摘要:

MAY, 1969 Vol. 94, No. I I18 THE ANALYST The pH Meter as a Hydrogen-ion Concentration Probe BY W. A. E. McBRYDE (Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada) Results obtained with three commercial pH meters, by measuring the e.m.f. of cells consisting of glass electrodes coupled with either calomel or silver chloride reference electrodes, are compared with the hydrogen-ion concentrations in solutions with three different background electrolytes at five or six different levels of salt concentration. The hydrogen-ion functions measured in this way can be empirically converted into concentrations for use in calculating equilibrium constants. When thermodynamic equilibrium constants cannot be calculated from such measurements it is preferable to report concentration quotients.Concordance is demonstrated among values of the second-stage ionisation constant for sulphosalicylic acid, measured in different ionic media and with different reference electrodes, when all values are converted into concentration quotients. RESULTS from measurements with commercial pH meters have been widely incorporated into calculations of equilibrium “constants” of two principal types. The first type are the dissociation constants of weak acids and bases, with which hydrolysis constants of various ionic species such as metal cations in aqueous solution can appropriately be grouped. The second type comprises formation constants of metal complexes, which may be obtained from the study of reactions that take #ace by proton displacement, with methods clearly enunciated by Bjerruml and widely applied since.The ionisation of a weak acid, HA, although perhaps better regarded as a proton-transfer reaction with the solvent, can be considered seciently well for our purposes as a simple dissociation, HA + H + A. (Charges will be omitted, if the meaning is not obscured, in the interest of simpler representation and of greater generalisation). The ionisation constant is defined by the equation Activities are represented by ai (concentrations are given by square brackets) and activity coefficients by yi. A concentration quotient Qa is defined by the equation in which Ya represents the appropriate quotient among the activity coefficients. It has become common practice to measure concentration quotients in media with a background salt at a constant and relatively high concentration, thus enabling work to be carried out with solutions of essentially fixed ionic strength in which Ya remains constant.Qa may thus be treated as a conditional ionisation constant2 for HA, with validity limited to solutions with the same background compositions as those in which the measurement of Qa was made. A third and related quotient has been described as a Bronsted acidity c~nstant,~ and is regarded as having the form .. .. .. . . (3). aH K, =- [HA] . a The justification for introducing a mixed quotient, containing both activity and concentration terms, is the supposition that the hydrogen-ion property, usually measured by the e.m.f. of a cell (or by comparison with buffers calibrated in this way), is the activity of this ion.0 SAC and the author. 337338 MCBRYDE: THE PH METER AS A HYDROGEN-ION CONCENTRATION PROBE [Artalyst, Vol. 94 This supposition amounts to identifying pH with -log aH. Many acidity constants recorded in the literature have been obtained by means of the familiar relationship .. and are assumed to fit into this category of mixed acidity constants. The ratio [HA]/[A] can be obtained with reasonable accuracy by various analytical methods (titration, spectro- photometry, etc.), and the pH value is that indicated by a glass electrode and meter, thus values of pKH are easily obtained. A problem arises with acidity constants measured in this way, which concerns the physical significance to be attributed to pH measurements, especially when measured with glass and calomel electrodes.This electrode assembly is calibrated with buffers, the com- positions of which have been carefully chosen to permit identification of pH with -log aH in those solutions,* but many of the recorded acidity constants have been measured in much more complex media than these, for instance, in solutions containing, say, 1 mole litre-l of a salt such as sodium perchlorate, or prepared in 50 per cent. v/v aqueous dioxan solution. The question arises of what significance is to be attached to the e.m.f. detected by a pH meter calibrated with buffers when the electrodes are immersed in these much more complex solutions. The assumptions concerning ionic activity coefficients and liquid junction potentials that may have been justifiable in dilute buffers cannot reasonably be expected to apply under such markedly different conditions as in the examples cited.It becomes evident that pH measurements made in such media cannot be related exactly to hydrogen-ion activities and, indeed, it is doubtful if any physical significance can be attributed to them. It has to be decided what pKi means when defined and determined by equation (4). It is true that for aqueous solutions with ionic strengths up to about 0.1 M, one may resort to making approximate activity corrections based on Debye - Huckel theory, with a view to converting mixed constants into either concentration quotients, Qa, or thermo- dynamic equilibrium constants.6 But to do so one must assume that pH values do relate to hydrogen-ion activities, and for such solutions this assumption probably offers a reasonably good approximation. However, numerous investigators of ionic equilibria in solution advocate that reactions should be studied in media with quite high concentrations of supporting electrolyte, with the object of blanketing any minor variations in ionic strength that may occur through shifts in the reactions under consideration and thereby maintaining the ionic strength almost constant.Because of this practice, several equilibria involving hydrogen ions have been studied under such conditions. Other investigations of this type have been conducted with solutions containing partially non-aqueous solvents. With solutions of this kind it is especially true that one should be most concerned about the interpretation of pH values and the significance of pKi values derived from them.In making use of published ionisation constants one must pay close attention to the kind of quotient that has been determined in the first place. While there are many meticu- lously determined thermodynamic ionisation constants, about the reliability of which there can be no question, there are also many quotients, particularly of the kind defined by equation (4) and based on pH measurements in solutions, in which it is impossible to establish any theoretical relationship with hydrogen-ion concentration or activity. It has been the practice in the determination of stability constants of metal complexes to combine equilibrium quotients for reactions such as with “practical” acidity constants for HjL to obtain the desired value.6 Provided the equilibria for the acid dissociation and for the proton displacement [reaction (5)] have been studied under compatible conditions, namely the same concentration of background salt and the same solvent composition, no exception can be taken to this practice, for the hydrogen- ion property cancels out of the final expression (unless protonated metal complexes should form).But instead of recording practical acidity constants of this kind, measured under unfavourable conditions, as characteristic of the acid in question, it is probably better to convert these into concentration quotients by making use of an empirical relationship between pHapparent and [HI. Some physical significance may at least in this way be attributed to the constant thus reported.M + rtHjL + MLn + rtjH . . .. .. ‘. (6)May, 19691 MCBRYDE: THE PH METER AS A HYDROGEN-ION CONCENTRATION PROBE 339 It is surprising to find in some contemporary papers values of acidity constants deter- mined by the use of equation (4), with pH values measured by a commercial pH meter, but reported as concentration quotients. This is tantamount to equating [HI with antilog (-PHapparent), and may be a rather misleading practice. Whether one’s preference lies in reporting acidity constants as mixed quotients or concentration quotients, it becomes necessary on occasion to be able to convert pH meter readings into hydrogen-ion concentrations. For instance, in the titration of a weak acid of total concentration CA (after adjustment for dilution at any stage) with a strong base, the added concentration of which, similarly adjusted, is CB, electrical neutrality imposes the condition on the mass-law relationship Equation (4) thus becomes a special case of equation (6), with the activity coefficients concealed in pKg and the concentrations [HI and [OH] assumed negligible by comparison with CB and CA.Should this latter assumption not be warranted, values of [HI or of [OH] may be required to evaluate pKi, and a bias may be introduced into the results by putting [HI = antilog(-pH) or [OH] = 14 - antilog(-pH). Also, if one is determining the degree of metal complex formation, FiM, in reaction (5) by pH titrations, the following relationship applies at constant ionic strength in which CL, CM, CH and COH represent total concentrations of ligand, metal, ionisable hydro- gen, and of strong base added to the solution, respectively.The item FiH refers to the degree of protonation of the ligand and is a function of pH and the acidity constants of the acid HjL. It may happen that [HI or [OH] is not negligible compared with CH - COH, and incorrect values for either of the former could vitiate the estimates of +iM. Several years ago this particular situation arose in the author’s laboratory; in measuring giiM in solutions of low pH, values of antilog (-pH) were used instead of [HI in evaluating equation (7) and, in consequence, misleading formation curves for a metal complex were obtained, which seemed dependent on hydrogen-ion concentration and therefore suggestive of the formation of a protonated complex.The experimental part of this paper describes efforts to relate values of H, defined as antilog (-PHapparent), to [HI in solutions containing various concentrations of certain back- ground electrolytes. With this information it should be practical to use pH meter readings for solutions in these media to determine hydrogen-ion concentrations. If one accepts that pH signifies -log aH, then it is clear that H should equal aH = yH[H], but if we are concerned about the medium affecting the value of yE, or about liquid junction potentials, we should perhaps look into how the ratio H/[H] is affected by changes in background electrolyte or by elimination of the liquid junction potential. At the same time, an answer is sought in the experimental investigation as to how precisely, under ordinary conditions and with ordinary equipment, one can expect to evaluate hydrogen-ion concentrations by means of a pH meter.Solutions were prepared with known concentrations of strong acid and various known concentrations of background electrolyte. The pH of each solution was measured at 25” C by three different pH meter assemblies, each involving a glass electrode combined with a saturated calomel electrode. Also, to eliminate the liquid junction associated with the calomel electrode, measurements were made on some of these solutions, to which a very small “spike” of sodium chloride had been added, with a silver chloride reference electrode. EXPERIMENTAL PREPARATION OF SOLUTIONS- All reasonable efforts were made to ensure that solutions were prepared free from contaminating substances.Freshly boiled distilled de-ionised water was used in all pre- parations. All chemicals used were of AnalaR, A.C.S. certified, or primary standard grade. Whenever possible, solutions were prepared with nominal concentrations by direct weighing. Solutions of acids and alkalis were carefully standardised by conventional titrimetric methods, and all solutions of alkali were freshly prepared to be as free from carbonate as possible,340 MCBRYDE: THE PH METER AS A HYDROGEN-ION CONCENTRATION PROBE [Analyst, Vol. 94 Sodium perchlorate solutions presented some problems because of the presence of car- bonate in commercially available batches of this salt. They were accordingly prepared by adding a slight excess of 60 per cent.aqueous perchloric acid solution to weighed amounts of anhydrous sodium carbonate; after prolonged boiling and cooling in an atmosphere of nitrogen, the excess of acid was nentralised with standard solution of alkali to pH 7. The final concentration could be calculated from the total amount of sodium ion taken or checked titrimetrically by acid displacement from an ion-exchange resin. All dilutions and preparations were carried out with Class A calibrated glassware. Solutions of nitric, perchloric and hydrochloric acids were prepared containing known concentrations of these acids ranging from 10-1 to M. Into these solutions were placed sufficient potassium nitrate, sodium perchlorate and potassium chloride, respectively, to form sets of solutions the total anion concentrations of which were, in turn, 0.05, 0.1, 0-2, 0.5 and 1.0 M.A further set of perchlorate solutions was prepared with a total anion concen- tration of 3 . 0 ~ . To the nitrate and perchlorate solutions sufficient sodium chloride was also added to yield a final molarity of 0.001 for the purpose indicated below. All solutions were controlled by thermostat at 25.0" & 0.1" C, and measurements on portions of these were made in beakers in jackets maintained at this temperature by circulating water. Buffer solutions were prepared as follows: (a) 0 - 0 5 ~ potassium tetroxalate, pH 1.68; (b) 0.01 M potassium tetroxalate, pH 2-15; (c) 0.05 M potassium hydrogen phthalate, pH 4.005; ( d ) Beckman buffer packet diluted as instructed, pH 6-865; and (e) Clark and Lubs' solution, 0.0102 M hydrochloric acid plus 0.05 M potassium chloride, pH 2.10.The first four of these solutions were, in addition, 0401 M with respect to sodium chloride; it was found that the presence of sodium chloride in the buffers at this concentration resulted in no perceptible change in pH as compared with buffers of the same concentration containing no sodium chloride. PREPARATION OF SILVER CHLORIDE ELECTRODES- These were prepared from Beckman No. 39261 silver billet electrodes, and were momen- tarily dipped in concentrated nitric acid, cleaned with a household scouring powder, rinsed clean, soaked in concentrated ammonia solution for 2 hours and then washed in distilled water for 24 hours.They were then anodised with chloride by electrolysis in 0.1 M hydro- chloric acid at a current density of 1.2 mA cm-2 for 30 minutes (platinum cathode). The electrodes were then washed in distilled water for 24 hours and soaked in 1 0 - 3 ~ sodium chloride solution for at least 6 hours before use. When not in use they were stored in M sodium chloride. Electrodes prepared in this way were used to make all the measurements in one back- ground salt solution, which required about 1 week. They were then cathodised to remove the chloride coating and re-prepared for further use as described above. pH MEASUREMENTS- The pH of each solution prepared as described above was measured on separate portions by the following three different pH meter assemblies.Cambridge fiortable fiattern fiH meter, Model 44237-This was modified by replacement of the four 1.5-V dry cells by a 6-V automobile battery, with marked improvement in stability. This is a null-type potentiometer, the slide wire encompassing 2.2 pH units or 220mV (according to setting), with each scale division equivalent to 0.02 pH or 2 mV. The electrodes used were E.I.L. GHS23/B all-purpose glass and RJ23 calomel. Radiometer $H meter, Model PHM4d-This is also a null-type potentiometer with full scale on the slide wire corresponding to 1.1 pH unit or 110mV, with each slide division equivalent to 0.1 pH or 1 mV. The electrodes used were Radiometer G202B glass and K 401 calomel. Orion research fiH meter, Model 801-This is a direct-reading potentiometer with digital read-out to 0.001 pH unit or 0.1 mV.The reading of the instrument is presented at 1-second intervals so that instability or drift of the measured signal is quickly seen. The electrodes used were Beckman 41263 glass and Leeds and Northrup 1199-31 calomel. These assemblies were standardised each time a set of solutions in a particular medium was to be measured, by means of buffers (c) and (b) or (e). For measurements on perchlorateMay, 19691 MCBRYDE: THE PH METER AS A HYDROGEN-ION CONCENTRATION PROBE 341 solutions and the standardisation preceding these, the calomel electrodes were fitted externally with a sleeve of glass tubing terminating in a sintered-glass disc, into which a 3-5 M ammonium. chloride - 3 per cent. agar preparation was placed.The effect was to interpose another com- partment of electrolyte between the perchlorate solutions and the saturated potassium chloride solution of the calomel electrode. This precaution avoided the problem of precipi- tation of potassium perchlorate at the site of the liquid junction, which can, on occasion, cause drifting readings. The silver chloride electrodes previously described were also used as reference electrodes with each meter. In these measurements the e.m.f. was determined in millivolts with each meter. It was realised that the cell now measured a hydrogen-ion function different from that measured by the cell with a calomel electrode. This is because