ISSN:0003-2654    

DOI:10.1039/AN97499FX013

出版商:RSC    

年代:1974

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN97499BX015

出版商:RSC    

年代:1974

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN97499FP037

出版商:RSC    

年代:1974

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN97499BP043

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

APRIL, 1974 Electrodes pH meter 1 Glass Vol. 99, No. 1177 PH ~ Pen 1 atlenuator THE ANALYST Bucking potential Reference The Simultaneous Measurement of pH, Chloride and I- Switch holding 1 1- Switch - Electrolytic Conductivity in Soil Suspensions ~- pH I Chloride I ' attenuator 2 . BY s. MCLEOD, H. c. -r. STACE, B. M. TUCKER AND P. BAKKER (Commonwealth Scientific and Industrial Research Organization, Division of Soils, A delaide, A ustralia) The measurements of pH, chloride and electrolytic conductivity have been simplified and made simultaneous for suspensions of soil in water through the use of a triple electrode system mounted in a single unit. A glass electrode and a silver - silver chloride electrode with a common reference electrode and two pH meters are used for the determination of pH and chloride, respectively.Electrolytic conductivity is measured by an ohm-meter principle using silver electrodes. The outputs of the three meters are recorded on a three-pen recorder with electrically independent channels. The pH, over the range 0 to 10, is read froin the chart while the values for chloride and electrolytic conductivity are obtained from graphs or tables. Once the instruments are set up they need little adjustment during the day. WE have developed an apparatus for the simultaneous measurement of pH, chloride and electrolytic conductivity in aqueous soil suspensions. This apparatus has greatly reduced the time taken for these analyses because previously these determinations were carried out separately on aliquots of the same suspension in a manner similar to that described by Piper.1 Soil suspensions are preferred to pastes, which are not adaptable to rapid standardised procedures and which have uncertainties in liquid junction potentials.The measuring electrodes are mounted together in a plastic holder. The glass electrode (for measurement of pH) and the silver - silver chloride electrode (for measurement of chloride) with a cormnon calomel reference electrode are connected to two pH meters and then through suitable attenuator circuits to two pens of a three-pen potentiometric recorder. The pens of the recorder must be electrically independent. The electrolytic conductivities are measured by the current between two silver electrodes from a solid-state oscillator using the ohm-meter principle, This technique is preferred to the use of a conductivity bridge because it requires no change in the ranges used.The oscillator current is recorded by the third pen of the Pen 2 - 3 Oscillator 193 Pen -194 [Analyst, Vol. 99 recorder. In order to prevent damage to the pens, the electrode system is automatically isolated from the recorder while samples are being changed. A block diagram of these arrangements is shown in Fig. 1. The pH and chloride attenuator circuits and the oscillator circuit, together with the bucking and base-line holding circuits, are incorporated in one unit, the attenuator box. The functions of the bucking and base-line holding circuits and the isolating switch are explained below. As some soil suspensions contaminate the silver - silver chloride electrode, this electrode is washed between determinations with a dilute solution of sodium chloride.The holder and electrodes are then washed with distilled water and the electrodes dried. The whole opera- tion can be carried out manually or semi-automatically. The precision of each measure- ment with this equipment has been found to be satisfactory. ELECTRODE SYSTEM ELECTRODE HOLDER- Two types of electrode holder have been developed, the first of which (type A) is of simple design in which the washings between determinations are effected by jets from wash- bottles and the system is dried with soft tissue-paper, while the second (type B) incorporates channels and small jets through which the wash liquids and air for drying the system are forced.These jets are positioned so that all surfaces of the electrodes and the holder are washed. The flow of wash liquids through the jets is controlled by non-return valves. MCLEOD et d.: SIMULTANEOUS MEASUREMENT OF PH, CHLORIDE .'u \ 3 ! 3 ' 4 5 d 8 7 5 4 ! : j I I I - Electrode holder Sample bottle 1 = Silver - silver chloride electrode with a 5 = Terminal block for conductivity electrodes 2 = Glass electrode with a rubber holding collar 3 = Calomel electrode with rubber holding 8 = Air vent collar inserted into block to prevent wash entering the filling hole 4 = Silver conductivity electrodes cemented into wall of holder Perspex holding collar 6 = Perspex block with O-ring to hold electrodes 7 = Wash holes 9 = Micro-switch with adjustable height 10 = Overload spring allowing for reasonable changes in bottle size Fig.2. Electrode holder, type A: (a) side elevation; (b) plan; and (c) micro-switch Type A-This holder is made in two parts from Perspex [Fig. 2 (a) and (41. The outer tube is 17 cm in length and 3 cm in diameter with a wall thickness of 3 mm, and has threeApril, 19741 AND ELECTROLYTIC CONDUCTIVITY I N SOIL SUSPENSIONS 195 holes for washing the electrodes between measurements. The silver electrolytic conductivity electrodes, made from 1.6 mm diameter silver wire, are set with Araldite in slots milled in the outer surface of this tube and tested for electrical leakage between them. They are about 2 mm apart and are bent so that their exposed faces are at the inner surface of the tube, a distance of about 5 mm from the bottom.They are connected to a terminal block at the top of the tube. Although platinum is the usual metal for use in conductivity electrodes, it could not be used in this work because it gave drifting readings. The glass, the silver - silver chloride and the calomel electrodes are mounted in holes in a block 2-3 cm in length. The glass and the calomel electrodes are held in place by rubber collars, and with the calomel electrode this collar also acts as a seal to prevent water from the washing procedure entering the filling hole of the electrode. The silver - silver chloride electrode is held in place with a Perspex collar. The block is supported in the outer tube by an O-ring seal and is provided with asmall air vent. Except for the electrolyticconductivity electrodes, these dimensions are not critical and can be altered to suit the suspension bottles provided that sufficient space is allowed to accommodate the electrodes.The glass and calomel electrodes are placed as far as possible from the electrolytic conductivity electrodes. This arrangement minimises the effect on the cell constant if the electrodes have to be replaced and reduces the electrical interference by the conductivity current on the electrode potentials. Type B-This holder is made of Perspex in five parts [Fig. 3 (a) and ( b ) ] , which are glued together with a suitable Perspex cement so as to form a complete unit. In our unit, the external diameter of the upper cylinder is 6.5 crn with a wall thickness of 6 mm and the length is 3.5 cm.The lower cylinder has an external diameter of 5.5 cm with a wall thickness of 3 mm and is 6.5 cm in length. This lower cylinder supports the electrolytic conductivity electrodes, which are made of silver wire as described for the type A holder, but in this unit these electrodes are soldered with silver to flexible insulated copper leads that pass through the various parts of the unit to one of the micro-switches (see Electrical circuits). A small air-vent is drilled near the top of the lower cylinder. The wall of the lower cylinder contains a channel of small diameter, which is connected through the various parts of the unit to a channel in the side-piece that supports the micro- switches. This channel is for the sodium chloride wash and it is also connected to the water and air supplies.The lower part of the channel has several small outlet jets, which are directed towards the silver - silver chloride electrode. The block that supports the glass, the silver - silver chloride and calomel electrodes in the same manner as described for the type A holder is held between these cylinders and is 2.5 cm in depth. This block contains an interconnected upper and lower set of channels. Each set has three channels, which radiate from the centre of the block as shown in Fig. 3 (b). These channels are connected to the water and air supplies by the side inlet 9. A channel 0-5 cm deep is machined into the bottom of the upper cylinder, and when the electrode block is fitted to the upper cylinder an annular space is formed to which the upper channels supply either water or air, as required. The lower set of channels arc also connected to a similar annular space between the block and the lower cylinder.Small jets, which enter the annular spaces, are placed at 0.5-cm intervals around the upper and lower rims of the electrode block and these jets are directed towards the outer and inner surfaces of the lower cylinder. The angle at which these holes are drilled is critical and should be 22 & &". This angle has been found the most satisfactory to ensure an even flow of water down the walls of the cylinder. Small jets that are directed towards the electrodes are also drilled in the lower set of channels. A small shield is placed at the back of the silver - silver chloride electrode so as to protect the other electrodes during the washing with sodium chloride solution.In the simplest arrangement that we have used for the washing procedure with this unit, the sodium chloride solution is dispensed from a bottle equipped with a squeeze-bulb, the distilled water is pumped from a large container by a small centrifugal pump and the flow is controlled by a solenoid valve. The pump and valve operate together when the pump is switched on for the water wash. Non-return valves are also placed in the lincs as close as possible to the holder and ensure that any liquid that remains in the lines does not flow into the unit during the measurements. The air for drying the whole system must be free from oil and is supplied at low pressure by a simple on - off valve.196 MCLEOD et al.: SIMULTANEOUS MEASUREMENT OF PH, CHLORIDE [Analyst, Vol. 99 5 6 1 1 12 7 8 5 (bJ 1 = Upper cylinder 6 = Glass electrode 10 = Water and air inlet 2 = Block 3 = Lower cylinder trode 12 = Lever arm 4 = Silver electrolytic conduc- 8 = Chloride electrode wash 5 = Calomel electrode 9 = NaCl and water inlet 7 = Silver - silver chioride elec- I 1 = Ganged rnicro-switches tivity electrodes shield Fig. 3. Electrode holder, type B: (a) side elevation; (17) plan GLASS ELECTRODE- The glass electrode is 15 cm long and has a 6 mm diameter stem with a bulb no larger than the stem diameter. The glass of the electrode is chosen so as to give a rapid response (105AR, Titron Instruments, Australia). An electrode of small dimensions is used in order to facilitate washing and to niinimise liquid hold-up.SILVER - SILVER CHLORIDE ELECTRODE- This electrode is made from a length of about 15 cm of silver wire of 1.6 mm diameter. A length of about 2 cm at one end of this wire is cleaned in dilute nitric acid until effervescence just begins, and is then washed thoroughly with distilled water. A length of about 1 cm of this cleaned portion is immersed in 0.1 N hydrochloric acid and a current of 3 to 4 mA is passed for 1 minute between this wire and a platinum cathode, which results in a thin brown coating of silver chloride that is less susceptible to contamination than a thick coating. After washing with distilled water, the electrode is ready for use. We have recently found that the stability of this electrode can be improved by standing it for a few hours (overnight) in distilled water.April, 19741 AND ELECTROLYTIC CONDUCTIVITY IN SOIL SUSPENSIONS 197 The silver chloride coating is gradually dissolved and eroded by continued use, but experience has shown that the performance of the electrode is satisfactory for up to 200 deter- minations.The used electrode is stripped by wiping it with a piece of soft material and re-coated. REFERENCE ELECTRODE- The reference electrode is common to the pH and chloride cells and is operated at earth potential. A calomel electrode (K401, Radiometer, Denmark) is used and it must have only a low flow-rate of potassium chloride solution through the liquid junction so as to prevent errors in the chloride and electrolytic conductivity measurements.A mercury - mercury(1) sulphate electrode (K601, Radiometer) was used originally as it avoids the risk of chloride contamination, which is possible from a calomel electrode. However, comparative measurements gave soil pH values from 0.3 to 0-6 unit lower than a conventional arrangement in which a calomel reference electrode was used. This error was found to be due to the use of potassium sulphate to form the liquid junction of the mercury electrode. The potassium and sulphate ions have unequal transference numbers, so that the liquid junction potential is not almost zero, as assumed with potassium chloride, but is dependent on the ionic content of the external solution. The error is made up of two com- ponents: a suspension effect, due to the presence of charged particles in the test solution and which is generally small, giving values about 0.1 pH unit too low, and a standardisation error due to differences in ionic content between the standardising buffer and the test solution, which gives pH values that are too low by about 0.2 to 0.5 unit.A similar effect could also occur in the measurement of chloride content with the silver - silver chloride electrode, but this would be partly compensated for by the empirical calibration of the electrode potentials against standard chloride solutions to the extent that the soluble salts may contain a large and fairly constant proportion of chloride. It is possible to minimise these liquid junction effects by making the soil suspensions in dilute salt solutions, but this expedient inhibits the use of conductivity measurements for soluble salts.The use of mercury - mercury( I) sulphate electrodes was therefore discontinued and the customary calomel electrodes were used. Tests on these calomel electrodes showed that detectable errors in the chloride concentration appeared only after standing for about 20 minutes and then only in the standard with the lowest chloride content. This result agreed with the earlier experi- ence that diffusion of salts from the liquid junction did not affect the conductivity measure- ments in unstirred suspensions over much longer periods than the 1 to 2 minutes required to complete the set of analyses. New reference electrodes must nevertheless be tested for excessive leakage before being put into regular service.MOUNTING THE HOLDERS- The electrode holder is mounted vertically, in an earthed aluminium box (open at the front), so that the electrodes are immersed in the top 1 cm of soil suspension when the bottle is in place on the stand. On no account should the electrodes enter the sediment, as the glass electrode may be damaged and the conductivity electrodes will give completely false readings. Preferably, the calomel reference electrode should just enter the liquid surface. The lever-arm of the micro-switch is adjusted so that the top of the bottle operates it when it is placed in position or removed. It is desirable that