摘要:

APRIL, 1969 THE ANALYST Vol. 94, No. I I 17 The Gas-chromatographic Analysis of Volatile Constituents in Polymers, with Particular Reference to Moisture Content* BY A. R. JEFFS (Hesearcl~ Uepavtuzent, Iwperial Chemical Industries Limited, Plastics Bavision, Bessemer Road, Welwyn Gardeiz City, Herts.) A versatile, laboratory gas-chromatographic unit is described for the determination of water and other volatile constituents in polymer powders. The sample, contained in a disposable glass tube, is heated in a closed loop to release the volatile constituents, which are subsequently swept on to the gas-chromatographic column for analysis. The factors influencing the design of the apparatus, calibration for water and accuracy of the method a t the various levels of moisture found in vinyl, acrylic and polyolefin polymers are discussed.THE determination of water in polymeric materials produced in the plastics industry has always been important because the presence of excessive amounts of water can adversely affect the properties of the material. In polymer used for moulding and film making, for example, surface blemishes on the finished article can be caused by water. Various methods for the determination of water in polymeric materials are available. They include the deter- mination of the loss in weight under defined drying conditions, Dean and Stark distillation methods,l Karl Fischer titration methods applied directly to the sample or to vacuum distillates obtained from the sample2 and infrared spectrophotometric methods. During the last few years, attention has been paid to the use of gas chromatography for water determinations, and Collier and Green3 have described a successful on-line analyser for the determination of water in free-flowing powder samples.A natural extension of this work was the consideration of a versatile laboratory instrument capable of dealing both with powder and granular samples under different temperature conditions. The possibility of determining other volatile constituents of polymers was also considered. This paper describes the development of such a chromatograph and its application to vinyl, acrylic and polyolefin homopolymers and copolymers. Further applications are being investigated, and since this work was initiated further reports have been published on this subject.4s6 APPARATUS- The final form of the general-purpose instrument produced is shown diagrammatically in Fig.1, and consists essentially of a sample tube, forming an external loop, connected to a gas chromatograph. This loop can be isolated and the sample heated to the required temperature. After an initial heating period the volatile constituents liberated from the sample are “flushed” on to the chromatographic column by a flow of carrier gas through, or over, the sample, and the required components separated and determined quantitatively. A pneumatic switch valve located in the chromatographic oven to prevent the condensation of volatile constituents within the valve, and a split heater mounted on a horizontal travel in a plane at right angles to the sample tube, are essential parts of the apparatus.The instrument is semi-automatic. A simpler instrument with manual timing and switching can be constructed. The carrier gas flows through the copper sulphate in tube D, which imparts a constant amount of water (about 3 p.p.m. w/v) to the helium. The “wet” carrier gas prevents the gas-flow lines from “drying out.” Dry pipework tends to adsorb moisture, which can then EXPERIMENTAL * Paper presented a t the SecondISAC Conference 1968, Nottingham. 0 SAC and the author. 24925) JEFFS : GAS-CHROMATOGRAPHIC ANALYSIS OF VOLATILE CONSTITUENTS [AfidySt, VOl. ‘34 be desorbed, thus leading to spurious results. The determination of water in a sample is unaffected as the “wet” carrier gas flows continuously through both the reference and analysis cells of the katharometer.The copper sulphate crystals are sieved and any fine powder rejected. The copper sulpliate tube is changed when two thirds of its length has visibly changed colour. A,, A, and A, = Edwards VPC I pressure controller B,, B, and B, = Pressure gauges 0 to 30 p.s.i. C,, C, and C3 = Rotameter-type flow gauges K, and K2 = Straight reducing couplings, captive seal type, for t o + inch 0.d. tubing (Drallim, Cat. No. L/50/D/B) L = Chromatographic column D = E = F = G = H = J = Clear plasticised PVC tubinn (5 foot lonz x M = Flame-ionisation detector 4 inch’ bore), Packed with crystals, CuS0,.5H,OP >44 mesh sulpkate N, and N,= Electrically actuated 3-port pilot valves (Martinair, Type 557C/IZ) Katharometer P = Nickel - chromium/nickel - aluminium Pneumatic sample valve (Pye Cat.No. I29OO), thermocouple embedded in the split heater fitted with P 9904 change-over block Q = Electronic temperature controller, pro- Internal IOOD portional type External loop made in part of 18-gauge stainless-steel capi I lary tubing Sample split heater R = Sample-valve time-delay unit, containing three synchronous timers and a semi-con- ductor, proportional energy controller for the sample heater Fig. 1. DetaiIs of the general-purpose instrument The pneumatic sample valve, F, operates merely as a switch valve, directing the carrier- gas flow either around the internal loop, G, or the external loop, H. The pilot valve, N,, operates sample valve, F. The sample split heater, J, shown in more detail in Fig.2, consists of a cylindrical aluminium block, 6 inches long with a 9-mm hole through the centre. The block is split axially and the two halves hinged. Each half of the block contains two cartridge-heater elements, each 5 i inches long, 9 inch o.d., and one half contains a thermocouple pocket to accommodate the thermocouple, P. The cartridge heaters are supplied by a semi-conductor energy controller contained in R, which is controlled by a galvanometer, two-position, temperature controller, Q. The heater is mounted on a horizontal travel in a plane at right angles to the sample tube. A jig for mounting the sample tubes is also part of the heater assembly. The #-inch coupling, K,, is brazed to its mounting bracket, which is rigidly attached to the heater base.This coupling is accurately centred with the central hole throughApril, 19691 251 the aluminium block. The coupling, K,, is brazed to the flexible carrier-gas inlet tube, and rests loosely in the second mounting bracket. The $-inch couplings are supplied with neo- prene, or butyl rubber captive seals. Neither of these inaterials is suitable for continuous use at elevated temperatures, especially as the seals are made and broken, up to 40 times a day. Alternative seals have been cut from silicone rubber tubing, 9-mm bore and 2-mm wall, and fitted inside the &inch coupling nuts. Such seals have given trouble-free operation for 2 or 3 days at heater temperatures up to 350" C before replacement was necessary. IN POLYMERS, WITH PARTICULAR REFERENCE TO MOISTURE CONTENT pressure release valve Fig.2. Split heater shown in open position advancing over sample tube The sample-heater assembly is placed on top of a Griffin and George oven, so that the L-bend of &-inch 0.d. S/S tubing disappears almost immediately into an opening on the top of the oven. In practice, both the inlet tube (6 inches long, &inch 0.d. and *-inch i.d. copper) and the exit tube (&inch 0.d.) are wrapped with heating tape and lagging to maintain the temperature of the whole assembly at about 100" C. The inlet tube is wrapped to a length of 4 to 5 inches and the L-bend of the exit tube is wrapped to a point 3 inches inside the oven. Both tubes are wrapped up to, and including, the &-inch thread of the Drallim coupling, leaving only the centre nut and the &inch coupling nut exposed so that the sample tubes can be readily changed.These two tape heaters are connected in series and supplied by a Type V-SHMPS Variac transformer, output 0 to 270 V (2 A). Without these heaters, volatile constituents condense at the coId ends of the sample tube and are swept away com- paratively slowly on switching the carrier gas through the sample. This leads to diffuse and unsymmetrical peaks in the chromatogram. Disposable sample tubes are made by cutting 9-mm 0.d. heat-resistant glass tubing into 'Ii-inch (&J5) lengths. The cut ends of the tubing are fire-polished in a flame. Such tubes will contain from 1.5 to 5 g of sample, depending on the polymer and its physical form. N, acts as a pressure release valve to the external loop, H.The sample-valve time-delay unit, R, contains three synchronous timers: (i) a variable 3 to 60-second synchronous process timer (Chronoset Type CF, D. Robinson & Co.); (ii) a miniature, fixed cycle, synchronous timer (Type MSC, Electrical Remote Control); and (iii) a variable 30-second to 12-minute synchronous process timer (Chronoset Type CF, D. Robinson & Co.). Timers (i) and (ii) are linked, and are initiated by a push button. Timer (iii) operates independently and again is started by a second push button. The unit also incorporates a re-set button that re-sets timers (i) and (iii). The oven contains the column, the katharometer and the pneumatic switch valve so that no condensation can take placein the switchvalve. Other items of the gas-chromatographic equipment not shown in Fig.1 include the katharometer bridge unit, the flame-ionisation ampfifier and the respectiv; potentiometric recorders for the- two detectors (Honeywell, 1 mV, f.s.d.).252 JEFFS : GAS-CHROMATOGRAPHIC ANALYSIS OF VOLATILE CONSTITUENTS piutalyst, vol. 94 The apparatus shown in Fig. 1 is fitted with katliarometer and flame-ionisation detectors. Although only one detector is necessary for any one specific method, e.g., a katharometer for the determination of water in polymer powder, it is invaluable to have both available (with separate recorders) in the preliminary stages to establish the conditions, i.e., in the above case to ensure that no organic components are being eluted at the same time as water, and thus contributing to the peak measurement.Fig. 3 shows chromatograms obtained simultaneously from t h e katharometer and the dried poly(viny1 chloride) powder, flame-ionisation detectors-on a partially P) ’inyl chloride n II II I I I I I I Fig. 3. Chromatograms obtained simultaneously with (a) the katharometer and (b) the flame-ionisation detectors on a partially dried poly(viny1 chloride) powder. In (b) the dotted line represents the position of the water peak as transposed from (a) The chromatographic columns used throughout this work for the determination of water in different polymers consist of 10 per cent. w/w stationary phase on polytetrafluoroethylene powder (I.C.I. “Fluon” CD4). The column packings were made up and packed in the manner described by KirklandeG PROCEDURE- Fit the glass sample tubes with a small quartz-wool plug at about 1 inch from one end, and store them in an oven until just before use.Transfer several tubes from the oven to a desiccator containing phosphorus pentoxide and allow to cool. (The desiccant should be examined at weekly intervals and renewed when necessary.) Use the tubes from the desiccator as required. Weigh a sample tube and pour the sample (powder or granule) into the tube. Re-weigh and insert a second quartz-wool plug with tweezers. These plugs are pre-prepared, oven-dried and stored in the desiccator. Place the sample tube in the heater jig between the couplings and tighten the couplings. Set the heater to the appropriate temperature for the sample under examination.At this point, one of two modes of operation is used, depending on the type of polymer under test. The reasons for this are discussed under Results. METHOD I (FOR VINYL AND ACRYLIC POLYMERS)- Set timer (iii) for 5 minutes. Open the split heater, move it forward on its travel and close it over the sample tube. Operate the push button for timer (iii). After the set time of 5 minutes the pilot valve, MI, is triggered, which operates the pneumatic switch valve, F, thus allowing helium carrier gas to flow around the external loop, H, through the sample tube and flushing the liberated volatile constituents on to the gas-chromatographic column.Apd, 19691 IN POLYMERS, WITH PARTICULAR REFERENCE TO MOISTURE CONTENT 263 The carrier gas remains routed around the external loop, H, until the re-set button is pressed.It is then diverted through the internal loop, G. Figs. 4 (a), (b) and (c) show typical chromato- grams that have been obtained in this way. Carry out a blank determination in a similar manner with an “empty” glass tube containing only two quartz-wool plugs, and subtract any blank from the water in the sample. Water - Blank Water A Sample Blank (b) Vinyl .’acetate c Sample Air / \ Methyl - meth- ac ry I at e Water / Blank .h-- Fig. 4. Typical chromatograms obtained with Method I for (a) poly(viny1 chloride) powder; (b) poly(viny1 chloride) - poly(viny1 acetate) copolymer powder; and (c) acrylic moulding powder METHOD 11 (FOR POLYOLEFIN POLYMERS)- Operate the first push button on unit R. This starts timer (i) set for 16 seconds and at the same time activates pilot valve, N,, which operates the pneumatic switch valve, allowing helium carrier gas to flow around loop H, through the sample tube (still at room temperature) for 15 seconds and sweep the sample free from air.At the end of this 15-second purge period the pneumatic switch valve closes, thus isolating the sample. After a further 2 seconds’ delay [fixed timer (ii)] the second pilot valve, N,, is activated for 2 seconds, allowing the excess of helium pressure in the closed sample loop to escape to atmosphere via the tee-piece and capillary tubing attached to N,. This valve sequence follows auto- matically on operating the first push button, and its completion is denoted by an indicator light. Open the split heater, move it over the sample and close.Operate the second push button for timer (iii). The sequence described under Method I then follows. Two chromatograms are obtained. The first, resulting from the 15-second purge of helium through the cold tube, shows a peak for air followed by a small peak for water. This is developed while the sample is being heated for 5 minutes. The second chromatogram, after the heating period, shows the liberated water followed by a peak for hydrocarbon diluent (the polymerisation medium) if present. Carry out a blank determination in the same manner on an “empty” tube containing two quartz-wool plugs only. Add the values for the two water peaks resulting from each determination and correct the sample figure for the blank value. Typical chromatograms obtained with polyolefin polymers are shown in Figs.5 (a) and (b). CALIBRATION- In Method I, for polymers other than polyolefins, the total moisture is liberated from the polymer powder at temperatures below the melting-point of the polymer. In these