ISSN:0003-2654    

DOI:10.1039/AN98409FX033

出版商:RSC    

年代:1984

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98409BX035

出版商:RSC    

年代:1984

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98409BP073

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST SEPTEMBER 1984 VOL. 109 1117 Quantitative Deter m i nation of Crysta I li ne Si lica in Res pi ra ble-size Dust Samples by Infrared Spectrophotometry Robert D. Foster and Ronald F. Walker Occupational Medicine and Hygiene Laboratories Health and Safety Executive 403-405 Edgware Road, Cricklewood London N W2 6L N UK A direct differential scanning infrared spectrophotometric method is described for the determination of crystalline silica in respirable-size atmospheric dust samples from a wide range of industrial environments. The test atmosphere is drawn through a cyclone elutriator fitted with a membrane filter at 1.9 I min-1 for, typically 1-4 h. The filter without further preparation is scanned from 1500 to 400 cm-1 and the atmospheric concentration of crystalline silica determined from the intensity of the quartz (799 780 and 695 cm-I), together with when present cristobalite (796 and 621 cm-1) absorption bands.Generation of standard crystalline silica atmospheres using a portable dust chamber is described. The effect of particle size variation on the absorbance intensity together with sample re-radiation and mineral interference effects are examined. Keywords Crystalline silica; quartz; cristobalite; infrared spectrophotometry; atmospheric pollution Quartz the most abundant polymorphic form of silica is used in the manufacture of glass silica brick pottery abrasives and mortar. Cristobalite and more rarely tridymite are formed when quartz or amorphous silica is heated and are sometimes employed as siliceous refractory materials.Consequently, exposure to crystalline silica dust can occur in a wide range of industrial environments including mining construction agri-culture and manufacturing. Silicosis the lung condition caused by inhalation of fine particles of crystalline silica has long been recognised as a health hazard in many of these industries. The tissue damage mechanism has been attributed to the physical shape solubility crystalline structure or cytotoxicity of the silica particles to macrophages. 1 However, the importance of crystalline silica in the pneumoconiosis developed by coalminers remains a controversial subject and many investigations have been carried out in an attempt to clarify the position.2-8 Pulmonary symptomatology associated with silicosis includes coughing dyspnea wheezing a pro-gressive diminution in lung function and repeated non-specific chest ailment .q - l ( ) The occupational exposure limits" for respirable dusts containing the different types of crystalline silica range from 5 to approximately 0.1 mg 117-3 for quartz-containing dusts and from 2.5 to approximately 0.05 mg m-3 for dusts containing cristobalite or tridymite. The practical limit for any particular respirable dust depends on its crystalline silica content. Because of the serious adverse pulmonary effects associated with crystalline silica inhalation many techniques have been developed for the quantitative analysis of the main poly-morphs especially quartz. These include gravimetry 12,13 spectrophotometry,1419 differential thermal a n a l y ~ i s ~ ( ~ * ~ X-ray diffractometry2s50and infrared spectrophotometry.51-68 Of these X-ray diffractometry and infrared spectropho-tometry are the principal techniques currently employed; each is capable of identifying free crystalline silica in the presence of silicates.69-71 In many laboratories infrared spectropho-tometry is the preferred technique in view of its relatively low cost; the compressed pellet ,5441 direct differential scanning (direct filter)"-@ or attenuated total reflectance (ATR)65-67 method is used.In the last two methods the membrane filter on which the sample dust is deposited is placed directly into the infrared beam or in an ATR unit without further sample preparation. In the compressed pellet method the sampled Crown Copyright.dust is recovered from the membrane filter by incineration and is mixed with potassium bromide or caesium iodide from which a pellet is prepared for infrared analysis. The compressed pellet method was originally used51 in the quantitative analysis of crystalline silica by infrared spectro-photometry and has subsequently been employed in the analysis of quartz-containing coal dust ,54,57,58 sedimentary rocks,55 granite dust59 and grinding wheel dust.60 However, the method is relatively time consuming and is subject to possible errors in the preparation of the pellet. Both the ATR and direct differential scanning methods eliminate the pre-parative stage of sample analysis but only the latter method lends itself to automated analysis. Consequently as the Health and Safety Executive's Occupational Medicine and Hygiene Laboratory processes over 2000 crystalline silica samples per annum the direct differential scanning method was examined in detail including an assessment of particle size sample re-radiation and mineral interference effects.Experimental Apparatus Infrared spectrophotometer. A Perkin-Elmer 580A ratio recording instrument was used fitted with a pair of Perkin-Elmer universal infrared multi-sampling accessories and linked on-line to a dedicated computer system. A supplemen-tary control interface was installed that permitted computer selection of either or both of the multi-sampling accessories. Computer system. A DEC PDPll/O5 mini-computer with 60 kbyte of usable memory equipped with two Plessey DDllB, RKOS-compatible dual hard disc systems and a Tektronix 4014-1 video display unit fitted with a Tektronix 4631 hard copy unit was controlled from a DEC LA36 terminal.Sampling device. BCIRA 25 mm diameter cyclone elutria-tor (Casella Ltd. type A7650/1). Samplingpump. Capable of drawing air through the filter at a flow-rate of 1.9 1 min-1. Filter. An acrylonitrile - PVC copolymer 25-mm diameter, 0.8-pm pore size (Gelman Sciences type DM-800). Standard atmosphere apparatus. Specially constructed (G. Farley Ltd. Enfield Middlesex). See text. Cascade impactor. Anderson 2000 Inc. 1 ACFM particle fractionating sampler (Anderson 2000 Inc. Atlanta GA, USA). Hydraulic press. 16-ton manually operated (Beckman-RIIC Ltd. type P-16) 1118 ANALYST SEPTEMBER 1984.VOL. 109 Evacuable die. 6-mm diameter (Specac Ltd.). Ball mill. Glen Creston Model 27-280 fitted with tungsten carbide-lined grinding cylinder and ball pestles. Agate mortar and pestle. Particle size measuring system. Coulter Counter Model TA 11 complete with population accessory and x - y recorder. Analytical balance. Perkin-Elmer AD-2Z auto-balance. Reagents Quartz. Sikron F-600 (a commercial product of Quartz-werke Frechen F.R.G.). Available from the HSE Occupa-tional Medicine and Hygiene Laboratories reference number A9950. Cristobalite. Supplied by Johns Manville Corp Denver. CO USA. Tridymite. Supplied by British Steel Corp. Middlesbor-ough Teesside. Additional minerals. See text (R. Parkinson & Co Ltd., Shepton Mallet Somerset).Potassium bromide. Anhydrous spectrograde material (Fisons) stored at 120 "C. Procedure Condition the membrane filters to be used in the analysis by distributing them across clean sheets of tissue paper and placing them in a laboratory oven at 50 "C for several hours. Ideally an entire box of 100 filters should be treated in one operation. Place the membrane filters into individual sample tins leaving the lids slightly ajar and allow to stand overnight in the balance room. Close the lids of the tins prior to weighing to prevent any mass change from moisture gain or loss during the relatively long period required to weigh all the filters. Weigh each filter to 2 1 pg on an electrobalance and replace in the sample tin until required. Select reference filters by mass in 0.3-mg increments so as to span the mass range of the batch of filters.Use the remaining filters for sampling. In an uncontaminated atmosphere fit a membrane filter into the BCIRA cyclone elutriator. Attach a portable sam-pling pump to the outlet of the cyclone and draw a sample of the test atmosphere through the device at a flow-rate of 1.9 1 min-1 for a half-shift period (nominally 4 h although in very dusty conditions shorter sampling times are necessary to prevent dust loss from the surface of the filter). After completion of sampling replace the membrane filter in the sample tin and carefully transport back to the laboratory in a horizontal position with the minimum of vibration. Leave the lids of the sample and reference tins slightly ajar overnight in the balance room and repeat the weighing procedure.Any mean variation in the reference filter masses is incorporated into the masses of the sample filters. Sort the sample filters into batches in which the original unused filter masses are within a 0.3-mg mass difference range and load into the sample beam infrared multi-sampling accessory. The filters are held in Perkin-Elmer rigid film holders (part No. 5101-4922) by pieces of folded card with aligned 20-mm diameter holes punched in them (Specac Ltd., part No. Q5764). The reference filters are placed in batch sequence in the reference beam multi-sampling accessory in a similar manner. In the absence of a second multi-sampling accessory and supplementary control interface a matched-mass sample filter batch may be scanned against a single reference filter.For small sample numbers it is convenient to scan each sample filter individually by mounting it and the corresponding reference filter in 25-mm diameter rotable sample holders (Specac Ltd. part No. 12600 HSE). Sample filters with eccentric dust depositions are scanned several times by rotating the filter holder through 60" or 90" increments and averaging the infrared spectral readings. Adjust the base-line setting for all the sample filters in the batch to be scanned to 95% transmittance between 400 and 1500 cm-1 and scan each filter five times over this wavenum-ber range. Using the computer facilities available process each data-averaged spectrum. Depending on the intensity and complexity of each spectrum smooth the data convert into absorbance ordinate scale expand and subtract the standard spectra of interfering minerals after ordinate scale ratioing.Measure the absorbances of the quartz doublet peaks at 780 and 799 cm-1 to kO.001 absorbance unit by drawing a base line tangentially between the minimum absorbances at approximately 730 and 850 cm-1 (although these values may vary depending on the shape of the spectral envelope) and determine the concentration from standard calibration graphs. Monitor the computer processing of the spectral data on the video display unit from which hard copies of relevant traces may be obtained for permanent retention. If cristobalite or tridymite is present determine their concentrations directly from the absorbance values at 621 and 567 cm-1 respectively.When additional overlapping absorp-tion bands are present in this region it is necessary to measure the 796 and 789 cm-1 band absorbances instead. In most instances quartz will also be present and can be computer-subtracted if desired. Alternatively measure quartz at 695 cm-1 or preferably 780 cm-1 (in the presence of cristobalite only) and calculate the absorbance at 799 cm-1 from the relevant calibration graph. The absorbance of cristobalite or tridymite at 796 or 789 cm-1 is then determined by difference. Using this procedure the concentration of cristobalite or tridymite can be determined to within k25% of the true value. However in excess of 5000 samples containing crystalline silica from a wide range of industrial processes have been analysed in these laboratories and to date tridymite has not been detected whereas cristobalite has been found in 10% of the samples.Although computer processing of the spectral data permits data averaging from multiple scanning together with signifi-cant spectral enhancement it is possible to determine crystalline silica concentrations using a standard double-beam infrared spectrophotometer alone. With instruments that do not have an asynchronously double-chopped optical system a matched pair of 9 pm germanium cut on/blocking filters (Specac Ltd. part No. 58.013) should be fitted prior to the sample and reference membrane filters. In this way sample re-radiation effects are reduced to a negligible level. Results and Discussion Selection of Filter Material An examination of the mid-infrared spectra of quartz, cristobalite and tridymite (Fig.1) indicates that the absorption bands most suitable for analysis lie between 550 and 850 cm-1. Between 850 and 1200 cm-1 and below 550 cm-1 all silicates exhibit strong absorption bands broadly characterised as asymmetric Si-0-Si stretching and 0-Si-0 bending vibra-tions respectively.72 The intermediate absorption region bands are of more use for identifying individual silicate structures and have been ascribed to Si-Si stretching,73 Si-0-Si symmetrical stretching74375 and activation of the symmetric stretch of the principal tetrahedral SiOj structural unit.72 Consequently the quartz doublet (780 and 799 cm-1) and singlet (695 cm-I) the cristobalite singlets (621 and 796 cm-1) and the tridymite singlet (789 cm-1) and shoulder (567 cm- 1) were used for all quantitative determinations of crystalline silica.A wide range of membrane filters from various manufactur-ers were examined to determine which if any would transmit infrared radiation between 550 and 850 cm-1. In addition it was considered important for the filter to transmit infrared radiation in the regions on either side of this waveband to facilitate identification of additional sample spectrum details ANALYST SEPTEMBER 1984. VOL. 109 1119 100 50 $ 0 6 c co c c 'E 50 E + 0 50 1500 1300 1100 900 700 500 Wavenumbericm ~ 1 Fig. 1. (b) cristobalite; and (c) tridymite Infrared absorption spectra of crystalline silica.(a) Quartz; Of the filter materials examined cellulose acetate (Millipore, types EA EH and EG) PVC (Sartorius type SM128; Gelman Sciences type VM) and an acrylonitrile - PVC copolymer (Gelman Sciences type DM) possess the required characteristics although each material has infrared absorption bands within the region of interest. The ideal filter material should also have a high collection efficiency be minimally affected by atmospheric moisture have a low electrostatic charge effect exhibit a smooth flat infrared base line when matched by mass and be suitable for X-ray diffractometric (XRD) analysis if possible. The infrared spectral matching is particularly important as the determination of low crystalline silica concentrations can involve using 10- or 15-times ordinate scale expansion.XRD compatibility is also a useful asset as correlation of data between the two techniques may be achieved using the same sample filter.76 Infrared spectral matching was carried out for each of the above filter materials by scanning weighed blank filters from 400 to 1500 cm-1 and measuring the degree of mis-matching, in absorbance in the wavenumber regions selected for crystalline silica analysis. Plots of absorbance against mass difference were obtained by retaining a filter in the reference beam and scanning against a range of filters of increasing mass in the sample beam. Cellulose acetate membrane filters were found to be relatively sensitive to mass differences an absorbance of 0.05 being obtained for a mass difference of 1 mg for the band produced between 900 and 800 cm-1.The accurate mass-matching of PVC-based filters was much less important filter mass differences of up to 0.5 mg producing no significant absorption bands in the region of interest. It was also found in some instances that scanning the same filter, regardless of filter type across different diameters produced slightly differing spectra. Therefore in order to achieve optimum filter cancellation it was necessary to scan and store the spectrum of a blank filter oriented in a known way with respect to the infrared beam. The filter was then used for sampling dust and subsequently re-scanned in the same orientation. In this way the filter acted as its own reference filter. Such a procedure although longer was used when preparing standard calibration lines.In the absence of any drying treatment the mass of membrane filters will be affected to a greater or lesser extent, by changes in atmospheric moisture levels. Certainly with paper filters the response to moisture after drying is so rapid as to render the filter unweighable. Therefore the mean variation in the masses of ten 25-mm filters with respect to changes in the atmospheric humidity of the laboratory was noted over a period of several weeks. The filters were stored in sample tins with their lids slightly ajar and were re-weighed six times during which period the relative humidity ranged from 36.5 to 50%. The average temperature was 21 "C. The three PVC-based filter materials produced mass changes from 0.04 to 0.07 mg and the cellulose acetate filters produced mass changes up to 0.25 mg.The cellulose acetate filters showed no mass-change trends with respect to relative humidity changes, unlike the PVC-based filters. These findings are in broad agreement with previous investigations.77 The filter types were further tested to assess their response to drying treatments by storing them in a desiccator or a laboratory oven at 50 "C for 1 2 and 24 h. A 1-h drying period in the laboratory oven was found to remove more moisture from all the filter types than storing them for 24 h in a desiccator. The optimum conditioning procedure was found, by trial and error to consist in drying the filters in a laboratory oven at 50 "C for 2 h and then allowing them to re-humidify in sample tins with the lids ajar overnight in the balance room.Of the PVC-based filter materials type DM exhibited the least electrostatic charge effect and could if necessary be weighed directly from the box without the need to remove any charge with a static eliminator. Conversely types VM and SM 128 not only required charge elimination prior to weighing, but were subsequently found to adhere to the sides of the sampling tins and cyclones. It is possible that the high electrical resistance of these filters77 prevents the complete elimination of the surface charge by slowing the migration of charges from the complex internal filter structure to the outer face. The presence of a residual electrostatic charge implies that some dust may be repelled from these filters during the sampling procedure.XRD compatability was assessed by scanning each filter material from 11.05 to 1.43 A using a Philips PW 1011/00 vertical powder diffractometer fitted with a graphite-crystal focusing monochromator and proportional counter. Copper Ka X-radiation was produced with generator settings of 44 kV, 36 mA 1 A divergence slit and 0.1 A anti-scatter receiving slit. All three PVC-based filter materials produced a weak continuous X-ray diffraction pattern from 11.05 to approxi-mately 3.00 ,& and were considered suitable for the XRD analysis of crystalline silica (for example quartz at 4.25 3.43, 1.82 and 1.54 A). Cellulose acetate filter material exhibited a much stronger continuum pattern over the same wavelength range and would be unsuitable for quartz analysis using either the primary or secondary diffraction lines at 3.34 and 4.25 A, respectively.On the basis of the above results type DM800 membrane filters were considered to be the most suitable for the determination of crystalline silica and they were therefore used in all the experiments. Fig. 2 shows the infrared spectrum of a DM800 filter from 400 to 1500 cm-1 together with the spectrum obtained from a mass-matched pair. Preparation of Mineral Standards Atmospheres of quartz cristobalite tridymite and a range of minerals commonly found with crystalline silica in occupa-tional environments were generated in a specially constructed glass dust cloud chamber (Fig. 3) housed in a fume cupboard.The top of the dust chamber was sealed with a 5 mm thick circle of clear Perspex into which had been machined a 2 mm deep grove that mated with the rim of the chamber. Four equidistant holes were drilled into the Perspex circle to accommodate up to four 25 mm diameter BCIRA cyclone elutriators. Approximately 1-2 g of the mineral of interest was placed in the bowl at the bottom of the chamber. When necessary mineral samples were initially ground to smal 1120 ANALYST SEPTEMBER 1984 VOL. 109 / I I I 80 -ai 5 60 -c. .