the micro-switch operates after the electrodes are immersed and before they emerge from the suspension. The box has a drain to take the wash liquids to waste.ELECTRICAL CIRCUITS The complete circuit is shown in Fig. 4; two Radiometer pH meters, Model PHX28, are used. The use of other meters will require appropriate changes in the values of the attenuator components. Meters with high impedance recorder outputs (above 30 ktZ) are not satis- factory with the present recorders. It is possible that such meters could be used with high input-impedance recorders, but we have not tried this arrangement. The potentiometric recorder has three electrically independent channels, each with a minimum span of 10mV and a full-scale zero adjustment (B34, Rika Denki). Chart speeds of 5, 10 and 20 mm min-l are suitable. The pH meter 1 is direct-reading with a recorder output. The meter scale is not folded in the soil pH range so as to avoid off-scale readings that might damage the meter.An198 MCLEOD et al.: SIMULTANEOUS MEASUREMENT OF PH, CHLORIDE [Analyst, Vol. 99 adjustable attenuator is used to reduce the output to 2 mV per pH unit, giving a working range of 20 mV for 10 pH units. The resistor values in the attenuator are matched to the pH meter output and the recorder input. When the soil suspension bottle is removed from the electrodes, the micro-switch [Figs. 2 (c), 3 (a) and 3 ( b ) ] substitutes a 20-rnV output from the base-line holding circuit (supplied from a 1-5-V dry cell) for the pH meter output to the recorder. This output drives the recorder pen to full scale in order to avoid irregular fluctu- ations during the washing of the electrodes and to separate the pen traces from each other.The pH channel normally works on the 20-mV range but its sensitivity can be doubled by switching to the 10-mV range. To glass electrode pH span PHM 28 Recorder lack ---- sleeve 0 --&-- OGlasr 4 0 3calomel 1 0 0 pH meter pH recorder 20 mV 8asrline holding To calomel reference electrode Chloride To silver recorder chloride electrode Bucking potential 0 10mV E.C. recorder 10 mV To conductivity electrodes 4.5-V d.c. (shielded twin cable) supply (dry cells) Fig. 4. Circuit and connections for attenuator box used with two Radiometer pH meters, Model PHM28. Switches: SW1 = 3-pole, 3-position and SW2 = 1-pole, 3-position (combined rotary); SW1 and SW2 off = off; SW1 and SW2 on = electrolytic conductivity calibrate; STY1 on and SLV2 off = read-out.SW3 = %pole, 2-position micro-switch; 1, recorder reads; 2, recorder standby The pH meter 2 is also direct-reading with a recorder output. The pH scale is used for log [Cl-] measurements so that the buffer adjustment can be used for standardisation. Like the glass electrode, the silver - silver chloride electrode is connected to the meter with a lead that is shielded to earth so as to protect it from electrical interference. A constant bucking potential of about 400mV is supplied from a dry cell through a high-resistance network in order to bring the electrode potential to a suitable value. An adjustable attenuator reduces the meter output to 4 mV per log [Cl-] unit, giving a working range of 2.5 units. On removal of the soil suspension bottle, the same micro-switch disconnects the meter from the attenuator so that the chloride recorder pen remains at its electrical zero during the washing process. The electrolytic conductivity oscillator circuit has been described by Tucker and Raupach.2 The meter in this original circuit is replaced with a 20-!2 resistor connected in parallel to the third recorder pen.Adjustment of the input voltage to the oscillator to give a full-scale (10 mV) reading on the recorder calibrates the maximum current to 0.5 mA. The oscillator is driven by three 1.5-V dry cells. A mains-operated rectifier supply was originally used but fluctuations in the mains voltage gave considerable instability, which could not be overcome by using a constant-voltage transformer. The micro-switch disconnects the electrodes during the washing cycle so that the recorder pen returns to zero between readings.April, 19741 AND ELECTROLYTIC CONDUCTIVITY IN SOIL SUSPENSIONS 199 CALIBRATION PH- The pH recorder is standardised against two buffers so that the pH can be read directly from the chart.The pH meter has a value for which the recorder output is zero, and the meter is standardised with a buffer whose pH is close to this zero by means of its buffer adjustment. For the Model PHM28 pH meter, the zero output is at yH 8 and a buffer of pH 7.78 is used. The recorder is adjusted to read the same value by means of its zero adjustment. After washing the electrodes with distilled water, a second buffer, with a pH as far from the first as possible (pH 4-00), is used to check the correctness of the pH meter reading.Small errors can be corrected by adjusting the sensitivity control or temperature setting; large errors indicate that the glass electrode is faulty. The recorder is then set to read the buffer value by means of its pH span control. When the recorder is first set up, it may be necessary to repeat this calibration procedure so as to obtain complete corre- spondence between the buffer pH values, pH meter reading and chart reading, but subse- quently it may only be necessary to make day-to-day calibrations with the buffer adjustment. A third buffer is used occasionally in order to check the settings, the other two buffers and the linearity of the glass electrode potentials. CHLORIDE- The chloride meter cannot be matched against the recorder because of the scale expansion used.The recorder is set to a fixed value for the lowest chloride standard using the recorder zero adjustment and the buffer adjustment of the meter. The pen is set to a second fixed value for a second chloride standard using the chloride span control. A third standard is used as a check. For soils, convenient values for these standards are 10,1000 and 100 mg 1-1 of chloride made from sodium chloride; these values are equivalent to 25, 2500 and 250 mg kg-1 of chloride in the soil when a ratio of soil to water in the suspension of 1 : 2-5 is used. A plot of chart reading against the logarithm of the chloride concentration in the soil is used to calculate the results. Provided that the chart is always standardised at the same points and that the silver - silver chloride electrode is kept in good condition, this plot is a straight line.It is then more convenient to use a table that relates chart reading to chloride content. The electrolytic conductivity recorder pen is set at zero with the oscillator switched off and it should remain at zero when the oscillator is switched on with the electrodes in distilled water. In the calibration setting (electrodes short-circuited), the recorder pen is set to full scale (10 mV) by its calibration control. A plot of chart reading against electrolytic conduc- tivity of standard salt solutions is used for the calculation of the results. The electrolytic con- ductivities of standard sodium chloride solutions at 20 and 25 "C are given in Table I.ELECTROLYTIC CONDUCTIVITY- TABLE I ELECTROLYTIC CONDUCTIVITY OF SODIUM CHLORIDE SOLUTIONS CALCULATED FROM DATA I N THE INTERNATIONAL CRITICAL TABLES Electrolytic conductivity/mS cm-I NaCl 7- concentration/M 20 "C 25 "C 0*0005 0.001 0.002 0.005 0.01 0-02 0.05 0.07 0.10 0.20 0.50 0.0562 0.1114 0.2208 0.542 1.065 2-082 4.995 6.86 9.60 18.30 42.2 0.0625 0.1241 0.2460 0.604 1.186 2-316 5.550 7.63 10.66 20.30 46.66 Regular checks on standard solutions are required so as to ensure that the true values are being recorded. For these calculations also, a table can be used instead of the chart.200 [Amlyst, Vol. 99 The state of the supply batteries is indicated by the position of the calibration control and they must be replaced as soon as they begin to fail, as the shape of the calibration curve may change when faulty batteries are used.The instrument is less precise at high conduc- tivities (above about 3 mS cm-l) and dilution of the suspension gives better values in such instances. SAMPLE MEASUREMENT Weigh the sample of soil into a suspension bottle that contains the required amount of aerated distilled water and shake the bottle for 1 hour. During this period, the apparatus is set up and calibrated ready for use. After shaking, allow the suspension to settle for about 0-5 hour. Place the first bottle under the electrodes and raise it to the correct height by using a stand; the rim of the bottle should then operate the micro-switch. Either when the readings are steady, or after 2 minutes if some drift occurs, the suspension bottle is lowered, thus releasing the micro-switch.The silver - silver chloride electrode is washed with a 2 g 1-1 solution of sodium chloride before washing the whole electrode system thoroughly with distilled water. Periodic checks on one pH buffer, one chloride standard and the electrolytic conductivity calibration should be made and, in addition, it is good practice to include a standard soil each day. COMPARISON WITH USUAL METHODS Comparisons of values for electrolytic conductivity and chloride content for twenty soils obtained with this simultaneous system and those obtained by the methods used previously in these laboratories are shown in Fig. 5 ( a ) and (b). The results, together with many others that are not shown, were considered to be acceptable and the simultaneous method has now been in use for 4 years.MCLEOD et al.: SIMULTANEOUS MEASUREMENT OF PH, CHLORIDE 100 1000 10000 Chloride by simultaneous systemhg I-' 0 1 2 3 4 5 6 Electrolytic conductivity by simultaneous system/m s cm-' Fig. 5. (a) Comparison of chloride determinations by electrometric titration and simultaneous system. (6) Comparison of electrolytic conductivity determinations by conductance bridge and simultaneous system PRECISION OF THE METHODS In order to estimate the precision of the values for pH, chloride content and electrolytic conductivity obtained by this simultaneous method, twenty soil samples were analysed by two operators on each of two occasions (about 1 month apart). Each occasion was divided into a morning and an afternoon run and the operators changed runs between the occasions.Each soil was analysed in duplicate during each run and a different random order for the observations was used for each run, that is, eight separate estimations for each variable were obtained for each soil sample. The results are presented in Table I1 in the form of analyses of variance showing the observed and expected mean squares assuming a random model for operators, occasions and repeat determinations made on the same day.3 The results showed no effect due to,4pril, 19741 AND ELECTROLYTIC CONDUCTIVITY IN SOIL SUSPENSIONS 20 1 operators, occasions or any interaction for pH and electrolytic conductivity. For pH, there was a significant effect due to repeat runs on the same day. For chloride, there was a signifi- cant effect due to repeat analyses on the same day and a significant interaction between operators and occasions, although neither effect alone was significant.TABLE I1 ANALYSIS OF VARIANCE AND DEFINITION OF VARIANCE COMPONENTS hIean square Degrees due to freedom Variation of Operators (OP) Occasions (OC) OP x oc Between runs within OP x oc Soils (S) s x OP s x oc s x OP x oc 4 19 19 19 18 Elecho- lytic Chloride conduc- p1-I content tivity 0,0073 44 0.0038 0.0034 10660 0.0016 0.0013 64 080* 0*0004 0-0166t 5227* 0.0018 14.0054 850 009 1.2296 0.0009 1494 0.0010 0.0013 2017 0.0023 0.00 17 1812 0.0006 Error 76 0.0028 1351 0.0013 u2 Total . . 159 Significance of the observed mzan squares: *, P < 0.05; f , P <: 0.01. Definitions of variance components : U2 = technical + sampling variance.2 ,5 x OP x oc 2 5 x OP 2 B x oc 2 2 OP x oc 2 oc 2 OP 0 = variance due to interaction of soils x operators x occasions. D = variance due to interaction of soils x operators. 0 == variance due to interaction of soils x occasions. 0 == variance associated with analyses repeated on same day a = variance due to the interaction of operators and occasions. - - variance associated with analyses carried out on different days. = variance associated with analyses carried out by different operators. U D Estimates of the variance components can be calculated by equating the observed and expected mean squares, and the precision of estimation for varying numbers of samples, operators and occasions can be derived. For example, for calculating the variance associated with analyses repeated on the same day (aR2) for the variable pH: Theref ore, 0' + 20 us2 = 0.0166 u2 = 0.0028 20 oR2 = 0.0138 ox2 = 0.00069 Similarly, estimates of all other variance components can be calculated.The usual practice is to assume that a component is not different from zero unless a significant F-test (= effect mean square/error mean square) is obtained. Hence, for the results in Table 11, the only variance components taken as greater than zero are those for ax2 for pH and chloride content and o:p oc for chloride content. The estimates of the variance components and standard deviations and coefficients of variation of single determinations are given in Table 111. The precision obtained for the determination of each variable was satisfactory.202 MCLEOD, STACE, TUCKER AND BAKKER TABLE I11 SUMMARY OF ESTIMATES OF ERROR Variance component ua (technical + sampling) .. .. .. .. .. 02(0P x OC) .. . . .. .. .. .. u; . . . . .. .. .. .. .. . . Total variance for a single sample . . .. .. Standard deviation (technical + sampling) . . . . Standard deviation (total) . . .. . . . . .. Coefficient of variation (total), pcr cent. .. . . . . Mean . . . . . . .. . . .. .. . . Coefficient of variation (technical + sampling), per cent. PH 0.0028 0.0007 0.0035 8.42 0.053 0.059 0-63 0.70 - Chloride content 1351 1421 194 2966 748 36.7 54.4 4.91 7-27 Electrolytic conductivity 0.0013 - - 0.0013 1-034 0.036 0.036 3.48 3.48 Variances associated with determinations made with different numbers of samples, operators and days can be derived. For example, if Y samples are analysed by s operators on each of t days, then: u2 + OR2 Variance (mean) = for pH r x s x t 0 2 + cTR2 r x s x t s x t for chloride content , C O P x oc Variance (mean) = u2 r x s x t Variance (mean) = _c__ for electrolytic conductivity CONCLUSIONS This method of simultaneous measurement has greatly reduced the time required for analyses, largely by eliminating the need for recording results at the bench and by avoiding the manipulation required for silver nitrate titrations and conductivity bridge measurements. The system has proved reliable over a period of 4 years and gives adequate precision. The instruments have required only slight day-to-day adjustments in order to maintain their standardisation. Throughout the development of the apparatus, the possibility of its incor- poration into an automatic system has been kept in mind, but this would not be justified unless a much greater output of results is required. We thank Mr. K. M. Cellier of the Division of Mathematical Statistics, who designed the precision tests and analysed the data for this study. We also thank Mr. L. Smith and Mr. B. Zarcinas, who were the operators in this study, and Mr. R. Sands, who made the electrode holders. REFERENCES 1. 2. 3. Piper, C. S., “Soil and Plant Analysis,” University of Adelaide, South Australia, 1942. Tucker, B. M., and Raupach, M., J . Aust. Inst. Agric. Sci., 1958, 24, 51. Anderson, R. L., and Bancroft, T. A., “Statistical Theory in Research,” McGraw-Hill Book Co. Received November Sth, 1973 Accepted November 19th, 1973 Inc., New York, 1952.

