either oxidative - reductive or reductive - oxidative detection can be achieved. The latter technique can be useful for the removal of oxygen interferences during reduction mode detection and this can improve the detection limits for compounds that reduce at very negative potentials.'51 252 Beauchamp et al.2ss designed a cell that can be used for dual detection and is very flexible in that a variety of electrodes can be employed.For example solid or D.M.E. electrodes can be used arid the former can be operated in either of the thin-layer (\durne 4.8 pl) or wall-jet (volume 0.2 plj configuration\. A micro-[IPLC sy5tem with a dual series electrode cell has also been developed recently.2sJ In most applications where a voltammetric cell has been used to monitor HPLC eluates d.c. voltages have usually been applied to the electrode to electrolyse the analyte. With this form of applied potentid the measured current actually consists of two components ( 1 ) Faradaic current (IF) which is due to the electron transfer at the electrode surface and this is proportional to the concentration of the electroactive species; and (ii) charging current ( I c ) which is due to the capacitive nature of the electrode surface.The charging current is undesirable as it contributes to the background current, thereby lowering the detection limit of the system. By applying a voltage to an electrode IF and Ic increase but Flow Parallel opposed Fig. 14. Electrode detection Parallel adjacent Series configurations for thin-layer dual electrod 694 ANALYST JUNE 1984 VOL. 109 if the voltage is removed Zc decays at a much faster rate than I F . A variety of pulsed voltage waveforms as shown in Table 9 have therefore been used to try to utilise this effect to increase the sensitivity of the detection system and yield other advantages such as decreased dependence of the measured current on flow-rate increased stability of the electrode surface and in some instances greater specificity.These waveforms have been applied to a variety of working electrode materials. Pulsed and reverse pulsed techniques are similar in opera-tion in that the current is measured just before the applied voltage is removed. The difference between the two tech-niques is the potential that is applied initially. Pulsed methods use a positive initial potential whereas reverse pulsed tech-niques use a large negative potential. The advantage of the latter is that the analytical signal is measured at a potential where dissolved oxygen is not reduced. In differential-pulse voltammetry (DPV) a pulsed potential is applied and the current is measured twice once before applying the pulse and once just before the end of the pulse.The chromatogram obtained is a plot of current difference versus time. One of the reported advantages of DPV over normal d.c. operation is that better specificity can be obtained because with DPV the detector is responsive only to species with oxidation potentials near to or between the initial and final applied potentials. To perform DPV at a dropping-mercury electrode the pulse has to be synchronised accurately with the drop formation and commercial instrumentation is now available for this technique.22s Discrimination between Faradaic and charging currents can be obtained with both a.c. and rapid-scan square-wave voltammetry. In the latter technique selectivity can be improved because potential and time resolution can be obtained from the same chromatographic run (three-dimensional chromatopolarogram).This technique cannot be considered suitable for routine analysis because of the cost and complexity of the associated instrumentation. The choice of the chromatographic eluent is important in any voltammetric detection method. In the introductory paragraph in this section it was stated that this detection method could be used in HPLC owing to the development of reversed-phase packing materials. Although aqueous eluents are often ideal for many separations they are not entirely suitable for voltammetry because water produces a high background current owing to its relatively low oxidation potential (+1.3 V). A recent study has shown that excellent separations of basic compounds which are suitable for analysis in the oxidative mode can be achieved on silica with non-aqueous eluents of methanol - perchloric acid.2hh The absence of water increases the oxidative range to about +1.6 V (oxidation potential of methanol) which improves the sensitivity.Whilst dealing with the subject of eluents it is a prerequisite that their delivery is pulseless or have a pulse frequency that is faster than the time constant of the detector - recorder in order that these detectors can be operated at high sensitivity. Many pumps that are reported to be pulseless are unsuitable but recently a pump with an extremely fast piston action (23 cycles s-1) has been introduced and is ideal for use with voltammetric detectors. Table 9. Some pulsed voltage waveforms used in HPLC voltammetric detection Pulsed waveform Reference Pulsed d.c.. . . . . . . . 255.256 Differential pulse . . . . . . 234.255-261 A.c. . . . . . . . . . . . . 262.263 Reverse pulsed d.c. . . . . . . 263 Rapid-scansquare-wave . . . . 265 The popularity of voltammetric detection in HPLC can be observed by the number of papers that are appearing in the current literature. Readers interested in techniques and applications are referred to an excellent bibliography that covers voltammetry in HPLC from its conception almost to the present day.267 One of the major application areas is the analysis of catecholamines at low levels and a review devoted to this topic has been published.2hX A similar review has been published on pesticides.'@ In addition to these papers, comparative studies between voltammetric and other non-voltammetric detectors have a~peared.2~~272 A very high proportion of these applications have been performed in the oxidative mode by d.c.amperometry on solid working electrode surfaces. The reasons for this are firstly that by operating in the oxidative mode oxygen interference problems do not occur and secondly that this type of detector can be built in a laboratory. With the numerous cell designs and associated electronic systems that have been described it is possible to build one of these detectors for under &500. Despite various claims of improved specificity and sensitivity that have been reported with pulsed voltage voltammetry many chromatographers have elected to use d.c. voltammetry as less complex instrumentation is generally required.In applications that require operation in the reductive mode particularly at low potentials the D.M.E. has generally to be used. The designs of these detectors are complex and users tend to purchase them from a manufacturer and consequently have to pay the current price of about f.5000. These detectors also require considerable practical expertise in order to obtain reproducible results. These drawbacks, coupled with the problem of oxygen interference severely limit the number of applications to which reductive mode detection could be applied. A method for the removal of oxygen has been devised recently by Lloyd,273 which could certainly improve the situation. Further mercury is not very desirable as a working electrode material (handling problems, purity toxicity etc.) and if an alternative could be found then reductive mode detection would receive considerable popular-ity.Finally voltammetric detection is still a comparatively new technique so this is one area in HPLC detection where developments should occur over the next few years. 1. 2. 3. 4. 5 . 6. 7. 8. 9. 10. 11. 12. 13. 14. 1s. 16. 17. 18. References Novotny M. Anal. Chem. 1981 53 1294A. Scott R. P. W. "Liquid Chromatography Detectors," Else-vier Amsterdam 1977. Wolf T. Fritz G. T. and Palmer L. R . 1. Chromatogr. Sci., 1981 19. 387. Knox J . H. Editor. "High Performance Liquid Chromato-graphy," Edinburgh University Press Edinburgh 1978. Parris N. A "Instrumental Liquid Chromatography." Else-vier Amsterdam 1976.Kirkland J . J . . Anal. Chem. 1968. 40 391. Horrath C. G. and Lipsky S . R Nature (London). 1966. 211 748. Little J. N and Fallick. G. J . 1. Chromatogr. 1975 112. 389. Higgins. J . W. J. Chromatogr 1978. 148. 335. Lawrence J. F. J . Chromatogr. 1981. 211. 144. White P. C J . Chromutogr. 1980 200. 271. Readman J . W Brown L. and Rhead. M. M. Analwt 1981. 106. 122. Smith R. N and Zetlein. M J . Chromutogr 1977. 130 314. Li K.-P and Arrington J . . Anal. Chem. 1979. 51. 287. Denkert M. Hackzell. L. Schill. G and Sjogren. E J . Chromurogr. 1981 218 31. Barber. W. E and Carr. P. W J. Chrornatogr. 1983.260.89. Small H . . and Miller. T. E. Anal. Chem 1982. 54. 462. Cochrane R. A and Hillman D. E J .Chromafogr. 1982, 241 392 ANALYST JUNE 1984 VOL. 109 69 5 19. 20. 21. 22. 23. 24. 2.5. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47 I 48. 49. SO. 51. 52. 53. 54. 5 5 . 56. 57 58 59 60 61 62. 63. 64. Larson J . R . and Pfieffer C. D . J . Chromatogr. 1983 259, 519. Larson J. R. and Pfieffer C. D . Anal. Chem. 1983,55.393. Wheals B. B. J . Chromatogr. 1983 262 61. Fell A . F. Anal. Proc. 1980 17 512. McDowell E. and Pardue H. L Anal. Chem. 1976 48, 1815. McDowell E . and Pardue H. L. Anal. Chem. 1977 49, 1171. Fell A. F. Anal. Proc. 1980 17 266. Horlick G . Appl. Spectrosc. 1976 30 113. Dessy R. E Nunn W. G . and Titus C. A .J. Chromatogr. Sci. 1976 14 195. Milano M. J . Lam S. and Grushka E J . Chromatogr., 1976 125. 31.5. Milano M. J. Lam S . Savonis M. Pauter D. B. Pav, J. W. and Grushka E. J . Chromatogr. 1978 149 599. Amita Y. Ichise M. and Kojima T. J . Chromatogr. 1982, 234. 89. Drenth B. F. H. Ghijsen R . T. and De Zeeuw R. A. J . Chromatogr. 1982 238 113. Ghijsen R . T . Drenth B. F. H. Overzet F. and De Zeeuw, R. A . J . High Resolut. Chromatogr. Chromatogr. Commun., 1982 5 192. Overzet. F. Ghijsen R. T. Drenth B. F. €I. and De Zeeuw, R. A . J . Chromatogr. 1982 240 190. Overzet F. J . High Resolut. Chromatogr. Chromatogr. Commun. 1982 11 604. Overzet F. Rurak A . Van der Voet H . Drenth B. F. H . , Ghijsen. R . T. and De Zeeuw R . A J . Chromatogr. 1983, 267 329.Fell A . F. Scott H. P Gill R. and Moffat A. C . J . Chromatogr. 1983 273 3. Willis B. G Hewlett-Packard Appl. Note No. 8450-5 1979. Borman S . A Anal. Chem. 1983 55 836A. Klatt L. N J . Chrornatogr. Sci. 1979 17 225. Catterick T. J . Chromatogr. 1983 259 59. Russell J. R . J . Chromatogr. 1983 280 370. White P. C. and Catterick T. J . Chromatogr. 1983 280, 376. Morris C. J . 0. R . and Morris P. Editors “Separation Methods in Biochemistry,” Second Edition Wiley New York, 1976. Jonker K. M. Poppe H . and Huber. J. F. K. Chromato-graphia 1978 11 123. Brown N. D . Sing H. C . Neeley W. E. and Koetitz S. E., Clin. Chem. 1977 23 1281. Alder H. Margoshes M. Snyder L. R. and Spitzer C J. Chromatogr. Biomed. Appl. 1977 143 125. Kaitaranta J .K. and Bessman S. P . Anal. Chem. 1981 53, 1232. Hirai Y Yoza N. and Ohashi S . J . Chromatogr. 1981,206. 501. Hirose. A Iwasaki Y. Iwata I. Ueda. K . and Ishii D . J . High Resolut. Chromatogr. Chromatogr. Commun. 1981 4, 530. Hwang J.-M. Shih. J . S . . Yeh Y .-C. and Wu S.-C Analysl, 1981 106 869. Zenki M. Anal. Chem. 1981 53 968. Borys. H . K and Karler. R. J . Chromatogr. 1981 205 303. Cassidy R . M. and Elchuk S. J . Chromatogr. Sci. 1981 19. 503. Guilbault. G. G . . “Practical Fluorescence,” Marcel Dekker, New York 1973. Udenfriend “Fluorescence Assay in Biology and Medicine,” Academic Press. New York 1962. Steichen J . C . . J . Chromutogr. 1975 104 39. Jadamec. J . R . Saner. W. A . . and Talmi Y. A n d . Chem., 1977. 49. 1316. Fogarty.M. P Shelley. D. C and Warner. I. M. J . High Resolut. Chromafogr. Chromatogr. Commun. 1981 4 561, Shelley. D. C. Fogarty M. P. and Warner I . M . . J . High Resolut. Chrornutogr. Chromutogr. Commun. 1981. 4. 616. Wright. J . C . and Wirth. M. J . Anal. Chem. 1980,52 1087A. Wright. J. C . . and Wirth M. J . A i d . Chem 1980,52. 9X8A. Yeung. E. S. and Sepaniak. M. J . . Anul. c‘hpm. 1980. 52, 1465A. Green R. B. A i d . Chem. 1983 55. 20A. Diebold. G. J . . and Zare. R. N . . Science. 1977. 196. 1439. 65. 66. 67. 68. 69. 70. Voigtman E. Jurgensen A. and Winefordner J. D. Anal. Chem. 1981 53 1921. Sepaniak M. J. and Yeung E. S. J . Chromatogr. 1981 190, 377. Sepaniak M. J. and Yeung E. S. J . Chromatogr. 1981,211, 9.5. Hershberger L.W. Callis J. B. and Christian G. D . Anal. Chem. 1979 51 1444. Sepaniak M. J. and Yeung E. S. Anal. Chem. 1977 49, 1554. Rauhut M. M. Bollyky L. J . Roberts B. G . Loy M., Whitman R. H. Iannotta A. V. Semsel A. M. and Clarke, R . A . . J . Am. Chem. Soc. 1967 89 651.5. Bollyky L. J . Whitman R . H. Roberts B. G . and Rauhut, M. M. J . Am. Chem. SOC. 1967,89 6.532. McCapra F. Prog. Org. Chem. 1971 8 231. McCapra. F. Pure Appl. Chem. 1974 24 611. Gundermann K.-D. Top. Curr. Chem. 1974. 46 61. Thore A . Sci. Tools 1979 26 30. Kobayashi S. and Imai K. Anal. Chem. 1980 52,424. Bostick W. D . Denton M. S . and Dinsmore S . R . Clin. Chem. 1980 26 712. Veazey R. L. and Nieman T. A. J . Chrornatogr. 1980,200. 153. Birks J . W. and Kuge M. C. Anal. Chem. 1980 52 897.Shoemaker B. and Birks J. W. J . Chromatogr. 1981 209, 251. Sigvardson K. W. and Birks J. W. Anal. Chem. 1983 55, 432. Kobayashi S . Sekino J . Honda. K. and Imai K. Anal. Biochem. 1981 112 99. Fine D . H. Huffman F. Rounbehler D. P. and Belcher, N. M. “Environmental N-Nitroso Compounds Analysis and Formation,” IARC Scientific Publications No. 14 IARC, Lyon 1976. Fan T. Y. Morrison J. Rounbehler D. P. Ross R. and Fine D . H. Science 1977 196 170. Fan T. Y. Herbst W. and Fine D . H . 37th Annual Meeting of the Institute of Food Technologists Philadelphia. 1977, Paper No. 1.52. 86. Spanggord R . J . and Keck. R. G . J . Pharm. Sci. 1980,69, 444. 87. Yu W. C. and Goff E. Anal. Chem. 1983 55 29. 88. Massey R. C . Crews C. McWeeny D. J .and Knowles, M. E. J . Chromatogr. 1982 236 527. 89. Lafleur A. L. and Morriseau B. D. Anal. Chem. 1980 52, 1313. 