-I-0 1500 1300 1100 900 700 500 I I I I 2oP ‘ I I Wavenurn ber/cm -Fig. 2. Infrared absorption spectrum of DM800 membrane filter A, single filter in sample beam; and B additional mass-matched filter in reference beam L-20 crn Lj e Fig.3. mineral dust atmospheres Glass dust cloud chamber used for generating and sampling particle size in a ball mill for 10-30 min; the ends of the tungsten carbide-lined grinding cylinder were taped with PVC tape in order to prevent leakage. The sampling pumps were connected to the loaded cyclones switched on and the flow-rate was adjusted to 1.9 1 min-1. Compressed air was applied for 2-3 s to one of the two side-arms to produce an atmosphere of dust in the main chamber sufficiently dense to appear cloudy. Suitable amounts (10-1000 pg) of dust were collected typically within 5-100 s. From each batch of mineral samples prepared a “standard“ mineral filter was selected whose maximum infrared spectral absorbance was between 0.5 and 0.7. To increase the signal to noise ratio of the spectrum each standard mineral filter was scanned five times between 400 and 1500 cm-1; the spectrum was stored on the computer hard disc system (digitised at 1.0 cm-1 increments).These spectra were subsequently used in the spectral manipulation (usually subtraction) of real samples where mineral interference presented a problem between 600 and 900 cm-1. Preparation of Calibration Graphs Using the dust cloud chamber techniques approximately 30 standard filters of each of the three crystalline silica poly-morphs were prepared covering the mass range 10-1000 pg. In addition to data averaging by scanning each filter five times, the stored spectra (1 cm-1 data point increments) were convoluted with a Savitsky - Golay quadratic - cubic smooth-0.02 a.uS k! _- A .rr I (b’ 0.02 a.u.T A I (d’ l- 0.02 a .u I 900 800 700 Wavenurnber/crn-’ Fig. 4. Infrared absorbance spectra of standard quartz dust depo-sited on DM800 membrane filters. ( a ) 15 pg; ( b ) 30 pg; (t) 75 pg and (4 134 CLg 0.6 1 fC 0.5 -c 0.4 -0 200 400 600 800 1000 1200 Mass of quartziFg Fig. 5. Calibration graphs for standard quartz dust deposited on DM800membrane filters. A. 695 cm-1; B. 780cm-l; andC 799cm-’ ing function’s data processing routine ( 5 to 25 point) to reduce spectral noise. In most instances 5 to 11 point smoothing was sufficient. The use of digital smoothing was found to attenuate the peak intensities slightly but not to an extent that significantly affected the results obtained. Typical absorbance spectra of quartz standards are shown in Fig.4; the positions of the base-line tangents (between the minimum absorbances at 730 and 830 cm-1 approximately) used in the preparation of the calibration graphs are also indicated. Calibration graphs for the quartz peaks at 799,780 and 695 cm-1 are shown in Fig. 5 . A linear relationship in conformity with the Beer - Lambert law was obtained in each instance. Above 1 mg (ca. 0.6 absorbance unit) the 799 and 780 cm-1 lines became non-linear. Particle Size Effects Theoretical studies79-82 have shown that materials of equi-valent chemical composition but different particle size have significantly different infrared spectra. They indicate that the absorbances of particles smaller than the incident wavelength ANALYST SEPTEMBER 1984 VOL.109 1121 increase with decreasing particle size and the true limiting absorbance at strongly absorbing wavelengths will only be achieved at particle sizes below 0.1 km. To determine the relationship between particle size and quartz absorbance eight quartz dust fractions of narrow particle size distribution were prepared using the glass dust cloud chamber previously described in conjunction with an Anderson 2000 cascade impactor. For this experiment a suitable diameter desiccator lid fitted with a B34/35 double cone adaptor was secured to the top of the main chamber and connected to the cascade impactor inlet with a short length of 25 mm i.d. plastic tubing. Sampling was carried out for 30 min, during which period compressed air was occasionally applied to the side-arm in order to maintain the dust cloud concentra-tion.The process was repeated several times to obtain sufficient stock material. After each sampling period the sized dust fractions were carefully removed from each collection plate to minimise possible re-entrainment. In addition a fresh supply of standard quartz was placed in the dust chamber reservoir to prevent the accumulation of a disproportionate amount of coarse particles. " 1 2 3 4 5 67891 2 3 4 567891 2 3 4 56789 Equivalent spherical diameter/pm Fig. 6 . graphs for a standard quartz sized fraction Percentage cumulative (A) and differential volume (B) 10.2 a.u. The particle size distribution of each sized fraction was determined by Coulter analysis using a Model TAII Coulter Counter.The instrument measures the volume distribution of the test particles when dispersed in a suitable e!ectrolyte, based on the assumption that the voltage pulse generated when a particle passes through a small orifice is directly proportional to the particle volume.83 The technique has previously been used successfully for measuring the size distributions of coal mine dusts.84-86 The Coulter analysis procedure consisted in adding approxi-mately 0.1 mg of dust to 200 ml of saline solution in a 250-ml beaker and immersing it in an ultrasonic bath for 1 min. The beaker was then presented to the orifice tube and the sample run until the instrument indicated that sufficient counts had been recorded to prepare a statistically viable histogram.A 50-pm orifice tube was used. Plots of percentage cumulative and differential volume were prepared (Fig. 6). The size parameter indicated in Fig. 6 is the equivalent spherical diameter below which 50% of the sample vohme (mass) occurs. Five potassium bromide pellets were prepared from 40 mg of anhydrous KBr and 100 k 1 pg of each sized quartz fraction using a 6-mm die. The pellets were scanned from 400 to 1500 cm-1 in the manner previously described and the mean absorbances of the 799 780 and 695 cm-1 bands determined for each sized fraction. Fig. 7 illustrates the marked particle size effect on the infrared spectrum of quartz over the respirable particle size range. Plots of absorbance in relation to particle size are shown in Fig. 8 together with the results of earlier work by Dodgson and Whittaker,58 the latter after normalisation at 5 pm to take account of differences in pellet size infrared beam areas and mass of quartz.T1:e results agree well down to a particle size of approximately 1.5 pm the lower limit of the Dodgson and Whittaker measurements. Below 1.5 pm the absorbance decreases with decreasing particle size, contrary to theoretical expectations. However it has been suggested,87 and subsequently verified,8&") that quartz par-ticles possess a thin amorphous surface layer that blends via a transitional layer into a crystalline core. Determinations of the thickness of the surface layer by several meth0ds87.~' indicate that the layer attains a constant maximum thickness of approximately 0.03 pm.Consequently it could be argued that 1300 1100 900 700 500 1300 1100 900 700 500 1300 1100 900 700 500 1300 1100 900 700 500 Wavenum bericm - I Fig. 7. Particle size effect on the infrared absorbance spectrum of 100 pg of standard quartz ( a ) 8.00 pm; (bj 6.50 pm; (cj 5.20 pm; ( d ) 3.75 prn; ( e ) 2.35 pm; and cf 1.70 pm. These values are the equivalent spherical diarnete,rs below which 50% by mass of the sized fractions occu 1122 ANALYST SEPTEMBER 1984. VOL. 109 as the percentage volume of such a layer rapidly increases with decreasing particle size (9% at 2 pm. 37% at 0.5 pm diameter. assuming a thickness of 0.03 pm) the attenuation in absor-bance is effeqtively due to a reduction in the crystalline quartz content of small particles. Indeed the profile of the percent-age crystallinity curve below 1.5 pm (Fig.8) closely follows the absorbance curves. A review by Harris and Revell'" agrees with the finding that attenuation occurs below 1.5 pm. Gade and Lufts3 have suggested a procedure for correcting absorbance measurements for particle size in the analysis of respirable dust from coal mines. This was based on the relationships between the particle size and the ratios of the absorbances at 799 and 780 cm-1 to the minimum absorbance between these two peaks at 790 cm-1. The present data when plotted in this manner (Fig. 9) show remarkably good agreement with Gade and Luft's original findings and the subsequent work of Dodgson and Whittaker,sx especially in view of the differences in sample preparation instrumentation and measurement.Distribution of Dust on Filters Dust collection in a cyclone elutriator is fundamentally an inertial process with some secondary effects attributable to the turbulent nature of the flow field inside the cyclone. Sampled air passes through the entrance slit spirals downwards along the cyclone wall and then reverses direction and flows upwards into the filter chamber and the outlet. The aerodynamic field inside the cyclone is generally described as a double vortex with a low-pressure highly turbulent vortex.9'-9i Conse-quently the helical path of the air emerging into the filter chamber may cause the dust to be thrown out from the central axis of the cyclone resulting in a lower deposition at the centre of the filter.Further the flow-rate through the filter chamber may be sufficiently high for a large proportion of the airborne dust to be impacted directly on to the central zone of the filter. In either instance it is extremely likely that there will be a varying mass distribution across the diameter of the filter. The shape of the concentration profile of sampled dust across the cyclone filter was examined by mounting a quartz-loaded filter in a Perkin-Elmer rigid film holder. which in turn was secured to the moving arm of a Vernier caliper gauge. A blank plate with a l-mm hole machined in its centre was secured to the fixed arm of the gauge so that the filter was totally covered and the hole aligned with the centre of the filter. The Vernier caliper gauge was rigidly clamped on to a sample holder in the sample well of the infrared spectrophotometer with the blank plate between the filter and the source and positioned so that the centre of the sample beam was coincident with the hole.A blank plate was positioned in a similar manner in the reference beam between a blank matched reference filter and the source. The sample filter was scanned five times between 600 and 900 cm-1 and the absorbance of the 780 and 799 cm- * quartz peaks was measured. This procedure was repeated at l-mm increments across several diameters of the filter by adjusting the caliper gauge; the mean values of the absor-bances were noted at each distance from the filter centre. Fig. 10 shows the significantly higher mass distribution in the central region of the filter.At the edges of the filter the mass distribution was approximately one third that in the central region. However as the absorbance of respirable size crystalline silica is dependent on particle size and as the helical path of the air will cause particles of smaller aerody-namic mass median diameter (i.e. the equivalent diameter of a unit density sphere having the same settling velocity as the particle under consideration) to impact further from the filter centre the absorbance relationship in Fig. 10 is not linear with respect to distance across the filter. An observed increase in the 7991min. and 780/min. absorbance ratios confirmed the decreasing particle size trend across the filter. A maximum 100 A - 80 1 2o 0 1 2 3 4 5 6 7 8 Particle size (equivalent spherical diameter1:pm Fig.8. Relationship between particle size and the infrared absor-bance of 100 pg of standard quartz for this work (0.695 cm I +. 780 cm-l; and 799 cm- 1 ) and. after normalisation a t 5 pm. for Dodgson and Whittaker's previous work>* (0.695 cm-l; 0.7XO cm I and @. 799 cm 1 ) . Curve A shows the percentage crvstallinity curve for ii particle with an amorphous silica surface layer 0.03 pm thick 3 1 1 I I I I I I I 0 1 2 3 4 5 6 7 8 Particle size (equivalent spherical diameter)+tm Fig. 9. Relationship between particle size and the quartz infrared absorbance ratios at 7801790 and 799/790 for this work (0 0) and previous work by Dodgson and Whittaker's ( W . 0) and Gade and L u f P (+. 0) E 2 e z z m 1 r-; I 1 I I 1 I 0 5 10 15 20 25 Distance across filterimm Fig.10 diameter DM800 membrane filter Mass distribution of sampled quartz dust across a 25 mm value was attained approximately 5 mm from the centre. the ratio decreasing further out probably owing to the decreasing mean particle size. The variation in mass and particle size distribution of real samples could in theory substantially affect the results obtained from fixed calibration lines. Further we have found that for real samples where crystalline silica usually rep-resents only a small percentage of the sample dust the absorbance ratios can be modified by a steeply sloping base line owing to the presence of strongly absorbing species such as oxides of iron or small absorbance band coincidences.Fortunately the mass distribution of the sampled dust is heavily skewed towards the upper particle size cut-off of the cyclone i.e. 4-7 pm where particle size effects are substan ANALYST. SEPTEMBER 1984. VOL. 109 100 80 8 6 0 -1123 --100 ,- -L- 7 i, 40 j 20. 5 0 - _ i L 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 . 2 3 4 5 Equivalent spherical diametedpm Fig. 11. foundry dust - hand fettling process; and ( c ) quarry dust Particle size distribution for ( N ) standard quartz dust; ( h ) tially less. Fig. 1 I shows typical particle size distributions for a range of industrial processes. together with A9950 standard quartz. and indicates that it is not necessary to apply particle size corrections when using a cyclone elutriator for sampling both samples and standards.The good correlation found between X-ray diffraction. where the particle size effect profile is significantly different ' ~ ~ ~ and infrared analysis con-firms these observations.7h In addition. any eccentricity in the mass distribution was generally found to be approximately +2% and scanning the filter five times in a fixed orientation was judged to be suitable for routine samples. The small percentage of sample filters with eccentric dust distributions were scanned three times at 60" or 90" increments by rotating the filter holder. Sample Re-radiation Effects Although many compounds that commonly occur with crystal-line silica in industrial environments have specific mid-infrared absorptions several possess extensive absorption continua.For example oxides of iron commonly found in foundry atmospheres are especially opaque to infrared radiation. Having absorbed substantial amounts of radiant energy such compounds will rapidly return to their vibra-tional ground states by collisional deactivation and radiation processes and effectively act as black-body radiators within this region of the electromagnetic spectrum. The infrared detector will therefore receive more radiant energy than the sample absorption processes would suggest owing to sample re-radiation which will effectively attenuate any absorption 0 1500 1300 1100 900 700 500 Wavenumbericm- ' Fig. 13. dust from a steel foundry Infrared absorption spectrum of a black sample of airborne bands of interest.In order to eliminate this problem several co m ni e r c i a 1 dispersive i n f r a re d spec t ro p h o t om e t e rs . i n c I u d i n g the Perkin-Elmer S8OA used in this work. are equipped with an asynchronously double-chopped optical system. Previous workhx has shown that most inexpensive instruments currently employed in the analysis of crystalline silica will produce increasingly erroneous results as the true transmission of the sample spectrum base line decreases. Such instruments may give up to 75% attenuation in the crystalline silica concentra-tion of samples containing high loadings of iron oxides ( 5 mg). Further with this combination of instrument and sample type it is possible to obtain a transmission spectrum of the filter material superimposed on the sample absorption spectrum to produce an almost unrecognisable spectrum.Sample re-radiation effects with a conventional dispersive instrument (Perkin-Elmer 399B) were reduced to a negligible level by fitting a matched pair of 25 mm diameter 9-pm germanium cut on/blocking filters (Specac) between the source and the sample and reference filters. The cut on/ blocking filters were mounted in the sample holders and separated from each membrane filter by a 5 mm thick 20 mm i.d. steel ring. The mid-infrared spectrum is shown in Fig. 12, together with the background spectrum obtained by inserting a matched filter in the reference beam. A typical mid-infrared spectrum of respirable-size airborne foundry dust obtained by the direct differential scanning method is shown in Fig.13, which clearly illustrates the increasing absorption at shorter wavelengths. Fitting a matched pair of cut on/blocking filters prior to the sample and reference membrane filters revealed that the true absorbance of the quartz doublet at 780 and 79 ANALYST. SEPTEMBER 1984. VOL. 109 10.02 a.u. 900 800 700 Wavenumber’cm-’ Fig. 14. Infrared absorbance spectrum of airborne iron foundry dust. (a) Without germanium cut oniblocking filter; and ( h ) with a filter in both sample and reference beams 100 80 8 $ 60 m c. - ._ $ 40 F I-c 20 0 ’ I I I I I 1500 1300 1100 900 700 500 Wavenumber cm-’ Fig. 16. steel foundry showing the presence of quartz and cristobalite Infrared absorption spectrum of respirable-size dust from a I - 1 I ‘ 1 , Mass of quartz big Fig.15. Correlation of quartz results for a typical batch of blach iron foundry samples bet\veen conventional dispersive (ordinate) and asynchronously double-chopped (abscissa) infrared instruments A. without cut oniblocking filters in former instrument and B. with filters cm-1 for this particular sample was a factor of 2.5 greater than a routine analysis using a simple dispersive instrument might suggest (Fig. 13). Comparison of the results obtained using a pair of cut on/blocking filters fitted to a simple dispersive instrument with an asynchronously double-chopped instru-ment and from XRD analysis showed that in the absence of mineral interferences the correlation between the three sets of results was within 10%.indicating that sample re-radiation was effectively eliminated. Fig. 1.5 illustrates the very good agree men t obtained bet w e e t i a sync h r o no us1 ly’ double -c h o ppe d and conventional dispersive instruments for a batch of black foundry samples containing zirconium silicate (zircon) and iron oxide when cut on/blocking filters were fitted to the latter instrument. Mineral Interference Effects The presence in the dust sample of any compound that possesses absorption band coincidences with the bands used in crystalline silica analysis may produce erroneous results. 