ISSN:0003-2654    

DOI:10.1039/AN9749900193

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

Analyst, April, 1974, Vol. 99,pp. 203-210 203 A Potentiometric Method for the Determination of Chloride in Boiler Waters in the Range 0-1 to 10 pg ml" of Chloride BY K. TORRANCE (Central EZectricity Research Laboratories, Kelvin Avenue, Leatherhead, Surrey) A potentiometric method for determining chloride in boiler water has been developed, which is based on the potential of a silver - silver chloride wire electrode versus a mercury(1) sulphate reference electrode immersed in a buffered solution of the sample. The method was tested in the range 0.1 to 10.0 pg ml-1 of chloride and the standard deviations of the results a t 0.1, 1.76 and 10.0 pg ml-l of chloride were approximately 0.04, 0.06 and 0.2 pg ml-1 of chloride, respectively. Substances normally present in boiler water do not interfere appreciably but the method is not suitable if octadecylamine is present.THE presence of chloride ion in the steam - water circuit of power stations can give rise to corrosive acidic conditions, and therefore its concentration in boiler waters has to be kept below a specified limit. The controlled chloride levels in many of the boilers used by the Central Electricity Generating Board lie in the concentration range 0.1 to 5.0 ,ug ml-l, a range in which, because of the solubility of the silver chloride, the normal Nernstian plot of the e.m.f. versus the logarithm of the chloride ion activity in the sample is curved. For most analytical purposes it is an advantage to have a linear calibration graph, and calibration functions that would provide this facility were investigated.The most convenient function of concentration to be plotted against e.m.f. for this concentration range was found to be log (c + l), where c pg ml-l is the concentration of chloride in the sample. THEORETICAL BASIS OF THE FUNCTION LOG (c + 1)- At chloride concentrations in the region of the solubility of silver chloride, d K s (where Ks is the solubility product of silver chloride), the contribution to the total chloride present from dissolution of the electrode surface becomes significant. Let the concentration of chloride present before dissolution be m and that dissolved from the electrode surface be s. The total chloride concentration in solution, T,,, is then (m + s). K , can be written .. * - (1) Ks = (m + s) (s) ys2 .. . .where yh is the mean univalent activity coefficient of the silver and chloride ions. By solving equation (1) for s, the real root, an expression for the total chloride present is The relationship between the e.m.f. of a silver - silver chloride electrode and the activity of chloride ions in solution, a,, is Equations of this form have been derived by other workers.lS2 Equation (3) can be rearranged such that @ SAC and the author.204 TORRANCE : POTENTIOMETRIC METHOD FOR THE [AnzaLyst, Vol. 99 On binomial expansion of the term raised to the power of one half, and considering the first two members only, the expression within the square brackets becomes At values of wb of less than 4 K s , this expression tends to which becomes + ll y+ pg ml-l of chloride [ 0-89 when expressed in concentration terms of micrograms per millilitre.At the ionic strength of the preferred buffer concentration, the mean univalent activity coefficient was calculated to be 0.8, thus the simple expression In [pgml-l of chloride + 13 was considered to be a practical calibration function. In order to confirm the application of this approximate but convenient expression the following procedure was adopted. Six standard chloride solutions were prepared, the log (c + 1) values of which were evenly spaced between 0.1 and 18.67 pg ml-l of chloride. The e.m.f. values of two portions of each solution were measured in a random order, as described under Method, and the resulting graph was tested statistically for linearity. As no significant difference was observed a t the 95 per cent.probability level the calibration graph was assumed to be essentially linear over the concentration range tested. EXPERIMENTAL EFFECT OF SODIUM HYDROXIDE- Alkaline conditions are maintained in boiler water in order to minimise corrosion and it was noted that the response of a silver - silver chloride electrode in solutions that had a constant chloride concentration varied with the pH. It was thought that this effect, prob- ably caused by the association of silver with liydroxyl ions, could be eliminated by adding an acidic buffer to the sample solutions before measuring their e.m.f. values. The choice of buffer solutions is discussed below. CHOICE OF BUFFER SOLUTION- An ammonium acetate - acetic acid buffer system was investigated as it has been used successfully in argentimetric titrations of chloride in boiler water.Two buffer concentrations were investigated : one solution contained 60 g 1-I and 78 g 1-1 of acid and salt, respectively, and the other was prepared with one tenth oi these concentrations. The experimental pro- cedure was that described under Method, in which 4 ml of buffer were added to 100 ml of sample, giving a final pH of approximately 4.7. The effect of adding 40 pg of sodium hydroxide per millilitre to 10 pgml-1 standard chloride solutions was studied by using both buffer ionic strengths; this concentration of alkali is in excess of that normally present in boiler waters, viz., 2 to 5 pg ml-l of sodium hydroxide. The e.m.f. values, for both buffer ionic strengths, of the 10pgml-l chloride solutions with and without the sodium hydroxide were measured alternately.A comparison of the two means indicated no statistically significant difference (for the 95 per cent. probability level and 10 degrees of freedom) between the standards when using the more concentrated buffer but a significant difference when the less concentrated buffer was used. The results are summarised in Table I. EFFECTS O F OTHER SU3STANCES- The effects of other substances were tested at two concentrations of chloride. The electrode potentials were measured alternately in standard chloride solutions and in the sarne concentration of standard with other substances added at concentrations greater than those normally found in boiler water. The experimental procedure was as describedApril, 19741 DETERMINATION OF CHLORIDE I N BOILER WATERS 205 TABLE I CHOICE OF CONCENTRATION OF ACETATE BUFFER hiean e.m.f.of 10 pg ml-1 Standard Calculated Mean e.m.f. of chloride solution +- deviation of value of t 10 pg ml-l standard 40 p g nil-' of sodium the difference (for 10 degrees chloride solution/mV hydroxide/mV of the two means/mV of freedom) -- Buffer A* Buffer Bt 191.4 192.8 - 0.34 4-12: 197.7 - 198.0 0.39 0.77 (N.S.) * Buffer A : 6.00 g 1-1 of acetic acid and 7-78 g 1-1 of ammonium acetate. t Buffer B: 60.0 g 1-1 of acetic acid and 77.8 g 1-1 of ammonium acetate. $ Significant at 1 per cent. but not significant at 0.1 per cent. probability level. N.S. denotes not significant. under Method and the results are shown in Table 11.None of the substances tested caused any effect that would interfere significantly with the analytical technique when present in the concentrations normally found in boiler water. TABLE I1 EFFECT OF OTHER SUBSTANCES Apparent" chloride concentration, p g ml-1, a t concentrations of Concentration of Substance substancelpg ml-l 1 p g ml-1 1 I .\mmonia . . . . .. 1.04 10.00 1.04 10.00 1-02 9-70 Cyclohexylamine . . . . Morpholine . . . . Hydrazine . . . . Na3P0, . . . . . . 10 SiO, . . . . . . 10 NaOII . . . . .. 20 1.03 10.00 Fe3+ . . . . .. 10 0.95 9.80 } 0.96 9.70 Na,SO, . . . . .. 1 0 Ca2+ . . . . . . 10 Mg2+ . . . . . . 10 1 0.1 l i Braces denote substances that were tested together. *If other substances had no effect the result would be expected to fall within the following ranges: 1-00 0.04 and 10.00 JI: 0.30 pg ml-l, for 95 per cent.confidence limits. The effect of iron(II1) ion on the electrode was significant at the 1 pg ml-l level but its concentration was considered to be greatly in excess of that expected under the alkaline reducing conditions of boiler water. EFFECT OF TEMPERATURE ON ELECTRODE ASSEMBLAGE- The effects of temperature on the electrode assemblage are complex. The potential of both the reference and the silver chloride electrode will vary according to the thermodynamic factor, -- . In addition, both the niercury(1) sulphate and the chloride electrode will have a temperature dependence related to the solubility of their respective sparingly soluble salts. No detailed study of temperature effects was undertaken as the problem is best solved by ensuring that all solutions are equilibrated in a water-bath thermostatically controlled to within h0.l "C.Experiments indicated that a decrease in temperature of 1 "C resulted in an increase in potential of approximately 1 mV at 25 "C, which would correspond to changes in concentration of 0.40 and 0-04 pg ml-l at the 10 and 1 pg ml-l of chloride levels, respectively. RT F EFFECT OF LIGHT- No effect of light was noted during the course of an analysis, but it is important to stress that the apparatus was screened from strong natural light.[Analyst, Vol. 99 206 TORRANCE POTENTIOMETRIC METHOD FOR THE METHOD Except where otherwise stated, analytical-reagent grade chemicals should be used. Water-Use water that has a low chloride content.Water of suitable quality (conductivity less than 10 pS m-l) can be obtained by passing distilled water through a mixed-bed de- ionisation unit. Acetic acid - ammonium acetate bufer solution-Dissolve 77.8 g of ammonium acetate in approximately 250 ml of water and add 57 ml of glacial acetic acid (sp. gr. 1.05). Dilute the mixture to 1 litre with water and store in a polythene bottle. This solution has been found to be stable for at least 9 weeks. Standard chloride solution A-Dry sodium chloride in an oven at 250 "C for 2 to 3 hours. Weigh 1.649 g of the dry salt, and wash it with water into a 1-litre calibrated flask, make the solution up to the mark with water and store it in a borosilicate glass bottle. This solution has been found to be stable for at least 1 year.RE AGENTS- 1 ml of solution A s 1000 pg of chloride. Standard chloride solution B-With a pipette, transfer 1.0 ml of solution A into a 1-litre calibrated flask and dilute to the mark with water. Transfer the contents of this flask into a polythene bottle and add 40ml of acetate buffer solution. This buffered standard has been found to be stable for at least 1 month. 1 ml of solution B (excluding buffer solution) = 1 pg of chloride. Standard chloride solution C-With a pipette, transfer 10.0 ml of solution A into a 1-litre calibrated flask and dilute to the mark with water. Add 40 ml of buffer solution and store as for solution €3. This buffered solution has been found to be stable for at least 1 month. 