90. Malcolm-Lawes D. J Warwick P. and Gifford L. A . J . Chromatogr. 1979 176 1.57. 91. Malcolm-Lawes D . J. and Warwick P. J . Chromatogr., 1980 200 47. 92. Lawrence J. F. and Frei R . W. “Chemical Derivatisation in Liquid Chromatography,” Elsevier Amsterdam 1976. 93. Twitchett P. J. and Williams P. L. U.S. Pat. 4233030 1978. 94. Twitchett P. J . Williams P. L. and Moffat A . C . J . Chromatogr. 1978 149 683. 95. Rhys-Williams A . T. Winfield S . A. and Belloli R. C. J . Chromatogr. 1982 235 461. 96. Iwaoka W. I. and Tannenbaum IARC Sci. Publ. 1979 14, 51. 97. Brinckman U. A . T. Welling P. L. M. De Vries G Scholten A. H . M. T. and Frei R.W. J . Chromatogr. 1981. 217 463. Scholten A. H . M. T. Brinkman U. A. T . and Frei R. W., Anal. Chim. Acfa 1980 114 137. Scholten A. H. M. T. Welling P. L. M. Brinkman U . A . T and Frei R . W. J . Chromatogr. 1980 199 239. Scholten A. H . M. T Brinkman U. A. T. andFrei R . W. J . Chromatogr. 1981. 205 229. Scholten A . H. M. T. and Frei R. W. J . Chromatogr. 1979, 176 349. Rodriguez. F. Kulakowski R. A . andClark. 0. K . Znd. Eng. Chem. Prod. Res. Dev. 1966 5 118. Ross. J. H. and Shank R. LAdv. Chem. Ser. 1973 No. 125, 108. Mirabella F. M. Johnson J. F and Barrall E. M. Am. Lab. 1975 10. 65. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 98. 99. 100. 101. 102. 103. 104. 105. Bartick. E.G . . J . Chromatow. Sci 1979. 17. 336 696 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120, 121. 122. 123. 124. 12s. 126. 127. 128. 129. 130. 131. 132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 142. 143. 144. 145. 146. 147. 148. Brown R. S . Hauser. D . W. andTaylor. L. T Anal. C'hein., 1980 52 1511. Payne-Wahl K Spencer G . F Plattner. R. D and Butterfield. R . C J . Chromutogr. 1981 209 6 1 . Cook. M. and Godfrey N. R. J . Chromatogr. 1982.237 15 1. Gloor R and Leidner H Anal. Chem. 1979 51. 635. Bell A. T. and Hair. M. L. Editors "Vibrational Spectrosco-pies for Absorbed Species." ACS Symposium Series. No. i37, American Chemical Society. Washington DC.1980. Vidrine. D . W and Mattson D. R. Appl. Spectrosc. 1978. 32 502. Combellas. C Bayart H. Jasse. B Caude. M. and Rossett, R . J . Chromatogr. 1983. 259. 21 1. Vidrine D . W "Fourier Transform Infrared Spectroscopy." Volume 2 Academic Press New York. 1979. Teraniae. N. and Tanaka. S Spectrosc. Lett. 1980. 13 117. Johnson. C. C . and Taylor. L. T Anal. Chem. 1983 55.336. Vidrine. D. W. J . Chromatogr. Sci. 1979. 17 477. Kuehl. D. T and Griffiths P. R . J . Chromafogr. Sci. 1979, 17 471. Kuehl D. T. and Griffiths P. R Anal. Chern. 1980. 52. 1394. Jinno K. and Fujimoto C J . High Resolut. C'hromatogr. C'hromutogr. Commun 1981 4 532. Christian. G . D . and Feldnian F. J . . "Atomic Absorption Spectroscopy Applications In Agriculture Biology and Medi-cine," Wiley-Interscience New York.1970. Ramirez-Munoz. J . "Atomic Absorption Spectroscopy and .4nalysis by Atomic Absorption Flame Photometry," Elsevier, Amsterdam. 1968. Kirkbright G . F. and Sargcnt. M . . "Atotnic Absorption and Fluorescence Spectroscopy," Academic Press New York, 1974. Manahan S . E . and Jones. D. R . Anal. Leti 1973. 6 245. Jones D . R l'ung. 1-1. C. and Manahan. S . E . . Atid. Chem., 1976 48. 7. Yoza N. and Ohashi S. Anal. Lett. 1973 6 595. Koropchak J . A . . and Coleman G . N . Anal. Chem. 1?80, 52 1252. Berndt. H . and Slavin W. At. Absorpt. JL'ewsl. 1978. 17. 109. Atwood J . G Schmidt. G. J . . and Slavin W. J . Chromar-ogr. 1979. 171 109. Slavin W and Schmidt G . J . . I . C'hromatugr. Sci. 1979 17, 610. Burns.D . T. Glocking F and Harriot. M Analj,st 1981, 106 921. Messman J. D. and Rains T. C. Anal. @hem. 1981. 53, 1632. Brinckman F. E. Blair N. R Jewett K. L. and Iverson, W. P . J . Chromatogr. Sci 1977 IS 493. Vickrey T. M. Buren M. S . and Howell. H. E . Anal. Lett., 1978 A l l 1075. Vickrey. T. M. Howell H. E . and Paradise. M. T Anal. Chem. 1979 51 1880. Brinckman F. E . Jewett K. L. Iverson. W. P Irgolic K. J . . Ehrhardt K. C. and Stockton. R. A J . Chromatogr. 1980. 191 31. Koizumi H . McLaughlin R . M. and Hadeishi T Anal. Chem. 1979 51 387. Vickrey T. M Howell. H . E. Harrison G . V. and Ramelow ti. J . Anal. Chem. 1980 52 1743. Tittarelli. P . and Mascherpa. A Anal. Chem. 1981.53 1466. Woolson E. A . and Aharonson. N . J . Assoc. Off.Anal. Chem. 1980 63 523. Jewett K. L. and Brincknian F. E. J . Chromutogr. Sci., 1981 19. 583. Parks E. J . Brinckman. F. E and Blair W. R J . Chromatogr. 1979 185. 563. Julin B. G . Vandenborn H . W. and Kirkland J . J . J. Chromatogr. 1975 112 433. McGuffin V. L. and Novotny. M. Anal. Chem. 1981. 53, 946. Chester T. L. Anal. Chem. 1980 52 638. Chester. T. L Anal. Chem. 1980 52. 1621. Gilbert P. T Anal. Chem. 1966 38. 1920. Folestad S and Josefsson. B. J . Chromutogr. 1981.203. 173. Gutsche B. Hermann R . Hohne M. and Rudiger. K Spectrochim. Acta Part A. 1979. 33 609. 149. 150. 151. 152. 153. 154. 155. 156. 157. 158. 159. 160. 161. 162. 163. 164. 165. 166. 167. 168. 169. 170. 171. 172. 173. 174. 175.176. 177. 178. 179. 180. 181. 182. 183. 184. 185. 186. 187. 188. 189. 190. 191. 102. 193. 194. Mrull. I . S . Panaro K. W . . and Gershatn. L. L . . .I. Chromatogr. Sci. 1983. 21. 360. Gast C. H Kraak. J . C. Poppe. H and Maessen. F. J . bl J . . J . Chromutogr. 1979. 185. 549. Hausler. D . W. and Tayior. 3'. T A n d . ('hem 1981. 53. 1223. Whaley I. B. S . . Snable. K. R . . and Browner. R . F ilrrml. Chem. 1982. 54. 162. Fraley. D. M. Yates. D and Manahan. S. E Anal. Chern., 1979. 51 2225. Hausler. D . W. and Taylor. 7'. T Anul. Chrm 1981. 53. 1227. Morita. M. and lJehiro T. Anal. C'htwz. 1981. 53. 1997. Morita M Uehiro T and Keiichiro F . . Anal. Chem. 1981, 53. 1806. Morita. M Uehiro. 'I and Fuwa. K Anal. Chem 1980.52, 349. Heine. D . R . Schlaback. T. D and Denton M. B . . Anal. Chem. 1982. 54. 81. McCarthy. J . P. Caruso. J . A . . and Fricke. F. L J . Chromatogr. Sci 1983. 21. 389. Yoshida. K Hasegawa T and Haraguchi H Anul. Chem 1983. 55 2106. Bayer. E Albert K . Nieder. M. Grom E and Kcller. T. J , Chromatogr. 1979 186. 497. Fyfe. C. A Cocivcr:i M. Danji S . W. H. Hostettcr. T. A , Sproat D . . and O'Brien. J . . J . Magn. Reson. 1976. 23. 377. Haw J . F. Glass T. E . Hausler D. W Motell. E and Dorn €I. C . Anal. Chem. 1980 52 1135. Haw J . F Glass. T. E and Dorn. H. C And. Chem. 1981. 53. 2327. Haw. J . F Glass. T. E . and Dorn. H. E . Anal. Chem. 1981. 53 2332. Haw. J . F Glass. T. E . and Dorn. H. E . Anal. Chem. 1983, 55,22. Reich T Chen K.S . . and Wan J . K. S . J . Chromatogr., 1981,206 139. Creber. K. A . M. and Wan J . K. S. J . A m . Chem. Soc 1981, 103 2101. Makino K J . Phvs. Chem. 1980 84 1016. Moriya F Makino. K. Suzuki. H . . Rokushkia. S . and Hatano H. J . Phys. Chem. 1980. 84 3085. Haderka S . J . Chromutogr. 1970. 52. 213. Vespalec. R and Hana K. J . Chromatogr. 1972. 65. 53. Alder J . F. and l h o e r A. J . Chromatogr. 1979. 178. 15. Alder J . F Drew. P. K. P and Fieiden P. R . J . Chromatogr. 1981. 212 167. Benningfield. L. V. and hlowery. R. A . J . Chromatogr. Sci., 1981. 19 115. Haderka S . . J . Chromatogr. 1971 54 357. Vespalec R. and Hana K. J . Chromatogr. 1972 65. 53. Poppe H and Kuysten. J . J . C,'hromatogr. 1977. 132 369. Slavik V. J . Chromatogr. 1978 148 117.Alder J. F. Drew P. K. P . . and Fielden P. R Arzal. Chem., 1983. 55. 256. Bade R. K. Benningfield L. V Mowery R. A . and Fuller. E . N. Int. Lab. NoviDec 1981 40. Small H. Stevens. T. S . and Bauman W. C. Anul. Chem., 1975 47 1801. Gjerde D . T. Fritz. J . S. and Schmuckler. G . J . Chromat-ogr. 1979 186 509. Fritz J . S . Gjerde D . T and Becker R . M. Anal. Chem., 1980 52. 1519. Gjerde D. T. Schmuckler. G . . and Fritz. J. S. J . Chromut-ogr. 1980. 187 35. Gjerde D . T and Fritz J . S. Anal. Chem. 1981. 53 2324. Thurman E. M. J . Chromatogr. 1979 185. 625. Molnar. I. Knauer. H and Wilk D. J . Chromatogr. 1980, 201 225. Johnson D. H . and Enke C. G . Anal. Chem. 1970,42,329. Keller J. M. Anal. Chem. 1981 53 344. Rodgers D. H . and Hall R .C. Pittsburgh Conference on Analytical Chemistry 1977. Paper No. 8. Popovich D . J . Dixon J . B. and Ehrlich B. J . J . Chromatogr. Sci. 1979. 17 643. McKinley W. A . Anal. Toxicol. 1981 5 209. Ciccioli P . Tappa R. and Guiducci A . A n d . Chem. 1981, 53. 1309 697 195. 196. 197. 198. 199. 200. 201. 202. 203. 204. 205. 206. 207. 208. 209. 210. 211. 212. 213. 214. 215. 216. 217. 218. 219. 220. 221. 222. 223. 224. 225. 226. 227. 228. 229. 230. 231. 232. 233. 234. 235. Lloyd R . J . . J . Chromarogr. 1981 216 127. Buck R. P Anal. Chern. 1974 46 28R. Schultz F. A and Mathis D. E . Anal. Chem. 1974,46.2253. Egashira S . J . Chromatogr. 1980 202 37. Deelder. R . S . . Linssen.H. A. J . . Kocn. J . G and Beescn. A . J . B J . Chromufogr 1981. 203 1 3 . Loscombe. C . R Cox. G. B and Dalziel. J . A. W J . Chromarogr. 1978. 166 403. Dorey R . C 111. Drss. Absrr. Znt. R 1980. 40. 1672. Alexander. P. W Haddad P. R . Low G. K. C . and Maitra, C . J . Chromatogr. 1981. 209 29. Sawyer D . T and Roberts. L. J . . "Experimental Electro-chemistry for Chemists." Wiley-Interscience New York. 1974. Adams R . N . J . Pharm. Sci 1969. 58 1171. Fleet. B and Little C. J . . J . Chromatogr. Sci. 1974 12. 747. Slais K and Krejci. M. J . Chromatogr. 1982 235 21. Rucki R. J . . Talanta 1980 27 147. Hanekamp. H. B and Van Nieuwkerk H . J . Anal. Chim. Acta 1980. 121 13. Anderson J . L Weisshaar D . E. and Tallman D. E . Anal. Chern. 1981 53 906.Adams. R . N. "Electrochemistry at Solid Electrodes," Marcel Dekker New York 1969. Pungor. E. Feher Z. S . . and Nagy. G . . Anal. Chim. Acfa, 1970 51 417. Nagy. G. Feher Z . and Pungor. E . Anal. Chim. Acta 1970. 52. 17. Zittel. H. E. and Miller F. J Anal. Chem. 1965 37 200. Wightman R . M. Paik. E. C. Borman. S . . and Dayton, M. A Anal. Chem. 1978 50. 1410. Hepler B. R . Weber. S . G. and Purdy W. C. Anal. Chim. Acra 1978 102 41. Fenn. R . J. Siggia S . and Curran D. J . Anal. Chem. 1978. 50. 1067. Armentrout D. N. McLean J . D. and Long M. W Anal. Chern. 1979 51 1039. Stulic K . Pacakova. V and Starkova B. J . Chromatogr., 1981 213 41. Weisshaar D. E . Tallman D. E. and Anderson. J . L. Anal. Chem. 1981. 53 1809. Takata Y . and Muto G .Anal. Chem. 1973 45. 1864. Kreuzig F and Frank. J . J . Chromatogr. 1981 218 615. Davenport R . J . and Johnson D . C. Anal. Chem. 1974,46, 1971. Buchta R . C . and Papa. L. J . J . Chromatogr. Sci. 1976 14, 213. Riggin R . M. and Kissinger P. T . in Kissinger P. T. Editor, "Laboratory Techniques in Electroanalytical Chemistry," Marcel Dekker New York 1975 p. 000. Peterson W. M. Am. Lab. December 1979. 69. Hanekamp H. B Bos P and Frei R. W. J . Chromatogr., 1979 186 489. Lown J. A . . Koile R and Johnson D . C . Anal. Chim. Acru. 1980 116 33. Yamada H. J . . and Matsuda H . J . . Elecrrounal. Chem. 1973. 144 189. Fleet B . and Little C. J. J . Chromatogr. Sci. 1974 12 747. Kissinger P. T . Refshauge C. J. Dreiling R. and Adams, R . N Anal. Lett.1973 6. 465. Stulic K. and Pacakova V. J . Chromatogr. 1981 208 269. Brurlt K. and Bruins C. H. P. J. Chrornatogr. 1979,172,37. Blaedel W. J . Olsen C. L and Sharma L. R. Anal. Chem., 1963 35 2103. Stulic K. and Pacakova V. J . Chromatogr. 1980 192 135. Oosterhuis B. Brunt K. Westerink B. H. C. and Doorn-bos. D. A Anal. Chem. 1980 52 203. 236. 237. 238. 239. 240. 241. 242. 243. 244. 245. 246. 247. 248. 249. 250. 251. 252. 253. 254. 255. 256. 257. 258. 259. 260. 261. 262. 263. 264. 265. 266. 267. 268. 269. 270. 271. 272, 273. Takata Y . and Fujita K. J . Chromatogr. 1975 108 255. Lankelma J . and Poppe H. J . Chromafogr. 1976 125 375. Taylor L. R . and Johnson D. C Anal. Chem. 1974.46,262.DeAngelis T. P. Hurst K. W. Yacynvch. A . M. Mark. H. B. Heineman W. R and Mattson J . S . And. Chem 1977. 49 1395. Pinkerton T. C Halizadeh. K Deutsch E . and Heineman, W. R . Anal. Chem. 1980 52 1542. Schieffer G. W Anal. Chem. 1979 51 1573. Schieffer G. W. Anal. Chem. 1981 53 126. Eggli R. and Asper R . Anal. Chim. Acra 1978 101. 253. Schieffer G . W. Anal. Chem. 1980 52 1994. King W. P. and Kissinger P. T . Clin. Chem. 1981,26. 1484. Brunt K. and Bruins C . H. P. J . Chromatogr. 1978 161, 310. Fenn R. J . Siggia S . and Curran D. J . Anal. Chem. 1978. 50 1067. Roston D . A and Kissinger. P. T . Anal. Chem. 1981 53, 1695. Blank C. L. J . Chromarogr. 1976 117. 35. Roston D. A . and Kissinger P. T. Anal. Chem. 1982. 54, 429. Bratin K. B. and Kissinger P. T. J. Liq. Chromutogr. 1981, 4 321. MacCrehan W. A . and Durst R . A Anal. Chem. 1981,53, 1700. Beauchamp R . Boinay P. Fombon J . J . Tacussel J., Breant M. Georges T. Porthault M. and Vittori 0 J . Chromatogr. 1981 204 123. Goto M. Nakamura T. and Ishii D. J . Chromatogr. 1981, 226 33. Mayer W. J. and Greenberg M. S . J . Chromatogr. Sci., 1979 17 614. Swartzfager. D. G . Anal. Chem. 1976 48 2189. Dieker J . W. Van der Linden W. E. and Poppe. W. E., Talunta 1979 26 51 1 . Hanekamp H. B Voogt W. H . . Bos P. and Frei R. W J. Liq. Chromatogr. 1980 3 1205. Hackman W. R . and Brooks M. A . J . Chromatogr. 1981, 222 179. Schieffer. G . W. ~ J . Chromatogr. 1981 202 405. MacCrehan W. A . . Anal. Chem. 1981 53 74. Hanekamp H. B. Voogt W. H. Frei R . W and Bos P., Anal. Chem. 1981 53 1362. Kemula W. and Kutner M. J . Chromatogr. 1981,204 131. Maitoza P. and Johnson D . C . Anal. Chim. Acta 1980,118, 233. Samuelsson P. O'Dea J. and Osteryoung J . Anal. Chern., 1980 52 2215. Flanagan R . J . Storey G . C. A . Bhamra. R . K. and Jane I., J . Chromatogr. 1982 247 15. Shoup R. E. "Recent Reports on Liquid Chromatography with Electrochemical Detection," Bioanalytical Systems West Lafayette IN. 1980. Felice L. J . Bruntlett C. S . Shoup R . E. and Kissinger, P. T. Nut. Bur. Stand. ( U . S . ) Spec. Publ. 1979 No. 519,391. Kissinger P. T. Bratin K. King W. P Rice J. R . ACS Symp. Ser. Pestic. Anal. Methodol. 1980 136 57. Brunt K. Bruins C. H. P. and Doornbos D . A . A n d . Chirn. Acta 1981 125 85. Bryson G. Clin. Chem. 1979 25 1188. IUPAC Pure Appl. Chem. 1979 51. 1175. Lloyd J. B. F. J . Chromatogr. 1983 256 323. Puper A4113 Received January 9th 1984 Accepted January 19th 198