1500 1300 1100 900 700 500 Wavenumber c m - ’ Fig. 17. Infrared absorbance spectrum of airborne dust from the brick and tile manufacturing industry showing substantial kaolinite interference.(a) Original spectrum of sample; ( b ) spectrum of kaolinite; and ( c ) sample spectrum after scaled absorbance substrac-tion of kaolinite Foundry atmosphere dusts were generally found to contain no significant interferences although the presence of oxides of iron or graphite in sufficient amounts produced a steeply sloping base line. Zircon calcite dolomite magnesite. forsterite and wollastonite were sometimes encountered in foundry atmospheres. but did not interfere in quartz analysis. However the strong singlet peak at 620 cm-1 in the spectrum of zircon totally enveloped the 622 cm-1 cristobalite band and when these compounds were present together it was necessary to use the 797 cm-1 cristobalite band for quantitative measurements.Although this singlet peak is three times more intense than the 622 cm-1 band. quartz. which is usually present will interfere. Indeed the presence of cristobalite in a foundry sample was generally identified by a high (B1.5) 799/780 cm-1 quartz doublet ratio (Fig. 16). In this instance, quartz was measured at 69.5 cm-1. when sufficient was present (>50 pg) or 780 cm-1. The absorbance at 799 cm-1 was calculated from the 799h9.5 cm-1 (4.19) or 7991780 cm-1 (1.17) calibration graph ratios and the absorbance of cristobal-ite then determined by difference. This procedure produced ANALYST SEPTEMBER 1984 VOL. 109 1125 uniform Gaussian-shaped band for cristobalite at 797 cm-1. Because the cristobalite band extends to 780 cm-l and beyond it was found preferable to use the 695 cm-1 band for measuring quartz whenever possible.Tridymite the other main crystalline silica polymorph has yet to be identified in this laboratory although over 5000 dust samples have been analysed. The most common mineral interferences encountered were certain clays and feldspars. They are used extensively in the ceramics industry and often occur in mining and quarrying environments. Kaolinite as a high percentage of total respirable dust mass was found in a substantial number of samples from the pottery and brick manufacturing industries. The interfering peaks at 754 and 795 cm-1 were quantitatively removed by scaled absorbance subtraction using the charac-teristically shaped doublet at 914 and 938 cm-1 as a reference band for obtaining a close approximation scaling factor.Scaled absorbance subtractions were performed by adjusting the scaling factor until the kaolinite band at 754 cm-1 was just eliminated (Fig. 17). In some instances this procedure left some residual absorbance at 914 cm-1 which was probably attributable to the presence of small amounts of either illite or montmorillonite both of which have weak absorptions in this region. For this reason it has been suggested97 that the 1010 and 1033 cm-1 peaks should be used for scaled absorbance subtraction in the analysis of kaolinite in coal. Although the layered structure of kaolinite is disrupted above 800 "C to produce mullite and cristobalite these minerals were only occasionally encountered in trace amounts in pottery environ-ments.Muscovite another clay mineral found in coal dust and used in the roofing felt industry required scaled absorbance subtraction at 829 cm-1 when present in sufficient amounts. Nepheline a framework silicate occasionally found in large amounts in certain ceramic industries was spectrally subtrac-ted (using the 593 or 645 cm-1 peaks) to eliminate the broad double band stretching from 800 down to 670 cm-1. Of the feldspar group of minerals orthoclase which is found particularly in granite quarries was the commonest interferent. Albite and anorthite were also encountered infrequently. In all three instances the spectra have a complex quadruple peak absorption between 810 and 700 cm-1 which extensively modifies the shape of the quartz doublet.Scaled absorbance subtraction was performed using the peak centred around 650 cm-1 for each mineral. Table 1 lists the positions of the main absorption bands for the interfering minerals most commonly encountered together with the peaks usually employed for scaled absorbance subtraction. Optimum subtraction was achieved by digitising spectral data at 1 cm-1 increments and using minimum smoothing compatible with reducing any noise to an acceptable level (usually 5 to 11 points). Digitisation at larger data point increments was found to attenuate the peak intensities without smoothing. However provided the calibration graphs are prepared using similar increments (4 cm-1) and smoothing, quantitative analysis may be performed at reduced sensitivity. Table 2 illustrates the reduction in peak height at the 695,780 and 799 cm-1 quartz bands assuming a value of 1.0 for each band using a 1 cm-1 data point increment and no smoothing.Standard mineral spectra were prepared using a cyclone sampler and the dust cloud chamber previously described in order to obtain a particle size distribution similar to that of real samples. Table 1. Positions of the main absorption bands for minerals that commonly interfere in crystalline silica analysis Mineral Albite . . Anorthite Cristobalite Daphnite . . Kaolinite . . Marcasite Mullite Muscovite Nepheline Orthoclase Quartz . . Ripidolite . . . . . . . . . . . . . . . . . . . . . . . . Steatite . . . . Tridymite . . Band wavelength*/cm- I 402 (m) 418 (m,sh) 432 (ms) 465 (m,d), 477 (m,d),535 (m),598(ms),614(m,sh), 652 (m) 726 (mw,q) 746 (m,q) 788 (m,q), 1005 (s,d) 1039 (s,d) 1105 (s,d) 1157 (s,d) 409 (m) 435 (msh) 481 (m) 538 (m) 577 (m), 625 (mw) 669 (mw) 728 (mw,d)) 758 (mw,d), 950 (s,d) 1021 (s,d) 1084 (ms sh) 1160 (ms) 496(ms),622(w) 797(mw) 1103(s) 1199 429 (ms,sh) 467 (ms) 539 (m) 610 (m) 667(m), 433 (m) ,474 (ms) ,547 (ms) ,695 (mw) 754 (ms sh) 771 (m) 798 (m) 991 (s) (w,d).795 (w,d) 914 (m) 938(m,sh) 1010 (s,d) 1033 (s,d) 1115 (m) 421 (mw) 533 (mw) 605 (mw,d) 625 (mw,d), 667(mw) 829 (m) 1016 (m) 1081 (s) 1153 (s) 468 (ms) 591 (s) 748 (s) 837 (s) 914 (s), 1135 (ms,d) 1162(ms,d) 481 (ms) 535 (ms) 695 (w) 751 (w) 802 (w), 829 (w) 935 (m,sh) 1029 (s) 1062 (ms,sh) 432 (s) 475 (s,sh) 512 (m,d) 542 (m,d),593 (ms) 645 (m) 719(ms) 762 (m,sh) 1014 (s), 1132 (s) 435 (ms) 464 (mw,sh) 540 (mw) 593 (m) 649 (mw) 728 (mw,t) 745 (mw,t).767 (mw,t) 788 (w,sh) 1046 (s) 1140 (s) 467 (ms) 524 (mw). 695 (w) 780 (m,d) 799 (m.d) 1087 (s) 1164 (ms,sh) 482 (ms) 571 (m) 647 (mw,d) 669(mw,d), 765 (mw),##O(m,d) 904(ms,d),992(s) 426 (ms,sh) 451 (s,d) 467 (s,d) 670 (ms) 761 (w) 826 (w) 960 (ms. sh) 1016 (s) 479 (ms) 568 (w) 789 (m) 1058 (s,d) 11 10 (s,d) * w = weak (<20% A) (A = absorption); mw = medium weak (2&40% A); m = medium (4040% A); ms = medium strong (60430% A); s = strong (>80% A); sh = shoulder; d = doublet; t = triplet; q = quadruplet. Bands in italics are used for scaled absorbance subtraction. Table 2. Effect of data point increment and smoothing on the peak intensities of the 695 780 and 799 cm- I quartz bands Data Degree of smoothing* Doint increment l c m - I : 695 780 799 2cm-1: 695 780 799 4cm-1: 695 780 799 0 1 .000 1 .000 1 ,000 0.9508 0.9930 0.9897 0.8197 0.9183 0.9279 5 0.9836 0.9982 0.9963 0.9508 0.9930 0.9890 0.8033 0.9156 0.9301 9 0.9836 0.9982 0.9963 0.90 16 0.9860 0.9890 0.7049 0.8699 0.9007 13 0.9705 0.9930 0.9926 0.8459 0.9666 0.9787 0.6066 0.8005 0.8389 17 21 25 0.9508 0.9246 0.8852 0.9921 0.9824 0.9710 0.9926 0.9890 0.9824 0.7705 0.7082 0.6492 0.9359 0.8919 0.8506 0.9603 0.9301 0.8934 0.5180 0.4590 0.4000 Doublet unresolved Doublet unresolved * Number of points using a Savitsky - Golay quadratic - cubic smoothing function data processing routine 1126 ANALYST.SEPTEMBER 1984. VOL. I09 Table 3. Reproducibility of the analytical method for raw and smoothed (1 1 point) spectra. x = Mean mass of quartz on filter* (pg); n - 1 = sample standard deviation (pg); n - 1 (rel.) = relative standard deviation (Yo) Mass of quartz on fi I t e r/ pg 32.6 49.2 105.9 187.6 288.9 399.7 IR peak/ cm I 780 799 7x0 799 780 799 780 799 780 799 780 799 Type Raw Smoothed Raw Smoothed Raw Smoothed Raw Smoothed Raw Smoothed Raw Smoothed Raw Smoot hed Raw Smoothed Raw Smoothed Raw Smoothed Raw Smoothed Raw Smoothed \ 3 1.90 3 1.43 33.13 32.88 47.58 45.44 50.74 49.39 10 1.76 103.70 109.Y5 110.38 186.79 182.93 188.40 187.16 285.77 276.97 29 1.95 289.28 395.48 390.33 403.84 399.40 I ? - 1 1.14 1.39 1.37 0.93 I .03 0.40 1.12 0.54 - 3 55 1.23 1.99 2.12 5.94 5.19 3.71 4.35 1.04 4.31 3.80 4.36 6.22 3.09 5.69 5.69 t? - 1 (rel.).% 3.57 4.42 4.13 3.83 2.16 0.88 2.21 1.09 2.5 1 1.19 1 .81 1.92 3.18 2.83 1.97 2.32 0.36 1 .56 1.30 1.51 1.57 0.79 1.31 1.42 * Based on seven measurements obtained by rotating the sample filter through 30" increments. Table 4. Analytical range for quartz and cristobalite Lower IR peak/ limit/ Material cm-1 Quartz . . . . . . 695 27 780 6 799 5 Cristobalite .. . . 621 16 796 6 ~~~~ Upper limit/ Pg 4 1 00 1 100 920 2400 1050 Range and Precision The precision of the analytical method was assessed by scanning several sample filters five times at 30" increments using the rotatable sample holder. The spectra were converted into absorbance and the characteristic quartz peaks measured in terms of peak height. The results are shown in Table 3. The upper and lower detection limits that may be deter-mined by the method are given in Table 4 for quartz and cristobalite. Concentrations above the upper detection limits may also be determined but the calibration graphs become curved above these values. In addition certain types of foundry samples especially those which contain iron oxide dust can become opaque to infrared radiation at such levels, and shorter sampling periods are necessary to obtain meaning-ful results.Conclusion The analysis of crystalline silica by direct differential scanning infrared spectrophotometry has been examined in detail. A glass dust cloud chamber has been described that is suitable for generating clouds of a wide range of dusts in addition to crystalline silica. Particle size effects have been shown to be unimportant when sampling with a cyclone elutriator. owing to the heavily skewed mass distribution of dust on the filter. Mineral interferences have been studied in a variety o f environments and found to be most significant within the ceramic industries . S ii m p I e re - r a d i ;i t i o n effects . o bse r ve d when using conventional dispersive infrared instruments.have been eliminated by the inclusion of matched germanium cut on/blocking filters in the optical system. The authors gratefully acknowledge the invaluable work of K. J. Pickard and D. N. Davis in modifying and interfacing the infrared spectrophotometer to the computer system and Dr. P. A. Ellwood for his assistance in the analysis o f crystalline silica samples. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. References Zaidi. S . H "Experimental Pneumoconiosis.'' Johns Hopkins Press. Baltimore lY69 p. 64. Jacobsen M Rae. S . Walton. W. H. and Rogan. J . M in "Inhaled Particles 111." Volume 11.Unwins. Old Woking. 1971 p. 903 Reisner. M. T. R in "Inhaled Particles 111." Volume 11, Unwins. Old Woking 1971 p. 921. Leiteritz. H. Bauer. D . and Bruckmann. E in "Inhaled Particles. 111." Volume 11. Unwins. Old Woking. 1971. p. 729. Casswell C. Bergman I . . and Rossiter. C. E . . in "Inhaled Particles. 111," Volume 11. Unwins. Old Woking. 1971. p. 713. Schlipkoter. H. W. Hilscher. W Pott F and Beck. E. G in "Inhaled Particles 111." Volume I Unwins Old Woking 1971, Le Bouffant. L Martin. J. C and Daniel. H. i n "Proceedings of the European Communities Commission. Conference on Technical Measures of Dust Prevention and Suppression in Mines Luxembourg 11-13 October. 1971." 1973. p. 127. Robock K and Klosterkotter. W "Inhaled Particles 111," Volume 1.Unwins Old Woking. 1971. p. 453. Mayer. E . . Arch. Environ. Health. 1963. 7. 130. "Occupational Exposure to Crystalline Silica." US Department of Health Education and Welfare Public Health Service, Centre for Disease Control National Institute for Occupa-tional Safety and Health Cincinnati Publication No. 75-120. 1975. Health and Safety Executive. "Occupational Exposure Limits 1984." Guidance Note EH 40184. HM Stationery Office, London 1984. Talvitie. N. A Anal. Chem 1951 23 623. Talvitie N. A . Am. Ind. Hyg. Assoc. J . 1964 25 169. Talvitie. N. A and Hyslop F. Am. Ind. Hyg. Assoc. J. 1958, 19. 54. Bennett H . and Reed R. A. Trans. Br. Ceram. SOC 1968, 68.57. Bennett H. and Hawley. W. G "Methods of Silicate Analysis," Academic Press London and New York 1965, p.51. Garrett H . E . and Walker. A. J. Analyst 1964. 89 642. Dobreva M Ann. Occup. Hyg 1975 18 121. Foner. H. A . and Gal. I . Analysr. 1981. 106 521. Berkelhomer. L. H. "Differential Thermal Analysis of Quartz. "US Bureau of Mines. Pittsburgh. PA R.I. 3763. 1944. Weiss. B. Hosts K . . and Boettner E. A Am. Ind. Hyg. Assoc. J 1973 34 193. Craig D. K. Ind. Hyg. 1 1961 12 434. White. I. G. and Grimshaw R. W. Trans. Br. Ceram. Soc 1970 69 175. Weiss B. Boettner E. A and Stenning M. Arch. Environ. Health 1970 20 37. Warne S. St. J . J . Inst. Fuel 1970 43 240. Ruzek J. and Linhova. R . Sb. Vys. S k . Chem. Technol. Praze. Anorg. Chem. Technol. 1969 13 227. Seiller W. J. Thermal Anal 1970. 2. 251. Gordon R. L. Griffin 0.G and Nagelschmidt G "The Determination of Quartz by X-ray Diffraction," Safety in Mines Research Establishment Sheffield. Research Report No. 52 1952. Gordon R. L. and Harris. G. W. Nature (London) 1955, 175 1135. p. 379 ANALYST. SEPTEMBER 1984. VOL. 109 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60, 61. 62. 63. 64. 65. 66. Talvitie. N . A and Brewer L Am. Ind. Hvg. Assoc. J . 1962, 23. 214. Kupka F. Silrcaty. 1967. 11 56. Bradley A. A J. Sci. Instrum. 1967 44 287. Oberg M. Environ. Sci. Technol 1968. 10 795. Crosby M. T. and Hamer. P. S Ann. Occup. Hyg. 1971,14. 65. Warner P. 0 Saad. L and Jackson.J. 0 J. Air follut. Control Assoc 1972. 22 887. Knight. G. Stefanich. W and Ireland. G. Am. Ind. Hyg. Assoc. J . . 1972 33. 469. Donovan D. T Knauber. J . W. and Vonder Heiden F. H., frog. Anal. Chem. 1973 6. 61. Bumsted H. E. Am. Ind. Hyg. Assoc. J . . 1973. 34. 150. Leroux J Lennox. D. H and Kay K. Anal. Chem 1973, 25 740. Leroux. J Davey A. B. C and Paillard A. Am. Ind. Hyg. Assoc. J . . 1973. 34 409. Stefanov D. D Minerrrl. Prtropr. AcIu 1972. 21. 49. Thatcher. J . W "The Determination of Free Silica in Airborne Dust Collected on Membrane Filters," US Department of the Interior Mining Enforcement and Safety Administration, Washington DC. Information Report 1021 1975. Edmonds. J. W. Henslee. W. W. and Guerra R. E. Anal. Chem 1977 49 2196. Altree-Williams S.Anal. Chem. 1977 49 429. Altree-Williams S. Lee. J . . and Mezin N . V Ann. Occup. Hvg. 1977 20. 109. Plowman C. Analyst. 1978 103. 982. Taylor M. Analyst. 1978. 103 1009. Altree-Williams S . A n d . Chem. 1979. 51. 304. Edmonds J . W Anal. Chem 1979 51. 305. Kohyama N. Ind. Health 1980. 18 69. Tuddenham W. M and Lyon. R. J . P. Anal. Chem. 1960, 32. 1630. Soda R. Bull. Nut. Inst. Ind. Health Jpn. 1960 4 29. Gade M. and Luft K. F. Staub 1963 23. 353. Estep P. A. Kovach. J . J . and Karr. C. Anal. Chem 1968, 40. 358. Chester R. and Green. R. N . Chem. Geol. 1968 3 199. Flick K. Arbeitsschutr 1969 7 161. Larsen D. J. von Doenhoff L. J. and Crable J . V . Am. Ind. Hyg. Assoc. J . 1972 33 367. Dodgson J. and Whittaker W. Ann. Occup. Hyg 1973 16, 373.Cares J . W. Goldin A. S Lynch. J . J and Burgess. W. A., Am. Ind. Hyg. Assoc. J . . 1973 34 298. Nieto J. L. Analyst. 1978. 103 128. Gonzalez E. and Nieto J . L. Ann. Occicp. Hyg 1978 21, 217. Gillieson. A. H. and Farrel. D. M. Can. Specfrosc 1971,61, 21. Toma S. Z. and Goldberg S. A, Anal. Chem. 1972,44,431. Dodgson J . and Whittaker W. Ann. Occup. Hyg. 1973 16, 389. Harrick N. J. and Bloxsom J . "Final Report for Study Program to Obtain the Infrared Internal Reflection Spectra of Powdered Rocks," Philips Laboratories Briarcliff Manor NY, Publication No. PL-3-IRS66-0603. 1966. McCue J. "Determination of Alpha Quartz in Airborne Dust by an Attenuated Total Reflection Technique," National Coal 67. 68. 69. 70. 71. 72. 73.74. 75. 76. 77 * 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 1127 Board. Scientific Department Scottish Area Edinburgh, Report No. H-2.78. 1976. Broxterman R. Anal. Tech. Occup. Chem. 1980 120 67. Foster R. D. and Walker R. F. Analyst 1981 106 1240. Anderson P. L. Am. Ind. Hyg. Assoc. J . 1975,36 767. Henry R. L. J . Opt. SOC. A m . 1948 38 777. Gade. H. and Reisner M. fneumonoconiosis Proc. Int. Conf. 1969 3. 636. Moenke H. H. W. in Farmer V. C. Editor "The Infrared Spectra of Minerals," Mineralogical Society Monograph 4, Bartholomew Press. Dorking 1974 p. 365. Lippencott E. R. van Valkenburg A. Weir. C. E. and Bunting E. N J . Res. Natl. Bur. Stand. 1958 61 61. Lazarev A. N. "Vibrational Spectra and Structure of Sili-cates," Plenum Press New York. 1972. Flanigen E. M. Khatami H. and Szymanmski H. A. Adv. Chem. Ser. 1971 No. 101 201. Pickard K. J. Walker R. F. and West N . G. to be published. Mark D. Ann. Occup. Hyg. 1974 17 35. Savitsky A. and Golay M. J . E. Anal. Chem. 1964 36, 1627. Jones R. N. J . Am. Chem. SOC. 1952 74 2681. Duyckaerts G. Spectrochim. Acta 1955 7 25. Otvos J . W. Stone H. and Harp W. R. Spectrochim. Acta, 1957 9. 148. Duyckaerts G. Analyst 1959 84 201. Allen T. "Particle Size Measurement," Chapman and Hall, London 1981 p. 392. Rendell R. E. G. and van Sittert G. C. H. in "Inhaled Particles 111," Volume 11 Unwins Old Woking 1971 p. 801. Dodgson J . Hadden G. G Jones C. O. and Walton W. H. in "Inhaled Particles 111." Volume 11 Unwins. Old Woking 1971 p. 757. Ogden T. L. and Rickmann A. M. Ann. Occup. Hyg. 1977, 20 257. Nagelschmidt G. Gordan R. L. and Griffin 0. G. Nature (London) 1952 169,534. Gordon R. L. and Harris G. W. Nature (London) 1952, 169 539. Dempster P. B. and Ritchie P. D. Nature (London) 1952, 169 538. Clelland D. W. Cummings W. M. and Ritchie P. D. J. Appl. Chem. 1952 2 31. Harris G. W. and Revell G. S. Health Saf. Work 1981 52. Chan T. and Lippmann M. Environ. Sci. Technol. 1977,11, 377. Helma J . Staub 1971 31 22. Wagner J. and Murphy R. S. Ind. Eng. Chem. Process Des. Dev. 1971 10. 346. Marinianscky E. and Cheng S. I. Proc. Summer Computer Simulation Conf. 1973 505. Gordon R. L. and Harris A. W. Nature (London) 1955, 175 1135. Painter P. C. Snyder R. W. Youtcheff J. Given P. H., Gong H. and Suhr N. Fuel 1980,59 364. Paper A4198 Received March 14th 1984 Accepted April 3rd 198