cases calibration is readily carried out by injecting known amounts of water by Hamilton syringe, 0 to 10 pl (or 0 to 50 pl), into a polymer sample that has already been dried by subjecting it to the test, i.e., after testing the sample uncouple the tube from the heater jig and cool254 JEFFS : GAS-CHROMATOGRAPHIC ANALYSIS OF VOLATILE CONSTITUENTS [AutabSt, VOI. 94 it in the desiccator. Inject a known amount of water and re-test. Repeat with different amounts of water. Construct a calibration graph relating the known amounts of water added to the peak height (or area) obtained.Carry out a spot check of at least one calibration point each day. /Air\ Blank Hydro- carbon I Sample Air / \ 4 ample Fig. 5. Typical chromatograms obtained with Method I1 for (a) polypropylene powder ; and (b) high pressure polyethylene For the polyolefins, the polymer is heated above its “melting-point” to obtain the total moisture; a horizontal sample tube is, therefore, used. The powder, provided it is not packed too tightly, contracts on melting and sinks to the bottom of the tube, thus allowing a free passage for the carrier gas above the sample. In this case calibration is best carried out with barium chloride crystals, BaC1,.2H20, weighed into empty sample tubes. Barium chloride loses its water of crystallisation, 14-75 per cent.w/w, at 115” C. Chromatograms obtained in this manner are similar to those obtained on samples, i.e., a small water peak, due to the atmospheric moisture, is obtained when the tube is purged with helium for 15 seconds, followed by a larger water peak after the heating period when the water is released from the crystals. If calibration is carried out by direct injection of water on to a quartz-wool plug in an “empty” tube with a Hamilton syringe, 75 to 90 per cent. of this water is purged on to the column during the 15-second period. Chromatograms are, therefore, obtained that are quite unlike those obtained for a sample, although the total amount of water should be unchanged. Weigh about 12 mg of barium chloride crystals into an empty sample tube and obtain chromatograms as indicated under Procedure.Repeat with 9, 6, 3 and 1-mg amounts of barium chloride crystals. Add the values for the two peaks obtained for water in each determination and correct for the blank. Construct a calibration graph relating milligrams of water to the corrected total peak height (or peak area) measured from the recorder chart. DISCUSSSION AND RESULTS OBTAINED WITH VARIOUS POLYMERS Before carrying out any quantitative work on the volatile constituents obtained from the polymer powder, a preliminary gas-chromatographic investigation of their complexity is necessary. For example, Fig. 3 (flame-ionisation detector trace) shows nine components other than air, water and the original monomer. These components are chlorinated hydro- carbons, such as 1,l- and 1,2-dichloroethane and cis- and trauts-dichloroethylene, that are present as impurities in the original monomer.Although these impurities are present only in p.p.ni. amounts in the original monomer they can be readily detected in the polymer. This is not surprising if one considers that 2.8 g of poly(viny1 chloride) polymer (an amountApril, 19691 255 readily examined in the apparatus) is equivalent to a 1-litre sample of vinyl chloride gas at N.T.P. If water is to be determined, then a suitable column has to be chosen so that the water is eluted free from organic constituents. At this stage the recording of simultaneous chromato- grams with the two different detectors is invaluable. Secondly, having chosen a suitable column, the temperature at which a given polymer liberates all of the moisture has to be found.This is determined by keeping the gas-chromato- graphic conditions constant and heating known amounts of polymer at various split-heater temperatures. The peak height (or peak area) per gram of sample is then plotted against the sample temperature. Figs. 6 and 7 are interesting examples of such graphs, which will be discussed under their individual headings. POLY (VINYL CHLORIDE) HOMOPOLYMERS- A graph of water liberated against sample temperature showed that a constant amount of water was evolved from the polymer at temperatures between 115" and 160" C. A heater temperature of 125" C was, therefore, chosen and the following gas-chromatographic conditions were used: column, 10 foot (& inch 0.d.) stainless-steel tubing, packed with 10 per cent.w/w poly(ethy1ene glycol) 400 on "Fluon" CD4; oven temperature, 90" C; and helium inlet pressure, 10 p s i . Fig. 4 (a) is a typical chromatogram obtained under these conditions by Method I. Studies on polymer drier efficiencies can be readily carried out with the apparatus, and it is interesting to note that the water in the polymer is more easily eliminated than the unchanged monomer, which boils at -13" C. Moisture figures in the range 0.02 to 2 per cent. w/w have been determined in homopolymers, although finished polymer powder normally contains 0.1 to 0.2 per cent. of water. VINYL CHLORIDE - VINYL ACETATE COPOLYMERS- The major volatile constituent in copolymers is unpolymerised vinyl acetate, and this monomer is difficult to eliminate from the polymer by ordinary drylng methods.Fig. 4 (b) shows a typical chromatogram obtained by Method I, the gas-chromatographic column and conditions being identical with those described under poly(viny1 chloride) homopolymers. I N POLYMERS, WITH PARTICULAR REFERENCE TO MOISTURE CONTENT M W L a E E c, J= M W c Y W .- a * O Y 40 ._._._._._.-._.-.-.-._._._._ I ._.-.-- I I I.;%.-.-*; I '-*-.-.r.CC ....................................................................... ... ...... .. ..... *.<.: ..... ::. 20 D# -- - J *.*.......... ... --- '1 I 5 I25 I35 I45 I55 I65 Sample temperature, OC Fig. 6. Graphs showing the amounts of (A), vinyl acetate; (B), water; (C), vinyl chloride; and (D), acet- aldehyde obtained from a sample, with increasing sampIe temperature The plot of volatile constituents (water, vinyl chloride monomer and vinyl acetate) against the sample temperature (Fig.6) shows that the maximum amount of vinyl acetate is eliminated from the sample under the test conditions, only in the temperature range 130" to 145" C. Above this temperature the amount of vinyl acetate detected starts to decrease. A similar effect is found for water. At the same time, another peak begins to appear in the chromatogram. This peak corresponds, in retention time, to acetaldehyde. This is consistent with the hydrolysis of the vinyl acetate monomer by water at the elevated temperatures, to give the unstable vinyl alcohol and, hence, acetaldehyde. A heater temperature of 135" C was adopted for these copolymers.256 JEFFS : GAS-CHROMATOGRAPHIC ANALYSIS OF VOLATILE CONSTITUENTS [AlzaZyst, Vol.94 ACRYLIC POLYMERS- In the case of acrylic mouldings powders based on methyl methacrylate, Method I is again used and a typical chromatogram is shown in Fig. 4 (c) . The chromatographic conditions used were similar to those used for poly(viny1 chloride) polymer but, as the chromatogram was less complex, a shorter column (4 foot long) of the same packing could be used. This resulted in a quicker throughput. A plot of moisture liberated against sample temperature indicates that 160°C is the most useful temperature to operate the split heater. A new problem becomes evident, however, on examining polymers of this type. Because of the polar end groups of the molecule, acrylic powders are more hygroscopic than the other polymers discussed and care must be taken in storing samples awaiting test.Again, in dis- pensing the sample for test the polymer should be exposed to the atmosphere for the minimum possible time. A sample dried to a water content of 0.05 per cent. will soon contain 0.2 per cent. w/w of moisture if stored in contact with the atmosphere. As Fig. 4 (c) shows, a peak due to unpolyrnerised methyl methacrylate is also obtained and, at first sight, this also appears to be a method for determining residual monomer as well as moisture. Unlike vinyl chloride and vinyl acetate, however, the unchanged methyl methacrylate in polymer appears to under- go further polymerisation during the 5-minute heating period, as the results obtained by the gas-chromatographic method are lower than the results obtained by infrared and chemical methods.POLYOLEFIN POLYMERS- As seen under Procedure, the difference between the two methods is that in Method I the polymer is heated in the presence of the residual air in the sample tube, and in Method I1 the residual air in the sample tube is replaced with inert carrier gas, helium, before the sample is heated. With the non-olefinic polymers already discussed the same results for water content are obtained by both methods, ie., it is not necessary to replace the air with inert gas. Polyolefins, however, are oxidised if heated in the presence of air, thus producing additional amounts of water. This is indicated in Fig.7, which shows the amounts of water liberated from polypropylene with increasing sample temperatures. The dotted line represents the results obtained by Method I, heating the polymer in the presence of air, and the continuous line represents the results obtained by replacing the air with inert gas, Method 11. aJ =a- € ,x ( b ) 3 I X -0.25 ?% $ %*<,-3-X (0) I I 3 L aJ a. I I I I 1 I X I E I I 3 (4 7% X/X0 (4 I , X-*” r aJ I I 2(‘ .- Ma- 09 s, - x z z . D 00 4 -0.25 g 3 -7April, 19691 IN POLYMERS, WITH PARTICULAR REFERENCE TO MOISTURE CONTENT 257 and cut into small pellets). The continuous-line plots obtained by Method I1 indicate that all the moisture is not released from the polymer until the “meltingpoint” is approached in the case of powder samples, and is exceeded in the case of granular samples.The heater tem- perature used for the determination of moisture in polypropylene samples was, therefore, chosen as 244” C, as this temperature is applicable to both powder and granular samples. The gas-chromatographic conditions used for the determination of moisture in poly- propylene were: column, 5 foot (4 inch 0.d.) stainless-steel tubing, packed with 10 per cent. w/w “Ucon” 50 HB 2000 fluid on “Fluon” CD4 powder; oven temperature, 100” C; and helium inlet pressure, 6 p.s.i. Fig. 5 (a) is a typical chromatogram for this determination. The normal range of moisture found in polypropylene polymers is from 0.01 to 0.05 per cent. w/w. Polyethylene, made by the high pressure process, contains only very small amounts of moisture and, as Fig.5 (b) shows, the water content in the blank determination is only marginally smaller than that of the sample. The difference between the two, on the sample illustrated, amounts to about 40 pg, or about 15 p.p.m. w/w (or 04015 per cent. w/w) of water in the sample. At this higher sensitivity and with lower amounts of water, the blank deter- minations can be somewhat erratic unless great care is taken with the storage of the dried tubes. For example, a tube that has been stored overnight in a desiccator over phosphorus pentoxide will give a much lower blank than a tube taken from the desiccator after a morning’s work when the desiccator has been opened and closed fifteen to twenty times in 2 or 3 hours. The reproducibility at these levels is, therefore, not good, and is in the region of +SO per cent.Plots relating the temperature of the sample to the amount of moisture liberated per gram of sample show that 150” C is the minimum temperature that should be used and that, in the presence of air, oxidation, giving rapid formation of water, starts at 170” C. It is, therefore, essential that Method I1 be used. The gas-chromatographic conditions used were as described for polypropylene and the heater temperature used for polyethylene was 200” C. The amounts of water involved in the polymers discussed (with the exception of poly- ethylene), in general, range from 0.02 to 0.5 per cent. w/w. Within this range the repro- ducibility given by the gas-chromatographic method is +5 per cent. relative. Table I shows the reproducibility on ten consecutive tests carried out on a sample of polypropylene powder.TABLE I TYPICAL REPRODUCIBILITY OF THE GAS-CHROMATOGRAPHIC METHOD Weight of sample, g 2.372 2.457 2.304 2.527 2.435 2.512 2.409 2.345 2.506 2.430 Peak height, mm 43.0 43.0 40.5 44.7 40.2 43.8 39.2 43.3 45.0 42.3 Water, per cent. w/w 0.028 0.027 0.027 0.027 0.025 0.027 0.025 0.028 0.028 0.027 The accuracy given by the method is more difficult to define, but the results obtained on acrylic moulding powders are in good agreement with a solution infrared method, and on polypropylene the results again show good agreement with the carefully carried out heating, trapping and Fischer titration method. Table II shows some typical results obtained by the gas-chromatographic method on polymer powders, compared with other methods of determining water.The Fisher titration method is applied to a vacuum distillate of the sample. Finally, the method has also been used to determine the unreacted secondary monomer in butadiene-based latices. In this instance a short length of glass tubing (1.5 to 2 inches), of suitable bore such that it can be slipped inside the normal sample tube, is packed loosely with cotton, glass or quartz-wool. Latex (5 to 10 drops) is weighed on to the wool and the258 JEFFS TABLE I1 COMPARISON OF THE PERCENTAGE OF WATER FOUND IN POLYMERS BY DIFFERENT METHODS Water, per cent. w/w by chromato- infrared Fischer Polymer P P h Y spectroscopy titration I A > gas Poly(methy1 methacrylate) Sample 1 . . 0.102 0.09 0.052 0.108 } 0.152 } Sample 2 .. 0.160 0.14 0.12 Poly (vinyl chloride Sample 1 . . 0.112 - 0.10 0.109 } 0.019 } Sample 2 . . 0.042 - 0.05 Polypropylene Sample 1 . . 0.033 - 0.03 0-041 } 0.035 } Sample 2 . . 0.022 - 0.02 short tube inserted into the sample tube. The sample tube is connected into the heater jig and the sample tested by Method I. With a suitable gas-chromatographic column, acrylo- nitrile and styrene can be determined. By re-weighing the sample tube after this treatment the total solids content of the latex can also be determined if required. I thank Mr. G. L. Collier for helpful discussions on the construction of the instrument and Mr. T. Kennedy who carried out work on the acrylic polymer powders. REFERENCES 1. 2. 3. 4. 5. 6. Dean, E. W., and Stark, D. D., Ind. Engng Chem., 1920, 12, 486. Haslam, J., and Clasper, M., Analyst, 1952, 77, 413. Collier, G. L., and Green, R. E., Measurement G. Control, 1968, 1, 18. Harrington, R. C., and Keister, D. C., Paper presented a t the Plastics-Paper Conference of the Technical Association of the Pulp and Paper Industry in New York City, September, 1966. Rice, D. D., and Trowell, J. M., Analyt. Chem., 1967, 39, 157. Kirkland, J. J., Ibid., 1963, 35, 2003. Received July 31st, 1968 Accepted October 16th, 1968

