Aszalyst, December, 1968, Vol. 93, $9. 817-820 81 7 The Continuous Polarographic Determination of Small Amounts of Nitroglycerine in Plant Effluent BY A. R. HOLLAND AND A. G. S. BENHAM (Imperial Metal Industries Limited, Summerfeld Research Station , Kidderminster) A fully automatic polarographic method is described for the continuous monitoring of nitroglycerine plant effluent, for which an arbitrary maximum level of 100 p.p.m. of nitroglycerine has been set. Conventional polarographic principles are involved, but the polarograph used has been specially designed, and incorporates some novel features that make it admirably suitable for this and similar work. Constructional details are given of the polarograph, and of an alarm system that operates when the permitted level of nitroglycerine in the effluent is exceeded.IN the manufacture of nitroglycerine by the injection nitration process (Nitroglycerine Aktiebolaget, Gyttorp), an aqueous solution is produced, the nitroglycerine content of which, after neutralisation, is about 0-5 per cent. This solution also contains about 2.5 per cent. of sodium nitrate, 1.5 per cent. of sodium carbonate and 0.1 per cent. of sodium sulphate. At this stage the nitroglycerine is in suspension, and most of it is recovered by a “settling out” process, the remainder (about 0.1 per cent.) being diluted with a 20-fold excess of water before it is discharged to waste. An arbitrary maximum level of nitroglycerine in this effluent has been set at 100 p.p.m., and the problem was how to provide a reliable analytical procedure so that in the event of this level being exceeded any failure at the dilution stage could be detected immediately and rectified.It was therefore considered necessary to have an automatic sampling and analysis system for the continuous determination of levels of nitroglycerine normally present in this effluent, and some form of warning (alarm) that would alert plant operators when the amount of nitroglycerine in the effluent was in excess of the permitted level. Small amounts of nitroglycerine can be determined by a spectrophotometric procedure in which N-1-naphthylethylenediamine is used, but the use of any colorimetric reagent is excluded because of the turbid nature of the effluent, especially during periods of heavy rainfall. Nitroglycerine produces three distinct polarographic waves (-0.25, -0.45 and -0-76 volt versus mercury-pool anode) in a quaternary ammonium halide solution, and a procedure involving the use of this electrolyte, by Williams and Kenyon,l was used in the following experimental work.EXPERIMENTAL CALIBRATION GRAPH- In preliminary experiments with a Cambridge photographic-recording polarograph, a linear calibration graph was obtained from the peak heights recorded at -0.75 volt when simulated effluent samples containing up to 150 p.p.m. of nitroglycerine were analysed; nitrogen was used to de-oxygenate the samples. In further experiments, an alternative (cheaper) base electrolyte (ammonium chloride - potassium chloride) was used; sodium sulphite was added to the electrolyte to replace the use of nitrogen, because it was envisaged that a gaseous de-oxidant would lead to complications in any fully automatic system.Satisfactory results were obtained, and this electrolyte was used in subsequent experiments. 0 SAC and the authors.818 HOLLAND AND BENHAM : CONTINUOUS POLAROGRAPHIC DETERMINATION [AflabSt, VOl. 93 DESIGN OF THE ANALYTICAL SYSTEM The following automatic stages were planned. (i) Sampling of effluent every 15 minutes. (ii) Addition and mixing of base electrolyte. (iii) Transfer of prepared sample to a polarographic cell. (iv) Analysis of prepared sample. (v) Removal of analysed sample from polarographic cell. The sampled eflluent is pumped through a filter into a header tank provided with an overflow. A week's supply of base electrolyte is also stored in another header tank, and this tank is provided with a floating lid to minimise evaporation losses.Supplies of base electrolyte and sample enter the polarographic cell by means of a gravity feed via the solenoid-controlled valves, S1 and S2 (Fig. l), and the chokes are adjusted so that equal volumes of both solutions can pass into the mixing chamber. In this way, the use of low-capacity pumps, which we have not found reliable in operation over long periods, is eliminated. A period of 30 seconds is allowed for the delivery of a 100-ml sample and also for the same volume of electrolyte; the system is self-purging. Electrolyte storage -ve - Sample '-1 Valve Valve 4 Mixer Filter Drain Fig. 1. Schematic layout ELECTRICAL CIRCUITS (FIG 2)- The two mercury cells are connected in series to provide the polarographic cell current via an adjustable resistor, R1, a potential divider, R3 (or R5), R4 and an adjustable resistor, R2. The recorder is connected across R2, and the oscillating potential produced by the dropping-mercury electrode is damped by a condenser connected across R2, so that a constant current of 10pA flows at full-scale deflection of the recorder. The polarographic current flows through the cell via a platinum wire, which is in electrical contact with the (negative) mercury reservoir, and a second platinum wire in similar contact with the (positive) mercury pool in the neck of the flask (Fig.1). The positive current is taken from the centre noint of the notential divider via the microswitch.MS.3. which makes OPERATION- The position of the valve voltmeter, V2, is shown in Fig. 2; the meter reading is adjusted to -1.00 volt by means of R1. The reading on V1 is noted on the moving-coil