ISSN:0003-2654    

DOI:10.1039/AN9840900677

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

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

ISSN:0003-2654    

DOI:10.1039/AN9840900699

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST. JUNE 1984 VOL. 109 703 lnverse Photometric Detector Based on Eriochrome Black T for Trace Metal Determination by High-performance Liquid Chromatography Philip Jones Philip J. Hobbs and Les Ebdon Department of Environmental Sciences Plymouth Polytechnic Drake Circus Plymouth Devon PL4 8AA, UK A simple post-column reactor based on Eriochrome Black T (EBT) has been developed to determine trace metals after separation on a silica-based cation-exchange column. Quantitative detection is performed by measuring the decrease in the absorbance of the post-column reactant. Quantitative analysis of nine metals in 16 min can be achieved. Linear calibrations up to 10 p.p.m. for cadmium cobalt copper indium iron(lll), iron(ll) magnesium manganese nickel and zinc are possible with limits of detection of 2-5 ng.The relationship of the cation ion-exchange chromatography and the organic acid eluent to the post-column reactor is discussed The system was used to determine cobalt copper iron manganese and nickel in standard alloys and the accuracy of the method is shown to be very good with coefficients of variation in the range 1-1 1%. Keywords lnverse photometric detection; trace metal determination; Eriochrome Black T; high-performance liquid chromatography; alloy analysis A large number of investigations now require the determina-tion of several elements in a single sample. In many instances the sample analysis time will be excessive if the elements have to be quantified individually. Thus there has been an increasing demand for sensitive techniques capable of rapid multi-element determinations.Atomic-emission spectro-scopic techniques have always been capable of simultaneous multi-element analysis and highly sophisticated inductively coupled plasma (ICP) instruments are now available for the determination of as many as 80 elements. An alternative approach involving much less capital cost and offering more scope for the investigation of chemical species is the use of liquid chromatographic techniques for rapid sequential multi-element determination. Until recently little attention had been focused on this area, particularly for the determination of groups of metals as classical liquid chromatography processes using columns were very slow and no convenient sensitive on-line metal detection systems were available.The development of highly efficient column packing materials for use in high-performance liquid chromatography (HPLC) systems has considerably increased the speed of liquid chromatographic analysis and lately several workers have demonstrated the feasibility of using HPLC instrumentation for the rapid sequential analysis of groups of metal cations. Most papers describe the use of ion-exchange systems to separate metal ions in aqueous media using polyfunctional carboxylic acids as eluting agents. Takata and Fujita' des-cribed the separation of seven metals on small particle size polystyrene-based cation-exchange resins using tartaric acid as the eluting system. Cassidy and Elchukz used a citrate eluent to compare column packing materials for the separation of metal ions and their subsequent applicability to trace enrich-ment methods.Metal detection systems still allow considerable scope for development and two approaches in particular have been successful. One involves electrochemical detection of metal complexes using coulometric principles3 and the other photo-metric detection of metal complexes.4 The latter approach appears to offer the most versatility considering the large amount of information available on the photometric determi-nation of trace metals. Metal cations separated on ion-exchange material in aqueous media will elute from the column generally as carboxylic acid complexes with little or no absorbance in the UV - visible region. These metal species can be converted into highly absorbing complexes by introducing a chromophoric reagent at the column outlet to mix with the effluent.After allowing sufficient time for reaction the mixture is passed through the detector and the absorbance of the metal complexes monitored at a specified wavelength. Post-column reactor systems although adding to the complex-ity of the apparatus allow a tremendous range of reagent type and conditions to be used and the determination of most metals if chromatographic systems can be developed to separate them. Many of the systems reported in the literature measure the absorbances of the metal complexes formed in the post-column reactor. Unfortunately for a particular group of metal complexes the wavelength maximum can vary over a wide range and it is difficult to choose a compromise wavelength to achieve optimum sensitivity for every metal.Another more novel approach is to monitor the decrease in the absorbance of the reagent as each metal reacts with it. This procedure was pioneered in our laboratories and a dithizone post-column reactor was reported that gave a good quantita-tive performance for six separated metal species.4 For good sensitivity and low detection limits monitoring based on measurement of the decrease in absorbance of the photo-metric reagent requires that the reagent itself absorbs very strongly and dithizone is ideal for this purpose. Dithizone is particularly suitable for the determination of the sulphophilic metals but reacts slightly or not at all with many of the more oxophilic metals.It was therefore considered that a more general metal detection system was needed and a number of chelating agents were investigated for use in post-column reactors. Eriochrome Black T (EBT) looked particularly promising as it reacts with a wide range of oxophilic and sulphophilic metals and is itself strongly absorbent at 610 nm. This paper describes the performance of an EBT post-column reaction system when used for the determination of a group of metals separated by cation exchange. Two certified reference metal alloy samples were analysed and the accuracy and precision of the detector assessed for both the major and minor constituents. Experimental Apparatus A typical isocratic high-performance liquid chromatographic system was used (Fig.1). A high-pressure pump (Consta-metric 111 Laboratory Data Control Riviera Beach FL, USA) at a constant flow-rate was connected via an injection valve (Rheodyne Model 7125 Rheodyne Cotati CA USA 704 ANALYST. JUNE 1983 VOL. 109 1 Injector Column -/-I _ I I I I [ E B L r U Fig. 1. Schematic diagram of the inverse photometric system, showing A analytical column pump; B post-column reagent pump; C microprocessor; D the pressure restrictor; and E post-column reactor coil. Solid lines represent liquid flows and dotted lines electrical connections fitted with a 100-pI loop to an analytical column (250 mm X 4.6 mm i.d. stainless-steel tubing). The effluent from the analytical cation-exchange column (Whatman Partisill0 SCX Whatman Ltd.Maidstone Kent) mixed with the post-column reactant at 180 "C with the greater outlet flow adjacent to the inlets. The reactant was pumped (Model AA Dual Piston Eldex pump Eldex Laboratories, Menlo Park CA USA) through a pressure restrictor to reduce the pump pulsations. A reactor coil of 0.6 ml capacity and 0.5 mm i.d. connected the T-junction to the double-beam UV - visible spectro-photometric flow monitor (Model SF770 Schoeffel Instru-ment Corp. Westwood NJ USA). The 1-cm length flow cell of 10 p1 capacity was monitored at 610 nm. Reagents Analytical-reagent grade reagents (BDH Chemicals Poole, Dorset) were used throughout. Stock Eriochrome Black Tsolution 0.4gper 100 ml of 0.1 M ammonia solution. Stock lactic acid solution 2 M . Prepared by mixing the acid with distilled water at this concentration so that biological growths did not form.Sodium hydroxide (2 M) - ammonia (2 M) solution (2 + 3 ) . Used for pH adjustment. Stock metal solutions. AnalaR grade metal (1 g) was dissolved in nitric acid to give a resultant 1000 p.p.m. solution in a 1 M acidity. Preparation of Samples and Standards The sample solutions were prepared by dissolving 1 g of the reference material (Bureau of Analysed Samples Ltd., Middlesborough) in a mixture of distilled water and AnalaR nitric acid (1 + 1). The resulting mixture was then made up to 1 1 in a calibrated flask with sufficient nitric acid to give a 1 M concentration. Mixed metal standard solutions were prepared to give metal concentrations in the ratio similar to that given in the certificates of analysis.Dilutions were prepared from these standards as required. Analysis of Magnetic Alloy This was analysed in two series of injections one at 100-fold dilution (10 p.p.m. nominal) for the major constituents copper iron(III) nickel and cobalt and one at 20-fold dilution (50 p.p.m. nominal) for the manganese determination. The resultant chromatograms were compared with those obtained from mixed metal standards diluted to give similar sized peaks. The microprocessor was programmed to change the amplification immediately after the cobalt peak was eluted (8 min) . Analysis of Monel Alloy This was analysed in a similar way to the above with injections of the stock solution diluted 100-fold. Injections of more concentrated solutions were not necessary to determine cobalt the minor constituent as sufficient sensitivity was available.The microprocessor was programmed to change the amplification just before the cobalt peak was eluted (6.2 min). Results and Discussion Chromatography Cation-exchange chromatography was considerably improved by the use of polyfunctional carboxylic acids in aqueous solution as eluents. Retention times were much shorter and the separation of ions with similar size and charge was enhanced. The speed of elution depended upon the fraction of the metal cation complexed with the carboxylic acid which in turn depended on the magnitude of the metal stability constants and the concentration of the carboxylate anion in the eluent. The capacity of the cation-exchange material was also of prime consideration high-capacity phases need carboxylic acids with larger metal stability constants to maintain short retention times.Thus cation-exchange materials based on high-capacity cross-linked polystyrene resins tend to be used with the more strongly chelating tartaric and citric acids.l.2 For this work maximum efficiency was needed to try and separate up to ten metal species in as short a time as possible. Therefore silica-based phases were investigated as they are more efficient than the resin-based for a similar particle size. Partisil 10 SCX is a 10-pm particle size silica phase with surface-bonded strong cation-exchange groups. The capacity of this silica phase is much lower than the Aminex resin used in previous work4 and so carboxylic acids such as tartaric and citric acids gave elution times that were too short and close to the solvent front to be of any value.A number of weaker chelating acids were studied namely acetic formic lactic and succinic acids. Lactic acid was found to give particularly good results with the Partisill0 SCX material. The concentration of the lactate ion critically affects the elution time and therefore, the concentration can be used to optimise the chromato-graphy in contrast to the other acids investigated where the concentration effect is smaller. Like most carboxylic acids, lactic acid is a weak acid and the concentration of the lactic anion depends on the pH. Therefore for a given total concentration of acid precise control of elution speed was achieved by pH adjustment.For this work two nominal concentrations were used namely 0.05 and 0.1 M and the pH was adjusted to between 3 and 4. The actual conditions chosen depended on the number of metal species to be separated and the capacity of the particular batch of Partisill0 SCX. Batch to batch variations in the manufacture of high efficiency ion-exchange materials can still be significant and it was found that a new batch sometimes required a different pH or even a different nominal concentration of lactic acid to achieve the same separation as the previous batch. The temperature of the chromatography column was a further consideration as in general temperatures above ambient improved column ef-ficiency and peak symmetry.A temperature of 40 "C proved a reasonable compromise for the Partisil 10 SCX column, because although higher temperatures would increase ef-ficiency a little more this would be at the expense of reduced column life resulting from the increased solubility of the silica matrix. Fig. 2 shows a chromatogram for nine metals obtained in 16 min. using 0.05 M lactic acid as eluent at a pH of 3.6. The column flow-rate was 1.5 ml min-1. The pH of the reactant mixture flowing through the detector cell was 9.9 and the absorbance 0.75 at 610 nm. Figs. 3 and 4 show the chromato-grams obtained from the analysis of Magnetic and Monel Alloys respectively using 100-pl injections of 10 p.p.rn ANALYST JUNE 1984 VOL. 109 705 cu -e(lll) Fe(l1) Mn I I 0 5 10 15 20 Tim e/m in Fig.2. Chromatogram ob-tained using a lactate eluent with a 100-pl injection of a solution containing 0.75 p. .m. of cobalt copper ironf;II), iron( II) magnesium man-ganese nickel and zinc and 4 p.p.m. of cadmium. Column temperature 40 "C; post-column reagent flow-rate 0.5 ml min-1; reactor coil 6 ml; and full-scale deflection 0.2 absorbance units I + Ni I CO Mr 0 5 10 Ti me/m in Fig. 3. Analysis of the Per-manent Magnetic Alloy BCS No. 384 Hycomax 111. 0.1 M lactate eluent pH 3.5 flow-rate 1.5 ml min-1; a post-column addition 0.7 ml min-*; absorbance 0.77 at 610 nm; and full-scale deflection 0.2 absorbance units amplified to 0.05 full-scale deflection for manganese nominal solutions. The conditions were the same as described in Fig.2 except that the lactic acid concentration was increased to 0.1 M to allow for the higher capacity of a new batch of Partisil 10 SCX. The detector sensitivity was set initially to 0.1 full-scale deflection for the Monel Alloy and changed to 0.05 absorbance using the microprocessor just before the cobalt and manganese were eluted. Similar sensitivity changes were also used for the Magnetic Alloy with amplification before the manganese was eluted but this time the manganese peak was too close to the background noise for accurate results; therefore a separate manganese determination was carried out using a 50 p.p.m. nominal alloy solution (Fig. 3). 0 cu Mn 5 10 Ti me/mi n Fig. 4. conditions as in Fig. 3. Absorbance 0.54 at 610 nm Chromatogram of the Monel Alloy 400 using the same Detector Performance Eriochrome Black T is one of the most frequently used metallochromic indicators for EDTA titrations and is known to react with aluminium cadmium cobalt copper indium, iron(II) iron(III) lead magnesium manganese mercury, titanium zinc and the platinum metals.Within the pH range 7-11 the above metal ions produce a colour change from blue to red. Preliminary investigations revealed that although the indicator reacted with many of the above metals both the molar absorptivities and the wavelength maxima of the resulting complexes varied over a wide range and so as discussed previously the eluted metals were detected by measuring the decrease in the absorbance of the EBT reagent.The most important parameters affecting detector perfor-mance were found to be the EBT concentration lactic acid concentration and pH 70h The detector system measures the decrease in absorbance of the EBT reagent as it reacts with a metal species and therefore a high EBT concentration is necessary to achieve a wide linear range. If the EBT concentration is too high for example if the base line absorbance is above 2 then the low-intensity radiation falling on the photocell will result in increased noise and hence poorer detection limits. However before this point was reached in the investigation. more serious noise problems arose from the pulsing of the pumps. This showed itself as a complex sine wave superimposed on the base line becoming particularly amplified above a base line absorbance of 1 with a consequent worsening of detection limits.If a wide linear range was not needed it might be expected that lower detection limits could be achieved by reducing the EBT concentration as this will result in proportionally less pump noise. This was indeed found to be the case but only down to an EBT concentration equivalent to a base line absorbance of 0.2. Below this absorbance detection limits started to worsen again as the metal response began to decrease. One possible explanation for this is that although metal lactate complexes have much lower stability constants than metal EBT com-plexes if the EBT concentration drops too low then the lactate will begin to compete more effectively for the metal ion thus reducing the sensitivity of the EBT reaction.Although different metals exhibited maximum sensitivity at different pH this effect was remarkably small and optimum response for the ten metal species investigated was found to be between pH 9.5 and 10.5. Above pH 10.5 reduced response was almost certainly due to the decrease in the molar absorptivity of the EBT peak at 620 nm. Below pH 9.5 the lower response was again due to the lactic acid effect where reduced dissociation of the EBT makes it more difficult to compete with the lactate ion. One further factor to be considered is the effect of the mixing ratio of the post-column reagent and column eluent on the detector performance. Increasing the percentage of the post-column reagent will dilute the lactic acid and so reduce any limiting effect on the metal EBT response.If the percentage addition becomes too large however this ben-eficial effect will be offset by the increased dilution of the metal peaks eluting from the column. Taking the above factors into account the following conditions were found to be a reasonable compromise, whereby low detection limits were achieved without excessive sacrifice in the linear concentration range. The column flow-rate was set to 2 ml min-I and the post-column reactor flow-rate to 0.7 ml min-1. Using an EBT solution comprising 6 ml of stock solution and 100 ml of 2 M ammonia solution made up to 250 ml the base line absorbance will be between 0.6 and 0.9 and the pH between 10 and 10.5. These conditions gave good linear calibrations up to 10 p.p.m.for cobalt cadmium indium iron(III) iron(II) magnesium, manganese nickel and zinc with absolute detection limits between 2 and 5 ng. Alloy Analysis For each alloy a 1-g sample was dissolved and after appropriate dilution eight replicate injections were carried out interspersed with mixed metal standards. The average concentration ("/o mim) was then calculated for each element together with the coefficient of variation. The results for the Hycomax I11 Permanent Magnetic Alloy analysis are shown in Table 1. The values found agree very closely with the certificate values and this is particularly encouraging as it applies to both major and minor elements whose concentra-tion ranges vary over nearly three orders of magnitude. There was no certificate value for iron and so the value was calculated by difference.The value found was lower but this is not surprising because all the impurities in the alloy may not be accounted for in the certificate. The precision as shown by the coefficients of variation is acceptable and as expected Table 1. Analysis of Hycomax I11 Permanent Magnetic Alloy (BSC No. 384) Coefficient Ce I t i f i e d Found. 7" of variation.* value. '1% Metal m/m "/O m!tn Cobalt . . . . . . 33.7 1.3 33.7 Copper . . . . 3.06 5.2 3.06 Iron . . . . . . 34.6 2 . 3 I Manganese . . . . 0.10 11.5 0.10 Nickel . . . . . . 14.8 3.3 14.6 * Relative standard deviation of 8 replicate injections. i- Iron has no certified value but the value calculated by difference was 35.6% Table 2.Analysis of Monel Alloy 400 (BSC No. 363!1) Coefficient of Found variation,* Metal YO mlm O/O Cobalt . . . . . . 0.033 8.3 Iron . . . . . . 1.73 9.2 Nickel . . . . . . 64.2 1.6 Copper . . . . 31.5 3.0 Manganese . . . . 1.17 9.4 * Relative standard deviation of 8 replicate injections. Certified value Yo tri 'm 0.032 1.86 1.26 31.9 64.7 becomes progressively poorer as the signal approaches the base-line noise for the elements with the lowest concentra-tions. The results for the Monel 400 Alloy are shown in Table 2. The percentage values found are not as close to the certificate values as obtained for the Hycomax I11 Alloy and this could be due to the coefficients of variation being generally higher, particularly for the major constituents.Nevertheless the results from this limited study clearly show the high stability of the post-column reactor system and that reasonable accuracy and precision can be obtained for a group of elements from the same sample injection whose concentrations may range over more than three orders of magnitude. Conclusion The EBT post-column reactor was simple to set up very stable in operation and as EBT and its metal complexes are water soluble no special precautions were necessary to prevent precipitation as with the dithizone post-column reactor.4 EBT does deteriorate slowly in solution owing to oxidation but the making up of post-column reactant solution daily seemed sufficient to maintain good detector performance. Calmagite, an analogue of EBT with the nitro group absent is much more stable and could be used in its place if very long-term stability is required.As presented in this study the quantitative performance of the detector system was very good and shows that the determination of metal species by monitoring the decrease in absorbance of the reagent is a useful concept and could be of general applicability. One particular advantage of this inverse photometric system is that all metals that react with the same stoicheiometry will give approximately the same molar response. Further any metal species that degrades or destroys the reagent rather than form a complex will also be detected. A disadvantage of inverse photometric detection is th ANALYST. JUNE 1983. VOL. 100 707 increased base-line noise as the system tends to amplify pump pulsations. This noise can be reduced by using pulse-free non-reciprocating pumps or alternatively very fast recipro-cating pumps where the pulse frequency is too fast for the detector to respond. The EBT post-column reactor system should be capable of determining many more metal species than those investigated here including aluminium titanium and the platinium metals. This will depend on the development of other chromato-graphic systems which are at present under active investi-gation. References 1. 2. 3. 4. Takata Y . and Fujita K. 1. Chromatogr. 1975. 108 25.5. Cassidy R. M. and Elchuk S . 1. Chromatogr. Sci. 1980 18, 217. Girard J . E. Anal. Chem. 1979 51 836. Jones. P. Hobbs P. J . and Ebdon L Anal. Chim. Actu, 1983 149 39. Paper A3141 4 Received November 28th 1983 Accepted January I6th I98