ISSN:0003-2654    

DOI:10.1039/AN9699400249

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

Analyst, April, 1969, Vol. 94, pp. 259-261 259 Charge-transf er Chromatography of Polycyclic Hydrocarbons BY G. D. SHORT* AND R. YOUNG7 (Royal Naval Scientijic Service, Admiralty Materials Laboratory, Holton Heath, Poole, Dorset) It is shown that the chromatographic behaviour of polycyclic hydro- carbons is modified by the relative strength of charge-transfer interactions when an insoluble electron acceptor, such as pyromellitic dianhydride, is dispersed in the solid phase. Improved separation of mixtures can be obtained, and brightly coloured spots fluorescing in the visible region alIow ready identification of components. THE ability of polycyclic hydrocarbons to form charge-transfer complexes can, in principle, form the basis of a method for their separation. Electron acceptors that have been used for this purpose are trinitrobenzene, picric acid, styphnic acid, trinitrofluorenone , chloranil and 9-dicyanomethylene-2,4,7-trinitrofluorenone.~~2 93 Cooper, Crowne and Farrell have also reported charge-transfer forces to be of importance in separating amines by gas - liquid chromatography, in which trinitrofluorenone is used in the stationary p h a ~ e .~ , ~ We have investigated the use of two acceptors of the anhydride type for the separation of polycyclic compounds by thin-layer chromatography, namely, pyromellitic dianhydride6 and 1,5,6,8- naphthalenetetracarboxylic acid dianhydride.7 Both compounds are commercially available , form fluorescent charge-transfer complexes with aromatic compounds8 and possess the advantage of being virtually insoluble in many solvents.Consequently, for chromatographic purposes it is convenient to introduce the acceptor material by mixing it with the slurry in fine powder form before coating the plates, thus ensuring a homogeneous, reproducible layer. A thin-layer chromatographic plate, the layer material of which contains 30 per cent. w/w of an electron acceptor (a value found to give optimum performance in this work) would be expected to give lower RF values, a prior& than an untreated plate, because most of the sample hydrocarbon would be rapidly converted into the charge-transfer complex on applica- tion to the treated plate. This reduces the competition between the eluent and the sample for interstices on the silica gel, and hence the eluent should tend to pass up the plate leaving the charge-transfer complex behind.The treated layer must be made slightly thicker to compensate for the presence of non-adsorptive acceptors, although RF values vary to some extent with layer thickness. To a first approximation the energy of charge transfer, which determines the colour of the complex, is proportional to Ip, the ionisation potential of the donor, for a series of complexes with the same acceptor. Hence a complex containing a donor of low ionisation potential will form a strongly bound complex that will absorb a t longer wavelengths than a corresponding complex formed by a donor of high Ip. Thus, a yellow complex would be less strongly bound than a purple one and, accordingly, the donor should possess a relatively higher RF value. Clearly, steric and solubility factors must also also play an important r61e in some instances.APPARATUS AND MATERIALS- All of the hydrocarbons used (Table I) were dissolved in either benzene or cyclohexane in concentrations of about 5 x M, and 5-p1 aliquots applied 3 cm apart, 1.5 cm from the lower edge of 20 x 20-cm plates. Pyromellitic dianhydride and naphthalenetetracarboxylic EXPERIMENTAL * Present address : I.C.I. Petrochemical and Polymer Laboratory, The Heath, Runcorn, Cheshire. t Present address : Department of Chemistry, University College, London. 0 SAC and the authors.260 SHORT AND YOUNG : CHARGE-TRANSFER CHROMATOGRAPHY [ANdySt, VOl. 94 dianhydride were each purified by two recrystallisations from dioxan. Basic apparatus was supplied by Camlab (Glass) Ltd., and prepared plates (Merck Laboratory Chemicals) by Anderman and Co. Ltd.A Pascall ceramic ball mill was used for milling and dispersion of acceptor material. TABLE I Eluent cyclohexane, with plates prepared by Method (b) RP VALUES AND RATIOS WITH RESPECT TO ANTHRACENE Treated plates * Hexamethylbenzene . . 0.89 1.10 N-Phenylcarbazole . . 0.81 1.00 1,3,6,8-Tetraphenylpyrene 0-73 0.90 Hydrocarbon R g value Rp ratio Naphthalene . . . . 0.97 1.20 f 0.05 Anthracene .. . . 0.81 1.00 Phenanthrene . . . . 0.81 1.00 Benzo[a]fluorene . . . . 0.69 0.85 Fluoranthene . . . . 0-57 0.70 Dibenzo[a,c]anthracene . . 0.53 0.66 Pyrene . . .. . . 0-53 0.65 Naphthacene . . . . 0.47 0.68 Perylene . . .. . . 0.45 0.56 Benzo[a]pyrene . . . . 0.32 0.40 Benzo[b]fluoranthene .. 0.32 0.40 Carbazole . . .. . . 0.16 0.20 9,lO-Dimethylanthracene 0 . o Untreated plates - Rp value R g ratio 0-78 1-30 f 0.06 0.72 1.20 0-6 1.0 0.54 0.90 0.24 0~40 0.42 0.70 0.45 0.75 0.54 0.90 0.24 0.40 0.6 1.0 0.39 0.65 0.39 0.66 0.36 0.60 0.39 0.65 0.06 0.10 0.56 0.93 Spot colour Yellow Yellow Pink Yellow Brick red Green Brown Yellow - green Orange Orange Olive green Olive green Orange - brown Orange - brown Orange Purple Ip, eVs910 8.1 7.9 7.4 8.15 - - 7-82 7.5 7-0 7.2 - I PLATE PREPARATION AND PROCEDURE- Preliminary experiments showed that naphthalenetetracarboxylic acid dianhydride was inferior as an electron acceptor because its charge-transfer complexes were unstable and had low fluorescence efficiencies. Pyromellitic dianhydride, on the other hand , formed brightly coloured complexes that showed strong fluorescence ; hence experimental work was concen- trated on the latter.Silica gel, on which hydrocarbons are poorly separated, was deliberately chosen as adsorbent so that results could more readily be analysed in terms of improved separation. RF values obtained from treated plates were directly compared with values from untreated plates by eluting identical samples with the same solvent under identical conditions. The distance travelled by the solvent front was 10cm in all instances. Anthracene was chosen as a reference substance, and all RF values obtained were compared with those of anthracene samples eluted simultaneously on the same plate. In this way allowance could be made for inevitable differences in layer thickness and composition between plates.Three basic methods of preparing plates were developed and are described below. Methods (b) and (c) were identical in performance, while method (a) was discarded because of poor reproducibility both in preparation and in results. The full series of RF values for samples eluted on plates prepared by method ( b ) are given in Table I. Method @)-A normal silica-gel plate was sprayed with a saturated solution of pyro- mellitic dianhydride in acetone, dried for 20 minutes, spotted with sample and eluted with pure cyclohexane. Method (b)-Pyromellitic dianhydride (60 g) was mixed with 130 ml of dry carbon tetra- chloride and milled for 3 hours, when the viscosity of the suspension ceased to increase.Silica gel (120 g) was added, together with an additional 250 ml of carbon tetrachloride, and the mixture milled again for 30 minutes. The resulting slurry was then used to prepare thin-layer chromatographic plates (12 x 2.5 x 0.4 cm) by a dipping method. Plates were dried for 20 minutes, spotted, dried for a further 10 minutes, eluted with cyclohexane, dried in darkness for 10 to 15 minutes and heated with hot air to complete dryness. Some plates were sprayed with a saturated solution of pyromellitic dianhydride in acetone to produce extra colour. The colours faded after about 5 days. The treated slurry was kept dry and away from light but lost its colour-forming properties in about 5 days. Re-milling and treat- ment with acetic anhydride did not restore the activity of the slurry.April, 19691 OF POLYCYCLIC HYDROCARBONS 261 Method (c)-A method of incorporating pyromellitic dianhydride into commercially produced thin-layer chromatographic plates that have exceptionally tough layers was evolved.The plates were soaked in a saturated solution of pyromellitic dianhydride in acetone and dried over hot air. This was repeated several times and thus enabled uniform layers of pyromellitic dianhydride to be applied as required, with no attendant “ageing” problems. DISCUSSION Table I shows the RF values obtained for the sample hydrocarbons. The values are qualitative because the layer parameters varied from plate to plate, but the results show that a significant change in elution characteristics can be obtained with most hydrocarbons.The stronger donors, such as 9,lO-dimethylanthracene which forms a purple charge-transfer complex with pyromellitic dianhydride, have lower RF values on the treated plates than on the untreated plates. Thus, complete separation of all other hydrocarbons from 9,lO-di- methylanthracene could be achieved by elution on treated plates. The weaker donors,such as hexamethylbenzene (yellow with pyromellitic dianhydride), have relatively large RF values on both treated and untreated plates. The most striking separation was achieved for pyrene - anthracene mixtures, which eluted identical distances on untreated plates, but were readily separated on pyromellitic dianhydride treated plates. Similarly, mixtures of N-phenyl- carbazole - dibenzo [a,c]anthracene, phenanthrene - fluoranthene and naphthacene - benzo- [blfluoranthene could be separated on treated plates while giving identical R F values on untreated plates.Phenanthrene has an RF ratio less than unity, although its ionisation potential is higher than that of anthracene. This is an instance in which steric and other factors predominate over charge-transfer complex formation. Similar considerations apply to carbazole. Naph- thacene and 1,3,6,8-tetraphenylpyrene, although highly coloured themselves, did not appear to undergo charge-transfer complex formation under our conditions. CONCLUSION The incorporation of an electron acceptor into a layer of silica gel can result in the improved separation of polycyclic hydrocarbons by thin-layer chromatography. By using pyromellitic dianhydride as acceptor, brightly coloured and highly fluorescent charge-transfer complexes are formed, thus readily allowing visual identification of the components of a mixture.The hope that RF values would be directly proportional to ionisation potential has been only partly fulfilled, showing that charge-transfer interactions are of the same order of magnitude as other chromatographic factors. However, use of an acceptor of higher electron affinity than pyromellitic dianhydride would render the charge-transfer effects of greater relative importance. Thus a further extension of this work, by using a compound with a high electron affinity and low solubility in cyclohexane, may well result in even better separations than have been obtained with pyromellitic dianhydride. This paper is published with the permission of the Navy Department, Ministry of Defence. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. Harvey, R. G., and Halonen, M., J . Chromat., 1966, 25, 294. Berg, A., and Lam, J., Ibid., 1964, 16, 157. Frank-Neumann, M., and Jossang, P., Ibid., 1964, 14, 280. Cooper, A. R., Crowne, C. W. P., and Farrell, P. G., Trans. Faraday Soc., 1966, 62, 2725. -,-,- , Ibid., 1967, 63, 447. Nakayama, Y., Ichikawa, Y., and Matsuo, T., Bull. Chem. S O ~ . Japan, 1965, 38, 1674. Jacquignon, P., Buu-Hoi, N. P., and Mangane, M., Bull. Chim. Soc. Fr., 1964, 10, 2517. Ilmet, I., and Berger, S. A., J . Phys. Chem., 1967, 71, 1534. Briegleb, G., Czekalla, J., and Reuss, G., 2. phys. Chem., 1961, 30, 333. Briegleb, G., and Czekalla, J., 2. Elektrochem., 1959, 63, 6. Received August 12th, 1968 Accepted November 18th, 1968