1 ml of solution C (excluding buffer solution) = 10 pg of chloride.Hydrochloric acid (1 + 99)-This solution should be freshly prepared each time a batch of electrodes is to be chloridised. APPARATUS- Measuring cylinders-Borosilicate glass measuring cylinders (100 ml) , fitted with glass stoppers, should be used to prepare samples for analysis. Water-bath-A water-bath of sufficient depth to cover approximately three quarters of the height of a 100-ml measuring cylinder should be used. It should be fitted with a circu- lating pump and a heating coil such that the temperature can be thermostatically controlled to within 50-1 "C. Magnetic stirrer-A variable-speed magnetic stirring unit should be used to stir the solution being analysed. A PTFE or polythene-covered magnet, 20 mm long, is suitable for use in a 50-ml borosilicate glass beaker and a constant stirring rate that is just sufficient to rotate the magnet should be chosen.Siher - silver chloride electrodes-The electrodes can be prepared from 19 s.w.g. (1.016 mm diameter) silver wire and a 120 to 140-mm length of this wire should be sealed into a soda- glass tube that is 5 to 7 mm in diameter and 100 mm long. As the seal may not be perfect, it is advisable to apply a thin coating of Araldite epoxy resin at the glass - silver surface. The length of wire protruding from the seal should be 10 to 15 mm. Prior to electrolysis, the silver wire should be cleaned by washing it first with several small amounts of benzene and then cautiously with dilute nitric acid. Finally, the electrodes should be washed thoroughly with water.The silver electrode should be chloridised by placing it in a solution of dilute hydrochloric acid (1 + 99) and electrolysing against a platinum cathode at a current of 0.4 mA for a period of 4 hours. After electrolysis, the silver chloride electrode should be washed thoroughly with distilled water and stored in water in the dark when not in use. Electrodes prepared in this manner were used daily over a period of 3 months. Mercwy(1) sulj5hate refeyence electrode-An Electronic Instruments Ltd,, Model RH 23/3, electrode with a ground-glass sleeve junction was used. The internal electrolyte used should be 1 M sodium sulphate solution. Millivoltmeter-A millivoltmeter capable of discriminating to k 0 . l mV should be used. A Beckman Research pH meter is suitable.PROCEDURE- Sample collection-Collect samples that have been cooled to below 35 "C in stopperedApril, 19741 DETERMINATION OF CHLORIDE I N BOILER WATERS 207 glass or polythene bottles, taking care to avoid contamination of the sample. Samples of boiler water have been found to be stable for at least 2 weeks. Analysis of sam$les-Pour 100 ml of sample into a measuring cylinder and add 4-0 ml of buffer solution. Replace the stopper, shake and place the cylinder in the water-bath, allowing approximately 10 minutes for equilibration of a sample whose initial temperature was 5 "C above or below that of the water-bath. Standard solutions B and C (with buffer added) should previously be temperature equilibrated in the water-bath. Clamp the reference and silver chloride electrodes together; if they are mounted on a retort stand, it is convenient to mark the height a t which the electrodes are clear of the stirrer.This level should be kept constant throughout the analysis. Rinse a 50-ml beaker three times with small portions of standard solution C, before adding sufficient solution for the potentiometric measurement, then immerse the electrodes in the solution and note the e.m.f. reading after 2 minutes. Take the electrodes out of the solution and remove the surplus solution from the reference electrode with a soft tissue. The silver chloride wire electrode will have very little solution on its surface and it is advisable to avoid damaging its surface by any mechanical contact. This measuring procedure is repeated for the following sequence of solutions : standard B, the sample solutions, standard C, standard B.Calculation of resuZts-Calculate the mean e.m.f. values of the two standard solutions and plot these values against their respective log (c + 1) values, i.e., 2 and 11. It is convenient to use semi-logarithmic paper for this purpose. The values of log (c + 1) corresponding to the measured e.m.f. of the samples are read from the calibration graph. When the chloride concentration in the samples is known to be less than 1 pg ml-l, greater precision will be obtained by using 0.1 and 1 pg ml-l chloride standards to define the calibration graph. RESULTS PRECISION- Six solutions containing known amounts of chloride in the range 0.1 to 10 pg ml-1 were prepared. Two independent 1 and 10 pg ml-l solutions were prepared as calibration standards, and to these were added ammonium acetate buffer solution in the recommended proportions.The e.m.f. values of these two solutions provided the within-batch calibration points through- out the precision tests. The e.m.f. values of two portions of each of the six solutions were measured. In addition, the e.m.f. value of each of the two standard solutions was measured before and after each batch. Readings were taken after 2 and 4 minutes and the within-batch, between-batch and total standard deviations were calculated for each electrode for both the 2 and 4-minute readings. The results are summarised in Table 111. TABLE I11 PRECISION OF DETERMINATION OF CHLORIDE Concentration Concentration Within-batch Between-batch of chloride of chloride standard standard added lpg ml-l found*/pg ml-l deviationlpg ml-l deviationlpg ml-1 Results after 2 minutes- 0.10 0.08 0.74 0.72 1.76 1.77 3.38 3.49 5.94 6-12 10.00 9.80 0.01 0.03 0.04 0.09 0.04 0.22 0-04 N.S.N.S. N.S. 0.14 N.S. Results after 4 minutes- 0.10 0.09 0.01 0.05 0.74 0.73 0.02 N.S. 1.76 1-78 0.04 N.S. 3.38 3.51 0.10 0 5-94 6-15 0.04 N.S. 10.00 9.82 0.23 0 * Mean of ten results. Total standard deviationfpg rnl-1 0.04 0.04 0.06 0.09 0.15 0.23 0.05 0.03 0.06 0.10 0-06 0.23 Although the internal calibration procedure described under Method was chosen because changes in the electrode potential were feared, the effect of using a fixed calibration graph on the precisions given in Table I11 was investigated.A mean calibration graph was con-208 TORRANCE : POTENTIOMETRIC METHOD FOR THE [Analyst, Vol. 99 structed from the 10 and 1 pg ml-l of chloride calibration points used in five batches. The values of the other six solutions were calculated from this calibration graph and the results, from the %minute readings, are presented in Table 11’. A comparison of the results presented in Tables 111 and IV shows that both the between- batch and the total standard deviations are increased by using a mean calibration graph. TABLE IV PRECISION OF DETERMINATION OF CHLORIDE FROM A MEAN CALIBRATION GRAPH Concentration of chloride addedlpg ml-l 0.10 0.74 1-76 3.38 5.94 10.00 Concentration of chloride found*/pg ml-l 0.09 0.74 1.78 3.49 6-10 9.82 Within-batch standard deviationlpg ml-1 0.01 0.03 0.03 0.09 0.04 0.11 * Mean of ten results.Between-batch standard deviationlpg ml- 0.06 0.13 0.15 0.19 0.25 0.93 Total standard deviation/pg ml-1 0.06 0.14 0.15 0.30 0.35 0.93 Three boiler waters from two power stations were also analysed by the described method. In each instance a batch consisted of five replicate analyses of the boiler water, together with five replicates of the same water plus a standard addition of chloride, e.m.f. readings being taken after 2 minutes. The results (shown in Table V) indicate satisfactory recovery of the added chloride and within-batch standard deviations similar to those obtained with standard solutions. TABLE V ANALYSIS OF BOILER WATER SAMPLES (AT CERL) Recovery oi Within- batch Chloride added chloride, standard Sample found/pg ml-1 per cent.deviationlpg ml-1 0-08 .. .. 3.08 - + 2.37 pg ml-l of chloride . . 5.46 100.4 0-06 Boiler water A .. Boiler water A’ .. . . .. 3.05 + 2.60 pg ml-1 of chloride . . 5.60 Boiler water B .. .. .. 3.00 + 2.50 pg ml-l of chloride . . 5.51 - 0.06 101.2 0.04 - 100.4 0.09 0.03 TABLE VI PRECISION OBTAINED WITH SYNTHETIC SOLUTIONS AT POWER STATIONS Chloride concentrationlpg rnl-1 A r I Total standard Station Added Found* deviation A B C D 0.1 0-5 2.0 6.0 0.1 0.5 2.0 6-0 - 0-51 2.08 6.14 0.10 0.49 2.04 - - 0.1 0.45 0.47 2.0 2.22 6-0 - - 0.09 0.1 1 0.16 0.03 0.04 0-07 - 0.06 0.09 0-1 0.09 0.03 0.5 0.48 0.05 2.0 2.05 0.13 6.0 * Mean of ten results. - -April, 19741 DETERMINATION OF CHLORIDE I N BOILER WATERS 209 The analytical method was used at a number of power stations for the determination of chloride in synthetic and boiler water samples.The precisions reported for synthetic solutions (Table VI) are similar to those reported in this laboratory. Table VII contains the results of the analyses of boiler waters and spiked boiler water samples and it can be seen that the recovery is satisfactory in all instances. TABLE VII ANALYSIS OF BOILER \f7ATER SAMPLES AT POWER STATIONS Recovery of Within-batch Chloride chloride added, standard Sample found1p.g ml-I per cent. deviationlpg ml-I Station A . . . . .. . . 6-83 - Nil + 6.1 pg ml-1 of chloride . . 12.62 95 0.32 . . .. 0.07 - Station B . . . . + 0.05 pg ml-1 of chloride . . 0.12 100 Station C . . .. .. .. 0.47 + 0.5 pg ml-l of chloride . . 0.97 Station D .. . . .. .. 0.22 + 0.25 pg ml-I of chloride . . 0-45 - 100 - 92 0.03 0.02 0.04 0.06 0.02 0.02 DISCUSSION The precisions reported in Table I11 were considered adequate for daily analysis of boiler water. In view of the similarities of the precisions found after 2 and 4 minutes, it was decided that the recommended procedure would be to measure the e.m.f. after 2 minutes. The time required for analysis depends principally on the batch size and the precision required by the analyst. One obvious method of improving the time taken for analysis would be to use a predetermined calibration graph. Estimates of the precisions that could be obtained by this procedure are reported in Table IV. A mean calibration graph was obtained from the ten standard solutions in five series of analyses. As only one reference electrode was used, the major difference in precision should be due to the small variations in the condition of the silver chloride electrode over a period of 2 weeks.Although the results obtained with a fixed calibration graph are much less precise than those given in Table 111, operational conditions in power stations can be foreseen when a rapid result of moderate accuracy is required in the shortest possible time. If the highest precision is sought, then measurements will involve a period of temperature equilibration, the duration of which will depend on the difference between the sample temperature and the temperature at which the procedure is carried out. 1 10 Chloride concentration (c + 1 )/pg ml-’ Fig. 1.Display of bias from calibra- tion function210 TORRANCE In the precision tests, a positive bias was observed in the region 1 to 10 pg ml-1, whit can be shown to be due to the choice of the function log (c + 1) and the method of constru tion of the calibration graph. An estimate of this bias was obtained by plotting the theoretic e.m.f. values calculated from equation (3) versus the logarithm of the chloride concentratic expressed as (c + 1). These results, together with the calibration graph that would E obtained by joining the 10 and 1 pg ml-1 calibration points, are displayed in Fig. 1. TI agreement between the theoretical and observed values cannot be checked unless accura values of K , and y k are known. a theoretical bias of +0.20 pg ml-l is calculated €or solutions containing 5-94 pg ml-1 r chloride, which agrees with the observed bias of +0.20 pg ml-l. The method is free of interference from most substances normally found in boiler water In the presence of octadecylamine, the signal from the electrodes drifted in the direction r increasing chloride concentration, a similar effect being noted with solutions containing sma amounts of commercial preparations used as corrosion inhibitors. The proposed method is considered to be more convenient and precise than the spectrr photometric methods of analysis for chloride i0ns~9~ in the range 0.1 to 10 pg ml-l in boil! waters. This work was carried out at the Central Electricity Research Laboratories and is pul lished by permission of the Central Electricity Generating Board. By using values of 1.6 x 10-lo for ITS3 and 0-8 for REFERENCES 1. 2. 3. 4. 5. Bardin, V. V., Ind. Lab., 1962, 28, 967. Florence, 1. M., J . Electroanalyt. Chem., 1971, 31, 77. Linke, W. F., “Solubilities: Inorganic and ;Iletal-Organic Compounds,” Volume 1, D. Van Nostrar Tomonari, A., Japan Chem. J., 1962, 83, 693. Florence, T. &I., and Farra Y . J., Analytica Chirn. Acta, 2971, 54, 373. Co. Inc., Princeton, N. J., New York, Toronto and London, 1958. Received September 17th, 19’ Accepted October lSth, 19’