a term including the activity coefficient of the chloride ion must now be introduced into the expression for the e.m.f.of the cell. To illustrate this, suppose we consider two solutions, 1 and 2, differing in hydrogen-ion concentration but containing the same concentration of chloride ion. Let the pH (= -log a,) of these be pH, and pH,, respectively. Suppose the e.m.f. of the cell, consisting of a glass electrode and a silver chloride electrode dipping into each solution in turn, is measured. The expression for the e.m.f. may conventionally be represented for 25" C by The difference between the two e.m.f .'s is Ecell = EogIass - E o A , a + 0.059 log [CI] + 0.059 log ycl- 0.059 pH. EcellW - Ecell(2) = 0.059 (PH, - 1% Yat2,) - 0.059 (pH1 - 1% YCl(lJ* The quantity (pH - log ya) is characteristic of each individual solution. It is equivalent to a quantity described and tabulated by Bates' as p(aHyCl) on the molal scale and, in the notation of this paper, amounts to p( [H]yHya).Hereafter it is described as pH'. Then we have If solution 1 is a standard buffer, values of pH' have been evaluated, and are given by Bates,8 for solutions containing no chloride, as p(aHycI)O. We have assumed that the presence of 0-001 M chloride in the standard buffers has not appreciably altered the tabulated values for this quantity, and that the latter could in such dilute solutions be applied on the molar scale. Accordingly, we are in a position to assign values of pH; to other solutions from the e.m.f.'s of the glass - silver chloride electrode cells. In practice, the meters were standardised with the four buffers (a) to (d), and the readings in millivolts were fitted to the best straight line by a least-squares plot of millivolts versus pH'.Sets of solutions differing in [HI, but with a common ionic background, were prepared, and three measurements of pH were made on each with the glass - calomel electrodes and, for the nitrate and perchlorate solutions, three additional measurements of pH' made with the glass - silver chloride electrodes. The range of [HI encompassed by these experiments was limited at the lower end of the concentration scale by evident difficulty in properly defining the actual hydrogen-ion concentration as this became smaller. It was observed that the ratio H/[H], which was manifestly constant within the range reported in this paper, usually fell off to smaller values as the pH approached 4.Hence we have restricted ourselves to reporting only those values for which we felt [HI was unambiguously fixed. It is also evident from our observation that the ratio H'/[H] is constant in a medium of fixed composition and consequently that it would be possible to calibrate the e.m.f. of the cell with glass and silver chloride electrodes directly in terms of colog [HI, without bothering with values of pH'. Plots of e.m.f. versus log [HI consist of a series of parallel straight lines, one for each background electrolyte, with slopes of 59 mV per log [HI at 25" C. The intercepts of such straight lines are governed by the values of Eogiass. (These were remarkably constant with two of our electrodes and drifted progressively with the third.) To avoid any implication that the instruments used were being "rated" we have, with our results that follow, indicated the meters used only by the letters A, B and C, and these are not in the order listed above.342 MCBRYDE: THE PH METER AS A HYDROGEN-ION CONCENTRATION PROBE [Artalyst, Vol.94 RESULTS AND DISCUSSION For each individual set of solutions the pH meter or the millivolt readings, converted into pH', were compared with the corresponding values of -log [HI. In each instance a least-squares fit to a straight line of the form was calculated, and the results are given in Tables I and 11. These include the standard deviation in pH or pH' about the straight line represented. Values are given for each meter assembly separately and a composite set for each group of measurements on the solutions with a certain background composition.pH or pH' = slope (-log [HI) + intercept TABLE I RELATIONSHIP BETWEEN APPARENT pH AND -LOG [HI WITH GLASS AND CALOMEL ELECTRODES pH = m (-log [HI) + b Ionic strength of background, M Nitrate solutions (number of & determinations) Meter m 10ab U ~ H A 0-983 B 0-996 0.998 C 0.972 0.951 Mean 0.980 A 0.982 B 0.994 0.998 C 0.976 1.029 Mean 0.996 A 0.988 B 1.001 1.000 C 0.985 0.983 Mean 0.991 A 0.992 B 1-002 C 0.978 Mean 0.990 A 0.996 B 0.999 c 0.989 Mean 0.996 A - B - c - Mean - - - 10.83 7.97 8-90 13-00 13.69 10-46 11.40 9-37 7.81 12.37 3.00 8.79 11.61 9-23 8.28 11.09 9.40 9.92 9.40 8.33 11-75 9-82 6-96 4.6 1 6.7 1 6.43 - - - - - - 0.01 1 0.004 0.006 0-016 0.004 0.020 0.013 0.004 0.007 0.01 8 0.029 0.02 1 0.01 1 0.004 0.006 0.012 0.013 0.016 0.005 0.003 0.020 0.016 0.010 0.004 0.018 0.014 - - - - - - Perchlorate solutions Chloride solutions 1 - m 0.981 0.984 0.983 0.989 0,983 0.994 0-986 0.978 0-994 0-976 0.991 0.981 0.994 0.986 0.983 0.9'79 0.983 0.982 0.990 0.994 0-996 0.993 1.003 0.999 1.001 1.001 0.990 0.992 0.996 0.992 - - 1 Oab 8.63 9.30 7.74 7.90 8.14 9.10 8.36 9.90 6-88 10.39 8.44 9.40 6.61 8.42 8.83 9.28 8.48 8-88 2.9 1 2-28 1.93 2.37 - 9.07 - 8.18 - 8.91 - 8.72 -60.17 - 60.10 - 60.97 - 60.42 - - *RE m 0-012 0.993 0.009 - 0.011 0.988 0.009 - 0.012 0.990 0.010 - 0.017 0.990 0-016 0.997 0.010 - 0.016 0-992 0.012 - 0.015 0-994 0.016 - 0.016 0.994 0.013 1.000 0.016 0.990 - .- 0.017 0-014 0.011 0.012 0.01 1 0.01 1 0.016 0.020 0.018 0.017 0.020 0.017 0.0 I6 0.018 - 0.994 0.996 0-987 0.989 0-986 0.987 0.996 0.994 0.999 0.996 - - - - - 10ab 8.32 9.32 8.70 8-78 10-04 11.03 10.31 10.48 8.64 10.12 9.64 9.67 11.06 10.66 11.42 11-01 4.90 6-09 4.09 4-69 - - - - - - - - - - - *pH 0.007 0.006 0.006 0.006 0.008 0.006 0.012 0.009 0.008 0.008 0.008 0.008 0.008 0.005 0.010 0.008 0.009 0.005 0.009 0,008 - - - - - - - - - - - - The payit may justifiab-y be mac ! that 1 e standard deviation shown in Tables I and I1 has been assigned entirely against the pH measurements by treating the nominal concen- trations of acid in all solutions as correct and, because of this, aPH may appear larger than it should be.It would require experiments of a different design to separate the various sources of error in this instance.However, in evaluating equilibrium constants among ionic reactions by the procedures previously referred to, one would normally insert the nominal values of "known" concentrations in equations such as (6) or (7), and attribute experimental uncertainties to the "measured" concentrations.May, 19691 MCBRYDE : THE PH METER AS A HYDROGEN-ION CONCENTRATION PROBE 343 TABLE I1 RELATIONSHIP BETWEEN APPARENT pH' AND -LOG SILVER CHLORIDE ELECTRODES pH' = m (--log [HI) + b Ionic strength of background, M (number of determinations) Meter A 0-06 (9) B C Mean A B C Mean A B C Mean A B C Mean A B C Mean A B C Mean Nitrate solutions P m 1.013 1-008 1-003 1.007 1.020 1.010 1.007 1.008 1.013 1.022 1.019 1.014 1.012 1.002 1.008 1.008 1.017 1.009 1.008 1.012 1.012 1.010 1.016 1.018 1.021 1.018 - - - - - - - 10ab 16.69 16.80 16-87 16-70 12.15 16.44 20.53 18.89 19-22 19.64 18.43 19.34 24-60 26-03 24.70 26.99 24.73 26.19 29-38 28-02 29.77 29.06 28-47 26.99 26-99 27.48 - - - - - - - - - UP= 0.012 0.003 0.006 0.006 0.010 0.01 1 0.004 0.004 0.006 0.011 0.008 0.014 0.007 0.006 0.008 0.004 0.013 0.014 0.003 0.003 0.016 0.01 1 0.009 0.007 0.01 1 0-010 - - - - [HI WITH GLASS AND Perchlorate solutions r ffl 1.007 1-015 1.010 1.01 1 1.000 1.007 1.008 1.006 1.020 1.010 1.019 1.026 1.026 1.017 1.004 1-012 1.019 1.012 1.018 1.016 1.013 1.016 1.030 1.029 1.030 1.029 1.013 1.032 1.020 1.026 1.023 1.030 1 -024 - - 10ab 16.70 13-96 12-60 14-72 17.70 18.12 15-63 20.58 16-85 18.46 17-04 16-41 16.26 17.60 23.34 20.78 20.35 21.49 22-64 21-06 22-19 21.96 14.64 12.92 12.78 13.45 - - - 32.63 - 33.90 - 33.90 - 34.69 - 34-34 - 31.06 - 33.40 0.002 0-002 0-003 0-003 0.010 0.006 0.016 0,002 0.006 0.004 0.006 0.013 0.010 0.010 0.004 0.002 0.010 0.009 0.006 0.004 0.007 0.010 0.01 1 0408 0.013 0.014 0.006 0.013 0.006 0.007 0.006 0.006 0.020 - - The relationship between H or H' and [HI, as applicable, was obtained for each set of solutions by finding, again by a least-squares treatment, the coefficients for a linear equation of the form for the solutions within that set.It is true that the quantity b in Tables I and I1 is close to -log (H/[H]) or -log (H'/[H]) as m is about 1, but it seemed preferable to evaluate the ratios indwidually to estimate with what precision they could be found from the pH readings. Values of the ratio I'= H/[H] or r' = H'/[H] for the various ionic media and electrode pairs are shown in Tables I11 and IV.Further studies of the application of the silver chloride electrode in conjunction with the glass electrode for the study of equilibria in complex-ion formation are in progress. It should be acknowledged that studies such as those of Lumb and Martellg have already demonstrated the feasibility of using this cell, although in solutions containing much greater chloride-ion concentration than proposed here. First, as the results in Tables I and I1 imply that pH or pH' values measured on a single instrument are subject to a standard deviation often in excess of 0.01 unit, this uncertainty is transmitted to the value of r or r'. Deviations in pH that appear relatively innocuous H = F[H]344 MCBRYDE: THE PH METER AS A HYDROGEN-ION CONCENTRATION PROBE [ATdySt, VOl.94 produce significant variations in r, and the values in Tables I11 and IV reveal the uncertainties to which careful evaluation of this ratio is subject. In the author's laboratory it has been the practice for several years to establish the ratio I' for particular background salt systems in which ionic equilibria are being studied, and it may be helpful to mention that the results presented here, which were collected within one special study, are satisfactorily corroborated by other accumulated results. TABLE I11 VALUES OF r= H/[H] WITH GLASS AND CALOMEL ELECTRODES Ionic strength of background, M (number of determinations) 0.05 (9) 0.20 (11) 0.50 (11) Meter A B C Mean S.E.* A B C Mean S.E.* A B C Mean S.E.* A B C Mean S.E.* A B C Mean S.E.* A B C Mean S.E.* Nitrate solutions & ll 0-814 0.843 0.856 0.794 0.864 0.834 0.777 0.813 0.830 0.765 0.808 0.799 0-777 0.801 0.819 0.782 0.819 0.800 0.819 0.819 0-765 0.801 0.865 0.899 0.860 0.875 - - - - - - - - - - - - or 0.046 0.016 0.013 0.078 0.097 0-060 0.009 0.068 0.018 0.0 18 0.094 0.064 0-05 1 0-007 0.043 0.009 0.013 0.058 0.064 0.043 0.006 0.024 0.006 0.097 0.045 0.008 0-036 0.009 0.077 0-043 0.007 - - - - - - - Perchlorate solutions - J 0.867 0.837 0.867 0.850 0.868 0.812 0.847 0.808 0.847 0.803 0-829 0.81 1 0.864 0.827 0.823 0.818 0.821 0.821 - - - - - 0-934 0-933 0.941 0.936 1.171 1.150 1-166 1.162 3.102 3.099 3.129 3.1 10 - - - or 0.063 0.048 0.058 0.040 0.060 0-032 0.055 0.007 0.088 0.041 0.092 0.041 0.083 0.057 0.067 0.008 0.072 0.085 0.081 0.077 0.013 0.054 0-056 0*060 0.051 0.009 0.068 0.086 0-076 0.074 0.013 0.289 0.247 0.220 0.245 0.043 - I Chloride solutions & J 0-837 0.836 0-848 0.840 0.789 0.786 0.789 0.787 0.807 0-791 0.805 0.801 0-792 0-803 0.781 0-792 - - - - - - - - - - - - 0.886 0-895 0.911 0.897 - - - - - - or 0.027 0.028 0.021 0.024 0-005 0.024 0.025 0.034 0.027 0.005 0.021 0.034 0.027 0.027 0.005 0.041 0.027 0.049 0.038 0.007 0.034 0.024 0.019 0.023 0.004 - - - - - - - - - - - - - * S.E.indicates the standard error of the mean of all determinations in a given category = O / ~ K where n determinations are averaged. It is perhaps worth drawing attention to the fact that the most consistent estimates of r, and those with generally the best precision, were found in potassium chloride solutions. The special advantage of this salt in minimising liquid junction potentials, associated with the about equal transference numbers of its ions, may have contributed to this observation.The worst precision among measurements with the calomel electrode was encountered with sodium perchlorate solutions, in spite of the fact that, by using the external sleeve on the reference electrode described above , pH readings of remarkable stability were encountered with these sets.May, 19691 MCBRYDE: THE PH METER AS A HYDROGEN-ION CONCENTRATION PROBE 345 The device of using a silver chloride reference electrode, with the addition of a constant low concentration of chloride ion to the test solutions, brought about a conspicuous improve- ment in the precision with which I” could be evaluated relative to the determination of r.It may be inferred that elimination of the liquid junction potential has a beneficial effect on measurements of this type, especially when the background salt is not potassium chloride. For investigations of ionic equilibria in which the presence of chloride at such a level of concentration can be tolerated, this electrode, used in this simple manner in conjunction with a glass electrode, may prove to possess advantages over the calomel electrode. TABLE IV VALUES OF r’ = H‘/[H] WITH GLASS AND SILVER Ionic strength of background, M (number of determinations) Meter A 0.05 (9) B C Mean S.E.* A 0.10 (11) 0.20 (11) 0-50 (11) B C Mean S.E.* A B C Mean S.E.* A B C Mean S.E.* A B C Mean S.E.* A B C Mean S.E.* Nitrate solutions & P 0.667 0.680 0.666 0.666 0.705 0.677 0.61 1 0.635 0.622 0.603 0.633 0.621 0.548 0.563 0.555 0.529 0.541 0-547 0.499 0.508 0.493 0.500 0.503 0.507 0.505 0.505 - - - - - - - - - - - - - - - or‘ 0.026 0.013 0.010 0.014 0-033 0.02 1 0.003 0.01 1 0.016 0.022 0-039 0.039 0.030 0.004 0.016 0.009 0.014 0.014 0-028 0.022 0-003 0.010 0.015 0.024 0.019 0.003 0.023 0.020 0-027 0.023 0.004 - - - - - I - - - - CHLORIDE ELECTRODES Perchlorate solutions - 1’ 0-666 0.685 0-72 1 0.684 0.668 0.638 0.677 0.6 10 0.638 0-638 0.641 0.647 0.637 0.635 0.578 0-602 0.591 0.590 0.566 0-588 0-676 0.676 0.651 0.681 0.678 0.670 2.053 1-988 2.082 2-068 2-092 1.869 2-024 - - - - - - - - or’ 0.01 1 0.019 0.014 0.015 0.015 0.009 0.026 0.004 0.009 0.029 0-017 0.029 0.048 0.037 0-029 0.004 0.008 0.019 0.029 0-020 0-003 0-026 0.023 0-018 0.024 0-004 0.042 0.043 0-046 0-044 0.008 0.077 0.148 0-117 0-128 0-137 0.127 0-142 0.017 - - * S.E.indicates the standard error of the mean of all determinations in a given category = 0/2/< where n deterxninations are averaged. ACIDITY CONSTANT FOR SULPHOSALICYLIC ACID- The principles embodied in this paper were used in a determination of the second-stage ionisation constant of sulphosalicylic acid (carboxyl hydrogen) in M salt solutions. Solutions were prepared in calibrated flasks containing the acid at a concentration CA of 9-989 x M, sodium hydroxide at concentrations CB of integral multiples from 2 to 12 of 1.284 x M346 MCBRYDE and, as background electrolyte, potassium chloride, potassium nitrate or sodium perchlorate at a concentration of 1 M.The pH of all these solutions was measured with the glass and calomel electrodes. For the latter two background solutions, to which sodium chloride was also added to give a molarity of 1.0 x lo-*, the pH‘ was also measured with the glass and silver chloride electrodes. Concentration quotients were calculated from all sets of measurements by the following relationship, which applies to this particular acid- For each calculation, values of r or r’ taken from Tables I11 or IV were used to convert H or H’ values for use in equation (8). The following is a summary of the results. Background electrolyte Reference electrode PQn (26’ C) I Potassium chloride Calomel 2.23 f 0.02 Potassium nitrate Calomel 2-24 f 0.01 Silver chloride 2.26 f 0.01 Sodium perchlorate Calomel 2-26 f 0.01 Silver chloride 2.31 f 0.01 Mean .... .. .. .. . . 2.26 f 0.02 It is evident that there is satisfactory concordance among the values for the concen- tration quotient determined in different background solutions and with different reference electrodes. Had the results been calculated in the form of mixed quotients incorporating values of H or H’, significant differences would have appeared among them. With those solutions for which reasonable estimates of activity coefficients are possible, and in which liquid junction potentials may be assumed of comparable magnitude to those arising in the standard buffer solution, mixed quotients can be converted by calculation into thermodynamic constants, a step that reduces all estimates to a common basis. In the absence of this possi- bility, conversion into concentration quotients seems to provide the only rational basis for comparing and recording information on ionisation equilibrium. The mixed or “practical” constants can only be justified as means to such ends, but have little intrinsic significance. The author gratefully acknowledges the financial assistance of a grant from the National Research Council of Canada, and the technical assistance of Miss Janet Rohr. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. Bjermm, J., “Metal Ammine Formation in Aqueous Solution,’’ P. Haase and Son, Copenhagen, 1941. Ringbom, A,, “Complexation in Analytical Chemistry, Chemical Analysis,” Volume 16, Inter- science Publishers, New York and London, 1963, p. 36. Brhsted, J. N., Chem. Rev., 1928, 5, 293. Bates, R. G., “Determination of pH: Theory and Practice,” John Wiley and Sons Inc., New York, 1964, Chapter 4. King, E. J., “Acid-Base Equilibria,” Pergamon Press, Oxford, 1966, Chapter 4. Irving, H. M., and Rossotti, H. S., J. Chem. SOC., 1954, 2904. Bates, R. G., op. d t . , p. 26. - , 09. cit., p. 409. Lumb, R. F., and Martell, R. E., J. Phys. Chem., 1953, 57, 690. Received Se9bember 26th, 1968 Accepted January loth, 1969

