摘要:

334 Analyst, April, 1979, Vol. 104, pp. 334-347 Reproducibility of Pyrolysis - Mass Spectrometry Using Three Different Pyrolysis Systems D. A. Hickman and 1. Jane Metropolitan Police Forewsic Science Laboratory, 109 Lambeth Road, London, SE 1 7 L P A study has been made into some of the factors that affect the reproducibility of pyrolysis - mass spectrometry. Three separate pyrolysis systems were examined and three sample types, a simple system of easily pyrolysable polymers, an acrylic paint and an alkyd paint, were employed in order to cover a range of ease of sample pyrolysis. These samples were also examined by pyrolysis - gas chromatography. The reproducibility of the pyrolysis - mass spectrometry system was found to vary according to sample type. The source of irreproducibility was identified as the pyrolysis process and not the mass spectrometric detection.Keywords : Reproducibility ; pyrolysis - mass spectrometry ; fijwolysis - gas chromatography ; paint analysis Although the combination of pyrolysis with mass spectrometry was investigated by Zemanyl in 1952, and by other w0rkers2,~ at the same time, it is only in recent years that pyrolysis - mass spectrometry (Py - MS) has become established as a technique for examining polymeric material. The pioneering work of Meuzelaar and co-workers concentrated on the field of complex bio-organic substrates, notably bacteria,*,5 but also humic acids6 and peptides.7 Bacterial identification by Py - MS has become an accepted procedure,8-10 and the technique has been used in examining other naturally occurring organic materials, such as poly- saccharides.11 The early work on synthetic polymers,2 including poly~tyrene,~ has been extended to include aromatic and aliphatic polyesters12 and vinyl polymers and copolymers.13 Zemanf4 has reported the pyrolysis - mass spectrometry of 25 commercially available polymers.The attraction of Py - MS as a technique providing useful data on small samples of polymeric material has led to its application in forensic science.l5-ls In this context the technique has been applied successfully to the identification and comparison of paint and fibre samples. In Py - MS a sample is broken down pyrolytically (either within or external to the ionisation source of the mass spectrometer) and then repetitive mass spectral scans of the pyrolysate are taken.Data processing facilities enable a number of scans to be accumulated, possibly followed by integration to produce a mass pyrogram of the ~amp1e.l~ If Py - MS is to be used on a routine basis, involving perhaps the comparison of a mass pyrogram of a sample with a library collection of mass pyrograms, then the reproducibility of the system becomes an important factor. Good analytical reproducibility will be essential for any computer-based method of comparing mass pyrograms, and it is especially important when electron-impact (EI) mass spectrometry is used. In this mode the fragmentation within the source of the mass spectrometer gives many small fragment ions and leads to complex mass pyrograms. Simpler mass pyrograms would result from the use of chemical- ionisation (CI) , field-ionisation (FI) or low-voltage EI mass spectrometric detection, as these methods give molecular ions.Although, at the qualitative level, it would be easier to detect differences between samples by visual inspection of CI mass pyrograms compared with the corresponding EI mass pyrograms, EI mass spectrometry has the advantages of better sensitivity, providing more information and ease of controlling the experimental conditions, particularly with respect to gas pressure within the ionisation source. The two processes occurring in a Py - MS analysis, the pyrolysis and the mass spectro- metric fragmentation, can give rise to irreproducibility, but few of the papers published in this field have investigated the reproducibility of the procedure. Reiner is quoted9 as saying that the reproducibility of mass spectrometry is not as good as that obtained with gas chromatography (both in conjunction with pyrolysis) owing to the tendency of the mass spectrometer ionisation source to become contaminated.Meuzelaar et aZ.,19 in characterisingHICKMAN AND JANE 335 bacteria, reported that the variations caused by culturing the bacteria, sampling and analysing by Py - MS were less than the differences between strains. The reproducibility was not measured, although the authors found that results taken prior to a complete overhaul of the mass spectrometer were similar to those obtained afterwards. Acceptable reproducibility was also claimed by Zeman14 when he subjected synthetic polymer samples to programmed heating on a direct probe in a mass spectrometer operating under EI conditions.He found that the polymers tested could be degraded reproducibly in the temperature range 160-500 "C. Quantitative measures of the reproducibility of Py - MS systems have been quoted by Schulten and Gortzll and by Saferstein and Manura.17 Schulten and Gortz used Curie- point pyrolysis in combination with high-resolution FI mass spectrometry and found the variation for repetitive pyrolyses of glycogen samples to be about -35% for peaks of intensity greater than 10%. Saferstein and Manura,l7 pyrolysing under CI conditions, found that their mass pyrograms were reproducible to &lo% at each mass unit, but this reproducibility could only be maintained over a 6-h period.Irwin and Slack20 have noted the lack of reproducibility data for Py - MS and as irrepro- ducibility in this type of analysis has been quotedleP1* as a limiting factor in its routine use in forensic work, the purpose of the present work was to assess the reproducibility of the technique. Three sample types were pyrolysed on three separate pyrolysis systems, coupled in turn to one mass spectrometer; concomitant analyses of the sample types were effected by pyrolysis - gas chromatography (Py - GC) . Experimental Equipment (a) Pyrolysis - gas chromatography Inst rumen t . . .. Column . . .. .. Carrier gas .. .. Temperatures . . .. Detector . . .. .. Pyrol yser .. .. (b) Mass spectrometry Inst rumen t .. .. Electron impact conditions : emission current ..accelerating voltage . . source temperature . . mass range . . .. scan rate .. .. Pyrogram read-out . . electron energy .. Data acquisition . . .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. Varian Aerograph, Series 1400 12 ft x 1/8 in 0.d. stainless steel packed with 15% Carbowax 20M on Gas-Chrom Q ($0-100 mesh) Nitrogen, flow-rate 25 ml min-l Flame ionisation Detector 220 "C; injector 200 "C; oven 190 "C isothermal for pyrolysis of wires coated with poly- vinyltoluene - polystyrene solution ; temperature programmed from 70 to 190 "C at 10 "C min-l for paint samples Pye Curie-point, 15-s pyrolysis at 610 "C VG Micromass 12F (a magnetic sector mass spectro- meter) 70 eV 100 pA 4 kV 230 "C approximately 35-250 a.m.u. 1 s per decade with magnet re-set time of 1 s VG 2040 data system Bryans 26000 A4 X - Y plotter (c) Pyrolysis systems f o r mass spectrometry (i) Pye Curie-point pyrolyser, 15-s pyrolysis a t 610 "C.(ii) Chemical Data Systems (CDS) Pyroprobe 100, 20-s pyrolysis at 700 "C (nominal temperature). The instrumental system employed for these two pyrolysers was influenced by the need to switch rapidly to normal gas chromatograph-mass spectrometer operation, and it is essentially the system of Hughes et aZ.15 The two pyrolysers could be connected alternately336 HICKMAN AND JANE : REPRODUCIBILITY OF PYROLYSIS - MASS Analyst, VoZ. 104 to an empty 45 cm x 6.35 mm o.d., 2 mm i d . glass column, maintained at 200 "C (in the GC oven), and were operated with a purging helium flow-rate of 5 ml min-l (mass flow).A 5 ml min-1 make-up flow of helium was introduced at the end of the glass column con- nected to a length of stainless-steel tubing, 0.159 mm 0.d. A restrictor (of stainless-steel tubing) was inserted between this stainless-steel tubing and the length of glass-lined stainless- steel microbore tubing that led to the jet separator of the mass spectrometer. The restrictor could be adjusted (e.g., after cleaning the jets) to ensure that the pyrolysate arrived in the mass spectrometer a t a reproducible time after pyrolysis. (iii) This probe,21 which fitted in place of the direct insertion probe, incorporated a Curie-point pyrolyser. The main difference between this system and systems (i) and (ii) was that the pyrolysis products passed directly into the ionisation source of the mass spectrometer.Pyrolyses with this system were carried out with no flow in the line connected to pyrolysis systems (i) or (ii) and with the dump valve closed. VG pyrolysis probe, 15-s pyrolysis at 610 "C. Samples Three sample types were chosen to cover a range of ease of sample pyrolysis. (a) A simple system of easily pyrolysable polymers. Polymer-coated wires were pre- pared by dipping Pye 610 "C Curie-point wires into a solution of polyvinyltoluene and polystyrene in tetrahydrofuran to a depth of approximately 1 cm. The sample loading was estimated as being about 200 ng. (b) An acrylic paint sample, known from Py - GC studies to pyrolyse reproducibly (International Pinchin Johnson, Canary Yellow). (c) An alkyd paint sample, known from Py - GC studies to pyrolyse in a variable manner (Permoglaze, White Gloss).For sample types (b) and (c) the masses of the fragments of dried paint used for the Py - GC studies were in the range 10-30 pg, while for the Py - MS studies the masses were in the range 3-12 pg. Method The reproducibility experiments were carried out over a 6-week period; during this time the mass spectrometer was also being used in its GC - MS mode for the routine identification of drugs in body fluids. On each day of the reproducibility study duplicate samples of the two paint types were run on each of the four pyrolysis systems, and five polymer-coated wires were run on the three Curie-point systems. Several instrumental variables were monitored each day in case an instrumental mal- function produced irreproducible results.Readings of the source and backing pressures were taken, and the levels of the nitrogen and oxygen peaks were monitored, in order to detect (and thence remove if necessary) any appreciable air leaks in the system. Each day, before use with the pyrolyser, the mass spectrometer was tuned on a sample of toluene (0.4 pl injected through the vapour-phase inlet) to give the spectrum m/e (%I) : 91 (100); 92 (71-74); 65 (11-13); 63 (5-6); 51 (5-6); and 93 (5). The procedure adopted in the Py - MS experiments was to introduce the sample into the pyrolyser and allow the system to stabilise for approximately 3 min before initiating the repetitive mass spectral scanning. Background spectra were then collected for 20 s before pyrolysing. Between 30 and 60 spectra were collected, depending on the pyrolysis system and sample type.Using the on-line computer facility, background subtraction was followed by data manipulation using the specially written FORTRAN IV pr0grams.1~ The intensities of the ions at m/e 28, 32, 40 and 44 were set to zero, and several scans were averaged to produce mass pyrograms. Twelve mass spectral scans were averaged for pyrolysis systems (i) and (ii), while 32 scans were averaged when the VG pyrolysis probe was used; this was increased to 52 scans for the pyrolysis of the alkyd paint samples, in order to increase the proportion of phthalic anhydride and benzoic acid detected. As some mass spectrometer facilities may not have access to computers for data reduction, single scans were examined at a fixed time interval after pyrolysis (in fact a scan shortly after the maximum total ion current had occurred).These scans would give a measure of the reproducibility attainable without use of a data system.April, 1979 SPECTROMETRY USING THREE DIFFERENT PYROLYSIS SYSTEMS Mass Spectrometer Tuning Results 337 Over the period of the experiment two sources were used. One source could be tuned easily to give a consistent standard toluene spectrum. This source has subsequently been removed from and returned to the mass spectrometer several times, for dismantling and cleaning, but it is still capable of being tuned to the standard toluene spectrum. The second source gave a lower sensitivity (by a factor of 3) at its optimum tuning position, but at this position it gave a different toluene spectrum.The standard toluene spectrum could be obtained by tuning this source, but a further (%fold) decrease in sensitivity occurred, and the spectrum suffered from erratic intensity ratios. As the pyrolysis results obtained with the second source were distinctly different from those obtained with the first source, only the results from the first source are considered in this reproducibility study. It has yet to be determined whether the source discrepancy is basic to the two sources, or whether the second source contained a fault. Pyrolysis of Polymer-coated Wires Under the experimental conditions the wires coated with polyvinyltoluene - polystyrene solution pyrolysed to give essentially the vinyltoluene and styrene monomers (see Fig.1 for the pyrolysis - gas chromatogram). The pyrolysate mass spectrum, m/e (%I) : 104 (loo), 118 (66), 117 (63), 103 (50) and 78 (34), included ions at m/e 104 and 103 arising from styrene and ions at m/e 118 and 117 from vinyltoluene [in each instance the ions represented the molecular ion and (molecular ion -1)]. Monitoring the ratios of ions 118/104 and 117/118 gave a measure of the inter- and intra-compound variation, respectively. The results of both the Py - GC and Py - MS experiments with this sample type are summarised in Table I. aJ t: n. a2 K B J 4 7 10 5 0 T ime/min Fig. 1. Pyrolysis - gas chromatogram of poly- mer-coated pyrolysis wire. A, Styrene; and B, vinyl- toluene. The results show the following: (i), averaging a pyrolysis run over 12 scans improves the reproducibility of the two Py - MS systems by a factor of two; (ii), the over-all reproducibility of ion measurement (after pyrolysis plus mass spectro- metric processes) for ions from a single compound is better than that for ions from different compounds ;338 HICKMAN AND JANE : REPRODUCIBILITY OF PYROLYSIS - MASS Analyst, VoZ.104 TABLE I CURIE-POINT PYROLYSIS FOR POLYMER-COATED WIRES Ratio, yo r 1 Pyrolysis into VG pyrolysis Pyrolysis - gas glass column probe chromatography Ion (30 samples) (30 samples) (32 samples) 118/104 67.3 (2.5%) 62.7 (5.5%) 56.7 (3.1%) 117/118 92.8 (1.8%) 90.8 (3.2%) 118/104 76.6 (6.8%) 62.4 (9.9%) 117/118 96.3 (6.2%) 92.4 (8.4%) Mass spectrometev Yuns averaged (12 scans), relative standard deviations given in parentheses- Single scan considered, relative standard deviations given in parentheses- (Scan No.18) (Scan No. 15) (iii), the reproducibility of measurement of the vinyltoluene - styrene ratio on Py - GC is approximately the same as that for Py - MS using the glass column and is approximately twice as good as Py - MS with the VG pyrolysis probe. The intra-compound reproducibility is also worse for the VG pyrolysis probe. This irreproducibility probably arises from the small sample sizes, and the longer time period of introduction of the pyrolysate into the mass spectrometer, for the pyrolysis probe (Fig. 2). I 1 Scan number Fig. 2. Total ion current profiles for two pyrolysis systems using A, VG pyrolysis probe; and B, Curie-point polymer-coated wires. pyrolyser into glass column.Pyrolysis of Acrylic Paint Samples Py - GC (see Fig. 3) indicated that pyrolysis of this sample gave three major components, vix., butan-1-01, butyl methacrylate and styrene, and the Py - MS results are consistent with this observation. The results for the Py - GC experiments are given in Table I1 and in Tables 111-V the Py - MS results are summarised. For the Py - MS study, ten ions were monitored with their intensities normalised to m/e 41 as 100~o: m/e 39, 41, 55, 56, 57, 69, 78, 87, 103 and 104. Some of these ions could arise from more than one compound, but the ions at m/e 103 and 104 were characteristic of styrene; the ratio of 103 to 104 was used to examine the intra-compound variation. Ions from this group appearing in the spectra of the major components are: [m/e (yo I ) ] butan-1-01 56 (loo), 41 (54), 55 (14); butyl meth- acrylate 41 (loo), 39 (59), 69 (58), 66 (33), 87 (32), 57 (11); and styrene 104 (loo), 103 (38), 78 (31), 39 (11).As a means of measuring the similarity of pairs of (averaged) mass pyrograms, FIT factors were calculated. The definition of the FIT factor15: w x 2 + I V 2 ) FIT = 1000 [l -April, 1979 SPECTROMETRY USING THREE DIFFERENT PYROLYSIS SYSTEMS 339 W C Q M a C 4 u 20 10 0 Time/mi n Fig. 3. Pyrolysis - gas chromatogram of A, Butan-1-01; B, butyl acrylic paint sample. methacrylate ; and C, styrene. TABLE I1 PYROLYSIS - GAS CHROMATOGRAPHY OF ACRYLIC PAINT Proportion of major components (relative to styrene), relative standard deviations given in parentheses. Variation Butan-1-01 Butyl methacrylate Styrene In-day variation (11 samples) .. 72.6 (1.8%) 66.2 (2.6%) 100 Long-term variation (14 samples) . . 71.5 (2.3%) 69.3 (1.9%) 100 TABLE I11 Py - MS OF ACRYLIC PAINT USING CURIE-POINT PYROLYSIS INTO GLASS TUBE In-day variation: 10 samples, 12 spectra averaged for each run. Ion (m/e) . . .. 39 41 55 56 57 69 78 87 103 104 103/104 Intensity . . . . 31.9 100 33.8 66.4 36.2 61.8 22.1 35.0 28.5 66.0 43.2 Relative standard deviation, yo . . . . 1.5 - 1.9 2.9 2.5 1.2 2.0 1.4 1.9 1.7 1.3 FIT factors (45 comparisons) average 999.8, worst 999. Long-term variation: 13 samples, 12 spectra averaged for each run. Ion (m/e) . . . . 39 41 55 56 57 69 78 87 103 104 103/104 Intensity . . . . 29.9 100 33.7 74.1 36.4 58.2 20.8 34.0 28.7 66.1 43.4 Relative standard deviation, :h .. . . 7.2 - 6.0 4.6 4.9 4.6 4.1 7.2 9.9 9.6 1.7 FIT factors (36 comparisons) average 996, worst 991. Four samples were omitted from the FIT factor calculations as they suffered interference from the m/e 43 ion during background subtraction. Long-term variation : 13 samples, single scan. Ion (m/e) . . .. 39 41 . 55 56 57 69 78 87 103 104 103/104 Intensity . . . . 31.3 100 37.8 62.5 43.8 66.5 25.1 40.1 35.6 81.7 42.5 Relative standard deviation, yo , . . . 9.7 - 12 6.7 10 6.6 13 7.6 17 16 3.1340 HICKMAN AND JANE : REPRODUCIBILITY OF PYROLYSIS - MASS Analyst, VoZ. 104 TABLE IV Py - MS OF ACRYLIC PAINT USING CDS PYROPROBE PYROLYSIS INTO GLASS TUBE In-day variation: 10 samples, 12 spectra averaged for each run. Ion (m/e) . .. . 39 41 55 56 57 69 78 87 103 104 103/104 Intensity .. . . 39.5 100 44.4 51.6 39.1 52.6 28.4 27.2 37.4 82.1 45.6 Relative standard deviation, % . . . . 