ISSN:0003-2654    

DOI:10.1039/AN9840900703

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST. JUNE 1984. VOL. 109 Distribution of Zinc Amongst Human Serum Globulins Determined by Gel Filtration - Affinity Chromatography and Atomic-absorption Spectrop hotometry John W. Foote and H. Trevor Delves 709 University Department of Chemical Pathology and Human Metabolism Southampton General Hospital, Southampton SO9 4XY UK Gel-filtration chromatography was combined with affinity chromatography (for albumin) and electrothermal atomic-absorption spectrophotometry to investigate the distribution of zinc amongst normal human serum globulins. Most of the zinc associated with serum globulins appeared to be bound to a2-macroglobulin and the observed zinc content of a2-macroglobulin was in very good agreement with that found using more complex techniques. No zinc was found in association with transferrin and the recovery of globulin-bound zinc was quantitative 104%.Keywords Zinc determination; serum globulins; affinity chromatography - gel-filtration chromatography; atomic-absorption spectrophotometry with electrothermal atomisation Appproximately 98% of the zinc in serum is bound to proteins.' Although it is widely accepted that this zinc is associated principally with albumin and a2-macroglobulin ,2 there are conflicting reports as to both the relative distribution of the metal between these protein+ and also its association with other protein species.69 The distribution of zinc amongst the serum proteins is of clinical relevance as low serum concentrations of zinc are found in many conditions in which no deficiency of the metal is suspected.'O Experiments in which the zinc flux through the plasma of stressed pigs was measured suggest that these reduced serum zinc concentra-tions may represent a normal physiological response to stress rather than an increased metabolic requirement for zinc,ll although it is also reasonable to presume that on many occasions these findings simply result from changes in the concentrations of the zinc-binding serum proteins.We have recently described a method that allows the rapid and simple determination of the human serum albumin-bound and globulin-bound zinc fractions which are separated by pseudo-ligand affinity chromatography on immobilised Ciba-cron Blue F3G-A.12 Two fractions are collected during this process the first containing globulins which elute spon-taneously from the Cibacron Blue column the second containing albumin which is subsequently released using thiocyanate.12 We have further evaluated the method by investigating the distribution of zinc amongst the globulin species present in samples of serum obtained from a normal human subject.In this study we submitted the serum globulins that were eluted spontaneously from the affinity column to a further separation procedure using gel filtration on Sephadex G-100. Experimental Apparatus The following apparatus was used for the separation of serum proteins a Varioperpex I1 Pump 2120 (LKB Produkter, Bromma Sweden); a 2138 Uvicord S fitted with filters for use at 280 nm and a 2138-100 flow cell incorporating a 5-mm light path (LKB Produkter) as ultraviolet (UV) monitor; a 2112 Redirac (LKB Produkter) for the collection of fractions of effluent from the chromatography columns; and a sample valve consisting of two four-way slide valves Altex Type 201-02 (Anachem Luton) fitted with a laboratory-made 2.0-ml sample loop.Affinity chromatography for albumin was performed on a 20 x 1.0 cm i.d. glass column Model MS-PC 1020 (Whatman Maidstone) and serum globulins were separated by gel filtration on a 95 X 2.6 cm i.d. glass column, Model 2187 (LKB Produkter). A Perkin-Elmer Model 2380 atomic-absorption spectropho-tometer was used with a Perkin-Elmer hollow-cathode lamp for the determination of zinc concentrations in the fractions obtained following the chromatographic procedures. The instrument was fitted with an AS-40 autosampler and an HGA-500 graphite furnace using standard (non-coated) fur-nace tubes.All samples tubes autosampler cups and other apparatus were made free of zinc using previously described proce-dures. 12 Reagents All reagents were of AnalaR grade (BDH Chemicals Poole) unless otherwise stated. Blue Sepharose CL-6B. Obtained from Pharmacia Fine Chemicals Hounslow. Octanoic acid. Specially pure grade. Sephadex G-100 (40-120 pm). Obtained from Pharmacia Sodium chloride. Sodium hydroxide 4 M AVS standard solution. Sodium dihydrogen orthophosphate dihydrate. Disodium hydrogen orthophosphate. Sodium thiocyanate. Laboratory-reagent grade. Tris( hydroxymethy1)rnethylamine. Fine Chemicals. Procedure Whole blood was obtained from an apparently healthy subject and the serum separated using the previously described technique that was designed to prevent contamination with exogenous zinc.12 Sephadex G-100 was swollen and made free of zinc using the method published by Gardiner et a1.6 Previously described methods12 were used for the preparation of Blue Sepharose CL-6B and for the preparation of a starting buffer solution consisting of 0.05 M sodium chloride in 0.05 M tris(hydroxy-methy1)methylamine - hydrochloric acid at pH 7:t and eluting and regeneration buffers containing 0.2 M sodium thiocyanate and 0.05 M sodium octanoate respectively in 0.05 M phos-phate buffer solution at pH 7.4. The apparatus was connected as shown in Fig. 1. Blue Sepharose and Sephadex were packed to heights of 6 and 8 710 ANALYST JUNE 1984 VOL.109 Sephadex Cibacron blue column column Sample valve I I I 1 ,--J-- ii Fractions Sample Buffer Fig. 1. proteins Schematic diagram of apparatus used for separation of serum cm. respectively and the resins were equilibrated by pumping 1 1 of starting buffer solution through the system. A 2.0-ml sample of serum was loaded into the sample loop and introduced into the columns in a stream of starting buffer solution at a flow-rate of 0.27 ml min-1. The passage of human serum through a bed of Blue Sepharose CL-6B under these conditions resulted in a virtually complete retention of albumin which is immobilised on the resin." The first 100 ml of buffer solution to be eluted from the Sephadex column were discarded and 2.7-ml fractions of the effluent were collected thereafter.When the absorbance of the effluent at 280 nm indicated that the elution of protein from the Sephadex column was complete the affinity column was removed from the system the retained albumin was released from the column and the Blue Sepharose was regenerated using previously published techniques. 17 The gel-filtration column was repacked with zinc-free Sephadex before further use. The fractions obtained following affinity chromatography -gel filtration were analysed for zinc by electrothermal atomic-absorption spectrophotometry as described.' using calibra-tion standards containing bovine albumin in order to over-come protein-related matrix interferences on zinc sensitiv-ity.13 Albumin and cx2-macroglobulin were determined in the fractions of effluent obtained from the columns using the previously described kinetic immunoturbidimetric tech-nique. 17 Immunoglobulin-G and transferrin were measured using simple modifications of this method. Results Protein Separation The elution profiles of the proteins of interest together with the absorbance at 280 nm found in the effluent obtained from the Sephadex column are shown in Fig. 2. The small albumin peak represented only 3.1% of the albumin present in the sample of serum applied to the Blue Sepharose column indicating a very good retention of albumin by the affinity column despite the very high loading with 2.0 ml of serum. A good separation of the remaining proteins was obtained.Elution of Zinc The elution profile of zinc found in the fractions of effluent that were collected is also shown in Fig. 2. Virtually all of the zinc recovered from the Sephadex column was eluted as a single peak which co-eluted with a,-macroglobulin. A small amount of zinc approximately 3% of the zinc recovered from f L 0.3 0.3 I 120 135 150 165 180 -&-0.75 r I -0.50 . 5. 0.25 ,! N 0 Elution volume/ml Fig. 2. Elution profiles of zinc albumin a,-macroglobulin (a,M), immunoglobulin-G (IgG) transferrin (Tf) and the absorbance of the effluent at 280 nm (A280) obtained after affinity chromatography - gel filtration. The concentrations of albumin a,-macroglobulin, immunoglobulin-G and transferrin were multiplied by 0.1 1 0.2 and 1 respectively before the data were plotted.The retention of albumin by Blue Sepharose was good no more than 3.1% of the albumin content of the sample being detected in the effluent from the gel-filtration column the gel-filtration column was eluted as a second peak that did not co-elute with any of the proteins under scrutiny (Fig. 2). No zinc was detected in fractions of effluent that contained transferrin or surprisingly in those fractions which contained the small amount of albumin that was carried over from the Blue Sepharose column. Using published methods,12.13 the total albumin-bound and globulin-bound zinc concentrations in the serum sample submitted for analysis expressed as means (95% confidence limits. number of determinations) were determined to be 18.5 pmol 1-1 (17.9-19.1 pmol 1-1 6) 14.2 pmol 1-1 (13.5-14.9 pmol 1-1 6) and 3.5 pmol 1-1 (3.3-3.7 pmol 1 - 1 6) respectively the recovery of zinc being 96% of the total content of the sample.The recovery of globulin-bound zinc from the gel-filtration column was 104%. Discussion Although the association of serum zinc with albumin and cxz-macroglobulin is widely accepted,2-10 there is no firm consensus of opinion as to the binding of zinc amongst other serum proteins. Difficulties may arise during an investigation into the distribution of zinc amongst the serum proteins if the delicate interaction between the metal and its ligands is disturbed as may occur during the performance of some procedures for protein separation or when inappropriate samples such as heparinised plasmal2.15 are submitted for analysis.The conflicting reports present in the literature must be viewed in the context of these considerations. Early reports of investigations using electrophoretic methods for protein separation described an association of zinc with a variety of globulin species.7.8.14 These results. however cannot be accepted without reservation as electro-phoretic procedures have been shown to disrupt the binding of zinc to serum proteins.' Using gel filtration on Sephadex G-150 combined with ion-exchange chromatography on DEAE-Sephadex to separate samples of heparinised plasma prepared from portal venous blood Evans and Winter9 observed an apparent association between transferrin and a significant proportion of the plasma zinc.These findings may have resulted from the displacement from albumin of a part of its bound zinc a process that has been shown to result from the use of DEAE-Sephadex protein separations16 or the use of plasma samples prepared with heparin. 17.1' More recently ANALYST. JUNE 1984 VOL. 109 71 1 Table 1. Comparison of the distribution of zinc in serum found using affinity chromatography on Blue Sepharose CL-6B” and that reported following polyethylene glycol precipitation4 and sucrose density gradient centrifugation’ Sucrose density chromatography” precipitation4 centrifugation5 Affinity Polyethylene glycol gradient Parameter ( n = 16) ( n = 28) ( n = 10) Totalserumzincipmoll . . . . 12.6 5 2.6* 12.9 It 1.4* 14.9 k 1.4* Albumin-bound zincipmol 1 .. 10.0 5 2.0 10.5 It 1.2 11.6k 1.5 Globulin-boundzincipmol I- 1 . . 2.4 5 0.3 2.5 k 0.St 3.3 k 0 . l t * Mean k standard deviation. t a,-Macroglobulin-bound zinc. Gardiner et a1.6 have used gel filtration on Sephadex G-100 to investigate the distribution of zinc in samples of human serum. The distribution of zinc obtained was not clear because, although these workers were able to confirm the binding of the greater part of the serum zinc to albumin and az-macroglobulin there additionally appeared to be small amounts of zinc associated with immunoglobulin-G and transferrin. The interpretation of these observations is not easy as Sephadex G-100 has an appreciable affinity for zinc, and its use in investigations such as those described here can result in both unacceptable contamination of serum samples with exogenous zinc,6 and in significant exchanges between zinc bound to the column matrix and that present in the sample.15 In addition the use of gel filtration on Sephadex G-100 under the conditions described results in a poor discrimination between albumin and transferrin because their relative molecular masses are similar 66 000 and 76 000, respectively.17 The procedure adopted in this study was designed to overcome these difficulties. Sephadex was made free of zinc before use by means of the rigorous purification procedures that were found effective by Gardiner et al. 6 The gel-filtration column was repacked with new Sephadex before each run as it is clear that zinc may accumulate in this resin as a result of the uptake of metal from samples submitted for separation or from the passage of large volumes of buffer solution through the column.and this zinc may re-equilibrate with various protein species during subsequent separations. The displace-ment of zinc from the albumin carried over from the Blue Sepharose column (Fig. 2) almost certainly occurred as a result of this process. The passage of serum through Blue Sepharose CL-6B under the previously described conditions results in a virtually complete immobilisation of albumin on the resin with negligible disruption of zinc - albumin interac-tion.12 The prior removal of albumin by chromatography on Blue Sepharose therefore prevents the incorporation of loosely bound albumin-associated zinc into the Sephadex resin and furthermore the separation of albumin from transferrin by gel filtration is no longer necessary.The data obtained in this study suggest that the globulin-associated zinc determined after the separation of serum by affinity chromatography for albumin on Blue Sepharose as previously described,l2 may legitimately be considered to be a2-macroglobulin-associated zinc as this is the only species that binds appreciable amounts of zinc in the globulin fraction. Further support for this hypothesis arises from the very good agreement between the zinc content of the globulin fraction obtained after separations on Blue Sepharose12 and the zinc content of cu2-macroglobulin found by other workers who isolated the protein using polyethylene glycol precipitation3 or sucrose density gradient centrifugation.5 These data are shown in Table 1.The nature of the remaining globulin-associated zinc represented by the second peak in the elution profile shown in Fig. 2 is not clear. It is possible that this zinc was associated with an unknown zinc-binding globulin but it is also possible that this peak represents a proportion of the zinc displaced from the small amounts of albumin that passed through the Sephadex column. Affinity chromatography for albumin on Blue Sepharose CL-6B coupled with electrothermal atomic-absorption spec-trophotometry as previously described,l2 provides a rapid and effective means for the determination of the distribution of zinc between albumin and az-macroglobulin in human serum. We thank Professor B.E. Clayton for her support and encouragement. We are also grateful to BDH Chemicals the Wessex Regional Health Authority’s Research Committee and the Trustees of the Welcome Trust for financial assistance given to one of us (J. W. F,). The atomic-absorption equipment used in this investigation was on loan from the Perkin-Elmer Corp. Norwalk CT USA and Bodenseewerke FRG. 1. 2. 3 . 4. 5 . 6. 7. 8. 9. 10. 11. 12. 13. 14. 1s. 16. 17. References Giroux. E. L and Henkin R. I. Biochim. Biophys. Actu, 1972 273 64. Parisi A. F and Vallee B. L. Biochemistry 1970 9 2421. Dawson J . B. Bahreyni-Toosi M. H Ellis. D. J . . and Hodgkinson. A . . Analyst 1981. 106. 351. Giroux. E. L. Durieux. M. and Schechter. P. J . . Rioinorg. Chem. 1976 5 21 I . Song M. K . and Adham N . F. Clin. Chim. Acta 1979 99, 13. Gardiner P. E. Ottaway J . M. Fell. G . S . and Burns R. R., Anal. Chim. Acta 1981 124 281. Prasad. A. S and Oberleas D J . Lab. Clin. Med. 1970. 76, 416. Boyett. J . D . and Sullivan. J . F. Metaholisrn 1970 19. 148. Evans G. W. and Winter. ‘I. W Riochem. Biophps. Res. Commun. 1975 66. 1218. Falchuk K. H. N . Engl. J. Med. 1977. 296 1129. Chesters J . K. and Will M Br. J . Nutr. 1981. 46 119. Foote J . W. and Delves H. T A n a l y t 1983; 108. 402. Foote J . W. and Delves H . T. Analyst 1982. 107. 1229. Dennes E. l’upper. R. and Wormall A . Biochem. J . 1962. 82. 466. Chesters. J . K . and Will M. Br. J . Nutr. 1981 46. 11 1 . Smith K. T. Failla M. L and Cousins. R. J . . Biachem. I . , 1979. 184 627. Peters T in Brown. S . S . . Mitchell. F. L and Young D. S . . Editors “Chemical Diagnosis of Disease.” ElsevieriNorth Holland Biomedical Press Amsterdam 1979. p. 317. Paper A3141 8 Received November 30th 198.3 Accepted January 9th 198