ISSN:0003-2654    

DOI:10.1039/AN9699400337

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

A~zalyst, May, 1969, Vol. 94, PP. 347-353 347 Preparation and Spectral Characteristics of Microwave-excited Electrodeless Discharge Tubes for Palladium, Silver, Platinum and Gold BY K. M. ALDOUS, R. M. DAGNALL AND T. S. WEST (Chemistry Depadment, Imfierid College, London, S . W.7) Electrodeless discharge tubes for palladium, silver, platinum and gold are prepared from the corresponding element and chlorine at a pressure of about 3 torr of carrier gas. This method of preparation allows the #-wave type of cavity to be used, which results in longer tube-life and increased stability of emitted radiation. The spectral characteristics of these line sources are also examined. THE application of microwave-excited electrodeless discharge tubes as spectral-line sources in atomic-fluorescence and atomic-absorption spectroscopy has recently been the subject of considerable investigation.lp2 Apart from the advantages associated with the spectral charac- teristics of these sources (such as their intense resonance-line emission, absence of carrier-gas background and higher excited lines) , the long life, ease of preparation for almost all elements and, particularly, the low cost of these sources are perhaps the most outstanding and attractive features.The preparation of hollow-cathode lamps for elements such as palladium, silver, platinum and gold can be difficult as well as expensive, and their intensity is such that they are usua.lly unsuitable for atomic-fluorescence spectroscopy. Good results have been obtained with some high intensity hollow-cathode lamps,3s4 but these are considerably more expensive than normal hollow-cathode lamps, and only a few are sufficiently intense to be of use for atomic-fluorescence spectroscopy.The comparative involatility of these noble metals and their compounds indicates that previously published preparation methods5 96 for electrodeless discharge tubes will not produce satisfactory sources. Tubes prepared from the elements alone give no emission apart from the carrier gas and they merely “plate out.” The metal fihs iodine6 method of preparation was attempted in this study, but the iodides are either difficult to prepare (e.g., gold) or they are too involatile (e.g. silver), and some molecular metal iodide band emission is observed at normal operating powers. The need to operate such tubes at high temperatures and, therefore, at high powers results in the tube-life often being short because of the tendency for the compound to dis- sociate irreversibly.The use of the #-wave (Evenson) type of cavity5 accelerates this behaviour. More recently we have tended to standardise on the 9-wave (Broida) cavitys because its tuning characteristics are not critical and because it acts as a thermostat for the discharge tube, which generally results in increased stability. This type of cavity is less efficient than the &wave cavity, and it therefore necessitates the use of a different set of preparative conditions. This paper describes, with special reference to palladium, silver, platinum and gold, a general method of preparing electrodeless discharge tubes from the corresponding metal chlorides, which are suitable for the $-wave cavity.We have found this particular cavity to be useful in atomic-fluorescence spectroscopy because it prevents radiation from the source from being directly reflected into the monochromator, thus resulting in improved limits of detection. Although it is much larger than the &wave cavity, we have had no difficulty in using it for atomic-absorption and atomic-fluorescence studies in most types of commercial atomic-absorption equipment. More recently, this cavity has been successfully used in conjunction with electronically modulated discharge tubes.’ 0 SAC and the authors.348 ALDOUS et aE. : PREPARATION AND SPECTRAL CHARACTERISTICS [Analyst, Vol. 94 The discharge tubes were prepared from transparent quartz tubing of 8 mm internal diameter, in the same manner as previously de~cribed.~ Surface etching of the quartz tubing, which was often experienced in the past, was removed by heating the quartz stubs to white heat in a Scorah oxygen - coal-gas burner.Only the significant differences in the methods of preparation for individual elements are given below. Spectral evaluation was assessed with a Unicam SPSOOA flame spectrophotometer, used in the “emission” mode of operation in conjunction with an E.M.I. 9601B photomultiplier and a Servoscribe pen-recorder. Microwave excitation was achieved as usual with a “Microtron 200” generator (Electro- Medical Supplies Limited) of output 2450 25 MHz, up to 200 watts, connected via a reflected power meter to either a $-wave (EMS 210L) or &wave (EMS 214L) type cavity.Initiation was obtained by using a “Tesla” vacuum tester. PALLADIUM- Palladium discharge tubes were made by introducing about 5mg of palladium metal (as foil) into a quartz bulb about 4 to 5 cm in length. The bulb and metal were de-gassed at red heat under vacuum with an oxy-propane hand torch. [Palladium has a high melting- point (1552” C), so that vacuum evaporation does not occur]. The tube was removed from the vacuum line, and chlorine gas from a cylinder was then introduced at atmospheric pressure into the quartz bulb. The metal chloride was formed with a minimum of heat to prevent loss of the chloride by volatilisation, and the chlorination process was stopped while there was still an excess of palladium metal in the quartz bulb.The tube was then cooled, re-evacu- ated and any free chlorine flushed out with argon. The tube was then sealed under an argon pressure of 3 torr. Tubes produced in this way initiated readily when placed in either of the microwave cavities used in this study. The optimum excitation powers were about 40 watts in the &wave cavity and about 60 watts in the $-wave cavity, with 2 and 5 watts reflected power, respectively. The higher power level required with the latter cavity arises because of its less efficient nature (lower Q value). EXPERIMENTAL TABLE I PALLADIUM DISCHARGE-TUBE SPECTRUM* Wavelength, nm Recorder readingt 244.7$ 18 247*6$ 16 276.3: 12 302.7 3 306-5 7 31 1.4 6 324.2: 63 326.1 14 325.8 7 330.2 24 337.3 11 342.1 20 343.3 22 344.1 10 346.0 13 348.1 14 348.9 12 351.6 37 356-3 28 367.1 9 360*9$ 62 363*4# 92 340.4# 100 * Tube operated at 40 watts, with 2-watts reflected power t Slit width 0-015 mm and gain 3,O; uncorrected for instru- Lines recommended for atomic-absorption spectroscopy.in the &wave cavity. ment response characteristics.May, 19691 OF MICROWAVE-EXCITED ELECTRODELESS DISCHARGE TUBES 349 Under these conditions, and in the absence of cooling, a stable, intense discharge showing no background emission from either argon or chlorine was obtained. The colour of the discharge was blue, and observation about 10cm from the entrance slit of the Unicam SPSOOA flame spectrophotometer showed that it contained all of the expected atomic lines for palladium, their relative intensities being similar to those obtained in the arc, rather than spark, spectrum.* 100 M - 73 2 3 0) 50 v) 0 Wavelength, nm Fig.1. Emission spectrum of palladium discharge tube. Excitation at 2460 MHz and 40-watts power in the )-wave cavity. Recorded on Unicam SPSOOA spectrophotometer with slit width 0.016 mm and gain 3,O Table I and Fig. 1 show the principal spectral lines that are to be observed near those lines which have been recommended for atomic absorption. I 1 40 0 10 20 30 I 40 Time, minutes Fig. 2. Stability and warm-up period. (a), Palladium A denotes discharge tube; and (b), platinum discharge tube. power on and B power off The discharge tubes required an initial “running-in” period of about 1 hour before good stability was reached.After this they would warm up in about 30 minutes, although this could be reduced to between 10 and 15 minutes by gently wanning the tube before initiation [Fig. 2 ( a ) ] . The stability with the $-wave cavity showed less than +2 per cent. variation in response, while with the &-wave cavity it was about +5 per cent. This indicated that better thermal stability was obtained in the former cavity because of the delocalised micro- wave field. Tubes prepared from palladium and iodine produced a stable, but very weak, discharge.360 [Artalyst, Vol. 94 SILVER- A quartz bulb, about 4 cm in length, was de-gassed at red heat under vacuum. When cool, about 6 mg of silver powder were introduced and, after flushing with argon, the tube was re-evacuated and the metal evaporated, by using an oxy-propane hand torch, to form a mirror on the inside walls of the quartz tube.This had the combined effect of de-gassing the metal and also providing a large surface area for chlorination. The quartz tube was then removed from the vacuum line and the metal chloride formed by direct combination with chlorine gas at atmospheric pressure. When some of the chloride had formed and while there was still an excess of metal, chlorination was stopped and the tube was then flushed with argon to remove free chlorine. The tube was finally sealed under an argon pressure of about 3 torr. The silver discharge was purple in colour, and was best operated in the $-wave cavity at about 40 watts, without cooling. The stability of these tubes was good and they possessed a clean spectrum free from band emission [Fig.3 (a)]. ALDOUS et al. : PREPARATION AND SPECTRAL CHARACTERISTICS - .b% 0 0 N N A, Wavelength, nm r" I I 30 60 Time, minutes Fig. 3. Emission characteristics of silver discharge tube: (a), emission spectrum; and (b), stability and warm-up period. A denotes power on. Excitation at 2460 MHz at 40-watts power in the &wave cavity. Measurements made on Unicam SPSOOA spectrophotometer with slit width 0.016 mm and gain 3,6 A silver discharge tube was also made in the usual manner-S with a sub-stoicheiometric amount of iodine. Although this initially gave an intense emission, after running for about 2 hours the emission from silver iodide molecular species caused high background radiation around the silver resonance line at 32800nm.TABLE I1 SILVER DISCHARGE-TUBE SPECTRUM* Wavelength, nm Recorder readingt 260-1 (1 260.9 <1 328*0$ 100 338-2$ 60 Tube operated at 40 watts, with 3-watts reflected power t Slit width 0.016 mm and gain 3,6; uncorrected for response $ Lines recommended for atomic-absorption and atomic- The resonance lines at 206.1 and 206.9 nm were observed by using the Unicam SPSOOA spectrophotometer but, mainly because of a poor photomultiplier response at these wave- lengths, their intensities were only weak in both metal chloride and iodide tubes (Table 11). in the *wave cavity. characteristics of Unicam SPOOOA spectrophotometer. fluorescence spectroscopy.May, 19691 OF MICROWAVE-EXCITED ELECTRODELESS DISCHARGE TUBES 351 Silver electrodeless discharge lamps, prepared as the chloride, required a “running-in” period of about 2 hours when first prepared.After this, their performance was reproducible and stable, and they required only a short stabilisation period of about 6 minutes [Fig. 3 (b)]. PLATINUM- Platinum is a very unreactive metal and will not undergo direct halogenation with iodine, even at high temperatures. Also, its reaction with chlorine is slow, and it was found difficult to produce much of the chloride by using fine platinum wire under the conditions previously described. Because platinum has a high melting-point (1769” C), it does not form a mirror under these conditions, and hence an alternative method was used to give a platinum mirror that would facilitate the above chlorination procedure.This was based on the observa- tion that the platinum chlorodiammine complex [Pt(NH,),CI,] on heating under vacuum deposited a platinum mirror on the quartz tube walls. The ammonium chloride formed could be sublimed out of the tube by using an oxy-propane hand torch. The following procedure proved successful. After de-gassing a tube, 4 to 5 cm in length, at red heat, about 10 mg of Pt(NH3),CI, were introduced into the tube and a platinum minor formed by heating it under vacuum wth a micro burner. The metal mirror was then de-gassed at red heat, and the large surface area of metal produced in this way was then chlorinated at atmospheric pressure, and at red heat, with chlorine gas from a cylinder. The process of chlorination was not allowed to proceed to completion, being stopped while there was still an excess of metal in the tube. Care was taken not to volatilise the chloride out of the tube during heating. The tube was then re-evacuated and flushed with helium to remove excess of chlorine and sealed under a helium pressure of about 1 torr.Helium was used as carrier gas because it gave a lower background emission than argon around the platinum resonance line at 265-9nm. Also, the higher ionisation potential of helium in comparison with argon caused a more intense emission of platinum resonance lines. Platinum tubes made in this way were examined with the Unicam SP9OOA spectrophoto- meter, as before, and were found to give all of the expected atomic lines for platinum (Fig. 4 and Table Wavelength, nm Fig.4. Emission spectrum of platinum discharge tube. Excitation at 2450 MHz at 40-watts power in the f-wave cavity. Measurements made on Unicam SPSOOA spectrophotometer, with slit width 0.02 mm and gain 3,5 The optimum running conditions were 40 watts in the &wave cavity (reflected power 2 watts) or 50 watts in the Q-wave cavity (reflected power 10 watts) ; the stability was slightly better in the latter cavity (less than +2 per cent. variation in signal response). Cooling the tube via a compressed air line caused a reduction in the intensity of the platinum emission and a corresponding increase in helium background emission. The colour of the discharge was orange-red, and the tube required a “running-in” period of about 2 hours when first prepared.When the tube was gently heated in a bunsen flame it subse- quently stabilised in about 15 minutes [Fig. 2 ( b ) ] .362 loo M 3 3 f 50- v) 0 ALDOUS et aZ. : PREPARATION AND SPECTRAZ. CHARACTERISTICS [Analyst, Vol. 94 TABLE I11 PLATINUM DISCHARGE-TUBE SPECTRUM* Wavelength, nm Recorder readingt 2 18.2 6 248.7 13 262.8 16 263-9 6 264.6; 24 265-9t 83 267.7 6 271.1 76 273.3 53 283.0 32 292.9 50 299.7 # 76 304-2 34 313.9 66 33 1-5 23 300.4: 100 * Tube operated at 40 watts, with 2-watts reflected power t Slit width 0.02 mm and gain 3,6; uncorrected for instru- $ Lines recommended for atomic-absorption spectroscopy. in the +-wave cavity. ment response characteristics. z rs - (4 I” a 9 x 8 I I /LJ\ 0 30 60 GOLD- Gold electrodeless discharge tubes were made in a manner similar to that for silver.A quartz bulb, 4 to 5 cm in length, was de-gassed at red heat under vacuum, and gold wire (about 6mg) was introduced and vacuum-evaporated on to the quartz tube walls. This effected the de-gassing of the metal and gave a large surface area for formation of the chloride. The chlorination procedure was performed at red heat at atmospheric pressure, and was allowed to continue until a sub-stoicheiometric amount of chlorine had reacted with the gold metal. The tube was then re-evacuated, flushed with argon to remove unreacted chlorine and finally sealed under an argon pressure of about 3 torr. These tubes required only a short “running-in” period of about half an hour, after which time a stable intense emission was obtained under optimum operation conditions of about 40-watts power in the &wave cavity (reflected power 1 watt) or 60 watts in the #-wave cavity (reflected power 5 watts).No cooling of the tube was necessary. Stability was better than +2 per cent. fluctuation in intensity at the resonance line at 267.6 nm in the Q-wave cavity and about +3 per cent. fluctuation in response in the &wave cavity.May, 19691 OF MICROWAVE-EXCITED ELECTRODELESS DISCHARGE TUBES 353 The spectrum obtained with the Unicam SPSOOA flame spectrophotometer was found to contain the expected gold atomic lines, together with mercury lines that arose from the mercury in the vacuostat gauge. Fig. 5 (a) shows the spectrum recorded at slit width 0.02 mm and gain 3,5 in a manner similar to that described above.Both of the gold atomic lines indicated have been recommended for atomic-absorption measurements. Fig. 5 (b) shows the stability and warm-up period for a gold tube operated after warming in a bunsen flame (A), and from cold (B). TABLE IV EFFECT OF MICROWAVE POWER ON EMISSION AT PRINCIPAL RESONANCE LINES Recorder reading* A r \ Power, Palladium Silver Platinum Gold watts (276.3 nm) (328.0 nm) (265.9 nm) (242.8 nm) 20 17 35 15 17 25 20 43 28 20 30 25 51 36 25 35 28 60 45 28 40 37 65 62 37 45 42 73 76 42 50 48 77 80 48 By using Unicam SP9OOA flame spectrometer, with slit width 0.02 mm Table IV shows the effect of microwave power on the emission intensity for all of the tubes examined in this study. Although the intensity of emission can be increased much more than that shown by using powers in excess of 50 watts, it is not considered advisable, because the stability of the emission is then less than can be tolerated in atomic-absorption spectroscopy.This problem is, however, not quite so severe at the level of the limits of detection in atomic-fluorescence spectroscopy. and gain 3,O and with the )-wave cavity. CONCLUSIONS Microwave-excited electrodeless discharge tubes can now be prepared for a wide range of elements, either from the element plus chlorine, as described above, or from the element alone or element $Zus iodine.6 Generally the metal iodide should be used, except in the few cases in which it niay be more convenient to use the metal itself, either because it is sufficiently volatile or because the iodide undergoes irreversible dissociation within the tube.In the latter instance, chlorides appear to be more suitable because of the tendency for chlorine and the metal to re-combine. The &wave cavity gives best results for tubes with a filler gas pressure of 1 torr and at a comparable power yields more intense spectra than the 2-wave cavity. A filler gas pressure of 3 tom is better for tubes operating in the $-wave cavity. Because of the uniform nature of the heating, tubes operated in the latter cavity show better stability, and less volatile compounds tend to give more stable output. In addition, because of the de-localisa- tion of the microwave field in the &wave cavity there is less tendency for hot spots and plating out to occur. Tubes containing very volatile elements or compounds can be excited in either cavity, provided sufficient cooling is supplied to control the vapour pressure within the tube. One of us (K.M.A.) thanks I.C.I. Agricultural Division, Billingham, for the award of a research grant to carry out this work. 1. 2. 3. 4. 5. 6. 7. 8. REFERENCES Dagnall, R. M., and West, T. S., A@$l. Opt., 1968, 7 , 1287. Mansfield, J. M., jun., Bratzel, M. P., jun., Norgordon, H. O., Knapp, D. O., Zacha, K. E., and Winefordner, J. D., Spectrochim. Acta, 1968, 23B, 389. West, T. S., and Williams, X. K., Analyt. Chem., 1968, 40, 335. -- , Analytica Chim. Acta, 1968, 42, 29. Dagdall, R. M., Thompson, K. C., and West, T. S., Talanta, 1967, 14, 551. Dagnall, R. M., Pfibil, R., jun., and West, T. S., Analyst, 1968, 93, 281. Browner, R. F., Dagnall, R. M., and West, T. S., Analytica Chim. Ada, 1969, 45, 163. Brode, W. R., “Chemical Spectroscopy,” Second Edition, J. Wiley and Sons Inc., New York, 1943. Received Sefitember 19th. 1968 Accepted November 12th, 1968