ISSN:0003-2654    

DOI:10.1039/AN9699400259

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

262 Analyst, April, 1969, Vol. 94, pp. 262-268 The Polarographic Behaviour of 0-Hydroxy-o’-carboxyazo Dyes and Their Beryllium Complexes* BY A. G. FOGG, J. L. KUMAR AND D. THORBURN BURNS (Defiartment of Chemistry, Loughborough University of Technology, Loughborough, Leicestershire) The basis of a method for the determination of beryllium by using the displaced polarographic wave obtained with Mordant red 74 (C.I. 16315) in the presence of beryllium ions is described. Aluminium ions do not give a displaced wave with this dye, and an amount of aluminium five times that of beryllium can be tolerated. Nickel and iron(II1) give displaced waves and interfere with the determination. THE polarographic half-wave potentials of metal - o,o’-dihydroxyazo dye complexes are, in many instances, more negative than those of the corresponding free dyes.The measurement of the height of these displaced waves has provided several convenient analytical methods for the determination of metals, in particular those of the more electropositive metals such as aluminium and magnesium. Recently, Latimer’ has reviewed these methods compre- hensively, and it will suffice to list the metals that have been shown to give displaced waves with o,o’-dihydroxyazo dyes: Mg, Al, Ca, Ti, V, Cr, Mn, Fe, Co, Ni, Zn, Ga, Y, Zr, Cd, In, La, Th, U and the rare earth metals Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb and Lu. As several metals do not give a displaced wave, some selectivity is possible in these methods; the method for determining aluminium in the presence of beryllium with Mordant violet 5 is particularly useful.2 Other metals, such as magnesium and calcium, form com- plexes, and give displaced waves, only in solutions with a high pH, and they do not interfere in the determination of other metals in solutions with a low pH.The displaced waves given by two metals with the same dye are not usually separated sufficiently for analytical purposes, although Rooney and McIveI3 have shown that iron(II1) and aluminium can be determined with Mordant violet 5 on one cathode-ray polarogram. Usually, the selection of a particular o,o’-dihydroxyazo dye for the determination of a metal is made because of the slightly superior polarograms obtained with this dye and not because of any improved selectivity that accrues from its use.Aluminium, iron(II1) and a number of other metals give displaced waves with all of the dyes of this group that have so far been studied. A. G. Fogg, C. M. Pyburn and G. F. Reynolds (unpublished work) and Pyburn,4 however, obtained a displaced wave for Mordant red 67 with calcium but not with magnesium, whereas a previous worker1 has shown that Mordant violet 5 gives a displaced wave with magnesium but not with calcium. Consideration of the advantages of these polarographic methods with o,o’-dihydroxyazo dyes led F o g , Pyburn and Reynolds to investigate complexes of several trichelating azo dyes, and they obtained displaced waves for iron(II1) with 4-(2-pyridylazo)resorcinol (PAR) and for beryllium with the o-hydroxy-0‘-carboxyazo dye, Mordant red 9 (C.I. 16105), I.The limited number of metals shown to give displaced waves with azo dyes other than those which are o,o’-dihydroxy-substituted makes these dyes quite selective. The possibility of using the displaced wave obtained with Mordant red 9 and beryllium is particularly attractive, as aluminium does not give a displaced wave with this dye. W I Paper presented a t the Second SAC Conference 0 SAC and the authors. Q so; 1968, Nottingham.FOGG, KUMAR AND BURNS 263 Mordant red 9 gives a 1 : l beryllium-dye complex,* which is formed in solutions a t pH between 4 and 7-5. In solutions at pH higher than 7-5, however, there is evidence of the hydrolysis of the complex, and polarography is restricted to solutions with a pH lower than 7.5. Part of the single dye wave (-0.45 V at pH 4) is displaced about 100 mV to more negative potentials in polarograms obtained with solutions with pH between 4 and 7.5, but the most clearly defined polarograms were obtained at pH 5.5.A curved calibration graph is obtained for the determination of beryllium, and the height of the displaced wave continues to increase up to a beryllium concentration of at least three times that of the dye. The dye wave has a polarographic maximum superimposed upon it, and this is transferred to the displaced wave in the presence of beryllium. In the present paper a polarographic study of the beryllium complexes of other o-hydroxy- 0’-carboxyazo dyes is described. Mordant red 74 is recommended as it is more satisfactory than Mordant red 9 for the polarography of beryllium, and the optimum solution conditions are given.EXPERIMENTAL The preliminary polarographic study of the dyes was made by using Commercial samples without prior purification. For the more detailed study of Mordant red 74, a sample of the acidic form of the dye was obtained by extraction of an acidified aqueous solution of the commercial dye with diethyl ether, and was shown potentiometrically to be 96 per cent. pure. Spectrophotometric and polarographic studies of the commercial Mordant red 74 and of the sample of the acidic form of the dye showed that the commercial sample contained 72 per cent. of Mordant red 74 as the disodium salt. The complexation of beryllium by the dyes was studied potentiometrically by using a Pye 290 pH meter, with an E.I.L. GHNS glass electrode.Polarograms were obtained by using a Cambridge Pen-recording polarograph and a Southern-Hamell, Mark 11, pulse polarograph. All half-wave potentials are quoted versm the mercury pool, except when otherwise indicated. PRELIMINARY STUDIES- The o-hydroxy-0’-carboxyazo dyes, Mordant red 60 (C.I. 15975), 11, Mordant yellow 8 (C.I. 18821), 111, and Mordant red 74 (C.I. 16315), IV, and their beryllium complexes, were examined polarographically. II so; Mordant red 60-Stock solutions of Mordant red 60 were prepared by adding just sufficient 0.1 M sodium hydroxide solution to dissolve the dye and then making the resulting solution up to volume with water. The dye exhibited a well defined polarographic wave of half-wave potential -0.35 V in solutions at pH 4.8.Polarograms were obtained for solutions 1 M with respect to buffer and 0.1 M with respect to potassium chloride at pH values between 4 and 7264 FOGG, KUMAR AND BURNS: POLAROGRAPHIC BEHAVIOUR OF [Artalyst, Vol. 94 (acetate buffer pH 4 to 5.5, phthalate buffer pH 5.5 to 7); the half-wave potentid of the wave was displaced the expected 59 mV (approximate value) to more negative potentials per unit increase in pH. At dye concentrations exceeding 4 x lO-4~, the dye wave had a polarographic maximum superimposed upon it, but this was not evident at or below this dye concentration. The addition of beryllium to the dye solution caused part of the dye wave to be displaced to a potential about 100 mV more negative than the main wave. When the polarographic maximum was present it was transferred to the displaced wave. The extent of the displacement was independent of the pH of the solution when between 4 and 6.The dye showed no apparent advantages over Mordant red 9 and was not studied further. Mordant yellow 8-The commercial sample of Mordant yellow 8 was soluble in water, and solutions of the dye M), 0.2 M with respect to buffer and 0-1 M with respect to potassium chloride, were polaxographed. In acetate buffer at pH 443 the dye gave a single polarographic wave (E, = -0.58 V), upon which a small polarographic maximum was superimposed. In ammonia solution buffer at pH 9.2 this maximum was much larger and, at this pH, the half-wave potential of the wave was -0078V. The addition of beryllium (2 x 10" M) to these solutions for polarography had no effect on the polarograms obtained; no displacement of the dye wave was observed.Potentiometric titrations indicated that a 1 : l beryllium-dye complex is formed in solutions with pH values between 4 and 7. It is probable that, in general, o-hydroxy-0'-carboxyazo dyes of the pyrazolone type do not give a displaced wave with beryllium. Mordant red 74-The commercial sample of Mordant red 74 was soluble in water up to a concentration of ~O-*M, and was very soluble in methanol. A stock solution of the dye (0.01 M) was prepared in a methanol - water mixture (1 + 4 v/v). The polarography of the dye was studied in buffered solutions with pH values between 4 and 7. The half-wave potential of the polarographic wave obtained with a solution at pH 5.5 was -0.41 V, and varied with pH in the usual manner.A much smaller step was obtained at -1.34 V, but this did not interfere in the study of the first wave. On the addition of beryllium, part of the dye step was displaced about 120 mV to more negative potentials. This is a slightly larger dis- placement than had been obtained with Mordant red 9 and Mordant red 60. On the addition of beryllium to the aqueous solution of the dye, the colour of the solution changed from reddish orange to yellow - orange and, under certain conditions, the complex was precipitated. To prevent this precipitation methanol was added to the polarographic solutions, and it was found subsequently that the addition of methanol up to an optimum concentration of 50 per cent. increased the height of the displaced wave.No polarographic maximum was observed with this dye; for this reason, together with the slightly larger displacement of the dye wave, the dye was chosen for a more detailed study. POTENTIOMETRIC STUDY OF MORDANT RED 74- To establish the region of pH in which the beryllium complex of Mordant red 74 is present, its formation was studied by pH titration, by using the method suggested by Irving and Rossotti5 for the determination of formation curves. Typical titration curves are shown in Fig. 1. The three solutions titrated contained the same amount of nitric acid (0.0004 M). Those used to obtain curves B and C contained the same amount of Mordant red 74 (O-OOO~M), and the solution used to obtain curve C also contained 0.0001 M beryllium.The dye used in this potentiometric study was the purified, free acidic form obtained by extracting the acidified commercial dye with ether. From the titration curves it can be calculated that the free acid is 96 per cent. pure. The carboxylic acid proton of the dye is lost when the pH is between 4 and 6, but the phenolic proton is not lost until the pH reaches 106. When the beryllium complex is formed the phenolic proton is lost, and it can be inferred that, as a further 0-5 ml of sodium hydroxide solution is required between pH 4.5 and 6.5 in the presence of beryllium, one extra proton is released per