ISSN:0003-2654    

DOI:10.1039/AN9749900203

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

Analyst, April, 1974, Vol. 99, pp. 211-213 21 1 Filter-papers as a Source of Error in Ammonia Determinations BY LIAM 0 HALMHAIN AND DONAL 0 DANACHAIR" (Analytical Chemistry Section, Institute for Industrial Research and Standards, Ballymun Road, Dublin 9) A selection of filter-papers was tested for the presence of ammonia. In both Hardened filter-papers The degree of contamination of filtrates was also determined. instances, significant levels of ammonia were found. contained significantly higher levels of ammonia. THE determination of ammonia, or the determination of protein and nitrogen contents via ammonia, in samples is an important test in many analytical laboratories. Prior filtration of the sample through a filter-paper is often carried out and indeed recommended in some instances.lP2 Actual digestion of the filter-paper with the sample is recommended in the determination of nitrogen in some types of sample.3 Our colleagues in our Water and Effluents Laboratory noted anomalously high ammonia contents in murky samples that had been filtered prior to carrying out the ammonia deter- minations, and it was suggested that ammonia leached from the filter-paper might be the source of the anomaly.A search of the literature revealed two references4r5 to such a source of error. As both references were rather old (18 and 27 years, respectively) and as no reference was apparently made to this problem in the more recent standard analytical texts and official methods of analysis, it was decided to investigate modern filter-papers. It is of interest to note that the authors of both of these previous papers claimed no originality for detecting the presence of ammonia in filter-papers, assuming it to be generally known, but remarked that they found no references to it elsewhere. Ammonia has certainly been used in the past in the production of hardened ashless grades of filter-paper.6 EXPERIMENTAL A range of Whatman filter-papers (from sealed cellophane-wrapped boxes) , a sample of old stock of Green's filter-paper and some Schleicher and Schiill Selecta filter-papers were tested as follows.TEST 1- Quantitative tests on all samples-About 20 g of filter-paper were weighed into a round- bottomed flask, which was fitted to a conventional macro-Kjeldahl distillation apparatus (the ammonia evolved was collected by dipping the outlet into 50 ml of 2 per cent.boric acid solution), 400 ml of 25 per cent. sodium hydroxide solution were then added and the flask was heated until 50 ml of distillate were collected. The distillate solution was made up to 250 ml with ammonia-free water and a suitable aliquot (1 to 5 ml) was further diluted to 50 ml; 2 ml of Nessler's reagent were added to the latter solution and the resulting colour was read after 15 minutes against the relevant disc in a Lovibond Nessleriser. By using phosphate buffer798 of pH 7-4 instead of the above sodium hydroxide solution, approximately one tenth of the ammonia evolved by the former method was collected. TEST 2- nitrate or nitrite, if present) was carried out on one sample (Whatman No. 541). The determination of total nitrogen (by the Kjeldahl method, which is exclusive of * Present address : Department of Biochemistry, University College, Dublin.@ SAC and the authors.212 TEST 3- Filter-papers were folded and placed in glass funnels in the usual mode for filtration and four successive 25-ml portions of ammonia-free water were passed through each paper. The ammonia content of each of these 25-ml portions was measured by the Nessleriser method. TEST 4- A qualitative spot test with Nessler’s reagent (BDH Chemicals Ltd.) applied directly on the paper gave positive results in all cases, with rather higher levels indicated for hardened ashless papers. RESULTS AND DISC~SSION 6 hALMHAIN AND 6 DANACHAIR: FILTER-PAPERS AS A [APdySt, Vol. 99 All results are expressed as parts per million ?iz/wz of ammonia in the filter-paper except when otherwise indicated. Test 1 showed that the ammonia content of analytical filter-papers ranged from SO t o 275 p.p.m.mlm. T,ower values were obtained for qualitative filter-papers such as Whatman No. 1. The lowest result (20 p.p.m.) was obtained on the sample from old stock of Green’s 401. The values increased with the degree of ashlessness and hardness of the filter-paper. Values of 210 p.p.m. were obtained for Whatman No. 542 and 275 p.p.m. for Schleicher and Schull No. 589/3 (Blue band). While values were constant within a given box of filter-papers, considerable variation existed between different boxes of the same grade, manufacturer and even batch number, e.g., two boxes of the same batch number of Whatman No.542 filter- papers gave 210 and 110 p.p.m. of ammonia. Test 2 showed that a Whatman No. 541 filter-paper, which had an ammonia content by Test 1 of 65 p.p.m., had a total nitrogen content (as NH,) of 160 p.p.m. The results of Test 3 on four different single papers (A to D) from the same box of Whatman No. 541 filter-papers are shown in Table I. Filter-papers from this box gave 50 p.p.m. of ammonia when subjected to Test 1. TABLE I RESULTS FOR AMMOMA DETERMINATION BY TEST 3 Amount of ammonia in the filtratelpg 7 A \ Portion of water (25 inl) passed through paper Paper A Paper I3 Paper C Paper D First . . .. . . . . .. . . 14 20 12 12 Second . . .. .. .. I . f . 5 8 8 7 Third . . .. .. .. .. .. 4 4 3 6 Fourth . . .. . . .. . ... 2 4 3 4 It is recommended2 that when filtering murky water samples the first 25 ml of filtrate be discarded. The above results show that more than 100 ml need to be discarded if filter- paper is used. When it is considered that the mass of a typical 12-5-cm filter-paper is slightly more than 1 g, the amount of ammonia, and indeed nitrogen, added in an analysis when the entire filter-paper is included with the sample can be of the order of 300 pg. Table I gives an indi- cation of the amounts added when filtering solutions through filter-paper. It is to be noted that Schleicher and Schull in their publication “Selecta Filter Papers- Description of Types” do state that hard filters are “unsuitable for all work where the filter must be burnt together with the sediments, in accordance with Kjeldahl” because of the nitrogen content that results from the manufacturing process.No warning is given of the unsuitability of any filter-papers for ammonia determinations or of other than hardened papers for nitrogen determinations. CONCLUSION While the ammonia content of filter-paper might be known to some analysts, a survey of the literature, and of standard and reference methods, indicates a lack of awareness of the necessity of avoiding the use of filter-papers when determining low levels of ammonia and nitrogen. The present work re-emphasises the necessity of carrying out blank determinations with all reagents including filter-paper and of taking the blank determination through all the stages of the analysis.April, 19741 SOURCE OF ERROR I N AhiMONIA DETERMINATIOWS 213 REFERENCES 1. “Colorimetric Cheniical Analytical Methods,” Seventh Edition, Tintometer Ltd., Salisbury, 1967, 2. “Standard Methods for Examination of Water and Wastewater,” Twelfth Edition, A.P.H.A., 3. Pearson, D., “The Cheniical Analysis of Foods,” Sixth Edition, Churchill, London, 1970, p. 9. 4. Leitch, J. I,., and Wells, L. A., J . Fya+zkli.P?, Inst., 1946, 241, 73. 5. Lucas, V., Reufa Bras. Farm., 1955, 36, 233. 6. IVhiteley, bl. A,, Editor, “Thorpe’s Dictionary of Applied Chemistry,” Fourth Edition, Longmans, 7. “Methods of Testing Water used in Industry,” B.S. 2690, British Standards Institution, 1965, p. 11. 8. “Standard Methods for Examination of Water and Wastewater,” Twelfth Edition, A.P.H.A., p. 86. A.W.W.A. and W.P.C.F., New York, 1965, p. 390. Green and Co., London, New York and Toronto, 1937, p. 182. A.W.W.A. and W.P.C.F., New York, 1965, p. 392. lieceivcd October 8tA, 1973 Accepted J a w a r y 16th, 1974