ISSN:0003-2654    

DOI:10.1039/AN9699400347

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

354 Analyst, May, 1969, Vol. 94, $$. 3M-368 Contributions to the Study of Functional Analytical Groups in the Determination of Palladium (II)* BY GH. BAIULESCU, C. GREFF AND F. DANET (Defiartmenf of Analytical Chemistry, University of Bucharest, Romania) A bisazoic derivative of chromotropic acid has been synthesised on the basis of a study of functional analytical groups in the determination of palladium(I1). The reagent thus obtained gives a very sensitive reaction with palladium(II), which enables between 0.1 and 4 pg of palladium per ml to be determined, both alone and in the presence of some other elements. Palladium(I1) combines with the reagent in the ratio 1 : 2, and the instability constant of the compound is 2.89 x lo-'. PREVIOUS investigations have shown that palladium( 11) reacts with organic reagents con- taining, as distinct from others, the two general types of analytical functional groups (A and B) shown below. and X where X represents -NO (ref.I) or -N=N-(ref. 2) where X represents -OH (ref. 2) or -COOH (ref. 2) It was established2 as early as 1959 that twenty-two organic compounds containing such groups react with palladium(II), and four of them, namely, benzyl orange, methyl orange, magneson I and methyl red, were proposed as reagents for its spectrophotometric deter- mination. In subsequent investigations other reagents containing such groups have been suggested for palladium( 11) determination, including Tropeolin 0 and Tropeolin 00.3 Pyrazolone derivatives also used were tartrazine,4 l'-phenyl-3'-methylpyrazole-5'-one-4'-yl- azobenzene-4-hydroxy-3-carboxylic acids and the potassium salt of 1 '-phenyl-3'-methyl- pyrazole-5'-one-4'-yl-azobenzene-2-hydroxy-5-sulphonic acid.6 All of these reagents contain the type B group.Among the derivatives of chromotropic acid used as reagents for palladium(I1) were arsenazo I,' mono and dinitrosochromotropic acids8 and Chromotrope 2Rmg * Paper presented at the Second SAC Conference 1968, Nottingham. 0 SAC and the authors.BAIULESCU, GREFF AND D ~ E T 355 As a result of recent work, 2,7-bis-(o-sulpho-~-methylphenylazo)-chromotropic acid (di- methylsulphonazo 111), I,l0J1 2,7-bis-(o-sulpho-~-bromophenylazo)-chromotropic acid, 11,1@ 2,7-bis-( o-sulpho+-nitrophenylazo) -chromotropic acid, I1 I,l0 2,7-bis- (0-sulpho-9-acet ylamino- pheny1azo)-chromotropic acid, IV,lO palladiazo, V,12 and arsenazo 111, VI,uJ4 have been proposed as reagents for palladium(I1).OH OH H,C ~ ~ N - ~ J ( ~ J - N ~ ~ ~ C H , H 0 3 S SO,H B r d H fi I H o b N=N / - \ N=N Br \ / HO,S S03H II OH OH HO~S SO,H V VI Only dimethylsulphonazo 111, I,11 palladiazo, V,12 and arsenazo 111, VI,13J4 have been used for the spectrophotometric determination of palladium(I1). Reagents V and VI have proved to be very sensitive for palladium(I1) determination. An examination of these reagents shows that they contain type B functional analytical groups, although the above papers make no mention of this.356 BAIULESCU et al.: CONTRIBUTIONS TO THE STUDY OF FUNCTIONAL [A?Za&St, VOl. 94 As it was observed2 that reagents containing both type A and B groups in their molecule, e.g., methyl red, give very sensitive reactions with palladium(I1) it was decided in the present work to synthesise a chromotropic acid derivative containing both group types within its molecule, the formula for which is shown below, VII.OH OH (H5C3,N = N N = N 0 N (C2H5), 7 - / \ NaO$ S03Na VII SYNTHESIS OF THE REAGENT- NN-diethyl-$-phenylenediamine sulphate (5.25 g; 0-02 mole) is dissolved in 20 ml of water containing 0-5 ml of concentrated sulphuic acid. The solution thus obtained is cooled in an ice-bath until it reaches a temperature of 0" to 5" C; 1.4 g (0.02 mole) of sodium nitrite dissolved in 5ml of water are added, dropwise, to ensure that the temperature does not rise above 5" C. The solution is left for 5 to 10 minutes at a low temperature.A solution containing 3-65 g (0.01 mole) of chromotropic acid (disodium salt) and 1.2 g (0.03 mole) of sodium hydroxide in 30 ml of water is also prepared and cooled to 5" C. To this solution, the solution of the diazonium salt previously prepared is added slowly and with constant stirring, the temperature being kept constant at 5" C. The mixture becomes dark blue and is maintained for 1 hour in the ice-bath, with intermittent stirring. The mixture thus obtained, which contains the maximum amount of dye in solution, is heated to boiling and saturated with sodium nitrate by addition of solid sodium nitrate. On cooling the dye crystallises out, together with sodium nitrate. It is then purified by recrystallisation from absolute ethanol.Black - blue crystals of the dye are thus obtained; the yield is 45.5 per cent. METHOD REAGENTS- Palladium chloride solution-A solution of palladium chloride, PdCl,, containing 100 pg of palladium per ml was prepared, the concentration of which was determined gravimetrically with dime th ylglyoxime. Reagent solution, 0-4 per cent., aqueous-This solution is stable for at least 1 month. All batches of reagent used were checked for purity by elemental and chromatographic analysis. Borate bufer solution, PH 9.24. Wavelength, nm Fig. 1. Absorption spectra: A, of reagent; and B, of the palladium(I1) compound PROCEDURE- Study of extinction variation wavelength function-B y using a Beckman spectrophoto- meter, Model DB, absorption spectra were recorded (Fig.l), both of the reagent and of the palladium(I1) compound obtained in the presence of borate buffer at pH 9.24.May, 19691 ANALYTICAL GROUPS I N THE DETERMINATION OF PALLADIUM(I1) 357 Examination of the spectra shows that the maximum absorption of the palladium(I1) compound obtained with reagent VII occurs at 695nm, and this wavelength was used throughout the present work. Study of extinction variation furtction of $alladium(II) cmcentration-In a study of the influence of various parameters of the palladium(I1) reaction with reagent VII, it was found that the best results are obtained by using a borate buffer at pH 9-24; and the reaction takes place at 80" to 90" C. To establish the concentration range in which the reaction obeys the Lambert - Beer law, the following procedure was adopted.W Palladium (I I ) per ml, pg Fig. 2. Extinction dependency on palla- dium (11) concentration Amounts of palladium chloride containing between 2.6; and 1OOpg of palladium(I1) are introduced, by pipette, into 25-ml graduated flasks; 1 O m l of sodium borate buffer solution (pH 9-24) and 4 ml of reagent solution (0.4 per cent.) are added. The volume is then made up to 25 ml with distilled water. The graduated flasks are heated between 80" and 90" C on a water-bath for 10 to 15 minutes, and then cooled to room temperature; determinations are performed at 595 nm, by using a Zeiss-Jena VSU-I spectrophotometer. The results obtained are shown in Fig. 2. NOTE- The conditions given for pH and temperature resulted from trials carried out with various buffer solutions over a range of pH values. An increase in the heating time beyond 16 minutes does not affect the absorption, and no differences were observed whether the samples were heated at 80" or 90" C for 10 or 16 minutes.This diagram shows that the proposed method enables palladium(I1) to be determined at concentrations between 0.1 and 4pg of palladium(I1) per ml, thus surpassing in sensi- tivity the reagents palladiazo and arsenazo I11 (E: = 2.89 x 104). S~ectro~htometric determination of palladium(II) in the presence of $latinum(IV)- Table I shows the results of palladium(I1) determinations in the presence of platinum(1V). TABLE I SPECTROPHOTOMETRIC DETERMINATIONS OF PALLADIUM(II) IN THE PRESENCE OF PLATINUM(IV) The order of addition of reagents is not important.Palladium (11) , Platinum(IV), Ratio of Pg Per ml PI? Per ml E Z palladium(I1) to platinum(1V) - 0.272 - 1 1 20 0-276 1:20 1 26 0.289 1:26 1 60 0.360 1:60 The results given in Table I establish that it is possible by this method to determine palladium(I1) in the presence of platinum(IV), up to the ratio 1 : 20. It may also be observed that, with this reagent, palladium(I1) can be determined in the presence of cobalt up to the ratio 1 : 16, copper 1 : 20, silver 1 : 5 and gold 1 : 2, with an error of less than 2 per cent. Interference by other elements has not been examined.358 BAIULESCU, GREFF AND DANET Determination of combination ratio and of the instability constant of the falladium(1l) - reagent VII compound-By using Job’s method of continuous variation,16 the ratio of palla- dium(I1) to reagent VII in the compound was found to be 1:2.With the use of the non- isomolar series, the average value for the instability constant was found to be equivalent to 2-89 x 10-7, which shows good stability. 1. 2. 3. 4. 6. 6. 7. 8. 9. 10. 11. 12. 13. 14. 16. REFERENCES Overholser, L. G., and Joe, J. H., J. Amer. Chem. Soc., 1941, 63, 3224. Popa, G., Negoiu, D., and Baiulescu, G., Zh. Analit. Khim., 1969, 14, 332. Popa, Gr., Negoiu, D., and Baiulescu. G., St. Cerc. Chim. Acad. R.P.R., 1969, 7, 73. Baiulescu, Gh., and Papagheorghe, M., Revtd Chim., 1969, 10, 710. Popa, Gr., Baiulescu, Gh., and Stoichitoiu, Lucia, St. Cerc. Chim. Acad. R.P.R., 1964, 13, 601. Baiulescu, Gh., Greff, C., Moldoveanu, S., and Diinet, F., Anal. Univ. Buc., 1967, 16, 161. Kuznetov, V. I., Zh. Analit. Khim., 1962, 7 , 226. Datta, S. K. and Ghose, P., Naturwissenschaften, 1968, 45, 616. Popa, Gr., Paralescu, I. and Baiulescu, G., St. Cerc. Chim. Acad. R.P.R., 1961, 9, 86. BudBSinsq, B., Vrzalova, D., and Bezdekova, A., Acta Chim. Hung., 1967, 52, 37. Budi%inskfr, B., and MenclovO, B., Talanta, 1968, in the press. PBrez-Bustamante, J. A., and Bumel-Marti, F., Analytica Chim. Ada, 1967, 37, 49. Sen Gupta, J. G., Analyt. Chem., 1967, 39, 18. Job, P., Justus Liebigs Annln Ckem., 1928, 10, 113. I , Ibid., 1967, 37, 62. -- First received July 16th, 1968 Amended October 28th, 1968 Accepted December 16th, 1968