1.9 - 1.8 2.5 4.2 5.7 2.9 5.9 3.3 2.6 1.2 FIT factors (45 comparisons) average 998, worst 993. Long-term variation: 14 samples, 12 spectra averaged for each run. Ion (m/e) .. . . 39 41 55 56 57 69 78 87 103 104 103/104 Intensity .. . . 35.8 100 41.9 54.7 38.3 54.5 25.9 28.0 36.6 81.3 45.1 Relative standard deviation, % . . . . 5.2 - 4.4 3.9 4.6 6.0 4.3 3.8 6.8 7.3 1.4 FIT factors (91 comparisons) average 996, worst 987. Long-term variation : 14 samples, single scan. Ion (m/e) .. .. 39 41 55 56 57 69 78 87 103 104 103/104 Intensity .. . . 37.4 100 44.3 54.8 39.0 50.9 29.7 25.1 42.6 93.2 45.8 Relative standard deviation, % .. . . 5.9 - 3.8 5.8 6.1 10 3.4 12 4.7 4.7 3.0 where I , and I , are the intensities of a given ion (normalised to the total ion current of a particular pyrogram) in the two sets of data, means that perfectly matched pyrograms would give an FIT factor of 1000. From Table V it can be seen that there are no results given for the in-day variation with the VG pyrolysis probe. The in-day experiment was carried out after the long-term experiments, and after the probe had been used for a considerable amount of pyrolysis - CI work. The variation obtained was extreme, and was attributed to a layer of carbon having been deposited on the inside of the probe line. If pyrolysis - mass spectrometry was to be carried out routinely with the VG pyrolysis probe, then regular cleaning of the line would probably be necessary for optimum performance.The results given in Tables II-V show: the paint sample pyrolyses reproducibly for the three major pyrolysis products, as measured by Py - GC; the m/e 103/104 ratio agrees between the three mass spectrometer pyrolysis systems over a long period of time, indicating that the variation originating in the mass spectrometer should affect the over-all reproducibility by no more than 1-2%; and good reproducibility was achieved with the two systems pyrolysing into the glass column for both the short-term and the long-term experiments. The VG pyrolysis probe gave poorer results than the other two Py - MS systems, and this could be caused by some variation in the line temperature of the probe.In addition, the TABLE V Py - MS OF ACRYLIC PAINT USING THE VG PYROLYSIS PROBE Long-term variation : I 1 samples, 32 spectra averaged for each run. Ion (m/e) .. .. 39 41 55 56 57 69 78 87 103 104 103/104 Intensity .. . . 37.9 100 36.9 50.5 30.2 50.9 27.1 30.0 34.5 76.3 44.4 Relative standard FIT factors (55 comparisons) average 995, worst 982. Long-term variation: 13 samples, 32 spectra averaged for each run. Relative standard FIT factors (78 comparisons) average 991, worst 966. Long-term variation : 13 samples, single scan. Ion (m/e) .. .. 39 41 55 56 57 69 78 87 103 104 103/104 Intensity .. . . 38.0 100 35.2 48.2 28.3 48.7 27.2 28.9 35.4 79.5 44.4 Relative standard deviation, yo . . . . 6.3 - 9.5 13 15 12 10 15 16 13 5.1 deviation, yo . . - . 4.4 - 4.4 6.8 6.8 6.3 6.7 7.0 10 9.7 2.1 deviation, yo .. . . 5.3 - 5.0 9.6 10 9.1 9.5 10 13 13 2.9April, 1979 SPECTROMETRY USING THREE DIFFERENT PYROLYSIS SYSTEMS 341 spread in the time the pyrolysate takes to reach the mass spectrometer (Fig. 2) means that small samples achieve only a small total ion current and give more variable ion ratios. The influence of this factor is shown in the long-term variation experiments with the VG pyro- lysis probe, where an improvement in reproducibility was observed after the omission of the data from two small samples that gave low total ion currents. Typical mass pyrograms for the acrylic paint samples on the three Py - MS systems are given in Fig. 4. FIT factors between these three mass pyrograms are less than 900. 5~ 0 4 J m/e l o / _I 0 m/e Fig.4. Mass pyrograms for yellow acrylic paint. (a), Curie-point pyrolysis into glass column; (b), CDS Pyroprobe into glass column; and (c), VG pyrolysis probe.342 HICKMAN AND JANE: REPRODUCIBILITY OF PYROLYSIS - MASS Analyst, VoZ. 104 x % a 0 n B I t 1 10 0 Tim e/m i n Fig. 5. Pyrolysis - gas chromatogram of alkyd paint A, Acrolein; B, methacrolein; C, benzene; and D, sample. unidentified component. Pyrolysis of Alkyd Paint Samples Py - GC (see Fig. 5 for a typical pyrolysis - gas chromatogram) showed that the pyrolysis of this sample took place in a variable manner, and that the major pyrolysis products were acrolein, methacrolein, benzene and an unidentified component. The Py - MS results showed, in addition to these components, the presence of phthalic anhydride and benzoic acid.Table VI lists the results of the P y - GC experiments, and the Py- MS results are given in Tables VII-IX. For the P y - MS study 12 ions were monitored, with their intensities normalised to m/e 41 as 100% : m/e 39, 41, 55, 56, 57, 67, 70, 76, 78, 91, 104 and 105. Ions from this group appearing in the spectra of the major components are shown at the top of p. 343. TABLE VI PYROLYSIS - GAS CHROMATOGRAPHY OF ALKYD PAINT Proportion of major components (relative to acrolein) , relative standard deviations given in parentheses. Unidentified Acrolein Methacrolein Benzene component (10 samples) . . .. .. 100 80.4 (5.6%) 41.8 (31%) 24.0 (5.3%) (14 samples) . . .. .. 100 84.1 (7.4%) 28.0 (44%) 24.5 (6.6%) In-day vanation Long-term variationApril, 1979 SPECTROMETRY USING THREE DIFFERENT PYROLYSIS SYSTEMS Component Mass spectrum (relevant ions only) m/e (%I) 343 Acrolein .. .. .. . . Methacrolein . . .. .. Benzene . . .. .. .. Unidentified component.. .. Benzoic acid .. .. Phthalic anhydride' . . .. 56 (loo), 55 (75) 41 (loo), 39 (70), 70 (70) 78 (100) 56 (loo), 41 (80) 105 (loo), 39 (11) 104 (loo), 76 (85) The ion a t m/e 57 could have originated from a variety of hydrocarbon materials, that at m/e 67 from a variety of cyclic or unsaturated hydrocarbon compounds and that at m/e 91 could have come from a compound containing a benzyl group. The results given in Tables VI-IX show that Py - GC does not monitor all the pyrolysis products ; acrolein, methacrolein and the unidentified component are formed reasonably reproducibly, while benzene is formed very irreproducibly.No significant amounts of phthalic anhydride or benzoic acid are observed in P y - MS when using the Curie-point TABLE VII Py - MS OF ALKYD PAINT USING CURIE-POINT PYROLYSIS INTO GLASS TUBE In-day variation: 10 samples, 12 spectra averaged for each run. Ion (m/e) . . .. 39 41 56 56 57 67 70 76 78 91 104 105 Intensity . . .. 56.8 100 52.9 44.1 33.6 24.2 25.3 2.1 9.7 14.4 4.1 8.0 Relative standard deviation, yo . . 6.3 - 2.6 1.3 4.6 3.1 1.7 43 18 3.9 43 13 FIT factors (45 comparisons) average 996, worst 988. Long-term variation: 14 samples, 12 spectra averaged for each run. Ion (mfe) . . .. 39 41 55 56 57 67 70 76 78 91 104 105 Intensity . . . . 53.0 100 54.0 49.3 36.0 22.6 29.7 - 10.9 13.2 5.2 7.0 Relative standard deviation, % .. 5.8 - 6.1 6.5 8.3 4.5 6.8 - 24 8.4 61 22 FIT factors (45 comparisons) average 994, worst 977. Four samples were omitted from the FIT factor calculations as they suffered interference from the m/e 43 ion during background subtraction. Long-term variation : 14 samples, single scan. Ion (m/e) . . . . 39 41 55 56 57 67 70 76 78 91 104 105 Intensity . . . . 53.2 100 59.0 50.4 43.9 26.5 32.0 - 9.8 18.3 15.5 13.7 Relative standard deviation, yo . . 7.5 - 8.9 8.4 11 4.5 7.8 - 23 13 51 23 TABLE VIII Py - MS OF ALKYD PAINT USING CDS PYROPROBE PYROLYSIS INTO GLASS COLUMN In-day variation: 10 samples, 12 spectra averaged for each run. Ion (m/e) . . .. 39 41 55 56 57 67 70 76 78 91 104 105 Intensity . . .. 55.0 100 69.6 37.2 37.7 33.7 24.2 74.7 20.9 21.8 81.1 36.2 Relative standard deviation, yo .. 3.7 - 3.8 3.5 5.7 3.9 3.6 10 16 4.1 10 8.0 FIT factors (45 comparisons) average 995, worst 983. Long-term variation: 12 samples, 12 spectra averaged for each run. Ion (m/e) . . . . 39 41 55 56 57 67 70 76 78 91 104 105 Intensity . . . . 58.6 100 62.0 38.9 33.7 32.4 25.0 58.4 18.6 20.4 61.5 21.3 Relative standard deviation, yo . . 2.6 - 3.7 4.5 3.1 3.9 3.5 25 22 7.0 27 10 FIT factors (66 comparisons) average 978, worst 909. Long-term variation : 12 samples, single scan. Ion (m/e) . . .. 39 41 55 56 57 67 70 76 78 91 104 105 Intensity . . . . 59.7 100 61.7 37.4 32.0 33.5 25.3 55.9 24.4 23.2 59.0 23.4 Relative standard deviation, yo . . 4.2 - 5.4 4.7 6.6 5.1 5.5 20 22 5.4 20 8.6344 HICKMAN AND JANE : REPRODUCIBILITY OF PYROLYSIS - MASS Anabst, vol.104 TABLE IX Py - MS OF ALKYD PAINT USING VG PYROLYSIS PROBE In-day variation : 9 samples, 52 spectra averaged for each run. Ion (m/e) . . . . 39 41 55 56 57 67 70 76 78 91 104 105 Intensity . . . . 61.2 100 56.6 44.3 25.3 24.7 25.9 18.8 28.4 14.0 22.0 11.2 Relative standard deviation, yo . . 4.0 - 4.9 3.0 2.5 7.0 3.6 13 17 2.8 9.5 4.3 FIT factors (36 comparisons) average 994, worst 981. Long-term variation : 12 samples, 52 spectra averaged for each run. Ion (m/e) . . .. 39 41 55 56 57 67 70 76 78 91 104 105 Intensity . . . . 57.7 100 61.9 45.5 34.2 22.6 27.4 44.7 9.3 11.4 47.9 13.3 Relative standard deviation, % . . 6.4 - 11 9.6 19 9.4 12 22 27 21 17 32 FIT factors (66 comparisons) average 976, worst 930.Long-term variation : 12 samples, single scan. Ion (m/e) . . .. 39 41 55 56 57 67 70 76 78 91 104 105 Intensity . . . . 55.7 100 59.5 44.8 30.5 23.0 27.5 26.9 12.2 12.4 29.6 12.1 Relative standard deviation, yo . . 9.5 - 13 9.6 18 6.6 8.8 17 20 13 15 17 pyrolyser with the glass column. This is presumably caused by these components being deposited on the cold silica liner in the pyrolysis unit. The ions from the other pyrolysis products are present in greater proportions, but the variations in ion intensities are some- what greater than the variations observed for the acrylic paint, reflecting the irreproducible pyrolysis of the alkyd paint. The over-all reproducibility, as measured by the FIT factors, is good, even though all possible information is not available for the FIT factor calculations.All the pyrolysis products from the paint appear to be observed in Py - MS when using the CDS Pyroprobe. Good reproducibility is measured for some ions, although for other ions, especially those associated with benzene, phthalic anhydride and benzoic acid, the reproducibility is poor. The FIT factors show the over-all reproducibility to be similar to that obtained when using the Curie-point pyrolyser with the glass column. The VG pyrolysis probe initially produced spectra similar to those from the Pyroprobe, but after work on pyrolysis-CI mass spectrometry using this probe, much less phthalic anhydride and benzoic acid was observed. This is caused, presumably, by carbon being deposited in the pyrolysis probe line. The long-term variation with this probe showed some irreproducibility, but this may be caused by line temperature variations that were observed with this prototype probe.The variation in the proportion of benzene observed, as monitored by the m/e 78 ion, is smaller for the three Py - MS systems than for the Py - GC system. Possibly there is some contribution to the intensity of the m/e 78 peak from components that are produced more reproducibly than benzene. Typical mass pyrograms for the alkyd paint samples on the three Py - MS systems are given in Fig. 6. FIT factors between these three mass pyrograms are less than 900. Conclusions The results obtained probably do not reflect the optimum Py - MS experimental situation as the VG pyrolysis probe was a prototype model, and some experimental difficulties were encountered in its use.Additionally, some disruption of the glass column occurred during the repeated changeovers from the Curie-point pyrolyser to the CDS Pyroprobe and this disruption probably degraded the measured reproducibility. The results, however, do give an indication of the kind of reproducibility to be expected from Py - MS and the relative merits of each of the pyrolysis systems examined. The conclusions can be summarised as follows. 1. The irreproducibility arising from the mass spectrometer fragmentation and detection, as measured by the variation of the intra-compound ion ratios, was no more than 2%, even over a period of several weeks. Although only one source was used, this was subjected toAerd, 1979 SPECTROMETRY USING THREE DIFFERENT PYROLYSIS SYSTEMS 345 m/e m/e 100 (C) m/e Fig.6. Mass pyrograms for white alkyd paint. (a), Curie-point pyrolysis into glass column; ( b ) , CDS Pyroprobe into glass column; and (c), VG pyrolysis probe. a considerable amount of GC - MS analysis during this period. This shows that the mass spectrometric detection of a pyrolysate is unlikely to introduce large errors; the major factor in the irreproducibility of Py - MS will be the pyrolysis process. As the EI fragmentation pattern is dependent on the instrumental tuning, it is recommended that the mass spectrometer is tuned to the spectrum of a reference compound, in order to achieve good instrumental reproducibility. Toluene spectra obtained on two other low-resolution magnetic sector mass spectrometers were very similar to the spectrum obtained with the present instrument.Hopefully pyrolysis spectra would also be similar on different instruments, but this will have to be tested by experiment. A study such as this would be necessary in order to determine whether an inter-laboratory exchange of data 2.346 HICKMAN AND JANE : REPRODUCIBILITY OF PYROLYSIS - MASS Analyst, VoZ. 104 is feasible; previous experience with Py - GC has shown that most laboratories build up in-house reference collections, making inter-laboratory reproducibility less important than intra-laboratory reproducibility. Samples that pyrolyse simply and reproducibly, e.g., the polymer-coated wires and the acrylic paint, gave good Py - MS reproducibility. If long-term reproducibility can be maintained, then for these types of substrate it would be feasible to compile a library system for identification and to employ FIT factors for quantitative comparisons.The alkyd paint gave a more complex mass pyrogram than the acrylic paint (compare Figs. 4 and 6), and good reproducibility was observed only for some ions in its spectrum. FIT factors (using all ions) coupled with a library system would only be able to discriminate between alkyd paints of different chemical type. It may be possible to distinguish alkyd paints of a given type by selecting certain ions (e.g., by discriminant analysis) for comparison, and this topic is currently under investigation. For quantitative comparisons, Py - MS is more limited with respect to sample size, when compared with Py - GC; the Py - MS systems used would only operate satisfactorily over a relatively narrow range of sample mass.Averaging several scans gave more reproducible results, compared with taking a single scan, for those ions that were already reasonably reproducible (approximately 5%). The reproducibility was not improved for those ions that showed large irreproducibility (>lo%). 7. The best experimental results were obtained with the CDS Pyroprobe connected in line to a glass column, and with the VG pyrolysis probe. The advantages of the CDS Pyroprobe are that it handles smaller samples better, the pyrolysate arrives in the mass spectrometer over a relatively short time interval and greater proportions of the less volatile components are produced during the pyrolysis.Disadvantages lie in the slowness of changing samples, the fact that the gas flows are interrupted during the sample introduction and that the pyrolysate enters the mass spectrometer through a restrictor and the jet separator. Advantages in the use of the VG pyrolysis probe are that the Curie-point wires employed give a reproducible pyrolysis temperature, which is important for inter-laboratory compari- sons. The probe is fairly quick in analysis time, and it is possible to mount several samples in the Curie-point wires before starting the experiment. Instrumentally, it is a simpler system, and the pyrolysate passes directly into the ionisation source. On the other hand, pyrolysate introduction takes place over a long period of time, and gives a lower absolute sensitivity, but some of the troubles encountered in its use may be ascribed to the probe being an experimental model.3. 4. 5. 6. We are grateful to the Home Office Central Research Establishment, Aldermaston, for We thank Dr. M. J. Whitehouse for useful advice on the loan of the VG pyrolysis probe. all aspects of mass spectrometry. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. References Zemany, P. D., Analyt. Chem., 1952, 24, 1709. Bua, E., and Maneressi, P., AnaZyt. Chem., 1959, 31, 2022. Brant, P., Dibeler, V. H., and Mohler, F. L., J . Res. Natn. Bur. Stand., 1953, 50, 201. Meuzelaar, H. L. C., and Kistemaker, P. G., Analyt. Chem., 1973, 45, 587. Posthumus, M. A., Boerboom, A. J. H., and Meuzelaar, H. L. C., Ado. Mass Spectrom., 1974, 6, 397. Meuzelaar, H. L. C., Haider, K., Nagar, B. R., and Martin, J. P., Geoderma, 1977, 17, 239. Meuzelaar, H. L. C., Posthumus, M. A., Kistemaker, P. G., and Kistemaker, J., Analyt. Chem., Anhalt, J. P., and Fenselau, C., Analyt. Chem., 1975, 47, 219. Maugh, T. H., Science, N. Y., 1976, 194, 1403. Simmonds, P. G., Appl. Microbiol., 1970, 20, 567. Schulten, H.-R., and Gortz, W., Analyt. Chem., 1978, 50, 428. Luderwald, I., and Urrutia, H., in Jones, C. E. R., and Cramers, C. A., Editors, “Analytical Pyro- lysis, ” Proceedings of 3rd International Symposium, Amsterdam, 1976, Elsevier, Amsterdam, 1977, p. 139. Hummel, D. O., in Jones, C. E. R., and Cramers, C. A., Editors, “Analytical Pyrolysis,” Proceedings of 3rd International Sympsoium, Amsterdam, 1976, Elsevier, Amsterdam, 1977, p. 117. Zeman, A., in Wiedmann, H. J., Editor, ‘Thermal Analysis,” Volume 3, Proceedings of 3rd Inter- national Conference on Thermal Analysis, Davos, 1971, Birkauser Verlag, Basle, 1972, p. 219. 1973, 45, 1546.April, 1979 SPECTROMETRY USING THREE DIFFERENT PYROLYSIS SYSTEMS 347 15. 16. 17. 18. 19. Hughes, J. C., Wheals, B. B., and Whitehouse, M. J., Forens. Sci., 1977, 10, 217. Hughes, J. C., Wheals, B. B., and Whitehouse, M. J., Analyst, 1978, 103, 482. Saferstein, R., and Manura, J. J.. J . Forens. Sci., 1977, 22, 748. Hughes, J. C., Wheals, B. B., and Whitehouse, M. J., Analyst, 1977, 102, 143. Meuzelaar, H. L. C., Kistemaker, P. G., Eshuis, W., and Engel, W. H. B., in Johnston, H. H., and Newsom, S. W. B., Editors, “Rapid Methods and Automation in Microbiology,” Proceedings of 2nd International Symposium on Rapid Methods and Automation in Microbiology, Cambridge, 1976, Learned Information (Europe) Ltd., 1976, p. 225. 20. 21. Irwin, W. J., and Slack, J . A., Analyst, 1978, 103, 673. Ardrey, R., Smalldon, K., Hickman, D. A., and Whitehouse, M. J., to be published. Received August 30th, 1978 Accepted October 18th, 1978