meter, then V2 is disconnected. Any variation in voltage can be corrected by restoring V2 to the original reading.December, 19681 OF SMALL AMOUNTS OF NITROGLYCERINE IN PLANT EFFLUENT 819 The sampling sequence is programmed by a 4 r.p.h. mains timer, which operates the microswitches, MS1, MS2 and MS3, by a camshaft. At zero time, the mains-operated solenoid valves, S1 and S2, are energised through MS1, thus permitting 1OOml of the sample and the same volume of the base electrolyte to pass into the mixing chamber. After a lapse of 9& minutes to allow the solution to de-oxygenate, the microswitch, MS3, makes contact and allows a potential of -1.00 volt to be applied instantaneously to the cell.The concen- tration of nitroglycerine is shown on the recorder (Fig. 2). After 5 minutes, MS3 breaks to stop the cell current and the sequence is repeated. Switches are provided to enable manual operations to be made, and a second mains timer (12 r.p.h.) is operated by MS2 to permit the slide-wire, R5, to be used when required. The nitroglycerine peak is displayed on a self-balancing potentiometer strip-chart recorder, which is fitted with a mercury switch that operates an alarm system when the nitroglycerine content of the sample exceeds 100 p.p.m. Fig. 1 shows a schematic layout of the unit, and Fig. 2 the electrical circuit.230V Alarm Potent iomet r ic R,, R, = 100-k i2 variable resistors R3, R, = 100-k I2 fixed resistors R2 = 500-k variable resistor SW = Single pole change over Fig. 2. Electrical circuits EFFECT OF VARIABLES- Sodizcm carbonate-At the 50 p.p.m. nitroglycerine level, the presence of 4 per cent. of sodium carbonate gave an apparent nitroglycerine level of 47.5 p.p.m. Under plant operating conditions, this level of sodium carbonate is most unlikely to occur, so that a normal variation in sodium carbonate content is not significant. Glycerol and sodiztm nitrate-The presence of 1 per cent. of glycerol in the effluent would indicate serious malfunctioning of the plant, but even at this level of glycerol the effect on the determined nitroglycerine content of the effluent is insignificant.The same observation was made in the presence of 1 per cent. of sodium nitrate. Calcizcm-Because it is possible that calcium could enter the effluent stream between the plant and the sampling point, the effect of 2 per cent. of calcium oxide was investigated. This produced a lowering of the recorded nitroglycerine content by about 2 per cent. TemfJeratwe-Wide fluctuations in the temperature of the effluent do not occur, and the small structure in which the polarograph is housed is controlled at the same temperature as that used to calibrate the instrument, i.e., 20" C.820 HOLLAND AND BENHAM It was, however, established that a 1” C rise above this temperature increased the peak height of the polarographic wave by about 1.5 per cent. Temperature compensation could be arranged in the electrical circuits, if necessary.METHOD REAGENTS- Base electrolyte-Dissolve 48-1 g of ammonium chloride, 7.5 g of potassium chloride, 0 4 g of sodium sulphite, Na2S0,.7H20, and 0 6 g of gelatine in warm distilled water and cool. Dilute the solution to 1 litre with distilled water and mix. Standard nitroglycerine solzdiovt-Dissolve 100 mg of nitroglycerine in 15 ml of methanol, transfer the solution to a l-litre calibrated flask, dilute to the calibration mark with distilled water and mix. This solution contains 100 p.p.m. of nitroglycerine. CALIBRATION- Establish that 100 ml of sample and 100 ml of electrolyte are discharged into the mixing chamber during each 15-minute cycle. Variable chokes are fitted in each delivery pipe to the mixing chamber, which should be adjusted, if necessary.Mix the base electrolyte and the standard nitroglycerine solution, adjust the temperature to 20” C, transfer the mixture to the polarographic cell via the mixing chamber and record the polarogram at a fixed potential of -1.00 volt. Dilute the standard nitroglycerine solution with distilled water to provide standard solutions containing 20, 50 and 75 p.p.m. of nitroglycerine, and similarly record the polaro- grams obtained from these solutions ; include a blank determination. PROCEDURE- sample by reference to the calibration graph. glycol dinitrate and is usually known as “A-type nitroglycerine.” Proceed as outlined under Calibration, and calculate the nitroglycerine content of the The nitroglycerine referred to throughout this paper does not contain any ethylene CONCLUSIONS The proposed procedure can be satisfactorily used for the continuous determination of nitroglycerine in a typical effluent at the levels of nitroglycerine normally present in these samples, i.e., less than 100 p.p.m., with an accuracy of 1 p.p.m. The polarogram can be recorded and interpreted visually, and the output of the recorder can be fitted with a mercury switch to operate an alarm system when the nitroglycerine content of the effluent exceeds 100 p.p.m. There is no obvious reason why a similar unit should not be used for the continuous determination of certain other constituents in flowing streams, e.g., metallic impurities. The unit described has been in continuous operation for 12 months. No serious instru- mental difficulties have been encountered, and periodic checks with solutions of known nitroglycerine content have confirmed the reliability of the method. The authors acknowledge the assistance given by Mr. D. Facer of this laboratory, who carried out most of the preliminary experiments. REFERENCE 1. Williams, A. F., and Kenyon, D., Talanta, 1959, 3, 160. Received August 16th, 1968