ISSN:0003-2654    

DOI:10.1039/AN9699400354

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

Astalyst, May, 1969, Vol. 94, $$. 359-363 359 Simultaneous Routine Determination of Copper and Zinc in Plants by Neutron-activation Analysis BY A. G. SOULIOTIS (Nuclear Research Center Democritos, Athens, Greece) Copper and zinc were determined simultaneously in plant leaves by neutron-activation analysis. After ashing the leaves, in the presence of copper and zinc carriers, with fuming nitric and 70 per cent. perchloric acids, the solution was evaporated almost to dryness. The residue was treated with Q N hydrochloric acid and the solution transferred to an ion-exchange column containing Dowex 1 x 10, 100 to 200-mesh anion-exchange resin in 6~ hydrochloric acid. Copper and zinc were quantitatively eluted with 3 and 0 . 0 0 6 ~ hydrochloric acid, respectively. The radioactivity of each fraction of solution was counted in a multi-channel andyser by using the 0.51 and 04-MeV peak areas of copper-64 and zinc-69m, respectively. Copper and zinc contents have been obtained for Merent kinds of plant leaves. AMONG the trace elements taken in from the soil by plants, copper and zinc represent pre- dominant factors governing plant nutrition and gr0wth.l Neutron-activation analysis is considered to be one of the most interesting methods for the determination of minute amounts of copper and zinc in biological materials, especially because of its sensitivity and accuracy, and the elimination of the need for a reagent blank.a*8*4~b Neutron-activation analysis has been used by several workers for the quantitative deter- mination of copper and zinc in plants by radiochemical separation, based on precipitation, solvent extraction and ion e~change.~ 9' An ion-exchange method has been reported for the routine determination of copper in biological materials with Chelex-10 (Bio-Rad) chelating resin.O The successful use of a system analogous to that described in this paper has also been reported for the routine activation analysis of manganese, copper and zinc in pine needleslo (The author participated in the project and determined the copper and zinc in pine needles by ion-exchange techniques.) Recently another ion-exchange method has been referred for the simultaneous determination of copper and zinc in biological materials with Amber- lite IRA-400 anion-exchange resin, followed by a precipitation technique.ll A comparative elemental study of a standard kale sample, including copper and zinc among other elements determined, has been performed by different laboratories from different parts of the world.12 This paper describes a simple routine and accurate activation-analysis method for the simultaneous determination of copper and zinc in plant leaves in a series of smples after one irradiation.Almost all published radiochemical procedures2J include the repeated precipitation of suitable copper and zinc compounds, an approach that does not lend itself to the simultaneous analysis of a number of samples. From this point of view, the ion-exchange technique is considered to be the most convenient and promising one. 0 SAC and the author.360 SOULIOTIS : SIMULTANEOUS ROUTINE DETERMINATION OF NUCLEAR REACTIONS INVOLVED IN COPPER AND ZINC DETERMINATIONS- listed in Table I being of interest in determinations of copper and zinc.[Analyst , Vol. 94 On irradiating any material with neutrons, various nuclear reactions take place, those Nuclide 6SCU 65CU 442n 6*Zn TABLE I NUCLEAR REACTIONS" *14 Activation cross- Abundance, section, Half- Reaction per cent.'* mb13 life *Vn p,y) 64Cu 69.10 3000 12.8 hours 65Cu (n,y) %u 30.90 660 5.1 minutes 64211 (n,y) Wn 48.89 220 245days W n (n,y) 6m Zn 18.56 18 13-8 hours Main y -radiation energies , MeV8 0.5 1 1 -04 1.1 1 0.44 Possible interferences from14 04Zn p,p) 64Cu 6ZNi [n,y) 63Ni+ 6eZn (n,p) Wu 6BGa (n,a) 66Cu 84Ni (n,y) 65Ni- 66Cu (n,y) Wu P- 63Cu (n,y)Wu- 64Zn (n,y) 66Zn OgGa (n,p) 8mZn "Ga (n,a) 6mZn [r- 63CU (11,y) 84Cu 0- Of the reactions given in Table I, the fast-neutron induced reaction YZn (n,p) 64Cu is unlikely to interfere in the determination of copper in plants.The amount of zinc in plants15 is usually less than 100pg per g of dry matter and, with the low-activation cross-section of the above reaction (less than 10 pb), the copper-64 produced is equivalent to only 0.07 pg of copper content per g of dry matter in plants. The other fast-neutron, as well as slow- neutron , competing nuclear reactions considered are found to be insignificant. EXPERIMENTAL REAGENTS- Materials were of analytical-reagent grade. Copper nitrate carrier solzltion-Prepare an aqueous solution of copper nitrate, Cu(N0,),.3H20, containing 2mg of copper per ml. Zinc acetate carrier solzctiorz-Prepare an aqueous solution of zinc acetate, Zn(CH3C00),.2H,0, containing 2mg of zinc per ml.Standard copper nitrate and zinc acetate soldion-After standardising the copper nitrate and zinc acetate carrier solutions described above, prepare a standard solution containing 100 pg of copper and 500 pg of zinc per ml. Nitric acid, fuming. Perchloric acid, 70 $er cent. HydrochZoric acid, 6, 4, 3, 0.5 and 0.005 N. APPARATUS- Ion-exchange columns of internal diameter 1 cm and height 20 cm were used, equipped with a socket and a stopcock. An electronic unit was used consisting of a high voltage supply, a pre-amplifier, an amplifier, a rate meter, a recorder and a 3 x 3-inch NaI (Tl) crystal shielded in a suitable lead cylinder. A y-spectrometer was used, consisting of a NaI (Tl) crystal (3 x 3 inch) coupled to an Intertechnique 400-channel pulse height analyser SA-40.IRRADIATION PREPARATION OF SAMPLES TO BE IRRADIATED- at 90" C for 25 hours. The dried samples were ground with a Perspex mill. The plant leaves were washed with detergent, rinsed with distilled water and oven-driedMay, 19691 361 Amounts of about 200mg of ground plant leaves were weighed into nine polythene snap-closure tubes (diameter 10 mm, height 36 mm) , which were then heat-sealed. An aliquot of about 3 0 0 ~ 1 of the standard copper and zinc solution was transferred by pipette into a separate polythene snapclosure tube with the same dimensions as those used for the samples. This tube was also heat-sealed. IRRADIATION CONDITIONS- The nine vials containing the samples, together with the vial containing the standard solution, were put into the holders of a rotating disc, which was rotating in a fixed position near the fuel elements of the reactor.The uniformity of the neutron flux was tested by neutron-flux measurements with gold chloride solutions. The irradiations were conducted in the “Democritos” swimming-pool nuclear reactor, operating at a power level of 1 MW. The neutron flux was 1.6 x 10l2 neutrons per cm2 per second, and the exposure time for the targets was between 8 and 9 hours. ANALYTICAL PROCEDURE^^-- About 12 hours after the end of the irradiation, the targets were taken from the reactor. They were quickly opened behind a lead shield and processed in the following way.(i) Ashing of plant leaves-Each plant-leaf sample was placed in a 100-ml tall-form beaker, each containing 3 ml of fuming nitric acid, 1 ml of the copper and 1 ml of the zinc carrier solutions. The beaker was then heated on a hot-plate. After the first 2 minutes of digestion, and when evolution of nitrous fumes had ceased, 2 ml of 70 per cent. perchloric acid were added, and the solution was slowly evaporated to dryness. The residue was dissolved in 5 ml of 4 N hydrochloric acid. (ii) Ion-exchange sej+aration of copper and zinc-Each solution was then placed on a column containing 14 cm of Dowex 1 x 10, 100 to 200-mesh anion-exchange resin, charged with 6 N hydrochloric acid. The stopcock of each ion-exchange column was regulated to allow an optimum flow-rate of 15 drops per minute.The supernatant solution of each column was allowed to pass through the resin, copper and zinc being retained on it. Thirty millilitres of 4 N hydrochloric acid were then passed through the column. The standard solution was subjected to exactly the same treatment with an identical column. The passage of any radio- activity was ascertained by use of the electronic recording apparatus. This enabled the elution of the different fractions of the standard solution passing through a glass tube under the shielded photocrystal connected with the electronic set to be checked. The 4 N hydrochloric acid fractions of the samples were rejected, as the corresponding fractions of the standard did not show any radioactivity. Forty millilitres of 3 N hydrochloric acid were passed through each column.After the first 10 ml of hydrochloric acid had passed, the recording system of the standard began to show an elution curve that was completed when the 40 ml had passed through. The respective fractions of plant-leaf samples collected and examined in a 400-channel analyser proved to be copper-64 of remarkable radiochemical purity. Thirty millilitres of 0.5 N hydrochloric acid were then passed through the standard and the samples. This fraction, showing no radioactivity for the standard, was rejected. Then 40 ml of 0.005 N hydrochloric acid were allowed to flow through. After the 10 ml of 0.005 N hydrochloric acid had been collected, the recorder also began to show an elutioncurve that was completed when 40 ml of 0.005 N hydrochloric acid had passed through.The respective fractions of the plant-leaf samples collected and also examined in a multi-channel analyser were found to contain zinc-69m of exceptional radiochemical purity. When the ion-exchange techniques described were used the chemical yields for copper and zinc determinations were studied by using tracers. These were found in all cases to be more than 99 per cent., with an optimum value for the elution rate of 0.75 ml per cm2 per minute. Experiments carried out with anion resins of different cross-linking and mesh, as well as with columns of different internal diameter, did not give such satisfactory results as those given below. RADIOACTIVITY MEASUREMENTS In order to have the same basis for counting, care should be taken to use beakers of the same dimensions for the same radioactive fractions, and to ensure that the volumes of the solutions are equal.Each fraction was counted for 5 minutes on a 3 x 3-inch NaI(T1) COPPER AND ZINC IN PLANTS BY NEUTRON-ACTIVATION ANALYSIS362 SOULIOTIS : SIMULTANEOUS ROUTINE DETERMINATION OF [Afla&St, VOl. 94 crystal connected with a 400-channel transistorised Intertechnique analyser, adjusted to count energies up to 2 MeV. A 3-mm Plexi-glass block was placed on the crystal to cut off bremsstrahlung. The copper-64 and zinc-69m photopeak areas (0.51 MeV and 0.44 MeV, respectively) were printed and compared with those of the standards. 7-SPECTROMETRIC EXAMINATION OF THE ISOLATED RADIOISOTOPES- The absence of any y-emitting contaminant in the isolated fractions of copper-64 and zinc-69m from the analysed sample would serve as an additional indication of their radio- chemical purity.The fractions were, therefore, qualitatively examined by scintillation count- ing: The energy spectra examined indicated the absence of any y-emitting contaminant ra&onuclide. DETERMINATION OF THE HALF-LIFE OF THE ISOLATED RADIOELEMENTS OF COPPER-64 AND ZINC-69- This was accomplished by plotting a decay graph on semi-log paper for the fractions of copper and zinc. The half-lives were obtained from the slope of the straight line, calcu- lated by the method of least squares. A value of 12-76; hours, with a standard error of +0-08 hour, was found for copper-64, and a value of 13-91 hours, with a standard error of k0.21 hour, for zinc-69m, instead of 12.8 and 13.8 hours, respectively, as reported in the literature.l3+ RESULTS The method described is a combination of activation and ion-exchange techniques.Copper and zinc contents were obtained for different plant-leaf samples. The values found are given in Table 11. The accuracy and reproducibility of the pro- cedure was checked by the use of a biological standard. The amounts of copper and zinc found in the biological standard were 4.2 and 32.2 p.p.m., respectively. These values are in good agreement with the copper and zinc contents found by other laboratories in the same sample.12 TABLE I1 CONCENTRATION OF COPPER AND ZINC IN PLANT LEAVES Plant Citrus sinensis Citrus nobilis Oka europaea Zea mays .. Beta vulgaris Nicotiana tabacum Medicago sativa Brassica oleracea Copper,* p.p.m... * . 6.3 .. .. 8.4 .. .. 9.7 .. .. 14.3 .. .. 18.8 .. .. 32.4 .. .. 9.8 .. .. 4-2 Zinc,* p.p.m. 20.7 34.6 28.9 39.6 97.6 33.2 19.3 32.2 * Mean value of duplicate analysis. DISCUSSION By using this method copper and zinc were determined simultaneously after one irradiation. The isolated copper and zinc fractions showed remarkable radiochemical purity. The results were reproducible within a relative error of less than k2.5 per cent. This simple and accurate method is convenient for the simultaneous determination of copper and zinc in a large number of samples, by using one standard solution eluted through a column and followed, under identical conditions, by the simultaneous elution of the sample solutions through a battery of columns.The same method can also be applied to the simultaneous determination of these two elements in almost all biological materials. I thank the reactor staff of the Nuclear Research Center “Democritos” for their helpful assistance throughout this work. REFERENCES 1. 2. 3. Stiles, W., in Ruhland, W., Editor, “Encyclopedia of Plant Physiology,” Volume IV, Springer- Dyer, F. F., and Leddicotte, G. W., “The Radiochemistry of Copper,” U.S. Academy of Sciences, Hicks, H. G., “The Radiochemistry of Zinc,” U.S. Academy of Sciences, Nuclear Science Series, Verlag, Berlin, 1968. Nuclear Science Series, NAS-NS-3027, Washington, 1961. NAS-NS-3016, Washington, 1960.May, 19691 COPPER AND ZINC IN PLANTS BY NEUTRON-ACTIVATION ANALYSIS 363 4. 5. 6. 7.8. 9. 10. 11. 12. 13. 14. 15. 16. Bowen, H. J. M., and Gibbons, D., “Radioactivation Analysis,” Clarendon Press, Oxford, 1963. Lyon, W. S., Jr., “Guide to Activation Analysis,” D. Van Nostrand Company Inc., New York, 1964. Bock-Werthmann, W., and Schulze, W., “AKtivierurtgs Analyse” Report AED-C-14-1, Hahn- Bock-Werthmann, W., Ibid., Report AED-C-14-2, Hahn-Meitner Institut fur Kernforschung, -, Ibid., Report AED-C-14-3, Hahn-Meitner Institut fur Kernforschung, Berlin, February, Fritze, K., Aspin, N., and Holmes, T. H., Radiochimica Acta, 1964, 3, 206. Yassoglou, N., Grimanis, A., Apostolakis, K., Nobeli, A., and Vrahamis, S., Annual Report of the Soils and Plant-nutrition Laboratory of Democritos NRC, 1964, Grant No. FG-Cr-101. Livingston, H. D., Smith, H., and Stojanovic, N., Talanta, 1967, 14, 606. Bowen, H. J. M., Analyst, 1967, 92, 124. Allen, R. A., Smith, D. B., and Hiscott, J. E., U.K. Atomic Energy Authority Report AERE-R 2938, Second Edition, 1961, pp. 63, 64, 194 and 196. Koch, R. C., “Activation Analysis Handbook,” Academic Press, New York, 1960, p. 80. Bowen, H. J. M., and Cawse, P. A., U.K. Atomic Energy Authority Report AERE-R-4309, 1963, Appendix 1. Kraus, K. A., and Nelson, P. M., “Metal Separations,” Symposium on Ion Exchange and Chro- matography in Analytical Chemistry, held in Atlantic City, N.J., June 18, 1966,” A.S.T.M., Special Technical Publication No. 195, p. 41. Received November 6th, 1967 Accepted October 8th, 1968 Meitner Institut fur Kernforschung, Berlin, December, 1961. Berlin, February, 1963. 1964.

ISSN:0003-2654    

DOI:10.1039/AN9699400359

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

364 Analyst, May, 1969, Vol. 94, $9. 364-368 The Determination of Dissolved Gases in Water Continuous Stripping and Gas Chromatography BY J. A. J. WALKER AND E. D. FRANCE ( U.K. Atomic Energy Authority, Reactor Materials Laboratory, Culcheth, Warrington, Lancs.) A device is described for the continuous removal of dissolved gases in water by mixing pure helium and water on a continuous-flow basis. The stripper unit has been tested by using a helium ionisation chromatograph and found satisfactory for recoveries of hydrogen, oxygen, methane, carbon monoxide and carbon dioxide a t the 0.1, 0.3, 0.16, 0-18 and 0-454 ml kg-1 levels, respectively, at 18OC. For a constant helium flow-rate of 1 ml minute-' and water flow-rates in the range 1 to 6 ml minute-1, the coefficient of variation is about 4 per cent., except for carbon monoxide and carbon dioxide. The maximum flow-rates of helium and of water are 10 ml minute-1 for this design of stripper; when the flow-rates of gas and water are equal over this greater range, the coefficient of variation indicated by the result is 10 per cent. Units consisting of several strippers fixed in a cabinet have been in routine use on experimental water loops for the past 2 years.The strippers have been operated on a continuous basis at concentrations from 0.001 to 1 ml of gas per kg of water. THE possible use of ferritic steel in the primary coolant system of commercial steam generating, heavy-water reactors necessitates investigation of the water chemistry associated with the presence of ferritic steel under boiling water conditions.In particular, dissolved gases (e.g. , oxygen) in the water may have a marked influence on the corrosion rate of this steel, and accurate measurements of the concentration of gases in the water are necessary. In published methods for the determination of gases in water, various devices are used for the removal of gaseous solute, either by equilibration of the constituent between an inert gas and the water1s2p3s4 or by using a static vacuum chambeIdps into which a discrete sample of water is introduced. The evolved gas is sampled and analysed by a gas-chromatographic method. The use of an inert stripping gas (e.g., helium) for the continuous removal of dissolved gas from high purity water, with the sample flow-rate restricted to about 1 ml minute-l and gas-in-water concentrations in the range 0-001 to 1 ml kg-l, was directly applicable to a particular analytical problem.A stripping unit was designed for this work and this, coupled with a helium ionisation gas ~hromatograph,~ was used to determine amounts of permanent gases present in water. The performance of the stripper has been evaluated at the Reactor Materials Laboratory, Culcheth, and on pilot plant over the past 2 years. EXPERIMENTAL APPARATUS- The stripper unit (Fig. 1) comprised a Perspex body machined to accommodate a glass- spiral tube up which heiium and water flow. A base and top Perspex section were drilled to provide inlet and outlet points for gas and water. The base, body and top section were bolted together to form a single unit, and O-ring seals were incorporated throughout to mini- mise air in-ledage.Uniform bubble formation was facilitated by insertion of a piece of poly(viny1 chloride) tubing (1.8 mm 0.d.) at the entrance to the glass spiral (Fig. 2) ; this spiral, which consists of twenty-eight coils of 6 cm diameter, is 5.3 m in length, with 0.4 cm 0.d. and 0.25 cm i.d. Both ends of the spiral are drawn out into capillaries about 2-6 cm in length, with 0.2 cm i.d. and 0-8 cm o.d., which pass through the Perspex container. The volume of the coil is 26ml. 0 SAC; Crown Copyright Reserved.WALKER AND FRANCE Perspex !id , Helium outlet A Fig. 1. Stripper unit (one third actual size) 365 Contamination of the pure cylinder helium used as the stripper gas was minimised by using a pressure regulator that had a leakage rate of not more than 4 x 1 hour-l.A stop valve and a capillary pressure reducer were sited between the regulator and helium inlet to the stripper. The capillary pressure reducer, capable of regulating the flow-rate of the stripper gas to about 1 ml minute-l, was made from stainless-steel tubing (600 cm in length, with 2 mm 0.d. and 0.36 mm bore) and designed for a pressure differential of 30 lb per incha to atmospheric pressure. Fig. 2. Enlarged view of A, in Fig. 1366 WALKER AND FRANCE: DETERMINATION OF DXSSOWED GASES IN [A?$U&St, VOl. 94 The sample (water) line was similarly fitted with a stop valve and capillary pressure reducer between the high-pressure (lo00 Ib per inch2) water source and the water inlet to the stripper.A stainless-steel tube, 220 cm in length, 1 mm 0.d. and 0.12 mm bore, gave a flow-rate of about 1 ml minute-l at the stripper inlet (to which it was connected by an O-ring pressure coupling) for a pressure differential of 1000 lb per inch2 to atmospheric pressure. A stainless-steel filter (porosity 10 p) in the water line between the stop valve and capillary tube removed suspended matter in the water, which might otherwise have blocked the capillary bore. The gas outlet part on the stripper was connected to the sample injection valve of a helium ionisat ion chromatograph . OUTLINE OF METHOD The separate parts of the apparatus are connected according to the schematic diagram (Fig. 3). The stop valves on the water and helium lines are opened and the gas - water mixture allowed to flow through the stripper.The flow-rates of helium and water are measured (for gas, by measuring the time taken for a soap bubble to move from one fixed mark to another, the two marks defining a known volume, and for water, by collection of a known volume in a measured time) after the gas and water have passed through the gas-chromato- graphic sample valve and the stripper, respectively. The flow-rates can be vaned for both helium and water by altering the length of the capillary pressure reducer. A time period of up to 2 hours may be required to purge the stripper and its associated pipework of air. From helium cylinder - 1 1 Capillary 4 pressure reducer IHelium + stripped components1 -- i to Sample valve Helium Sample gas J LStripping carrier to vent 1-4 I coil gas Capillary pressureb reducer I valve Fig.3. Schematic diagram of apparatus The sampling of the helium from the stripper is started and the components chromato- graphed. By reference to a calibration graph of the amount of constituent verszcs the chromatographic peak height, the mount of constituent present in the sample volume taken can be ascertained. The concentration of the constituent in the water is calculated from the following relationship- Millilitres of constituent - amount of constituent from chromatogram in 10” litres per kilogram of water - volume of sample in millilitres flow factor lo00 flow-rate of helium where the flow factor = flow-rate of water PRACTICAL CONSIDERATIONS- A critical factor in the efficient operation of the stripper unit is the effect of air in-leakage to the gas - water lines.At low gas - water flow-rates small leaks caused by incorrectly made O-ring compression joints can produce gross errors in the results for oxygen and nitrogen.May, 19691 WATER BY CONTINUOUS STRIPPING AND GAS CHROMATOGRAPHY 367 To minimise the dead volume in the pipework connecting the stripper to the gas-chromato- graphic sampling valve, 0-030 inch bore stainless-steel tubing (4 inch 0.d.) is used. Accidental “carry-over” of water (e.g., at high flow-rates) leads to a blockage of the stainless-steel tube and prevents water from reaching the sampling valve. If a blockage occurs both gas and water leave the stripper by the water exit. The possibility of “carry-over” can be obviated by allowing sufficient dead volume at the end of the spiral stripper tube for a clean separation of helium and water.Bubble geometry for gas and water is not critical, except when complete breakdown of the water bubble occurs. There appears to be a continuous film of water on the inside surface of the stripper spiral, and interchange of water between a water bubble and the film will take place as the bubbles move up the tube. The process of stripping may be con- sidered a dynamic one, with new water surfaces continually being exposed to the helium stripper gas. STANDARD GAS - WATER MIXTURES- To test the efficiency of the stripper it was necessary to prepare standard gas - water solutions for hydrogen, oxygen, methane, carbon monoxide and carbon dioxide.A 5-litre capacity bottle (Fig. 4) was fitted with a leak-tight stopper, through which passed an inlet and outlet tube for gas and a tube connected to a Kelvin pressure gauge. A mixture (of known composition) of the standard gas in helium was passed through the distilled water in the bottle until equilibrium was established, when the solubility of the standard gas was calculated at the temperature of the experiment by using Henry’s law.* The standard appara- tus was then connected to the stripper unit and the water sample syphoned from the bottle through the stripper. During the analysis, to obviate entrainment of gas mixture bubbles in the outlet pipe to the stripper and to maintain the standard mixture in contact with the surface of the solution, the gas inlet was connected to tube 2.Substitution of the gas mixture by pure helium showed efficient de-gassing of the distilled water to a residual concentration of 0401 ml kg-l for any one added constituent. Gas unit Fig. 4. water mixtures Apparatus for preparation of standard gas-in- RESULTS Various solutions of permanent gases in de-gassed water were prepared and analysed by using the standard technique with variations in helium and water flow. Results of these analyses, carried out at a temperature of 18” C (constant to within 1” C), are shown in Table I, and the standard deviation is quoted (as a The precision of these results depends partly on the constancy of flow-rates for helium and water. This in turn is dependent on the flow-control device; a needle valve was used for these experiments.Lack of uniform flow-rate could explain the imprecision shown in the recoveries of carbon monoxide when the flow-rate of water was increased while the helium was kept constant. figure) after each mean recovery.368 WALKER AND FRANCE TABLE I RECOVERIES OF STANDARD GAS ADDITIONS Amount added, ml kg-1 Constituent (in water) Hydrogen . . .. 0-096 Oxygen . . .. 0.286 Methane .. .. 0.155 Carbon monoxide . . 0.180 Carbon dioxide . . 0.454 Amount recovered, ml kg-l (in water) 0-094 f 0.003 (7 results) 0-097 & 0-009 (8 results) 0.283 & 0.010 (4 results) 0.297 f 0-014 (5 results) 0.156 f 0.007 (7 results) 0.156 f 0,015 (8 results) 0.163 f 0.02 (7 results) 0.183 & 0.006 (7 results) 0.267 0.316 0.390 0.480 Flow characteristics, ml minute-’ - Water Helium 1.3 1 to 5.7 Equal flow-rates 3 to 12 1 1.5 to 6.8 Equal flow-rates 1.5 to 11.5 1.4 1.1 to 5.0 Equal flow-rates 2 to 10 1.3 1.9 to 6.5 Equal flow-rates 2-0 to 10.0 The results of the recovery experiment in which a standard solution of carbon dioxide in water was used show that the flow-rate of helium should be five times the water flow-rate for complete recovery.Carbon dioxide is classed as a soluble gas relative to the other gases investigated and, for a helium and water flow-rate of 1 ml minute-l, the residual carbon dioxide remaining in the water is about 46 per cent. at equilibration. For hydrogen, oxygen, carbon monoxide and methane, the residual gas is between 2 and 3 per cent. at equilibration, at a flow-rate of 1 ml minute-l for helium and water. Although a nitrogen standard has not been used it is assumed that there will be little difference between nitrogen and oxygen with respect to efficient stripping. This assumption is based on the fact that the solubilities of these two gases are similar. 1. 2. 3. 4. 6. 6. 7. 8. REFERENCES Gibson, S. P., Allison, G. M., and Atherley, J. P., “Proceeding: of 4th Conference on Analytical Chemistry in Nuclear Reactor Technology, Gatlingburg, 1960, U.S. Atomic Energy Commission, Re+ort TID-7606, pp. 330-343. Singler, M., and Snyder, D. T., “Proceedings of 6th Conference on Analytical Chemistry in Nuclear Reactor Technology, Gatlinburg, 1962,” U.S. A tomic Ertergy Commission, Re$& TID-7655, Swinnerton, J. W., Linnenbom, V. J., and Cheek, C. H., Analyt. Chem., 1962, 34, 483. Williams, D. D., and Miller, R. R., Ibid., 1962, 34, 657. Gaunt, H., and Shanks, C., Chem. 6. Ind., 1964, 651. -- , Ibid., 1965, 328. Berry, R., in van Swaay, M., Editor, “Gas Chromatography 1962,” Butterworths & Co. (Publishers) “Handbook of Chemistry and Physics,” Fourth Edition, Chemical Rubber Publishing Co., Cleve- First received Decembev 14th, 1967 Amended March 4th, 1968 Accepted December 16th, 1968 pp. 147-176. Ltd., London, 1962, p. 321. land, Ohio, 1960, p. 1706.