ISSN:0003-2654    

DOI:10.1039/AN9790400334

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

348 Analyst, April, 1979, Vol. 104, pp. 348-357 Interference Films on the Sensor Membranes of Solid-state Copper( I I) Ion-selective Electrodes G. J. Moody, N. S. Nassory and J. D. R. Thomas D. Betteridge, P. Szepesvary” and B. J. Wright Chemistry Department, University of Wales Institute of Science and Technology, Cardifl, CF1 3N U Chemistry Department, University College of Swansea, Singleton Park, Swansea, SA 2 8PP Copper (11) ion-selective electrodes based on copper(I1) sulphide - silver sulphide sensor membranes, which showed anomalous responses with copper(I1) nitrate in the presence of chloride, have been examined by Auger spectrometry. In some electrodes exposed to solutions of potassium chloride the chloride is found to have penetrated the bulk of the membrane matrix, whilst in others only a surface contamination is observed.The anomalous electrode response is exhibited when exposure to chloride is in the presence of copper nitrate. The Auger signal alters during the duration of the spectrum as a consequence of electron bombardment. The effects of argon-ion and electron bombardment are compared. Keywords : Ion-selective electrodes for copper(II) ; interference films on ion- selective electrodes ; A uger spectrometry A common type of interference encountered with solid-state ion-selective electrodes1 involves interaction between the interferent and sensor membrane component (s) to form new com- pound(s) of varying degrees of solubility. Thus, the Orion 94-35 bromide electrode can tolerate thiocyanate to a level dictated in principle by the quotient of the respective silver salt solubility products, &oAgSCN/K~oAgBr m 3.Beyond this value the silver thiocyanate film formed on the surface of the silver sulphide - silver bromide sensor solicits a near- Nernstian response to thiocyanate SCN- + AgBr(s) C+ AgSCN(s) + Br- Fortunately, no permanent damage is incurred and the offending film is removed with emery cloth when the “thiocyanate” electrode reverts to its original bromide-sensing role.1 However, interference from iodide ions can be considerably less than would be predicted on the basis of solubility product ratios according to the extent of the interference deposited on the sensor surface of a chloride ion-selective electrode.2 An interesting effect is observed with the Orion 94-29A copper solid-state ion-selective electrode immersed in copper(I1) nitrate solution. Chloride leakage from a calomel reference electrode, or deliberate addition of the chloride ion, rapidly dulls the normal shiny sensor surface and the electrode gives a chloride, not copper, ion activity response.Nevertheless, the original copper function is completely restored after polishing about half of the dulled sensor surface, This copper to chloride to copper response cycle has been repeated many times and with different Orion 94-29A electrodes3 The chemical nature of this chloride-responsive electrode film remains unidentified, but it may consist of a chloride species arising by, say Ag,S(s) + Cu2+(aq) + 2C1- e CuS(s) + ZAgCl(s) Preliminary attempts to identify chloride complexes in simulated silver sulphide - copper( 11) sulphide materials were unsu~cessful.~ Therefore, a variety of copper( 11) sulphide - silver sulphide - potassium tetrachlorocuprate( 11) discs have been exposed to potassium chloride or copper( 11) nitrate - potassium chloride solutions and their surfaces examined by Auger spectroscopy* in a further attempt to establish the nature of the interference. In addition, the potentiometric behaviour of some selected discs has been measured before and after such * On leave from The Institute of Isotopes, Budapest, Hungary.MOODY, NASSORY, THOMAS, BETTERIDGE, SZEPESVARY AND WRIGHT 349 chemical exposures (Table I) but, owing to the fragile nature of the discs, it was not possible to examine their potentiometric behaviour after Auger examination.TABLE I COMPOSITION AND TREATMENT OF ELECTRODE DISCS PRIOR TO AUGER* EXAMINATION Slope, S, (in mV per decade) and lower linear calibration range (R) of electrode in standard solutions at 25 "C Disc Disc composition I A 1 series (molar ratio) Soaking treatment After soaking Before soaking Orion CuS: Ag,S (2 : 3)t KCl (1.3 M) for 1 week No C1- response in KCI S = 30.8; R = M in Cu(NO,), M in CuSO, M in Cu(CI04), 0.5 cms KCI (1.33 M) for cUso, S = 27.0; R = M in Cu(OAc), 10 min S = 31.0; R = 10-6 M in Cu(NO,), S = 28.0; R = No Cu*+ response in Cu(NO,), or S = 28.5; R = C1- response in both KCI and CuCI, standards. S-61; R-10-4 M S = 29.8; R = 2 x 10-6 M in 20 cma Cu(NOa)a (lo-' M) 4- S = 29.8; R = 2 X 10-6r.1 in Cu(NOa)s 1 Cu(NO,), - Cu(NO,), but S = 25 mV above No Cu'+ response below 10-8 M No C1- response in KCI and Cul+ re&onse restored after Cu(NO,), / 94-29 electrode A B CuS: Ag,S (1 : 1) Control C CuS : Ag,S (1 : 1) KCl (1.3 M) for 5 min CuS: Ag,S (1 : 1) 20 a* Cu(NO,), (10-1 M) + S = 30.0; R = 2 x M in 0.5 cm3 KCI (4 M) for 10 min M Cu(NO,), CuCl standards polishing membrane D CuS:Ag,S (1:9) KCl(l.3 M) for 5min S = 30.3; R = 5 x 10-LM in S = 30.3; R = 5 x M in E CuS:Ag,S (1:9) + KCl (1.3 M) for 5 min F CuS:Ag,S (1:l) + KCI (1.3 M) for 5 min C W O s h Cu(NO,), S = 10; R = 10-4 M in Cu(NO,), S = 15; R = M in Cu(NO,), S = 1 2 ; R = M in Cu(NO,), S = 14; R = lo-' M in Cu(NO,),.K,[CuC141 (1% mlm) K,[CuCl4I (1% mlm) No C1- response No C1- response * Auger examination not feasible with the whole electrode assembly.t See Frant and Ross.5 Experimental Electrodes Discs of various compositions were prepared by pressing3 from mixed metal sulphides after co-precipitation from solution and very thorough washing with de-ionised water. Electrodes were then assembled from hollow poly(viny1 chloride) (PVC) bodies with the sensor discs sealed into one end, using the epoxy resin cement, Araldite.g An internal silver - silver chloride reference electrode immersed in a copper(I1) nitrate (10-1 M) and potassium chloride (10-3 M) internal filling solution was then sealed in a t the other end of the PVC stem. With only copper chloride as the internal solution calibration graph slopes were less than 20mV per decade. All of the electrodes were calibrated in solutions of copper(I1) sulphate, copper(I1) acetate, copper perchlorate, copper(I1) nitrate or potassium chloride, as appropriate, in conjunction with an Orion double-junction reference electrode containing 4 M potassium nitrate solution.The potentials of these cells were then recorded with a Beckman 4500 digital pH meter by using various test solutions kept at 25 0.1 "C. Auger Electron Spectroscopy (AES) Discs for AES were treated as in column 3 of Table I, drained, air-dried and examined on a retarding field analyser Auger spectrometer made by Vacuum Generators, Sussex. The spectrometer was supplemented by an additional electronic device controlling the potentials of electrostatic plates, thus ensuring convenient deflection and improved focusing of the primary electron beam.The observed energies of the emerging electrons were displayed by an electronic digital voltmeter and the Auger spectra drawn by an X-Y plotter (Advance- HR-2000). The pressure in the analyser was about 1.3 x 10-8 Pa (10-lO Torr). The specimen was bombarded with 1.1-keV electrons at an angle of incidence of about 15", the beam of the primary electrons being about 1 mm in diameter. The spot of the specimen surface hit by the electrons showed, therefore, a nearly elliptical shape with axis lengths of 4 and 1 mm. The primary beam current was about 70pA and the sample bias 90 v.350 MOODY et aE. : INTERFERENCE FILMS ON THE SENSOR MEMBRANES Analyst, Vol. 104 Spectra were normally taken in the 0400 eV energy range (some up to 1000 eV) and repeated several times at different modulating voltages (RMS).The sweep time applied was usually 1000 s per scan, that is, 0.4 eV s-l. The focusing of the incident electron beam was carried out by maximising the intensity of the 272-eV carbon signal. Neither the position of the specimen nor the electrical settings were changed in the course of the investigation of a given sample. The Auger transitions used are listed in Table 11. Representative spectra are shown in Figs. 1-3. The copper peak at 63 eV was not clearly resolved in many of the spectra because of the electron-energy analyser incorporated in the spectrometer. Consequently, no quantitative measurements of copper were attempted. TABLE I1 AUGER SIGNALS CHOSEN FOR THE ELEMENTS' Element cu S c1 K C Ag tg Z 29 16 17 19 6 47 47 8 Nominal energy1eV Auger transition 63 M3M4M4 262 L,MZ,,MZ,, 272 KL&, 304 M*N2,3M4,5 366 M,N4,ciN4,, 610 KLZLZ 162 L3M2,3M%3 181 L3M2,3M2,3 The carbon peak is a well known problem in surface-sensitive techniques and arises principally from surface adsorbed carbon dioxide and, possibly, diffusion pump oil vapour.It is usually removed during the early stages of argon-ion bombardment. Persistence of carbon through a series of bombardments (for example, Fig. 6) may therefore be taken to indicate the presence of carbon in the sample, for example, in a carbonate impurity. How- ever, it is not posqible to draw firm qualitative or quantitative conclusions from the available evidence. Argon-ion bombardment (Fig. 2) for semi-quantitative depth profiling was performed at argon pressures of about 1.3 x Pa (lo-' Torr).The source of the argon ions was the ionisation gauge, built-in for pressure measurements. The usual bombarding current was 20-30pA. In the initial work the specimen was turned to face the ion gun in order to achieve a perpendicular ion impact; later, however, its position was fixed at an incident angle of about 20". The bombardment period (5-10 min) was repeated 8-10 times. I t I I 200 360 4a Energy Scale/ eV Fig. 1. Representative Auger spectrum of disc A with hand-drawn energy scale. Scan speed, 0.4 eV s-1; RMS, 3 V.April, 1979 OF SOLID-STATE COPPER(II) ION-SELECTIVE ELECTRODES 351 Fig. 2. Auger spectra of disc A with hand- drawn scales. (a), Before argon-ion bombard- ment; (b), after 740 pA min argon ion bom- bardment.Scan speed, 0.4 eV s-l; RMS 4 V. T Fig. 3. Auger spectra of disc A with hand-drawn scales showing evidence of a new peak a t 218 eV attributed to trapped argon (between the C1 and K signals) in the later electron irradiation exposure of the right-hand spectrum. Scan speed, 0.4 eV s-1; RMS, 4 v.352 Auger Spectroscopy The Auger signal is proportional to the amount of the element from which it is derived but the proportionality factor depends on the element and the particular transition. In this work the changes in signal are taken to indicate changes in element concentration but no attempt has been made to relate this numerically to the actual percentage chemical composition. As a guide, an argon-ion etching of 1 pA min corresponds roughly to the removal of 0.1 nm of material.However, in these samples the combined effects of electron bombardment by the source and argon-ion etching make the application of simple guide lines difficult. MOODY et at. : INTERFERENCE FILMS ON THE SENSOR MEMBRANES Analyst, VoZ. 104 Results and Discussion Efects qf electron bombardment In electron-bombardment examinations of an electrode disc (A in Table I) by AES, the same spot on the sample was repeatedly scanned and all signal intensities changed with time (Figs. 1 and 4). The signals for chloride and potassium decreased at unequal rates (Fig. 4), while carbon and silver showed more marginal changes (Fig. 4). I I 20 40 60 80 100 120 Time/min Fig. 4. Summarising diagram of changes of Auger signal intensities with time for disc A on electron bombardment.Scan speed, 0.4 eV s-l; RMS, 3 V. 0, Silver; 0, chlorine; v, carbon; and +, potassium. Any vacuum effects, for example, evaporation or sublimation, were discounted because a sample does not necessarily sputter evenly and because different spots on the same disc gave similar patterns. The huge increase in the sulphur signal can cause overloading and for this reason it is not shown in all of the figures. However, the general shape of the overloaded sulphur peaks indicated substantial increases in the sulphur signals. Clearly, the surface undergoes alterations as a result of electron irradiation during the time of recording the Auger spectrum. The resultant pitting of the surface was confirmed by a scanning electron microscope.Hence, AES was supplemented by argon-ion bombard- ment. Efect of argon-ion bombardment The usual practice, in Auger spectroscopy, for removing surface contaminants is to fracture the specimen, scrape its surface in a very low vacuum or to sputter the surface with highly energetic ions of noble gases. This last method is likewise applicable to surface depthA + d , 1979 OF SOLID-STATE COPPER(II) ION-SELECTIVE ELECTRODES 353 profiling, as noble gas ions hitting the surface with appropriate kinetic energy are able to remove surface layers many hundred hgstroms in thickness, which is at least ten-fold greater than the thickness of the layers effectively analysed by the electron beam. In this work, argon-ion sputtering was used and the following points were qualitatively concluded for disc A: that the surface layer of carbon is removed during sputtering; that sputtering tends to “restore” the surface and slightly moderate the changes induced by electron bombardment (effects that lie on the border between significant and insignificant) ; after each sputtering, that is, on each newly exposed surface, electrons affect the specimen in the same way as before; and that changes induced by electrons in various scans are comparable to those caused by ion bombardment.These conclusions are illustrated for the combined effects of electron and argon-ion bombardment for disc A in Figs. 5 and 7 and disc D in Fig. 6. It will be noted that despite the effects of electron bombardment there are over-all trends that are made evident by argon-ion bombardment and these are shown in Fig.8. Some general results are presented in Table 111. These are obviously tempered by the considerations given above, but are taken to refer to profiles from 0.01 to 1Opm. It should be noted that the ratios in the ordinates of Figs. 5 , 6 and 8 relate to signal intensities and not to actual chemical composition. I+ 200 400 600 800 1 000 Electron irradiation t i m e h i n Fig. 5. Changes, for disc A, with electron irradiation time interspersed by argon-ion bombardments of contributions of individual signals expressed as percentages of total of signals for the various species. 0, Chlorine; 0, silver; +, potassium; v, sulphur; v, carbon. From the above it can be deduced that both potassium and chloride ions have penetrated the membrane (A) to an extent approaching 0.1 pm.This distance can be estimated on the basis of removing 0.1-nm layers for each 1 p A min of argon-ion bombardment. How- ever, in view of the changes that can take place in the sample consequent on bombardment, the calculations are, at best, a p p r o ~ i m a t e . ~ ~ ~ In principle, Auger spectroscopy can distinguish oxidation states of elements, but the resolution of the present instrument was unfortunately too low for this purpose. The results for the several discs examined are summarised in Table IV.354 MOODY et al. : INTERFERENCE FILMS ON THE SENSOR MEMBRANES Analyst, VoZ. 104 Potentiometric Measurements and Their Relationship to AES Results The potentiometric performance of electrodes with sensor discs composed of copper(I1) sulphide - silver sulphide in a 1 : 1 molar ratio (discs A, B and C) compared favourably with 5 - 4 - Y 2 G 3- \ 2 - 1 - 2ol t I l l 1 0 0 .0 . + +++ + TI' $$I I 1 I 4 5 6 7 8 9 ;v 600 00 V V 30 t + V V VV 0 O 0 0 o+ + + +O 0 0 I 1 800 vv 0 0 00 ++ I 1000 V a 0 + - v v 0 . 0 0 + + Electron irradiation t imehi n Fig. 6. Changes, for disc D, with electron irradiation time interspersed by argon-ion bombard- ments of contributions of individual signals expressed as percentages of total of signals for the various species. 0, Chlorine; 0, silver; +, potassium; D, sulphur; and v, carbon. the commercial Orion 94-29 electrode (Table I). Disc D, with a 1 : 9 molar ratio of copper(I1) sulphide and silver sulphide, behaved similarly (Table I), but sensor discs of 9 : 1 molar ratio of copper(I1) sulphide and silver sulphide were too fragile to be useful for electrode con- Electron irradiation time/min Fig.7. Change of ratio of chlorine: potassium signal intensities with time for disc A on both argon-ion and electron bombardment. The vertical lines indicate argon ion bombardment. p A min: (1) 130; (2) 200; (3) 208; (4) 200.April, 1979 OF SOLID-STATE COPPER(II) ION-SELECTIVE ELECTRODES 355 Argon ion exposure/mA rnin Fig. 8. Effect of argon-ion bombardment on composition of discs A and D. The values correspond to the first spectrum obtained after the bombardment and are normalised to percentages of total K, S, C1, C and Ag signals. (a), Disc D; and (b), disc A. 0, Chlorine; 0, silver; +, potassium; v , sulphur; and v, carbon.struction and in any case responded poorly to copper standards. This fact illustrates the importance of incorporating sufficient silver sulphide into the sensor matrices of many solid- state ion-selective electrode^.^ In addition, the resistivity of the discs (measured with a Multimeter) is a function of the silver sulphide content as observed for silver chloride - silver sulphides and silver bromide - silver sulphide6 sensor discs. TABLE I11 EFFECT OF ARGON ION BOMBARDMENT ON ELECTRODE DISC A Signal intensitieslmm Time of electron RMS A \ irradiation/min voltage C1 K C I C l / I K Comment 214 2 100.5 20.5 10.5 4.9 490 2 28.8 14.5 10.8 1.99 51 1 4 136 79.2 61 1.72 567 2 31.2 1.46 11.5 2.14 After 130 p A rnin Ar+ 649 668 2 4 15 13 63.1 71 TS* 9 i::; }After a further 200 pA min Ar+ 802 2 20 10.8 TS* Refocus, further 208 pA min Ar+ 814 4 83.8 64.4 17.7 bombardment 846 4 65.5 61.2 16.1 864 4 48.6 81.8 948 2 44.2 18.5 0 2.39 After further 200 pA min Ar+ 988 3 97.4 52.2 0 1.87 }bombardment * TS = Too small to measure.Electrodes with discs (E and F) containing 1% m/m of potassium tetrachlorocuprate(I1) dihydrate respond poorly to copper(I1) nitrate. The possibility that this complex may be related to the primary interferent species must be tempered by the fact that electrodes incorporating it do not respond to chloride (Table I). Also, infrared spectroscopic examina- tion (2004000 cm-1) of this finely ground disc material in potassium bromide matrices showed no definite evidence of copper chloride complexes.Auger spectroscopy could not provide evidence on this matter either and there was, in addition, an inability to detect potassium owing to overlapping carbon signals (Table IV). The response of various copper(I1) ion-selective electrodes is modified by fluoride and chloride to an extent dependent on age, condition of membrane surface, the presence of356 MOODY et al. : INTERFERENCE FILMS ON THE SENSOR MEMBRANES Analyst, VoZ. 104 acids and their associated anions, oxygen tension and exposure timeslo The absence of short-term chloride effects for new discs (discs A and D) has also been found for freshly polished membranes.1° Significant chloride interference arises only when potassium chloride is added to copper( 11) nitrate solutions in which the Orion 94-29 electrode or the models fitted with various pressed silver sulphide - copper(I1) sulphide discs are immersed or when these electrodes are immersed in copper(I1) nitrate - potassium chloride solutions.This behaviour is apparent from the results given in Table I for the Orion 94-29 electrode and that made from disc C, for which the Auger spectroscopic data are given in Table IV. Similar behaviour was also observed for an electrode made from a disc of copper(I1) sulphide - silver sulphide (1 : 9) and soaked in 20 cm3 of copper(I1) nitrate solution (10-1 M) and 0.5 cm3 of potassium chloride solution ( 4 ~ ) . The presence of copper(I1) nitrate in some way sensitises the membrane to inter- ference from chloride. Auger spectroscopy can confirm this insofar as nitrate is present in the membrane matrix for disc C.Disc series TABLE IV RESULTS OF AUGER SPECTROSCOPIC EXAMINATION OF VARIOUS SENSOR DISCS Comments on electron and argon-ion impact experiments A Sulphide signals increase with depth. Reference disc: evidence of copper, silver, sulphide and carbon only Steady pattern with depth, showing evidence of silver, copper, potassium, sulphide and chloride ions. Nitrogen (from nitrate) is observable and constant after a 250 pA min argon-ion bombardment Potassium and chloride ions penetrate the membrane to an extent approaching 100 nm B C D Little evidence of sulphide at the surface where potassium and chloride are predominant. Evidence for sulphide constant with depth. Potassium was not detected, possibly because of Sulphide increases with depth, while chloride falls.Potassium not detected, possibly because Sulphide increases with depth while chloride and potassium decrease overlapping carbon signal. of overlapping carbon signal. E Chloride, silver and carbon remain constant with depth Silver and carbon levels remain constant with depth F There is no clue to the operative mechanism, although it probably involves copper in conjunction with chloride, as soaking the Orion 94-29 electrode, or electrodes made with pressed membranes of copper(I1) sulphide - silver sulphide (1 : l), in potassium chloride solution alone or in potassium chloride plus potassium nitrate solution does not affect Cali- bration with copper(I1) nitrate. On the other hand, the observation noted above, con- cerning the more rapid decrease in the chloride signal in comparison with that for potassium during electron and argon-ion bombardments of disc A, suggests that some complexation of chloride takes place in the surface disc layers.This may be upset by the presence of copper( 11) nitrate, although the observed patterns neither confirm nor discount this. The massive rise in the Auger sulphur signal with depth (Table IV) could be caused partly by surface sulphur arising from membrane oxidation,ll which either diffuses away or is reflected in a relatively weaker elemental sulphur signal. CuS(s) + go, + 2H,O+ -+ Cu2+(aq) + 3H,O + S(s) Unfortunately, AES cannot distinguish sulphur from sulphide in this work. Also, there is no evidence of the e.m.f. values of the copper ion selective - reference electrodes system changing in a positive direction, as would be expected from a release of copper ions.Conclusion The study of surface layers of ion-selective electrode discs of copper(I1) sulphide and silver sulphide by use of Auger spectrometry, involving argon-ion and electron bombard- ment, shows that chloride penetration of the membrane disc occurs. This phenomenonApril, 1979 OF SOLID-STATE COPPER(II) ION-SELECTIVE ELECTRODES 357 can be responsible for interference with the copper ion-selective electrodes by chloride, although the presence of copper(I1) nitrate appears to be necessary for stimulating such interferences. The authors thank the University of Technology, Baghdad, Iraq, for paid leave of absence (to N. S. N.), The Royal Society for a Visiting Fellowship (to P. S.), the University of Wales for a Studentship (to B. J. W.), Mrs. P. Connors for technical assistance and M. Jones (Department of Mechanical Engineering, University College, Swansea) for Auger spectro- meter facilities. 1. 2. 3. 4. 6. 6. 7. 8. 9. 10. 11. References Orion Research Inc. Newsl., 1969, 1, 29. Klasens, H. A., and Goossens, J., Analytica Chim. Acta, 1977, 88, 41. Crombie, D. J., Moody, G. J., and Thomas, J . D. R., Talanta, 1974, 21, 1094. Taylor, N. J., in Burnham, R. E., Editor, “Techniques of Metal Research,” Volume VII, Part 1, Frant, M. S., and Ross, W. J., German Offen., 1942 379 (cl. G 01 n), March 12th, 1970. Fyfield, R. C., Moody, G. J., and Thomas, J. D. R., unpubli$ed work. Briggs, D., in West, A. R., Editor, “Molecular Spectroscopy, Heyden, London, 1977, pp. 468-481. Holm, R., and Storp, S., in West, A. R.. Editor, “Molecular Spectroscopy,” Heyden, London, 1977, Orion Research Inc., Br. Pat. 1 150 698, April 30th, 1969. Midgley, D., Analytica Chim. Acta, 1976, 87, 19. Biswas, A. K., and Mohan, N. P., J . Appl. Chem. Biotechnol., 1971, 21, 15. Interscience, New York, 1972, pp. 216-219. pp. 482-502. Received August 21st, 1978 Accepted November 2nd, 1978