ISSN:0003-2654    

DOI:10.1039/AN9749900211

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

214 Analyst, April, 1974, Vol. 99, pp. 2144217 Determination of Impurities in Oxygen by Mass Spectrometry BY J. P. COMPTON AND S. D. WARD (Department of Trade and Industry, National Physical Laboratory, Teddington, Middlesex, T W11 OL W ) The difficulties in analysing high-purity gaseous oxygen are overcome by using a copper-based catalyst to remove the oxygen completely, allowing impurities to be readily determined by mass spectrometry. The sensitivity is estimated to be 1 p.p.m. and the uncertainty is at most 5 per cent. of the concentration. The equipment is described, together with the results of experiments confirming the accuracy, linearity and reproducibility for the particular impurities nitrogen, argon and krypton. THIS paper describes a method of determining a number of impurities in oxygen with an uncertainty of up to 5 per cent.of the concentration and a sensitivity in the region of 1 p.p.m. atomic. The method is based on the use of a mass spectrometer to determine impurities following the complete removal of oxygen by means of a copper-based catalyst. The need for such measurements arose in the course of work to establish the low-tem- perature fixed points of the 1968 International Practical Temperature Scale. These require high-purity gases (less than 10 p.p.m. of total impurities) if the temperatures are to be realised with sufficient accuracy (better than 0.3 mK). In the case of oxygen, the significant impurities are nitrogen, argon and krypton. The more volatile gases, such as helium, hydrogen and neon, are improbable impurities, although important if present, while the less volatile gases, such as carbon dioxide and water vapour, are likely to be frozen out without influencing the fixed-point temperatures.As is well known,lr2 it is difficult to analyse oxygen mass spectrometrically owing to the reaction between oxygen and carbon on the filament. The life of the filament is drastically reduced as a result of this reaction, and the copious amount of carbon monoxide produced masks the main peak of nitrogen at m/e 28. However, the removal of oxygen by chemical means, in order to permit determination of the residual gases, is well established. Many absorbents have been used, including phosph~rus,~ chromium(II1) ~ulphate,~ hot copper turnings5 and sodium - potassium alloy.6 The present work utilises the copper-based catalyst R3-11, manufactured by BASF, which removes oxygen by the reaction Cu + 40, -+ CuO (H = -37.1 kcal) The advantages of such a catalyst are cheapness, ease of handling in the oxidised form, straightforward regeneration to the reduced form and facility for operation at room tem- perature.However, as the oxygen absorption capacity increases with temperature, the present system used a nominal catalyst temperature of 150 "C. APPARATUS- The apparatus, shown in Fig. 1, consists essentially of two 60-cm3 metering chambers, V and W, from either of which gas can be passed through the catalyst and thence into the mass spectrometer batch inlet system. Simple Bourdon gauges monitor the pressures in V and W. In normal use, V alone is used to define the volume of sample gas, which is admitted to it through the needle valve N,.The second chamber is used, as described later, in experiments to test the linearity and accuracy of the system. All interconnections in the system are of 6 mm diameter tubing. The catalyst is held in a chamber, 60 cm in length and 2 cm in diameter, supported vertically and filled to within 5 cm of the top with 140 g of catalyst, which is in the form of pellets of diameter 5 mni and length 3 mm. A plate, P, pierced with small holes, retains the catalyst in the chamber, while a 7-,urn filter, F, prevents the transport of dust into the @ SAC; Crown Copyright Reserved.COMPTON AND WARD 215 inlet system of the mass spectrometer. A heating tape is wrapped around the outside of the catalyst chamber, and the temperature, indicated by two monitoring thermocouples, is easily maintained within 5 "C of the nominal 150 "C.Taps T, and T, provide connections to a diffusion pumping system with a liquid nitrogen trap and allow evacuation of all or part of the equipment as desired. With the exception of taps TI, T,, T, and T,, and the Bourdon gauges, the entire system is constructed of stainless steel, the parts having been carefully cleaned before assembly. To pump Bourdon gauges , c ~ s p e c t r o m e t e r -.+TO mass Catalyst chamber 'c h e r m oco u p 1 e Plate, P Fig. 1. Schematic representation of the analysis system Before initial use, and infrequently thereafter, the catalyst niust be reduced by passing hydrogen at 200 "C through it, setting up the reaction CuO + H, + Cu + H,O (H = -20.7 kcal) The water is removed from the catalyst chamber by baking and pumping it which, at the same time, serves to remove other adsorbed contaminants.Although the manufacturers advise against exceeding a temperature of 250 "C, an overnight bake at 300 "C caused no discernible impairment of the catalyst. Regeneration of the catalyst was necessary after the analysis of approximately eighty oxygen samples (each of 60 cm3 at atmospheric pressure). EXPERIMENTAL METHOD- The system is normally kept under a high vacuum, and an analysis is carried out in the following way. A sample of oxygen is admitted to the metering chamber V, filling it to a measured pressure (normally 1 atm). Under the control of the needle valve N,, this sample is then bled into the catalyst chamber over a 3-minute period (the bleeding is carried out slowly so as to avoid possible local overheating of the catalyst, which is not made to withstand exposure to almost pure oxygen).The oxygen is now completely absorbed and, after an interval of 10 minutes in order to ensure equilibrium at the low pressure of the residual impurities, these remaining gases are admitted to the spectrometer batch inlet system. Peak heights are measured at m/e 28, 40 and 84 and are then converted, by means of calibration data, into the partial pressures of nitrogen, argon and krypton present in the original oxygen sample. At low impurity levels (less than 100 p.p.m.), a correction is applied for outgassing of nitrogen in the spectrometer inlet system, the correction being determined in a separate experiment.The system is readily calibrated by using pure samples of nitrogen, argon or krypton in order to determine the relationship between spectrometer peak height and the initial pressure in V. Except when the catalyst requires regeneration, the oxygen present does not exceed its detection limit, estimated to be 0-5 p.p.m., while the estimated detection limits for the three impurities under consideration are of a similar magnitude.216 [Analyst, 1701. 99 VERIFICATION OF ACCURACY- In order to check the linearity of the system for each of the three major contaminants, tests were carried out in the following manner. Chamber V was filled with one contaminant at a measured pressure, Po, of approximately 0.1 atm.This gas was then expanded into the empty chamber, W, which, after closing tap T,, was evacuated. The gas remaining in V was again expanded into W, which was once more isolated and evacuated. This expansion process was repeated as many times as desired until, after the final expansion, gas was allowed to remain in W, which thus contained contaminant at the same pressure P as did V. After again closing T,, the gas from W was fed through the catalyst chamber to the spectrometer, thus yielding a measurement of P. This value for pressure could be com- pared with that calculated from the known volume expansion factor, D, because the measured pressure, P, is related to the number of expansions, n, by the relationship PIP, = Dn. D itself was determined by measurement of the pressure change upon performing an expansion from 1 atm.In Fig. 2, P (on a logarithmic scale) is plotted against n for the representative case of nitrogen. Each point represents a separate set of PZ expansions, while the line shows the theoretical relationship based upon the direct determination of D. COMPTON AND WARD : DETERMINATION OF IMPURITIES 0 5 10 15 Number of expansions ( n ) Fig. 2. Pressure as a function of the number of expansions of the sample gas. The crosses are experi- mental points, while the straight line represents the calculated pressure based upon the measured expansion factor In the second stage of the test, the purest available oxygen was added to the con- taminant still retained in V, making up the total pressure to 1 atm.This doped oxygen was then analysed, yielding the over-all partial pressure of the contaminant. Subtraction of the pressure previously found to be present in the identical volume W yielded the partial pressure of contaminant intrinsically present in the oxygen, which could be compared with that measured directly in similar, but undoped, oxygen. The results of these tests are shown in Table I. The agreement of the results to within 1 p.p.m. or 5 per cent., whichever is the greater, together with the satisfactory agreement shown in the expansion experiments, are taken to confirm the linearity of the system and the absence of perturbing absorption effects for the three gases nitrogen, argon and krypton. It is possible that the exothermic absorption of oxygen promotes the desorption of small amounts of gas from the catalyst.This effect is unlikely in view of the thorough baking to which the catalyst has been subjected, and certainly cannot exceed the amounts measured in high-purity oxygen and given in Table I.April, 197-41 I N OXYGEN BY MASS SPECTROMETRY 217 Carbon monoxide is another possible and potentially important contaminant of liigh- purity oxygen. It was hoped that carbon monoxide might be detected after oxidation to carbon dioxide, but it was found to be largely absorbed, yielding only an insignificant amount of carbon dioxide. The present equipment is therefore unsuitable for the determination of carbon monoxide in oxygen. TABLE I RESULTS OF MEASUREMENTS ON PURE AND DOPED OXYGEN Partial pressure of test gas/N m--2 A r > Test With 1 atm of added As measured in gas oxygen in V -Alone in W Difference oxygen alone Argon 2.0 1-9 0.1 0.1 1.0 0.9 0.1 2.9 2-2 0-7 1-4 0.8 0.6 5.6 3-6 2-0 3.3 1.4 1.9 Krypton 4.9 4.4 0.6 0-6 Nitrogen 9.3 7.2 2.1 1.7 CONCLUSIONS As a result of the absorption of oxygen by a cheap, convenient and re-usable catalyst, it is possible to use a mass spectrometer to measure impurity concentrations in oxygen with a sensitivity of 1 p.p.m. and an accuracy of 5 per cent. Self-consistency checks indicate that calibration results for nitrogen, argon and krypton can be reliably extrapolated over several orders of magnitude. The absence of perturbing sorption effects has also been demonstrated for these three gases. REFERENCES 1. Crable, G. E., and lierr, N. IT., A?zalyt. Chem., 1957, 29, 1281. 2. Ross, P. J., Ibid., 1966, 38, 175. 3. Hughes, E. E., and Dorke, W. D., Ibid., 1968, 40, 866. 4. Arthur, P., Ibid., 1964, 36, 701. 5. Talakin, 0. G., Golovin, Yu. I., and Dykhno, N. M., J . Analyt. Chem., U.S.S.R., 1971, 26, Part 11, 6. Roboz, J., Analyt. Claern., 1967, 39, 176. 1094. Received October 25th, 1973 Accepted December 28th, 1973