ISSN:0003-2654    

DOI:10.1039/AN9699400364

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

Analyst, May, 1969, Vol. 94, $$I. 369-376 369 A Field Method for the Determination of Organic Aromatic Isocyanates in Air BY D. W. MEDDLE, D. W. RADFORD AND R. WOOD (Ministry of Technology, Laboratory of the Government Chemist, Cornwall House, Stamford Street, London, S.E.l) A general field method is described for the determination in air of some organic aromatic isocyanates commonly used in industry. Collection of the isocyanate vapour or aerosol, or both, is effected in a mixture of hydrochloric acid and dimethylformamide. The amine produced is diazotised and coupled with N- 1-naphthylethylenediamine to form a coloured complex, the intensity of which is compared with a set of permanent colour standards appropriate to the isocyanate being determined. The method enables the determination of up to 0.04 p.p.m.v/v of isocyanate in air to be made. The procedure is simple and the time required for a complete analysis is about 35 minutes. MOST polyurethane products, such as cellular plastics and adhesives, were originally based on the use of tolylene-2,4-di-isocyanate (TDI) as starting material. The toxic hazard involved in handling this volatile material was appreciated and industry sought to use, when possible, other organic aromatic isocyanates possessing lower vapour pressures. The following are among the more widely used of these compounds in the United Kingdom. 4,4’-Di-isocyanatodiphenylmethane, otherwise known as methylenediphenyl iso- cyanate (MDI), which is used in the manufacture of rigid and semi-rigid foams and surf ace-coating lacquers.Polymethylenepolyphenyl isocyanate (PAPI), which is also used for manufacturing rigid foam, casting non-cellular plastics and as a functional component in adhesive systems such as the bonding of synthetic fibres to synthetic rubbers. Naphthylene-l,5-di-isocyanate (NDI), which is used in the production of synthetic rubbers of high tensile strength and resistance to oxygen, ozone, petroleum and mineral oil. 3,3’-Dimethoxy4,4‘-di-isocyanatobiphenyl, normally referred to as dianisidine di- isocyanate (DADI), which is a slow reacting compound used in certain polymer and adhesive systems. The above compounds, although not an exhaustive list, provide a good cross-section of the various chemical types occurring within the organic isocyanate class used in the production of polyurethane plastics.Despite the low vapour pressures of MDI, PAPI, NDI and DADI, hazardous concen- trations of these isocyanates may occur as droplets or as aerosols in industrial atmospheres. This can arise as a result of the malfunctioning of the spraying or injection equipment used, allied to inadequate ventilation. The currently recommended threshold limit value for both TDI and MDI is 0.02 p.p.m. v/v in air.1 Although at present there are no threshold limit values for PAPI, NDI or DADI, it seemed reasonable to assume that they would be no less hazardous. Methods have been proposed for the determination of individual aromatic isocyanates in air, but no attempt has hitherto been made to develop a single general procedure that might be of use for determining any organic aromatic isocyanate at the 0.02 p.p.m.v/v level. In view of the wide use in industry of isocyanates other than TDI, a need for such a procedure appeared to exist. The method originally proposed by Marcali2 for the determination of TDI in air, and subsequently modified by other workers,as4 was selected as being suitable for adaptation to a general field-test procedure. This colorimetric method depends on the hydrolysis of the TDI to the corresponding amine; the latter is diazotised and finally coupled with N-l-naph- thylethylenediamine to form a coloured complex, which is measured spectrophotometrically. It was decided that a critical examination of the various steps of the Marcali2 method was required. This, it was hoped, would enable the optimum conditions for the collection and determination of any organic aromatic isocyanate to be established.It was realised at the 0 SAC; Crown Copyright Reserved.370 MEDDLE, RADFORD AND WOOD: A FIELD METHOD FOR THE [AHdySt, VOl. 94 outset that compromise would inevitably have to be made when attempting to match the optimum experimental parameters required for two, or more, different isocyanates. Conse- quently, TDI and MDI, examples of a volatile and a non-volatile isocyanate, respectively, were selected as the test isocyanates with which any detailed investigations on each step of the procedure would be carried out. After establishing the particular optimum conditions for the above, these would be applied to the other isocyanates being studied, viz., NDI, PAPI and DADI, and an ultimate compromise condition reached.EXPERIMENTAL PREPARATION AND CALIBRATION OF STANDARD ATMOSPHERES- Atmospheres containing known concentrations of the various isocyanates were required, both to assess the efficiency of the sampling technique and for use in the development of a field method. Dynamic atmospheres of the relatively volatile TDI were prepared by bubbling dry nitrogen at a controlled rate through the liquid isocyanate, which was maintained at a steady temperature, and diluting with a metered stream of nitrogen to the required concentration. The atmospheres were sampled in duplicate, one sample being analysed by the proposed colorimetric method and the other by a gas-chromatographic rneth~d.~ The results obtained are summarised in Table I.TABLE I RESULTS OF DUPLICATE ANALYSES OF TDI STANDARD ATMOSPHERES Average concentration of TDI atmospheres per run, pg litre-1 r 1 No. of runs Colorimetricall y By gas chromatography 24 0.065 f 0.01* 0.070 f 0.01 9 0.080 f 0.017 0.087 f 0-018 10 0.23 f 0.028 0.26 f 0-029 Standard deviation of mean concentration. Attempts to produce standard atmospheres of MDI and PAPI by a similar method were unsuccessful. The concentrations of MDI in atmospheres produced by bubbling dry nitrogen through liquid MDI maintained at 50" C diminished rapidly. At the same time, the dimer of MDI was formed in the vessel containing the liquid isocyanate. In view of this and as all of the isocyanates, apart from TDI, had very low vapour pressures and, consequently, were likely to present a hazard mainly in the form of an aerosol rather than as a vapour, a method of generating consistent aerosol atmospheres was devised.This work, including the results obtained, has been described fully elsewhere.6 Besides being applicable to the production of reproducible atmospheres of the non-volatile isocyanates, the method was also found to be more suitable for the generation of standard atmospheres of TDI than the vapour saturation technique described above. CHOICE OF ABSORBING SOLUTION- A solution was required which, as well as providing a high efficiency of collection of isocyanates (see Choice of absorber below), would also enable a colour of sufficiently high intensity to be developed. By using TDI atmospheres, preliminary work had indicated that an absorbing solution of dimethylformamide - dilute hydrochloric acid, as used by ReillyJ4 was superior in both respects to a dilute acetic acid - hydrochloric acid mixture, identical absorbers being used in each instance.Reilly' had also reported that the addition of acetic acid to the dilute hydrochloric acid absorbent used did not improve on the 75 per cent. collection efficiency of MDI vapour in the first of two absorbers in series. In view of these observations the dimethylformamide - dilute hydrochloric acid mixture was selected for further study. It was noted that fairly rapid deterioration took place in the dimethylformamide - dilute hydrochloric acid mixture. When isocyanate solutions were added to a freshly prepared mixture the colours developed were more intense than those obtained when it was a few days old; this was examined further.A 6 + 6 v/v mixture of dimethylformamide and 1-36 M hydrochloric acid was prepared. Each day over a period of 10 days, 6-65 pg of MDI were added (see Preparation of calibration graphs and colour standards below) to 5-ml aliquotsMay, 19691 DETERMINATION OF ORGANIC AROMATIC ISOCYANATES IN AIR 37 1 of the mixture and the colour developed. The optical density dropped gradually from0-715 to 0.545, while the pH of the mixture rose from 1-04 to 3-0, and dimethylamine hydrochloride was formed as a result of the reaction between the two species present. The use of an aged mixture, with reduced acidity, to collect isocyanates, presumably resulted in two competing reactions taking place, the hydrolysis of the isocyanate to the corresponding iunine and the formation of a complex between the isocyanate and dimethylamine. Several variations in the composition of the absorbing solution were tried to avoid this difficulty.These included varying the amount of water present during the hydrolysis of isocyanate to amine stage. Although a greater colour response was noted with reduced water content, the age of the mixture used was still the critical factor. It was subsequently shown that to obtain reproducible results it was advisable to prepare the absorbing mixture freshly before each determination. A volume of 5 d , consisting of 3 ml of dimethylformamide and 2 ml of 1-65 M hydrochloric acid, was found suitable for use with the absorber chosen for the field method.This method of preparing a mixture was found to obviate the necessity of redistilling the dimethylformamide before use, as had been pre- viously re~ommended.~ The results obtained by successively sampling a TDI atmosphere with absorbing solutions containing undistilled dimethylformamide of four different origins showed a variation in optical density ranging from 0.464 to 0.466. Similar experiments with these four samples after re-distillation gave slightly less consistent results, with a range of optical densities from 0-462 to 0.470. CHOICE OF ABSORBER- To achieve maximum sensitivity with the proposed method, it seemed logical to keep the volumes of reagents to a minimum. The narrow, impinger-type absorbers designed by Reill9 were useful in achieving this.However, while proving satisfactory for the virtually quantitative trapping of a vapour atmosphere such as TDI, they were found to be only 63 per cent. as efficient as a sintered absorber for trapping an aerosol - vapour atmosphere such as MDI. It appeared that the aerosol particles passed through an impinger in the bubbles formed. The introduction of a second similar trap in series with the first provided no improvement in collection efficiency. A domed sintered absorber, of the type shown in Fig. 1, containing 5 ml of the proposed absorbing solution, was found to trap virtually 100 per cent. of an aerosol or vapour standard isocyanate atmosphere.6 The sinter acted as a filter in removing the aerosol particles, and it was necessary to allow time for these to dissolve and hydrolyse after the end of the sampling period. Table I1 shows the effect of this digestion time on the recovery of various isocyanates from aerosol atmospheres; a period of 10 minutes was selected as a suitable compromise.Sintered domes of porosity 1 were chosen to allow a sampling rate of 1 litre minute-l. TABLE I1 EFFECT OF DIGESTION TIME ON THE RECOVERY OF ISOCYANATES FROM AEROSOL ATMOSPHERES Isocyanate recovery, pg litre-' Digestion time, minutes 0 6 10 15 20 30 M b I PAP1 DAD1 N6I 0.36 0.70 0.30 - 0.42 0.72 - 0.60 0.46 0.74 0.42 0.62 0-43 0-79 - 0.60 - 0.78 0.43 0.64 - 0-77 - 0.62 In the sampling of any aerosol-containing isocyanate atmosphere it was found necessary to use a straight inlet tube mounted in a vertical position, otherwise some loss of sample occurred on the sides of the probe.CHOICE OF COUPLING CONDITIONS- N-l-naphthylethylenediamine dihydrochloride had already been shown to be an effective coupling reagent for TDI, full colour development occurring within 3 minutes.2 Initial work with the other isocyanates under study, in which a similar 0.1 per cent. coupling reagent372 MEDDLE, RADFORD AND WOOD: A FIELD METHOD FOR THE [A"ZdySt, VOl. 94 solution was used, revealed that a much slower coupling rate was operating in all instances. Full colour development was not reached after 1 hour. It was also noted that the speed of coupling increased with temperature. In a method for the determination of MDI, based closely on the original method2 for TDI, Grim and Linch3 found that the time for colour production was reduced from 2 hours to 15 minutes by increasing the concentration of the coupling solution.A series of experiments was carried out with each isocyanate with in- creasing concentrations of N-l-naphthylethylenediamine solution. The results are shown in Table 111. It can be seen that, by using 0.5 ml of a 0.75 or 1.0 per cent. v/v N-l-naphthyl- TABLE I11 EFFECT OF VARYING CONCENTRATION OF COUPLING REAGENT SOLUTION ON RATE OF FULL COLOUR DEVELOPMENT TDI MDI PAP1 I & r N- 1 aaphthylethylene- diamine concentration, gper 100 ml . , . . 0.25 0.5 0.75 1.0 0.25 0.5 0.75 1.0 0.25 0.5 0.76 1.0 Isocyanate, pg . . . . 6.98 6.98 6.98 6.98 6-65 8-65 6-65 6.65 6.75 6.75 6.75 6.75 Optical density- After 5 minutes . . 0-434 0.444 0.528 0.499 0.462 0.615 0.663 0-670 0-345 0.421 0.477 0.493 After 10 minutes .. 0.460 0.455 0.539 0.502 0.577 0-678 0.710 0.710 0.429 0.486 0.519 0.519 After 15 minutes . . 0.464 0.456 0.541 0.510 0.637 0.700 0.713 0.713 0.468 0.511 0.531 0.530 Wavelength of maximum absorbance after 10 minutes' develop- ment, nm . . .. 556 650 550 550 586 590 590 590 575 575 580 580 DAD1 NDI & & N-1 -naphthylethylene- diamine concentration, gper 100 ml . . . . 0.25 0-5 0.75 1.0 0.25 0.5 0.75 1.0 Isocyanate, pg . . . . 7-0 7.0 7.0 7.0 6.36 6-36 6.36 6.36 Optical density- After 5 minutes . . 0-258 0.301 0.325 0.319 0.277 0.321 0.329 0.345 After 10 minutes . . 0-295 0.307 0.325 0.320 0.315 0.337 0.336 0.347 After 16 minutes . . 0.305 0.307 0,325 0.320 0.326 0-337 0.336 0.347 Wavelength of maximum absorbance after 10 minutes' develop- ment, nm .... 600 605 615 620 686 586 586 585 ethylenediamine solution, 98 per cent. of the full colour development was obtained within 10 minutes, a time considered satisfactory for the purposes of the proposed field test. This choice of coupling solution and time of colour development also made the reaction independent of temperature between 20" and 25" C, the range within which the temperature of the reaction solution should be at this stage of the procedure. An examination of the results in Table I11 suggested that the use of a 0.75 per cent. v/v solution of the amine provided the best com- promise. Although the 2,4TDI isomer was used in these experiments, later work confirmed that the 80 + 20 and 65 + 35 mixtures of the 2,4- and 2,6-TDI isomers normally used in industry reacted similarly.It can also be seen from Table I11 that increasing the concen- tration of the amine solution had the effect of altering the wavelength of maximum absorbance of all of the isocyanates, apart from NDI. CHOICE OF DIAZOTISATION CONDITIONS- In previous methods for determining individual isocyanates a 3 per cent. w/v sodium nitrite solution containing 5 per cent. w/v of sodium bromide, the latter acting as a reaction catalyst, was used for diazotisation. In the present work, tests with MDI indicated that the use of 0.2 ml of this solution, and allowing a 2-minute diazotisation period, produced an optimum response; a longer period gave no enhancement. For convenience in the field, 0-5ml of nitrite solution was selected for use.This reduced the final optical density by a factor inversely proportional to the increase in volume of the reaction solution, Le., by about 8 per cent.May, 19691 DETERMINATION OF ORGANIC AROMATIC ISOCYANATES I N AIR 373 An investigation of the optimum temperature conditions required for the diazotisation of the various isocyanates was also carried out. Known amounts of each isocyanate were added to 5ml of freshly prepared absorbing mixture and the temperature adjusted to a point within the range 10" to 25" C. The temperature control was removed and the remainder of the colour development procedure carried out immediately, the starting temperature being taken as that of diazotisation. The results obtained are listed in Table IV, and it can be seen that the optical densities produced with MDI and PAPI were independent of diazo- tisation temperature over the range studied.Diazotisation of the respective amines derived from NDI and 2,4-TDI produced a slight reduction in optical density when carried out above 20" C. The 80 + 20 mixture of the 2,4- and 2,6-TDI isomers, and to a greater extent the 65 + 35 mixture, exhibited a more pronounced decrease in colour when the diazotisation temperature was above 20" C. This latter observation suggested that the diazo compound of the amine derived from 2,6-TDI was particularly susceptible to breakdown above 20" C. In contrast to the other isocyanates, the formation of the diazo compound of DADI appeared to take place slowly, reaching its maximum at 24" C.By allowing a waiting period of 5 instead of 2 minutes, the diazotisation of this isocyanate proceeded to completion at a lower temperature. TABLE IV EFFECT OF DIAZOTISATION TEMPERATURE ON DEVELOPMENT OF COLOUR Optical density (from starting diazotisation temperature) Weight, r L > Isocyanate pg 10°C 15°C 17OC 2OoC 21°C 22°C 23°C 24°C 25°C - 0.669 NDI 7-94 0-490 0.490 0.490 0.483 0.481 0.480 0.478 0.476 0.474 2,4-TDI 5.8 0.470 0.470 - 0.470 0.468 0,461 0.453 0.447 0.440 TDI 9.6 0.625 0.625 - 0.625 0.618 0.610 0.602 0.696 0.588 0.758 - - - MDI 7.09 0.768 0.758 - 0.758 - PAPI 7-94 0.655 0.657 - 0.657 - - - 2,4- and 2.6- (80 + 20 mixture) 2,4- and 2,s- TDI 11.28 0.680 0.680 - 0.680 0.666 0.630 0.597 0664 0.630 (65 + 35 mixture) DADI* 8.95 0.238 0.265 - 0.380 0.402 0.426 0.450 0.465 0.465 DADIt 8.95 - 0.417 - 0.452 0.458 0.463 0.464 0.465 0.465 * 2-minute diazotisation period. t 6-minute diazotisation period.On the basis of the results given in Table IV, a temperature of 20" C of the absorbing solution was selected as the optimum for the diazotisation step for all of the isocyanates; when determining DADI the diazotisation time was increased to 5 minutes. This temperature condition was readily achieved, as it was found that the temperature of a freshly prepared absorbing solution was about 37" C and dropped to about 20" C after the passage of 10 litres of atmosphere sampled at 1 litre minute-l, the ambient temperature being in the range 20" to 25" C. However, if a result of the highest precision is required, and the ambient air tempera- ture is above 25" C, it is advisable to ensure that the absorbing solution temperature is not above 20" C before proceeding with the diazotisation step.PREPARATION OF CALIBRATION GRAPHS AND COLOUR STANDARDS- It was desirable to prepare calibration graphs for the various isocyanates by using, if possible, a solvent of similar composition to that of the absorbing solution. It was found that the amounts (20 to 50 mg) of the isocyanates, liquids or solids, required for this were soluble only with difficulty in the absorbing solution. The dissolution of the isocyanates in dimethylformamide, followed by the addition of the requisite amount of dilute hydrochloric acid, was also unsatisfactory. Some of the isocyanate normally reacted with impurities in the dimethylformamide before the acid was added.These difficulties were circumvented by dispensing aliquots of standard toluene solutions of the various isocyanates from an Agla micrometer syringe directly into the absorbing solution ; 5 ml of absorbing solution tolerated the addition of 0.08 ml of toluene without becoming opalescent. In practice, the concentra- tions of the toluene solutions were adjusted so that 0.06 ml contained the maximum isocyanate374 MEDDLE, RADFORD AND WOOD: A FIELD METHOD FOR THE [ArtdJ6t, VOl. 94 concentration required for the production of any calibration graph. Calibration graphs, over the range 0 to 16 pg, were prepared for all of the isocyanates, by using the optimum conditions determined for each stage of the procedure.Each isocyanate exhibited a linear response between concentration and optical density. Table V shows the optical densities, measured on a Unicam SP600 spectrophotometer at the appropriate wavelengths, with 20-mm cells, obtained by the collection of 10-litre samples of 0.02 pep.". v/v atmospheres of each isocyanate. TABLE V OPTICAL DENSITIES PRODUCED BY COLLECTION OF 10 LITRES OF 0.02 P.P.M. V/V ISOCYANATE ATMOSPHERES 2,4- and 2,6-TDI mixtures - MDI PAPI DAD1 NDI 2,4-TDI (80 + 20) (65 + 36) Weight, pg . . . . 2.0 3.3 2.4 1.7 1.4 1.4 1.4 Optical density (20-mm cell) . . .. . . 0.216 0.23 0.126 0.100 0.135 0.117 0-113 The values given in Table V indicated the possibility of preparing permanent colour standards for the various isocyanates, thus eliminating the need for a portable spectrophoto- meter in the field.With the co-operation of Tintometer Ltd., sets of standard discs were prepared representing the intensity of the colours produced by collecting 10-litre samples of 0-01,0.02 and 0.04 p.p.m. v/v atmospheres of MDI, PAPI, NDI and the 65 + 35 mixture of 2,P and 2,6-TDI when viewed vertically through a solution depth of 60mm. The use of the colour standards for the 66 + 36 mixture of 2,P and 2,6-TDI for the determination of pure 2,PTDI or the 80 + 20 mixture of TDI isomers results in a slight over-estimation of the isocyanate concentration. PROPOSED FIELD METHOD FOR THE DETERMINATION OF ORGANIC AROMATIC ISOCYANATES IN AIR APPARATUS- AZZ-gZms absorbers-These are of the type shown in Fig. 1. To pump *=I + Sinter,' porosity I Fig.1. All-glass absorberMay, 19691 DETERMINATION OF ORGANIC AROMATIC ISOCYANATES IN AIR 375 Glass tubes-Flat-bottomed, 64mm in length and 1Omm in diameter. Sampling pum$-This should be capable of drawing air at the rate of 1 litre minute-1 through the absorber. Colour standards-comparator discs containing coloured glass standards for this test, for use with the Lovibond "1000" Comparator, are available from Tintometer Ltd., Salisbury. Discs are available for MDI, NDI, PAP1 and TDI. REAGENTS- All reagents should be of analytical-reagent grade when possible. Dilute hydrochZoric acid-Dilute 15 ml of concentrated hydrochloric acid (sp.gr., at 20" C, Dimethyl formamide-This reagent can be used as received. Diazotisation solution-Dissolve 3 g of sodium nitrite and 5 g of sodium bromide in water Sulphamic acid solution, 10 per cent.w/v, aqueous. N-1-Naphthylethylertediamine solutiort--Dissolve 0.75 g of N-1-naphthylethylenediamine dihydrochloride in water, add 2 ml of concentrated hydrochloric acid and dilute to 100 ml with water. Prepare a fresh solution after 2 days. 1.19) to 100 ml with distilled water. and dilute to 100 ml. PROCEDURE- Place 3 ml of dimethylformamide and 2 ml of dilute hydrochloric acid in the absorber and insert the inlet tube. Mount and fur the absorber at the sampling site in a vertical position (this is especially important when sampling an isocyanate atmosphere in the form of an aerosol). Attach the pump to the absorber and draw 10 litres of the test atmosphere through the absorbing solution at a rate of 1 litre minute-1.Allow the absorbing solution to stand for 10 minutes to ensure complete hydrolysis of all isocyanate collected. (This waiting period is not required when TDI is collected.) Lift the inlet tube so that the sinter is clear of the liquid and expel the liquid trapped in the domed sinter as completely as possible. Ensure that the temperature of the absorbing solution is not above 20" C, then add 0-5 ml of the diazotisation solution. Shake the absorber to mix the contents and allow to stand for 2 minutes (5 minutes if DAD1 is being determined). Add 0.5 ml of the sulphamic acid solution and shake the mixture until the effervescence has ceased; 2 minutes after the addition of the sulphamic acid solution, add 0.5 ml of N-1-naphthylethylenediamine solution and mix well.Pour sufficient of the test solution into a 64 x 10-mm flat-bottomed glass tube to fill it to a depth of 50 mm. Fill a similar tube to the same depth with water. Insert both tubes into the comparator. By using the appropriate set of colour standards for the isocyanate under test, view through the depth of the liquid. Obtain the nearest colour match between standards and sample 10 minutes after addition of the N-1-naphthylethylenediamine solution. Should a level higher than 0.04 p.p.m. v/v (it?., the top colour standard) be indicated, a smaller sample of atmosphere should be taken at the same sampling rate. The resulting colour match in parts per million units for this second run is then multiplied by -(where x is the sample volume in litres) to obtain the isocyanate concentration in the atmosphere.If a spectrophotometer is available, a graph of the optical densities against concentration can be prepared (see Preparation of calibration graphs and colour standards) for any of the isocyanates. From the increased optical density of the sample at the appropriate wavelength it is possible to make a more precise determination of the isocyanate concentration in the atmosphere sampled. 10 X DISCUSSION AND APPLICATION OF METHOD No detailed examination was carried out on possible interferences with the proposed field method; any primary aromatic a.mine collected from an atmosphere would interfere by producing a colour. Also, as the isocyanate grouping showed a tendency to react readily with most other organic groupings and with water, the presence of such species in an isocyanate atmosphere might cause the concentration of the latter to be reduced.Both of these factors operate in the interest of safety. With water vapour a further complication arose because376 MEDDLE, RADFORD AND WOOD one reaction with aromatic isocyanates, I, involves the formation of amine, 11, before the production of a carbanilide, III- RN.CO 2> H O RNH.COOH , Z -co RNH, f RN CO RNH.CO.NHR I I1 I11 Thus, with the proposed test, when applied to a moist atmosphere containing isocyanate, amine could be determined as isocyanate and, consequently, atmospheres that appeared to be hazardous might not necessarily be so. The carbanilide, 111, does not produce a colour in the proposed test.This theory was tested by preparing standard, dynamic atmospheres of TDI with moist air and collecting duplicate samples. One sample was collected in an absorbing solution of dimethylformamide containing 75 pg ml-l of hexamethylenediamine to which was added dilute hydrochloric acid after the completion of sampling, and any colour developed in the normal way. By using this technique, the TDI was inactivated by reaction with the hexamethylenediamine, and only aromatic amine, derived from TDI, would produce a colour. The other sample was collected in the normal absorbing solution and both TDI and its related amine would produce a colour under these conditions. Thus, by measuring the optical densities of the colour produced in each absorbing solution an estimate of the TDI, and related amine concentrations, in any atmosphere could be made.In experiments conducted with air, with a water content ranging from 4 to 19 g m4 at 22” C, a reduction of 6 to 16 per cent., respectively, in the TDI concentration of the atmospheres was noted compared with the theoretical atmospheres possible. No evidence was found of amine being produced in these experiments. This suggested that any free isocyanate in a moist atmosphere would be determined by the proposed method without interference. Isocyanate and an aromatic primary amine could occur together in an atmosphere when the latter is used as the polymerisation catalyst. The above technique of duplicate sampling with two different absorbing solutions can be used to determine the concentration of isocyanate in the presence of amine.In this instance, a spectrophotometer must be used to measure the optical densities of the colour produced in such a determination. The proposed method has been satisfactorily assessed under field conditions at industrial establishments where TDI and MDI were being used. No method of check testing was possible but, by using the test, it was found that atmospheric contamination by isocyanates diminished rapidly as the distance from the source of contamination increased. The apparatus required for the test is portable and requires only an external electrical power supply to operate the pump. A determination can be completed in about 35 minutes. Permanent glass standards are available, which allow the determination in air of the most commonly used isocyanates in industry, over the range 0 to 0.04 p.p.m. v/v (ie., up to twice the present threshold limit value). With the use of a portable spectrophotometer and a previously prepared calibration graph, the proposed method is suitable for more precise determinations over the range 0 to 1-5 pg litre-l. This work was carried out on behalf of the Department of Employment and Productivity Committee on Tests for Toxic Substances in Air. We are grateful to the Government Chemist for permission to publish this paper, and to H. M. Factory Inspectorate for arranging facilities for the field tests. REFERENCES 1. 2. 3. 4. 5. 6. 7. “Ministry of Labour, Safety, Health and Welfare, New Series No. 8, Dusts and Fumes in Factory Atmospheres,” Third Edition, H.M. Stationery Office, London, 1966. Marcali, K., Analyt. Chem., 1957, 29, 552. Grim, K. E., and Linch, A. L., Amer. Ind. Hyg. Ass. J., 1964, 25, 285. Reilly, D. A., Analyst, 1963, 88, 732. Wheals, B. B., and Thomson, J., Chem. & Ind., 1967, 753. Meddle, D. W., and Wood, R., Ibid., 1968, 1635. Reilly, D. A., Analyst, 1967, 92, 513. Received November LSth, 1968 Accepted November 29th, 1968