beryllium ion and a 1 : 1 beryllium - Mordant red 74 complex is formed. Above pH 6.5 hydrolysis of the beryllium occurs. To obtain a more balanced graph for measurement purposes, the beryllium concentration was increased to 0-O003~ but, unfortunately, the distance between curves B and C nowApril, 19691 0-HYDROXY-0'-CARBOXYAZO DYES AND THEIR BERYLLIUM COMPLEXES 265 increased to a value equivalent to more than 16 protons per beryllium ion.SnavelyS ex- perienced similar difficulties with the potentiometric titration of beryllium ions in solution. Thus, there must be some hydrolysis of the beryllium when its concentration approaches that of the dye. 0-OIM Sodium hydroxide solution, mi Fig. 1. pH Titration curves: curve A, nitric acid; curve B, nitric acid and Mordant red 74; and curve C, nitric acid, Mordant red 74 and beryllium nitrate DETAILED POLAROGRAPHIC STUDY OF MORDANT RED 74 AND ITS BERYLLIUM COMPLEX- The graphs in Figs.2 and 3 show the dependence of the dye and displaced dye steps on the methanol and buffer concentrations. The height of the displaced step given in Fig. 2 increases to a maximum at a concentration of about 50 per cent. v/v of methanol. As shown in Fig. 3 there is a marked increase in the height of the dye wave and a decrease in the height of the displaced dye wave with increase in buffer concentration. Because of these findings Methanol concentration, per cent. v/v Fig. 2. Effect of the concentra- tion of methanol on the height of the displaced step. Buffer concentration M, beryllium nitrate concentration 4 x l o - 4 ~ and Mordant red 74 con- centration 1 x 10-3 M : 0, displaced step; and x, free dye step 1 1 I 1 1 1 Buffer concentration, M 0.5 I .o Fig.3. Effect of the concentra- tion of acetate buffer on the height of the displaced step. Methanol con- centration 40 per cent. v/v, beryllium nitrate concentration 4 x M and Mordant red 74 concentration 1 x 10-8 M : 0, displaced step; and x , free dye step266 FOGG, KUMAR AND BURNS: POLAROGRAPHIC BEHAVIOUR OF [AHdySt, VOl. 94 subsequent polarograms were obtained with solutions 0.1 M with respect to potassium chloride and 0-1 M with respect to acetate buffer containing 50 per cent. v/v of methanol.* The height of the displaced dye step was found to be linear with beryllium concentration only in the presence of at least about a 10-fold excess of dye. It was thought that the non- linearity might arise from difficulties in measuring the displaced step height at higher beryllium concentrations, by using the Cambridge polarograph.To check this, use was made of the greater resolving power, and ultimately of the greater sensitivity, of the Sout hern-Hanvell, Mark 11, pulse polarograph. In the event, the calibration graphs were of similar nature. i I I ! I I 0.5 I .o I .o 2.0 Beryllium concentration x lo-' M Beryllium concentration x lod M Fig. 4. Calibration graph: Mor- Fig. 6. Calibration graph: Mor- dant red 74 concentration 2 x lo-' M dant red 74 concentration 1 x lo-' M The calibration graphs given in Figs. 4 and 5 were obtained from pulse polarograms. Only twice as much dye as beryllium is present in the solution of maximum beryllium con- centration represented in Fig. 4, and the calibration graph is curved. On the other hand, at least a 50-fold excess of dye over beryllium is present in the solutions used to obtain the linear calibration graph shown in Fig.5. It was found that 1 0 - 6 ~ beryllium was about the minimum concentration that could be determined, although a concentration of l o - 7 ~ could be detected. Studies of the effect of mercury head on the polarograms, obtained with solutions of Mordant red 74 and of Mordant red 74 and beryllium, have shown that the heights of both the dye and displaced dye steps are diffusion controlled. The height of both steps was proportional to the square root of the corrected mercury head. EXPERIMENTAL PROCEDURE FOR CALIBRATION GRAPHS final solution) by either conventional or pulse polarography is as follows. REAGENTS- The recoxnmended procedure for the polarography of up to 46 pg of beryllium (in the Methanol, analytical-reagent grade.Potassium chloride solution, M. Bufer solution-Prepare a solution 0.3 M with respect to sodium acetate and 0.7 M with respect to acetic acid. This solution buffers at pH 5 in aqueous solution, but the pH meter reading obtained for aqueous methanol solutions prepared as described below is about 5.5. Mordarzt red 74 solution, 065fier cent. (about 0.01 M) in 20fier cent. w/v methanoLDissolve 0.55 g of Mordant red 74 in 20 ml of methanol and dilute to 100 ml with water. This solution must be prepared fresh daily. Standard beryllium sulphate solution, 0.01 M-uSe analytical-reagent grade beryllium sulphate and make the solution 0 . 0 4 ~ with respect to nitric acid.Care should be taken when handling beryllium-containing materials, especially finely divided solids. 98 * It was found that the addition of potassium chloride improved the shape of the polarograms. The half-wave potential of the dye wave under these solution conditions was - 0-41 V o e m u the mercury pool (- 0.39 V V ~ V S U S the S.C.E.).April, 19691 0-HYDROXY-0'-CARBOXYAZO DYES AND THEIR BERYLLIUM COMPLEXES 267 PROCEDURE- Place 5 ml each of potassium chloride solution, buffer solution and Mordant red 74 solution in a 60-ml calibrated flask. Then, with a pipette, introduce into the flask an aliquot of the standard solution containing not more than 45 pg of beryllium. Add 25 ml of methanol and make the solution up to 6Oml with distilled water. Heat the solution for 5 minutes in a water-bath at 60" C to obtain complete complex formation, and then cool it to room tempera- ture.De-oxygenate the solution for at least 5 minutes and polarograph it in the usual way. For the polarography of up to 46pg of beryllium with the pulse polarograph, the procedure is as above, except that 5 ml of a 10-5 M Mordant red 74 solution is used. Typical polarograms are shown in Fig. 6. U C L 3 U K L 6 1 1 -0.2 -0.4 -0% -0.8 Applied potential, V -0.6 -0.3 Applied potential, V Fig. 6. Typical polarograms : (a), Cambridge polarogram, beryllium nitrate concentration 4 x 10-4 M and Mordant red 74 concentration 4 x M; and (b), pulse polarogram, beryllium nitrate concentration 2 x lo-& M and Mordant red 74 concentration 2 x M INTERFERENCES- Metals that form complexes with Mordant red 74 at the pH at which polarography is carried out will lower the effective concentration of the dye and will interfere if present in sufficient concentration.The use of a large excess of dye in the procedure described above is advantageous in reducing the effect of this interference. Anions, such as phosphate, which complex beryllium more strongly than Mordant red 74, must be absent. Aluminium does not give a displaced wave with Mordant red 74, and it has been shown that aluminium up to about one half of the Mordant red 74 concentration does not interfere with the beryllium wave. This corresponds to an aluminium-to-beryllium ratio of at least 5 to 1. Unfortunately, several other metals have been shown to give displaced waves with Mordant red 74.In the presence of nickel the dye wave is displaced 120mV and, in the presence of iron(III), two post waves are obtained corresponding to displacements of 210 and 320mV.268 FOGG, KUMAR AND BURNS DISCUSSION Solutions of simple beryllium salts give two polarographic waves, one at -1.4 V (versus S.C.E.) caused by the discharge of hydrogen, and a second wave at -1.8 V (versus S.C.E.) corresponding to the reduction of the hydrated beryllium ion to the metal.g The use in analysis of a polarographic wave occurring at such a high negative potential, however, is generally unsatisfactory, and complexation of the beryllium will simply displace the beryllium reduction wave to an even more negative potential. Beryllium is extractable into acetylacetone, and Dehn, Gutmann and SchoberlO have developed a polarographic method for determining beryllium based on this.They polarograph the beryllium in a 90 per cent. dimethyl sulphoxide and 10 per cent. acetylacetone solution, with tetraethylammonium perchlorate as base electrolyte and gelatin as maximum suppressor. Under these solution conditions the half-wave potential of the beryllium wave is at -1.05 V. A drop time of 8.5 seconds is necessary to avoid a polarographic maximum that is not suppressed by gelatin. The alternative to the use of the reduction wave of the hydrated or complexed metal ion is to use a displaced ligand wave. Two criteria are necessary; the ligand must be reduced at a convenient potential and the reduction of the complexed ligand must take place at a more negative potential than the reduction of the free ligand. Thus, the o,o’-dihydroxyazo dyes have proved useful in the determination of electropositive metals such as aluminium and magnesium.The procedure for the polarography of beryllium, by using o-hydroxy-0’-carboxyazo dyes, described in the present paper has the advantage that the reduction is carried out in an aqueous system. (The handling of beryllium in dimethyl sulphoxide, which has been shown to be a particularly penetrating solvent, could well be hazardous.) The extent of thedis- placement of the Mordant red 74 wave, from -0.41 to -0.53 V, although not large, is sufficient for accurate measurements to be made on both conventional and pulse polarographs. It was hoped that interferences would be few and, indeed, it has been shown that moderate amounts of aluminium do not interfere with the beryllium wave. Nickel, however, gives an identical displaced wave with Mordant red 74, and thus constitutes a serious interference. The polarographic behaviour of metal complexes of other o-hydroxy-0’-carboxyazo dyes will be investigated with a view to improving the existing procedure and to limiting the interferences still further. Hopkin & Williams Ltd. kindly supplied a sample of Mordant red 74. 1. 2. 3. 4. 6. 6. 7 . 8. 9 . 10. REFERENCES Latimer, G. W., TaZanta, 1968, 15, 1. Willard, H. H., and Dean, J. A., AnaZyt. Chem., 1960, 22, 1264. Rooney, R. C., and McIver, P. J., Analyst, 1962, 87, 895. Pyburn, C. M., Ph.D. Thesis, Loughborough, 1967. Irving, H. M., and Rossotti, H. S., J . Chem. Soc., 1963, 2904. Snavely, F. A., Ph.D. Thesis, Pennsylvania State University, 1962. Everest, D. A., “The Chemistry of Beryllium, ” Elsevier Publishing Company, Amsterdam, London and New York, 1964. Tepper, L. B., Hardy, H. L., and Chamberlin, R. I., “Toxicity of Beryllium Compounds,” Elsevier Publishing Company, Amsterdam, London and New York, 1961. Kemula, W., and Michalski, M., CoZZ. Czech. Chem. Commun., 1933, 5, 436. Dehn, H., Gutmann, V., and Schober, G., Mh. Chem., 1962, 93, 877. Received August 26tk, 1968 Accepted October 25th, 1968