ISSN:0003-2654    

DOI:10.1039/AN9790400348

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

358 Analyst, April, 1979, Vol. 104, pp. 358-366 Determination of Nitrite Ion in Unused Cutting Fluids and Cutting Oils Using a Gas-sensing Electrode Method Ferrers R. S. Clark and Hart B. MacPherson" Product Safety Laboratory, Department of Consumer and Corporate Afjairs, Tunney's Pasture, Ottawa, Ontario, Canada, K1A OC9 Modifications of the Orion NO, gas-sensing electrode method that were made to determine nitrite ion in unused cutting fluids and cutting oils are described. Detection limits for both types of lubricants of the order of 15pg g-l of NO,- ion were obtained. Previous analysis of six cutting fluids, collected in the Ottawa region and analysed by spectrophotometry, confirmed the presence of high levels of nitrite ion and showed fair agreement between results. Analysis of six other cutting fluids and 20 cutting oils collected in the same region showed the presence of nitrite ion in only three instances.The operation of the electrode, interferences, the use of standard-addition and -subtraction methods and the possibility of applying this method to used cutting lubricants are discussed. Keywords : Nitrite-ion determination; gas-sensing electrode ; nitrogen oxide electrode ; cutting f l u i d s ; cutting oils Metal-working lubricants are used in cutting, grinding, rolling and drawing operations in many industrial and trade applications.'y2 They are generally referred to as oils if they contain a hydrocarbon base (either a petroleum fraction or animal fat) and fluids or solutions if they are synthetic chemical solutions without a hydrocarbon base.Semi-synthetic lubri- cants are mixtures of oils and fluids. The oil-type lubricants can be made to mix with water by addition of a suitable emulsifier, or can be left non-emulsifiable. Fluid-type lubricants are usually water soluble. All of these products are available in Canada and some typical formulations have been ~ o m p i l e d . ~ - ~ The presence of nitrite salts, added to cutting fluids as rust inhibitors (and therefore present in semi-synthetic cutting oils), has recently become of interest to health regulatory and consumer protection agencies owing to the reaction between nitrite ion (or nitrous acid) and the common components, diethanolamine and triethanolamine (acting as either rust inhibitors or as part of the emulsifying agent) to form N-nitrosodiethanolamine,*J regarded as one of the most potent of animal carcinogens.Danger to both the animal and human organism has been reviewed in detail6*' and, indeed, deaths in China have been attributed to this compound.* In the Ottawa region, up to 45% of triethanolamine, 18% of sodium nitrite and 3% of N-nitrosodiethanolamine have been found in cutting fl~ids,~,lO levels that are considered to present a real danger to machinists through skin contact and mist inhala- tion. Other precursors can participate in the formation of N-nitrosamines under widely varying conditions.6*11 ,12 While nitrite salts have been used as rust inhibitors in cutting fluids, nitrites have not been added to non-emulsifiable lubricants. In addition, most companies faced with the possibility of legislation in the USA and Canada have removed nitrites from lubricant formulations or have removed the product from the market.Ethanolamine emulsifiers, however, have been retained. This laboratory was charged recently with the analysis of unused cutting oils (this term will subsequently refer to water-emulsifiable cutting oils) and cutting fluids available in Canada for N-nitrosodiethanolamhe. As this material may appear some time after mixing the p r e c ~ r s o r s , ~ ~ ~ ~ one method of control considered was the prohibition of the simultaneous presence of nitrite ion and di- or triethanolamine. Sensitive methods of detection of these * Present address : National Research Council of Canada, Chemistry Division, Analytical Section, Ottawa, Ontario, Canada, K1A OR9.CLARK AND MACPHERSON 359 precursors were sought because their safe concentration in cutting oils or fluids is not known, but is thought to be very low.The methods considered for nitrite-ion determination were restricted to those for free aqueous nitrite ion, or nitrite ion bound sufficiently weakly to other agents such that a simple room-temperature separation method would release it for analysis. Although spectrophotometric methods are sensitive (detection limits are in the low micrograms or nanograms per millilitre range), n u m e r ~ u s , ~ ~ - ~ ~ and are used as the basis of several standard r n e t h ~ d s , l ~ - ~ ~ it was envisaged that the water-soluble or micellar material in cutting oils could lead to spectral interference, unless significant dilutions of the sample were made ; hence spectrophotometric methods were considered inadequate.Ion-exchange liquid chromatographic methods using either conductimetric21*22 or ion-selective electrode23 detection, capable of measuring concentrations down to 5 and 46pgml-l of nitrite ion, respectively, were not investigated because the detection limits were too high. While the indirect, differential pulse-polarographic method reported by Boese et aZ.24 can optimally detect 0.046 pg ml-l of nitrite ion, there was no reason to believe that the probability of encountering interfering electroactive compounds in cutting fluids or oils would be low and, further, the method of standard additions is not possible because of curvature of the calibra- tion graph.Although a calorimetric method was reported by Hansen et aZ.25 to be highly selective, the sensitivity (detection limit about 69 pg ml-l of nitrite ion) was not adequate. On the other hand, nitrogen oxide (NO,) gas-sensing electrodes (e.g., Orion, Model 95-46) are economical, can detect concentrations of nitrite ion between 0.046 and 460 pg ml-l and have analysis times comparable to those of the above methods. Additionally, the Orion electrode method has been shown to give increased precision and accuracy over two standard spectrophotometric methods, and not to respond to colour, turbidity and many diverse ions.26-28 The subject of this paper is the adaptation of the Orion electrode method to the quantita- tive analysis of unused cutting fluids and cutting oils for labile nitrite ion.Results of the analysis of several fluids and oils are presented, together with a summary of the electrode operation and interferences. Also, the use of standard-addition or -subtraction methods is discussed. Experimental Instrumentation An Orion NO, electrode (Model 95-46), a digital pH meter (Fisher Accumet, Model 520) and a strip-chart recorder (Perkin-Elmer, Model 56) were used to make all measurements. A magnetic stirrer (Cole-Parmer Spin-Master, Model 4803), adjustable micropipette (Pipet- man, 0-200 p1) and centrifuge (Daman/IEC, Model UV) were also used. The Merckoquant nitrite-test papers were obtained from BDH (M10057-01). Reagents A 0.1 M sodium nitrite stock solution (Orion, number 95-46-08) was refrigerated at all times.Daily, a M sodium nitrite working solution was prepared by accurately diluting 1 ml of the stock solution to 100 ml with distilled water. The acid - sulphate buffer was prepared as directed by the Orion manual.26 All chemicals and solvents were of reagent grade except where indicated otherwise. Methods Calibration graphs were prepared by adding appropriate micro-aliquots of the working solution to 50 ml of buffered distilled water. Analysis of water-diluted cutting fluids (diluted to about 300 V/m) was performed by adding about 0.2 g of potassium dichromate and then applying the modified Orion electrode method described below. The cutting oil, weighed accurately (0.2-0.4 g), was dissolved in 100 ml of practical-grade 2,ZJ4-trimethylpentane (Fisher P-2396).Saturated calcium sulphate solution (100 ml) was added and the mixture was agitated for 30 s; the phases were then allowed to separate for a minimum of 1 h. The Analysis of cutting oils was performed as follows.360 CLARK AND MACPHERSON: DETERMINATION OF NITRITE ION IN UNUSED Artalyst, ‘vol. 104 aqueous phase was isolated and residual organic matter was separated from it by centri- fugation at 16 s-1 for 600 s. To 45 ml of aqueous phase, about 0.2 g of potassium dichromate was added and the Orion electrode method26 applied. After the electrode response had equilibrated, the potential was recorded and 100 p1 of 1% sulphamic acid solution (Fisher A-295) were added to remove the nitrous acid selectively and quantitatively. After the electrode response had re-equilibrated, the potential was recorded again.The difference in the concentration, determined by comparison with a calibration graph before and after addition of sulphamic acid, was attributed to nitrite ion. The calibration graph was deter- mined by analysing aqueous standards before and after analysing the set of samples. When cutting oils were examined, the membrane was replaced after immersion for 45 min. Much longer exposure to water-diluted cutting fluids was possible. Results and Discussion Operation of NOx Electrode Various aspects of gas-sensing electrodes, including the NO, type, can be found in the literature. These include the theory of operation and construction,2&-28 the possibility of increasing the sensitivity (with the ammonia electrode) by using a more dilute inner filling solution (IFS),29 the rates of loss of nitrous acid or nitrite ion from standard solutions30 and the extent of interferences from various diverse species affecting the Orion NO, ele~trode.~6J1932 In summary, the response of the NO, electrode obeys the Nernstian-like equation where Xi represents an interfering species and K , is its selectivity constant.During this work, the linear range of the Orion NO, electrode (Fig. 1) appeared to be between 0.23 and 230 pg ml-l (or 5 x M, respectively) of nitrite ion with curvature at the lower end owing to dissolved carbon dioxide, according to the Orion method manuaL26 Using the same bottle of IFS and batch of membranes, the average slope and intercept obtained (Table I) were significantly different from those reported by Orion.26 Each slope and intercept shown in Table I was determined by simple linear regression of the data into equation (1).and 5 x Fig. 1. Typical calibration graph using the Orion NO, gas-sensing electrode. The time required for electrode-potential equilibration was, on average, longer than that About 4-12 min were required, depending on the membrane used, the In general, longer reported by Orion. magnitude and direction of concentration change and the sample type. times were required with the metal-working lubricants analysed during this work.April, 1979 CUTTING FLUIDS AND CUTTING OILS WITH A GAS-SENSING ELECTRODE TABLE I TYPICAL SLOPE (S) AND INTERCEPT (KO) VALUES OBTAINED USING THE ORION NO, ELECTRODE 361 SlopelmV 45.8 43.1 41 .O 43.1 47.4 41.0 42.8 InterceptlmV 233 228 229 232 245 227 23 1 Mean .. .. .. . . 43.5 232 Standard deviation . . . . 2.4 6 With respect to this work, interfering species could be described under three general categories: those which reacted with nitrous acid or nitrite ion in the sample before immersion of the electrode, those which dissolved the membrane or removed its hydrophobic activity (commonly called wetting) and those which diffused across the membrane to alter the pH of the IFS. Oxidants (e.g., hydrogen peroxide) and reductants (e.g., sulphamic acid) are examples of the first type. This type of interference was observed during this work when several nitrite-free cutting oils were spiked with sodium nitrite solution during analysis and the electrode response was seen to increase for about 20s at first and then to decrease quickly as the nitrous acid was removed (it is possible that sulphonate salts caused this effect).The presence of this type of interference did not pose any problem for this laboratory but ensured the absence of nitrite ion in the lubricants. Micellar or water-soluble organic material, examples of the second type of interference, effected a serious restriction on the over-all sensitivity of the method. In short, early wetting of the electrode membrane in water-diluted cutting fluids and oils occurred, preventing analysis. With cutting fluids, large dilutions were made and, with cutting oils, solvent extraction was used partially to break the emulsion. The effect of the presence of this type of interference was to cause the electrode to respond as if nitrite ion were present.The selectivity constants of some weak acids have been determined26 and their effects during this work were avoided by a standard-subtraction method discussed in the next section. Because iodide salts are included in the general formulations of cutting fluid^,^ the effect of 0.1 M potassium iodide solution was investigated and found to be a weak but positive interference. The effect of an iodine solution was similar. For the same reason, mono-, di- and triethanolamine were investigated in the same manner and also by adding small amounts to a 0.46pgml-l nitrite ion standard solution during analysis. Several other salts were investigated to identify a suitable salting-out reagent to be used in a solvent-extraction emulsion-breaking step.Saturated calcium sulphate (about 0.02 M), 0.1 M potassium dichromate and 0.1 M sodium sulphate solutions were found not to interfere in the analysis of 0.046 and 0.46 pg ml-l nitrite ion solutions. However, 0.1 M sodium chloride solution did interfere positively. This agrees only partially with Tabatabai,32 who reported that 0.1 M sodium chloride solution did not interfere but a 1 M solution did. Methods were chosen in order to avoid early wetting. These methods will be discussed further in the following sections. The third type of interference (e.g., weak acids) was also encountered. The selectivity constant was determined to be about 0.012. No interferences were observed during 15 min.Calibration of the Electrode Because the method of standard additions has been recommended for use with electro- analytical techniques by several worker~,3~~3* it was considered for use during this work. While this method can, for any technique, compensate for enhancement or reduction in the response of the analyte brought about by any unmatched part of the matrix (direct compari- son with similar standards cannot do this), it cannot in any technique compensate for an362 CLARK AND MACPHERSON: DETERMINATION OF NITRITE ION IN UNUSED Analyst, Vol. 104 interfering species that causes a response like that of the analyte or one that removes the analyte before meas~rernent.