ISSN:0003-2654    

DOI:10.1039/AN9699400369

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

Analyst, May, 1969, Vol. 94, $9.377-386 377 A Rapid Method for the Production Control of the Non-ionic Component in Synthetic Detergents BY B. M. MILWIDSKY (Zohur Delergetzt Fuctwy, Huifa, Israel) A dilute solution of the detergent is slightly acidified and mixed with 20 ml of the cobaltothiocyanate reagent and 20 ml of dichloromethane, the mixture is shaken and allowed to separate. The intensity of the colour of the dichloromethane is measured and the percentage of non-ionic detergent is calculated from straight-line cali- bration graphs. Graphs are given for varying amounts of ethylene oxide units per molecule and also for varying hydrophobic molecules. In addition, poly(ethy1ene glycols) of varying molecular weights were checked by this method, together with synthesised fatty acid - poly(ethy1ene glycol) esters in an attempt to find a relationship between molecular weight and the slope of the graph.It was found that both the hydrophobic and hydrophilic portions of the molecule affect the slope. Anionic detergents interfere in this determination, but a simple method to overcome this interference is given. THE published methods of analysis of non-ionic detergents, which generally are ethylene oxide condensation products, are not suitable for the routine analytical control of the manu- facture of ready-made detergent liquids or powders, particularly when continuous methods of production are being used and the result is required within minutes of sampling. Gravimetric methods and titrimetric methods involving filtration are unsuitable because of the length of time required, and even longer periods are necessary for those non-ionic detergents which have to be separated from inorganic salts. It was found by Gnamml and van der Hoeve2 that polyoxyethylene compounds form a blue, water-insoluble complex with ammonium cobaltothiocyanate.Brown and Hayes* extracted the complex into chloroform and measured the intensity of the colour of the chloroform solution. In the method developed by Brown and Hayes, the complex was formed and after 5 minutes it was extracted three times with chloroform, the extracts were combined and made up to standard volume. The solution was centrifuged to clear it of water-haze, before measuring the colour. It was decided to investigate whether this method could be shortened to be suitable for routine manufacturing control.Various solvents and solvent mixtures heavier than water were tested and it was found that dichloromethane gave the most rapid separation of the phases with minimum entrain- ment of water. Although dichloromethane is extremely volatile it could be used during the hot Israeli summer without appreciable change in the colour reading, even if the open cell was left in the colorimeter for 5 minutes. Extraction was carried out only once in accordance with the theory that if constant volumes and conditions are used the first extraction will extract a definite and constant proportion of the complex. If the detergent solution was alkaline, as it most often is, precipitation of a basic cobalt salt occurred. The ammonium cobaltothiocyanate solution has a pH of 5 and, therefore, it was necessary to adjust the solution to the same pH before the formation of the complex.Under these conditions, with dichloromethane as the solvent for the complex, no difference in colour reading was found if the complex was formed, allowed to stand for 5 minutes and the dichloromethane added, or if all the reagents are added initially, the mixture shaken and the colour read immediately after the separation had taken place. 0 SAC and the author.378 APPARATUS- inverse logarithmic scale giving values of extinction coefficients, E, where MILWIDSKY: A RAPID METHOD FOR THE PRODUCTION CONTROL [Anuiyst, Vol. 94 EXPERIMENTAL Calorimeter-The instrument used was a Gallenkamp photo-electric colorirneter with an incident light, per cent.= loglo transmitted light, per cent.' The values of E were used in every instance and were not converted into optical densities. Cells-Optical test cells with an optical depth of 10mm were used. Filters-The instrument is normally equipped with Ilford Bright Spectrum Filters. Filter No. 607, which gives a maximum transmission at 600 mp, was found to give the best sensitivity. In addition a heat-absorbing filter, Chance No. ON20, is used. Se@rating funnel, 100m2 cajbacity-The stem of the funnel should not be longer than 1 cm. REAGENTS- Ammonium cobaltothiocyanate solutiort-Dissolve 30 g of cobalt nitrate and 200 g of am- monium thiocyanate in 1 litre of water. Dichloromethane-Chemically pure. Standard acid solution-Either 0.1 N or N, depending on the detergent being tested.MethyZ red irtdicator solution. PROCEDURE- Weigh l o g of detergent, dissolve it in water then transfer it to a 500-ml graduated flask and make up to the mark with distilled water. (Note that aliquots can be taken from this solution for all the other determinations required.) Take a 50-ml aliquot of this solution, titrate with standard acid, which can be either 0.1 N or N, depending on the alkalinity of the material, with methyl red indicator. To another 50-ml aliquot in a 100-ml graduated flask add a further amount of standard acid, equal to that used in the above titration and make up to the mark with distilled water. From this diluted, acidified solution transfer a 10-ml aliquot into a 100-ml separating flask and add, by means of a pipette, 20 ml of ammonium cobaltothiocyanate solution and 20ml of dichloromethane.Shake the separating funnel for 1 minute and allow the phases to separate ; this takes place almost immediately. Discard the first 1 ml of the lower layer and then fill a test cell directly from the funnel with the dichloromethane (lower) solution and read the colour in the colorimeter. The amount of detergent present is then read off from a graph previously prepared for the particular material being used. If l o g exactly of the detergent are weighed initially and all the above dilutions are observed exactly and the calibration graph is plotted with milligrams vers?ds E (colour), the number of milligrams obtained from the graph will be the percentage of detergent present in the original material.If the colour reading is outside the upper limit of the scale, the 10-ml aliquot should be replaced by a 5-ml aliquot and 5 ml of distilled water added. If the colour is at the lower limit of the scale, the original sample taken should be greater than log. It is good practice, when rinsing the separating funnel after the determination, to rinse with the stopcock closed and to allow the funnel to drain upside down with the stopcock open not to enclose water in the bore of the stopcock. It was found that traces of water tend to destroy the complex in the dichloromethane solution. A separating funnel with a short stem was, therefore, recommended. PREPARATION OF THE CALIBRATION GRAPH- Weigh accurately 1 to 2 g of the non-ionic detergent, dissolve it in water and dilute to 250 ml in a graduated flask.Titrate a 50-ml aliquot with 0-1 N acid, with methyl red indicator. To another 50-ml aliquot, add the amount of acid required for the above titration and dilute to 100 ml. This is solution A.May, 19691 OF THE NON-IONIC COMPONENT IN SYNTHETIC DETERGENTS 379 Add to the separating funnel 1 ml of solution A and 9 ml of distilled water. Proceed as above. Continue this procedure with 2 ml of solution A and 8 ml of water, 3 ml of solution A and 7 ml of water and so on until at least 5 points have been obtained. From the readings obtained plot a graph of E versus milligrams of detergent present in each increment. RESULTS Because pure materials of this type are not readily obtainable, most of the experimental work was carried out on commercial samples.In addition various poly(ethy1ene glycols) were checked. As manufacturers of non-ionic detergents do not always publish exact details of their products, fatty-acid poly(ethy1ene glycol) esters were synthesised in the laboratory to give reference figures, which were also checked by this method. For each of twenty-five different samples, straight-line graphs were obtained on calibra- tion, with two minor deviations as explained below. An ethylene oxide condensation product, which has a nominal 9 molecules of ethylene oxide in the larger molecule, can have some molecules with 4 ethylene oxide units and others with 15 and, depending on the conditions of manufacture, the spread can be either wide or narrow. In addition, varying amounts of poly(ethy1ene glycols) will be formed during the manufacturing process.To complicate the matter further the hydrophobic portion of the molecule can be a mixture of both isomers and homologues. As will become evident below, each of these factors affects the slope of the calibration graph. In Fig. 1, eight calibration graphs are shown for different materials all of which contain about 9 ethylene oxide units. All of the materials plotted in Fig. 1 have virtually the same number of ethylene oxide units in the molecule but, as can be seen, there are wide variations in the curves, and even for A and B, which are nominally the same material, there is a difference caused by the different manufacturing procedures. 1 1 1 1 I I I I I I 1 I I I ' I I I I 20 30 40 Non-ionic, mg Fig.1. Calibration graphs for the following materials: A, nonylphenol with 9 molecules of ethylene oxide, estimated from physical charac- teristics; B, nonylphenyl with 9 molecules of ethylene oxide, different manufacturer from A; C, fatty alcohol C,, to C, with 9 molecules of ethyIene oxide; D, oleyl alcohol with 8 to 9 molecules of ethylene oxide; E, oleic acid with 9 molecules of ethylene oxide, synthesised from Carbowax 400; F, launc acid, 9 molecules of ethylene oxide, synthesised from Carbowax 400 ; G, poly(propy1ene glycol), with 8 to 9 molecules of ethylene oxide, co-condensate of ethylene oxide and propylene oxide; H, with 9 molecules of ethylene oxide, Carbowax 400 Non-ionic, mg Fig. 2. Calibration graphs of 4 fatty alcohol ethoxylates: A, Gl to C, containing 9 ethylene oxide units; B, GI to C, containing 12 ethylene oxide units; C, Gl to G6 containing 7 ethylene oxide units; D, oleyl alcohol containing 9 ethylene oxide units; E, Gl to G6 containing 3 ethylene oxide units380 MILWIDSKY: A RAPID METHOD FOR THE PRODUCTION CONTROL [.4utdySt, VOl.94 Fig. 2 shows calibration graphs of four different fatty alcohol ethoxylates containing 3, 7, 9 and 12 ethylene oxide units, all from the same manufacturer and all using as the hydrophobe an alcohol with an average carbon number of 13, ranging from C1, to C1, alcohols. In addition, oleyl alcohol with 8 to 9 ethylene oxide units (from a different source), is shown for comparison. It will be noted that the graphs increase for the 3, 7 and 9 samples, fall for the 12 sample, and for the oleyl alcohol product they are similar to those for the 7 sample.In Fig. 3 five alkylphenol products are shown; two graphs for nonylphenol with 9 ethylene oxide units (graphs A and B in Fig. 1) and in addition nonylphenol with 5 and with 12 ethylene oxide units and a dialkylphenol with 6 to 7 ethylene oxide units: again no regularity is found. 60 60 - 50 40- 30 - 50 - 20 Lu 20 10 - - 20 30 49 I 1 r 1 / F ‘B c / - /D - / G H , I I I I I I I l l 10 20 30 40May, 19691 OF THE NON-IONIC COMPONENT IN SYNTHETIC DETERGENTS 381 wax 200) and five ethylene oxide units. The products tested were commercial materials that have a random distribution of the number of ethylene oxide units around a definite mean, and it is therefore considered that the colours obtained were caused only by those fractions containing at least six ethylene oxide units per molecule.However, it does appear that a commercial product with low (average) ethylene oxide content can still produce a reaction and this method can, therefore, be used for its determination. From the results obtained with methoxypoly(ethy1ene glycols) it appears that the functional terminal group (in this instance OH) plays a part in the complex formation; Carbowax 550 has half the number of ethylene oxide units of Carbowax 1000, but almost identical graphs were obtained for these two materials. Carbowax 1000 has two functional groups whereas Carbowax 550 has only one. This suggests that removal of a free OH group by esterification increases the degree of colour and poses the question whether interaction of the second OH group will increase further the colour developed.Carbowax 350 was esteri- fied with lauric acid, but it was found that this ester gave only a weak reaction, considerably weaker than the pure Carbowax 350 by itself. In Fig. 5, graphs for Carbowaxes 350, 550 and 750 are shown, together with the lauric acid ester of Carbowax 350. Also in Fig. 5 are shown oleic acid esters of Carbowaxes 600 and 4-00 and lauric acid esters of Carbowaxes 1540 and 400. By comparing Figs. 4 and 5 it will be noted that the graph of the oleic acid ester of Carbowax 600 falls fairly close to the graph of Carbowax 600 itself, and the two esters of Carbowax 400 fall on either side of the graph of the unesterified material.It therefore seems that one terminal functional group is essential for the reaction and when two are available the reaction is weaker. In the case of the methoxy Carbowaxes a hydrophilic group was added and this enhanced the colour. The addition of a hydrophobic group to the poly(ethy1ene glycols) did not give a regular pattern. The colour was either enhanced, not affected seriously or decreased. A further anomaly occurs in the two esters of Carbowax 400. Lauric acid is slightly less hydrophobic than oleic acid, but the lauric acid ester gives a weaker colour than the oleic acid ester. Finally the “tailing-off” effect observed with unesterified Carbowax 1540 was not observed when this material was esterified with lauric acid to produce a detergent.28 u1 26- 24 30 20 20 30 40 LLI - -. Non-ionic, mg Fig. 5. Carbowax esters: A, Carbowax 750; B, Carbowax 550; C, oleic acid ester of Carbowax 600; D, lauric acid ester of Carbowax 1540; E, Carbowax 350; F, oleic acid ester of Carbowax 400; G, lauric acid ester of Carbowax 400; H, lauric acid ester of Carbowax 350 Anionic, mg Fig. 6. Interference of anionic detergent : A, 20 mg of non-ionic; B, 12 mg of non-ionic; C, 9 mg of non-ionic INTERFERENCE OF ANIONIC DETERGENTS- Soap when mixed with non-ionic detergents did not in any way interfere with the results, because the determination is carried out at an acidic pH and soap, as such, no longer exists at this pH. Synthetic anionic detergents when mixed with non-ionic detergents interfere and Nadeau and Siggia6 suggest that this is caused by complex formation.382 MILWIDSKY: A RAPID METHOD FOR THE PRODUCTION CONTROL [AutUbSt, Vol.94 It was found, however, that the method could be adapted to determine non-ionic in the presence of anionic detergents. The separation into two phases was not immediate and the colour was considerably less than when anionic detergents were absent. On mixing varying amounts of alkylbenzenesulphonate with a constant volume of non-ionic detergents it was found that the colour developed in the dichloromethane layer fell, initially, with the amount of anionic detergent and then reached a constant value as shown in Fig. 6. This observation was used for the routine determination of non-ionic detergents in the presence of anionic detergents, in that the concentration of both is arranged so that the concentration of the anionic detergent is on the level portion of the graph. DETERMINATION OF NON-IONIC DETERGENTS IN THE PRESENCE OF ANIONIC DETERGENTS- Prepare a calibration graph for the non-ionic to be determined as described above and keep the remainder of solution A.Weigh accurately sufficient of the type of anionic detergent being mixed with the non-ionic, so that about 1 6 g of 100 per cent. active anionic detergent is obtained (the percentage of active matter needs to be determined previously), dissolve in water and dilute to 250ml. Take a 50-ml aliquot of this solution and titrate with acid with methyl red indicator. Take another 50-ml aliquot of this solution, add the same amount of acid used in the titration and dilute to 100 ml.This is solution B. Into the separating funnel introduce 2 ml of solution B, 1 ml of solution A and 7 ml of distilled water. Proceed as described above for the determination of non-ionics. Then use 2 ml of solution B, 2 ml of solution A and 6 ml of water, then another 2 ml of solution B, 3 ml of solution A and 5 ml of water, and finally 2 ml of solution B, 4 ml of solution A and 4 ml of water. Plot a graph for this operation noting the constant amount of anionic detergent present in the separating funnel. Repeat the above procedure with 4 ml of solution B, and varying amounts of solution A and water, keeping the total constant at 10 ml. Again repeat the above procedure with 6ml of solution B and varying amounts of solution A and water, but keeping the sum total at 10 ml.For normal detergents it will be found that the graphs of the last two determinations coincide. If not (e.g., non-ionics with low amounts of ethylene oxide) it is necessary to weigh larger amounts than previously mentioned. The two coinciding graphs are taken as the calibration graph for this particular non-ionic mixed with this particular anionic detergent, and the minimum amount of anionic detergent for this graph is noted. For routine determination of a non-ionic in the presence of an anionic detergent, 10 g of the detergent are weighed and diluted to 500 ml. On an aliquot of this solution the anionic detergent is determined in the normal manner. It is then necessary to calculate whether the dilution given in the method for the determination of non-ionics will leave sufficient anionic detergent in the 10-ml aliquot of solution A so that the non-ionic readings will fall on the calibration graph.If so, the procedure is as above, if not, a larger aliquot is required to make up solution A, allowance being made for the acid addition or, alternatively, a larger amount of detergent is weighed. CATIONIC DETERGENTS- Cationic detergents interfere and this method is not applicable in their presence. Some of the fatty acid alkylolamides (the dialkylolamides) exhibit non-ionic detergent characteristics. It was noted (in a private communication from J. Tidhar) that a qualitative reaction is obtained by using the cobaltothiocyanate reagent with lauric acid diethanolamide and this reaction was investigated quantitatively.Alkylolamides are available commercially as either coconut fatty acid or lauric acid condensates with one of monoethanolamine, mono-isopropanolamine, diethanolamine or di- isopropanolamine and other than the fact that the ethanolamines are synthesised from ethylene oxide and the isopropanolamines from propylene oxide they have no ethylene oxide grouping. FATTY ACID ALKYLOLAMIDES-May, 19691 OF THE NON-IONIC COMPONENT IN SYNTHETIC DETERGENTS 383 Neither the condensation products of mono-ethanolamine nor mono-isopropanolamine with either of the fatty acids gave the reaction. Diethanolamine condensed with either of the fatty acids gave the colour reaction weakly, the graph having a tangent approximately of the order of a fatty alcohol with three ethylene oxide units (Fig.2). It was observed that when the constituents were mixed initially no colour was developed. Further observations were made as the condensation proceeded. Initially no colour was developed, but when the condensation was about half completed a colour started to appear and became stronger as the amount of amide increased, until it reached a maximum when the condensation was completed. When this ethanolamide reacted with a further molecule of diethanolamine to produce the 1 : 2 molecular ratio ethanolamide the colour intensity decreased proportionately, indi- cating that no further products that react with the cobaltothiocyanate reagent are produced. Condensation of the fatty acids with di-isopropanolamine also gave a colour reaction but considerably weaker than that of the diethanolamide.Alkylolamides axe occasionally used with non-ionic detergents in liquid formulations, but when they are thus used their proportion of the total active matter is low, usually about 10 per cent. Mixtures of some typical non-ionic detergents were made with a coconut fatty- acid diethanolamide, the mixtures were allowed to react with the cobaltothiocyanate reagent, as above, and the intensity of the colours produced was measured to determine the extent of their interference of the alkylolamide in the non-ionic determination. The non-ionic detergents chosen were product A of Fig. 1 and the dialkylphenol plotted in Fig. 3. The results are shown in Table I. TABLE I INTERFERENCE OF ALKYLOLAMIDES IN NON-IONIC DETERMINATIONS Kon-ionic, mg Product A 10 - - 10 10 10 10 15 15 15 Dialk ylphenol 17 17 17 Alkylolamide, mg - 45 22.5 45 11-5 5.7 2.8 5-7 2-8 - - 2-7 2.1 * Reading not reliable.Colour intensity 27 10 36 32 28 27 42 44 42 2* 35 36 35 From Table I it will be observed that when about 3 mg of alkylolamide are present in the aliquot , irrespective of the amount or type of polyethoxylated non-ionic, the interference can be ignored. No rapid method for the determination of alkylolamides is yet available, but if a detergent contains not more than 3 per cent. of alkylolamide and the dilutions described above are adhered to the interference of the alkylolamide can be ignored for routine work. If the detergent contains more than 3 per cent. of alkylolamide (of the diethanol- amide type only), dilutions should be arranged so that the final 10-rnl aliquot contains less than 3 mg of the alkylolamide when using this method for routine control of detergents. The amount of a lauric acid diethanolamide required to give a definite colour reading is about ten times that of the average polyethoxylated detergent used commercially.This method is, therefore, not suggested as a means of control of ethanolamide content in detergents because the amounts normally present are low, but it is suggested that this reaction might be a useful method for the production control of the condensation. Methods normally used for the control are based on negative results, Le., the drop in concentration of free amine, ester or fatty acid. By the use of this reaction the actual “appearance” of a product being synthesised can be observed.However, as yet no clear picture has emerged of what material is giving this cobaltothiocyanate reaction.384 MILWIDSKY: A RAPID METHOD FOR THE PRODUCTION CONTROL [Anahst, VOl. 94 INTERFERENCES OF INORGANIC ANIONS AND OTHER BUILDERS- No interference was found to take place in the presence of sodium tripolyphosphate, tetrasodium pyrophosphate, sodium perborate, sodium sulphate, sodium chloride, sodium silicate, sodium carboxymethylcellulose (CMC) and optical brighteners. PRECISION OBTAINABLE- Experimental error and accuracy were checked in various ways. First, two calibration graphs for different batches of the same material were prepared at an interval of 3 months. The graphs were plotted on different scales.Four cardinal points on the colour ordinate were chosen and the amount of material necessary to produce this colour from each graph was read. The results are shown in Table 11. TABLE I1 EXPERIMENTAL ERROR IN CALIBRATION OF GRAPHS Material, mg r Percentage Colour intensity Graph I Graph I1 deviation 50 22.1 22.4 + 1.3 40 18.8 18.6 - 1.1 30 14-1 14.6 + 2.1 20 10.1 10.3 + 2.0 Discrepancies in the plotting of the calibration graphs were also checked. The points measured for material A from Table I (an alkylphenol with 9 ethylene oxide units) are given in Table 111. TABLE I11 CALIBRATION OF MATERIAL A Material, mg. . . . 3-06 6.12 9-18 12.24 15.30 18.36 21.42 Colour intensity . . 6-5 15 25 33 44 64 60 These points are not all on a perfectly straight line but an average line was drawn through them.Four of the above points are, however, on a dead straight line and a graph was drawn through these points ignoring the others. Again four cardinal points were chosen, and the amount of material necessary to produce this colour was read off from each graph. The results are given in Table IV. TABLE IV DIFFERENCES IN READINGS FOR DIFFERENT METHODS OF DRAWING GRAPHS Material, mg v Percentage Colour intensity Average graph Special graph deviation 60 17-9 17.2 - 4.0 40 14.0 14-5 - 3.6 30 10.8 11.2 - 3.7 20 7.8 8-1 - 3.7 In this instance the deviation is higher than that given in Table 11, but it must be remem- bered that the special graph is artificial and not normal practice. The calibration graphs produced in the course of this work were used for the determination of the non-ionic content in compounded detergents, both liquid made in the laboratory and powders made both in the laboratory and on a commercial scale.For samples made in the laboratory the standard of comparison taken was the amount of active matter added to the blend. For the commercial powders the materials were dried in an oven at 105” C and the “total active detergent matter’’ was determined by extracting with isopropyl alcohol in a Soxhlet extraction apparatus. Anionic detergents, if present (either soap or synthetic) were determined separately, the synthetic portion determined by the method of Eptons and soap by the method of Milwidsky and Holtzman,’ and these results were subtracted from the total active matter.Results are given in Table V.TABLE V 0 r ACCURACY ACHIEVED Non-ionic & Soap Added Determined determined 42 4043 - Isopropyl alcohol Percentage extract deviation - 2.8 - Anionic Total active determined determined - 40.8 Detergent type Liquid Type Alkylphenol with 9 molecules of ethylene oxide Alkylphenol with 12 molecules of ethylene oxide Alkylphenol with 12 molecules of ethylene oxide Alkylphenol with 0 molecules of ethylene oxide Alkylphenol with 9 molecules of ethylene oxide Alkylphenol with 6 molecules of ethylene oxide Blend of alkylphenol with 6 mole- cules of ethylene oxide and oleyl alcohol with 8 molecules of ethylene oxide Alkylphenol with 9 molecules of ethylene oxide Unknown 12 12.7 - 26* 38.7 - + 1.9 Liquid - 11-4 - 0.9 - Alkaline liquid 11.6 11.4 - Liquid plus solvent 24 24.4 - I 24.4 - + 1.7 Heavy-duty powder, laboratory blend Heavy-duty powder, laboratory blend Heavy-duty powder, spra y-dried 10 10.2 - 6 6.2 5* - 10.2 - + 2.0 6* 15.2 - + 1.3 17.1 - 1.7 10-4 16.7 - 11.2 - 16.1 11.5 - 2.6 16-6 + 3.1 Heavy-duty powder, Heavy-duty powder, spray-dried spray-dried - 11.2 - w 0 M 3 v, - 10.9 6.7 * Added.386 MILWIDSKY From the figures in the last column it will be seen that the results achieved are quite satisfactory for all practical purposes. These results take into account all the factors: personal error, experimental error and the influence of other materials commonly present in detergents. It should also be noted that the last item was a powder commercially available, containing a non-ionic detergent, the type of which was unknown to us.A good approximation for the amount of this material was nevertheless obtained. INTERFERENCE OF ORGANIC SOLVENTS- No interference was found from any of the simple solvents that are also soluble in water such as alcohols, ketones, ethylene glycol, diethylene glycol and Cellosolve. Diglycol- amine, which can be considered to be monoethanolamine with one ethylene oxide unit added, also did not react. Carbitol which is the lowest of the series of methoxypoly(ethy1ene glycols) interferes. DETERMINATION OF AN UNKNOWN NON-IONIC DETERGENT- It has been stressed that no relationship can be found between the slope of the graph and the structure of the detergent. However, if the rapid determination of the percentage of an unknown detergent present in a ready-made material is required, a good approximation can be found by using the following method. It is necessary to have available a series of calibration graphs for various detergents used commercially. This should present no difficulty to an industrial laboratory. Proceed as above and then repeat the procedure with a 6-ml aliquot of solution A and 5 ml of water. A graph is then selected with a slope such that the first colour reading gives an amount of detergent exactly double that of the second colour. This can be used as a good approximation to the percentage of non-ionic detergent present but gives no clue as to the type. A quick and reliable method has been developed, which can be used with confidence for routine analytical determinations of non-ionic detergents, and in addition rapid approxi- mations can be made for unknown non-ionics. REFERENCES 1. Gnamm, H., “Die LBsungs und Weichmachungsmittel, ” Sixth Edition, Wissenschaftliche Verlags- 2. van der Hoeve, J . A., Rec. Trans. Chim. Pays-Bas, 1948, 67, 649. 3. Brown, E. G., and Hayes, T. J., Analyst, 1965, 80, 756. 4. Crabb, N. T., and Persinger, H. E., J . Amer. Oil Chern. Soc., 1964, 41, 752. 6. Nadeau, H. G., and Siggia, S., in Schick, M. J., Editor, “Nonionic Surfactants,” Marcel Dekker 6. Epton, S. R., Trans. Favaday SOL, 1948, 44, 226. 7. Milwidsky, B. M., and Holtzman, S., Soap, 1966, 41 (6)] 83. gesellschaft, Stuttgart, 1960, p. 336. Inc., New York, 1967, p. 841. Received Jury 16th, 1968 Accepted November 4th, 1968

ISSN:0003-2654    

DOI:10.1039/AN9699400377

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

Analyst, May, 1969, Vol. 94, PP. 387-388 387 The Spectrophotometric Determination of Cationic Surfactants with Picric Acid BY I. SHEIHAM AND T. A. PINFOLD (Department of Chemistry including Biochemistry, University of the Witwatersrand, Johannesburg, Republic of South Africa) Bromophenol blue and bromothymol blue have been used in the past for the spectrophotometric determination of cationic surfactants, by extrac- tion into 1,2-dichloroethane of the blue complex formed between the dye and the surfactant. The method is shown to be unsatisfactory because the dyes can exist in both blue and yellow forms, and often an uncontrollable change of colour occurs in the organic phase. Picric acid, which produces a stable, intensely coloured complex with surfactants, is suggested as a substitute.The pH of the aqueous solutions is restricted to the range 4 to 12; high ionic strengths are not deleterious. THE colorimetric determination of surfactants is well known; it is usually achieved by forming a complex between the surfactant ion and a dye of opposite charge. The complex is then extracted into an organic solvent and determined spectrophotometrically. The method for cationic surfactants was first introduced by Auerbach,lJ who determined quaternary ammonium detergents by using bromophenol blue and bromothymol blue as dyes, and 1,2-dichloroethane and benzene as solvents. This procedure was not effective because the presence of dispersed water in the organic phase impaired the spectrophotometric measurement and the extraction of the complex into benzene was incomplete.Van Steveninck and MasS showed that the dispersed water could be removed by centrifuging, and concluded that 1,2-dichloroethane was an entirely suitable solvent for use with bromophenol blue. We cannot agree with this conclusion because on occasions we found the colour of the complex between bromophenol blue and three surfactants, dodecylpyridinium chloride, hexadecylpyridinium bromide and hexadecyltrimethylammonium chloride, to change appre- ciably with time. Solutions that were initially blue and absorbed at 600nm gradually became yellow and absorbed at 385 nm. The colour change took place at a rate that was neither constant nor reproducible, the reduction in absorption at 600 nm being up to 14 per cent. per hour. When compared with samples that showed no fading, however, this decrease was appreciably greater.It is well known that a blue bivalent form and a yellow univalent form of the bromo- phenol blue ion exist, and complexes of both forms with hexadecyltrimethylammonium chloride have been isolated by Mukerjee4 and Mukerjee and My~els,~ who extracted the blue form from solutions of pH 12 and the yellow form from solutions of pH 2. Although they do not mention any colour change occurring in the organic solvent, we have found that in the pH range 4 to 12 both forms occur in 1,2-dichloroethane, and that as this cannot be avoided spectrophotometric analysis is, therefore, not possible. As conditions of high acidity or basicity in the sample are sometimes difficult to establish, a method suitable for use in the intermediate pH range is needed.The deterioration of the blue colour was always attended by an increase in that of the yellow colour, and it suggested that a change from the basic to the acidic form was occurring. This was confirmed by introducing carbon dioxide into the solvent, when the acidic properties of the gas greatly encouraged the transformation. Although the reason for the colour change is clear, it was not evident why it occurred on some occasions but not on others. Prior distillation of the solvent often facilitated the change, although no difference could be found in the infrared spectrum before and after distillation. Further, it was found that two different samples of dye behaved similarly, showing that the effect was not caused by an impurity.As bromophenol blue was clearly unsuitable and bromothymol blue showed the same behaviour to an even greater extent, both were abandoned. 0 SAC and the authors.388 SHEIHAM AND PINFOLD A suitable dye for use with 1,2-dichloroethane should be a strong organic acid, highly coloured and of simple structure. Orange I1 @-(/3-naphthylazo) benzenesulphonic acid] has been used6 for the purpose; there are objections to this reagent that will be discussed shortly. We, on the other hand, used $-nitrophenol with some success and found picric acid (2,4,6-trinitrophenol) to be eminently suitable. EXPERIMENTAL A known volume of surfactant solution was made up to about 40ml with water and about 1 ml of a 0.1 per cent. w/w solution of picric acid in 09002 M sodium hydroxide added.The whole was shaken with 20 ml of doubly distilled 1,2-&chloroethane for 6 minutes. Part of the organic layer was then removed and centrifuged at 3500 r.p.m. for 5 minutes; the optical density was measured at 376 nm in 4-cm glass cells. A calibration graph adhered closely to Beer’s law and gave an optical density of 1.2 for 0-4 pmoles of any of the three surfactants. The extraction did not depend on the ionic strength of the surfactant solution and no interference was encountered with solutions M with respect to sodium chloride. The three surfactants investigated showed the same behaviour, and the method was therefore quite unselective. Bromophenol blue and bromothymol blue behave in the same way. No optical absorption was found for picric acid alone in the pH range 4 to 12.At pH 2, however, an absorption peak occurred at 340 nm because of partition of the un-ionised acid. Picric acid is more suitable than Orange 11, which is not a readily available reagent and is susceptible to photolytic decomposition. This reagent is useful in the pH range 2 to 8. It appears, therefore, that provided measurements are confined to pH range 4 to 12, the use of picric acid affords a reliable method of determining cationic surfactants. The authors are indebted to the South African Atomic Energy Board for a research grant; also to Mr. N. Sparrow for assisting with part of the investigation. The precision attained throughout was within 3 per cent. REFERENCES 1. Auerbach, M. E., Analyt. Chem., 1943, 15, 492. 2. - , Ibid., 1944, 16, 739. 3. Van Steveninck, J., and Maas, M., Red. Trav. Chim. Pays-Bas Belg., 1966, 84, 1166. 4. Mukerjee, P., Analyt. Chem., 1966, 28, 870. 5. Mukerjee, P., and Mysels, K. J., J . Amer. Chem. SOC., 1966, 77, 2937. 6. Few, A. V., and Ottewill, R. H., J . Colloid Sci., 1966, 11, 34. Received August 20th, 1968 Accepted November 7th, 1968