ISSN:0003-2654    

DOI:10.1039/AN9699400262

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

Analyst, April, 1969, Vol. 94, pp. 269-274 268 Polarographic Determination of Calcium and Magnesium BY B. FLEET, SOE WIN AND T. S. WEST (Chemistry Department, Imperial College of Science and Technology, London, S . W.7) A method for the determination of calcium and magnesium based on the decrease in the height of the anodic polarographic waves of ethylene- diaminetetra-acetic acid (EDTA) and ethyleneglycol-bis- @-aminoethylether) - NNN’N’-tetra-acetic acid is proposed. The sum of calcium and magnesium is determined from the height of the anodic wave of EDTA and subsequently calcium is determined on a separate aliquot from the decrease in the anodic EDTA wave. SEVERAL studies of the polarographic behaviour of metal-ion - ethylenediaminetetra-acetic acid (EDTA) systems have been carried out, the aim being, in most instances, to establish the optimum conditions for amperometric titration.132 In the simplest example in which the sought metal ion gives a well defined reduction wave, the decrease in the height of the diffusion current for the metal ion is followed as the standard ligand is added.For metal ions that are not normally electroactive, it is possible to follow the titration by the addition of an electroactive indicator ion, which forms a weaker complex than the sought metal ion. In this case the end-point is indicated by a sudden decrease in the cathodic current caused by the reduction of the indicator ion. It is also possible to follow the course of the titration by monitoring the anodic wave of the complexone. EDTA gives an anodic wave at the dropping-mercury electrode, the electrode reaction at pH 10, where EDTA exists mainly in the form HYS-, being- HY3- + Hgo -+ HgY3- + H+ + 2e-.In the present study this principle has been applied to the polarographic determination of calcium and magnesium although it is apparent that by a critical selection of masking agent - buffer conditions a wide range of metal ions could be determined. Ethyleneglycol-bis-(~-aminoethylether)-NNN’N’-tetra-acetic acid (EGTA)3 has recently been used for the selective titration of calcium in the presence of magnesium.436 Although there is a large difference in the stability constants (log KCaY = 11.0, log K, = 5.2) the end-point in the visual procedure is not distinct. This situation can be improved by the addition of an auxiliary masking agent for magnesium6 or by using an end-point detection technique that does not rely on measurements made near the end-p0int.*3~3~3~ Potentio- metric indication with a mercury - mercury-EGTA electrode was used by Schmid and ReilleyP; amperometric indication738 has also been used.In both procedures a J-type mercury electrode was used. The aim of the present method was to determine calcium in the presence of magnesium by measuring the decrease in the height of the anodic wave of EGTA as a result of complex formation. The sum of calcium and magnesium is then determined on a separate sample from the decrease in the anodic wave of EDTA and the magnesium content obtained by difference . EXPERIMENTAL REAGENTS- All reagents are of analytical-reagent grade unless otherwise stated.EDTA, 10-2 M stock solzction-A 3-723-g sample of EDTA disodium salt dihydrate was dissolved in distilled water and diluted to 1 litre. EGTA , M stock solzltion-A 3-8-g sample of ethyleneglycol-bis-( /3-aminoethylether)- NNN’N‘-tetra-acetic acid (Fisons Ltd.) was dissolved in 10 ml of 2 N sodium hydroxide and diluted to 1 litre with distilled water. Bufler, 3-6 M ammonia solution - 1.2 M ammonium flitrate-Two-hundred millilitres of 0.88 ammonia solution (about 35 per cent.) and 96.1 g of ammonium nitrate were diluted to 1 litre. 0 SAC and the authors,270 FLEET, SOE WIN AND WEST : POLAROGRAPHIC Triton X-100 soiution, 0-01 per ceqat. Pentasodium tri$kos$kate solzction , M-A M sodium tripolyphosphate , Na5P301, , analytical-reagent grade solution was prepared, assuming a 100 per cent.Na,P,O,, content, by dissolving 3.68 g in 1 litre of distilled water. Synthetic samples of calcium and magnesium were prepared from calcium carbonate and magnesium sulphate heptahydrate, respectively. APPARATUS- Polarograms were measured on a HeyrovskJi, type LP55A, polarograph. The photo- graphic recording galvanometer was omitted and the potential drop across a 4-7-kQ resistor was displayed on a Servoscribe recorder (Type RE511). The response characteristics of this recorder are suitable for polarography. Sufficient damping can be achieved with a 2000-pF capacitor connected in parallel with the input. Polarograms were measured in a 10-ml capacity Kalousek cell against a saturated calomel reference electrode connected to the sample compartment by means of a salt bridge.Capillary characteristics were as follows: h = 60 cm, t (at OV versus S.C.E.) = 4*ls, and m (at OV verszls S.C.E.) = 1.965 mg s-1. [Analyst, VOl. 94 PROCEDURE- were prepared in 50-ml calibrated flasks. Analysis for the sum of calcium and magnesium-Solutions with the following composition Sample calcium + magnesium . . 0.1 to 0.8 mM EDTA .. .. .. . . 1.0 mM Buffer .. .. . . 0.18 M ammonia solution - 0-06 M ammonium nitrate Triton X-100 . . .. . . 0.002 per cent. Polarographic measurements were then made at zero applied volts against S.C.E. on 10-ml aliquots of the above solutions, and the concentrations were read off from a calibration graph constructed under identical conditions by using calcium and magnesium standards.Analysis for calcium in the presence of magnesium-Solutions with the following com- position were prepared in 50-ml calibrated flasks. Calcium . . . . 0.1 to 0-6 mM Magnesium . . . . 0.0 to 2-0 mM Sample { EGTA .. .. .. . . 1.0 mM Buffer . . .. .. . . 0.18 M ammonia solution - 0.06 M ammonium nitrate Pentasodium triphosphate . . 2.5 x 10-8 M Polarographic measurements were then made at zero applied volts against S.C.E. on IO-ml aliquots of the above solutions, and the concentrations were read off from a calibration graph constructed under identical conditions by using calcium standards in the absence of magnesium. For sample concentrations differing much from the above, suitable adjustments of the concentrations of the complexones, masking agent, etc., can be made bearing in mind that a 50 per cent.or more decrease in diffusion current gives the most accurate results. RESULTS AND DISCUSSION POLAROGRAPHIC BEHAVIOUR OF EDTA AND EGTA- Both EDTA and EGTA give anodic waves at the dropping-mercury electrode (Figs. 1 and 2), with the wave of EGTA not being as well defined as the EDTA wave. To establish the best conditions for the measurement of the wave the effects of the composition and concentration of the buffered supporting electrolyte were studied. The choice of the buffer composition is critical as the anodic limit of the dropping-mercury electrode is governed by mercury complex formation with the buffer component, e.g., C1- or NH,. The dependence of the half-wave potential for EDTA is evident from the electrode reaction, a shift of E, to more negative potentials with increasing pH being observed.Thus it is possible to shift the wave of the complexone away from the anodic wave of the buffer by carrying out the measurement at a high pH value. In the ammonia solution - ammonium nitrate buffer the anodic limit of the dropping-mercury electrode was +0*1 V. This wasApril, 19691 DETERMINATION OF CALCIUM AND MAGNESIUM 271 sufficient to allow the EDTA wave to be completely developed. It can be seen from Fig. 1: that the limiting current plateau for EDTA occurs at zero volts vmus the normal calomel electrode (N.C.E.). This fact is of importance for the development of a continuous procedure and also for an amperometric titration procedure. Potential Fig.1. Dependence of anodic wave of EDTA on the sum of calcium and magnesium concentra- tion. 1, 1 mM EDTA; 2 to 6, 0-2, 0.4, 0.6, 0.8 and 1.0 mM added calciumpZus magnesium in 0 . 1 8 ~ ammonia solution - 0.06 M ammonium nitrate with 0.002 per cent. Triton X-100. Starting potential: -0.3 V; f.s.d. &SPA. Potent i ai Fig. 2. Dependence of anodic wave of EGTA on calcium concentration. 1, 1 mM EGTA; 2 to 7 0-1, 0.2, 0.3, 0.4, 0.6 and 1.0 mM calcium in the presence of 2.0 mM magnesium 0.18 M in ammonia solution - 0.06 M ammonium nitrate buffer with 2.6 mM pentasodium triphosphate added. Starting potential: -0-26 V; f.s.d. 843pA. A maximum, which was observed on the EDTA wave at de-polariser concentrations above 1 x lo4 M and also when the buffer concentration was below a certain critical level (0-18 M ammonia solution, 0.06 M ammonium nitrate), was effectively suppressed by the addition of 0-002 per cent. of Triton X-100.The decrease in the anodic wave of EDTA with the addition of calcium and magnesium is shown in Fig. 1. The dependence of the decrease in anodic current (id - i) on the sum of calcium and magnesium concentration is rectilinear (Fig. 3). Concentration, m M Fig. 3. Dependence of the decrease in anodic current (id-i) of EDTA and EGTA on calcium plus magnesium concentrations. 0 Anodic current decrease of EDTA with added equimolar calcium plus magnesium 0 Anodic current decrease of EGTA with addition of calcium. Theoretical curve, i.e., magnesium absent Anodic current decrease of EGTA with addition of calcium in the presence of 2-6 m M pentasodium triphosphate and 2.0 mM magnesium. i d for EDTA a t 1 mM = 6.7 PA, for EGTA a t 1 mM = 4-24 pA272 The anodic polarographic behaviour of EGTA is similar to EDTA except that the wave is less well defined and the diffusion current constant I = - id is about 15 per cent.lower. c . ma 13. t ) Despite the low value of the stability constant of the magnesium - EGTA it was found that the presence of magnesium ions caused a distortion in the anodic wave of EGTA. This was caused by a shift of the wave to more positive potentials when it merged into the large anodic current because of mercury - ammine complex formation. This shift is analogous to the shift in the E+ value of the cathodic reduction of a metal ion to more negative values on complex formation.To prevent this interference several possible masking agents, which would prevent magnesium but not calcium from reacting with EGTA, i.e., log KMgL > 5-2 and log KcaL < 11.0, were examined. Of some thirty ligands selected on the basis of their known stability constants (including conventional metallochromic indicators), three were found to be suitable, pentasodium tri- phosphate, Tiron and Eriochrome black T. The relative masking efficiencies of these ligands are shown in Fig. 4. The observed current is plotted against concentration of masking agent. FLEET, SOE WIN AND WEST : POLAROGRAPHIC [Analyst, Vol. 94 I I I I I 1 O '0 2 4 6 8 10 Masking agent concentration, mM Fig. 4. Relative masking efficiency for mag- nesium; anodic current for 0 - 6 m EGTA with 0.2 mM added magnesium.Pentasodium triphosphate, 9 Eriochrome black T, 0 Tiron and - - - theoretical 100 per cent. efficient masking agent It can be seen that pentasodium triphosphate is the most effective and requires an excess of only about 25 per cent. over the theoretical amount necessary to react with magnesium before masking is complete. This is shown by the coincidence of the observed limiting current for a solution of 0 . 6 m ~ EGTA, 0 . 2 m ~ with respect to magnesium, with the theoretical value 01 I r 1 I J Masking agent concentration, mM 0 2 4 6 8 10 Fig. 6. Effect of masking agents (for mag- nesium) on the reaction between calcium and EGTA. 0.6 m EGTA with 0 4 m ~ calcium added. --- Anodic wave heightf or EGTA alone, pentasodium triphosphate 8 Eriochrome Black T 0 TironApril, 19691 DETERMINATION OF CALCIUM AND MAGNESIUM 273 at a concentration of 0.25 mM pentasodium triphosphate.No interference with the definition of the anodic wave of EGTA was observed for concentrations of masking agent up to M. The effect of the added masking agent on the reaction between calcium and EGTA was studied. A theoretical decrease in anodic current was observed (Fig. 5), showing that the masking agents selected had no influence on the reaction. A rectilinear dependence of the decrease in anodic current of EGTA as a function of calcium concentration was observed (Fig. 3). The theoretical graph obtained by measuring the decrease caused by calcium in the absence of magnesium and masking agent (open circle) is almost coincident with the calibration graph obtained in the presence of 2 - 5 m ~ penta- sodium triphosphate and 2.0 mM magnesium (full point).By using the two calibration graphs established previously, the analysis of a range of synthetic samples was carried out. The results obtained over a range of calcium-to-magnesium ratios are shown in Tables I and 11. As most natural waters contain calcium-to-magnesium ratios between 10 : 1 and 10 : 3, with a total metal-ion concentration of the order of lO-3~, the recommended procedure is directly applicable to water analysis. TABLE I DETERMINATION OF THE SUM OF CALCIUM AND MAGNESIUM Present, m M 0.10 0.10 0.10 0.50 0.50 0.50 0.80 0-80 0.80 Found, m M 0.08 0.08 0.08 0.49 0.49 0.50 0-79 0.79 0.79 Deviation, per cent.- 20.0 - 20.0 - 20.0 - 2.0 - 2.0 0 - 1.2 - 1.2 - 1.2 TABLE I1 DETERMINATION OF CALCIUM IN THE PRESENCE OF MAGNESIUM (a) Magnesium present, m M 0.50 0.50 0-50 1.00 1.00 1 -00 2.00 2.00 2-00 0-50 0.50 0.50 1.00 1.00 1.00 2.00 2.00 2.00 0-50 0.50 0.50 1.00 1 -00 1.00 2.00 Calcium present, m M 0-50 0-50 0-50 0.50 0-50 0.50 0.50 0.50 0.50 0.20 0-20 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.10 0.10 0.10 0.10 0.10 0.10 0.10 Calcium found, m M 0.48 0.50 0.50 0.50 0.51 0.50 0.5 1 0.50 0.52 0.23 0.23 0-23 0.2 1 0.22 0.20 0.24 0.23 0.23 0.10 0-11 0.1 1 0.11 0.11 0.09 - Deviation, per cent. - 4.0 0 0 0 - 2.0 0 + 2.0 0 + 4.0 + 15.0 + 15.0 + 15.0 + 5.0 + 10.0 0 + 20.0 + 15.0 + 15.0 0 + 10.0 + 10.0 + 10.0 + 10.0 - 10.0 -274 FLEET, SOE WIN AND WEST INTERFERENCES- At the high pH value used most of the potentially interfering metal ions are precipitated as hydroxides.In addition the pentasodium triphosphate will mask barium, strontium and manganese( 11). Although triethanolamine can be used, strong mercury complex-forming masking agents, such as CN-, cannot be tolerated. Copper, zinc, cobalt and cadmium cause interference and must be removed. However, the present work is aimed at the determinations of calcium and magnesium in natural waters and possibly biological fluids with which there is no interference from these elements. 1. 2. 3. 4. 5. 6. 7. 8. 9. REFERENCES Reilley, C. N., Scribner, W. G., and Temple, C., Analyt. Chem., 1956, 28, 450. Stock, J. T., “Amperometric Titrations, ” Interscience Publishers, a division of John Wiley & Schwarzenbach, G. , “Die Komplexometrische Titration, ” Second Edition , Ferdinand Enke, Stutt- Schmid, R. W., and Reilley, C . N., Analyt. Chem., 1957, 29, 264. Ringbom, A., Pensar, G., and Wanninen, E., Analytica Chim. Acta, 1958, 19, 525. Fleet, B., Soe Win, and West, T. S., Talanta, 1968, 15, 333. Monnier, D., and Roueche, A., Helv. Chim. Acta, 1966, 35, 231. Monnier, D., Delphin, G., and Haerdi, W., Analytica Chim. Acta, 1964, 47, 103. Jordan, J., and Alleman, T. G., Analyt. Chem., 1957, 29, 9. Sons Inc., New York, London and Sydney, 1965. gart, 1965. First received August 21st, 1968 Revised December 4th, 1968 Accepted December 19th, 1968