ISSN:0003-2654    

DOI:10.1039/AN9749900214

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

218 Analyst, April, 1974, 1701. 99, p p . 218-221 Analysis of Steroids Part XXIV.* A Specific Method for the Spectrophotometric Determination of 17- Ethynyl Steroids BY G. SZEPESI AND S. GoRoG (Chemical Works, G. Richter Ltd., Budapest X , Hungary) The investigation of the equilibrium between 17-keto steroids, acetylene and 17-ethynyl steroids has afforded two possibilities for its analytical appli- cation. On the one hand, the 17-ethynyl steroids can be quantitatively converted into 17-keto steroids and the latter determined spectrophoto- metrically as their 16-glyoxalyl derivatives (Amax. = 294 nm, E = 10 700 and u = 1-1 per cent.). On the other hand, a t 0 “C the 17-keto steroid contami- nants in 17-ethynyl steroids can be selectively determined on the same principle. These methods cannot be applied in the investigation of 3-keto steroids.THE introduction of the 17-ethynyl group into various steroid hormones greatly enhances their hormonal activity. The resulting “super gestogens” and “super oestrogens” can be used in very small doses, hence the micro-determination of these compounds is very important. Most of the existing micro-scale methods for the determination of 17-ethynyl steroids involve ultravioletl*2 and infrared3 spectrophotometric, colorimetric4-0 and fluorimetric6 pro- cedures, which usually depend on other functional groups in the steroid molecule such as the a,p-unsaturated 3-keto group of the phenolic A-ring. These and many other methods of similar nature are very sensitive, but are not, however, specific for the ethynyl group, The specific determination of the 17-ethynyl group can be performed by argentimetric - acidimetric titration,’,* but obviously this method cannot be used on the micro-scale.Gas-chromato- graphic method^^-^^ have also proved to be suitable for the micro-determination of 17-ethynyl steroids. In this paper, a selective spectrophotometric method is reported for the determination of 17-ethynyl steroids, which is based on the “de-ethynylation” reaction of the 17-ethynyl steroid leading to the formation of the 17-keto steroid, and determination of the 17-keto group by our previously described diethyl oxalate method.13-16 Examination of the equilibrium between the 17-ethynyl and 17-keto steroids has also suggested the possibility of determining 17-keto steroid contamination in 17-ethynyl steroids.EXPERIMENTAL APPARATUS- A Spektromom 202 spectrophotometer with 1-cm quartz cells was used. REAGENTS AND MATERIALS- Solvent mixture-Mix 900 ml of analytical-reagent grade t-butyl alcohol, distilled over sodium, with 100 ml of analytical-reagent grade cyclohexane. The water content of this mixture should not exceed 0.05 per cent., as determined by the Karl Fischer method. Sodium t-butoxide reagent, 0-25 N-Dissolve 2.88 g of sodium in 400 ml of the above mixture of t-butyl alcohol and cyclohexane by boiling the mixture, cool and dilute the solution to 500 ml. Diethyl oxalate reagent, 1 N-Dilute 75 g (67.3 ml) of redistilled diethyl oxalate t o 500 ml with the above solvent mixture. Ethanol, 96 +er cent. VlV. Hydrochloric acid, 0.5 N.Cyclo hexane. The steroids used in this study were products of G. Richter Ltd., Budapest, and their purity was checked by titration of the ethynyl group7 and thin-layer chromatography. The * For details of Part XXIII of this series, see J . CWromat., 1973, 76, 502. @ SAC and the authors.SZEPESI AND GOROG 219 three ethynyl steroids used in the control experiments for tlie determination of 17-keto contamination (see Table 11) were highly purified materials with 17-keto contents of less than 0.05 per cent. PROCEDURE FOR THE DETERMINATIOK OF 17-ETHYNYL STEROIDS- Dissolve an accurately weighed amount of the sample equivalent to about 0.03 g of 17-ethynyl steroid in the solvent mixture and dilute this solution to 50ml in a calibrated flask. Transfer 2-0 ml of this stock solution into a carefully dried 50-ml flask.Add 5 ml of solvent mixture, 2-5ml of sodium t-butoxide reagent and reflux the mixture for 2 to 3 minutes. Cool the mixture and add 0.5ml of diethyl oxalate reagent. Allow the mixture to stand at room temperature for 15 minutes, then add 3 ml of 0.5 N hydrochloric acid, transfer the mixture into a 50-ml calibrated flask and dilute to volume with ethanol. Prepare the reference solution in a similar manner, except that the 2-0-ml aliquot of the stock solution is added after the addition of the hydrochloric acid. Determine tlie absorbance against the reference solution at 294nm and calculate the 17-ethynyl steroid content on the basis of the absorbance of a similarly treated standard solution or from the molar absorptivities given in Table I.TABLE I SPECTRAL DATA FOR THE DETERMINATION OF 17-ETHYNYL STEROIDS AFTER DE-ETHYNYLATION AND FORMATION OF 16-GLYOXALYL DERIVATIVES Molar Relative Purity,’ absorptivity standard deviation,* per cent. (294 nm) per cent. Mestranol . . . . . . . . . . 99.9 10 700 fl.1 17-a-Ethynyl-androst-5-en-3/3,17-diol . . 99-8 10 800 & 1.0 17-a-Ethynyloestradiol . . .. . . 99.9 10 700 + 1.0 Ethynodiol djacetate . . . . . . 99.5 10 600 - Ethynodiol . . . . . . . . . * 98.9 10 700 - * Six parallel runs. PROCEDURE FOR DETERMINING THE 17-KETO STEROID CONTENT IN 17-ETHYNYL STEROID- Accurately weigh about 0.03 to 0.05 g of the sample to be tested into a 50-ml calibrated flask. Dissolve the material in a mixture of 2.5 ml of cyclohexane, 2-6 ml of solvent mixture and 0.5m.l of diethyl oxalate reagent.Before the addition of sodium t-butoxide reagent, place the flask containing the mixture in an ice-bath and allow it to stand for 15 minutes. Then add 2.5 ml of sodium t-butoxide reagent and allow the mixture to stand for an additional period of 30 minutes in the ice-bath. Add 3 ml of 0.5 N hydrochloric acid and dilute to volume with ethanol. Prepare the reference solution in a similar manner but in this case weigh the substance to be tested into the flask after the addition of the hydrochloric acid. The difference between the two weighings must not exceed 1 per cent. Calculate the 17-keto content on the basis of the extinction recorded at 294 nm by using the molar absorptivities given in Table I.In instances of very low 17-keto contents the use of 2-cm cells is advisable. RESULTS AND DISCUSSION The results for the determination of various 17-ethynyl steroids by the recommended method are summarised in Table I, and those for the determination of 17-keto steroid contaminants in the 17-ethynyl steroids are given in Table 11. The 17-ethynyl steroids are produced commercially from 17-keto steroids by treating the latter with sodium acetylide according to the following reaction. acetylide CH,O220 SZEPESI AND GGROG [Analyst, Vol. 99 TABLE I1 DETERMINATION OF 17-KETO STEROID CONTENT IN 17-ETHYNYL STEROIDS Amount of contamination 7 - 7 Taken, Found, Standard Main component Contaminant per cent. per cent. deviation* Mes tranol .. .. . . . . Oestrone-3- 0.060 methyl ether 0.2 12 0.320 0.420 17-a-Ethynyloestradiol. .. . . . Oestroiie 0.180 0.480 0.620 1 7-a- Ethynyl-andros t-5-en-3 /3,17-diol D ehydroepi- 0.200 androsterone * Six parallel runs. 0.064 0.005 0.210 0.320 4 0.05 0.424 0-182 0.482 5 0 . 0 3 0.622 0.205 -+O*Ol In the presence of a large excess of sodium acetylide at low temperatures, this equilibrium is shifted to the right. At high temperatures, however, in the presence of a large excess of sodium alcoholate and in the absence of acetylene, this equilibrium can be shifted in the opposite direction. In the procedure described above, the elimination of the ethynyl group followed by the determination of the 17-keto group formed offers the possibility of deter- mining the 17-ethynyl steroids as shown by the reactions given below.OH ... C S C H Sodium t-butoxide Diethyl oxalate 0 0 The effect of the temperature on the equilibrium in the presence of a 100-fold excess of sodium t-butoxide was investigated, and the results obtained are shown in Table 111. At 0 "C, this equilibrium shifts towards the formation of the ethynyl group, but with increasing temperature the extent of the conversion by the de-ethynylation reaction also increases. At 81 "C, which is the boiling-point of the solvent mixture used, the latter reaction. becomes quantitative. The results given in Table I11 were obtained by the spectrophoto- metric determination of the 17-keto group after its conversion into the corresponding 16-glyox- alyl derivative as described under Experimental. At high temperatures, the equilibrium is established almost instantaneously.TABLE I11 EFFECT OF TEMPERATURE ON THE EQUILIBRIUM TernperaturelOC . . .. . . .. .. .. 0 5 25 50 81 Conversion of the de-ethynylation reaction, per cent. 0 1-1 4-9 53-0 100.0April, 19741 ANALYSIS OF STEROIDS, PART XXIV 221 The quantitative nature of the de-ethynylation reaction at the boiling-point temperature was proved by subjecting the reaction mixture to thin-layer chromatography, which showed that the 17-ethynyl steroids had completely disappeared and that the 17-keto steroids were formed exclusively. It was therefore concluded that the problem of the quantitative con- version of 17-ethynyl into 17-keto steroids could easily be solved. As both the de-ethynylation reaction and the formation thereafter of the glyoxalyl derivative occur in the presence of sodium t-butoxide, the favourable situation exists that after the completion of the elimination reaction, the conversion of the 17-keto steroids into their 16-glyoxalyl derivatives with diethyl oxalate can be carried out in the same reaction mixture without the need for any separation step.The latter reaction reaches completion within 15 minutes at room temperature. The 16-glyoxalyl derivatives of 17-keto steroids are very stable and possess a strong absorption band at 294 nm, at about pH 2.13 In Table I, the spectral data of some 17-ethynyl steroids that were investigated after the above treatment with sodium t-butoxide and diethyl oxalate are summarised. With oestrogens the ring “A” is aromatic and it has an absorption band near the peak of the 16-glyoxalyl-l7-keto group, the effect of which can be nullified by our recently described differential spectrophotometric method.14 For the purpose of uniformity, the latter procedure was used for all materials investigated.It is to be noted, however, that the method described is not suitable for the determination of 17-ethynyl steroids that contain an unsaturated 3-keto group in the A-ring, e.g., ethisterone and norethisterone, as these compounds undergo oxidation in boiling alkali solution giving 6-keto derivatives.l7JS As mentioned in the introduction, examination of the equilibrium enabled us to determine the 17-ketone content of 17-ethynyl steroids, which determination cannot be carried out by the standard diethyl oxalate method13 because the de-ethynylation reaction discussed above occurs at room temperature.As can be seen from the results given in Table 111, this interference can be eIiminated by carrying out the reaction at 0 “C. By using this method, trace amounts of 17-ketone impurity in 17-ethynyl steroids can be determined, which procedure is of value for analytical control in the manufacture of some 17-ethynyl steroids. Some results obtained are collected in Table 11. Obviously, the limitation described in the determination of 17-ethynyl steroids applies also to this determination. The relative standard deviations given in Tables I and I1 are characteristic of the precision of the method. The authors thank Mrs. S. V. Ramakrishnan for her technical assistance. 1 .2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 1s. REFERENCES Klein, S., James, A., and Tuckermann, M., J . Amer. Pharm. Ass., Sci. Edn, 1960, 49, 314. Stenlake, J. B., and Williams, W. D., J . Pharm. Pharmac., 1957, 9, 908. Rosenkrantz, H., Potvini, R., and Skogstrom, P., Analyt. Chem., 1958, 30, 975. Becker, A., and Ehringer, F., 2. analyt. Chem., 1963, 198, 162. Tsilifonis, D. C., and Chafetz, L., J . Pharm. Sci., 1967, 56, 625. Comer, J. P., Hartsaw, P., and Stevenson, C. E., Ibid., 1968, 57, 147. Gorog, S., Acta China. Hung., 1966, 47, 7. “British Pharmacopoeia 1968,” The Pharmaceutical Press, London, 1968, p. 395. Boughton, 0. D., Bryant, R., Ludvig, W. J., and Timma, D. L., J . Pharm. Sci., 1966, 55, 951. Schroff, A. P., and Grodsky, J., Ibid., 1967, 56, 460. Schulz, E. P., Ibid., 1965, 54, 144. Talmage, J. M., and Penner, M. H., Ibid., 1967, 56, 657. Gorog, S., A m l y t . Chem., 1970, 42, 660. -, Analyst, 1971, 96, 437. Gorog, S., and Szepesi, G., Acta Pharm. Hung., 1971, 41, 25. Szepesi, G., and Gorog, S., Ibid., 1971, 41, 30. Cross, J. M., Eisen, H., and Kedersha, R. G., Analyl. Chem., 1952, 24, 1049. Langenbach, R. J., and Knoche, H. W., Steroids, 1968, 11, 123. NOTE-References 7 and 14 are to Parts I1 and XVII of this series, respectively. Received January 8th, 1973 Amended October 23rd, 1973 Accepted November 15th, 1973