~~ Hence, during this work standard additions could not compensate for the presence of any of the above interfering weak acids. An additional complication arising when the method of standard additions is applied to potentiometric measurements is that compensation for enhancement of the analyte response due to the unmatched part of the sample matrix can only be accomplished by using an iterative computer program.27 s36-39 In order to avoid interfering weak acids, present during this work, a standard-subtraction method was adopted.The agent used to remove nitrous acid was sulphamic acid, chosen because the reaction was reported to be selective, rapid and q ~ a n t i t a t i v e . ~ ~ Interfering species of the other type, Le., those which cause enhancement or suppression of the response of the analyte, were investigated by analysing several spiked nitrite-free cutting oils.This was a necessary criterion in order to ensure accuracy for this standard-subtraction method. The addition of sulphamic acid did not seem to have any serious effect on the electrode performance. However, the inner pH electrode was raised and lowered between each analysis and then equilibrated in a standard sodium nitrite solution to minimise the memory effects of species that could have moved across the membrane. Sulphamic acid additions had no effect on the electrode potential when only the following positive interfering species, listed in order of increasing K , values, were analysed : potassium iodide, sodium chloride, acetic acid, hydrofluoric acid, formic acid, lactic acid and pyruvic acid. No significant effects were observed.Analysis of Cutting Fluids and Cutting Oils To our knowledge only three methods of nitrite detection have been applied to cutting fluids and cutting oils, but none has been demonstrated to be useful at low microgram per millilitre levels, a requirement with which this laboratory was charged. Merckoquant nitrite test papers have been employed,40 which in our experience turned various shades of pink after being dipped briefly into standard sodium nitrite solutions of 1 pg ml-l of nitrite ion or higher. However, their limit of detection was raised considerably when applied directly to coloured cutting fluids or oils. In our opinion, they could safely be used for lubricants containing 3Opgg-l of nitrite ion or higher. The evolution of brown fumes (nitrogen dioxide) after addition of a strong acid has also been employed41 and this could have been the basis of a sensitive and selective method if a suitable technique for measuring nitrogen dioxide, such as headspace-gas chromatographic analysis, was used.Lastly, Williams et aL9 applied a standard spectrophotometric procedure for water and waste water samples20 to several high-nitrite cutting fluids collected in the Ottawa region. Cutting Juids We were fortunate to obtain the same samples that were analysed by Williams et al. for our use. Our analyses of them (Table 11) confirmed the presence of high levels of nitrite salts in these cutting fluids. Slightly lower concentrations were found in two out of six additional cutting fluids tested. It was not surprising to find four of our results compara- tively lower than those of Williams et al., as triethanolamine is also present in these cutting fluidsg and this compound can react slowly over a period of r n o n t h ~ ~ ~ ~ ~ ~ 0 with nitrite ion in the product itself.While direct analysis of the cutting fluids was made after a large dilution with distilled water (to about 300 V/m) because of the high nitrite ion concentrations, smaller dilutions and hence lower detection limits might be possible. However, an optimum dilution factor existed for these samples as membrane wetting took place quickly when immersion into a fluid diluted by the factor of 10 V/m was made. Assuming a dilution factor of 300 V/m and a lower limit of aqueous nitrite detection of 0.046 pg ml-l, the lower detection limit for nitrite ion in cutting fluids was 17 pg g-l.During the analysis of these cutting fluids it was demonstrated that the standard- subtraction step was necessary to avoid erroneously high results (Table 11) in six instances. The Merckoquant test papers gave a positive indication of nitrite ion presence in fluids 1-8 and negative indication in fluids 9-12. About 20 cutting fluids could be analysed in 8 h.A p d , 1979 CUTTING FLUIDS AND CUTTING OILS WITH A GAS-SENSING ELECTRODE 363 TABLE I1 CONCENTRATION OF NO,- ION FOUND IN DIFFERENT CUTTING FLUIDS NO,- concentration/pg g-l r A 7 This work A Cutting I 7 fluid SS* used SS* not used 1 2 3 4 5 6 7 8 9 10 11 12 6.0 x 104 - 2.9 x 104 - 9.0 x 104 - 2.0 x 103 - 1.3 x 105 - 3.1 x 104 - 1.8 x 102 1.9 x 103 2.3 x lo2 2.0 x 103 < 14 6.5 x lo2 < 14 85 < 16 1.5 x 103 <I6 1.4 x 103 Williams et aZ.@ 8.2 x 104 3.4 x 104 9.8 x 104 2.0 x 103 8.6 x lo4 3.8 x 104 - - - - - - * SS = Standard-subtraction method.Cutting oils Simple dilution and measurement of cutting oils proved unsuccessful during this work. Wetting of the membrane took place within 60 s of immersion in a cutting oil with a dilution factor of 1000 V/m. As a result, a solvent-extraction step was developed in order to separate the nitrite ion from the wetting agents. Although more efficient separations could have possibly been used, this separation fulfilled the sensitivity requirement placed on this laboratory. After determining that cleaner aqueous phases were obtained when the cutting oil was dissolved in the solvent phase first, rather than in the aqueous phase, the utility of several organic solvents was investigated by agitating 100 ml of 4.6 pg ml-l standard nitrite ion solution with 100ml of organic solvent and then analysing the aqueous phase. p-Xylene (Eastman 277), hexane mixture (Fisher H-292) and s-tetrachloroethane (Fisher A-31) gave, for undetermined reasons, recoveries of only 10% or less.Of the suitable solvents practical- grade 2,2,4-trimethylpentane (TMP) was used because it was inexpensive and available in large amounts. A restriction on the sample size of the cutting oil was observed to limit the sensitivity of the method just as with cutting fluids. Using 100-ml volumes of TMP and aqueous phases, aliquots of cutting oil of 1 g or greater caused the electrode potential to be erratic and the membrane life to be greatly reduced.The optimum aliquot seemed to be between 0.2 and O.4gJ but even then the membrane required changing after about 45 min of immersion. With some oils or with smaller aliquots longer immersion times were possible. The efficiency of the solvent-extraction step was investigated by analysing, in quadrupli- cate, a nitrite-free cutting oil, which was spiked so that after extraction the aqueous phase contained 2.3pgml-l of nitrite ion. As aqueous phases, 0.05 M sodium sulphate and 0.01 M potassium dichromate solutions were found to be unsuitable because they allowed recoveries of only 89% and 72% of nitrite ion, respectively, whereas distilled water and calcium sulphate solution allowed about 100 yo recoveries (Table 111).Saturated calcium sulphate solution is recommended for use as it produced, according to the electrode performance, slightly cleaner aqueous phases possibly owing to its emulsion-breaking properties. To investigate the recovery at a much lower nitrite ion concentration, i.e., 0.046 pg-ml-l in the aqueous phase, another spiked nitrite-free cutting oil was analysed in quadruplicate, using saturated calcium sulphate solution as the aqueous phase. The average recovery was lOOyo but a larger relative standard deviation was observed (Table IV). The solvent-extraction step was also shown to be suitable for cutting fluids if the need, however unlikely, ever arose (Table V). All three aforementioned aqueous phases were used to analyse, in quadruplicate, cutting fluid 8.The results were not statistically different from each other (using analysis of variance test with a 95% confidence limit).364 CLARK AND MACPHERSON : DETERMINATION OF NITRITE ION IN UNUSED Analyst, VoZ. 104 TABLE I11 CONCENTRATION OF NO2- FOUND IN AQUEOUS PHASES AFTER EXTRACTION CONTAINED 2.3 pg ml-l OF NO2- ION OF A NITRITE-FREE CUTTING OIL, SPIKED SO THAT THE AQUEOUS PHASES NO2- concentration in aqueous phase*/pg ml-1 DW A B C 2.14 2.04 1.96 1.72 2.28 2.25 2.14 1.79 2.28 2.37 2.04 1.61 2.41 2.56 2.08 1.46 I A \ Mean . . .. . . . . 2.28 2.31 2.05 1.65 Standard deviation . . * . 0.11 0.22 0.08 0.14 Recovery, yo . . . . 99 100 89 72 C = 0.01 M K,Cr,O,. * DW = distilled water; A = saturated CaSO,; B = 0.05 M Na2S0,; Of the 20 cutting oils analysed only one was found to contain a detectable nitrite ion concentration (16 pg g-l).The other oils, by definition, contained less than about 15 pg g-f of nitrite ion. Also, in only two instances (neither of the above) was the standard- subtraction step necessary to avoid erroneous results equivalent to 150 and 41 pgg-1 of nitrite ion. TABLE IV CONCENTRATION OF NO2- ION FOUND IN A SATURATED CaSO, AQUEOUS PHASE AQUEOUS PHASE CONTAINED 0.046 pg ml-l OF NO2- ION AFTER EXTRACTION OF A NITRITE-FREE CUTTING OIL SPIKED SO THAT THE Cutting NO,- concentration in Recovery, 32 0.041 89 24 0.044 96 23 0.052 113 25 0.046 100 oil aqueous phase/pg ml-l % Mean . . .. .. .. . . 0.046 Standard deviation . . .. .. 0.005 100 10 The Merckoquant test papers did not indicate the presence of nitrite ion in any of the About 10 cutting oils could be analysed during 8 h.cutting oils. TABLE V CONCENTRATION OF NO2- ION FOUND IN CUTTING FLUID 8 USING DIFFERENT AQUEOUS PHASES NO,- concentration in aqueous phase*/pg g-' DW A B C WD 215 203 23 1 220 238 235 216 244 234 246 238 211 238 230 247 227 213 245 236 245 r L > Meant . . .. .. 229 211 240 231 244 Standard deviation? . . 10 6 6 8 4 * As in Table 111; WD = water-diluted sample. Over-all mean = 231 pg g-1; over-all standard deviation = 13 pg g-l.April, 1979 CUTTING FLUIDS AND CUTTING OILS WITH A GAS-SENSING ELECTRODE 365 Used cutting JEuids and oils No attempt was made to apply these methods to used cutting fluids or used cutting oils. To do so would bring up the following two considerations, depending on the particular use of the lubricant.The most important consideration would be that the used lubricant would then be an oil in water emulsion that would have to be broken in such a way as to release the nitrite ion quantitatively to aqueous solution. As this could not be done efficiently enough for our needs (stated previously) by agitating the emulsion with TMP or other solvents, another emulsion-breaking step would have to be used. Examples of such techniques are the addition of ethanol, the addition of salts, boiling, freezing and vigorous agitation. We feel that the last two would be the most promising. The second consideration would be that additional contaminants would be added to the emulsion during use of the lubricant.Three main contaminants could be wear metals, surfactants (possibly used to clean the emulsion off the worked material) and tap water (used to make the emulsion). Only the presence of surfactants (wetting agents) would prevent the optimum use of this method but they should also be separated from nitrite ion with the proper choice of emulsion-breaking step. Conclusion We have found the methods of nitrite ion determination described herein to meet our needs for rapidity, economy and sensitivity. The detection limits for cutting fluids and oils were about 15 pg g-l. In addition, positive interference from weak acids could be avoided by a standard-subtraction method using sulphamic acid. The major drawbacks of the methods were the restrictions placed on the sensitivities due to the required sample dilutions and the extra time required for membrane replacement.We thank David T. Williams (Environmental Health Directorate, Health Protection Branch, Ottawa, Canada) for providing us with several previously analysed cutting fluids, the several lubricant manufacturing companies who provided us with numerous other lubricants and John A. Page (QueenJs University, Kingston, Canada) for helpful discussions. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. References Sutcliffe, T., “Machining Coolants,” E. F. Houghton and Co., Philadelphia, Pa., 1977, pp. 30-38. Forbes, W. G., “Lubrication of Industrial and Marine Machining,” Second Edition, John Wiley, G-osselin, R. E., Hodge, H. C., Smith, R. P., and Gleason, M.N., “Clinical Toxicology of Commercial March, J ., “Advanced Organic Chemistry : Reaction, Mechanism and Structure,” McGraw-Hill, Smith, P. A., and Leoppky, R. N., J . Am. Chem. Soc., 1967, 89, 1147. “Scientific and Technical Assessment Report on Nitrosamines,” EPA-600/6-77-001, US Environ- Alexander, M. , Chairman, “Assessment of Scientific Information on Nitrosamines,” Report to the Alexander, M., Chairman, “Assessment of Scientific Information on Nitrosamines,” Report to the Williams, D. T., Benoit, F., and Muzika, K., Bull. Toxic. Envir. Contam., 1977, 20, 206. Fan, T. Y., Morrison, J., Rounbehler, D. P., Ross, R., Fine, D. H., Miles, W., and Sen, N. P., Walters, C. L., Chemy Brit., 1977, 13, 140. Wishnok, J. S., and Tannenbaum, S. R., Analyt. Chem., 1977, 49, 715A.Boltz, D. F., Crit. Rev. Analyt. Chem., 1973, 3, 147. Fawcett, R., Tame, D. A., and Johnson, T. E., J . Ass. Publ. Analysts, 1976, 14, 23. Gabbay, J., Almog, Y., Davidson, M., and Donagi, A. E., Analyst, 1977, 102, 371. Toei, K., and Kiyose, T., Analytica Cham. Acta, 1977, 88, 125. Horwitz, W., Editor, “Official Methods of Analysis of the Association of Official Analytical Chemists, ” Fiddler, R. N., J . Ass, Off. Analyt. Chem., 1977, 60, 594. Hamilton, J. E., J . Ass. Off. Analyt. Chem., 1976, 59, 284. American Public Health Association, American Water Works Association and Water Pollution Control Federation, “Standard Methods for the Examination of Water and Wastewater,’ ’ Fourteenth Edition, American Public Health Association, Washington, D.C. , 1975.New York, 1954, Chapter 32. Products,” Fourth Edition, Williams and Wilkins Co., Baltimore, 1974, p. 158. New York, 1968, p. 486. mental Protection Agency, Washington, D.C. , 1976. US Environmental Protection Agency, Washington, D.C., 1976. US Environmental Protection Agency, Washington, D.C., 1976, p. 6. Science. N.Y., 1977, 196, 70. Twelfth Edition, Association of Official Analytical Chemists, Washington, D.C., 1975, p. 228.366 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 36. 36. 37. 38. 39. 40. 41. CLARK AND MACPHERSON Small, H., Analyt. Chem., 1975, 47, 1801. Anderson, C., Clin. Chem., 1976, 22, 1424. Schultz, F. A., and Mathis, D. E., Analyt. Chem., 1974, 46, 2253. Boese, S. W., Archer, V. S., and O’Laughlin, J. W., Analyt. Chem., 1977, 49, 479. Hansen, L. E., Richter, B. E., and Eatough, D. E., Analyt. Chem., 1977, 49, 1779. “Instruction Manual-The Nitrogen Oxide Electrode, Model 95-46,” Orion Research Inc., Cam- Bailey, P. L., “Analysis with Ion-Selective Electrodes,” Heyden, London, 1976. Ross, J. W., Riseman, J. H., and Krueger, J. A., Pure Appl. Chem., 1973, 36, 473. Bailey, P. L., and Riley, M., Analyst, 1977, 102, 213. Zafirion, 0. C., and True, H. B., Analytica Chim. Acta, 1977, 92, 223. Sherkan, S., J. Ass. Ofl. Analyt. Chem., 1976, 59, 971. Tabatabai, M. A., Commurt. Soil Sci. Plant Anal., 1974, 5 , 569. Page, J . A., Van Loon, G.. 2nd Joint CIC/ACS Conference, Montreal, 1977, Lecture Anal. 64. Toth, K., and Pungor, E., Am. Lab., 1976, June, 3. Klein, R., and Hatch, C., Am. Lab., 1977, July, 21. Orion Research, Orion Newsl., 1970, 2, 50. Orion Research, Orion Newsl., 1971, 3, 1. Branc, M. J. D., and Rechnitz, G. A., Analyt. Chem., 1970, 42, 1172. Zipper, J. J., Fleet, B., and Perrons, S. P., Analyt. Chem., 1974, 46, 2111. Yang, C. C., Cincinnati Milacron Co., Cincinnati, Ohio, personal communication. Abrams, E., E. F. Houghton and Co., Philadelphia, Pa., personal communication. bridge, Mass., 1974. Received August loth, 1978 Accepted October 17tk, 1978