ISSN:0003-2654    

DOI:10.1039/AN9699400269

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

Analyst, April, 1969, Vol. 94, pp. 275-283 275 A General Method for the Determination of Organophosphorus Pesticide Residues in River Waters and Effluents by Gas, Thin-layer and Gel Chromatography BY J. ASKEW, J. H. RUZICKA AND B. B. WHEALS (Ministry of Technology, Laboratory of the Government Chemist, Cornwall House, Stamford Street, London, S.E. 1) A general, comprehensive scheme for the extraction of organophosphorus pesticides from river waters and sewage effluents is described. The pesticides, after extraction with chloroform, are determined by gas and thin-layer chromatography. The procedure includes details of an improvement that enables all of the pesticides to be detected on thin-layer chromatoplates with a phosphorus-specific ammonium molybdate spray. The use of gel chromato- graphy on columns of Sephadex LH 20 as an adjunct to identification is also described.ORGANOPHOSPHORUS pesticides are finding increasing use in agriculture as the trend to supplement and replace organochlorine pesticides continues. At present about forty organo- phosphorus pesticides have satisfied the conditions laid down in the Ministry of Agriculture’s “Pesticide Safety Precautions Scheme,” and many of them are already in substantial use in this country. These pesticides can theoretically be anticipated to give rise to sixty, or more, oxidation products, but if such oxidation occurs it will probably be some time after applica- tion, and it is improbable that these metabolites will be responsible for any major water pollution problems. The analyst attempting to detect and characterise residues of these pesticides and their metabolites in samples of unknown history, as he is frequently required to do in this laboratory, is confronted with a most complex task.Only by the multiple use of chromatographic techniques is it possible to accomplish a general analysis of this type successfully. The pollution of water by pesticides as a result of intentional application, carelessness, accident or natural leaching from sprayed areas has attracted much attention, but in the past this has primarily been concerned with the organochlorine insecticides and herbicides.1 We have shown in a recent study of hydrolysis rates that the organophosphorus pesticides, although undoubtedly much more labile than their organochlorine counterparts, decay in river water at rates sufficiently slow to constitute a potential pollution hazard.2 It seems probable that river authorities will increasingly be called upon to undertake analyses for organophosphorus pesticides.This paper describes the separation and identification of some of these compounds in samples of river waters and sewage effluents by the use of gas, thin-layer and gel chromatography. EXPERIMENTAL EXTWCTION- The organophosphorus pesticides vary greatly in their polarity, and the extent of their extraction from aqueous samples is markedly dependent on the nature of the solvent used. Table I gives a comparison of the efficacies of three solvents in removing these pesticides from water, and it is apparent that a polar solvent such as chloroform is the most generally useful.Menazon is structurally unique among the pesticides examined, in that it possesses primary amino groups. These confer a property that may be used to advantage in the selective removal of menazon from chloroform solution with dilute acid. 0 SAC; Crown Copyright Reserved.76 ASKEW et al. : A GENERAL METHOD FOR THE DETERMINATION OF [Analyst, Vol. 94 CLEAN-UP- In our experience most water extracts are found to be sufficiently low in co-extractives to require no clean-up for gas-chromatographic purposes. If difficulties are encountered at the thin-layer stage, e.g., charring of co-extractives resulting in obscuration of the molybdenum- blue colour that characterises an organophosphorus compound, then a preliminary clean-up will be necessary.The nature of the co-extractives will undoubtedly vary with the location from which the samples are taken, and no completely comprehensive clean-up procedure can be recommended. In our studies, however, a column containing 1 g of Nuchar carbon eluted with chloroform gave the most useful clean-up. The following pesticides are exceptional in being retained by the column described : azinphos-ethyl, azinphos-methyl, coumaphos, dichlorvos, haloxon, menazon, phosalone and vamidothion. When these pesticides are en- countered, a clean-up on alumina3 or magnesium oxide4 would be preferable. TABLE I EXTRACTION OF PESTICIDES FROM AQUEOUS SOLUTION BY ORGANIC SOLVENTS Results obtained by extracting 1 litre of a 0.01 p.p.m. pesticide solution with 20ml of solvent.Determination made by gas chromatography Pesticide extracted, per cent. Pesticide Chlorfenvinphos Demeton-S-methyl Dimethoate . . Pyrimithate . . r 1 Hexane Benzene Chloroform .. .. 92 95 87 .. .. 0 76 98 .. .. 0 0 41 .. .. 92 93 90 GAS CHROMATOGRAPHY- Many of the organophosphorus pesticides can be separated and quantitatively deter- mined by gas chromat~graphy.~ Some of these pesticides have strong electron-capturing properties and, provided samples do not contain appreciable amounts of electron-capturing co-extractives, this property affords a sensitive means of detection.s More useful are the recently developed “phosphorus-specific” detectors, which are highly sensitive and relatively s ~ e c i f i c . ~ ~ ~ * ~ Fig. 1 gives a comparison of the responses of an electron-capture detector and a phosphorus detector (flame thermionic) to a river-water extract before and after fortification with pesticides, an Apiezon L column being used for the separation. Note the enhanced response for the chlorine-containing pesticide, chlorfenvinphos, on electron capture and the poor response for demeton-S-methyl. The variation in the relative response of the electron- capture detector for organophosphorus compounds is a disadvantage for general screening and has been described elsewhere.lOJ1 The interpretation of gas chromatograms obtained when using a phosphorus-specific detector of the Hartmann’ type must be made with caution.A chromatographic peak may often be caused by the presence of relatively large amounts of compounds that do not contain phosphorus ; volatile nitrogen-containing compounds, in particular, can be detected at the microgram level. Additional confirmation of a characterisation needs to be made with another technique, such as the thin-layer chromatographic identification described below, The pesticides amenable to gas chromatography can be separated on a variety of stationary phases; the retention times shown in Table I1 are based on columns that are commonly used for separating organochlorine pesticides.12 The retention times vary greatly, and it is necessary to use either temperature programming or isothermal funs at more than one temperature for successful screening. The technique can be rendered quantitative down to sub-nanogram levels, and the comparison of relative retention times on two or more columns enables a tentative pesticide characterisation to be made.THIN-LAYER CHROMATOGRAPHY- In the general screening method of Bates4 a paper-chromatographic separation was used to differentiate between the various pesticides present, but the technique is slow, and a muchApd, 19691 ORGANOPHOSPHORUS PESTICIDE RESIDUES IN RIVER WATERS 277 0 Time, minutes (11) Fig. 1. Comparison of gas chromatograms ob- tained with (I) a phosphorus detector and (11) an electron-capture detector : (a), an extract equivalent to 1 litre of Thames water; and (b), the same extract fortified with pesticides. Peaks correspond to 6 ng of each compound : 1, demeton-S-methyl; 2, dimetho- ate; 3, pyrimithate; 4, parathion; and 5, chlorfenvin- phos more rapid procedure based on thin-layer chromatography has recently been described.13 The merits of thin-layer chromatography in the separation of organophosphorus pesticide residues are reviewed e1~ewhere.l~~~~ (i) Hexane - acetone (5 + 1).(ii) Chloroform - acetone (9 + 1). (iii) Chloroform - acetic acid (9 + 1). The first of these systems is generally useful with the RF values of the forty pesticides spreading over the range 0 to 0.9, whereas solvent (ii) is intended for the resolution of those pesticides with lower RF values in solvent (i). The third solvent, which is particularly useful for moving menazon from the base-line, is applicable to the same group of pesticides that show a separation in solvent (ii). The observed RF values are shown in Table 11.Chromogenic reagertts-Many spray reagents have been recommended for the detection of organophosphorus pesticides after separation on thin-layer plates, and are discussed by Watts,le Abbott and Egan14 and Ragab,l7 Most of the chromogenic reagents used are non- specific and of limited usefulness for screening applications. Because of their limitations, efforts were made to improve the phosphorus-specific ammonium molybdate spray reagent. Solvent systems-Three solvent systems have been investigated.278 ASKEW et al. : A GENERAL METHOD FOR THE DETERMINATION OF [Amzlyst, Vol. 94 In 1949, Hanes and Ishenvoodl* developed a chromogenic reagent containing ammonium molybdate in a perchloric acid - hydrochloric acid mixture, with which they were able to detect certain phosphoric acid esters on paper chromatograms.The reaction involves acid hydrolysis to orthophosphate, followed by formation of molybdophosphoric acid and reduction TABLE I1 USAGE AND CHROMATOGRAPHIC RESULTS OF ORGANOPHOSPHORUS PESTICIDES Gas-chromatographic Thin-layer chromatographic r r-- chromato- results results Gel- Usage Azinphos-ethyl . . A, H Azinphos-methyl . . A, H Bromophos .. .. v Carbophenothion .. V Chlorfenvinphos . . A, H, V Coumaphos .. .. v Crufomate . . .. v Demeton-S . . .. A,H Demeton-S-methyl . . A, H Diazinon . . . . A, H, F, V Dibrom . . .. A,H Dichlofenthion . . V Dichlorvos .. . . A, H, F, V Dimefox .. .. A,H Dimethoate . . .. A,H Disulfoton . . .. A,H Ethion . . . . A, H Ethoate-methyl . . A, H Fenchlorphos Fenitrothion .. F Formothion .. .. A,H Haloxon . . .. v Malathion . . . . A, H, F, V Mecarbam . . .. A,H Menazon .. .. A,H Mevinphos . . .. A,H Morphothion . . .. A,H Oxydemeton-methyl A, H Parathion . . .. A,H Phenkapton . . .. A,H Phorate . . .. A,H Phosalone . . .. A,H Phosphamidon .. A, H Pyrimithate . . .. A,H Schradan . . .. A,H Sulfotep . . .. A,H TePP .. .. A,H Thionazin . . .. A,H Trichlorphon .. A, H, V Vamidothion .. A,H . . A, H, F, V Apiezon L 995* 840* 135 385* 129 N.D. N.D. 19 22 38 26 66 3$ 1: 43 47 224* 47 86 81 62 N.D. 66 117 N.D. 8,10$ 212* N.D. 100 640* 31 730* 41, 55 78 78 19 N.D. 21 N.D. 2: S.E. 30 970* 870* 102 2702 140 N.D. 164 21 33 41 6 58 4: 2: 65 45 220* 68 70 88 82 N.D. 85 127 N.D. 13, 16$ 285* N.D. 100 420* 29 600* 67, 91 79 130 26 N.D. 25 N.D. 4': soivent solvent X.E.60 (i) (ii) 870* 0.33 0.90 870* 0.19 0-88 63 0.85 0-93 187* 0.83 0.96 102 0.24 0.79 N.D. 0.33 0.90 141 0.06 0.43 13 0.33 0.93 27 0.17 0.73 18 0.61 0-95 25 0 to 0.22t 0 to 0-89t 29 0-77 0.96 3-5t 0.22 to 0.27 0.73 2: 95 26 156* 95 40 95 129 N.D. 75 116 N.D. 12, 15$ 356* N.D. 100 290* 18 685* 73, 110 42 74 18 26 16 N.D. N.D.S 0.08 0.05 0.82 0-77 0-07 0.84 0.49 0.15 0.04 0.37 0.42 0 0.10 0.06 0 0.57 0.74 0.80 0.39 0.04 0.62 0 0-75 0 0.45 0-03 0.01 0.44 0-37 0.97 0.97 0.61 0.93 0.91 0.75 0-71 0.95 0.95 0.02 0.64 0-49 0.05 0.91 0.97 0.97 0.97 0-34 0.96 0.02 0.92 0 to 0.50t 0.92 0.18 0.16 - 0.60 0.16 0.03, 0.62 0.61 0-30 - - - graphic results 113 133 100 98 77 99 73 79 86 77 86 89 83 71 95 87 82 87 101 114 112 86 85 79 106 79 100 104 89 104 72 81 59 75 73 89 80 73 - - Usage: A denotes agricultural: V veterinary; F food storage; and H horticultural.Gas-chromatographic results : the values shown are retention times relative to that of parathion = 100. The columns used contained the following stationary phases: (i) Apiezon L, 2 per cent. and Epikote 1001, 0.2 per cent.; (ii) S.E. 30,4 per cent. and Epikote 1001, 0-4 per cent.: and (iii) X.E. 60, 2 per cent. and Epikote 1001, 0.2 per cent., coated on acid-washed, dimethyl- dichlorosilane-treated, 80 to 100-mesh Chromosorb G. All columns were 150 cm in length with 0-3 cm 0.d. Retention times were determined at 195" C, except where marked * = 220" C and $ = 150" C. The retention times of parathion on the three columns were (i) 220°C 1.70 minutes, 196°C 4-0 minutes and 150°C 19 minutes; (ii) 220°C 1.80 minutes, 195°C 4.50 minutes and 15OOC 22 minutes; and (iii) 220" C 1.60 minutes, 195" C 4.25 minutes and 150" C 21 minutes.N.D. denotes not detected. Thin-layer chromatographic results: the values shown are the RE. values in the solvent (i) hexane - acetone (5 + l ) , (ii) chloroform - acetone (9 + 1) and (iii) chloroform - acetic acid (9 + 1). t Denotes streaking. Gel-chromatographic results: the values shown are the elution volumes relative to that of parathion = 100 when eluting with ethanol from a Sephadex LH 20 column.April, 19691 ORGANOPHOSPHORUS PESTICIDE RESIDUES IN RIVER WATERS 279 of the latter to molybdenum blue. This reagent has been applied to pesticide analysis with only limited success, as the compounds vary widely in their ease of hydrolysis; Bunyan19 has reported the difficulty of hydrolysing pesticides on thin-layer plates with this reagent.Barney20 advocated the use of a preliminary spray containing a hydriodic acid - acetic acid mixture, which Saliman21 had previously shown to be effective in hydrolysing organophos- phorus pesticides. By use of this spray and heating on a hot-plate to 250" C, it is possible to break down many pesticides. In our hands the procedure was not always successful, and we attributed this to losses caused by volatilisation at this high temperature and to the dry nature of the environment produced being unfavourable to hydrolysis. We have modified the procedure advocated by Barney20 to give a method capable of detecting all forty pesticides.By the simple expedient of clipping a second glass plate over the sprayed plate before heating, it is possible to hydrolyse the pesticides in an oven at the lower temperature of 180" C. We ascribe the success of this modification to the moist environ- ment maintained between the plates during the heating stage. Benzidine, the reductant used by Barney, is reputed to be carcinogenic, and we have replaced this by tin(I1) chloride, which was found to be more effective than ascorbic acid, hydrazinium sulphate or hydro- quinone. By using the procedure described below we have been able to detect 1 pg or less of all of the pesticides listed, and analogous sensitivities have been obtained with those oxidation products examined. Trimethyl, triethyl and tributyl phosphates can be similarly detected, but triphenyl and tricresyl phosphates do not respond, presumably because of the greater resistance to hydrolysis of the latter compounds.GEL CHROMATOGRAPHY- In a recent publication we have described the separation of organophosphorus pesticides by gel chromatography on columns containing Sephadex LH20.22 This technique can provide a useful adjunct to the other chromatographic techniques if used to separate pesticides into groups based on their elution sequence, and has the merit of effecting separations by using completely different principles. A useful feature of gel chromatography is that a column can be used repeatedly over long periods of time without any detectable change in the elution volume of pesticides and, in almost all instances, with quantitative recovery of the compounds. Only menazon, which is absorbed strongly, and formothion, which breaks down in ethanol solution on heating, cannot be successfully recovered by gel chromatography. When the presence of menazon is suspected in a sample a preliminary acidic extraction is advised.The elution volumes of the various pesticides related to