ISSN:0003-2654    

DOI:10.1039/AN9840901117

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST SEPTEMBER 1984 VOL. 109 1129 Determination of Olivine and Serpentine in Kimberlite by X-ray Diffraction Mark Hodgson* and A. William L. Dudeney Department of Mineral Resources Engineering Imperial College South Kensington London S W7 2BP, UK Olivine and serpentine in kimberlite from the Wesselton pipe have been quantitatively determined by a modified X-ray diffraction procedure in which concentrates of the minerals were separated from the kimberlite for use as analytical standards. The olivine extracted was essentially mono-mineralic but the serpentine contained 25% olivine as finely disseminated inclusions. Equations were deduced t o interrelate measured X-ray diffraction intensities with the mineral content and particle size distributions of ground kim berlite samples and standards.Particle size distributions were characterised by means of appropriate absolute size constants. For Wesselton kimberlite containing about 35 and 48% of olivine and serpentine, respectively a relative precision of 6% was achieved. The total olivine and serpentine content (83%) agreed well with that determined by particle-point counting (88%) and thin-section microscopy (83%). Keywords X-ray diffraction; olivine determination; serpentine determination; kimberlite; analytical standards In the course of a study of the hydrothermal alteration of diamondiferous kimberlite1.2 it was required to determine the proportions of olivine and serpentine in representative labora-tory samples taken from a bulk sample of the Wesselton pipe.3 Preliminary work indicated that this material was less complex in mineral associations than many kimberlites4 but neverthe-less represented a complex starting material for fundamental test work.In particular microscopic examination of screened size fractions indicated that the major phenocrystic and matrix minerals (olivine serpentine calcite and mica) were liberated only at particle sizes less than 150 pm. Qualitative X-ray diffraction of apparently liberated olivine and serpentine grains (taken from a -150+106 pm fraction and ground to -75 pm) showed that while the olivine was essentially mono-mineralic the serpentine contained a considerable proportion of finely dispersed olivine. Four methods were considered for the simultaneous deter-mination of olivine and serpentine particle-point counting, phase area estimation in the thin section electron probe microanalysis and X-ray diffraction.The reliabilities of the two visual methods were in doubt because of the poor liberation characteristics mentioned above. Use of the elec-tron probe was too laborious and expensive for routine analyses of the type required. 1 In principle X-ray diffraction was a satisfactory method but a suitable procedure remained to be developed. X-ray diffraction methods are sometimes employed for quantitative analyses in well defined systems. For example, Leroux et al.5 gave a theoretical relationship between the intensity of a characteristic diffraction in the quartz spectrum and the proportion of the mineral in a simple binary mixture, and Mangia6 described a rapid routine procedure for the determination of anhydrous and hydrated tripolyphosphate.In more complex systems particularly those involving mix-tures of minerals the approach is normally qualitative or semi-quantitative. This is in part because of the large group of factors ,significantly affecting the precision and accuracy of X-ray measurements. Thus in addition to instrumental factors consideration must be given to the state and proper-* Present address J . Roy Gordon Research Laboratory, International Nickel Corporation Ltd Sheridan Park Mississauga, Ontario Canada. ties of the particles making up the powdered sample in terms of bulk mixing mount thickness surface orientation and “roughness ,” size distribution degree of crystallinity chem-ical homogeneity polytypism and X-ray diffraction/ absorption ratio.It is also the result of the general increase in complexity and signal to noise ratio apparent in diffractograms of multi-mineral samples and to the lack in many instances of suitably representative analytical standards. This paper describes a quantitative X-ray diffraction method for olivine and serpentine in which the emphasis is placed on the provision of analytical standards by extracting them directly from the sample to be analysed. Account is taken of impurities remaining in the serpentine extracted. The concept of the absolute size constant7 is employed to account for differences in particle size distribution in the standards and samples. Experimental Analytical Standards As the kimberlite minerals were variable in composition and not completely liberated the aim was to produce representa-tive sample concentrates of olivine and serpentine present.The sequence of steps shown in the separation scheme illustrated in Fig. 1 ensured that individual grains visibly free of microscopically observable impurity were collected in the most appropriate concentrate. Grains of other minerals or composite grains were rejected. The 500-g fraction shown in Fig. 1 was sampled from the 80-kg bulk lot’ using well established procedures.8 It was assumed that the -150+106 pm fraction obtained was mineralogically and chemically representative of the bulk. Standards extraction was carried out as shown in Fig. 1. Details of the procedures employed are described elsewhere.1 Qualitative X-ray diffractograms of the standards obtained are shown in Fig. 2. The position of the (112) line for olivine occurred at 2.46 8 indicating that the olivine was primarily forsteritic. The serpentine peaks were relatively ill defined, indicating a poorly cyrstalline phase that could not be attributed soIely to one of the three main alfotropes chryso-tile antigorite and lizardite.!O.l’ Comparison of the patterns of peaks in Fig. 2 indicated the presence of olivine in the serpentine standard 1130 ANALYST SEPTEMBER 1984 VOL. 109 -150+106 pm fraction (500 g) I SHAKING 7 TABLE l(110 g) fraction (370 a) I DENSE n'rn'A SEPARi Sp. gr :2.96 I Mica ( i o g) " FI oat" Sp. gr. :3.23 Sp. gr. :2.71 LOW-INTENSITY MAGNETIC SEPARATION Locked Sp.gr. :2.58 I particles I i- .c Magnetics: ilmenite, spinels (10 9) Non-magnetics Magnetics Noh-(20 9) magnetics I HNOB LEACH IN G I HCI LEACHING ELUTRiATlON SEPARATION * Calcite (10 a) 1 ELUTRIATION* Serpentine Mica Serpentine Mica (2 9) I (2 9) I I . J J Serpentine SEPARATION ON PAPERt T - mica I I OLIVINE STANDARD (100 g) SERPENTINE S~ANDARD (150 9) ~ i c & 9) Fig. 1. Flow sheet for standards extraction. * Separation on the basis of shape.9 t Electrostatic separation; mica particles "stick" to paper 60 50 40 30 20 10 4 Fig. 2. X-ray diffractograms of the olivine and serpentine standards showin the peaks employed at 36.8" 28 (2.46 A) and 12.0" 28 (7.36 4 respectively "20 Cu Kor Sample Preparation and X-ray Diffraction Measurement Typically the composition of the analysed mixture was 0.6000 g of ground sample 0.0050 g of molybdenum sulphide (internal standard'*) 0.2950 g of gum arabic (X-ray absorption suppressor13) and 0.1000 g of sodium fluoride (diluent or inert additive) which was employed to give a convenient and constant total mass of 1.0000 k 0.0005 g.The sample was mixed for 30 s in a Fritsch laboratory orbital mill charged with glass balls. This tumbling action was to provide adequate bulk mixing and reduce any preferential orientation of the minerals without causing a reduction in particle size. A mass of 0.800 g of the mixture was placed on a Philips automatic powder diffraction mount smoothed and compac-ted carefully with the aid of a glass slide. Measurements of X-ray diffraction intensity were made with a Philips PW 1310 X-ray generator and a PW 1050/25 goniometer using Cu Ka radiation in conjunction with a 2" slit a 2-s time constant ANALYST SEPTEMBER 1984 VOL.109 1131 1000 counts s-1 recording rate (Philips PW 1352 counter and scaler) and a 2" 28 min-1 scanning rate. The intensities of selected peaks (normally those at 2.46 A for olivine and 7.3 A for serpentine) were recorded as the ratio of the area under a peak of interest to that under the molybdenum disulphide peak at 6.03 A. It was assumed that the peak at 2.46 8 could be attributed to olivine although certain varieties of chrysotile serpentine can give a line at or near this spacing. In general this peak was sufficiently well resolved for the area to be measured with confidence.To prepare standard samples for particles size calibration convenient amounts of the standard olivine and serpentine concentrates were ground separately in ethanol in a Tema vibratory agate mill for periods varying systematically from 2 to 16 min. A Coulter Counter Model ZBl1 was used to determine the particle size distribution in each product. The results were interpreted in terms of the Rosin - Rammler function7 defined by WR = 100 exp[-(dld,)"] . . . . (1) where WR is the cumulative mass percentage coarser than particle size d (given as a diameter in metres) d is the absolute size constant (a measure of average particle size) and n is the uniformity constant. The absolute size constant was computed and employed to characterise a particular size distribution by means of a single number.This was possible as the distributions of the ground samples obtained had similar values of n (1.5 k 0.1). In all instances the correlation coefficients of data fitted to the Rosin - Rammler function exceeded 0.95. The samples were prepared for X-ray diffrac-tion measurement as above except that for olivine 0.0900 g of mineral was employed together with 0.6100 g of sodium fluoride and for serpentine 0.3000 g of mineral with 0.4000 g of sodium fluoride. These amounts retained the same over-all mass (1.0000 g) and over-all proportions of gum arabic and molybdenum disulphide. They also reflected the greater proportion of serpentine in Wesselton kimberlite and the lower intensity of diffraction caused by this mineral at the d-spacings employed.To prepare standard samples for phase proportion calibra-tion each standard (in the form of sub-samples of one of the Tema-ground samples mentioned above) was mixed in vary-ing proportions with sodium fluoride other factors being as before. Results and Discussion Fig. 3 shows the effects of (a) particle size and ( b ) mineral proportions on the relative peak intensities measured for the olivine standard. The intensity of diffraction at 2.46 A increased markedly with decrease in size in the range 20-50 pm for the 15% (by mass) olivine sample employed and reached a maximum at 10-20 pm as described by Gordon and Harris14 for quartz. A similar size dependence was assumed to exist for other proportions of olivine.Other peaks notably 5.10,2.78 and 1.75 A were similarly tested but no acceptable correlation could be found relating peak intensity to particle size, The relative intensity ( I ) increased linearly with percentage of olivine (p) at d = 42 pm [equation (2)] and was expected to increase similarly for other values of the absolute size constant. Z(p)(42) = -0.095 + 0.035~ (do = 42 pm) . . (2) The calibration graphs (a) and (b) in Fig. 3 were employed together with equation (3) to determine intensities [I(p)(dO)] relevant to various olivine proportions p and particle size distributions characterised by the absolute size constant d,. W ( ~ M W d O ) = I(p)(42)/1(15)(42) . . (3) In equation (3) 1(15)(d,) was taken from Fig. 3(a) for a 0.9 0.8 0.7 0.6 > 0.5 4- .-In c $ 0.4 .-Y ([I P) a , n Absolute size constant/pm 1.2 -0.8 -0.4 -I I I I I I 0 10 20 30 40 50 Fig.3. Graphs of relative peak intensity at 2.46 8 for olivine against (a) absolute size constant (olivine content 15% mlm) and (6) percentage olivine (absolute size constant 42 pm) Olivine ,YO 0.9 0.7 0.5 > In c .-0.3 .-Y ([I n p 1.2 a 1.0 .- CI ([I a -0.8 0.6 0.4 0.2 20 30 40 50 Absolute size constant/pm 0 20 40 60 80 Serpentine YO Fig. 4. Graphs of relative peak intensity at 7.1 A for serpentine corrected for olivine content) a ainst (a) absolute size constant [serpentine content 37.2%) and (bl percentage serpentine (absolute size constant 39 pm) selected value of do Z(p)(42) from Fig.3(b) for a selected value of p and 1(15)(42) from Fig 3(a) or (6). Examination of the X-ray diffraction trace of the serpentine standard revealed the presence of finely disseminated olivine. The proportion of olivine was determined as follows. A 50% serpentine sample was prepared employing 0.3000 g of th 1132 ANALYST SEPTEMBER 1984 VOL. 109 2.5 .- 5 2.0 v) C al c. .-Y m 1.5 al .- c -2 !?? -.- z 1.0 +-al L I-0.5 0 0.5 1 .o 1.5 Experimental relative peak intensity Fig. 5. Error curve for olivine 0 0.2 0.4 0.6 0.8 1 .o Experimental relative peak intensity Error curve for serpentine Fig. 6 . ground standard (do = 39 pm) and the measured relative peak intensity at 2.46 8 was 0.438. Using equations (1) and (3) a value of Z(p)(42) and hence of p could be calculated: 0.438 = 0.55( -0.095 + 0.035p)/0.43 giving p = 12.5%.Thus the undiluted serpentine was determined to contain approximately 25% of olivine. The average of 20 such determinations was 25.6 k 2.6% of olivine and by difference the average serpentine content was 74.4% (37.2% in the 50% sample considered). By using the methods previously described for olivine Fig. 4 was obtained which shows the effect of ( a ) particle size and ( b ) corrected mineral proportion on the relative peak intensi-ties of the serpentine standard at 7.3 A. Regression analyses produced the following equations the symbols employed being the same as for olivine but now applied to serpentine: Z(p)(39) = 0.010 + 0.016~ (do = 39 pm) . . . . ( 5 ) In order to check the validity of these relationships when olivine and serpentine were present together in assay samples, relative intensities were calculated for various known mixtures of the two standards (in the range 040% of olivine and &6O% of serpentine) and compared with corresponding intensities that had been measured directly.For example a sample containing 24% of olivine and 36% of serpentine (i. e. 0.3600 g of a ground mixture of 0.320 g of olivine and 1.280 g of serpentine do = 32 pm; with 0.2400 g of sodium fluoride) gave an experimental relative peak intensity for olivine of 1.08. The Z(37.2)(dO) = 1.454 - 0.0231do (p = 37.2%) . . . . (4) calculated value from equation (3) was 1.35. Fig. 5 shows the resulting graph of experimental versus calculated relative peak intensity the error curve for olivine.As can be seen, deviations were small at low olivine proportions and increased with olivine content the calculation giving an overestimate of about 36% at 24% of olivine. Fig. 5 was "fitted" to a suitable polynominal and set equal to Z(p)(d,) in equation (3); letting ZE represent an experimentally measured value of relative intensity the combined expression was given by equation (6). Z(15)(d0)Z(p)(42)/0.430 = 0.135 + 0.474 ZE + 0.731 IE2 (6) Substituting for Z(p)(42) from equation (2) and rearranging, equation (7) was obtained such that the proportion of olivine could be expressed in terms of relative X-ray diffraction peak intensity and particle size: p = [1.66 + 5.81ZE 4- 8,96Z~~)/Z(15)(d~)] 4- 2.73 .. (7) The data obtained for serpentine were treated in a similar manner. The graph relating experimental and calculated relative peak intensity is shown in Fig. 6. Using ZE to denote the measured peak intensity the theoretical peak intensity could be equated with the experimentally found value as follows: Z(37.2)(do)Z(p)(39)/Z(37.2)(39) = -0.022 + 0.8151 . . (8) Noting that Z(37.2)(39) = 0.56 equations (4) (5) and (8) were combined and rearranged to give equation (9) relating serpentine content to diffraction intensity and particle size. p = [(52.21 - 1.41)/(2.45 - 0.039do)] - 0.65 . . (9) In comparison with olivine the deviations evident in Fig. 6 increased steadily from low values as the serpentine propor-tion increased the calculation giving an underestimate of about 21% at 25% of serpentine (for do = 39 pm).It is clear that the X-ray diffraction characteristics of the two minerals are interdependent when in admixture. This is likely to arise mainly from differences in orientation between the two principal minerals and the internal standard but a detailed explanation is not attempted here. Provided "unknown" samples behave similarly as binary mixtures equations (7) and (9) should remain valid. Replicate samples of Wesselton kimberlite were analysed by X-ray diffraction employing equations (7) and (9) and by established methods of particulate microscopy and thin-section microscopy. 15 The latter also gave results for the proportions of other minerals present. The results (Table 1) reflected the presence of finely divided olivine in the serpentine and were closely comparable 83 88 and 83%, respectively when considering the combined percentage of olivine and serpentine.This good agreement suggested that any effect on the determination of olivine and serpentine due to the presence of calcite mica etc. was negligible. Table 1. Comparison of analyses of Wesselton kimberlite Serpentine Olivine Calcite Mica, Method O/O O/O o/o etc.,% - - X-ray diffraction . . . . 48 35 Particulate microscopy . . 68 20 9 3 Thin-section microscopy . . 57 26 11 6 As a further check a partial elemental mass balance was made (Table 2). The elemental percentages of silicon, magnesium iron calcium and aluminium were determined for the standards and kimberlite employing established methods of sodium peroxide fusion and atomic-absorption spectrometry.Taking the rock to contain 48% serpentine [(Mgl .soFeo BsAlo.34Cao.28Mo.o4)si~05(0H)41 and 35% olivine [(Mg1.62Feo,z,Mo 1oCao "7A1".01)Si0413 the elemental propor-tions of the kimberlite were then calculated and compare ANALYST SEPTEMBER 1984 VOL. 109 with the experimental values (the term M represents eknents not inchded; semi-quantitative XRF analyses indicated that nickel was likely to be the main contributor). As shown in Table 2 the agreement for the observed and calculated elemental head assays was good for magnesium and iron. Similar agreement was evident for calcium and silicon when account was taken of about 10% of calcite and 5% of micaceous silicate. 1133 Table 2.Elemental analyses of serpentine olivine and Wesselton kimberlite Kimberlite O/O Element Serpentine o/o Olivine o/o Observed Calculated Si . . . . . . 19.4 18.5 16.3 15.8 Mg . . . . . . 15.0 25.9 16.1 16.3 Fe . . . . . . 7.3 7.7 6.1 6.1 Ca . . . . . . 3.8 1.7 6.3 2.4 A1 . . . . . . 3.2 0.1 1 .o 1.5 On the basis of four replicate measurements the relative precision of X-ray diffraction analysis was +6% for both olivine and serpentine. The systematic error associated with the measurements could not be reliably estimated but was probably low because of the (deliberate ) close similarity between the standard mixtures and the kimberlite analysed. This similarity could not have been achieved with any commercially available analytical reference material.Con-fidence in the results obtained was increased by their consistent agreement with optical and mass balance measure-ments. Although the extraction of the analytical standards and the calibrations were time consuming (30 man-days being required) the method was routinely applicable. A time of 90 min was required for each analysis of serpentine and olivine, including sample preparation particle size measurement and X-ray diffraction measurement. In order to achieve accept-able results however it was essential to follow the prescribed procedures rigorously. The method described has been employed to follow trends in the olivine and serpentine contents of kimberlite samples after various hydrothermal reactions. The other products of reaction did not give significant diffractions at the d-spacings in this work.These trends will be described and discussed elsewhere. It should be noted that the method deals only with binary mixtures of olivine and serpentine in which the mutual proportions of the two minerals vary while their individual elemental proportions remain unchanged at least within the scope of the over-all relative error quoted ( 2 6 % ) . Thus, representative sprnpling by Gy’s methods yields analytical samples having the same average mineralogical and elemental distributions as in the original kimberlite even though individual grains may vary in composition. The olivine is similar compositionally to the fosterite end-member of the fosterite - fayalite series and the serpentine has an elemental composition typical of this mineral even though in a poorly crystalline form (as an alteration product of olivine’).Separa-tion of a representative kimberlite size fraction (Fig. 1) yields analytical standards retaining the same average characteristics individually as before the bulk of the grains rejected are olivine - serpentine composites calcite and muscovite mica. Any minor iron bearing phases removed are too low in proportion (ca. 5%) to affect seriously the measured XRD intensities. Application to other kimberlite bodies and olivine - serpen-tine mixtures would require further attention to the composi-tional relationships between the standards samples and minor phases. Significant biases could arise from compositional variations in the main minerals and from changes in the nature and proportions of the minor phases.In particular the position and intensity of the (112) peak (and other peaks) in the olivine spectrum vary with the magnesium to iron ratio. 16-17 The sensitivity and resolution of the analytical peak would need critical re-examination for each new material. Within these constraints the present method based on (i) taking standards directly from the samples to be analysed; (ii) establishing calibration graphs and equations to deal simul-taneously with variations in particle size and mineral propor-tions for each mineral; and (iii) correcting the calibration equations via “error” curves to take account of X-ray interdependence between the main minerals when in admix-ture provides an approach to the analysis of complex kimberlite bodies.The authors thank Nancy Judge for typing the manuscript. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. References Hodgson M. PhD Thesis University of London 1981,449 pp. Hodgson M. and Dudeney A . W. L. Clays Clay Miner., 1984 32 19. Scott B. M. personal communication. Ahrens L. H. Dawson J. B . Duncan A. K. and Erlank, A. J . Editors “The Physics and Chemistry of the Earth,” Volume 9 Pergamon Press Oxford 1975 p. 51. Leroux J. Lennox D . M. and Kay K. Anal. Chem. 1953, 25 740. Mangia A . Analyst 1983 108 540. Orr C. Jr. “Particulate Technology,” Macmillan New York, 1966 p. 17. Gy P. H in “Sampling in the Mineral and Metallurgical Processing Industries,” Processing Symposium Institution of Mining and Metallurgy London 1973 p. 16. Beaven C. H. J . and Napier-Munn. T. J . unpublished results, 1978. Brindley G . W. and Brown G. Editors “Crystal Structures of Clay Minerals and Their X-ray Identification,” Minerolog-ical Society London 1961. Berry L. G. Editor “Selected Powder Diffraction Data For Minerals,” Joint Committee on Powder Diffraction Standards, Swarthmore PA 1974. Cody R. D. and Thompson G. L. Clays Clay Miner. 1976, 11 224. van der Marel H. R. Contrib. Mineral. Petrol. 1966 12 96. Gordon R. L. and Harris G. W. Nature (London) 1955, 175 1135. Zussman J . Editor “Physical Methods in Determinative Mineralogy,” Academic Press London 1967 p. 31. Yoder H . S and Sahama Th. G. Am. Mineral. 1957,42,475. Jahanbagloo I. C . Am. Mineral. 1969 54 246. Paper A31292 Received August 25th 1983 Accepted March 23rd 198