ISSN:0003-2654    

DOI:10.1039/AN9840900709

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST J U N E 1983. VOL. 109 713 Extraction - Atomic-absorption Spectrophotometric Determination of Lead by Hydride Generation in Non-aqueous Media Jose Aznarez Francisco Palacios Juan Carlos Vidal and Javier Galban Department of Anaiyticai Chemistry Faculty of Sciences University of Paragoza Zaragoza Spain A method for lead hydride generation in a non-aqueous extraction phase is proposed. The lead hydride generation is carried out in an aliquot of lead pyrrolidine-I-carbodithioate extract in chlorcform by the addition of sodium tetrahydroborate(ll1) solution in dimethylformamide. Lead is determined by atomic-absorption spectrophotometry at 217.0 nm. The proposed method gives improved sensitivity and eliminates interferences which is necessary because lead hydride generation in aqueous solution occurs in the presence of oxidants such as potassium dichromate or ammonium peroxydisulphate.The method was applied to the determination of lead in BCS standard steels and air particulates. Good accuracy and precision were obtained. Keywords Lead determination; hydride generation; atomic-absorption spectrophotometry; non-aqueous media Difficulties in the lead hydride generation procedure are well known even when using sodium tetrahydroborate(II1) for the determination of lead by atomic-absorption spectro-photometry.1 The causes are the small yield of the volatile hydride and its low thermal stability. In the last few years it has been shown that the yield in the lead hydride generation procedure increases when there are oxidants in solution such as potassium dichromate,2 malic acid (0.3 M) - K2Cr207, HN03 (0.5 M) - Hz02 (lo%) or HN03 (0.2 M) - (NH4)2S208 (1 M ) .~ Yields obtained for the lead hydride generation are between 16 and 60% in these solutions. Detection limits for the determination of lead by AAS (using 10% NaBH4) in the above-mentioned solutions are 3.2 1.7 and 1.1 ng ml-1, respectively.3 Lead hydride generation with NaBH4 at higher than ambient temperature (up to 65 "C) has also been used.4 There are also interferences in the AAS determination of lead with hydride generation.3 At the 0. I-pg level of lead Au, Ag Cd and Cu interfere seriously; As Bi Ge and Sn can also interfere if they are present at concentration levels greater than 5 pg. The interference of up to 100 pg of each element cited can be eliminated by extraction of lead with dithizone in chloroform.Lead is subsequently back-extracted in water where it is determined by AAS following hydride generation with NaBH4. Generally an electrically heated silica tube has been used as the atomising device. A linear peak height has been found with contents of up to 100 ng ml-1 of lead. In this paper a method for lead hydride generation in non-aqueous media and its determination by AAS is pro-posed. Lead was separated and concentrated by extraction with ammonium pyrrolidine-1-carbodithioate or sodium diethyldithiocarbamate into chloroform. An aliquot of the chloroform extract was mixed with NaBH4 solution in N,N-dimethylformamide (DMF) and anhydrous acetic acid. Lead hydride was easily generated and drawn along by the hydrogen simultaneously evolved.The procedure can be carried out in an inexpensive generator described elsewhere.5 The lead hydride was taken directly into the atomisation device of the spectrophotometer without the need for a recollecting system. In addition to the elimination of inter-ferences and lead pre-concentration due to the extraction the AAS signal of lead was about four times higher than that obtained by hydride generation in aqueous media in the presence of 0.5 M HN03 - 10% H202 using the same instrumental parameters. The proposed method has been applied to the determina-tion of lead in BCS standard steels and air particulates and can be applied to the determination of lead in petroleum products without previous mineralisation which will be the subject of a later paper.Experimental Apparaus The following apparatus was used a Pye Unicam SP-9 atomic-absorption spectrophotometer fitted with a three-slot air - acetylene burner; a Perkin-Elmer lead hollow-cathode lamp with 10-mA maximum current; an Orion Research Microprocessor IonAnalyzer 901 for pH measurement a McLeod CA-1 apparatus for airborne particulate samples; a Haaker mechanical shaker; and separating funnels of 100-ml capacity. Reagents and Solutions All chemicals were of analytical-reagent grade. obtained from Merck. Standard lead solution 1000 pg ml-1. Prepared by dissol-ving 1.614 8 g of lead nitrate in distilled water and diluting to 1 1 in a calibrated flask. The working solutions were prepared by diluting this solution immediately before use.Potassium acidphthalate buffer solution 0.05 M andpH 4.0. Prepared according to reference 6. Ammoniiim pyrrolidine-1-carbodithioate solution 0.2% mlV in distilled water. Prepared just before use. Sodium diethyldithiocarbumate solution 0.2% miV in dis-tilled water. Prepared just before use. Sodium tetrahydroborate(III) solution 4% mlV in N,N-dimethylformamide. This solution can be used for 2 weeks. The N N-dimethylformamide was purified by distillation and the fraction boiling between 148 and 150 "C was collected. Ascorbic acid solution 10% mlV in distilled water. Prepared just before use. Anhydrous acetic acid - concentrated sulphuric acid 3 + 1 VIV. Prepared just before use. Procedure Place a suitable aliquot of sample solution (pH 1-3) containing up to 10 pg of lead in a 100-ml separating funnel.Add 2 ml of ascorbic acid solution 5 ml of ammonium pyrrolidine-l-carbodithioate solution and 10 ml of pH 4 buffer solution. Extract with 10 ml of chloroform by mechanical shaking for 5 min. Leave the phases to separate. Place 1 ml of the chloroform extract into the hydride generator and add 2 ml of acetic acid - sulphuric acid (3 + 1). Inject 3 ml of NaBH4 solution in DMF through the septum membrane of the generator and record and measure the AAS peak height of lead at 217.0 nm. Prepare a calibration graph for known amounts of lead using the standard lead solutio 714 ANALYST JUNE 1984 VOL. 109 and by following the same procedure. Carry out blank determinations to check the possible contamination of lead by the reagents used.Results and Discussion Solubility and Stability of NaBH4 Solution in DMF The solubility of sodium tetrahydroborate(II1) in DMF at room temperature is about 6% miV.7 For lead hydride generation in non-aqueous media the most suitable concentra-tion was found to be 4% mlV of NaBHj in DMF. This solution posesses a slight turbidity that does not need to be eliminated by filtration as the solution can be used for up to 2 weeks as confirmed by the peak height obtained during this period using the same solution. Acids for Lead Hydride Generation As in the antimony hydride generation in non-aqueous media,' a miscible acid is needed in the organic phase for lead hydride generation.Anhydrous acetic acid and mixtures of anhydrous acetic acid - concentrated sulphuric acid (1 + 1,2 + 1 and 3 + 1) have been studied. The best results as shown by the atomic-absorption signals and the miscibility with the solvents were found using the 3 + 1 mixture. The volatile lead hydride was generated immediately, together with the hydrogen evolved and reduction was complete in less than 30 s. Therefore there was no need for collection devices and the hydride was carried directly into the atomising system of the atomic-absorption spectro-photometer using the inexpensive generator described else-where.' A white crystalline precipitate due to the formation of Na2S04 was observed in the organic phase. This precipi-tate or turbidity did not interfere with lead hydride evolu-tion.Optimisation of the Atomic-absorption Signal The optimisation of the atomic-absorption signal with this hydride generator has been discussed in previous papers.'.' Recommended instrumental parameters were as follows: acetylene flow-rate 0.9 1 min-1; air flow-rate 5.5 1 min-1; burner height. 4 mm; wavelength 217.0 nm; spectral slit width 1 nm; and lamp current intensity 10 mA. Obviously better results can be obtained using an electro-thermal atomisation system such as an electrically heated silica tube. but we do not think it necessary. Optimisation of Lead Hydride Generation The conditions required for lead hydride generation in non-aqueous media were studied. The best conditions can be obtained as follows place 1 ml of chloroform extract of lead -pyrrolidine-1-carbodithioate complex into the generator add 2 ml of acetic acid - sulphuric acid (3 + 1) and inject 3 ml of the NaBHj solution in DMF (4% mlV) through the septum membrane of the generator.Lead Extraction Dithizone sodium diethyldithiocarbamate and ammonium pyrrolidine 1-carbodithioate have been used for lead extrac-tion in chloroform. It was found that lead hydride generation was independent of the extraction reagent used. if the conditions were suitable for the quantitative extraction of lead with each reagent. Finally. ammonium pyrrolidine-l-carbodithioate was chosen because of its greater stability in acidic solutions. Conditions for lead extraction with this reagent have been studied previously.8 Calibration Graph Sensitivity and Detection Limit The peak height of the atomic-absorption signal for lead was linear between 0.03 and 1 pg of lead (contained in 1 ml of the chloroform extract) with a regression coefficient R = 0.999 1.The sensitivity (1% absorption) and detection limit using setting 4 for the expansion scale in the spectrophotometer were 3 and 1 ng ml-1 of lead in the chloroform extract, respectively. Ten replicate determinations of 0.5 pg of lead in 1 ml of extract gave an average of 0.48 v g of lead with a relative standard deviation of 3.5%. The peak height obtained by non-aqueous hydride generation for lead was about four times greater than in water - nitric acid (0.5 M) - H202 (10Y0) with the same generator and under the same spectropho-tometer conditions.The pre-concentration and elimination of interferents due to the extraction were other advantages of the proposed method. Interference Study Interferences from different elements in the determination of lead were studied (Table 1). Many ions did not interfere in the determination of 5 pg of lead (0.5 pg ml-1 of lead in the chloroform extract) with 200 mg (giving a ratio of 4 X 104 1) as the interference limit. Some ions such as Mo(VI) V(V) and Fe(II1) interfered in the determination of lead owing to the oxidation of ammonium pyrrolidine-1-carbodithioate in the extraction. These interferents can be easily eliminated by addition of ascorbic acid solution to the aqueous solution before extraction. Those elements which could generate volatile hydride and were partially extracted presented some interference in the determination of lead as shown in Table 1.Applications The proposed method has been applied to the determination of lead in BCS standard steels (Table 2) and atmospheric particulates (Table 3). Steel samples were dissolved as indicated elsewhere.8 The recovery of 5 pg of lead added to spiked sample solutions of BCS 329 mild steel was 98.6% with a relative standard deviation of 2.8% (five determinations). Table 1. Interferents in the determination of lead Ratio of Interference interferent limit"/ to lead Interferent mg (rnim) Ca. Mg Al Cu Zn. Mn(I1). Cr(II1). Fe(II) Nb(V). Co. NOi-. PO4' , SiO,? - . . . . . . . . . . 200 4 x 1 0 4 Ge(IV). Sn(I1). Sn(IV) SOA'- . . 25 5 x 10' Bi(II1).As(1II). As(V) . . . . . . 1 0 2 x 10" Mo( VI). V(V). Fe( 111) . . . . . . 5 1 x 10" Sb(II1). Sb(V) Se(1V). Te(1V). The interference limit is the amount of foreign ion tolerated before an error of not more than 3.S% is caused in the determination of 5 pg of lead. Table 2. Results for the determination of lead in standard steels Amount of lead. '10 Re la t i ve standard Standard steel Certified Found" deLiation. ''c, BCS 329 mild steel . . 0.050 0.051 2.2 BCS417mildsteel . . 0.006 0.006 3.0 BCS 458 mild steel . . 0.016 0.016 2.5 de t e rmin a t i o ns . * Mean and relatike standard deviation for ten replicat ANALYST JUNE 1984 VOL. 109 715 Table 3. Determination of lead in airborne particulates in urban atmosphere Filter location A . . . .. . B . . . . . . c . . . . . . D . . . . . . E . . . . . . F . . . . . . G . . . . . . Amount of lead found/vg 17.35 12.50 13.50 12.45 12.54 10.10 12.60 Volume of air filtered/m3 4.375 5.138 3.742 6.578 4.873 4.005 4.614 Concentration of lead foundipg m-3 3.97 2.43 3.61 1.89 2.57 2.52 2.73 The method has also been applied to the determination of lead in airborne particulates in urban atmosphere. Cellulose filters (37 mm diameter; 0.045 pm) were treated by dissolution with nitric acid - perchloric acid as previously reported.8 The method was checked by the addition of standard lead solution (equivalent to 0.05-1.0 pg m-3 of lead in the atmosphere) to the filter prior to its dissolution. The lead recovery was 98.2% with a relative standard deviation of 3.0% (five determina-tions). References 1. 2. 3. 4. 5. 6. 7 . 8. Godden R. G . and Thomerson D. R. Analyst 1980 105, 1137. Smith R. A t . Spectrosc. 1981 2 155. Jin K. and Taga M . Anal. Chirn. A m 1982 143 229. Jin K. and Taga M. Bunseki Kagaku 1980 29 522. Castillo J. R. Lanaja J. Martinez Ma. C. and Aznarez J., Analyst 1982 107 1488. Meites L. “Handbook of Analytical Chemistry,” McGraw-Hill New York 1963. Aznhrez J. Palacios F . and Vidal J . C. Analyst 1984 109, 123. Aznhrez J. Palacios F . and Vidal J. C. A t . Spectrosc. 1982, 3 192. Paper A31421 Received December 2nd I983 Accepted January lst 198