ISSN:0003-2654    

DOI:10.1039/AN9790400358

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

Analyst, April, 1979 SHORT PAPERS 367 Dicarboxidine [y,y'-(4,4'-Diamino-3,3'- biphenyly1enedioxy)dibutyric Acid] as a Reagent for the Spectrophotometric Determination of Cyanide and Chlorine Kerstin Groningsson Research Department, Analytical Chemistry. A B KABI, S-112 87 Stockholm, Sweden Keywords : Dicarboxidine [ y , 7'-( 4,4'-diamino-3,3'-biphenylylenedioxy)dibutyric acid] chromogen ; cyanide determination ; chlorine determination ; spectro- photometry Benzidine, o-tolidine and o-dianisidine are reagents with wide applications, e.g., in the detection of blood,lS2 and the spectrophotometric determination of ~yanide,~ chlorine,* compounds containing the -CO-NH- group5p6 and manganese.' However, the carcinogenic properties of these chromogens necessitate the search for safer alternatives.Dicarboxidine [y,y'-(4,4'-diamino-3,3'-biphenylylenedioxy)dibutyric acid] is described by Jonsson et a1.8 as being significantly less carcinogenic than benzidine and its derivatives. Svahn and Gyllanderg have been able to replace o-tolidine and o-dianisidine with dicarbo- xidine in the detection of N-protected amino acids, peptides, barbiturates and meprobamate on thin-layer chromatograms. This paper describes the use of dicarboxidine as a reagent for the spectrophotometric determination of cyanide and chlorine. Optimum reaction conditions are evaluated and the sensitivity of the method is compared with those using benzidine and o-tolidine as the chrornogen.3~~ The reaction with cyanide permits the determination of cyanocobalamin in tablets after liberation of hydrogen cyanide.Experimental Apparatus The absorption spectra were recorded on a Shimadzu, Model MPS-5000, Multipurpose Recording Spectrophotometer. A Zeiss PMQ I1 spectrophotometer with l-cm cuvettes was used. Reagents tion. electrophoresis at pH 3.6. All chemicals were of analytical-reagent grade and were used without additional purifica- The purity of the dicarboxidine dihydrochloride was verified by high-voltage paper For the determination of cyanide Orthophosphoric acid, 0.04 M. Bromine solution, not necessarily saturated. Sodium arsenite solution, 2% m/V. Butan-2-01, Ethanol, 95%. Pyridine solution. Dicarboxidine solution. Pyridine - dicarboxidine reagent. Standard cyanide solzctions. Dissolve 125 ml of pyridine in 375 ml of water and add 10 ml of Dissolve 0.3 g of dicarboxidine dihydrochloride in 25 ml of water Immediately before use, dilute 5.00 ml of dicarboxidine Prepare a stock solution containing 150 mg 1-1 of potassium concentrated hydrochloric acid.and filter if necessary. solution to 50.0 ml with pyridine solution. cyanide and dilute aliquots as required.368 SHORT PAPERS Analyst, Vol. 104 For the determination of chlorine Dicarboxidine reagent. Dissolve 3.3 g of dicarboxidine dihydrochloride in 1000 ml of 3 x 10-3 M hydrochloric acid; this solution is stable for several days. Standard chlorine solutions. Prepare a stock solution of chlorine gas in water, standardise this solution by iodimetric titration with thiosulphate immediately before use and dilute aliquots with water as required.Procedures Determination of cyanide Mix 2.00 ml of 0.04 M orthophosphoric acid with 2.00 ml of sample (containing 0.4-2 mg 1-1 of cyanide in water) in a centrifuge tube. Add 1 drop of bromine water and 2 drops of sodium arsenite solution followed by 6.00 ml of butan-2-01. Shake the mixture, add 2.00 ml of pyridine - dicarboxidine reagent, then shake the mixture vigorously for 2 min and allow it to stand at room temperature for 15 min. Centrifuge, then mix 4.00 ml of the organic layer with 1.00 ml of ethanol and measure the absorbance of this solution at 490 nm against a reagent blank prepared from distilled water. Prepare a standard graph from the standard cyanide solutions. Determination of chlorine To 50.00 ml of sample (containing 0.5-3 mg 1-1 of chlorine) add 2.50 ml of 3 M nitric acid and 2.50 ml of dicarboxidine reagent.Allow the solution to stand at room temperature for 30min and measure the absorbance at 460nm against a reagent blank prepared from chlorine-free water. Prepare a standard graph from the standard chlorine solutions. Results and Discussion Determination of Cyanide The yellowish reaction product between cyanide and pyridine - benzidine has an absor- bance maximum at 480 nm3 whereas pyridine - dicarboxidine gives a red product with an absorbance maximum at 490 nm. The stability and development time of the two colours are similar. The colour with pyridine - dicarboxidine reaches maximum intensity almost immediately after addition of the pyridine - dicarboxidine reagent and fades slowly with time.Thirty minutes after mixing the sample and colour reagent, the absorbance decreases to 96.5% of ____ -~ 1 Wavelength/nm Fig. 1. Stability of the colour developed from cyanide and pyridine - dicarboxidine. Sample solution : 0.84 mg 1-1 of cyanide treated according to the described method. Spectra : 1, immediately after mixing; 2, 1.5 h after mixing; 3, 3 h after mixing; 4, 4.5 h after mixing; 5 , 6 h after mixing; 6, 22 h after mixing.April, 1979 SHORT PAPERS 369 the value read after the 15-min period proposed in the method. Fig. 1 illustrates the stability of the colour, and as can be seen the absorbance maximum is shifted with time to longer wavelengths. The calibration graph is rectilinear in the concentration range 0.2-2.7 mg 1-1 of cyanide (in the working solution).The sensitivity is about 70% of that given by the pyridine - benzidine reagent. The concentration of dicarboxidine dihydrochloride in the reagent appears not to be critical, as concentrations between 1.3 x and 2.6 x low3 mol 1-1 of dicarboxidine in the pyridine - dicarboxidine reagent gave identical absorbance values. The method is used in our laboratory for the routine determination of cyanocobalamin in tablets. Hydrogen cyanide is first quantitatively liberated by reduction with tin(I1) chloride according to the procedure of Dessouky and Pungor.lo Amounts of 50 pg of cyanocobalamin are used in the determination and the method has a relative standard deviation of about 2%. Determination of chlorine In the determination of free chlorine using the o-tolidine method, sodium bis(2-ethyl- hexy1)sulphosuccinate is added in order to stabilise the blue quinhydroneimino product formed.When this stabilised neutral o-tolidine procedure* was applied using dicarboxidine dihydrochloride (7 x 10-3 moll-1) instead of o-tolidine, a yellow colour with an absorbance maximum at 438 nm was obtained, which, however, was too faint to be used in practical measurements. Some modifications of the method, omitting the sodium bis(2-ethylhexyl)sulphosuccinate stabili~er,~ were therefore tried. By simply mixing the sample solution, reagent and phosphate buffer (pH 7.4), a greenish colour was obtained with an absorbance maximum at about 460nm. The absorbance, however, was low and the standard graph was not recti- linear below 0.1 mg 1-1 of chlorine.Changing the pH from 7.4 to 2 resulted in an unstable, intensely red colour with an absorbance maximum at 460 nm. By further lowering the pH to about 1 with nitric acid, as indicated under Procedure, a more stable red colour with an absorbance maximum at 460 nm was obtained. A spectrum recorded immediately after mixing a sample containing 3 mg 1-1 of chlorine with nitric acid and dicarboxidine reagent is presented in Fig. 2. The absorbance decreases with time, but for more dilute chlorine solutions (0.5-1 mg 1-l) the absorbance increases with time. This is illustrated in Fig. 3, which shows standard graphs obtained 2, 35 and 45 min after addition of acid and dicarboxidine reagent. After 35 min, the graph is rectilinear and passes through the origin, while deviations from the origin and over-all lower absorbance values are obtained after 45 min.The absorbance maximum does not change during the time studied. At a 01 I I I I 360 400 440 480 520 560 600 Wavelengthhm Fig. 2. Spectrum of the dicarboxidine product obtained from the reaction with chlorine. Sample solution: 2.95 mg 1-1 of chlorine treated according to the proposed method a t pH 1. The spectrum was recorded within a few minutes after the addition of dicarboxidine reagent.370 SHORT PAPERS Analyst, Vol. 104 Concentration of chlorine/mg I-’ Fig. 3. Standard graphs for chlorine deter- mined with dicarboxidine. Time after addition of dicarboxidine reagent: A, 2 ; B, 35; and C, 45 min. concentration level of 3 mg 1-1 of chlorine the absorbance after 35 min had decreased to 82% and after 45 min to 69% of the value obtained after 2 min.For practical measurements it would be preferable to wait 30 min before measuring the absorbance. The dicarboxidine method described above appears to give about the same sensitivity as o-tolidine used as described in the literat~re.~ The reason for the different behaviour a t lower and higher chlorine concentrations has not yet been investigated and needs additional studies. Although the method is useful in its present form, it is obvious that a suitable stabiliser for the colour would be advantageous. No attempts have hitherto been made to effect this improvement. The author thanks Mr. Lars Jansson for skilful technical assistance and Dr. N. A. Jonsson for valuable discussions of the manuscript. 1 . 2. 3. 4. 6. 6. 7 . 8 . 9. 10. References Adler, O., and Adlet, R., Hoppe-Seyler’s Z . Physiol. Chem., 1904, 41, 59. Culliford, B. J., and Nickolls, L. C., J. Forens. Sci., 1964, 9, 175. Nusbaum, J., and Skupeko, P., Sewage Ind. Wastes, 1951, 2, 875. Johnson, J. D., and Overby, R., Analyt. Chem., 1969, 41, 1744. Reindel, F., and Hoppe, W., Chem. Ber., 1954, 87, 1103. Krebs, K. G., Heusser, D., and Wimmer, H., i n Stahl, E., Editor, “Thin-Layer Chromatography,” Maljr, J., and Fadrus, H., Analyst, 1974, 99, 128. Jonsson, N. A., Groningsson, K., Pavlu, B., Vessman, J.. and Westlund, L. E., Arzneimittel-Forsck., Svahn, C. M., and Gyllander, J., J . Chromat., 1979, 170, 292. Dessouky, Y . M., and Pungor, E., Analyst, 1971, 96, 442. Springer-Verlag, Berlin, 1969, p. 854. 1979, 29, in the press. Received October 19th, 1978 Accepted November 7th, 1978