ISSN:0003-2654    

DOI:10.1039/AN9840901129

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1984, VOL. 109 1135 he,,= 575 nm Spectrofluorimetric Determination of Aluminium(lll) in Portland Cement and Aluminium Bronze with 1 -Hydroxy=2=carboxyanthraquinone F. Salinas, A. Muiioz de la Peiia and J. A. Murillo Department of Analytical Chemistry, Faculty of Sciences, University of Extremadura, Badajoz, Spain kex. = 465 nm The fluorescent chelate of aluminium with 1 -hydroxy-2-carboxyanthraquinone is used to provide a sensitive determination of aluminium. The detection limit is 2 ng ml-1 and the range of application is between 3 and 250 ng ml-1. The stoicheiometry of the complex is 2: 1 (reagent to metal). The method has been applied satisfactorily to the determination of aluminium in Portland cement and aluminium bronze. Keywords: Aluminium determination; spectrofluorimetry; 1-hydroxy-2-carboxyanthraquinone; Portland cement; aluminium bronze The fluorescent reactions of 1-hydroxy-2-carboxyanthra- quinone have been studied and fluorimetric methods for the determination of Be(II),* Mg(II)2 and Y(III)3 have been proposed in previous papers.Of anthraquinone derivatives, only 1,2-dihydroxy- anthraquinone-3-sulphonate (Alizarin Red S),4 1,4-dihy- droxyanthraquinone-2-sulphonate (2-quinizarin su1phonate)s and 1,2,4-trihydroxyanthraquinone-3-sulphonate (purpurin su1phonate)b have been proposed for the fluorimetric determi- nation of aluminium. The ranges of application are between 80 and 760, 11 and 54 and 20 and 100 ng ml-1, respectively. This paper is concerned with the fluorescent characteristics of the complex formed between 1-hydroxy-2-carboxy- anthraquinone and Al(II1) and a spectrofluorimetric method for the determination of this ion.Experimental Reagents All experiments were performed with analytical-reagent grade chemicals and pure solvents. Doubly distilled and demineral- ised water was used throughout. Aluminium ion stock solution, 1 g 1-1. Prepared from A12(S0&. 18H20 (Merck) and standardised gravimetrically as the oxide.7 1-Hydroxy-2-carboxyanthraquinone. Prepared by diazo- tisation and further hydrolysis of 1-amino-2-carboxyan- thraquinone according to Scholl’s method,8 and purified by recrystallisation in glacial acetic acid. A reagent solution (10-3 M) was prepared by exact weighing of l-hydroxy-2- carboxyanthraquinone in ethanol. Other solutions, obtained by exact dilution, were also prepared.Buffer solution, p H 3.5. Prepared from chloroacetic acid and sodium chloroacetate of concentration 0.1 M. Apparatus All fluorimetric measurements were performed on a Perkin- Elmer MPF-43 fluorescence spectrophotometer , equipped with an Osram BO 150-W xenon lamp, excitation and emission grating monochromators, 1.0 x 1.0 cm quartz cells, an R-508 photomultiplier and a Perkin-Elmer 056 recorder. M solution of Rhodamine B) gave a scale reading of 60 units, with the following parameters: sensitivity, 1 coarse and 7 fine; slits, 4 nm for excitation and 3 nm for emission; temperature, 20 “C; wavelength of excitation, 480 nm; and wavelength of emission, 580 nm. The fluorescence spectra are given without spectral correction.A standard fluorescent stick (equivalent to a 3 x A Crison 74 pH meter with a glass - saturated calomel combination electrode was also used. Procedure for the Determination of Aluminium(II1) Transfer a suitable aliquot containing 0.075-0.625 pg of aluminium into a 25-ml calibrated flask. Add enough water t o ensure a final water content of 40% V/V, 1 ml of chloracetic acid - sodium chloroacetate buffer solution of pH 3.5 and 4 ml of 2 x 10-4 M ethanolic reagent solution and dilute the mixture to the mark with ethanol. Measure the fluorescence intensity at 575 nm, with excitation at 465 nm (sensitivity, 10 coarse and 5 fine; slits, 12 nm for excitation and 12 nm for emission) against a reagent blank prepared in a similar way but without aluminium. A similar procedure can be employed for the determination of 0.625-6.25 pg of aluminium (sensitivity, 10 coarse and 4 fine; slits, 8 nm for excitation and 8 nm for emission).Results and Discussion Spectral Characteristics and Effects of Experimental Variables The excitation and emission fluorescence spectra of the complex with Al(II1) and of the reagent are shown in Fig. 1. The wavelengths chosen were 465 and 575 nm for excitation 200 400 500 700 Unm Fig. 1. Fluorescence excitation and emission spectra of the complex 1-hydroxy-2-carboxyanthra uinone - AI(II1) (1,2) and of the reagent 1”. 2’). Apparent pH = 4 3 ; 60% ethanol; [AI(III)] = 100 ng m1-l; reagent] = 8 x 10-5 M1136 ANALYST, SEPTEMBER 1984, VOL. 109 and emission, respectively, where the differences between the emission of the complex and the reagent are maximum.Studies on the effect of acidity on the fluorescence intensity showed that this variable considerably affects the complex formation. As shown in Fig. 2, for 70% V/V ethanol in the medium, the fluorescence intensity remained maximum and unchanged in the apparent pH range between 4.3 and 5.2. Study of the effect of the water content in the medium (Fig. 3) indicated that the complex showed maximum emission with 60% VWof ethanol. Under these conditions, the optimum pH range is reduced to 4.6-4.9. The pH can be suitably adjusted by addition of 1 ml of the buffer solution of pH 3.5. Under these conditions, the complex formation is instantaneous and the complex remains stable for at least 6 h. The fluorescence intensity is not affected by changes in the order of addition of reagents.The dependence of the fluorescence intensity on temperature is critical. A tempera- ture increase from 10 to 60 "C diminishes the fluorescence intensity by 1.4% "C-1. This effect can be explained by the higher internal conversion as the temperature increases, facilitating non-radiative deactivation of the excited singlet state. It is therefore recommended that a thermostat is used, choosing a measurement temperature of 20 "C, i.e., about room temperature. The effect of the reagent concentration on the fluorescence intensity of solutions containing 25 ng ml-1 of Al(II1) was studied under conditions similar to those of the recommended method. The fluorescence intensity increased with increasing reagent concentration up to 3 x 10-5 M , remained constant between 3 x 10-5 and 6 x 10-5 M and decreased again above 6 X 10-5 M. This effect was attributed to fluorescence inversion phenomena.In accordance with this, a 3.2 x 10-5 M reagent concentration was chosen as the optimum (4 ml of 2 x 10-4 M reagent solution in a final volume of 25 ml). Composition of the Complex The stoicheiometry of the complex was studied under the established conditions by the classical methods of Job (Fig. 4) and Yoe and Jones. From these studies the composition of the complex was concluded to be 2 : 1 (ligand to metal). According to the structure of the ligand and taking into account the 2 3 4 5 6 7 PH Fig. 2. Graph of relative fluorescence intensity versus pH for the l-hydroxy-2-carboxyanthraquinone - Al(II1) complex.70% ethanol; [AI(III)] = 100 ng mlkl; [reagent] = 4 x 10-5 M 8 10; 10 30 50 70 H20, '10 Fig. 3. Effect of solvent composition on the fluorescence intensity of the l-hydroxy-2-carboxyanthraquinone - AI(II1) complex. [AI(III)] = 100 ng ml-1; [reagent] = 4 x 10-5 M proposed structures for complexes of aluminium with l-hydroxy derivatives of anthraquinone,"l4 we propose struc- ture I for the complex. ,fAY0 0 I Analytical Parameters Under the operating conditions outlined in the recommended procedure, there is a satisfactory linear relationship between fluorescence intensity and Al(II1) concentration for two ranges of concentration covering a total interval of 3-250 ng ml-1 (Fig. 5). For three series of ten measurements on 10, 20 and 200 ng ml-1 of Al(III), relative errors of 1.93, 1.09 and 0.72% and relative standard deviations of 0.85, 1.07 and 0.74%, respec- tively, were obtained (95% confidence level).The detection limit is 2 ng ml-1 when defined as the analyte concentration leading to a luminescence intensity that is three times the blank standard deviation.15 I , t I I 0.1 0.3 0.5 0.7 0.9 [Al(lll)]/[Al~lll~] + IRI Fig. 4. [AI(III)] = 2 x 10-5 M Stoicheiometry of the complex (Job's method). [Reagent] + B 90 - A >. cn 0, 4- .- 70 - .- 0 5 10 15 20 25 3 A 0 50 100 150 200 250 300 B [Al(lll)l/ng rnl-' Fig. 5. Calibration graphs. A, [Reagent] = 3.2 x 10-5 M ; sensitivity, 10 coarse, 5 fine; slits, excitation 12 and emission 12 nm. B. [Reagent] = 3.2 X 10-5 M; sensitivity, 10 coarse, 4 fine; slits, excitation 8 and emission 8 nmANALYST, SEPTEMBER 1984, VOL.109 1137 Table 1. Tolerance limits for various ions in the determination of aluminium at the 10 ng ml-* level Foreign ion Maximum tolerance: foreign ion to AI(II1) ratio (mlm) Li+, Na+, K+, Ag+, Sr2+, Ba2+, Co2+, Cu2+, Ni2+, Pb2+, Cd2+, Hg2+, NH4+, C1-, Br-, N03- 9 BO2- 3 SO42- . . . . . . . . . . 100 Ca2+, Mn*+ , Bi3+, Po43- . . . . . . . . 50 Zn2+, Ga3+ . . 10 F-, tartrate, citrate . . . . . . . . . . 1 EDTA, c2042- . . . . . . . . . . . . 0.5 Be2+ . . . . . . . . . . . . . . . . ~ 0 . 5 Mg2+,Sn2+,Y3+,Fe3+,~r3+;La3+‘ * “ . . . . ’ . 5 ~ ~~ ~ Table 2. Spectrofluorimetric determination of aluminium in Portland cement and aluminium bronze by the use of l-hydroxy-2- carboxyanthraquinone AI(II1) AI(II1) Standard Certified certified, found, deviation, Standard sample composition, YO Portland cement, BCS No.372 . . SO2, 21 -3; Ti02, 0.33; Fe203, 2.49; Mn20,, 0.06; CaO, 65.8; MgO. 1.3; Na20, 0.21; K,O, 0.62; SO?, 2.35; P2O5, 0.19 Aluminium bronze, BAS No. 32a . . Cu, 85.9; Zn, 0.94; Fe, 2.67; Mn, 0.27; Ni, 1.16 YO o * Yo 10 5.351- 5.50t 0.08 8.8 8.93 0.09 * Each value is the average of six separate determinations. t AS A1703. Effect of Foreign Ions The effect of various ions on the determination of 10 ng ml-l of Al(II1) was studied by first testing a 100 : 1 (m/m) ratio of interferent ion to aluminium and, if interference occurred, reducing the ratio progressively until interference ceased. Higher ratios were not tested. The criterion for interference was a variation of fluorescence intensity of more than 2% from the value expected for aluminium alone.The results obtained are shown in Table 1. Applications The recommended procedure has been applied satisfactorily to the determination of aluminium in two standard samples, a Portland cement and an aluminium bronze, without previous separation. The samples were dissolved according to the following procedures. For Portland cement, take about 0.25 g of the sample and 0.5 g of ammonium chloride and add 5 ml of hydrochloric acid. Heat in a steam-bath for 30 min, then add 20 ml of hot water to the solution for complete dissolution of the soluble components of the cement sample. Filter the solution and wash the precipitate with hot water. Cool the filtrate and dilute the solution to 250 ml in a calibrated flask.For aluminium bronze, take about 0.25 g of the sample, add 5 ml of nitric acid and heat until evolution of nitrogen oxides ceases. Cool and dilute the solution to 250 ml in a calibrated flask. Aliquots of the above solutions were taken and aluminium was determined as described. The results are given in Table 2. Conclusions The proposed ligand, l-hydroxy-2-carboxyanthraquinone, compares advantageously with other anthraquinone deriva- tives proposed for the spectrofluorimetric determination of aluminium. The methods using 2-quinizarin sulphonate5 and purpurin sulphonateb recommend making the fluorescence measure- ments 1 h after mixing the reagents in order to develop the complexes completely. Interferences were not reported.The complex with Alizarin Red S4 is stable for up to 80 min, and 11 of the 31 ions tested interfered in the method. None of the methods has been applied to real samples. The method proposed in this paper, based on the fluor- escent complex formed between 1-hydroxy-2-car- boxyanthraquinone and Al(III), is more sensitive and shows a wider range of application than the above-mentioned methods. The complex formation is instantaneous and the complex remains stable for at least 6 h. The method is relatively free from interferences and has been applied satisfactorily to the determination of aluminium in Portland cement and aluminium bronze. 1. 2. 3. 4. 5 . 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. References Capitan, F., Salinas, F., and Franquelo, L. M., Anal. Lett., 1975,8, 753. Capitin, F., Salinas, F., and Franquelo, L. M., Quim. Anal., 1977, 31, 275. Salinas, F., Munoz de la Pena, A., and Murillo, J. A., Anal. Len., 1984, in the press. Arhmedli, M. K., Efendiev, D. A., and Ruvinova, F. I . , Uch. Zap. Aterb. Gos. Univ., Ser. Khim. Nauk, 1973, No. 4, 10; Anal. Abstr., 1975, 29, 4B65. Capitin, F., Roman, M., and Guiraum, A., An. Quim., 1974, 70, 507. Capitan, F., Roman, M . , and Alvarez-Manzaneda, E., Afini- dad, 1973, 50, 623. Erdey, L., “Gravimetric Analysis, Part 2,” Pergamon Press, Oxford, 1965. Scholl, R., Monasfh. Chem., 1913, 34, 1023. Beech, W. F., and Drew, H. D. K., J . Chem. SOC., 1940,603. Flagg, J . F., Liebhafsky, H. A., and Winslow, E. H., J . Am. Chem. Soc., 1949, 71, 3630. Parker, C. A., and Goddard, A. P., Anal. Chim. Acta, 1950,4, 517. Dorta-Schaeppi, Y., Hurzeler, H . , and Treadwell, W. D., Helv. Chim. Acta, 1951, 34, 797. Larsen, E. M., and Hirozawa, S. T., J. Inorg. Nucl. Chem., 1956,6, 198. Biryuk, E. A., Nazarenko, V. A., and Ravitskaya, R. V . , Zh. Anal. Khim., 1968, 23, 1795. IUPAC Nomenclature, Symbols, Units and Their Usage in Spectrochemical Analysis, Pure Appl. Chem., 1976, 45, 105. Paper A31457 Received December 30th, 1983 Accepted March 29th, 1984