ISSN:0003-2654    

DOI:10.1039/AN9840900713

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

Fluorimetric Determination of Manganese at the Nanogram Level by Catalytic Oxidation of Pyridoxal 2-Pyridylhydrazone by Hydrogen Peroxide Soledad Rubio, Agustina Gornez-Hens and Miguel Valcarcel Department of Analytical Chemistry, Faculty of Sciences, University of Cordoba, Cbrdoba, Spain The development and optimisation of a fluorimetric reaction rate method for the determination of Mn(ll) is described in detail. The method is based on the enhancement by Mn(ll) of the reaction between pyridoxal 2-pyridyl hydrazone and hydrogen peroxide. Equations for the kinetics of the catalytic and non-catalytic reaction are developed, and the appropriate energies of activation are determined. Using several kinetic methods (tangent, differential, fixed-time and fixed-intensity) micro-amounts of Mn(l1) (0.05-8 ng ml-1) can be determined.The influence of foreign ions on the accuracy of the results is investigated and the method is applied to the analysis of manganese in food samples. The characteristics of the fluorimetric methods are critically compared with those described previously. Keywords : Manganese determination; kine tic flu0 rimetric method; p yridoxal 2-p yrid ylh ydrazon e; dairy products Recently, fluorimetric reaction-rate methods for the determi- nation of inorganic ions have been reviewed. 1 This methodol- ogy permits the development of fluorimetric procedures for the determination of transition metal ions that generally cause fluorescence quenching by the non-radiation deactivation of excited states. Cu(II),2 Co(II),3 Au(1II)J and other ions have been determined by their catalytic effect on the oxidation of azomethine compounds to yield highly fluorescent products.For the determination of metal ions a statistical study of the kinetic catalytic methods proposed has been carried out (Fig. 1). Copper has been the ion most detected, followed by cobalt, vanadium and manganese. In most of these methods photometric detection is used. Only four kinetic fluorimetric methods for Mn(I1) appear to have been described. Morgen et al.5 found that this ion accelerates the aerial oxidation of the beryllium - morin complex kinetically and they established the determination of Mn(I1) in the range 5-50 ng ml-1. Other ions such as Cu(1I) , Cr(III), Fe(II), Fe(III), Co(II), Ni(1I) and Ag(1) also catalyse this reaction. Biddle and Whery6 reported the fluorimetric determination of Mn(I1) via enzymatic oxidation of 2,3- diketogulonate.The method can be applied to the determina- tion of this ion in the range 8-50 PM, but some ions such as Cu(I1) and Fe(I1) cause significant interferences. Guilbault et al.' have described a procedure in which Mn(I1) was assayed by its inhibition of the fluorimetric determination of hydrogen peroxide with the peroxidase enzyme. From 0.3 to 12 pg ml-1 of Mn(I1) could be detected, but the method was not specific for this ion. The oxidation of salicylaldehyde thiosemicarb- azone with hydrogen peroxide is catalysed with Mn(II).s A kinetic method is proposed to determine 2-9 ng ml-1 of Mn(I1). The interferences of the method have been eliminated using masking agents, but Fe(II1) interferes when its concen- tration is in a ten-fold excess. Therefore, this paper describes a sensitive and selective fluorimetric method based on the catalytic effect of Mn(I1) for the oxidation of pyridoxal 2-pyridyl hydrazone with hydrogen peroxide in a basic medium.The reaction rate is followed by measuring the increase in the fluorescence intensity of the product of oxidation formed with time. Exper h e n tal Apparatus The fluorescence measurements were made using a Perkin- Elmer MPF-43A spectrofluorimeter fitted with a device for kinetic measurements, which permits direct recording of fluorescence versus time graphs at fixed excitation and emission wavelengths. The cell compartment of the spectro- fluorimeter was thermostated by circulating water.All measurements were recorded with a sensitivity of 0.1 and excitation and emission slits of 5-nm spectral band pass. A set of fluorescence polymer samples was used daily to adjust the spectrofluorimet?r to compensate for changes in source A Perkin-Elmer 380 atomic-absorption spectrophotometer, equipped with a standard air - acetylene burner head and a manganese hollow-cathode lamp, was used for measurement of the manganese contents in the analysed samples. intensity. '(7 Reagents All chemicals were of analytical-reagent grade and the solutions were stored in black PTFE bottles. All glassware and bottles were cleaned according to the method of Thiers,g using acidic solutions and distilled water. Pyridoxal2-pyridylhydruzone (PPH) solution, 9.7 x 10-4 M.Prepared by dissolving the pure compound in ethanol. The synthesis of this reagent is very easy and has been described previously. 1" Manganese(II) stock solution, 1 .OOO g 1-1. The appropriate amount of manganese metal was dissolved in a minimum volume of nitric acid (1 -t 1) and diluted to 1 1 with 1% V/V hydrochloric acid. Buffer solution, p H 10.5. Prepared by dissolving 6.183 g of boric acid in water, adding 2 g of sodium hydroxide and diluting to 1 1. - z & (ow Y.2 4o I .; I $0" 20 se c .- r u J o w LU .- * b € 0 E CU Co V Mn Fe Mo Ru 0 s Cr Ag Hg Ir W NiTi ion determined Fig. 1. Statistical study of kinetic methods for the determination of metallic ions718 ANALYST, JUNE 1984. VOL. 109 Procedure To 1.2 ml of 9.7 X 10-4 M PPH solution in a 10-ml calibrated flask was added the volume of manganese(I1) solution needed to ensure a final cation concentration of between 0.05 and 8 ng ml-1, 2.5 ml of 0.1 M borate solution (pH 10.5) and 2 ml of 0.6 M hydrogen peroxide.The stop-clock was started. The solution was diluted to the mark with doubly distilled water. A portion of the reaction mixture was transferred into a thermostated 1.0-cm fluorescence cell at 65 k 0.1 "C and the measurements were begun 5 min after preparation of the samples. The reaction was followed fluorimetrically by measuring the rate of change in the fluorescence of PPH. The excitation wavelength was 355 nm and the emission was measured at 425 nm. The net fluorescence of the catalysed reaction values was obtained by subtracting the measurements of a blank solution prepared in a similar manner without manganese.Determination of Manganese in Dairy Products and Orgeat After drying the sample at 100"C, 2&30 g were taken for analysis, depending on the ash content. Dry ashing was carried out in a porcelain capsule in a furnace at 500 "C for 24 h. The ash was then moistened with water and 5 ml of perchloric acid were added. The sample was digested in a sand-bath until dry. A 5-ml volume of hydrochloric acid was then added and the sample was again evaporated to dryness. If the sample did not dissolve in hydrochloric acid, the perchloric - hydrochloric acid treatment was repeated. The final evaporation product was dissolved in 5 ml of hydrochloric acid and the clear solution transferred into a SO-ml calibratc I flask and the reaction rate measured as described above.Isolation of Oxidation Products PPH (0.5 g) was dissolved in 350 ml of concentrated ammonia solution. A 250-ml volume of 30% hydrogen peroxide solution was added. The mixture was heated on a steam-bath for 72 h, adding ammonia to keep the pH between 9.5 and 10. It was concentrated nearly to dryness on a rotary evaporator. Methanol was added and the mixture was cooled to room temperature. Under these conditions a white precipitate appeared. Results and Discussion Kinetics and Optimisation The fluorescence maxima of PPH (he, 398 and he, 475 nm) showed a hypsochromic change (hex 355 and he, 425 nm) when hydrogen peroxide was present in the solution. In the presence of trace amounts of manganese(I1) the oxidation rate of PPH increased owing to the catalytic effect of this ion.This effect is shown in Fig. 2. Several tests were carried out to characterise the oxidation product of PPH. The fluorescence spectra of pyridoxal, 2-pyridylhydrazine and PPH, as well as those of their oxidation products, are plotted in Fig. 3 from which the following can be inferred. (i) The fluorescence spectra of the initial reagents show large differences whereas the spectra of the oxidation products have very similar fluorescence maxima. (ii) The fluorescence characteristics of the oxidised pyri- doxal agree with those of the 4-pyridoxic acid lactone according to Bridges et al., 11 who attributed to this compound a he, of 360 nm and a he, of 430 nm in basic medium.(iii) The slight increase in the fluorescence intensity of the oxidation product of the 2-pyridylhydrazine and the batho- chromic shift of its fluorescence maxima can be attributed to the formation of an N-oxide of the pyridine and an intra- molecular hydrogen bond. This effect gives a planar structure to the molecule. 300 380 380 460 540 Wavelengthinm Fig. 2. Excitation and emission spectra showing the catalytic effect of manganese(I1) on the oxidation on PPH by hydrogen peroxide. 1 and l ' , PPH; 2 and 2', PPH - hydrogen peroxide; 3 and 3 , PPH - hydrogen peroxide - Mn(I1). [PPH] = 9.7 X l t s M: [H20,] = 0.35 M ; [Mn(II)] = 0.5 ng ml-I; reaction time = 30 min; and temperature = c 5 8 0 1 .L I 11, / T P A A , 350 450 0 2 50 Wavelengthinm Fig. 3. Broken lines: excitation fluorescence spectra of (a) pyridoxal 2-pyridylhydrazone, ( b ) pyridoxal and (c) 2-pyridylhydrazine.Bold lines: emission fluorescence spectra of the oxidation products of (a-c), obtained using solutions containing 1.5 M hydrogen peroxide solution. Temperature, 50 "C (iv) The fluorescence characteristics of the oxidation product of PPH may be attributable to the contribution of both oxidation products of the initial reagents, but the lactone formation seems to be the main cause of the fluorescence of the oxidation of PPH according to information inferred from the relative fluorescence characteristics of the spectra of Fig. 3. Therefore, the oxidation mechanism probably consists of hydrolytic cleavage of the azomethine group and oxidation of the resulting compounds.The results of the elemental analysis of the oxidation products of PPH are not conclusive because these products are contaminated with reagent and sodium hydroxide.719 stability of the diluted solution of manganese(II) Because the proposed method is very sensitive, we prepared a dilute solution of Mn(I1) (100 ng m - I ) from the stock solution, but some problems were encountered with the stability of this dilute solution. Optimum container materials and solution conditions were selected to eliminate losses of trace amounts of manganese by adsorption on the surface of the container. The stability of this solution was investigated using glass and PTFE containers, at three pH values and in the presence o r absence of tartrate ions as a complexing agent. to prevent Mn(1I) losses.Several samples were prepared by the pro- cedure described above, but using 1.2 ml of 100 ng ml-1 Mn(I1) solution. prepared and stored in different ways. The reaction rates were measured at different starting times from the preparation of stock solutions. Fig. 4 shows the relative reaction rate of each sample. When the dilute Mn(I1) solution was prepared without tartrate ions, the concentration of Mn(I1) decreased quickly as the sample prepared at 1 h had a relative reaction rate 65% less than the sample freshly prepared from dilute Mn(I1) solution. There was also a variation in the concentration of Mn(I1) when the dilute solution had a tartrate ion concentration of 1 1.18 ml-l. The dilute Mn(I1) solution was only stabilised when it was stored in a PTFE bottle at pH 3.7 and in the presence of 10 pg ml-1 of tartrate ions. In this instance the relative reaction rate of the samples prepared from this solution at different times did not change.at least for 50 h, but the value of the reaction rate was 20% less than that of the sample prepared from the dilute iMn(I1) solution without tartrate ions, probably owing to the complexing effect of these ions. 100 -/ 1 a, Y 2 .- 5 80 f 60 a, 40 c 0 m > m cc .