ISSN:0003-2654    

DOI:10.1039/AN9749900218

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

222 Analyst, April, 1974, Voi. 99, pp. 222-224 A Stable and Sensitive Colorimetric Method for the Determination of Ergocalciferol (Vitamin D2) by Using Trifluoroacetic Acid BY SAMIR A. GHARBO AND LEO A. GOSSER ( Warren-Teed Research Center, Warren-Teed Pharmaceuticals Inc., 582, West Goodale Street, Columbus, Ohio 43215, U.S.A.) A sensitive and fairly stable colour for the spectrophotonietric determina- tion of ergocalciferol (vitamin D2) following reaction with trifluoroacetic acid has been developed. In the method hydroquinone was used to stabilise the yellow colour, which reached maximum absorbance in not less than 80 s and re- mained stable for a t least 20 s, thus giving the analyst convenient time to carry out the determination. Maximum absorbance was reached a t a wavelength of 490 nm.The colour could be destroyed within 2 minutes by the addition of hydrogen peroxide solution, which property was used to advantage in differential spectroscopy. The absorbance is linear for ergocalciferol solutions containing from 0 to 32 p g , optimum results being obtained in the 5 to 15-pg range. The relative standard deviation of the method is 1-1-76 per cent. ERGOCALCIFEROL (vitamin D,) is isomerised to isotachysterol by reaction with antimony trichloride,l acetyl chloride2 and trifluoroacetic acid.3 This reaction has provided the basis for several chemical methods of assay for ergocalciferol. The method that is widely used is the colorimetric method involving the use of antimony tri~hloride,~~~ but it suffers from several disadvantages6 Clements, Forbes, Olliff and Rogers’ used the fairly stable yellow colour produced with trifluoroacetic acid at 403 nm for the determination ; they also reported that a maximum appeared at 496nm during the first 90s and then rapidly disappeared. Recently, Sklan and 13udowski3 reported that trifluoroacetic acid produced a yellow colour, which reached maximum absorption at 500 nm after 25 s and faded rapidly, but which could be analytically useful.They used the ultraviolet absorbance of the esterified isotachysterol at 290 nm for their determinations after first extracting the ester. During the investigation of chemical methods for the determination of ergocalciferol in pharmaceutical preparations it was noted that the stability of the absorbance at 496 and 403 nm when using trifluoroacetic acid was affected by oxidation in air.It was found that low concentrations of a solution of hydroquinone in chloroform, which acts as an antioxidant, increased the sensitivity of the yellow colour and stabilised the absorbance at 490 nm. This absorbance could be abolished rapidly by the addition of a small volume of hydrogen peroxide solution. This paper briefly describes the optimum conditions for utilising the reaction with trifluoroacetic acid for the determination of ergocalciferol. EXPERIMENTAL REAGENTS- Tri3uoroacetic acid, “OR” grade. CAUTION-Trifluoroacetic acid is an extremely corrosive acid and should be handled Chloroform, analytical-reagent grade. Hydroqztinone solution, 0.1 per cent.-Dissolve 50 mg of crystalline hydroquinone in 1 ml of diethyl ether and dilute to 50 ml with chloroform.Hydrogen peroxide, 30 per cent. solution, reagent grade. Ergocalciferol , U.S. P. Reference Standard. Perkin-Elmer, Coleman Model 124, spectrophotometer. Rotary evaporator. @ SAC and the authors. with care. Prepare a fresh solution daily. INSTRUMENTATION-GHARBO AND GOSSER 223 PROCEDURE- Mix the sample under test, containing 5 to 15 pg of ergocalciferol, with 2 ml of the 0.1 per cent. hydroquinone solution in an organic solvent contained in a 25-ml Erlenmeyer flask. Completely evaporate the mixture under vacuum at 35 to 40 "C. To the residue add 0-5 ml of chloroform, mix, and add 2-0 ml of trifluoroacetic acid, mixing them well and then transfer- ring the mixture to a l-cm spectrophotometer cuvette within 50s.Measure the highest absorbance reached at 490 nm within 1 to 3 minutes from mixing, zeroing the instrument with a solvent blank of chloroform - trifluoroacetic acid (1 + 4). After measurement add 2 drops of hydrogen peroxide solution to the cuvette, mix, and measure the absorbance 2 minutes &lo s after the addition of hydrogen peroxide. Subtract the latter measurement from the former and calculate the amount of ergocalciferol by comparison with a similarly treated standard solution. RESULTS AND DISCUSSION Various parameters were studied in an attempt to increase the intensity and stability of the yellow colour developed with trifluoroacetic acid. The presence of chloroform was necessary for the maximum development of the sensitive colour, while alcohol and, to a lesser extent, ether, hindered its development.The amount of chloroform could be increased from 0.5 to 2.0 ml with no significant effect on the stability of the colour but with a corresponding decrease in sensitivity due to the larger volume of solution. The 4: 1 ratio of chloroform to trifluoroacetic acid recommended by Clements et a1.' resulted in an appreciable decrease in both sensitivity and stability. Hydroquinone in higher concentrations than 0.1 per cent. tended to crystallise out from the chloroform on standing; however, 2 ml of the 0.1 per cent. solution were sufficient to stabilise the colour. The residue obtained by evaporating this solution in chloroform to dryness on a rotary evaporator resulted in only a 4 per cent. decrease in the yield of colour if the residue was maintained under vacuum at 40 "C for 45 minutes.The addition of 30 per cent. hydrogen peroxide solution was sufficient to cause the colour at 490 nm to fade rapidly in the first minute and then to reach a stable intensity after 90 s. The slight change in volume on addition of hydrogen peroxide solution had a negligible effect on the blank value. Nitric acid and potassium permanganate were found to be unsuitable for destroying the colour; one drop of concentrated nitric acid gave an initial incomplete oxidation, followed by a slight increase in absorbance at 490 nm while potassium permanganate dissolved in trifluoroacetic acid, had a slow, decreasing effect on the absorbance at 490 nm with the appearance of a new maximum at 448 nm.Various amounts and dilutions of U.S.P. Reference Standard ergocalciferol in chloroform were used to construct a linear calibration graph. The maximum absorbance attained at 0 10 20 30 32 Concentration of ergocalciferol /pg Fig. 1. Linear relationship of two different samples and their dilutions of reference stan- dard ergocalciferol to the colour development in the trifluoro- acetic acid procedure224 GHARBO AND GOSSER 490 nm was rectilinear with the amount of ergocalciferol over the range from 0 to 32 pg (Fig. 1 ) . The relative standard deviation of the absorbance for six identical aliquots containing 9.1 pg of ergocalciferol was & 1.76 per cent. The visible spectrum obtained approximately 4 minutes after the addition of trifluoroacetic acid (Fig.2) illustrates the absorbance band at 490 nm with only a slight contribution at 403 nm. The maximum intensity of absorbance at 490 nm was attained in not less than 80 s, and was stable for at least 20 s before decreasing slowly, thus leaving ample time for the determinations. Ergocalciferol similarly treated with trifluoroacetic acid but in the absence of hydroquinone, attained its maximum absorbance in less than 1 minute, the value decreasing rapidly (Fig. 3). 350 400 450 500 550 Wavelength/nm Fig. 2. A typical spec- trum of the colour developed in the reaction of ergocalci- ferol with trifluoroacetic acid in the presence of hydro- quinone before (A) and after (B) the addition of hydrogen peroxide t I I A ‘ A f Ti me/ mi nu tes Fig. 3. Time course of the colour formation of ergocalciferol with tri- fluoroacetic acid : A, with hydroquinone (at three concentrations of ergocalci- ferol) ; €3, without hydroquinone; and C, with hydrogen peroxide The method for the determination of ergocalciferol described in this paper holds several important advantages over the colorimetric methods reported in the literature.The colour produced, in addition to being proportional to the concentration, is more sensitive and more stable than those produced by similar procedures. The addition of hydroquinone in order to stabilise the colour overcomes the technical difficulty of obtaining an accurate absorbance value from an unstable colour. The procedure is simple and rapid and requires no special skills ; moreover, in the presence of spectral interferences of a general background nature, the use of hydrogen peroxide for differential spectroscopy provides a distinct advantage. No significant variation in results was observed by different analysts using similar samples. The use of this method for the determination of ergocalciferol in pharmaceutical forrnula- tions containing other vitamins after preliminary purification is currently under investigation. REFERENCES 1. 2. 3. 4. 5. 6. 7. Murray, T. I<., Erdody, P., and Panalaks, T., J . Ass. 08. Analyt. Chem., 1968, 51 (4), 839. Sheppard, A. J., LaCroix, D. E., and Prosser, A. R., Ibid, 1968, 51(4), 834. Sklan, D., and Budowski, P., Analyt. Biochem., 1973, 52, 584. “The British Pharmacopoeia 1968,” The Pharmaceutical Press, London, 1968, p. 140. “The United States Pharmacopeia,” XVIII Revision, Mack Publishing Co., Easton, Pa., 1970, p. 915. Stross, P. S., and Brealey, L., J . Phaym. Pharmac., 1966, 7, 739. Clements, J. A., Forbes, St. J., Olliff, C. J., and Rogers, A. R., Ibid., 1967, 19 (supplement), 1035. Received August 13th, 1973 Accepted October 16th, 1973

ISSN:0003-2654    

DOI:10.1039/AN9749900222

出版商:RSC    

年代:1974

数据来源: RSC