ISSN:0003-2654    

DOI:10.1039/AN9840901135

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST SEPTEMBER 1984 VOL. 109 1139 Synthesis and Chromogenic Properties of Water-soluble Hydrazones, Spectrophotometric and Analogue Derivative Spectrophotometric Determination of Trace Amounts of Iron with 2-Pyridyl-3’-sulphophenylmethanone 2-Pyrimidylhydrazone (PSPmH) and Kinetic Studies of the Complex Formation of Iron(ll) with PSPmH Takeshi Aita Tsugikatsu Odashima and Hajime lshii Chemical Research Institute of Non-Aqueous Solutions Tohoku University Katahira Sendai 980 Japan Three water-soluble hydrazones have been synthesised and their chromogenic properties and reactivities with metal ions investigated. One of them 2-pyridyl-3’-sulphophenylmethanone 2-pyrimidylhydrazone (PSPmH) has been utilised for the spectrophotometric determination of iron. Iron(ll) reacts with PSPmH to form a stable 1 2 (metal to ligand) complex having absorption maxima at 379 and 580 nm in the pH range 7.3-10.0.The apparent molar absorptivities of the complex are 4.79 x 104 and 1.18 x lo4 I mol-1 cm-1 at 379 and 580 nm respectively. A selective spectrophotometric method is proposed for the determination of iron at parts per million levels and has been applied successfully to the analysis of water samples. The sensitivity of the method can be increased significantly by employing analogue derivative spectrophotometry iron determination at parts per billion levels being feasible. Further the formation reaction of the iron(l1) - PSPmH complex has been investigated by a stopped-flow spectrophotometric technique. The rate for the complex formation is proportional to each of the concentrations of iron(ll) and PSPmH and independent of the hydrogen ion concentration.On this basis the 1 1 complex formation between Fez+ and undissociated PSPmH is assumed to be the rate-determining step. The rate constant for the complex formation at an ionic strength of 0.2 and at 25 “C is 1.3 x 105 I mol-1 s-1. The activation parameters for the complex formation are as follows € = 44.3 kJ mol-1 AW = 41.8 kJ mol-1 A 9 = -6.48 J K-1 mol-1 and AG* = 43.9 kJ mol-1. Keywords Iron de te rm ina tion; 2-p yrid yl-3 ’-sulp h op hen ylm e than on e 2 -p yrim id yl h ydrazon e; spectro-photometry; analogue derivative spectrophotometry; kinetics of iron complex formation Various hydrazones have been used as highly sensitive and/or selective spectrophotometric or fluorimetric reagents for metal ions.However most of them and their metal complexes are insoluble in water so that the determination of metals with hydrazones has usually been performed in an aqueous ethanol medium or in the presence of surfactants or accompanied by solvent extraction. About 10 years ago Going and co-workers synthesised a water-soluble hydrazone 2-pyridyl-3’-sulphophenylmethanone 2-pyridylhydrazone (PSPH) investi-gated its complex formation with several metal ions spectro-photometrically and potentiometrically 1 and applied it to the pre-concentration of trace amounts of metal ions by a combined complexation - anion exchange technique .2 Since then however no paper concerning PSPH and its analogues has been reported. In the work presented here three water-soluble hydrazones derived from 2-(3’-sulphobenzoy1)pyridine including PSPH, 2-pyridyl-3’-sulphophenylmethanone 2-benzothiazolylhyd-razone (PSBH) and 2-pyridyl-3’-sulphophenylmethanone 2-pyrimidylhydrazone (PSPmH) were synthesised and their chromogenic properties and reactivities with various metal ions were investigated and compared.The results for PSPmH seemed to be superior to the others as a chromogenic reagent for iron(II) and its complex formation with iron(I1) has been studied both spectrophotometrically and analogue derivative spectrophotometrically for the determination of trace amounts of iron. Further the complex formation has been investigated kinetically by a stopped-flow spectrophotometric technique. Experimental Synthesis of Hydrazones Equimolar amounts (0.01 mol) of 2-(3’-sulphobenzoy1)-pyridine prepared by sulphonation of 2-benzoylpyridine by the procedure of Bradsher et al.,3 and the corresponding hydrazine were dissolved in 80 ml of 50% aqueous ethanol and heated under reflux for 3 h.After cooling to room temperature the precipitated crude compound was filtered off and recrystallised from ethanol - water. Reagents All reagents used were of analytical-reagent grade unless stated otherwise. All solutions were prepared with distilled, de-ionised water. PSBH PSPH and PSPmH solutions 1 x 10-2 M. Prepared by dissolving the required mass of each hydrazone in 0.01 M sodium hydroxide solution. These solutions were further diluted with water if necessary.Zron(1rr) stundurd solution. Prepared with iron( 111) am-monium sulphate as described earlier.4 Iron(I1) solutions were prepared by reducing the iron(II1) solution with sodium ascorbate. Buffer solutions. 1 M acetic acid - 1 M sodium acetate d 1 5 potassium dihydrogen phosphate - ~ / 1 5 disodium hydrogen phosphate 0.2 M boric acid + 0.05 M sodium chloride - 0.05 M sodium borate 1 M aqueous ammonia solution - 1 M am-monium chloride and 0.05 M sodium carbonate - 0.1 M sodium hydrogen carbonate systems were used according to the pH values required. Apparatus For measurements of the absorbance and the absorption spectrum a Hitachi 139 and a Hitachi 200-10 spectropho-tometer respectively were used. To obtain the derivative spectrum a modified Hitachi 200-0576 derivative unit com-posed of two analogue differentiation circuits (each having six different time constants) was connected in series between a Hitachi 556 dual-wavelength spectrophotometer’s output an 1140 ANALYST SEPTEMBER 1984 VOL.109 Table 1. Structures and physical properties of hydrazones Elemental analysis.*o/o Hydrazone R M.p./"C vc=N/cm I C H N S 54.2 3.1 13.2 15.0 57.6 4.0 15.1 8.8 . . . . . . 1580 (55.59) (3.44) (13.65) (15.62) 1610 (57.61) (3.99) (15.81) (9.05) PSBH 2-Benzothiazolyl 280t PSPH 2-Pyridyl 2501 . . . . . . PSPrnH . . . . 53.7 3.9 19.2 8.7 1580 (54.08) (3.69) (19.71) (9.02) 2-Pyrimidyl 235: * Results in parentheses indicate calculated values. t Decomposed. Table 2. Acid dissociation constants of hydrazones. Ionic strength = 0.2; and temperature = 25 f 0.1 "C Acid dissociation constant the reagent at 580 nm and at 25 & 0.1 "C using the stopped-flow spectrophotometer.The ionic strength was maintained at 0.2 with sodium perchlorate throughout the measurements. Hydrazone Pka Pkd Pkd, PSBH . . . . . . . . 10.34 4.34 1.35 PSPH . . . . . . . . 14.93" 5.88 3.41 PSPmH . . . . . . . . 14.68* 4.91 1.67 * Ionic strength = 3.0 calculated by simplex method with a personal computer. a Hitachi 057 X - Y recorder's input the former being used as an ordinary double-beam spectrophotometer. The details of this apparatus and the principle and characteristics of analogue-derivative spectrophotometry have already been described."h For kinetic measurements a Union Giken RA-401 stopped-flow spectrophotometer was used.Cells of 10 mm optical path length were used throughout all the measurements unless stated otherwise. Procedure Ordinary spectrophotometry Place a sample or standard solution containing less than 30 vg (or 120 pg) of iron in a 25-ml calibrated flask and add 1 ml of 1% sodium ascorbate solution 1 ml (or 3 ml) of 2.5 X 10-3 M PSPmH solution and 3 ml of 0.05 M borate buffer solution (pH 8) and dilute to the mark with water. Measure the absorbance of the resultant solution at 379 (or 580) nm against a reagent blank. Second-derivative spectrophotornetry When the iron content of the coloured solution prepared by the procedure described above is too low to give a measurable absorbance record the second-derivative spectrum from 650 to 450 nm against a reagent blank by using a combination of both first- and second-order differentiation circuits of No.6 and a scan speed of 300 nm min-1 and measure the second-derivative value (the vertical distance from a peak to a trough or that from the base line to a trough of the peak). Kinetic measurements Reaction curves for the complex formation of iron(I1) with PSPmH were measured under a pseudo-first-order excess of Results and Discussion Properties and Characteristics of the Hydrazones The infrared spectra of the svnthesised PSBH PSPH and PSPmH were measured with potassium bromide discs in order to confirm their structures. The spectra had absorption peaks assigned to the stretching vibration of an azomethine bond (-N=C<) around 1600 cm-1.7 On the basis of these results, and those of the elemental analysis shown in Table 1 the synthesised- hydrazones are presumed to have the structures shown in Table 1.PSBH PSPH and PSPmH are appreciably soluble in water and very soluble in alkaline solution. The acid dissociation constants ( k O l k, and ka3) were determined spectrophotometrically at an ionic strength of 0.2 and at 25 2 0.1 "C. The results are given in Table 2. PSBH, PSPH and PSPmH exist in solution in any of the following forms depending on the pH: where L denotes the undissociable parts of the hydrazones. Usually because the sulphonic acid group is dissociated completely in solution its acid dissociation constant could not be determined. The k, and k may be caused by the protonation of the nitrogen atom 08 the hydrazine part and the pyridine nitrogen atom respectively whilst k, is due to the deprotonation of the imino group.Reactivity of the Hydrazones The reactivities of PSBH PSPH and PSPmH with various metal ions at pH 4 7 and 10 are summarised in Tables 3 4 and 5 respectively. Every hydrazone is found to react with cobalt(II) copper(II) iron(II) nickel(I1) and palladium(I1) to give coloured complexes having large molar absorptivities. Iron(I1) complexes have two absorption maxima one of which has a relatively large molar absorptivity whilst another is separated from absorption maxima of other metal complexes. This suggests the usefulness of the synthesised hydrazones for the sensitive or selective determination of iron so that the iron(I1) complexes were examined in subsequent studies ANALYST SEPTEMBER 1984 VOL.109 1141 ~~~ Table 3. Reactivity of PSBH with different metal ions PH 4 PH 7 pH 10 Metal ion h,,,,/nrn Ag(1) . . . . -Cd(I1) . . . . -Cr(II1) . . . . -Cr(V1). . . . -Co(1I) . . . . 490 Cu(I1) . . . . 483 Fe(I1) . . . . 585 Fe(II1) . . . . 500 Hg(I1) . . . . -Mn(I1) . . . . -Pb(I1) . . . . -Ti(1V) . . . . -V(IV) . . . . 495 Zn(I1) . . . . -Ni(I1) . . . . 465 Pd(I1) . . . . 535 V(V) . . . . 498 * E = molar absorptivity. &*/I mol-1 cm-1 --24 200 --10 400 9 400 16300 --3 100 8 700 9 800 12 700 ---Lax./nm 403 430 483 --438 409 600 429 427 433 440 448 535 483 429 --&/Imol-lcm-l 14 700 52 800 28 700 --25 OOO 35 100 13 200 21 900 43 100 11 700 41 800 18 900 11 300 5 800 51 400 --hlll,X./nm 415 44 1 483 --455 400 600 456 432 442 447 460 530 ---435 &/I mol-1 cm-1 7 500 32 500 26 400 --20 800 15 OOO 12 700 9 500 34 OOO 35 600 31 500 12 100 12 200 ---38 100 Table 4.Reactivity of PSPH with different metals ions PH 4 Metal ion h,,,./nm &/I mol-1 cm-Ag(1) . . . . - -Cd(I1) . . . . -Cr(II1) . . . . - -Cr(V1). . . . - -Cu(I1) . . . . - --Co(I1) . . . . 480 31 300 Fe(1I) . . . . 541 13 200 Fe(II1) . . . 490 15 300 Hg(I1) . . . . - -Mn(I1). . . . - -Ni(I1) . . . . -Ti(1V) . . . . - -V(1V) . . . . V(V) * .. . Zn(I1) . . . . - --- - Pb(I1) . . . . Pd(I1) . . . . 440 23 200 - -- -PH 7 pH 10 &/I mol ~ I cm- 1 --31 300 --16 700 20 500 9 200 2200. --5 200 18 500 7 800 ---Y? -L a x ./nm -443 480 --468 385 59 1 ---442 462 485 ---440 &/1 mol-1 cm-1 -13 500 30 600 --18 700 40 900 10 400 ---38 300 2 700 12700 ---35 300 Table 5. Reactivity of PSPmH with different metal ions PH 4 PH 7 pH 10 Metal ion h,,,./nm Ag(1) . . . . -Cd(I1) . . . . -Cr(TI1) . . . . -Cr(V1). . . . -Cu(I1) . . . . 453 Co(I1) . . . . 452 Fe(I1) . . . . 525 Fe(II1) . . . . 416 Hg(I1) . . . . -Mn(I1) . . . . -Ni(I1) . . . . -Pb(I1) . . . . -Ti(1V) . . .. -Zn(I1) . . . . -Pd(I1) . . . . 490 V(1V) . . . . 470 V(V) . . . . 465 dl mol-1 cm-1 --26 800 --4 700 11 200 19 300 ----11 400 5 000 7 900 --Lax./nm -417 452 --435 379 578 --425 418 490 455 416 ---&/I mol-I cm-1 9 400 26 900 ---16 100 46 100 11 200 --2000 35 800 12 000 3 500 26 600 ---hmax./nm -417 452 --42 1 379 580 420 425 418 435 490 ----416 d l mof-1 cm-l 36 500 26 900 ---25 700 47 900 11 800 34 700 36 200 26 800 11 500 15 300 ----42 60 1142 0.6 a 0 m n 6 0.4 2 0.2 ANALYST SEPTEMBER 1984 VOL. 109 ---Spectrophotometric Studies of the Iron(I1) Complexes Absorption spectra Absorption spectra of the iron(I1) complexes with PSBH, PSPH and PSPmH at each optimum pH (see the following paragraph) are shown in Fig.1. Every complex has two absorption maxima at 409 and 600 nm for the PSBH complex, at 385 and 591 nm for the PSPH complex and at 379 and 580 nm for the PSPmH complex. However in lower pH regions the absorption maximum of each complex at longer wavelengths was blue-shifted by 20-40 nm on lowering the pH whilst that at shorter wavelengths did not change, although the absorbance decreased gradually. This may be caused by the formation of protonated complexes. Effect of p H Fig. 2 shows the effect of pH on the formation of the iron(I1) complexes measured at each absorption maximum at shorter wavelengths from which it was found that constant absor-bances are obtained in the pH ranges 7.3-10.0 6.5-8.5 and 9.0-11.4 for PSPmH PSBH and PSPH complexes respec-tively.Measurements at each absorption maximum at longer wavelengths gave almost the same results. Composition of the Complexes The composition of the complexes was determined by the continuous variations method and the molar ratio method. The results revealed the formation of a 1 2 (metal to ligand) complex for the PSPmH complex and 1 3 complexes for the PSBH and PSPH complexes. Taking into account the sensitivity selectivity and optimum pH range on the basis of the results described above PSPmH was thought to be the best reagent of the three for the determination of iron. Therefore other conditions for its use were investigated further.Spectrophotometric Studies for the Determination of Iron with PSPmH Effect of PSPmH concentration and stability of the complex A 3-fold molar excess of PSPmH was the minimum required for obtaining a constant absorbance. and an excess of PSPrnH did not interfere. The complex formed under the recommended conditions was very stable so that the absorbance remained constant even after 5 h. Calibration graph sensitiviry and precision The calibration graphs obtained by the recommended pro-cedure were linear over the ranges 1-30 and 4-1 20 pg of iron at 379 and 580 nm respectively. The equations of the lines obtained by a least-squares treatment were Fe (p.p.m.) = 1.17A (at379nm) . . (1) Fe (p.p.m.) = 4.73A (at580nm) . . (2) where A is the absorbance.The sensitivities for an absorbance of 0.001 and the molar absorptivities calculated from equa-tions (1) and (2) were l . 17 ng cm-2 and 4.79 x lo4 l mol- 1 cm-l at 3.79 nm and 4.73 ng cm- and 1.18 x 104 1 mol-1 cm-1 at 580 nm respectively. Thirteen standard solutions containing 20.6 pg of iron were analysed by the recommended procedure. The results gave relative standard deviations of 0.42 and 0.36% at 379 and 580 nm respectively. Effect of diverse ions Solutions containing 10.5 pg of iron and various amounts of other ions were prepared and the recommended procedure was followed. An error of k3% in the absorbance reading was 0.8 0.6 z ([I n $ 0.4 n a In 0.2 0 400 500 600 Wavelengthhm Fig. 1. Absorption spectra of hydrazones and their iron(I1) complex solutions.Fe(II) 1.5 x 10-5 M; hydrazones 2.0 X M . A B C. PSPmH PSPH and PSBH complexes respectively; D. E F PSPmH, PSPH and PSBH alone respectively. pH A D 8.0; B E 9.8; C F, 7.6. Reference A B C reagent blank; D E F water I I I I 1 O L d 6 8 10 12 PH Fig. 2. Effect of pH. Fe(II) 1 .5 X lop5 M ; hydrazones. 2.0 x 10-4 M ; reterence. reagent blank. A PSBH complex at 409 nm; B. PSPmH complex at 379 nm; C PSPH complex at 385 nm considered tolerable. The results are summarised in Table 6, from which it can be seen that measurements at 580 nm are more selective although less sensitive than those at 379 nm and in the presence of masking agents the method in which the absorbance is measured at 580 nm is selective for iron(I1) although the tolerance limits for cobalt(I1) and nickel(I1) are low.Anions such as fluoride chloride bromide iodide nitrate, perchlorate thiocyanate sulphate tartrate and citrate did not interfere (or if so scarcely) even when they were added up to 0.04 M as the final concentration. Practical applications The proposed method was applied to the determination of iron in water samples. The results are shown in Table 7 together with those obtained by atomic-absorption spectro-photometry carried out for comparison. Both sets of results are in good agreement. Sensitisation by Employing Analogue Derivative Spectro-photometry Derivative spectrophotometry using the analogue differentia-tion circuit is extremely effective for enhancing the sensitivity of ordinary spectrophotometry.5.6 As an example of sensitisa-tion the second-derivative spectrophotometric determination of iron with PSPmH is described here.Selection of conditions for measurements As the second-derivative value (the vertical distance from a peak to a trough or that from the base line to a trough of the spectrum) depends on both the time constant of the analogu ANALYST SEPTEMBER 1984 VOL. 109 1143 Table 6. Tolerance limits for other metal ions. Amount of iron(I1) taken = 10.5 pg Metal ion 379 nm Tolerance Iimitlpg No masking agent added Masking agent added lo00 CNII) Mg(I1) 100 Ag(I) AI(III) Cr(III) Pb(II),* Cd(II),* Zn(II),* Cr(VI) Pd(II),t Ti(1V) V(IV),S V(V)$ Below 10 Cd(II) Co(II) Cu(I1) Co(II),* Ni(II)* Mn(II) Ni(II) Zn(I1) * 1 ml of 1% sodium thioglycollate solution was added.t Precipitated. $. 1 ml of 1% sodium citrate solution was added: 580 nm No masking agent added Ag(I) A W ) Cd(II), Pd(II),t Ti(IV) Zn(I1) Co(II) c u (11) Hg(II), Ni(II) Pb(II) V(IV), Masking agent added Ca(II) Mg(W Cr(III) Cr(VI) Mn(II) V(IV)?$ V(V)+ Cu(I1) ,* Pb(I1) ,* Hg(II)* Co(I1) ,* Ni( II)* V(V) Table 7. Determination of iron in water samples Iron content p.p.m. Proposed method Atomic-Sample values Average method Individual absorption Tapwater . . . . 0.315 0.316 0.309 0.315 0.318 Wellwater . . . . 0 . S 3 0.554 0.554 0.553 0.555 Hiroseriverwater . . 0.23* 0.233 0.229 0.232 0.234 0.2 0, > m 2 0.1 .- c > m .- L = o 00 m 0.1 0.; [ a ) 450 550 650 b) L I 450 550 651 Wavelength/nrn Fig.3. Influence of ( a ) scan speed (with circuits all No. 6) and (6) circuit number (with scan speed 300 nm min- 1 ) on second-derivative spectra of iron(I1) - PSPmH complex solution. Fe(II) 103 p.p.b.; PSPmH 2.5 x 10-5 M; pH 8.3; slit width. 1 nm; reference reagent blank. Numerical values indicate scan speed in ( a ) and first- and second-differentiation circuit numbers in ( b ) differentiation circuit and the scan speed of the spectropho-tometer in second-derivative spectrophotometry these need to be selected to give a well resolved large peak ( to give good selectivity and higher sensitivity). Fig. 3 shows the second-derivative spectra of the iron( 11) - PSPmH complex measured with various circuit numbers and scan speeds.A combination of circuit No. 6 (which has the largest time constant the time constant increasing with increasing circuit number in our apparatus) and a scan speed of 300 nm min-1 seems to give the best sensitivity and resolution ( i . e. selectivity) for the iron determination. Calibration graph sensitivity precision and accuracy The calibration graph prepared by plotting the second-derivative value versus the iron concentration an example of which is shown in Fig. 4 is linear and passes through the origin when the peak to trough values or the base line to trough values were plotted. The equation for each line measured with a combination of circuit No. 6 and a scan speed of 300 nm min- was Fe (p.p.b.) = 58.50 . . . . .. (3) respectively where D is the second-derivative value represen-ted by the conversion of the value into absorbance. It will be seen that as little as 50 ng of iron can easily be determined in this way. Ten standard solutions containing 517 ng of iron were analysed. The results gave a relative standard deviation of 1.2% and were in good agreement with results obtained by non-flame atomic-absorption specrophotometry carried out for comparison. Fe (p.p.b.) = 1120 . . . . . . (4) Interferences The selectivity in second-derivative spectrophotometry is better than that in ordinary spectrophotometry. The following ions at the levels indicated did not interfere in the determina-tion of 517 ng of iron when the same masking agents as those shown in Table 6 were used copper(II) lead(I1) and vanadium(1V) and(V) 10 pg; mercury(II) 5 pg and cobalt(I1) and nickel(II) 1 pg.Kinetic Studies for the Complex Formation of Iron(I1) with PSPmH Generally because the rate of complex formation defined as the change with time in the concentration of the components in the complexation reaction depends on the concentration 1144 0.8 Q a 2 0.6 > Q .- t. > a .- L z c 0.4 v) 8 0.2 0 10 20 30 40 50 Fe,p.p.b. Fig. 4. Calibration graph for iron in second-derivative spectropho-tometry. Fe A 5.2; B 10.3; C 20.7; D 33.1; and E 51.7 p.p.b. Circuits all No. 6; scan speed. 300 nm min-1; slit width 1 nm; recorder sensitivit XI; cells 50 mm; reference reagent blank. F, Peak to trough vaLes plotted; G base line to trough values plotted 4.2 - - d I 3 3.8 -C 1 I 3.4 -0.02 0.04 0.06 0.08 9 .W React ion t imels Fig. 5. Graph of -In(A - A,) versus reaction time. Fe(II) 3.7 X 10-6 M; PSPmH 2.0 x 10-4 M; pH 7.3; ionic strength = 0.2; temperature 25 k 0.1 "C; wavelength 580 nm of metal ion ligand and hydrogen ion the rate of formation of the iron(I1) - PSPmH complex can be assumed to be -- d[Fe(ll)l - k' [Fe(II)]@[PSPmH]b[H+]c . . (5) dt where t is the reaction time k' is the rate constant and a b and c are the reaction orders with respect to the concentrations of iron(II) PSPmH and hydrogen ion repectively . Reaction order with respect to iron( IZ) concentration Under a pseudo-first-order excess of PSPmH and at a constant pH the reaction rate is expressed as - d[Fe(ll)l = kOb,.[Fe(II)]@ .. . . ( 6 ) dt where kobs is the observed pseudo-first-order rate constant. If ANALYST SEPTEMBER 1984 VOL. 109 50 40 7 30 v) . 2 20 30 40 0 10 20 R~ x 1 0 4 1 ~ Fig. 6. strength = 0.2; temperature 25 k 0.1 "C Graph of koh\ versus R 1 . Fe(II) 3.7 X M ; pH 7.8; ionic 1 I 7.0 7.2 7.4 7.6 7.8 4.8 ' PH Fig. 7. Graph of log (k,bslRr) versus pH. Fe(II) 3.7 X PSPmH 2.0 x M ; M; ionic strength = 0.2; temperature 25 5 0.1 "C 12.5 12.0 s t -I 11.5 11.0 I I 1 I 3.30 3.35 3.40 3.45 3.E 11T x lO3/K-' I Fig. 8. pH 7.4; ionic strength = 0.2 Arrhenius plot. Fe(II) 3.7 X 10-6 M; PSPmH 2.5 X M; a = 1 by integrating equation (6) and then applying Beer's law equation (7) is derived: In (A - A,) = In A" - kobs.t . . . . (7) where Ao A and A are the absorbance of the reaction system at t = 0 t and 00 respectively. Equation (7) means that by experimentally ascertaining the linearity of a graph of -ln(A, - A,) versus t we can establish whether a = 1 or not and the kobs. value from the slope of the graph. Thus a kinetic graph for the complex formation was measured in the presence of a large excess of PSPmH and at pH 7.3 and -ln(A - A,) values were plotted against t (Fig. 5). The plot gives a straight line of slope 21.1 which indicates that the complex formation reaction is first order with respect to the iron(I1) concentra-tion i.e. a = 1 and the observed rate constant kobs. is 21.1 s-l under the experimental conditions given in Fig. 5 ANALYST SEPTEMBER 1984 VOL.109 1145 Reaction order with respect to PSPm H concentration In order to determine the reaction order with respect to the PSPmH concentration the effect of the PSPmH concentration on observed rate constant kobs. was examined in the total PSPmH concentration (RT) range of 4.0 x 10-4-5.0 x 10-5 M. The results are shown in Fig. 6. A graph of kobs versus RT gives a straight line of slope I which passes through the origin. This indicates that the complex formation reaction is first order with respect to the PSPmH concentration i.e. b = 1. Reaction order with respect to hydrogen ion concentration The effect of the hydrogen ion concentration on the observed rate constant kobs. was studied in the pH range 7.0-8.0. As is clear from the results in Fig.7 the hydrogen ion concentration is independent of the observed rate constant in this pH region, i.e. c = 0. Above pH 8 the reaction curve could not be obtained presumably because of the hydrolysis of iron(I1). Effect of other variables The concentrations of borate buffer and sodium ascorbate solutions used were independent of the observed rate con-stant. Rate-determining step and rate constant for the complex formation On the basis of the results described above the rate law for the iron(T1) - PSPmH complex formation can be expressed as -- d[Fe(ll)l - k,[Fe(II)][PSPmH] . . A t where kf denotes thy rate constant for the complex formation. Equation (8) suggests that a 1 1 complex formation reaction between Fez+ and undissociated PSPmH is the rate-determining step.Its rate constant kf was calculated and a value of 1.3 x 105 1 mol-1 s-1 was found. Activation parameters Fig. 8 illustrates the temperature dependence of the rate constant kf in the range 15-30 “C. The graph of In kf versus 1/T (where T denotes absolute temperature) gives a straight line (Arrhenius plot). From this Arrhenius plot the activation parameters were calculated to give the following results: activation energy E = 44.3 kJ mol-1 activation enthalpy AH* = 41.8 kJ mol-1 activation entropy ASS = -6.48 J K-I mol-1 and activation free energy AGt = 43.9 kJ mol-1. Conclusions Three water-soluble hydrazones were synthesised and the complex formation of one of them PSPmH with iron(I1) was investigated in detail with respect to its chemical equilibrium and reaction kinetics. On the basis of the results spectropho-tometric and analogue derivative spectrophotometric methods for the determination of trace amounts of iron have been proposed and applied successfully to the analysis of water samples. The proposed methods offer the advantages of simplicity rapidity reasonable selectivity and high sensitivity. Very high sensitivity was attained on employing second-derivative spectrophotometry. 1. 2. 3. 4. 5. 6. 7. References Going J. E. and Sykora C. Anal. Chim. Ada 1974,70,127. Going J . E. Wesenberg G. and Andrejat G. Anal. Chim. A m 1976 81 349. Bradsher C. K. Parham J . C. and Turner J. D. J . Heterocycl. Chem. 1965 2 228. Singh R. B. Odashima T. and Ishii H. Analyst 1983 108, 1120. Ishii H. and Koh H. Nippon Kagaku Kaishi 1980 203. Ishii H. and Satoh K. Fresenius 2. Anal. Chem. 1982,312, 114. Silverstein R. M. and Bassler G. C. “Spectrometric Identifi-cation of Organic Compounds,” Second Edition 1967 Wiley, New York; translated by Araki S . and Mashiko Y. Tokyo Kagaku Dojin Tokyo 1969 p. 102. Paper A4181 Received February 27th 1984 Accepted March 26th 198