ISSN:0003-2654    

DOI:10.1039/AN9790400367

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

April, 1979 SHORT PAPERS Semi-automatic Determination of Manganese in 371 Natural Waters and Plant Digests by Flow Injection Analysis M. F. Gine, E. A. G. Zagatto and H. Bergamin Filho Centro de Energia Nuclear na Agricultura, CEP 13400 Piracicaba, Sdo Paulo, Brazil Keywords : Manganese determination ; water analysis ; plant material analysis ; flow injection analysis ; spectropJzotometry The formaldoxime methodl has been applied extensively to the spectrophotometric deter- mination of manganese, because of its high sensitivity and the rapid formation of a stable coloured complex. However, iron also forms a strongly coloured complex with formaldoxime and this causes serious interference. Formaldoxime (H,C=NOH) can be obtained in a solution acidified with hydrochloric acid by reaction of hydroxylammonium chloride with formaldehyde.Formaldoxime is dissociated in an alkaline medium, producing anions (H,C=NO-) that form coloured com- plexes with multivalent metals (manganese, iron, copper, cobalt, nickel, cerium and vanadium). The physical and chemical conditions of this reaction permit the use of the continuous- flow injection system developed by RfiiiEka and co-~orkers.~-5 This technique utilises the introduction of the sample into a continuously flowing carrier stream of water or reagent. When injected, the sample is pushed by this stream and dispersed into the reagent stream, whereupon the required reaction takes place. The coloured complex is then carried into a spectrophotometric flow cell, where the absorbance is measured after an exactly defined time interval.Experiment a1 Reagents 3 ml of formaldehyde solution (37% m/V) and dilute to 100 ml with water. must be prepared weekly and stored in a glass bottle. Formaldoxime solution, 1 M. Dissolve 7 g of hydroxylammonium chloride in water, add This solution Ascorbic acid solution, 5% m/V. Potassium cyanide solution, 10% m/V. Methyl orange solution, 0.00270 m/V in phosphate bufer, pH 7.0. Phosphate bufer. Dissolve 0.512 g of potassium dihydrogen orthophosphate and 1.034 g This Sodium hydroxide solution, 6 M. Hydrochloric acid, 0.4 M. of dipotassium hydrogen orthophosphate in water and dilute to 1 1 with water. solution is 0.01 M in phosphate, pH 7.0. Standards Dissolve 0.308 g of manganese(I1) sulphate (MnSO,.H,O) in about 800 ml of water, add 13 ml of concentrated nitric acid and dilute to 1 1 with water.Working solutions containing manganese in the range 0.05- 5 p.p.m. are prepared weekly by appropriate dilution of the stock solution. Standard manganese solution, 100 p.p.m. Working standard solutions. Apparatus and Procedure The system employed for this work is shown in Fig. 1. Polyethylene tubing of 0.86 mm i.d. was used and all connectors were made from Perspex. The pump was a Technicon AAII peristaltic pump, fitted with Tygon pumping tubes. The samples were injected by means of a proportional injector.6 The sample was aspirated to fill a loop, which exactly defined the injected volume; this loop was then placed as part372 SHORT PAPERS Analyst, Vol. 104 S Fig. 1. Flow diagram of the system.P, peristaltic pump; S, injection port; R, reaction coil (length 30 cm) ; M, mixing coil (length 60cm); and W, waste. The numbers in the pump are the flow-rates in millilitres per minute of the carrier, reduction, reagent, neutralisation and masking streams, which correspond to (a), (b), (c), (d) and (e), respectively. The distances S-C and C-B are approximately 2 cm. For details, see text. of the carrier stream. Water samples were injected without any pre-treatment, other than preservation (5 ml of 2 M sulphuric acid were added per 1000 ml 'of sample, immediately after collection) and plant samples were digested with nitric - perchloric acid using a Technicon BD-40 block digestor.' Formaldoxime, masking and neutralisation streams are added to each other at point A (Fig.1) and mixed by passage through coil M. At point B, the reagents meet the sample zone, which had previously received a reduction stream of ascorbic acid at point C. The formation of the coloured complex takes place in the reaction coil (R) and the absorbance is measured at 455 nm in a Beckman, Model 25, spectrophotometer, connected to a Beckman, Model 24-25 ACC, recorder, and equipped with a Hellma Type 178 flow cell, light path 10 mm, volume 0.08 ml. For the analysis of waters, the carrier stream was water, while for the analysis of plant digests, the acidic conditions of the samples were maintained in the carrier stream. Conse- quently, the sodium hydroxide concentration in the neutralisation stream was changed, in order to achieve the required alkaline conditions in the final stream.In order to prevent the reaction between iron(I1) and formaldoxime, another iron complex- ing agent (potassium cyanide) was used, in the presence of a reductant (ascorbic acid) that reduces iron(II1) to iron(I1). Aluminium, titanium, uranium, molybdenum and chromium also form light-coloured complexes that normally do not interfere in the determination of manganese in water or plant material by this method. If the aluminium or titanium concentrations are higher than 40 p.p.m. an additional masking flow of tartrate is recom- mendeda2 The effect of the presence of suspended and coloured materials in the sample was evaluated by replacing the formaldoxime with water and running the samples again, so as to obtain blank values.Results and Discussion The proposed system (Fig. 1) was designed in order to optimise the sampling rate, reagent composition and consumption of sample and reagent. The carrier stream was chosen to have a flow-rate of 5 ml min-l, in order to obtain a desirable speed of analysis with a satis- factory degree of sample dispersion, as will be discussed later. The other flow-rates were fixed as specified in Fig. 1. Under these conditions, the reagent streams have little influence on the sampling rate and dilute the carrier stream (or sample zone) by a theoretical factor of only 0.86, which was confirmed by tests using a methyl orange solution. The composition of the reagents was established as follows. Carrier stream. In order to avoid loss of sensitivity caused by differences between the refractive index of sample and carrier stream,s the latter should be water for the analysis of water samples and 0.4 M hydrochloric acid for plant samples.In order Neutralisation stream. Incomplete neutralisation results in a distorted peak.April, 1979 SHORT PAPERS 373 TABLE I INFLUENCE OF INJECTED VOLUME ON SENSITIVITY AND SAMPLING RATE Standard solution contained 2 p.p.m. of manganese. Volume injected/ ml Absorbance 0.058 0 0.115 0.161 1 0.189 0.232 3 0.309 0.348 5 0.399 0.464 7 0.468 a: 0.637 Relative Sampling rate/ sensitivity, yo * samples per hour? 16.3 230 25.8 167 42.9 143 54.1 136 64.7 120 88.0 - * Data obtained with methyl orange solution.6 t Data obtained after examination of the shape of the peak in Fig. 2.9 : “Infinite volume” represents the conditions when the carrier stream was replaced by sample.to maintain an adequate alkalinity (pH greater than 9.0), 6.0 or 1 . 0 ~ sodium hydroxide solution was used as the neutralisation stream for the analysis of plant extracts or water samples, respectively. Masking stream. The interference of iron was minimised by using ascorbic acid and potassium cyanide in the amounts specified above. In this situation, the iron( 111) interference was suppressed when its concentration was below 20 p.p.m. In practice, the concentration of this interferent in nitric - perchloric acid digests of plants is less than 15 p.p.m. and seldom reaches 1 p.p.m. in natural waters. In special instances of samples with higher contents of iron( 111) the concentration of masking reagent must be increased.0.4 0) C -2 2 I] 0.: Time I I 0.25 0.50 I nj ected volurne/ml Fig. 2. (a), Peaks corresponding to different injected S is the injection (b), Shows the influence of the injected volume on the volumes of a 2 p.p.m. manganese standard. point. peak height. For details, see Table I. Formaldoxime solution. The effect of formaldoxime concentration on the slope and linearity of calibration graphs was studied by varying the reagent concentration (0.5, 1.0, 1.5 and 2.0 M). Linear calibration graphs from 0.1 to 2.0 p.p.m. of manganese were obtained374 SHORT PAPERS Analyst, Vol. 104 by using formaldoxime solutions in the concentration range between 1.0 and 2.0 M. Higher blank values (probably due to refractive index differences)8 were detected when the formald- oxime concentration was increased and loss of sensitivity resulted when more diluted formaldoxime solutions were used. As a compromise, a 1.0 M formaldoxime solution was chosen as standard.8 0.2 e 0, a m 13 0 0.6 b - I A - L . ~ 1 .o Amount of manganese, p.p.rn. Fig. 3. (a), Calibration peaks. The values on the peaks represent ( b ) , Corresponding p.p.m. of manganese. Chart speed: 5 min in-'. calibration graph. The dispersion of the sample, which is related to the volume injected, is a critical para- meter that influences the sensitivity and sampling rate.5 The results presented in Table I and Fig. 2 were obtained with the system shown in Fig. 1 and indicate the dependence of the sensitivity and sampling rate on the injected volume, which was changed by variation of the sample loop in the proportional injector. The relative sensitivity is the ratio between the dispersion of the actual injected volume and that of an infinite volume and was deter- mined by using a methyl orange solution according to the procedure described by RfiiiEka and H a n ~ e n .~ The sampling rate was evaluated after examination of the recorded peak TABLE I1 COMPARISON OF RESULTS OBTAINED BY CONTINUOUS FLOW INJECTION AND ATOMIC-ABSORPTION SPECTROPHOTOMETRY Concentration of manganese, p.p.m. Water samples Plant digests Flow injection AAS* Flow injection AAS* 0.08 0.07 1.63 1.58 0.20 0.22 1.58 1.53 0.28 0.30 0.50 0.55 0.08 0.09 0.46 0.53 0.07 0.06 0.70 0.85 0.12 0.12 0.50 0.53 0.07 0.07 1.61 1.63 A I \ ----z---+-----~ 1 ID Atomic-absorption spectrophotometry.April, 1979 SHORT PAPERS 375 shape (Fig.2) as recommended by RuWka and S t e ~ a r t . ~ The good correlation between the results in columns 2 and 3 (Table I) indicates that completion of the chemical reactions is achieved. A series of determinations were performed using the system shown in Fig. 1, with an injected volume of 0.35 ml. Under these conditions, about 135 samples per hour could be analysed, with a standard deviation of better than 1% over the range 0.2-2 p.p.m. of manganese. Fig. 3 shows a typical calibration run and the corresponding calibration graph. In order to check the accuracy of the proposed method, a comparison of this method with atomic-absorption spectrophotometrylO was made. Some individual differences were found and no attempt was made to study the problem.The results are presented in Table I1 and indicate no statistical difference at the 1% level, for both plant material and water analysis. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. References Marczenko, Z., “Spectrophotometric Determination of Elements,” translated by C. G. Ransay, Marczenko, Z., Analytica Chim. Acta, 1964, 31, 224. RbiiCka, J., and Hansen, E. H., AnaJytica Chim. Acta, 1975, 78, 145. Krug, F. J., Bergamin Filho, H., Zagatto, E. A. G., and Jcrrgensen, S. S., Analyst, 1977, 102, 503. RbiiCka, J., and Hansen, E. H., Analytica Chim. Acta, 1978, 99, 37. Bergamin Filho, H., Medeiros, J. X., Reis, B. F., and Zagatto, E. A. G., Analytica Chim. Acta, 1978, 101, 9. “Digestion and Sample Preparation for the Analysis of Total Kjeldahl Nitrogen and/or Total Phosphorus in Food and Agricultural Products using the Technicon BD-20/40 Block Digestor,” Technicon Industrial Method No. 369-75A, Technicon Instruments Corp., New York, 1977. Bergamin Filho, H., Reis, B. F., and Zagatto, E. A. G., Analytica Chim. A d a , 1978, 97, 427. RhiiCka, J., and Stewart, J . W. B., Avzalytica Chim. Acta, 1975, 79, 79. “Analytical Methods for Atomic Absorption Spectrophotometry,” Perkin-Elmer Corp., Norwalk, Received August Ist, 1978 Accepted September 27th, 1978 John Wiley, New York, 1975, p. 342. Conn., 1973, pp. 1-15.

ISSN:0003-2654    

DOI:10.1039/AN9790400371

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

April, 1979 SHORT PAPERS 375 Gravimetric Determination of Copper( I I) and Cobalt( II) by Selective Precipitation with Benzi midazole K. N. Upadhyaya Chemistry Department, University of Day es Salaam, P.O. Box 35061, Day es Salaam, Tanzania Keywords : Copper determination ; cobalt determination ; benzimidazole ; gravirnetry Fiegl and Gleichl investigated the formation of salts with benzimidazole by some metal ions, such as mercury(II), copper(II), cadmium, zinc and cobalt (11). According to these workers the copper salt is precipitated from an ammoniacal solution of a copper salt and the cobalt salt is precipitated from an ammoniacal solution in the presence of ammonium chloride. The violet - blue crystalline precipitate obtained in the presence of cobalt has been used for its detection; nickel does not interfere as the nickel salt is not precipitated immediately under similar conditions, but on standing for over 12 h a grey - violet precipitate Yogada3 has used the reaction between cobalt ions and benzimidazole as the basis of a semi- quantitative method for the determination of cobalt by a spot test.Ghosh and Ghosh4 reported the formation of a stable cobalt(I1) complex with benzimidazole and on that basis determined cobalt gravimetrically. According to these workers the precipitation is carried out at pH 10 and therefore all elements forming insoluble hydroxides must be absent.376 SHORT PAPERS Analyst, Vol. 104 From a review of the literature it appears that although benzimidazole forms insoluble precipitates with several metal ions, its role as a quantitative precipitant has not been studied systematically. The present work was therefore undertaken with a view to determining copper(I1) and cobalt(I1) separately and in a single solution by the use of benzimidazole as a quantitative precipitant under controlled pH conditions.Determinations of copper and cobalt were carried out separately and then in solutions containing both ions. Interferences by Ni2+, Cd2+, Zn2+, Bi3+, Sb3+ and Be2+ were examined, Nickel did not interfere but serious interferences were observed with the other ions. Experiment a1 Reagents Benzimidazole was prepared and purified by the method described in the 1iteratu1-e.~ A stock solution (0.04 M) of the reagent was prepared in doubly distilled water and was made slightly acidic with hydrochloric acid.The solutions of copper(I1) and cobalt(I1) were prepared from their sulphate salts and were standardised before use. All chemicals used were of analytical-reagent grade. Apparatus electrodes. pH measurements were made using a Metrohm Herison pH meter, Model E 520, and Methods Determination of copper in a solution containing only coP$er(II) ions Place 5-20ml of the copper solution (5-20mg of Cu) in a 400-ml beaker, add 1040ml of the reagent solution, dilute to about 150 ml and heat the solution almost to boiling. Raise the pH of the solution slowly by adding dilute ammonia solution, drop by drop, with constant stirring until an additional drop does not form a precipitate. It was found that precipitation started a t pH 4.8 and was quantitative at pH 5.5.Heat the mixture on a hot water-bath for 15 min, then filter through a weighed sintered-glass crucible, porosity G4, wash the precipitate with 0.1% reagent solution, then three times with distilled water and finally with 50% ethanol. The conversion factor for Cu(C,H,N,), to copper is 0.21339. Dry at 110 "C for 2 h and weigh (Table I). Determination of cobalt in a solution containing only cobalt(II) ions Place 5-20 ml of the cobalt solution (5-20 mg of Co) in a 400-ml beaker, add 1040 ml of the reagent solution, dilute to about 150 ml and heat the solution almost to boiling. Raise the pH of the solution by adding ammonia solution, drop by drop, with constant stirring until an additional drop of ammonia solution does not form a precipitate.The precipitation in this instance starts at pH 7 and is quantitative at pH 8.2. After the precipitation is complete follow the same procedure as outlined for copper (Table I). The conversion factor for Co(C,H,N,), to cobalt is 0.2010. Determination of copper and cobalt when present in the same solution To a solution containing a mixture of copper(I1) and cobalt(I1) ions, add an amount of the reagent solution nearly double that necessary for complete precipitation of the metal ions, dilute the solution and heat it almost to boiling. Add ammonia solution, drop by drop, with constant stirring until the pH of the solution reaches 5.5. Heat the solution on a water-bath, filter, wash, dry and weigh the precipitate as described above for copper. Collect the filtrate and washings, boil the solution and add a further amount of ammonia solution dropwise with constant stirring until the solution attains a pH of 8.2.Heat, filter, wash, dry and weigh the precipitate as described for cobalt (Table 11). NOTE- The pH values are critical for the precipitation of both copper and cobalt. The freshly precipitated copper compound tends to dissolve in ammonia solution above pH 7 and the cobalt compound above pH 10.April, 1979 SHORT PAPERS TABLE I GRAVIMETRIC DETERMINATION OF COPPER( 11) AND COBALT(II) IN SEPARATE SOLUTIONS USING BENZIMIDAZOLE AS PRECIPITANT All results reported are the means of three determinations. 377 Calculated Mass of Mass of mass of Cu taken/ precipitate/ Cu(C,H,N,),/ mg mg mg Error, % 10 47.0 46.830 t 0 .3 6 15 70.5 70.245 +0.36 20 93.9 93.660 t 0 . 2 5 5 23.3 23.415 -0.45 Calculated Mass of mass of Mass of Co precipitate/ Co(C,H,N,),/ takenlmg mg mg Error, yo 5 24.8 24.850 -0.2 10 49.9 49.700 + 0.40 15 74.3 74.550 -0.34 20 99.8 99.400 +0.40 Results and Discussion As is evident from the results in Tables I and 11, benzimidazole can be used as an effective quantitative precipitant for copper(I1) and cobalt(I1) ions. By careful control of the pH the determination of these two ions can be carried out in a single solution. The fact that nickel does not interfere is an added advantage of the method because nickel is found together with these metals in many materials. The pH ranges are selective for both copper and cobalt. Whereas the freshly precipitated copper compound tends to dissolve in ammonia solution above pH 7, that of cobalt is not affected up to pH 10.The results of the analysis are broadly in agreement with the compositions of the complexes reported for these metals ear1ier.l The colour of the copper complex was chocolate brown and that of the cobalt blue - violet. The compounds were found to be stable and no loss of mass was noticed when either compound was heated to 300 “C. TABLE I1 GRAVIMETRIC DETERMINATION OF COPPER(II) AND COBALT( 11) IN A SINGLE SOLUTION USING BENZIMIDAZOLE AS PRECIPITANT All results reported are the means of three determinations. Mass of Mass Mass Approxi- copper Calculated of Cu of Co mate molar precipitate mass of taken/ taken/ ratio of (pH 4.8-5.5)/ Cu(C,H,N,)/ 10 10 1 : l 46.6 46.830 20 10 2: 1 93.3 93.660 10 20 1:2 47.0 46.830 mg mg CutoCo mg mg Mass of cobalt Calculated precipitate mass of Error, % (PH 7-8.211 ~(GH,N,),/ - mg mg c u co 49.6 49.700 -0.50 -0.20 49.9 49.700 -0.38 + 0.40 99.7 1 99.400 +0.36 + 0.30 References 1. 2. 3. 4. 5. Feigl, F., and Gleich, H., Mh. Chem., 1928, 49, 385. Feigl, F., “Quantitative Analysis mit Hilfe von Tupfelreaktionen,” Second Edition, Akademische Yagoda, H., Mikrochsmie, 1938, 24, 117. Ghosh, S. P., and Ghosh, H. M., J . Indian Chem. SOC., 1956, 33, 899. Org. Synth., 1941, Collect. Vol. 11, 65. Verlagsgesellschaft, Leipzig, 1936, p. 68. Received June 14th, 1978 Accepted August 15th, 1978