ISSN:0003-2654    

DOI:10.1039/AN9699400387

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

Analyst, May, 1969, Vol. 94, $$. 389-391 389 The Determination of Calcium in Chrome Refractories by Atomic-absorption Spectroscopy BY N. SIDIROPOULOS (Iscor Research Laboratories, P.O. Box 460, Pretoria, South Africa) A method is described for the determination of 0.40 to 8.0 per cent. of calcium in chrome refractories by atomic-absorption spectroscopy. Separa- tion of those elements which interfere chemically is not necessary if lanthanum is added to suppress their interference. SOME refractory materials contain less than 2 per cent. of silica, and the amount of lime present can have a significant bearing on the mechanical strength of the refract0ry.l For magnesite bricks the lime-to-silica ratio is usually specified as 2 (k0.4) to 1, with similar tolerances for other chrome-bearing refractories.Methods for determining silica in these materials are reasonably well established, but conventional methods currently used for determining calcium oxide at about the 0-5 per cent. level are slow and much less reliable. Calcium has been determined by atomic-absorption spectroscopy in plant materials,2 and in silicates: and it seemed reasonable to assume that a similar procedure could be applied to the determination of calcium in chrome-bearing materials. EXPERIMENTAL APPARATUS- The equipment used was a single-beam, Techtron AA-4 atomic-absorption spectrophoto- meter, with a calcium hollow-cathode lamp. Operating conditions were: slit width, 0.05 mm; wavelength, 422.7 nm; and lamp current, 9 mA. A fuel-rich, pre-mixed air - acetylene flame, with an AB-41 stainless-steel burner, was also used.SPECIAL SOLUTIONS- Standard calcium soZutiofi-Dissolve 2.497 g of calcium carbonate in 10 ml of hydrochloric acid (1 + 4), dilute the solution with water to 1 litre and mix. Transfer a 100-ml aliquot of this solution into a 1-litre calibrated flask, dilute the solution to the mark with water and mix. Lanthanum sohtion, 5 $er cent.-Dissolve 58.65 g of lanthanum oxide (La203) in 250 ml 1 ml of solution = 100 pg of calcium. of concentrated hydrochloric acid, dilute the solution to 1 litre with water and mix. PREPARATION OF CALIBRATION GRAPH- Transfer 1.5 g of anhydrous sodium carbonate and 1.5 g of anhydrous disodium tetra- borate (Na,B,O,) into a beaker and dissolve these reagents in 50 ml of 2 N sulphuric acid.Boil the solution gently to expel carbon dioxide, cool, dilute to 250 ml in a calibrated flask and mix. Transfer 10-ml aliquots into a series of 100-ml calibrated flasks, and to each add 20 ml of the lanthanum solution (5 per cent.). Add, successively, 0, 0-5, 1.0, 2.0, 3.0, 4-0, 5.0, 6.0, 7-0 and 8-0 ml of the standard calcium solution to cover the range 0 to 8 p.p.m. of calcium in the final solution. Dilute each solution to the mark and mix. Measure the absorbances at 422.7 nm. The calibration graph is usually linear and passes through the origin. 0 SAC and the author.390 SIDIROPOULOS DETERMINATION OF CALCIUM IN CHROME [A?Za&St, VOl. 94 PROCEDURE- Transfer 250mg (200 mesh) of the sample into a s m d platinum dish, add 1-5g of anhydrous sodium carbonate and 1 6 g of anhydrous disodium tetraborale, mix, then fuse the mixture and cool.Dissolve the melt in 50 ml of 2 N sulphuric acid and proceed as described under Preparation of calibration graph. RESULTS AND DISCUSSION Table I shows the calcium oxide values obtained on the N.B.S. No. 103 chrome-refractory sample ; values are averages of triplicate determinations on five independently weighed samples. TABLE I ANALYSIS OF N.B.S. CHROME-REFRACTORY SAMPLE NO. 103 Experiment No. 1 .. .. .. 2 .. .. .. 3 .. .. .. 4 .. .. .. 6 .. .. .. Mean . . .. N.B.S. certified Lean N.B.S. standard deviation .. .. .. .. .. .. * . .. .. .. .. .. .. .. .. .. Calcium oxide, per cent. 0.76 0.81 0.78 0.80 0.80 0.79 0-79 0.02 The above sample has the following percentage composition: Cr203, 36-97 ; Al,O,, 20-83 ; total iron, as FeO, 14-39; MnO, 0.21; MgO, 16-27; CaO, 0.79; SiO,, 8-24; TiO,, 0-93; ZrO,, 0.07; and P,O,, 0.7.Of the elements present in this type of sample, aluminium, titanium, silicon and zir- conium are likely to cause the greatest interference, but this can be eliminated either by a prior isolation of the calcium, or the addition of a suppressing agent to the sample solution. The latter has the advantage of rapidity and simplicity; efficient suppressors have been suggested by Elwell and Gidley,* and these include EDTA, strontium and lanthanum. The effect of elements present in chrome refractories was investigated, and Table I1 shows the decrease in absorption caused by the presence of individual cations and the com- bined effect of all cations, each at levels normally present in chrome refractories.The amount of lanthanum used is adequate to eliminate the effect of the interfering cations. TABLE I1 ABSORPTION OF SOLUTIONS CONTAINING 4 P.P.M. OF CALCIUM WITH OTHER ADDITIONS Element added Nil .. .. .. Zirconium .. .. Titanium . . .. Aluminium . . .. Silicon .. .. Chromium .. .. Manganese .. .. Iron . . . . .. Magnesium . . .. r Amount, p.p.m. Absorption, per cent. .. - 43 .. 0.05 <1 0.5 <1 .. 6 13 60 <1 .. 6 18 60 2 .. 60 2 .. 60 28 .. 60 33 .. 60 41 .. 60 48 .. 10 23 .. .. .. .. .. .. .. .. .. 60 10 60 4 63 63 M* represents a combination of all cations in the expected concentra- tion in chrome refractories, in the presence and absence of 1 per cent. of lanthanum.May, 19691 REFRACTORIES BY ATOMIC-ABSORPTION SPECTROSCOPY 391 Table I11 summarises the effect of acid concentration on the absorption of a solution containing 4 p.p.m.of calcium and shows that absorption varies with the concentration of hydrochloric acid, but remains constant in a nitric acid solution of between 0.5 and 2 N. TABLE I11 EFFECT OF ACIDITY Absorbance of a 4 p.p.m. calcium solution Aqueous .. .. .. .. 0-26 Hydrochloric acid, 0.5 to 2 N Nitric acid, 0.5 to 2 N . . .. 0-17 Sulphuric acid, N . . .. .. 0-14 . . 0-21 to 0-13 The author thanks the South African Iron and Steel Industrial Corporation for per- mission to publish this work, and also thanks Mr. P. J. Malan for supplying the samples and Mr. J. S. du Toit for his assistance. REFERENCES 1. Kriek, H. J. S., and Segal, B. B., Trans. Brit. Ceram. Soc., 1967, 66, 66. 2. David, D. J., Analyst, 1969, 84, 636. 3. Trent, D. J., and Salvin, W., Atomic Absorption Newsletter, 1964, 3, 118. 4. Elwell, W. T., and Gidley, J. A. F., “Atomic-absorption Spectrophotometry,” Pergamon Press First received August 4th, 1967 Amended December 16tlr, 1968 Accepted December 16th, 1968 Ltd., Oxford, London, New York and Paris, 1961, p. 86.

ISSN:0003-2654    

DOI:10.1039/AN9699400389

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

392 Analyst, May, 1969, Vol. 94, $9. 392-396 The Chromatographic Separation of Sodium, Potassium, Calcium and Magnesium Ions Applied to the Identification of Soluble Salts in Clays and Clay Products, and of Efflorescent Salts BY H. A. FONER* (The Houldsworth School of Agglied Science, The University, Leeds 2) Paper and thin-layer chromatographic methods have been used to detect and identify the soluble salts in clays and in burnt clay bricks, and to charac- terise the efflorescent salts on clay building bricks. Both techniques are simple and have advantages over conventional methods, especially with large numbers of samples. The novel thin-layer separation of the ions is more rapid and sensitive than that achieved by paper chromatography. TWO common faults in brick structures, namely, the appearance of efflorescent salts on the bricks and sulphate expansion of the Portland cement mortar, can be attributed to excessive amounts of soluble salts in the bricks.The soluble salts present in fired bricks, raw clays and efflorescences are the sulphates of sodium, potassium, calcium and magnesium, with occasionally a little iron and silica.1 There is no simple relationship between the composition and amount of efflorescent salts on a brick and the soluble salt content of either the fired brick or the raw clay from which it is made.2~3 The likelihood of efflorescence developing depends not only on the soluble salt content of the raw material, but on such variables as the temperature and duration of firing, the weathering conditions in service and the mortar that bonds the bricks.This also applies to sulphate expansion.2~3 Further, the amount of salts extracted from a clay or brick depends markedly on the method of extraction. Various methods have been proposed, which involve stirring or shaking6 the sample with water. The current British Standard method for bricks6 is similar to that described in the Procedure (below), except that 10-g instead of 20-g samples are used. Obviously, quantitative values for soluble salts can be compared only for standardised extraction conditions. In general, it is known that excessive amounts of soluble salts can be troublesome, and it is usual to check the soluble salt contents of both clays and bricks. Similarly, it is often necessary to analyse efflorescent salts as a check on their origin.For the reasons mentioned above, it is rarely necessary to make an accurate determination of each constituent. A simple and rapid identification of the ions present and an approximate estimation of their relative abundance are all that is required. The anion present is usually sulphate, and simple tests are readily available for it and also for any iron present. It seemed desirable to develop simple and rapid methods for separating sodium, potas- sium, calcium and magnesium ions in soluble salt and efflorescence analysis. This paper describes a paper-chromatographic method suitable for such samples and the separation of the ions by thin-layer chromatography on a cellulose substrate. Both methods are simple, rapid and suitable for a large number of samples. The latter technique is both faster and more sensitive than paper chromatography. The relative abundance of the ions present can be gauged approximately from the size of the spots on the chromatogram.6 The chromatographic techniques described are much less costly than alternatives such as atomic-absorption or flame spectroscopy.* Present address : Geochemistry Department, Geological Survey of Israel, 30 Malchei Israel Street, Jerusalem . 0 SAC and the author.FONER 393 METHOD Anion-exchange resin-Amberlite IRA400 (Cl) . Standard solutions-Prepare solutions of the acetates of sodium, potassium, calcium and magnesium containing 1.0 per cent. w/v of the metal as follows. Add a slight excess of 5 N acetic acid to the amounts given for the following analytical- reagent grade compounds: calcium carbonate, 2-5 g; sodium carbonate, 2.3 g ; potassium carbonate, 1.8 g; and magnesium oxide, 1-7 g.Boil to remove carbon dioxide and make up to 1OOml with distilled water. Standard acetate mixture-Mix 1 O m l of each of the above standard solutions. Violuric acid (5-isonitrosobarbituric acid) reagenb-Add about 043 g of the reagent to 100 ml of ethanol - water (1 + 1). Stir thoroughly and remove any excess by filtration. Solvent for chromatography-Mix 80 ml of ethanol with 20 ml of 2 N acetic acid. The mixture should be made up daily to avoid esterification. Ammonz'zm acetate solution, 10 per cent. w/v, aqueous. PROCEDURE- Preparation of the acetate form of the ion-exchange resivt-Make a slurry of about 40 g of Amberlite IRA400 (Cl) resin and pour it into a 30-cm ion-exchange column.Elute slowly with ammonium acetate solution until all of the chloride has been removed (about 2& litres are needed). Then wash with freshly distilled water until the eluate is free from ammonium ions. REAGENTS- NOTE-It is essential to use carbon dioxide free water for washing the resin. Preparation of the brick or clay extract-Place 20.0 g of the powdered sample (B.S. 100 mesh) into a 200-ml polythene bottle. Add 100 ml of distilled water and shake the bottle mechani- cally for 1 hour. Filter through a dry paper (Whatman No. 42); do not wash the residue. If necessary, repeat the filtration until a clear filtrate is obtained. Make up a small column (7 x l-cm diameter is suitable) of the acetate form of Amberlite IRA400. Use carbon dioxide free water in the preparation of the column and in all subsequent operations.Pass the brick or clay extract slowly through the column and wash with 60 to 100 ml of water. Test a small amount of the eluate to ensure that it is free from sulphate and evaporate the eluate to about 1 ml on a water-bath. Withdraw the concentrated solution with a hypodermic syringe fitted with an unpointed needle. Rinse the evaporating basin with 1 ml of water, drawing this also into the syringe, and mix the two portions of liquid. Make the volume up to a standard amount (e.g., 2.0 or 2-5ml) and mix again. Expel the solution into a test-tube and reserve for chromatographic analysis. Eflorescent salts-Dissolve 0.10 g of the salts in 10ml of water.Pass the solution through the ion-exchange column and wash a s in the preceding section. Evaporate the eluate and make the volume up to 5.0 ml. Paper chromatography-Spot the solutions (see Table I) on to Whatman No. 1 filter-paper and allow the spots to dry. Equilibrate the filter-paper for about 2 hours over the solvent. Dip the paper into the solvent and develop the chromatograph by ascension for about 20 hours; during this time the solvent front rises about 35 cm. Spray with violuric acid reagent and then dry at 55" to 60" C. The spots resulting from the various ions present are now visible and can be recognised by their position and colour. Although the R, values are reproducible, it is advisable to run both standard and mixed acetate solutions in parallel with the test solution.The amounts of solutions used, RF values and colours produced are shown in Table I. TABLE I Remove the filter-paper from the tank and dry at 50" to 60" C. PAPER CHROMATOGRAPHY Volume applied, Weight of metal RF Solution ions present, pg value Colour Standard sodium acetate . . 5 60 0.51 Lilac Standard potassium acetate 6 50 0-42 Lilac Standard calcium acetate . . 5 50 0.60 Red - orange Standard magnesium acetate 5 50 0-68 Yellow Standard acetate mixture . . 10 100 - - Test solution . . .. .. 10 - - -394 FONER : CHROMATOGRAPHIC SEPARATION OF [Analyst, VOl. 94 The calcium and magnesium spots can be rendered more prominent by holding the filter-paper over concentrated ammonia solution, but this causes the sodium and potassium spots to disappear.Thiut-layer chromatogra~hy-Prepare thin-layer plates with a cellulose powder containing no binder. The results shown below (Table 11) were obtained with 300-pm thick layers of Macherey Nagel cellulose, Type MN300, mixed according to the makers' instructions. A "Silverson" high speed mixer was found particularly useful for dispersing the cellulose. Any deviation in materials and techniques will give different results (see under Discussion). Spot the plates with appropriate solutions in a manner analogous to that described in the paper-chromatographic method. Develop the plates in a sandwich chamber by using the solvent ethanol - 2 N acetic acid (80 + 20, v/v). The solvent front takes about 80 minutes to run 10 cm. When development is complete, dry the plate at 50" to 60" C.Allow the plate to cool a little, spray with violuric acid reagent and warm again to 60" C, if necessary. Typical results are given in Table 11. TABLE I1 THIN-LAYER CHROMATOGRAPHY Volume applied, Weight of metal RF Solution Pl ions present, pg value Standard sodium acetate . . 1 10 0-41 Standard potassium acetate 1 10 0-30 Standard calcium acetate . . 1 10 0.46 Standard magnesium acetate 1 10 0.62 Standard acetate mixture . . 2 20 Test solution . . .. .. 2 - - - Colour Lilac - blue Lilac Orange Yellow - brown - - DISCUSSION THE SEPARATION OF ALKALINE EARTH AND ALKALI METAL IONS- Several ion-exchange and paper-chromatographic separations of the alkaline earth and alkali metal ions had been reported at the start of this work. The paper-chromatographic method appeared to have great advantages, especially if it were possible to find a single reagent capable of identifying all of the constituents of interest.Methods were available7~* that depended on the chromatography of the metal chlorides and their detection by exposure to light or photographic developers after spraying with silver nitrate solution. More promising alternatives were the method of Erlenmeyer, von Hahn and Sorkin9 and similar methods,loJ1 involving the chromatography of the acetates and their detection and identification with violuric acid. The present method is based on this work. More recently, some work has appeared on the separation of alkali metal and alkaline earth ions by thin-layer chromatography, and this is discussed below.Butterworthlz has published an elegant method for the microscopic identification of efflorescent salts, and discussed their formation.13 An electrophoretic method for the determination of soluble salts in building materials has also been described.l* COLOUR-DEVELOPING REAGENT- Only two reagents appear to form coloured compounds with all of the metals of interest, zliz., violuric acid, which has been known for over a century,l5 and lithium tetracyanoquho- dimethanide, described by Drudingls in 1963. The latter reagent seems to have no advantage for the present work, especially as Druding states that one of the disadvantages of violuric acid is the necessity of heating the chromatograms for 30 minutes at 120" C to develop the colour. The author, and other^,^ have found that the colours appear more quickly and at much lower temperatures. The properties of violuric acid have been the subject of controversy.Solutions of the reagent have been variously reported as blue17 and colourless.18 There also seem to be doubt about the colours of the various salts of violuric acid. Table I11 shows the colours reported by various authors. Some of the confusion may be caused by the formation of mono and dibasic salts.19May, 19691 SODIUM, POTASSIUM, CALCIUM AND MAGNESIUM IONS 395 TABLE I11 COLOURS REPORTED BY VARIOUS AUTHORS FOR VIOLURIC ACID SALTS Author Sodium Potassium Calcium Magnesium Welcher20 . . .. .. . . Red Blue Red Carmine red . . . . Orange Red Orange Yellow Erlenmeyer, von Hahn and Sorkino . . Violet Violet Orange Yellow - red Finaral .. .. .. .. .. - Blue - Purple Foner (see Table I) . . .. . . Lilac Lilac Red - orange Yellow Foner (see Table 11) . . .. . . Lilac - blue Lilac Orange Yellow - brown Druding16 . . .. Violuric acid can be purchased or readily synthesised from alloxan and hydroxyl- amine.22s23s24 Fresh solutions are colourless, but gradually turn blue if kept in glass. PAPER CHROMATOGRAPHY- The method used is basically that of Erlenmeyer and his ~o-workers.9,~O The relationship between RF values and solvent composition is similar to that discussed under Thin-layer chromatography. A typical separation is shown in Fig. 1. Iron salts remain on the base-line of the chromatogram. 0 0 0 0 0 0 Na+ K+ Ca2+ Md' Mixture i Na' K+ Ca" Mg2+ Mixture Fig. 2. Thin-layer chromato- Fig.1. Paper-chromatographic graphic separation of sodium, potas- separation of sodium, potassium, cal- sium. calcium and magnesium ions cium and magnesium ions individually Time and as a mixture. Time 20 hours 80 minutes individually and as a mixture. THIN-LAYER CHROMATOGRAPHY- The application of thin-layer chromatography to inorganic analysis has been relatively neglected. The field has been reviewed,2692s and a few schemes for the analysis of metal ions on silica-gel layers have been de~cribed.27,~~ Other workers have separated alkali metals on heteropoly acids,29 alkali metals and magnesium,3O and alkali metals and alkaline earths16 on silica gel. In the last two instances it was found necessary to purify the commercial silica gel before use.To obviate the purification of the adsorbent and to obtain better separation, it was decided to adapt the paper-chromatographic method described above to the thin-layer. technique by using a cellulose substrate. The 20 x 20-cm plates were mostly developed in a Camag sandwich chamber. Experi- ments showed that although the RF values obtained are sensitive to solvent composition,396 FONER the separation of ions is almost constant in the range ethanol - 2 N acetic acid (9 + 1, v/v) to (7 + 3, v/v). A typical separation is shown in Fig. 2. A slight displacement effect occurs when a mixture of ions is chromatographed. The need for rigorously standardised conditions was avoided by always running a known mixture with the unknown samples. The type of absorbent used markedly affects development times and RF values.31 Most of the chromatograms in this work were run using Machery Nagel MN300 cellulose, which formed adherent layers.Whatman cellulose powder CC41 also gave good separations with a faster running time (about 40 minutes), but the layers were f r e e . Both of these adsorbents could be used without purification. Much better resolution was obtained on cellulose layers than on silica gel. The author thanks Griffin and George Ltd. for donating the Camag thin-layer chromato- graphic equipment used in this work. 1. 9 ”. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 26. 26. 27. 28. 29. 30. 31. REFERENCES Bennett, H., and Hawley, W. G., “Methods of Silicate Analysis,” Second Edition, Academic Press, British $;tandard 3921 : 1966, “Specification for Bricks and Blocks of Fixed Brickearth, Clay or Ragsdale, L. A., and Raynham, R.A., “Building Materials Practice,” Edward Arnold Ltd., British Standard 1257 : 1946, now replaced by British Standard 3921 : 1966, reference (2). British StaEdard 3921 : 1966, “Specifications for Bricks and Blocks of Fixed Brickearth, Clay Purdy, S. J., and Truter, E. V., Lab. Pract., 1964, 13, 500. Tristram, D. R., and Phillips, C. S. G., J. Chem. Soc., 1955, 680. Burstall, F. H., Davies, G. R., Linstead, R. P., and Wells, R. A., Ibid., 1960, 616. Erlenmeyer, H., von Hahn, H., and Sorkin, E., Helv. Chim. Actu, 1961, 34, 1419. Seiler, H., Sorkin, E., and Erlenmeyer, H., Ibid., 1952, 35, 120. Butterworth, B., Trans. Brit.Ceram. S O ~ . , 1933, 32, 270. Howard, J. M. H., Chem. 6. Ind., 1967, 688. Baeyer, A., Justus Liebigs Annln Chem, 1863, 127, 199. Druding, L. F., Analyt. Chem., 1963, 35, 1682. Donnan, F. G., and Schneider, W., J. Chem. SOL, 1909, 95, 966. Hartley, W. N., Ibid., 1906, 87, 1796. Hantzch, A., and Isherwood, P. C. C., Ber. dt. chem. Ges., 1909, 42, 986. Welcher, F. J., “Organic Analytical Reagents,’’ D. Van Nostrand Co. Inc., New York; Macmillan Finar, I. L., “Organic Chemistry,” Longmans, Green & Co., London, 1956, Volume 11, p. 461. Ceresole, M., Ber. dt. chem. Ges., 1883, 16, 1133. Guinchard, J., Ibid., 1899, 32, 1723. Pollard, F. H., and McOmie, J. F. W., “Chromatographic Methods of Inorganic Analysis: with Special Reference to Paper Chromatography,’ ’ Butterworths Scientific Publications Ltd., London, 1953, p. 57. Pollard, F. H., Burton, K. W. C., and Lyons, D., Lab. Pract., 1964, 13, 605. Truter, E. V., “Thin Film Chromatography,” Cleaver-Hume Press Ltd., London, 1963, p. 178. Takitani, S., Ja@an Analyst, 1963, 12, 1166; Anulyt. Abstr., 1965, 12, 1057. Seiler, H., in Stahl, E, S., Editor, “Thin Layer Chromatography,” Academic Press, New York Lesigang, M., Mikrochim. Acta, 1964, 34, 43. Seiler, H., and Rothweiler, W., Helv. Chim. Acta, 1961, 44, 941. Wollenweber, P., Lab. Pract., 1964, 13, 1194. London and New York, 1966, p. 103. Shale, London, 1964, p. 110. or Shale, p. 34. British Standards Institute, London, p. 9. , 3 , Ibid., 1962, 35, 2483. --- -, Ibid., 1954, 53, 600. & Co. Ltd., London, 1947, Volume 111, p. 287. and London, 1965, p. 469. First received December 18th, 1967 Amended November Id, 1968 Accepted December 16th, 1968

ISSN:0003-2654    

DOI:10.1039/AN9699400392

出版商:RSC    

年代:1969

数据来源: RSC