ISSN:0003-2654    

DOI:10.1039/AN9840901139

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1984, VOL. 109 Kinetic - Photometric Determination of EDTA, Zinc and Bismuth Interchange Reactions of )C=N- Groups Angel Rios and Miguel Valcarcel Department of Analytical Chemistry, Faculty of Sciences, University of Cordoba, Cordoba, Spain 1147 bY The in situ formation of the nickel(l1) - 6-methylpicolinaldehyde thiosemicarbazone complex by means of interchange reactions of X=N- groups occurs when nickel(ll), 6-methylpicolinaldehyde azine and thio- semicarbazide are mixed at pH 4.5. The reaction is monitored spectrophotometrically a t 396 nm. Several species cause interferences, and these effects can be used for their indirect determination. The most important compound in this context is EDTA. This Ni(ll) - EDTA - 6-methylpicolinaldehyde azine - thiosemicarbazide system also permits the indirect kinetic - photometric determination of trace amounts of zinc(ll) and bismuth(ll1).Keywords: Indirect kinetic - photometric determination; XXV- interchange reactions; EDTA, zinc(//) and bismuth (111) determination New analytical applications of > C=N- interchange reactions are proposed in this paper. Although these reactions were known in organic'-3 and inorganic66 chemistry, their implica- tions in analytical chemistry were first reported in 1973 by Valcarcel and Pin07 as the cause of the instability of photometric ligands in solution. Subsequently, these reactions were employed for the homogeneous precipitation of pallad- ium dimethylglyoximate,8 using several azomethine deriva- tives of, diacetyl monoxime in the presence of an excess of hydroxylamine.They have also been used for the kinetic - photometric determination of copper(I), cobalt(I1) and nickel(I1) by an in situ formation reaction of chelates.9 The interchange reactions of X=N- groupslo: R \ R \ ,C=N-X + Y-NH2 ,C=N-Y + X-NHZ. R- R- consist of a net change of X for the Y radical by means of the action of an excess of amine on an azomethine compound. These reactions occur at a reasonable speed in aqueous medium at an appropriate pH. In this work we have used the Ni(I1) - 6-methylpicolin- aldehyde azine - thiosemicarbazide system. Earlier we had employed this system for the determination of trace amounts of Ni(I1) by the action of the thiosemicarbazide on the 6-methylpicolinaldehyde azine (6-Me-PAA).g The reaction was carried out in the presence of nickel(I1) at pH 4.5 so that the 6-methylpicolinaldehyde thiosemicarbazone (6-Me-PAT) was formed in situ in its own reaction medium: complexes at pH 4.5.Zinc(I1) and bismuth(II1) ions displace the nickel(I1) ions complexed with EDTA: NiY2- + Zn2+ Z ZnYZ- + Ni*+ and NiY2- + Bi3+ BiY- + Ni2+ The liberated nickel(I1) ions form a complex with the 6-Me-PAT yielded by a X=N- interchange reaction: 6-Me-PAA + 2H2NNHCSNH2 + 2(6-Me-PAT) + H2NNH2 Ni2+ + 2(6-Me-PAT) -+ Ni(6-Me-PAT)22+ The Ni(6-Me-PAT)2*+ complex is monitored photometrically (Amax, = 396 nm) by recording absorbance - time curves. Experimental Apparatus A Perkin-Elmer 575 ultraviolet - visible spectrophotometer with 1.0-cm glass cells and equipped with an electrothermic thermostat (Peltier principle) was used for the kinetic measurements.A Radiometer PHM 62 pH meter with a combined glass - calomel electrode was also used. Reagents All reagents were of analytical-reagent grade and solutions were prepared with distilled water. 6-Methylpicolinaldehyde azine, 0.1 % mlV solution in EDTA interferes in the determination of nickel(I1) by formation of the NiY2- complex, but this suppressive effect can be employed for the indirect kinetic - photometric determination of trace amounts of EDTA. The new system established, Ni(I1) - EDTA - 6-Me-PAA - H2NNHCSNH2, is interesting because it permits new indirect determinations of zinc(I1) and bismuth(II1) ions. The indirect determination of zinc( 11) and bismuth( 111) by means of the NiY2- - 6-Me-PAA - H2NNHCSNH2 system is based on the different stability of NiY2-, ZnY2- and BiY- ethanol.The reagent was easily synthesised by the condensa- tion of 6-methylpicolinaldehyde with hydrazine. I1q12 Thiosemicarbazide solution, 5 x 10-2 M. Standard solutions of nickel(ZI) (98.9 pg ml-I), zinc(U) (99.8 pg ml-1) and bismuth(Z1I) (100.1 pg ml-1). Prepared by the dissolution of appropriate amounts of nickel(I1) nitrate, zinc and bismuth metal, respectively, in water. EDTA standard solution, 0.101 M. Sodium acetate - acetic acid buffer solution, p H 4.5. Total concentration = 1.4 M.1148 ANALYST, SEPTEMBER 1984, VOL. 109 Procedures Kinetic and photometric methods for the indirect determination of EDTA To a 10-ml calibrated flask containing 38.8 pg of nickel(II), add EDTA solution (between 5 x 10-6 and 3.5 x 10-5 M), 1.5 ml of a 0.1% mlV solution of 6-Me-PAA in ethanol, 4 ml of sodium acetate - acetic acid buffer solution (pH 4.5) and an appropriate volume of distilled water so that when 0.3 ml of 5 X 10-2 M thiosemicarbazide solution is finally added, the total volume becomes 10.0 ml.A portion of these samples is transferred into a 1 .O-cm cell thermostated at 35 k 0.1 "C. The change in absorbance at 396 nm is monitored by recording absorbance - time graphs from 45 s after the samples are prepared. Two types of determinations are possible: before the equilibrium (kinetic determination) and when the equilib- rium of reaction is obtained (photometric method). Kinetic and photometric methods for the indirect determination of zinc(II) and bismuth(III) Prepare the samples as described above, but initially add zinc(I1) (3-40 pg) or bismuth(II1) (10-80 pg) ions.The concentration of EDTA is 2 X 10-5 M. Preparation of Samples Detergent A 1.00-g amount of detergent sample (Ariel, courtesy of Procter and Gamble, containing 0.27% EDTA), 1.00 g of sodium sulphite and 25 nil of 30% mlV sodium hydroxide solution were mixed. The solution was boiled for 30 min and distilled water was added to replace the evaporated amount. The solution was transferred into a 100-ml calibrated flask and diluted to the mark with distilled water. Blende mineral A 0.50-g amount of mineral (BAS No. 41dG: Zn 51.4, S 31.2, Fe 10.0, Pb 0.91, Cu 0.10, As 0.13, Mn 1.12, Cd 0.22 and CaF 0.03%) was added to 30 ml of concentrated hydrochloric acid - nitric acid mixture (1 + 1 V / V ) and warmed in order to complete the solution.This solution was filtered and trans- ferred into a 500-ml calibrated flask and diluted to the mark with distilled water. Suppositories One unit (approximattely 3.0 g) (Supobismut: 500 mg of sulphamethoxazol, 100 mg of trimethoprim, 150 mg of bismuth canphocarbonate, 50 mg of benzidamine chlorohy- drate, plus excipient) was taken, 50 ml of concentrated 0.8 1 0.6 9) C f 0.4 0 a 0.2 I I 400 450 Wavelengthhm 0" Fig. 1. In situ formation of nickel(I1) - 6-methylpicolinaldehyde thiosemicarbazone complex by Ni(I1) - 6-Me-PAA - H2NNHCS- NH2 system at pH 4.5 and 35 "C. Nickel, 3.88 pg ml-1; 6-Me-PAA, 6.3 X M; and thiosemicarbazide, 1.5 x 1 0 - 3 M. Graphs recorded over 3 min sulphuric acid - nitric acid mixture (1 + 1 V/V) were added and the mixture was heated under reflux for 1 h.The destruction of the organic matter was completed with several drops of concentrated hydrogen peroxide. After cooling, the solution was transferred into a 100-ml calibrated flask and diluted to the mark with distilled water. Results and Discussion Kinetic Study of Ni(I1) - 6-Me-PAA - HZNNHCSNHZ System The in situ formation of a nickel(I1) - 6-Me-PAT complex is observed spectrophotometrically at pH 4.5 (Fig. 1). The kinetic study of this reaction, made by recording absorbance - time graphs at 396 nm and 35 k 0.1 "C, gave the following results. The partial orders of the reaction were calculated, by plotting log(initia1 rate) against log(concentration), and the rate equation at pH 4.5 is suggested to be V = kl[Ni*+][6-Me- PAA]i[H2NNHCSNH2] + k2[6-Me-PAA] [H2NNHCSNH2] where the second term on the right-hand side corresponds to the reaction in the absence of nickel(I1). Initial rates of reactions were obtained and the influence of different experimental variables was observed, as follows.Influence of p H The reaction proceeds most favourably in the pH range 4.34.8 (Fig. 2); at neutral or alkaline pH the interchange reaction does not take place. Influence of thiosemicarbazide concentration The optimum thiosemicarbazide concentration was found to be 1.0-2.0 x l o - 3 ~ . Influence of nickel( II) concentration Between 1 and 4 pg ml-1 of nickel(I1) a straight-line graph of rate versus [Ni2+] is obtained. Therefore, the kinetic determi- nation of nickel(I1) is possible in this range.Also, when reaction equilibrium is attained, one can carry out the determination of nickel( 11) by the photometric method (0.5-0.8 pg ml-1). The molar absorptivity of the nickel(I1) - 6-Me-PAT complex obtained by the in situ technique is 7.7 X 103 1 mol-1 cm-1. Influence of temperature An increase in temperature results in an increase in the rate constant ( k ) , in accordance with the Arrhenius equation. By plotting log k against T-1, the activation energy for this reaction was calculated and was found to be 8.80 kcal mol-l. For the reaction in the absence of nickel(I1) the value was 5.27 kcal mol-1. Therefore, the reaction in the absence of metallic ion is most favoured. 10-4 3.0 4.0 5.0 6.0 PH Fig.2. Influence of pH on the initial rate reaction of the Ni(I1) - 6-Me-PAA - H2NNHCSNH2 system. Experimental conditions as in Fig. 1ANALYST, SEPTEMBER 1984, VOL. 109 1149 Table 1. Comparison of Zn(I1)- and Bi(III)(NiY2-) - 6-Me-PAA - HzNNHCSNHz systems with the Ni(I1) - 6-Me-PAA - H2NNHCSNH2 system. [6-Me-PAA] = 6.3 x 10-4 M; [H2NNHCSNH2] = 1.5 x 10-3 M; and [EDTA] = 2 X 10-5 M Final [Zn2+] = [Ni2+] = 6.6 x 10-5 M Ni(I1) - 6-Me-PAA - HzNNHCSNHz 116.0 0.565 Zn(II)(NiY*-) - 6-Me-PAA - HzNNHCSNH2 107.8 0.486 [Bi3+] = [Ni2+] = 3.8 x 10-5 M Ni(I1) - 6-Me-PAA - H2NNHCSNH2 68.5 0.337 Bi(III)(NiY2-) - 6-Me-PAA - H2NNHCSNH2 67.0 0.337 Concentration System V x 10-3/min- l absorbance Table 2. Parameters for the determination of EDTA, zinc and bismuth Method of Species determination Concentration range EDTA .. . . Kinetic 5.0 x 10-6 - 2.5 x 10-5 M 5.0 x 10-6 - 3.5 x 10-5 M Zinc(I1) . . . . Kinetic 0.3 - 2.0 pg ml-’ 0.5 - 4.0 pg ml-I Bismuth(II1) . . Kinetic 1.0 - 5.0 pg ml-I 1.0 - 8.0 pg ml-1 Photometric Photometric Photometric Relative standard deviation, O/O 0.83 0.43 0.51 0.83 0.78 0.87 Study of the Systems in the Presence of EDTA EDTA forms a complex with Ni(I1) that is more stable than the Ni(1I) - 6-Me-PAT complex. Therefore, the addition of EDTA to the Ni(I1) - 6-Me-PAA - H2NNHCSNH2 system decreases the initial rate and the final absorbance. The amounts of EDTA added to the samples are limited to high concentrations owing to its total interference with the Ni(I1) - 6-Me-PAA - H2NNHCSNH2 system, and to low concentra- tions owing to its non-interference.This effect is illustrated in Fig. 3. Several ions, such as Zn(I1) and Bi(III), displace nickel(I1) ions coordinated with EDTA at pH 4.5 and an increase in the zinc(I1) or bismuth(II1) concentration brings these new systems near to the initial Ni(I1) - 6-Me-PAA - H2NNHCSNH2 system again. Thus, in the presence of zinc(II), we can represent this situation by the scheme shown below, according to the absorbance - time graphs (Fig. 4). A similar scheme could be formulated for bismuth(II1). Ni(ll) - 6-Me-PAA - H2NNHCSNH2 ? When [Zn*-I=[Ni2+1 / Zn(ll)(NiY*-) - 6-Me-PAA - H2NNHCSNH2 \ \ Ni(ll) - EDTA - 6-Me-PAA - H2NNHCSNH2 When [Zn2+]=0 A comparison of the reaction when there are equimolar concentrations of Zn(I1) or Bi(II1) and Ni(I1) with the Ni(I1) - 6-Me-PAA - HzNNHCSNH2 system is interesting (Table 1).These results show that the displacement is almost total with bismuth(III), whereas for the system with zinc(I1) the difference between the Ni(I1) - 6-Me-PAA - H2NNHCSNH2 and Zn(II)(NiY2-) - 6-Me-PAA - H2NNHCSNH2 systems indicates that between 0.27 and 0.50 pg ml-1 of nickel(I1) a complex with 6-Me-PAT is not formed. The nickel(I1) will be present as NiY2- or free nickel. At this pH we have not observed the formation of Ni(I1) - thiosemicarbazide com- plexes photometrically. These observations are in agreement with literature data13 on the different stabilities of ZnY2- and BiY- complexes at pH 4.5. Analytical Applications Table 2 shows the results obtained for the indirect determina- tion (kinetic and photometric methods) of EDTA, zinc(I1) 0.6 al 6 0.4 5 2 n a I B C 1 D rE 0.2 0 2 4 6 8 10 Timeim in Fig.3. Absorbance versus time graphs at 396 nm for different concentrations of EDTA: A, 0; B. 9.7 X 1 0 F M ; C, 1.9 X lo-’ M ; D, 2.9 x 10 M ; and E. 3.9 x IW’ M . txperimental conditions as in Fig. 1 0.3 Q) c m 5 0.2 2 n Q 0.1 B I I 0 2 4 6 8 1 Timelm in Fig. 4. Absorbance versus time graphs at 396 nm for A, Ni(I1) - 6-Me-PAA - H,NNHCSNH, s stem; B, NiY2- - 6-Me-PAA - H,NNHCSNH, system. (C),&(DY and (E) for the Zn(II)(NiY2-) - 6-Me-PAA - H,NNHCSNH, system at different concentrations of zinc(I1): (C) 1 pg ml-1; (D),-2 pg ml-1; and (E), 3 pg ml-l. Experi- mental conditions as in Fig. 1 (concentration of EDTA, 2.0 X M)1150 ANALYST.SEPTEMBER 1984. VOL. 109 and bismuth(II1) by using the techniques proposed above. Measurements for the photometric method can be made in 12-14 min when reaction equilibrium has been reached. The relative standard deviation (P = 0.05) of the method was calculated for 11 determinations in each instance. The deviations are not too high in spite of the fact that the chemical systems are fairly complicated. The kinetic determination (initial rate method) is more rapid than the photometric method, as reaction equilibrium is not necessary. Interferences The presence of foreign species in the reaction medium affects all systems in a similar way, and tolerances are given in Table 3. The main interferences are due to species that react with Table 3. Interference of foreign ions in the photometric and kinetic determination of EDTA, zinc(I1) and bismuth(II1) by interchange reactions of X=N- groups with 6-Me-PAA and thiosemicarbazide Ion added Tolerance ratio (foreign ion to analysed species) Li+.Na+, K+, Be’+.Ba”, Ca2+. Mg2+. Mn”, AP+, As3+, Ce4+. F-, C1-, Br-, I-, NO1-, NO1 , P04i-, acetate . . . . . . . . . . . . 104-, S20,2-, citrate, tartrate , . . . . . . . 10 Fez+, Pd2+, Pt4+, Moo4’- , . . . . . . . S042-, SO?’-, S 2 0 7 ’ - , COT’-, SCN , ClOj , CIO,-, 100 Pb2+, Cd*+ , Cr3+, perborate . . . . . . . . 50 1 < I Cut, Co2+ , Ag+ , Hg2+, V03-, Mn0,- . Cr20,z- Table 4. Determination of EDTA. zinc(I1) and bismuth(II1) in different samples Amount Species Sample determined analysed Present Found EDTA . . . . . . Detergent* 0.27 O/o Photometric method: 0.53% Zinc(1I) .. . . Blendemineral 51 .4% Photometric method: 51.9% Bismuth(II1) . . . . Suppositories 0.075 gunit-’ Photometric method: 0.068 gunit-’ Kinetic method: 0.30% Kinetic method: 51.5% Kinetic method 0.077 g unit- 1 (BAS No. 41dG) (Supobismut) * Courtesy of Procter & Gamble. azine (generally oxidising), species that react with the metal ions (complex-formation reactions) and species that give coloured solutions (Cu+, Co2+, Fe2+, Pd2+, Ptj+ and V03-), etc. Some of these interferences can be eliminated by suitable treatment of the sample (addition of reducing agents, masking agents, etc.), in order to make these methods applicable to real systems. Applications The methods reported have been used to determine EDTA, zinc(I1) and bismuth(II1) in several samples. The results are given in Table 4. For these determinations the standard additions method was applied. The best results were obtained by the kinetic (initial rate) method. References 1. 2. Curtius, T., and Lubin, A., Chem. Rer., 1900, 33, 2460. Hinman, R. L., J. Org. Chem., 1960, 25, 1775. 3. 4. 5 . 6 . 7. 8. 9. 10. 11. 12. 13. Szmant, H. H . , and McGinnis, C., J. Am. Chem. Soc., 1950. 72, 2890. Muto, Y., Nippon Kagaku Zasshi, 1955, 76, 252. Verter, H. S . , and Frost, A. E., J. Am. Chem. SOC., 1960.82, 85. Martin, D. F., Ad. Chem. Ser., 1963, No. 37, 192. Valcarcel, M., and Pino, F., Talanta, 1973, 20. 224. Rios, A , , and Valcarcel, M., Analyst, 1982, 107, 737. Rios, A., and Valcarcel, M., Quim. Anal., 1983, 1, 227. Patai, ‘S., “The Chemistry of the Carbon-Nitrogen Double Bond.” Interscience, London, 1970. p. 81. Allan, F. J., and Allan, G. G . , J . Org. Chem., 1958. 23. 639. Valcarcel. M., Bendito, D. P., and Pino. F.. If. Quim. Anal., 1971, 25, 1. Ringbom, A , . “Complexation in Analytical Chemistry,” Interescience. New York, p. 332. Paper A411 Received January 3rd, 1984 Accepted March 20th, 1984