ISSN:0003-2654    

DOI:10.1039/AN9790400375

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

378 SHORT PAPERS Analyst, Vol. 104 Routine Determination of Nitrogen by Kjeldahl Digestion Without Use of Catalyst Eric Florence and Douglas Frank Milner National Institute for Research in Dairying, Shinfield, Reading, RG2 9A T Keywords : Nitrogen determination ; non-toxic Kjeldahl digestion ; hydrogen peroxide oxidation The use of a catalyst in Kjeldahl digestion has been widely examined; its presence substantially accelerates oxidation and completes the digestion to allow the subsequent determination of nitrogen. Results from many investigations have been reported by Bradstreet .l These results have established mercury as being the most effective catalyst and it is universally used in routine laboratory work. However, it is very toxic and expensive to use, and consequently much research has been undertaken to find an efficient substitute of low toxicity and cost.Stirrup and Hartley2 and Williams3 used copper(I1) sulphate in combination with non- toxic titanium dioxide and found this mixture to be a suitable replacement for mercury in the analysis of feedingstuff s and cereal grains. Glowa4 tested non-toxic zirconium dioxide alone and also combined with copper(I1) sulphate as a possible replacement for the mercury catalyst used in the AOAC5 standard method for the analysis of feedingstuffs and fertilisers. A longer clearing period in the initial stage of digestion was required for zirconium dioxide when used alone and consequently the use of a mixed catalyst was recommended. Tinpall6 used a copper catalyst with high-temperature digestion at 405 "C; good recoveries of nitrogen from amino acids were obtained. Mercury can be replaced by other catalysts provided that certain parameters of the digestion procedure are modified.These parameters may be the extension of the digestion period, the elevation of boiling-point by appropriate adjustment of salt to acid ratio and the use of temperature-controlled apparatus. However, the use of a pollution-free digest for the determination of nitrogen in the routine laboratory is very desirable. Trials have been undertaken to establish an oxidation scheme yielding such a digest with the use of hydrogen peroxide. Bradstreetl has reported on the use of hydrogen peroxide without a catalyst in the macro- digestion of samples taken from a variety of products.The digestion time is dependent on the type of product and large volumes of hydrogen peroxide solution and sulphuric acid are used. A recent paper by Tomonari et a1.' has described the use of hydrogen peroxide without a catalyst for the deter- mination of nitrogen in feedingstuffs. The results obtained were in good agreement with those from a standard method, except for fishpaste, and a bad recovery of nitrogen from tryptophan was reported. In this paper a procedure is described for the routine determination of nitrogen in feed- stuffs and biological products on a semi-micro scale by use of Kjeldahl digestion but without the use of a catalyst. Multiple additions of hydrogen peroxide solution are made in order to accelerate decomposition and clarification. Digestion for a further 30 min gives complete conversion.Nitrogen in the digest is determined by using an AutoAnalyzer and the results are processed by a computer. These requirements are not compatible with routine analysis. Experimental Apparatus Digestion tube. Digestor. kamp, London). AutoAnaZyzer. Datalogger. Com$mter. These were 150 x 16 mm in size, rimless, Pyrex and calibrated at 25 ml. This was an electrically heated micro-Kjeldahl range, holding 24 tubes (Gallen- A Technicon Mark I analyser (Technicon, Basingstoke). A Venture model (Digitronix Ltd., Milton Keynes). A PDP 11/10 (Digital Equipment, Galway, Ireland).April, 1979 SHORT PAPERS 379 Reagents Anti- bumping granules. Hydrogen peroxide solution, 30% m/ V , analytical-reagent grade. Sulphuric acid, d = 1.84, analytical-reagent grade.Potassium sulphate tablets, each weighing 1.2 g. These were obtained from Thompson and Mercury catalyst tablets, each containing 0.05 g of red mercury(II) oxide and 1.2 g of potassium Capper Ltd., Cheshire. sulphate. Kjeltabs FM (Thompson and Capper Ltd.) were used. Procedure Preparation of sample hammer mill (Christy and Norris Ltd., Chelmsford). and samples for analysis drawn out by use of a syringe fitted with a wide-bore needle. Digestion A 1-g sample was taken in the instance of acid faeces and 1 ml was taken by pipette for other liquid samples. A few anti-bumping granules were added. Feedingstuffs were ground to pass through a 1 mm screen aperture in an 8-in laboratory Acidified faeces were homogenised A 0.1-g amount of dried sample was weighed directly into a tube.Test method One tablet of potassium sulphate was added to the sample, followed by 3 ml of sulphuric acid. The tube was supported at an angle of approximately 45" on this range and was completely visible to the operator. Approximately 1 ml of hydrogen peroxide solution was added by using a Pasteur pipette and the reaction allowed to take place at ambient temperature. When this reaction was completed (2-3 min) the tube was heated and the rate of digestion controlled by selecting appropriate heat settings. Low and high settings are provided for the heating spiral at the bottom of the tube and a high setting only for the spiral a t the side-wall of the tube. Hydrogen peroxide solution was then added one drop at a time by means of the Pasteur pipette.These additions were made carefully, allowing the drop to trickle down the inside of the wall of the tube and to flow slowly into the acid mixture, thereby minimising the effervescence. It was important to allow this reaction to be completed for each drop before making a further addition as the reaction can be violent and dangerous when an excessive volume of peroxide is added. Additions were continued until the clearing stage was com- pleted. This object was achieved when the digest became permanently discoloured and white acid fumes appeared. Digestion was continued for 30 min in order to complete the process. The tube was then placed on the range without any application of heat. No further use of hydrogen peroxide was made.Standard method acid. method. One tablet of mercury catalyst was added to the sample followed by 3 ml of sulphuric The tube was heated on the range with the rate of digestion controlled as in Test Digestion was continued for 1 h after clearing. Autoanalysis The digest was diluted to 25 ml with de-ionised water and an aliquot analysed to deter- mine nitrogen by using the method described on Technicon Method Sheet N-3b. Ammonium sulphate solutions were used for calibration. The presence of mercury increases the fonna- tion of indophenol dye and this enhancement of colour intensity has to be compensated for as described by Davidson et a1.8 Consequently, standard solutions for calibration (see Standard method) contained 2 mg of mercury(I1) oxide per millilitre of solution.Data processing this tape fed to the programmed computer. The signal response for peak height was recorded on a cassette tape by datalogger and Calibration standards were run both a t the380 SHORT PAPERS Analyst, Vol. 104 beginning and at the end of the run in order to make a correction for base-line drift, assuming that this drift was constant with time. Results and Discussion Preliminary investigations were carried out to determine the digestion time required by Analysis It This period was adopted the test method using a sample of white fishmea1.l by use of the standard method gave a nitrogen content of fishmeall of 10.10 &- 0.10%. is evident that a 30-min period following clarification is adequate. for all further digestions by the test method.Results are given in Table I. TABLE I DIGESTION TIMES FOR NITROGEN DETERMINATION I N WHITE FISHMEAL’ BY USE OF THE TEST METHOD Time after clearing/ Nitrogen,* Standard min Y O deviation, yo 30 10.21 0 07 45 10.17 0 07 60 10.09 0.06 75 10.18 0.05 * Mean of four replicates. Samples taken from a variety of feedingstuffs and biological materials were analysed in There triplicate by both test and standard methods. is good agreement for all of the products. The results are given in Table 11. TABLE I1 NITROGEN DETERMINATION IN FEEDINGSTUFFS AND BIOLOGICAL PRODUCTS BY TEST AND STANDARD METHODS Test method I Sample I Barley . . . . .. Hay . . .. .. .. Grass meal . . . . .. Concentrates. . . . . . Weatings . . . . .. White fishmeal2 . . .. Skim-milk powder . . .. Whole milk .. .. .. Urine (cow) . . . . .. Acidified faeces (cow) . . Nitrogen, yo - 1.87, 1.86, 1.85 1.43, 1.46, 1.46 1.87, 1.82, 1.80 2.48, 2.47, 2.45 2.33, 2.33, 2.36 10.44, 10.64, 10.64 5.25, 5.48, 6.49 0.592, 0.572, 0.588 0.528, 0.528, 0.529 0.152, 0.151, 0.152 a Mean, % 1.86 1.45 1.83 2.47 2.34 10.57 4. 5.41 0.584 0.528 0.152 Standard method h I \ Nitrogen, % & Mean, % 1.87, 1.84, 1.85 1.85 1.41, 1.49, 1.39 1.43 1.82, 1.85, 1.84 1.84 2.47, 2.44, 2.48 2.46 2.37, 2.37, 2.37 2.37 10.54, 10.66, 10.62 10.61 5.30, 5.36, 5.42 5.36 0.587, 0.596, 0.606 0.596 0.521, 0.528, 0.526 0.525 0.154, 0.149, 0.150 0.151 Digestion by means of the test method was further examined by measuring the recovery of nitrogen from tryptophan, which is considered a good reference standard for assessing the efficiency of digestion.Results for both the test and standard methods are given in Table 111. These results compare well for both accuracy and precision. TABLE I11 RECOVERY OF NITROGEN FROM TRYPTOPHAN BY TEST AND STANDARD METHODS Nitrogen, yo Number of Standard Recovery, r analyses Mean deviation %* Test method . . . . 14 13.55 0.29 98.8 Standard method . . . . 15 13.59 0.35 99.1 * Theoretical value of 13.72%.April, 1979 SHORT PAPERS 381 The method described in this paper is most suitable for carrying out routine laboratory work, where one operator can work four digestion units (96 determinations) simultaneously. The stages of digestion are staggered, with each unit employed in a cycle of events as follows : initial cold reaction with hydrogen peroxide solution ; preliminary clearing with hydrogen peroxide solution ; final clearing with hydrogen peroxide solution ; completion of digestion without the use of hydrogen peroxide solution.This cycle permits the operator to handle a maximum of two units at one time for the additions of hydrogen peroxide solution to hot digest. When using the method frothing was minimised by the continuing reaction of hydrogen peroxide, and any residue on the side-wall of the tube was easily washed down by the small additions of hydrogen peroxide solution. The initial clearing stage was rapid and the total digestion time averaged 60min for the products tested. The cost of reagent chemicals required for a single digestion by test and standard methods was 2.1 and 0.9 p, respectively. The extra cost of digestion by use of the test method can be offset by the cost of labour and equipment required for the recovery of mercury from digest solutions obtained when using the standard method in order to prevent environmental pollution. Calculation of the results by computer was fast, reliable and accurate. 1. 2 . 3. 4 . 5 . 6 . 7 . 8 References Bradstreet, R. B., “The Kjeldahl Method for Organic Nitrogen,” Academic Press, London, 1965. Stirrup, J. E., and Hartley, A. W., J . Ass. Publ. Analysts, 1976, 13, 72. Williams, P. C., J . Sci. Fd Agric., 1973, 24, 343. Glowa, W., J . Ass. 08. Analyt. Chem., 1974, 57, 1228. “Official Methods of Analysis of the Association of Official Analytical Chemists,” Eleventh Edition, Tingvall, P., Analyst, 1978, 103, 406. Tomonari, M., Tsukagoshi, Y., Furuuchi, M., Bushimata, K., and Kani, T., Tokyotaritsu Eisei Davidson, J., Mathieson, J., and Boyne, A. W., Analyst, 1970, 95, 181. Association of Official Analytical Chemists, Washington, D.C., 1970. Kenkyusho Kenkyu Nempo, 1976, 27, 210. Received August 14th, 1978 Accepted October 26th, 1978

ISSN:0003-2654    

DOI:10.1039/AN9790400378

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

382 Book Reviews Analyst, April, 1979 HIGH PERFORMANCE MASS SPECTROMETRY : CHEMICAL APPLICATIONS. By MICHAEL L. GROSS. A symfiosium co-sponsored by the University of Nebraska, Lincoln, the National Science Foundation, A EI Scientific, and INCOS Corp., Lincoln, November 3-5, 1976. A CS Symfiosium Series No. 70. Washington, D.C. : American Chemical Society. 1978. Price $28. The objective of the book is to survey current research in high-performance mass spectrometry. It is divided into two sections, the first representing about a quarter of the book. This describes methods for acquiring metastable ion data which relate to stable and low energy decomposing ions and the application of such methods to determine ion structures, properties and potential energy surfaces. Methods for investigating very rapid ion reactions using field ionisation kinetics are also covered.The second and much longer part of the book considers the use of high-performance mass spectrometry for tackling analytical problems. I t covers both qualitative and quantitative mass spectrometry, using a variety of ionisation techniques, including simultaneous positive and negative ion chemical ionisation. For those who regard GC - MS as the only method for the analysis of organic mixtures, the chapter on trace analysis by evaporation from the direct inlet probe and detection by single ion monitoring is a useful reminder that this approach can be a valuable additional method. As a vehicle for demonstrating this, the author describes the analysis of a number of commodities for trace amounts of aflatoxins.GC - MS is, of course, vital for many applications and the subsequent chapter gives an intro- duction to the technique. The particular considerations for GC with high-resolution mass spectro- metry are noted, and the relevance of selecting appropriate GC conditions is considered. In one of the chapters on chemical ionisation mass spectrometry, a method for the generation of exact mass data using a quadrupole instrument is given. The use of field desorption with chemical ionisation, and in a later chapter with electron impact, is described. The use of various forms of mass spectrometry in biomedical applications is well presented, with some examples of selected ion monitoring for the detection of drug metabolites in urine extracts.A well balanced chapter on field desorption helps to distinguish myth from fact and is backed by examples of biomedical applications. There are occasions when deliberate pyrolysis coupled with electron-impact fragmentation can be put to advantage, and one paper describes such a procedure for identifying some nucleosides in DNA. The value of ultra-high-resolution measurements in the analysis of petroleum products is demon- strated, and emphasises the need for such a procedure combined with high-performance liquid chromatography. There is only one paper on inorganic mass spectrometry, and the use of isotope dilution for trace analysis. The last two chapters of the book emphasise the need for a computer system in order to make full use of the vast amount of data that is generated per unit time by a mass spectrometer.The book represents a useful collection of good quality papers and is certainly recommended reading to those concerned with the analytical aspects of mass spectrometry. Presumably in some attempt to imply rapid publication, all of the contributions are dated December 30th, 1977. In fact, the contributions were presented at a symposium in November, 1976, and the manuscripts have been up-dated. Would it not have been more useful to have published the original papers much sooner? T. A. GOUGH Pp. x + 358. Direct inlet and GC - MS are covered. Pyrolytic fragmentation is an effect which spectrocopists often try to avoid. TRACE ELEMENTS IN HUMAN HAIR. By VLADO VALKOVIC. Pp. x + 194. New York and Amongst its interesting properties is the ability along with other keratinous materials to concentrate trace elements in its structure.Hence the potential exists for studying the recent past of a person in terms of the distribution of these elements along the shaft of a growing hair. Any book giving information on this topic is therefore to be welcomed, but in fact reading “Trace Elements in Human Hair” provokes very mixed feelings. The positive contributions are offset by some serious deficiencies. London: Garland STPM Press. 1977. Price $19. Hair is a complex biological material.BOOK REVIEWS 383 There are chapters on hair growth; structure of hair; trace elements in hair; the role of trace elements ; applications of trace-element measurements ; and methods for measurement of trace- element levels in hair. There is also a comprehensive set of references, but unfortunately this only covers the period up to 1975, with few references from 1976 and later.Because of this, most of the work described in the book on the quantitative analysis of hair is concerned with “bulk” samples, which can give only an average result for a particular person. More recent studies, e.g., that by Maes and Pate, J . Forens. Sci., 1976, 21, 127, have shown that frequently there is as much variation in composition between hairs taken from different parts of the head of one person as there is between different people. Hence a more critical examination by the author is needed, particularly in assessing the significance of the results which he reports from the many research workers in this field.The chapter on methods of measurement deals exclusively with the techniques available a year or so ago, viz., atomic-absorption spectroscopy, neutron-activation analysis, X-ray fluorescence emission induced by proton beams or radioactive sources, and spark-source mass spectrometry. It would have been worthwhile enlarging this section to include firstly electron probe work, particularly with plasma-ashed samples and using wavelength-dispersive analysers to increase the signal to noise ratio, and secondly fine-beam X-ray fluorescence methods. For example, the statement “hair not dependent on steroid hormones . . . but sometimes inhibited by androgenic hormone” immediately suggests the riddle, when is a steroid not a steroid? A cursory check revealed over 40 misspellings; the absence of adequate proof reading, even though the book appears to have been produced from typescript, is quite inexcusable.R. L. WILLIAMS The book contains a number of factual errors. Typographical errors abound. IARC MONOGRAPHS ON THE EVALUATION OF CARCINOGENIC RISK OF CHEMICALS TO HUMANS. Volume 17. SOME N-NITROSO COMPOUNDS. Pp. 365. Lyon: International Agency for Research on Cancer. Available in the UK through HM Stationery Office. 1978. Price SwFrBO; $25. Distributed by the World Health Organization. The Monographs were begun in 1972 in order to collect all available relevant experimental and epidemiological data on chemicals to which man is exposed and to assess the data from the point of view of risk. The working group consists of experts in the various appropriate fields, notably pathology, microbiology, toxicology, physiology, chemistry and epidemiology.A new group, frequently with different contributors, is convened for each monograph. The present collection of monographs is devoted exclusively to N-nitroso compounds, some of which have been discussed in earlier publications in this series (Volumes 1 and 4). A vast amount of work on these compounds has been undertaken in the last few years and the appearance of this book, updating and expanding the earlier volumes, is most welcome. The first part of the volume describes the background to the series, the objectives and scope and some comments on the selection of chemicals. The general principles for evaluating the carcinogenic risk of the selected chemicals is described. This includes an assessment of the various aspects of the published data based on animal experiments, any evidence of carcinogenicity in humans and epidemiological data.The second part of the book consists of a monograph on each of 17 N-nitroso compounds. This includes the simple aliphatic nitrosamines, and some hetero- cyclic nitrosamines which have been the subject of many studies on their occurrence in foodstuffs. N-Nitrosodiethanolamine and N-nitrosomorpholine, which have been found in industrial chemicals and formulations, are included, as is N-nitrosonornicotine, which appears to be unique to tobacco products. There are also several nitrosoamino acids and nitrosofolic acid, the latter apparently prepared only as a laboratory experimental chemical.Two nitrosoalkylureas, which have been used for many years for the laboratory preparation of the corresponding diazo compounds, are listed. Their use for this purpose is no longer encouraged, and one has been studied for use as a cancer chemotherapeutic agent. One nitroso compound (steptozotocin) occurs naturally and is formed microbiologically, and has application in the study of diabetes. For each of the above compounds the following information is given, where available : 1. 2. synonyms, trade-names, molecular formula and basic physical properties ; production, both in the laboratory and commercially, and uses, which for this group of compounds are very restricted ;384 BOOK REVIEWS Analyst, Vol. 104 occurrence, which is a particularly interesting section with respect to the nitrosodialkyl- amines as they have been found in food, water, air, industrial chemical formulations, tobacco smoke and drugs; N-nitrosodimethylamine is by far the most common and a detailed list of levels of occurrence in foods is tabulated; this nitrosamine is also formed in vivo and there is some evidence for its occurrence in human biological fluids ; analysis, which is very brief as the IARC has recently published a separate monograph covering the volatile nitrosamines,l and there are few methods for the remaining compounds ; biological data, only included if they are relevant to the assessment of risk, and normally covering carcinogenicity and related studies in animals ; observations on humans are included in this section although in most instances no data are available; 6.the final section of each monograph summarises the data and attempts to make an evalua- tion, clearly a most difficult task. Each monograph has a set of references from which the source material was obtained. The book, although a paperback, is well presented and no errors were detected. It is commendably up to date and was published in May, 1978, only 7 months after the Working Group had met. 3. 4. 5. Reference 1. Egan, H., (Editor-in-Chief), “Environmental Carcinogens : Selected Methods of Analysis,” Volume I, “Analysis of Volatile Nitrosamines in Food,” Scientific Publication No. 18, IARC, Lyon, 1978. T. A. GOUGH ION-SELECTIVE ELECTRODES. CONFERENCE HELD AT BUDAPEST, HUNGARY, 5-9 SEPTEMBER 1977.Edited by E. PUNGOR and I. BuzAs. Pp. x + 613. Amsterdam, Oxford and New York: Elsevier. Budapest: Akademiai Kiad6. 1978. Price $75; k45.45; DF1180. This volume arrived for review hard on the heels of that from the conference held a year earlier in Budapest (see Analyst, 1978, 103, 1007). Again we have camera-ready copy of variable quality but a much fatter volume with a more than commensurate increase in price, which one suspects is not unrelated to the distribution in the West having been taken over by an international house. Like the so-called discussion lectures (there is no discussion included), these are arranged in the book in alphabetical order of the first named contributor. Some gathering together under topic headings would have been helpful to the reader.Some of the Plenary Lectures (Bates, Dolidze) present new work, others review their own or others’. Pungor’s lecture, placed sixth, is more in the nature of an introduction to current problems. The discussion lectures fall into the categories of mechanistic studies (Baucke, Koryta, Boksay, Belijustin, Nagy), analytical usage (the majority) and clinical (Durst, Fuchs, Havas, Mascini). It may be noted that six of the papers are on pH glass electrodes, of which the most notable is that of Baucke, who reviews his work on the ion-sputtering technique of studying concentration profiles in surface layers and presents a new mechanism based on silanol group dissociation. Other papers presenting significant advances are those of Nagy, Fjeldly and Johannessen, who show that there is a fluoride depleted layer on the surface of the lanthanum fluoride electrode from Auger spectroscopy, Tacussel and Fombon, who discuss new phosphate and chloride sensors, and Neshkova and Sheytanov, who describe plated chalcogenide electrodes. In short, a mixed bag of a more international flavour than at the previous meeting, which all interested in ISE’s will need to borrow from their institution’s library; it could only be recom- mended for private bookshelves at half the price. There are the texts of seven Plenary Lectures, occupying about one third of the book. A. K. COVINGTON

ISSN:0003-2654    

DOI:10.1039/AN9790400382

出版商:RSC    

年代:1979

数据来源: RSC