- Y - 20 30 40 50 1 " 10 Tirneih Fig. 4. Effect of preparation conditions on the stability of the dilute manganese(I1) solution. A. Prepared at pH 0.6 or 4.6 and stored in glass or PTFE bottles; R , prepared at pH 0.6 with 1 yg ml-1 of tartrate ions and stored in a PTFE bottle; C, prepared at pH 3.7 with 1 pg ml-1 of tartrate ions and stored in a PTFE bottle; D, prepared at pH 3.7 with 10 yg ml-1 of tartrate ions and stored in a PTFE bottle 260 1 180 8 140 100 20 6o t Effects of reaction variables The influence of the temperature on the reaction rate (dl+/dt = tan a, where If is the intensity of fluorescence) was studied in the range 20-70 "C (Fig.5 ) . The reaction rate increased almost linearly when the temperature was increased to between 40 and 70 "C. Temperatures higher than 70 "C were not tested because they were near to the boiling-point of the ethanol. A temperature of 65 "C was selected for further studies. A linear relation between the logarithm of the rate constant and the reciprocal of the absolute temperature was found for both the catalytic and the non-catalytic reaction.The activation ener- gies were found to be 17.4 -t 0.2 kcal mol-1 for the catalytic reaction and 22.9 k 0.3 kcal mol-1 for the non-catalytic reaction. The pH dependence of the system was studied using sodium hydroxide and hydrochloric acid solutions. The reaction rate did not depend on the pH in the range 10.2-10.9 (Fig. 5 ) . The reaction rate decreased at lower or higher pH values. A study was made to determine the dependence of the system on the type of buffer solution used. Several solutions were tested. The slopes of the fluorescence versus time graphs decreased when ammonia - ammonium chloride or disodium phosphate - sodium hydroxide buffer solutions were used. The slopes increased using sodium hydrogen carbonate - sodium carbo- nate or boric acid - sodium hydroxide buffer solution, but with the latter better results were obtained. The fluorescence intensity maximum decreased slightly when the concentration of this buffer solution increased, but the rate of reaction did not change when the concentration was between 2 x 1 0 - 2 and The rate of the catalytic reaction increased linearly with a rise in the PPH concentration from 5.5 x 10-5 to 10-4 M, but it remained constant between 10-4 and 2 x 10-4 M.The effect of hydrogen peroxide concentration was tested in the range 1.5 x 10-2-3 X 10-1 M. The reaction rate depended linearly on the concentration of this variable up to 5 x 10-2 M, it did not change between this concentration and 2 x 10-1 M and decreased as the hydrogen peroxide concentration continued 5 x M.20 30 40 50 60 70 80 Tern peratureK I 1 I I L 1 I 9.7 9.9 10.1 10.3 10.5 10.7 10.9 11.1 -1 PH I I I I 5 7 9 11 13 15 17 [PPH]/M X lop5 I I I I Fig. 5. Effect of the variables on the reaction rate of PPH - hydrogen peroxide - Mn(I1). A, Temperature; B, pH; C, PPH concentration; and D, H20, concentration increasing. The effect of both variables is also shown in Fig. 5. The fluorescence intensity versus time graphs for solutions containing different Mn(I1) concentrations were recorded. The initial slopes indicated a first-order reaction with respect to manganese. The net catalysed reaction values, obtained by the measurements of a blank solution prepared in a similar manner without manganese, are shown in Fig. 6. The various kinetic dependences are summarised in Table 1.On the basis of our kinetic investigation, equation (1) is suggested for the oxidation of PPH (1.2 x 10-4 M) by hydrogen peroxide (0.12 M) at pH 10.4 in the presence of 2.5 x 10-2 M borate buffer solution and with manganese as a catalyst. d[PPH],,ldt = K[Mn*+] . . . . (1) Equation (2) gives the kinetics for the non-catalytic reaction under the same conditions: where K and KO are the rate constants for the catalytic and non-catalytic reactions, respectively and [PPH],, is the concentration of the oxidised reagent. d[PPH],,ldt = KO [PPHI-'" . . . . (2) Kinetic Determination of Manganese Calibration graphs The tangent method was used to calculate the rate of the catalysed reaction, which was plotted as a function of the manganese concentration.The differential , I 2 the fixed-time720 ANALYST, JUNE 1984, VOL. 109 Table 1. Summary of Mn(1I) - PPH - hydrogen peroxide kinetic data Background reaction Mn(I1)-enhanced reaction less background reaction Dependenct of V , on . . [H-'] . . [Buffer]' [Buffer 10 [PPH] I / z [PPH] 5 . . [ H 2 0 2 J 1 . . [H202]0 . . [H' 1'' . . " a 1 Concentration rangehi 6.0 X 1 0 "1.5 X 10- I ' 1.5 x 10 "-1.0 x 10 '0 1.0 x 1(k"'-4.0 x 10"' 5.0 x 1 v i -2.0 x 10-2 . . 2.0 x 1Vz -5.0 x 102 5.0 x 1 0 5 -1.2 x 1 w 1.2 x l ( P -1.5 x 10-4 2.0 x l(k2 -5.0 x 1 ( P 5.0 X 10-2 -1.5 X l(k' 1.5 x 10-1 -3.0 x 10 1 . . . . . . . . . . . . . . . . . . Dependence of V,, on Concentration rangeh 6.0 x lP1L1.O x 10-" HA]?/? H']" . . . . 1.0 X l t ' l - 1 . 0 X l&") H i ] 2 .. . . 1.0 X 10-l"-2.0 X 1@ 10 PPH]"' PPH]" . . . . 1.0 X 1W -2.0 X 1W H202]'' . . . . . . . . BufferI"3 . . 1.5 X 1 P -2.0 X 10 2 Buffer]" . . 2.0 x 1 P -5.0 x 10' 5.5 x 1@5 -1.0 x 1 w 2.0 X 10 2 -5.0 X 10 2 5.0 X 10' -2.0 X 10 ' 2.0 X 1 t ' -3.0 X 10-1 . . H20,]"2 HzO,] ' . . . . 0 3 5 7 9 1 1 Ti me/mi n Fig. 6 . Fluorescence - tlme graphs obtained for ( 1 ) PPH; (2) PPH - H1O,; and (3) PPI1 - H,O,y Mn(1I) systems. A,, = 355 nni; h,,, = 425 nm; [PPH] = 9.7 X l ( r M ; [H202] = 0.12 M ; pH = 10.5; and temperature = 50 "C and the fixed-intensity methods13 were also used. For the fixed-time method, measurements were made after 5.5 min. For the fixed-intensity method, the inverse of the time necessary to obtain a relative fluorescence intensity of 30"/~1 was plotted against the manganese concentration.In all instances, the calibration graphs were linear in the concentra- tion range indicated in Table 2. The precision of the four methods applied to the determinations of 1 ng ml-1 of Mn are given in Table 2. The fixed-time method was the most sensitive, the fixed- intensity method had the higher range of application, but the tangent method was the most suitable for the kinetic determi- nation of manganese, because of its greater precision. In te rf eren ce stu d j The interference effects of53 cations and anions on the kinetic determination of 1 ng ml-1 of manganese(I1) were examined and the tolerance limits for interfering ions are summarised in Table 3. Some masking agents (potassium cyanide, sodium fluoride and thioglycollic acid) were used to eliminate the more serious interferences.Pb(I1). Cu(II), Ag(I), Ni(I1) and Co(I1) decreased the reaction rate, when their concentrations were higher than those indicated in Table 3, because these ions form complexes with PPH. Ti(IV), Zr(IV), Hf(1V). Sb(II1) and Fe(I1) ions in higher concentrations accelerated the reaction because they also catalyse the oxidation of PPH. No ion interfered at the same level as Mn(I1). Only Co(I1) caused perturbation when its concentration was in a 6Wf0ld excess over Mn( 11). Table 2. Kinetic fluorimetric determination of manganese(I1) Mn(I1) Relative standard concentration deviation Method rangeing ml-l ( n = l o ) , %I Tangent. . . . . . 0.2-3.0 Fixed-intensity . . 0.2-8.0 Differential . . . . 0.1-2.0 Fixed-time .. . . 0.05-2.5 0.62 2.05 2.44 2.22 Table 3. Effect of various ions on the determination of 1 ng ml-' of manganese( 11) Tolerance ratio of ion to Mn(I1) Ion added Na(I), K(I), Be(II), Ca(II), Sr(II), Ba(II)? Li(1). Mg(II), Tl(I), Se(IV), La(III), Cd(II), Pt(1V). Cr(II1). In(III), Sn(II), Au(III), V(V), Zn( 11). Zr( IV) .* Hg(II), AI(III), Ce( IV). W(VI), Cr(VI), Mo(VI), As(II1). NO3-. NOz-, SO,'-, C0,'-, C,O,Z-, P o p , P2076, P30,0S-, C1-, Br-, I-, CN-, C104-, acetate. citrate . . . . Cu(I1) ,i- Ni( 11)$ . . . . . . . . . . . . 750 Fe(I1)t . . . . . . . . . . . . . . . . 250 Co(I1)i- . . . . . . . . . . . . . . . . 60 1 000 Sb(III), Ti(IV),* Hf(1V). Pb(II), Ag(I)B . . . . 500 * Sodium fluoride concentration. 10 pg ml-I. -t Potassium cyanide concentration.5 pg ml-I. $ Potassium cyanide concentration. 10 pg m-I. B Thioglycollic acid concentration, 2 pg ml-I. Determination of manganese in dairy products and orgeat The kinetic method was applied satisfactorily to the determi- nation of manganese in different commercial dairy products and orgeat samples. Table 3 shows the results obtained for each sample. Compared with the atomic-absorption method the fluorescence reaction-rate method had a higher sensitivity for the determination of manganese (55 ng ml-1) and did not incur the error of the atomic-absorption method, in which manganese measurement was carried out in the ultraviolet region (279.5 nm). Conclusion To evaluate the reliability of the proposed method, a systematic study of the kinetic methods for the determination of Mn(I1) was carried out.Fig. 7 shows several characteristics of these methods.5-8.l+lY Most of them are based on the oxidation reaction of organic reagents by hydrogen peroxide or periodate. Only one method14 is more sensitive than the method proposed here. It is based on the measurement of theANALYST, JUNE 1984. VOL. 109 ~ Table 4. Results of the kinetic determination of manganese in dairy products and orgeat. The results are the averages of six determina- tions Manganese found/ Sample ug per 100 g Pasteurisedmilk . , . . 3.9 4 0.1 Skimmed milk (0.05% fat) . . 8.3 k 0.1 Sterilised milk (3.20% fat) . . 5.6 2 0.1 Whipping cream . . . . 4.3 4 0.5 Ice cream . . . . . . . . 8.9 ? 0.3 Yoghurt . . . . . . . . 5.3 * 0.1 Cheese . . . . . . . . 32.0 ? 0.4 Orgeat .. . . . . . . 12.7 k 0.2 Atomic-absorption spectrophotometric method/ PE Per 100 E 3.8 8.3 5.6 4.4 9.0 5.1 32.1 12.6 I ~ Y! 5 0 I U 0 c c m 0 6 N 0, I -6 t -1 49 ------I 48 1- 47 - 46 - 45 - 44 - 5 I 1 43 - 42 - 41 - 40 - 39 +-----I 38 - 37 I I 36 I 1 35 I ‘ 34 - 33 - 32 t 4 31 - 30 29 - 7 I 4 28 t-----+ 27 - 26 - 25 -8 I I 24 r ‘ 2: - 22 I I21 - 20 - 19 1 i 18 b I 16 I17 t 115 Lr-. 3 , , ’ - - ’ ’ ‘ I ” “ J - 14 1 - 2- -5 -4 -3 - 2 -1 u 1 Fig. 7. Characteristics of the kinetic methods proposed for thc determination of Mn(I1) bv photometric and fluorimetric techniques. Thin lines. photometric method; bold lincs. fluorimetric method: and striped line. proposed method. Numbers on lines are refercnccs reaction rate of the oxidation of purpurin by hydrogen peroxide, catalysed by Mn(I1) and in the presence of ethylenediamine as an activator, but this method presents more interferences than ours because Pb(1I).Fe(II1) and I- cause perturbations at the same Mn(I1) level. Another method’s with a similar sensitivity is based on the catalytic effect of Mn(1I) on a Nitrosil Yellow - hydrogen peroxide system activated with o-phenanthroline, but Pb(TI), Fe(TI) and pyrophosphate interfere at a lower concentration than in this method. 72 1 The method proposed has a very high sensitivity (detection limit 0.05 ng ml-I), high selectivity [no ion interferes at the same level of Mn(II)] and the precision is very acceptable for the determination of low ranges of Mn(I1). This method can be applied to the determination of Mn(I1) in food samples.1. 2. 3. 4. 5 . 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 3 5 . 36. 37. References Grases, F . , and ValcArcel. M., Talanta, 1983. 30. 139. Grases, F., Garcia-Sanchez, F.. and Valcarcel, M., Anal. Chim. Acta, 1980, 119, 359. Grases. F., Garcia-Sanchez, F., and Valcrircel, M., Anal. Quim., 1980, 76(B). 402. Grases, F., Garcia-Sanchez, F., and Valcarcel, M., Anal. Lett., 1979. 12, 803. Morgen, E. A . , Vlasov, N. A.. and Kozhenyakina, L. A . , Zh. Anal. Khim., 1972, 27, 2064. Biddle. V. L., and Whery, E. L., Anal. Chrm., 1978, 50, 867. Guilbault. G . G.. Brignac, P . , and Zimmer, M., Anal. Chem., 1968. 40, 190. Moreno, A . , Silva, M., Perez-Bendito, D., and Valcarcel. M.. Talanta, 1983, 30, 107.Thiers. R. E., Methods Riochern. Anal., 1957. 5 . Rubio, S., G6mez-Hens, A , , and Valcrircel. M., Anal. Quim., 1983, 79B, 72. Bridges, J . W., Davies, D . S., and Williams, R. T., Microchem. J . , 1966, 98, 45. Yatsimirskii, K. B., “Kinetic Methods of Analysis,” Pergamon Press, Oxford, 1966. Chapter 3. Yatsimirskii. K. B.. and Tikhonova, L. P.. “Essays on Analytical Chemistry,” Pergamon Prcss, Oxford, 1977, p. 529. Bartkus, P., and Nauekaitis. A . , Nauchn. Konf. Khim. -Anal. Pribalt. Resp. B. SSR, (7ezisy Dokl.). lst, 1974, 190. Bartkus, P.. Nauchn. Kmj: Khim.-Anal. Pribalt. Resp. B. SSR, (Tezisy Dokl.), 1st. 1974. 193. Bartkus, P . , and Daugirdiene, D.. h‘auchn. Konf. Khim.-Anal. Pribalt. Rcs. R. SSR, (Tezisy Dokl.), ist, 1974, 187. Janjic, T., Milanovic, G .A., and Celap, M. B . , Anal. Chern., 1970. 27, 42. Bartkus. P . , and Jasinskiene. E. I.. Zh. .4nal. Khirn., 1968.23, 1622. Sychev, A. Y., and Isak. V. G . , Zh. Anal. Khim., 1978, 33, 1351. Yamane. T., and Fukasawa. T.. Bunseki Kclgaku, 1977, 26. 300. Bartkus, P., Kaleniskaite, S . , and Jasinskiene, E. I., Tr. Akad. Naiik Lit. S S I I , 1969. 10. IS: Rqf: Zh., Khirn.. 1070. 19GD. 3. 3G12.1. Blank, A. B., and Voronova. A.. Zavod. L,ub., 1956.31. 1299. Sychev, A. Y . , and Isak, V. G.. Izv. Akud. Nauk. Mold. SSR, Ser. B i d . Khim. Nauk, 1979, 4, 88. Dolmanova, I. F . , Zolotova, G. A , . and Ratina, M. A . , Zh. Anal. Khim., 1978. 33. 1356. Bartkus, P., and Jasinskiene, E., Khim. Khim. Tekhnol.. 1970, 11, 83. Shigeki, A , , Kunio, ‘I,., and Tsutomu, M..Anal. Chim. Acta, 1975. 80, 135. Sekheta, M. A , . Milovanovic. G . A . , and Janjic. T. J . . Mikrochim. Acta, 1978, 297. Weis, H., and Ludwig, € I . , Anal. Chim. Acta, 1972. 62, 125. Sychev, A,, and Tiginyanu, Y., Zh. Anal. Khim., 1969. 24, 1842. Dolmanova, I. F., Poddubienko, V. P.. and Peshkova, V. M., Zh. Anal. Khim. 1970, 25. 2146. Mottola, H. A , , and Harrison, 0. R., Talanta, 1971, 18, 683. Nikolelis, D . P.. and Hadjiioannou, T. P., Analyst, 1977. 102, 591. Tiginyanu, Y. D., and Oprya, V. I., Zh. And. Khim., 1973,28, 2206. Bognar, J . , Mag!,. Tud. Akad. Kern. Tud. O s z t . Kozl., 1958,7, 335. Nikolelis, D. P . , and Hadjiioannou, T. P.. Anal. Chem., 1978, 50. 205. Szebelledy. L., and Bartfay. M.. Fresenius Z. Anal. Chem., 1936, 106. 408 Hadjiioannou, ’r. P., arid Kephalas. T. A . . Mikrochim. Acta, 1969. 6. 1215.722 ASALYS?. .JUNE 1984. VOL. 109 38. 39. 40. 41. 42. 43. 44. Dittel, F . , Fresenius Z . Anal. Chem., 1967. 229. 193. Efstathiou, C., and Hadjiioannou, T. P., Talanta, 1977, 24, 270. FernAndez, A . , Sobel, C., and Jakobs, S., Anal. Chem., 1963, 35, 1721. Fukasawa, T.. Yamane, T.. and Yamazaki, T., Bunseki Kagaku, 1977, 26, 200. Bartkus, P . , Nauekaitis, A , . and Jasinskiene, E., Nauchn. Tr. Vyssh. Uchebn. Zaved. Lit. SSR, Khim. Khim. Tekhnol., 1975, 16. 19; Ref. Zh., Khirn., 1976, 19GD, 3, 3G141. Dolmanova, I. F., Yatsimirskaya, N. T., and Peshkova, V. M., Zh. Anal. Khim., 1973, 28. 112. Pantaler, R. P . , Alfimova, L. D., and Bulgakova, H. M., Zh. Anal. Khim., 1975, 30, 1584. 45. 46. 47. 48. 49. Siokawa, T., and Suzuki, S . , J. Chem. Soc. Jpn., 1951, 12, 72. Perez-Bendito, D., Valcarcel, M.. Ternero. M.. and Pino, F., Anal. Chim. Acta, 1977. 94. 405. Belcher. R., Hasani, S . M. T., and Townshend, A . , Egypt. J. Chem., 1973, 131. Korenman, I. M., and Lebedeva, A. N . , Zh. Anal. Khim.. 1963, 19GDE, 20. Sanchez-Pedreno, C., and Arias. J . J . , Quim. Anal., 1974, 28, 184. Paper A311 79 Received June 20th, I983 Accepted January 18th, 1984