ISSN:0003-2654    

DOI:10.1039/AN9840901147

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1984, VOL. 109 1151 Trifluoroethylxanthate as a Reagent for the Determination of Gold Mohamand Farid Hussain, Raj Kumar Bansal and Bal Krishan Puri* Department of Chemistry, Indian Institute of Technology, Hauz Khas, New Delhi- 110 016, India Masatada Satake Faculty of Engineering, Fukui University, Fukui 910, Japan A direct spectrophotometric method has been developed for the determination of gold using trifluoroethylxanthate as the complexing agent. Gold(ll1) reacts with this reagent in the ratio 1 : 3 (metal to ligand) over the pH range 2&11 .O. The complex absorbs strongly at 452 nm and Beer's law is obeyed over the concentration range 1.3-7.5 p.p.m. of gold. The molar absorptivity and Sandell's sensitivity are 1.09 x 103 1 mol-1 cm-1 and 0.018 p.g cm-2 of Au, respectively.Ten replicate determinations of 45 pg of gold gave a mean absorbance of 0.250 with a standard deviation of 1.66 x 10-3 and a relative standard deviation of 0.66%. The interference of various ions has been studied and conditions were developed for the determination of gold in certain synthetic samples. Keywords: Gold(ll1) determination; potassium trifluoroeth ylxanthate; synthetic samples analysis Gold( 111) may be determined using various complexing agents such as diethyldithiocarbamate,' dithizone,' 8-mercaptoqui- noline,2 2-quinolyl aldoxime,3 thioamides4 and aniline .s In most of the examples, the sensitivity is low and the colour fades after a few minutes and, in some instances, the complex is formed only on heating for a long period. Xanthates have been tried for the determination of gold after their extraction into various solvents; however, the stability, sensitivity and selectivity of this method are low because the extraction is often incomplete.Gold has also been extracted as ethylxan- thate in highly acidic media.6.7 In this instance, the extraction is reasonable but the selectivity is again low because many metal ions are extractable under these conditions. In this work trifluoroethylxanthate has been studied as an analytical reagent for the direct spectrophotometric determination of gold(II1) as it is stabler at low pH than more common xanthates. Various parameters such as pH, reagent concentra- tion, standing time, electrolyte concentration and interference of different ions have been studied. This method has been applied to the determination of gold(II1) in certain synthetic samples.The reagent has been found to be fairly sensitive and selective compared with the other complexing agents men- tioned above. Experimental Apparatus An Elico pH meter and Pye Unicam SP-SOO and SP-700 spectrophotometers were used. Reagents All solvents and reagents were of analytical-reagent grade. Gold(ZZZ) standard solution, 10-2 M. Prepare a stock solution of gold(II1) by dissolving the requisite amount of gold(II1) chloride in distilled water and add hydrochloric acid to make the solution 1 M with respect to the acid. Standardise gravimetrically by the hydroquinone method .s Potassium trifli-loroethylxanthate, 0.2%. Prepare by the method of DeWitt and Roper.' Dissolve the reagent in 2 ml of dimethylformamide and adjust the volume with distilled water. Perchloric acid solution, 1 M.Ammonia solution, 1 M . * To whom correspondence should be addressed. Ammonia - ammonium acetate buffer solution, p H 9.4. Mix 0.2 M ammonium acetate solution and 0.2 M ammonia solution in the proportions 1 + 2. Dimethylformamide. Check the purity spectrophotometric- ally before use. Preparation of synthetic samples. Mix an appropriate amount of various metal salts so that the final composition of the mixture corresponds to the standard alloys/minerals. Place the mixture (0.6 g) in a beaker, add 10 ml of aqua regia and heat gently to effect complete dissolution. Then add carefully 15 ml of concentrated hydrochloric acid in 2-ml steps and evaporate the solution to dryness on a steam-bath.Dissolve the residue in 10 ml of 1 M hydrochloric acid, dilute with water and filter if silver is present in the sample. Dilute the filtrate to 500 ml with distilled water in a calibrated flask. General Procedure To an aliquot of gold solution in a beaker add 2 ml of the reagent solution (0.2%). Measure the pH and adjust it if necessary to lie within the range 2.g11.0 by adding 1 M perchloric acid solution or 1 M ammonia solution. Then add 2 ml of buffer solution and 2 ml of dimethylformamide and dilute to 10 ml in a calibrated flask. Place a portion of this solution in a 1-cm glass cell and measure the absorbance at 452 nm. Prepare a calibration graph under similar conditions. Results and Discussion Absorption Spectra The absorption spectra of potassium trifluoroethylxanthate and the gold(II1) xanthate were recorded in water - dimethylformamide solution against water and the reagent blank, respectively.Gold(II1) - trifluoroethylxanthate shows absorption in the range 450-455 nm, the absorption of the reagent being negligible at this wavelength (Fig. 1). Thus, in all instances the absorption was measured at 452 nm. Effect of pH The determination of gold(II1) was studied over the pH range 1.0-12.0 using perchloric acid and ammonia solutions. The absorbance was constant in the pH range 2.0-11.0 (Fig. 2). Effect of Reagent At optimum pH, absorbance measurements were carried out by varying the reagent concentrations. It was observed that1152 ANALYST, SEPTEMBER 1984, VOL.109 the addition of more than 0.5 ml of 0.2% reagent solution was required to obtain maximum and reproducible absorbance for 45 pg of gold. Smaller amounts gave incomplete complex formation. Therefore, 2 ml of a 0.2% solution of the reagent were used throughout the study. Effect of the Volume of Dimethylformamide The effect of the volume of dimethylformamide on the absorbance was studied under the optimum conditions. It was noted that the absorbance remained constant when a mini- mum of 1.5 ml of dimethylformamide were added in 10 ml of the final solution. In all instances, 2 ml of dimethylformamide were used in order to dissolve the chelate in the aqueous phase. A higher volume of dimethylformamide had no effect on the absorbance. Hence, 2 ml of dimethylformamide were preferred.Effect of Standing Time The absorbance of gold xanthate in water - dimethylformamide solution was constant for 2 h. In all instances the absorbance was measured after 15 min. Effect of Electrolyte Various electrolytes, such as sodium chloride, potassium nitrate and sodium acetate (0.01-0.1 M ) , caused no improve- ment in the absorbance, indicating the absence of the salting effect. Composition of the Metal Complex The composition of gold(II1) xanthate was established by Job’s method of continuous variations, A sharp peak at a 0.25 molar fraction suggested the formation of Au(CF3- CH20CS2)3 under these conditions. Further, the composition was also ascertained by the logarithmic method.1° The required values of log[MR,]/[Mtl+] and log(HR1 for the logarithmic method were calculated and a graph was plotted as shown in Fig.3. A straight line with a gradient of 3.0 was obtained, which confirmed the formation of Au(CF3- CH20CS2)3 under these conditions. Choice of Solvent Solvents such as acetone, acetonitrile and dimethylformamide for dissolving the gold(II1) complex were tried. It was found that the absorbance values were low with the first two solvents whereas the results were satisfactory and reproducible with dimethylformamide . Beer’s Law and Sensitivity A calibration graph was constructed under the optimum conditions described above. The graph obeyed Beer’s law over the concentration range 1.3-7.5 p.p.m. of gold. Ten replicate determinations on 45 pg of gold gave a mean absorbance of 0.250 with a standard deviation of 1.66 x 10-3 and a relative standard deviation of 0.66%.The molar absorptivity was calculated to be 1.09 x 103 1 mol-1 cm-1 and Sandell’s sensitivity was 0.018 pg cm-2 of Au at 452 nm for an absorbance of 0.001. Effect of Diverse Ions A 100-mg mass of salt of the anions was added to an aliquot containing 45 pg of gold and the determination was carried out as given under General Procedure. Amongst the anions examined (Table 1) only oxalate, nitrate and EDTA inter- fered and they could be tolerated in relatively small amounts. Among the cations examined (Table 2) only Gulf, Fe3+, Ni2+ and Pb*+ interfered. The interference from iron(II1) was eliminated by masking at pH 8.0 with 0.5 ml of 10% ascorbic acid solution, whereas copper(II), nickel( 11) and lead( 11) were masked with 0.5 ml of 10% EDTA solution at pH 8.0.The method was fairly selective for the direct spectrophotometric determination of gold(II1) in certain synthetic samples corre- sponding to alloys and minerals (Tables 3 and 4). A,,,, = 450 - 455 nm /09 One of the authors (M.F.H.) is grateful to the Council of Scientific and Industrial Research. New Delhi. for the award of a Fellowship. Wavelengthhm Fig. 1. Absorption spectra of A , potassium trifluoroethylxanthate measured against water; and B, old xanthate measured against reagent blank for 45 pg of gold. Zonditions: reagent, 2 ml of 2% solution; dimethylformamide, 2 ml; pH, 9.4; and total volume, 10 ml 0 2 4 6 8 10 12 PH Fig. 2. Effect of pH an absorbance. Au, 45 pg; reagent, 2 ml0.2% solution; dimethylformamide, 2 ml; total volume, 10 ml; reference, reagent blank; and wavelength, 452 nm Log [ H Rl Fig.3. Logarithmic method of formula determination. Initial concentration of gold, 0.25 x 10-3 M; potassium trifluoroethylxan- thate, 0.25 x 1 0 - 3 M; pH, 9.4; other conditions as in Fig. 2ANALYST, SEPTEMBER 1984, VOL. 109 1153 Table 1. Effect of anions on the absorbance of the gold complex. Measurements were carried out in dmmonia - ammonium acetate buffer solution (pH 9.4) using 45 pg of gold and 2 ml of 2% reagent solution Amount added/ Absorbance at Salt added mg 452 nm - - 0.249 Sodium nitrite . . . . . . . . 100 0.248 Sodium nitrate . . . . . . . . 40 0.245 100 0.180 Sodiumfluoride . . . . . . 100 0.250 Potassiumchloride . . . . . .100 0.250 Potassiumbromide . . . . . . 100 0.250 Potassiumiodide . . . . . . 100 0.25u Sodiumcarbonate , . . . . . 100 0.240 Sodiumphosphate . . . . . . 100 0.248 Potassiumsulphate . . . . . . 100 0.250 Potassium thiocyanate . . . . 100 0.240 Sodium acetate . . . . . . . . 100 0.250 Sodium oxalate . . , , . . . . 60 0.242 100 0.234 Sodium citrate . . . . . . . . 100 0.250 Ammoniumtartrate . . . . . . 100 0.250 DisodiumEDTA . . . . . . 80 0.240 100 0.230 Table 2. Effect of cations on the absorbance of the gold complex. I Metal ion Absorbance Conditions as in Table Metal salt added - Copper(I1) chloride Cobalt(I1) chloride Nickelchloride . . Lead(I1) nitrate Palladium(I1) chloride Zincnitrate . . . . . . . . . . . . . * I . addedlyg Antimony(II1) chloride . . Sodiumarsenite .. . . Iron(II1)chloride . . . . Potassium tellurite . . Ammonium molybdate . . Platinum(1V) chloride . . Iridium(II1) chloride . . Sodiumtungstate . . . . Osmium(VII1) oxide . . Ruthenium(II1) chloride Rhodium(II1) chloride . . Silver(1) nitrate . . . . Indium(II1) sulphate . . Thallium(1) nitrate . . Sodiumselenite . . . . Mercury(I1) chloride . . Ammonium metavanadate(V Chromiumnitrate . . . . Cadmiumchloride . . Calciumchloride . . . . Aluminium nitrate . . Magnesium chloride . . Potassium titanium(1V) oxalate . . . . . . Silicate . . . . . . . . - 250 100 100 100 200 100 200 200 200 500 200 200 100 200 300 100 100 500 200 300 200 200 200 200 4000 4000 4000 ) 200 4000 4000 at 452 nm Remarks 0.249 0.255 0.258 0.252 0.248 0.256 0.240 0.246 0.250 0.248 0.252 0.250 0.245 0.255 0.250 0.255 0.253 0.255 0.246 0.250 0.250 0.250 0.248 0.246 0.250 0.238 0.248 0.250 0.248 0.245 0.248 After masking with 0.5 ml of 10% EDTA at pH 8.0 After masking with 0.5 ml of 10% EDTA at pH 8.0 After masking with 0.5 ml of 10% EDTA at pH 8.0 After masking with 0.5 ml of 10% ascorbic acid at pH 8.0 After precipitat- ing as AgCl with 1 MHCI Table 3.Determination of gold in synthetic gold - copper - silver samples Amount of Amount of Relative gold taken/ gold standard Composition, o/o ug found*/ pg deviation, YO Au-Cu-Ag(50+20+30) . . 21.4 21.55 + 1.20 36.7 36.90 +0.93 53.5 53.45 -0.31 64.2 63.80 -0.59 Au - CU - Ag (60 + 35 + 5) . . 18.7 19.05 +0.98 30.2 30.50 +0.28 49.4 49.25 -0.27 67.5 67.10 -0.41 AU - CU - Ag (40 + 55.5 + 4.5) 20.4 20.45 + 1.23 32.8 32.95 + 1 .05 46.3 46.60 +0.87 52.5 52.60 +0.99 AU - CU - Ag (20 + 70 + 10) .. 30.5 30.70 +1.12 42.9 42.85 -0.78 58.0 58.20 +0.59 66.2 65.75 -0.38 * Average of five determinations. Table 4. Determination of gold in synthetic mixtures corresponding to alloys and minerals Re I at ive Synthetic Amount of Amount of standard Alloy or composition, gold taken/ gold found*/ deviation, mineral Y O Pg CIg Y O Au - Pd alloy Sylvanite mineral . . Porpezite mineral . . Au, 50.0: Pd, 50.0 Au, 24.5; Te,62.1; Ag, 13.4 Au, 7.84; Pd. 89.13; Pt, 0.505; Rh, 0.505; Fe, 0.505: c u , 0.505; Zn, 0.505; Pb, 0.505 20.5 38.0 52.9 68.2 30.5 43.0 57.2 73.6 32.5 45.0 50.1 65.6 * Average of five determinations. 21.15 38.90 53.10 68.65 30.78 42.90 57.05 73.90 33.25 45.70 50.60 66.30 +0.80 +0.69 +1.18 +0.72 +0.64 -0.37 -0.52 +0.89 + 1.42 +0.88 +0.86 + 1.04 1. 2. 3. 4. 5. 6. 7. 8. 9. 10 References De, A. K . , Khopkar, S. M., and Chalmers, R. A . , “Solvent Extraction of Metals,” Van Nostrand Reinhold, New York, 1970. Suprunovish, V. I., and Shevchenko, Yu. I., Zh. Anal. Khim., 1979, 34, 1738. Dutta, N. K., and Dhar, S . N., J. Znst. Chem. Calcutta, 1978, 50,83; Chem. Abstr., 1978, 89, 139922h. Radushev, A . V., and Golomolzin, €3. V . , Zh. Anal. Khim., 1979, 34, 742. Rzeszutko, W., and Kopec, T., Fresenius Z . Anal. Chem., 1977, 285, 125; Chem. Abstr., 1977, 87,94987~. Donaldson, E. M., Talanta, 1976, 23,411. Donaldson, E. M., Talanta, 1982, 29, 663. Vogel, A. I., “A Text Book of Quantitative Inorganic Analysis,” Third Edition, Longmans, London, 1969. DeWitt, C . C . , and Roper, E . E., J. Am. Chem. SOC., 1932,54, 444. Zolotov, Yu. A., “Extraction of Chelate Compounds,” Ann Arbor Science Publishers, Ann Arbor, MI, 1970, p. 122. Paper A31445 Received December 19th, 1983 Accepted February 4th, 1984

ISSN:0003-2654    

DOI:10.1039/AN9840901151

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST, SEPTEMBER 1984, VOL. 109 1155 Extraction - Spectrophotometric Determination of Boron with 2-Methylpentane-2,4-diol and 8-Hydroxy- 1 -(2-hydroxy-l- naphthylmethyleneamino)naphthalene-3,6-disulphonic Acid in Toluene Jose Aznarez and Jose M. Mir* Analytical Chemistry Department, Faculty of Sciences, University of Zaragoza, Zaragoza, Spain A spectrophotometric method for the determination of boron after its extraction into toluene containing 2-methylpentane-2,4-diol has been developed. In a portion of the extract boron is determined spectro- photometrically at 460 nm in N,N-dimethylformamide medium after formation of the 8-hydroxy-l-(2- hydroxyl-l-naphthylmethyleneamino)naphthalene-3,6-disulphonic acid - boron complex, in 40% glacial acetic acid, which has been heated at 80 "C for 60 min.The molar absorptivity is 1.54 x 103 I mol-1 cm-1 and the range is linear between 1 and 10 pg ml-1 of boron. Interferences have also been studied. Keywords: Boron determination; Z-methylpentane-2,4-diol extraction; 8-hydroxy-l-( 2- hydroxy-1 -naphth yl- methyleneamino)naphthalene-3,6-disulphonic acid; spectrophotometry The need for a more selective and sensitive determination of boron in natural products created the necessity to study different methods of boron separation.' Of the different existing methods, extraction into an organic phase is one of the techniques more generally used and better results have been obtained,2 owing to the fast extraction equilibrium and its simple handling. A large number of extraction studies have been carried out on both high- and low-polarity organic phases,s6 the latter having some advantages because of its selective extraction thereby avoiding extracting cationic species such as Fe( 111) by the formation of ion pairs or oxonium compounds in the solvent.' Chloroform has been generally used as an extraction medium, but because of its toxicity, high cost and low boiling temperature, it was decided that another solvent such as toluene, which has lower toxicity, should be used.8 Among the different possible extraction reagents, it was decided to use 2-methylpentane-2,4-diol (MPD), which has been used previ- ously in this department.9 The utilisation of the molecular UV - visible absorption technique was adopted and of the different reagents (curcu- min, 10 1,l '-dianthrimide, 11 carminic acid, 12 Azomethine H.13 etc.) 8-hydroxy-l-(2-hydroxy-l-naphthylmethyleneamino)- naphthalene-3,6-disulphonic acid (HSNHN) (I) was chosen. OH N II CH Hoy--lJJ Experimental Apparatus The work was carried out with a Pye Unicam SP8-100 UV - visible spectrophotometer, a Kotterman mechanical shaker and a Gebruder-Haake thermostatic bath. Plastic, PTFE and platinum materials were used in order to avoid possible contamination from boron in glass. Reagents All products used were of analytical-reagent grade and obtained from Merck, unless stated otherwise. 2-Methylpentane-2,4-diol solution, 10% VJV in toluene. Obtained from Eastman Kodak. H S Y H N solution, 20% in dimethylformairiide. Standard boric acid solution. 1000 pg ml-1. The acid was Azomethine H solution, 20% in water.dried at 110 "C before dissolving. Procedure Take a sample volume containing from 30 to 3000 pg of boron, as boric acid. Add 10 ml of HCl - KCl buffer solution (pH 1.5) and extract with 10 ml of extractant solution (10% MPD solution in toluene). Shake the mixture mechanically for 4 min and allow the phases to separate. Dry the organic phase with approximately 1 g of anhydrous sodium sulphate and to 3 ml of this extract add 3 ml of HSNHN solution in dimethylformamide and 3 ml of glacial acetic acid. Heat the solution in a water-bath at 80 "C for 60 min. Allow it to cool to room temperature in darkness and dilute to 10 ml with glacial acetic acid. Measure the absor- bance at 460 nm against a reagent blank prepared in the same way but without boron.Simultaneously, prepare a calibration graph, following the same procedure, with known boron solutions obtained by dilution with the standard boric acid solution. This compound is an azomethine derivative, synthesised in our department, and has already been applied in the aqueous phase. 1J *To whom correspondence should be addressed. Results and Discussion Boron Extraction Boron reacted with MPD forming a compound that was extractable in toluene. The effect of different parameters on the extraction yield (pH of the medium, shaking time and amount of diol) was studied. The effect of pH is shown in Table 1. A 10 pg ml-I boron solution was used and the boron1156 a C 0.4 e $ 4 0.3 a 0.2 ANALYST. SEPTEMBER 1984, VOL. 109 - - - Table 1. Effect of pH on the extraction yield.Boron was determined in the aqueous phase using the Azomethine H method. Total boron, 75 pg rnl-1 PH 5 4.5 4 3.5 3 2.5 2 1.5 1 Absorbance 0.925 0.699 0.489 0.209 0.073 0.071 0.072 0.070 0.069 Boron in aqueous phase/pg rnl-' 7.6 5.3 2.3 0.7 0.5 0.6 0.4 0.4 10 Boron extracted, % 87 90 93 97 99.1 99.3 99.2 99.3 99.5 0 500 450 400 Wavelengthlnm Fig. 1. Absorption spectrum of (1 j HSNHN reagent in dimethylfor- rnarnide - acetic acid - toluene and (2) B - HSNHN complex in dimethylformarnide - acetic acid - toluene 0 . 7 ! 7 - - - 7 c 0.6 0.5 1 O.' t / / I I I I I 1 2 3 4 5 Volume of DMF/ml Fig. 2. Absorption spectra for varyin acetic acid : dirnethylforrnam- ide (DMF) : toluene ratios; total vofume = 10 ml: A, 4 ml of acetic acid; B, 3 rnl of acetic acid; C , 2 ml of acetic acid; D, 1 rnl of acetic acid and E, no acetic acid existing in the aqueous phase was determined by the Azome- thine H method.The optimum diol concentration was 10% V/V, with which reproducible results were obtained. A 15% diol concentration or higher produced difficulties because of the high viscosity of the solution and because of the belated colour development, which interfered with accurate spectro- photometric measurements. The extraction can be considered quantitative as an extraction yield higher than 97% was obtained when the pH of the solution was between 1 and 3. Hydrochloric acid was used because it is the common acid used in the dissolution of natural samples, although sulphuric acid can also be used, giving the same results.The compound formed was extracted quickly and total extraction was obtained by shaking for 4 min. The aqueous to organic phase ratio used was 3 : 1. It was observed that the recovery of boron was quantitative for a single extraction using the conditions indicated and a yield of 99.5% was obtained. The procedure involved the determination of boron in the aqueous phase, after extraction, using the Azomethine H method. In a previous paper14 the reactivity of the HSNHN reagent was studied. In this work the application of this reagent was extended to the spectrophotometric determination of boron in a toluene extraction phase. In Fig. 1 the spectra of the free reagent (two maxima at 400 and 470 nm) and the boron - reagent compound (one maximum at 460 nm) are shown.The presence in the organic phase of a miscible acid was required in order to form the boron - reagent compound. Of the different acids tested the best results were obtained with glacial acetic acid. The highest absorbance values at 460 nm were attained with 40% of acetic acid in the final solution. This is shown in Fig. 2. Temperature and Heating Time The formation of the compound was very slow at room temperature. More than 5 h were needed to reach a constant absorbance. As a consequence, the temperature and heating time were studied in order to reduce the time of colour development. The maximum absorbance value varies very little with temperature and it was found that a temperature of 80 "C and a heating time of 60 min were suitable conditions for the spectrophotometric determination.At higher tempera- tures the absorbances were lower, owing to the decomposition of the compound and at lower temperatures a long heating time was needed. Compound Stability Variation of the absorbance with time after heating at 80 "C for 1 h and cooling to room temperature showed that the absorbance remained constant for 5 h and then started to decrease slowly. Beer's Law Once the optimum conditions for the determination of boron with HSNHN in toluene were established, the concentration range in which Beer's law is obeyed was studied. This range was linear between 1 and 10 pg ml-1 with a calibration graph of A = 1.55 X lO3x + 0.0035 where A is the absorbance value at 460 nm and x is the concentration of boron (pg ml-1) in the organic phase.The correlation coefficient was 0.999 and the molar absorptivity 1.54 x 103 1 mol-1 cm-1. Sandell's sensitivity was 7.45 x 10-4 pg cm-2 of boron in the organic phase. Interferences Chloroform as the organic phase in the extraction has the advantage of extracting few elements, other than boron. IfANALYST. SEPTEMBER 1984. VOL. 109 1157 Table 2. Interference study. Boron determination with HSNHN; total boron, 2 pg ml-1 In terferent NH4+ . . . . Na + . . . . K’ . . . . . . Ca2+ . . . . Mg2+ . . . . Al’+ . . . . Mn’+ . . , , Fe3+ . . . . cuz+ . . . . Zn’+ . . . . Ni2+ . . . . Ba2+ . . . . P o p . . * . sod2- . . . . Total content/mg 60 60 60 60 10 40 10 10 1 1 1 1 1 10 Interferent: boron ratio (mlm) 3 x lod 3 x 104 3 x lo4 3 x lo4 5 x 103 2 x 104 5 x 10-7 5x103 5 x 10’ 5 x 10’ 5x10’ 5 x 102 5 x 102 5x103 the organic phase has a high polarity5 other cations such as Fe(II1) and Sb(V) can be extracted by formation of ion pairs.The results in Table 2 show that none of the cations mentioned previously are extracted when toluene is used as the extraction medium. Only F- produces an appreciable interference in the determination because BF4- is formed, thus avoiding the extraction of boron. Interference can be eliminated by adding ZrC14 to the aqueous phase. 1 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. References Nemodruk, A. A., and Karalova, Z. K., “Analytical Chemistry of the Elements,” Ann Arbor Science Publishers, Ann Arbor, MI, 1969. Nikolaeva, A. V . . and Urdanova, A. G., Tr. Konf. Khim. Bora, Ego Soedin., 1958, 157. Svarcs, E., Latv. PSR Zinat. Akad. Vestis Khim. Ser., 1981,5, 557. Wear, J. I . , in Black, E. A,, “Methods of Soil Analysis, Part 2,” American Society of Agronomy, Madison, WI, 1965, chapter 75, p. 1059. Maeck, W. J . , Booman, G. L., Kussv, M. E., and Rein, J. E., Anal. Chem., 1961, 33, 1775. Lanza. P., and Buldini, P., Anal. Chim. Acta, 1974, 70, 341. Aznarez, J., and Bonilla. A , , An. Quim., 1958, 74, 756. Weast, R. C., Editor, “CRC Handbook of Chemistry and Physics,” Forty-eighth Edition, CRC Press, Cleveland, OH, 1968. Aznarez, J . , and Mir, J . M., Analyst., 1983, 109, 368. Wikner, B., Commun. Soil. Sci. Plant Anal., 1981, 12, 697. Villanova, A . , Metal. ABM, 1980, 36. 271. Ogner. G., Commun. Soil Sci. Plant Anal., 1980, 11, 1209. Melton. J . R., J . Assoc. Off. Anal. Chem.. 1982, 65. 234. Mir, J. M., Thesis, Faculty of Science, University of Zaragoza, Spain, 1982. Paper A31451 Received December 22nd, 1983 Accepted March 21st, 1984

ISSN:0003-2654    

DOI:10.1039/AN9840901155

出版商:RSC    

年代:1984

数据来源: RSC