ISSN:0003-2654    

DOI:10.1039/AN9840900717

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

.AY.,\LYST. JUNE 1984. VOL. 109 723 Simple Electrochemical Cell and Microcell for Spectrofluorimeters Brian L. Cousins Jodi L. Fausnaugh and Theodore L. Miller Department of Chemistry Ohio Wesle yan University Delaware OH 430 75 USA A microcell and an electrochemical cell for fluorescence studies are described. The cells are inexpensive and can be constructed in a few minutes without special equipment. They can be used in commercial instruments without special cell adapters and without the need of alignment. Tests show that high-resolution spectra can be recorded with the cells and that picogram amounts of material can be determined. The major advantage of both cells is the small sample volume required. The sample volume has been reduced t o 2 PI for the microcell and 90 UI for the electrochemical cell.Simplicity of construction accuracy and small sample volume make these cells suitable for many applications. Keywords Electrochemical cell; microcell; spectrofluorimetry Yildiz et al. 1 designed an optically transparent thin-layer electrode (OTTLE) for fluorescence experiments over 15 years ago. They illustrated that perylene could easily be monitored by a thin-layer electrode with a commercial spectrofluorimeter at concentration levels several orders of magnitude lower than values detectable with absorption instruments. Most spectroelectrochemical experiments since 1968 have focused on absorption measurements. However, the use of the inherent fluorescence of amino acids in biological systems2 has re kindled an interest in fluorescence.Simone et al.3 recently used a conventional OTTLE for fluorescence studies of cytochrome c. Because a number of technical difficulties have complicated their preliminary stud-ies they have developed a new long optical path electro-chemical cell for future investigations.4 As the conventional OTTLE offers a number of advantages we have developed a simple microcell for fluorescence investigations that allowed us to evaluate thoroughly the fluorescence characteristics of the OTTLE and to remedy the technical difficulties outlined by Simone et al.3.4 Experimental Apparatus Absorption spectra were recorded at room temperature on a Varian Cary 219 spectrophotometer. All luminescent measurements were taken at room temperature with an Aminco-Bowman Katio I1 spectrofluorimeter.Cyclic voitammograms were obtained with a Princeton Applied Research (PAR) Model 175 universal programmer coupled to a PAR Model 173 potentiostat - galvanostat with a Model 179 digital coulometer Model 178 electrometer probe and a Hewlett-Packard Model 7045B X - Y recorder. All reported potentials are relative to a saturated calomel electrode (S.C.E.). Reagents Terbium chloride and cerium(II1) sulphate (99.9%) were purchased from Alfa Inorganics. All other compounds were of analytical-reagent grade and used as obtained. Fresh stock solutions were prepared for each luminescence study. Cell Design The optically transparent thin-layer electrode construction is similar to the electrochemical cell for absorption.5-6 Two non-fluorescent quartz plates (50.8 x 16 x 0.1 mm) replace the microscope slides and an internal electrical connection is used.In this design one of the PTFE spacers is moved about 2 mm towards the centre of the cell. A strip of aluminium foil (4 X 70 mm) is folded around one end of the minigrid. The final fold leaves a 1 x 70 mm strip of foil with a strong connection to the minigrid. Next the aluminium foil strip is placed along the edge of the quartz plate in a space made by moving the PTFE spacer. The minigrid lies across the cell about 2 mm from the bottom as shown in Fig. l(a). A thin bead of epoxy cement is placed between the aluminium foil strip and the PTFE tape. When the second quartz plate is clamped on top the epoxy seals the aluminium foil and prevents solution contact with the foil when the cell is in use.The aluminium foil strip extends through the top of the cell allowing effective electrical contact. The microcell is constructed from two quartz plates (50.8 x 16 X 1 mm) 0.002-in adhesive PTFE tape spacers (Fluorofilm DF-1200 PTFE tape Fluorocarbon Dilectrix Division Lock-port NY USA) and epoxy cement. Strips of PTFE tape 6 mm wide are cut and placed along the 50.8 mm sides of a quartz plate. A second quartz plate is then laid on top to form a “sandwich,” which is clamped into place. Epoxy cement is applied along the taped edges and allowed to cure. The optical portion of the cell between the spacers is about 4 mm wide. E\ t Em. Fig. 1. Electrochemical cell and microcell design and use. (a) Assembly of the OTTLE; and ( b ) top of the microcell in a standard 1-cm sample compartment.(A) Quartz plates; (B) PTFE tape spacers; (C) gold minigrid; (D) aluminium foil strip (E) standard 1-cm sample compartment; and (F) microcel 733 ANALYST. JUNE 1984 VOL. 109 When a 5-pl volume is used the sample rises in the cell to a height of 13 mm. The excitation and emission slits are positioned 3 mm above the cell bottom and are 5 mm high. Therefore the sample height in the microcell must be greater than 8 mm. By moving the spacers so that there is 2 mm between them. sample volumes of 2 p1 may be used. Under these conditions the sample height is 10 mm. Reducing the optical width to less than 2 mni would require the excitation slit to be less than 1 mm. The sample is introduced into the thin layer between the quartz plates by capillary action.A plastic microweighing dish is convenient for this operation. When aqueous solutions are placed on the surface of the plastic dish they form spherical drops making it easy to transfer the sample to the microcell. The solution quickly enters the cell when the open end of the cell is placed on the drop. Non-aqueous solutions tend to spread out on the plastic surface and may require slightly larger volumes. Capillary action holds the sample in the cell. In fact the force is so strong that the sample migrates slowly up the cell when a 2-1.11 sample is used in a cell with spacers 2 mm apart. The solution can be removed from the cell by stretching a thin-walled rubber tube over the end of the cell and gently aspirating.This technique can also be used to rinse the cell. When the cells with samples are ready they are placed diagonally into a standard 1-cm sample compartment as illustrated in Fig. l(b). It may be necessary to file the epoxy cement on the edges to allow the cell to fit into the sample compartment. It is also possible to insert the cell between the compartment corners not used in Fig. l(b). In this alternate geometry the exciting and emitted radiation enter and leave the same sample face. The front-face geometry is necessary for optically thick samples. However the scattering of the incident radiation is greatly increased so that the front-face geometry should only be employed with optically thick samples. Mirrors should also be removed from the sample holder to avoid increased scattering of the incident radiation.If the optical section of the cell requires a thorough cleaning after extensive use the epoxy cement can be easily removed by warming the cell in dilute HNQ. A solution cup for the OTTLE was constructed from a short section of 12 mm o.d. plastic test-tube that was attached to a small square of aluminium or PTFE ( a standard 1-cm fluorescence cell lid can be used). The square was cut to fit snugly in the sample compartment of the fluorinieter and the test-tube section was cemented to the square with epoxy cement. In operation the solution cup replaces the nitrogen purge system in the Aniinco-Bowman sample compartment and the OTTLE is placed diagonally in the compartment as shown in Fig.l(h) on top of the solution cup. The cell must be completely filled with solution before it is placed on the solution cup because the solution will migrate up inside the cell by capillary action breaking the external solution contact. The reference (S.C.E. with a long 4-mm o.d. tip) and auxiliary electrodes are placed on opposite sides of the cell in the corner away from the excitation and emission beams [upper right-hand corner in Fig. l(b)]. It is convenient to tape the auxiliary electrode to the cell. Using the solution cup described above. a total sample volume of about 1 ml is required; a 1-cm PTFE block with a thin epoxy ring on the top can replace this solution cup and the total sample volume can be reduced to about 90 p1 with this arrangement.Results and Discussion The microcell with 4 mm between spacers uras tested with several samples in three solvents. The spectrum for a terbium reagent [dipicolinic acid (DPA) to Tb(II1) ratio 3 1 Tb(II1). 0.318 mhi in piperazine buffer at pH 6.51 using the microcell is presented in Fig. 2A. The excitation wavelength was 230 nm with the excitation and emission slits at 1 mm (5.5 nm band pass) and 0.5 mm (2.75 nni band pass) respectively. Trace B was recorded for the same sample in a standard 1-cm cell under identical instrumental conditions except the excitation wavelength was 250 nm. When an excitation wavelength of 250 nm is used for the microcell the relative intensity is about one eighth of the value for the standard cell as observed for all of the other samples tested.Terbium emission is based on non-radiative intramolecu!ar energy transfer from the DPA ligand and the increased luminescence in the microcell with 230 nm excitation is probably due to an inner filter effect in the standard 1-cm cell pushing the apparent excitation wavelength to higher values. The peaks at 230 and 460 nm in trace A result from scattering of the incident radiation from the surface of the quartz. Although non-fluorescent quartz plates were used to construct the cell. there is a very weak fluorescence in the 300-450-nm range from the quartz which is visible in Fig. 2 (h,,, for the excitation of this fluorescence is below 200 nm but the edge of the band extends above 200 nm). A 3.23 yg ml-1 sample of Tb(II1) was determined by spectrofluorimetric analysis using DPA in the microcell with a -0.6"/" erior when excited at 250 11111.7 The results presented above show that the detection limit can be lowered by a factor of eight when a 230-nm excitation wavelength is used with the microcell.Although this study has focused on spectro-fluorimetric measurements the absorption spectrum for the terbium reagent was recorded on a standard spectro-photometer with the microcell. The emission spectrum for quinine sulphate (20 pg ml-1 in 0.05 M H2S0,) is shown in Fig. 3. The excitation uavelength was 250 nm and the excitation and emission slits were 1 mm (5.5 nm band pass) and 0.5 mm (2.75 nm band pass), respectively. The peaks at 250 500 and 750 nm in trace B result from scattering of the incident radiation.A sharp cut filter placed in the emission beam to cut off the exciting radiation can be used to remove the peaks as shown in trace B in Fig. 3. The filter must transmit the luminescence (a Corning 0-51 filter was used for this sample). Using the 0-51 filter the emission spectra for quinine sulphate samples at 0.2 pg ml-1 (250 nm excitation) and 20 11.g 1-1 (350 nm excitation) were recorded. It should be noted that the &51 sharp cut filter removec most of the weak luminescence from the quartz. However the end of the band above 350 nm is transmitted by the filter and appears as a shoulder on the spectrum of quinine loo -U 200 400 600 Wavelength nm Fig. 2. Lurninesccncc spectra of a terbium reagent in piperazine buffer at pH 6 .5 . The spectrum \$as recorded in ( A ) the inicroccll and (I31 ;I standard 1-crn cell. Both spectra were recorded a t room t e in pe r a t u r ANALYST. J U N E 1984. VOL. 109 725 sulphate. As the luminescence intensity from the quartz drops off sharply with excitation above 250 nm greater sensitivity is achieved for quinine sulphate when excited at 350 nm. Changing the excitation wavelength can also be used when the luminescence overlaps with the scatter peakb. In the terbium reagent spectrum (Fig. 2A) excitation at 230 nm causes the second-order band from the emission monochro-mator to occur at 460 rim which is below the terbium emission. However excitation at 250 rim (hmds for the standard 1-cm cell) would give a 500-nm scatter peak, completely obscuring the first peak in the emi5sion spectrum.Generally there is only a slight decrease in sensitivity when a different excitation wavelength is used because the excitation band is wide. Quinine sulphate samples were used to evaluate the reproducibility of the cell path length and cell position within the sample compartment. Firstly the relative intensity for a 2.00 pg ml-1 sample was monitored as the microcell was repeatedly inserted into the cell compartment to record the emission intensity then removed and replaced. The average value for ten trials was 8.54 with a standard deviation of 0.01. The relative intensity of five samples in the 20 pg ml-1 to 2 pg 1-1 range and a blank were recorded at 450 nm using the instrumental conditions outlined for the microcell spectrum shown in Fig.3. A correlation coefficient of 0.999 was obtained for the linear regression of the relative intensity data for the microcell. Six samples containing quinine sulphate were determined at two different concentrations using the regression data and another sample was analysed in six different microcells. The results are sumrnarised as follows. I I I I I 200 300 400 500 600 700 800 Wavelengthlnm Fig. 3. Luminescence p x t r u m ot quinine sulphate in 0 05 hi H2SOI. The spectra Rere recorded at room temperature (A) with and (B) mithout a filter When the actual concentration of the solution was 2.00 pg [nl-l? the average value for the six samples was 1.99 pg ml-I with a standard deviation of 0.008 pg ml-1; for a 2.00 prig 1-1 sample the average value was 2.3 pg 1-1 with a 1.1 pg 1-1 standard deviation.When a 20.0 pg ml-1 sample was analysed with six different microcells the average value was found to be 20.0 pg ml-1 with a standard deviation of 0.8 pg ml-1. As the detection limit is approached the error increases, as shown by the 2.00 pg 1-1 sample but even at this level one can determine 10 pg (based on a volume of 5 pl) with reasonable accuracy in a standard commercial instrument without a special cell adaptor and without the need of alignment. Matched cells can be prepared if necessary. The cell was evaluated for fluoroelectrochemical use with cerium(I11) in sulphuric acid. Cerium(II1) ions luminesce with a A of 352 nm when excited at 252 nm.8 This luminescence has been used for the determination of Ce(II1) in the presence of the non-fluorescent Ce(1V) ions." The electrochemistry of the Ce(1V) - Ce(II1) system has recently been investigated in sulphuric and perchloric acid solutions.10.11 A typical cyclic voltammogram for 1.08 mM Ce(II1) in 0.5 M H2S04 is shown in Fig. 4(b). The peak at +0.95 V is due to the reduction of gold oxide formed on the electrode surface. 12 The change in relative luminescence intensity (hex. = 252 nm hem, = 352 nm) was recorded simultaneously and is presented in Fig. 4(a). As Ce(II1) is oxidised to Ce(1V) the intensity decreases and at +lS V with only Ce(lV) near the minigrid, the intensity drops to the background level. The relative intensity returns to the initial level on the reverse scan.The luminescence spectra of the Ce(II1) solution at various applied potentials are presented in Fig. 5. The Ce(II1) luminescence illustrates the optimum condi-tions for the electrochemical or microcell because the emission band lies near the centre of the scatter peaks. Consequently, filters were not required for the cerium analysis. As the OTTLE yields a relative luminescence intensity that is about 90% of the value observed in the microcell under the same conditions all of the test results presented earlier for the microcell can be applied to the electrochemical cell. There-fore dilute solutions can be studied with fluorescence filters and the electrochemical cell outlined in this paper. Conclusion The microcell described has been used in our laboratory for more than a year.The microcell is much easier to construct than those reported earlier.13-15 The components cost about half as much as a standard 1 -cm fluorescence cell and it can be constructed without special equipment in a few minutes. Although the microcell cannot detect femtogram amounts as the nanolitre cells designed fur HPLC detectors can,16 it can be used in commercial fluorescence instruments and test +20 ' I 1.5 1.4 1.3 1.2 1.1 1.0 0.9 0.8 0 -10 - 20 -30 M-1.5 1.4 1.3 1.2 1.1 1.0 0.9 0.8 ENvs. S.C.E. Fig. 4. 0 . 5 M H2S0,. E ,,, ,', = + 1.1 V vs. S.C.E. with the initial scan positive and scan rate 5 mV s I Simultaneous collection o f ( a ) the relative emission intensity data and ( h ) the cyclic voltammogram for 1.08 mM Ce(II1) i 726 ANAI-YST.JtJNE 1984. VOL. 109 70 60 50 > v) + .-40 4-S al .-.- c 30 [r 20 10 0 200 I 300 400 500 600 Wavelengthinm 700 800 Fig. 5. (D) 1.275 (E) 1.300 (F) 1.350 (G) 1.400 and (H) 1.500 V Luminescence spectra of 1.08 mM Ce(II1) in 0.5 M H,SO at various applied potentials. (A) 1.100 (B) 1.200. (C) 1.250, results show that it can be used without special cell adaptors and without the need of alignment. It can be used in the front-face geometry for optically thick samples and can also be used in a standard spectrometer for absorption measurements. The electrochemical cell provides the enhanced sensitivity of fluorescence spectroscopy coupled with simple construc-tion. The conventional OTTLE design greatly decreases the time for complete electrolysis when compared with a long optical path electrochemical cell.4 The major advantage of the electrochemical cell and microcell is the small sample volume.This feature will be especially attractive for very small biological and forensic samples that cannot be diluted. Acknowledgement is made to the Donors of the Petroleum Research Fund administered by the American Chemical Society for the support of this research (PRF No. 11999-B3), and to the National Science Foundation for providing an instrumentation grant to purchase the Aminco-Bowman spectrofluorimeter (CDP-7923794). 1. 2. References Yildiz A. Kissinger. P. T and Reilley C. N. Anal. Chem., 1968 40 1018. Galley W. C. and Milton J . G. Photochem. Photohiol., 1979 29 179.3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. Simone M. J . . Heineman W. R . and Kreishman. G. P. J . Colloid Interface Sci. 1982 86 295. Simone M. J . Heineman W. R . and Kreishman G. P. Anal. Chem. 1982. 54 2382. DeAngelis T. P. and Heineman. W. R . J . Chern Educ., 1976 53 594. Heineman W. R. Anderson C. W. Halsall H . B. H u n t , M. M Johnson J. M. Kreishman G . P. Norris. B. J . , Simone M. J . and Su C.-H. Adv. Chem. Ser. 1982 No. 201, 1. Miller. T. L. and Senkfor. S. I. Anal. Chem. 1982 54,2022. Mazza L. Ann. Chim. (Rome) 1940 30 47. Armstrong. W. A. Grant. D. W and Humphreys W. G Anal. Chem. 1963 35 1301. Kravtsov V. I. Stolyarov. G. K. and Izhak 0. A . Russ. J . Inorg. Chem. 1979 24 1812. Bishop. E. and Cofrk P. Analyst 1981 106 316. Adams R. N. “Electrochemistry at Solid Electrodes,” Marcel Dekker New York 1969 p. 194. Pillai C. G. and Patel. R. C. Anal. Biochem. 1980. 106 506. Ansevin A. T. and Vizard D. L. Anal. Biochem 1979. 97. 136. Eisinger J. and Flores J. Anal. Biochem. 1979 94 IS. Hirschy. L. Smith B. Voigtman E. and Winefordner J . D., Anal. Chem. 1982 54 2387. Paper A31251 Received August 8th 1983 Accepted January l s t 198

ISSN:0003-2654    

DOI:10.1039/AN9840900723

出版商:RSC    

年代:1984

数据来源: RSC