that of parathion are shown in Table 11. On a column of 75-ml bed volume, parathion has an elution volume of about 75 ml and, on first setting up a Sephadex column, the elution volume of parathion should be checked by collecting suitable fractions and examining them by gas or thin-layer chromato- graphy. The fractions collected by using the analytical scheme shown below are based on the elution of parathion at 75 ml.Band spread of the eluting peaks naturally occurs; the spread is gaussian and usually extends 3 to 4 ml each side of the maximum. On the same type of column, organochlorine pesticides elute over the range 65 to 85ml. ANALYTICAL SCHEME Stage (a) is applied as a general screening technique, and stage (b) is incorporated only Stage (a)- when suspected pesticides are encountered. Sample Chloroform extraction* Column clean-up (only if co-extractives required removal) Chloroform removed and replaced by ethanol or acetone Gas-chromatographic examination (see Table 11) Thin-layer chromatography [gel plates and solvent (i)] I I I I I * If menazon is suspected remove by acidic extraction at this stage.280 ASKEW et al. : A GENERAL METHOD FOR THE DETERMINATION OF [Analyst, Vol.94 Stage (b)-Separate the extract before gas or thin-layer chromatography as follows- Ethanol solution Separation on Sephadex LH 20 (column of 75-mi bed volume) I I Fraction 1 (Elution volume 40 to 60 ml, RV 50 to 80) Chlorfenvinphos Crufomate Demeton-S Diazinon Dimefox Mevinphos Oxydemeton-meth yl Phosphamidon Schradan Sulfotep Vamidothion also Trimethyl phosphate Triethyl phosphate Tributyl phosphate TePP. Fractibn 2 (Elution volume 61 to 70 ml, RV 81 to 93) Demeton-O-meth yl Demeton-S-methyl Dibrom Dichlof enthion Dichlorvos Disulfoton Ethion Ethoate-methyl Malathion Mecarbam Phorate Pyrimithate Thionazin Trichlophon Fraction 3 (Elution volume 71 to 110 ml RV 94 to 145) Azinphos-ethyl Azinphos-meth yl Bromophos Carbophenothion Coumaphos Dimethoate Fenchlorphos Fenitrothion Haloxon Morphothion Parathion Phenkapton Phosalone Examine by gas - liquid and thin-layer chromatography. METHOD APPARATUS- Gas-chromatographic equipment-A single or dual column instrument, fitted with a phosphorus-specific detector (Varian Aerograph, Model 205-B) , was used.Thin-layer chromatographic equipment-Suitable for the preparation of 250 p thick, silica-gel chromatoplates, 20 x 20 cm. Ovens-Two are required, one at 120' C and the other at'180' C. Development chamber-A glass tank, 22 x 21 x 9 cm, with well fitting lid. Kuderna-Danish evaporator-This is fitted with a 10-ml pear-shaped flask. Glass chromatographic column, 2.5 x 30 cm. Thin-layer sfiray equipment-A Shandon laboratory spray gun was used. REAGENTS- All reagents should be of recognised analytical grade whenever possible.Chloroform. Ethanol. Ammonia solution, sp.gr. 0.88. Hydrochzloric acid, N. Sodium hydroxide solution, 2 N. Silica gel &For thin-layer chromatography (obtainable from E. Merck Ltd., Darm- stadt). Nuchar carbon-A grade with particles of about 1-mm diameter was used (obtainable from Eastman Kodak Ltd.). Sodium sulphate, granular, anhydrous. Hydriodic acid spray-Mix 25 ml of hydriodic acid (sp.gr. 107)~ 25 ml of glacial acetic acid and 50 ml of water. The solution is stable for several weeks. Ammonium molybdate sfiray-Dissolve 2 g of ammonium molybdate, NH4Mo,O2*.4H20, in 20 ml of water - concentrated hydrochloric acid (1 + 1) with gentle heating, and adjust volume to 1 O O m l with water. The solution is stable for several weeks.Tin(l1) chloride sfway-Dissolve 1 g of tin(I1) chloride, SnC12.2H,0 , by heating with 10 ml of concentrated hydrochloric acid. Add 40 ml of water and 50 ml of acetone. Prepare freshly each day.April, 19691 ORGANOPHOSPHORUS PESTICIDE RESIDUES IN RIVER WATERS 281 EXTRACTION- Extract 1 litre of water or sewage effluent with three portions each of 50 ml of chloroform (centrifuge the emulsified portion of the sample if stable emulsions form; the extent of emulsion formation can often be reduced by dissolving about 20g of sodium sulphate in the sample solution before extraction). Dry the combined chloroform extracts by passing through a 10 x 1-cm column containing granular anhydrous sodium sulphate, and collect the eluate in a Kuderna-Danish evaporator fitted with a 10-ml pear-shaped flask.Wash the column with a further 25 ml of chloroform and evaporate the solution to small volume on a steam-bath. When menazon is suspected to be present in a sample proceed as follows. Shake the combined chloroform extracts with two portions of 25ml of N hydrochloric acid, retaining both phases. Wash the acidic extract with 25 ml of chloroform adding this to the previously extracted chloroform. Add 30 ml of 2 N sodium hydroxide to the acidic solution and back- extract the menazon by shaking with two portions of 25 ml of chloroform. Dry the chloro- form solutions by using the procedure described above. The extract containing menazon should be examined by thin-layer chromatography; proceed as follows with the other solution. CLEAN-UP- This is necessary only when co-extractives interfere in the subsequent chromatographic determinations.De-gas 1 g of Nuchar carbon by evacuating in the presence of about 25 ml of chloroform, and add the slurry to a 1 cm diameter x 30cm long column containing about 0-5g of Celite 545 (the latter prevents charcoal fines from eluting from the column). Place the chloroform sample solution (1 ml) on the column and elute with 100 ml of chloroform, collecting the eluate in a Kuderna-Danish evaporator. Evaporate to small volume on a steam-bath. GAS CHROMATOGRAPHY- Remove all traces of chloroform from the sample solution by subjecting it to a gentle stream of air or nitrogen until dry. To prevent losses by evaporation of volatile pesticides, such as dichlorvos, dimefox and mevinphos, remove the sample from the air stream im- mediately after the last traces of solvent have evaporated; alternatively, use a micro Snyder column,23 evaporating successively with acetone or ethanol to remove final traces of chloro- form.Add 1 ml of ethanol or acetone to the residue, swirl it to effect dissolution and inject 5 pl aliquots on to suitable gas-chromatographic columns. Use an injection liner (e.g., Pyrex glass) if uncleaned extracts are injected, to minimise column contaminations. Determine the elution time of any chromatographic peaks relative to a suitable standard pesticide, such as parathion, and compare the results with those previously obtained with pesticide standards (see Table 11). THIN-LAYER CHROMATOGRAPHY- Prepare 250 p thick layers of silica gel G on 20 x 20-cm glass carrier plates and activate by heating at 120" C for at least 2 hours.Spot on the sample solution and develop the plate in a suitable solvent system by ascending chromatography. [For general screening use solvent (i); if compounds with low RF values in this solvent are encountered, use solvent (ii) for a second determination. For menazon, use solvent (iii).] Remove from the tank when the solvent has travelled about 10 cm from the origin. When the plate is dry, spray uniformly with the hydriodic acid spray. Clip a similar glass carrier plate over the sprayed surface and heat in an oven at 180" C for 30 minutes, preferably with the plates standing in a vertical plane to aid heat flow over the surfaces.Remove the plates from the oven and unclip the cover-plate. (Carry this out under a fume hood, as iodine fumes are evolved.) When the plate is cool, spray with ammonium molybdate solution and replace in the oven for 5 minutes. Remove and, when cool, spray with tin(I1) chloride solution. Bleach the background by placing the plate in a tank containing an atmosphere of ammonia vapour. Organophosphorus pesticides appear as blue spots on a buff background. Compounds containing no phosphorus do not give this reaction, but large amounts of some co-extractives can appear as light brown, charred areas, which may some- times mask the blue coloration. Determine the Rp values of any blue spots and attempt a characterisation by reference to standards (see Table 11).282 ASKEW et at?.: A GENERAL METHOD FOR THE DETERMINATION OF [Artalyst, Vol. 94 GEL CHROMATOGRAPHY- This technique need only be applied as an adjunct to identification when the presence of pesticides is indicated by one or both of the previously mentioned chromatographic techniques. Soak beads of Sephadex LH20 in absolute ethanol for 24 hours and pour the swollen gel slurry into a column, 2.6; cm in diameter x 30 cm long, until a packed bed volume of 75ml is obtained. Run the ethanol from the column until the meniscus just touches the top of the gel bed and add the 1 ml of ethanol sample solution without disturbing the column surface. Run it until the meniscus again touches the surface, and add a further 1 ml of ethanol and repeat. Add 6; ml of ethanol to the column, attach a reservoir containing about 150 ml of ethanol and allow the column to elute at a rate of about 2 ml minute-1, by using pressure if necessary.Collect the fractions eluting between 40 to 60, 61 to 70 and 70 to 110 ml. Do not allow the column to run dry, as it can be used repeatedly. Note that the elution fractions are based on an elution volume for parathion of 75-6 mI, which we have found to be constant on several different columns: should parathion show a different elution volume, the new fractions to be collected can readily be evaluated from the relative elution volumes shown in Table I1 and stage (b) of the analytical scheme. Evaporate each fraction to small volume in a Kuderna-Danish evaporator and examine by gas and thin-layer chromatography as previously described.QUANTITATIVE ASPECTS- The analytical scheme described is essentially a qualitative procedure but can readily be rendered quantitative. Those pesticides giving gas-chromatographic peaks can be quanti- tatively determined by reference to a standard peak height - concentration curve. A standard curve should be prepared the same day as the analysis is made; dilute sub-standards, which decompose fairly readily at room temperature and on exposure to light, are prepared from a refrigerated stock solution (about 1000 pg ml-I). The thin-layer procedure can be rendered crudely quantitative by a visual comparison of spot intensity or area with a standard series. A better method is to run a second plate with a fresh extract, scrape off the layer without spraying and determine the phosphorus content colorimetrically (see Abbott, Burridge, Thomson and Webbls).RESULTS The thin-layer procedure capable of detecting down to 1 pg can be applied to waters containing 0.001 p.p.m. of pesticide, while the more sensitive gas-chromatographic procedure can usually be used to detect levels a factor of 10 or more lower. The procedure has been applied successfully in this laboratory to numerous river waters fortified at these levels. TABLE I11 RECOVERY OF PESTICIDES PASSED THROUGH THE ANALYTICAL PROCEDURE One-litre water samples containing 0-001 p.p.m. of pesticide extracted with chloroform and the pesticide subsequently eluted from a Sephadex LH20 column with ethanol. Determinations were made by gas chromatography Pesticide Recovery, per cent.Chlorfenvinphos . . .. .. 82, 87 Demeton-S-methyl . . .. 71, 75 Dimethoate . . .. .. 89, 86 Malathion . . .. .. .. 92, 92 Parathion . . .. .. .. 90, 92 Phorate . . .. .. .. 72, 75 Pyrimithate . . .. .. 75, 79 Table I11 shows the recoveries of pesticides from l-litre samples of water “spiked” at a level of 0.001 p.p.m. with various pesticides. The values were obtained from samples extracted with chloroform and then subjected to separation on a Sephadex column; all of the pesticides quoted in Table I1 gave recoveries of greater than 70 per cent. When additional clean-up is required lower recoveries are encountered.April, 19691 ORGANOPHOSPHORUS PESTICIDE RESIDUES IN RIVER WATERS 283 Of interest to the analyst is the fate of pesticides when present in water containing large amounts of suspended solids.A laboratory experiment was carried out with samples of a mixed liquor from an activated sludge plant; these were “spiked” with a mixture of five pesticides at the 0.01 p.p.m. level. After 24 hours, the solids were filtered off and the aqueous phase examined for pesticides. Assuming that the losses are caused by adsorption rather than hydrolysis, and our previous work supports this assumption, the results indicate that the fate of a pesticide in such solutions is dependent on its structure. No evidence of oxidation products was found. The results are shown in Table IV. TABLE IV RECOVERY OF PESTICIDES FROM SOLUTIONS WITH HIGH SUSPENDED SOLIDS The solutions were “spiked” at the 0.01 p.p.m. level and stored for 24 hours.The solids were filtered off and the aqueous phase analysed by gas chro- matography for the presence of pesticides Pesticide recovered in aqueous phase, per cent. r \ A Pesticide Sample 1* Sample 2t Chlorfenvinphos . . .. 33, 36 37, 40 Demeton-S-methyl . . .. 82, 80 75, 80 Dimethoate . . .. .. 62, 68 60, 62 Parathion . . .. .. 0, 0 0, 0 Pyrvnithate . . .. .. 40, 40 36, 32 on ignition 2760 mg per litre. on ignition 4080 mg per litre. Sample 1. Total suspended solids 4200 mg per litre; and solids volatile t Sample 2. Total suspended solids 4600 mg per litre; and solids volatile Permission to publish this paper has been given by the Government Chemist, Ministry of Technology. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. REFERENCES Faust, S. D., and Suffet, I. H., in Gunther, F. A,, Editor, “Residue Reviews,” Sprhger-Verlag, Ruzicka, J. H., Thomson, J., and Wheals, B. B., J . Chvomat., 1967, 31, 37. Laws, E. Q., and Webley, D. J.. Analyst, 1961, 86, 249. Bates, J. A. R., Ibid., 1965, 90, 453. Ruzicka, J. H., Thomson, J., and Wheals, B. B., J . Chromat., 1967, 30, 92. Egan, H., Hammond, E. W., and Thomson, J., Analyst, 1964, 89, 175. Hartmann, H. C., Bull. Envir. Contam. Toxic., 1966, 1, 159. Giuffrida, L., J . Ass. Of. Agric. Chem., 1964, 47, 293. Brody, S. S., and Chaney, J. E., J . Gas Chromat., 1966, 42. Suffet, I. H., Faust, D. S . , and Carey, W. F., Enviv. Sci. Tech., 1967, 1, 639. Cook, C. E., Stanley, C. W., and Barney, J. E., AnaZyt. Chem., 1964, 36, 2354. Simmons, J. H., and Tatton, J. O’G., J . Chromat., 1967, 27, 253. Abbott, D. C., Burridge, A. S., Thomson, J., and Webb, K. S., Analyst, 1967, 92, 170. Abbott, D. C., and Egan, H., Ibid., 1967, 92, 475. Abbott, D. C., and Thomson, J., in Gunther, F. A., Editor, o p . cit., Volume 11, 1965, p. 1. Watts, R. R., in Gunther, F. A., Editor, op. cit., Volume 18, 1967, p. 105. Ragab, M. T. H., J . Ass. 08. Analyt. Chem., 1967, 50, 1088. Hanes, C. S., and Isherwood, F. A., Nature, 1949, 164, 1107. Bunyan, P. J., Analyst, 1964, 89, 615. Barney, J. E., J . Chromat., 1965, 20, 334. Saliman, P. M., AnaZyt. Chem., 1964, 36, 112. Ruzicka, J. H., Thomson, J., Wheals, B. B., and Wood, N. F., J . Chromat., 1968, 34, 14. Burke, J. A., Mills, P. A., Bostwick, D. C . , J . Ass. OH. Analyt. Chern., 1966, 49, 999. Berlin, Gottingen and Heidelberg, Volume 15, 1966, p. 44. First received March 21st, 1968 Amended August 14th, 1968 Accepted October 29th, 1968

ISSN:0003-2654    

DOI:10.1039/AN9699400275

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

284 Analyst, April, 1969, Vol. 94, pp. 284-291 Determination of Free Hydrogen Cyanide in River Water by a Solvent-extraction Method BY H. A. C. MONTGOMERY, DEIRDRE K. GARDINER AND J. G. G. GREGORY* (Water Pollution Research Laboratory, Elder Way, Stevenage, Herts.) The method described for the determination of undissociated hydrogen cyanide in river waters is designed to avoid disturbance of the equilibria between hydrogen cyanide, cyanide ion and complex cyanides during the determination. A small proportion of the hydrogen cyanide present is extracted by equilibrating the sample with 1, 1, 1-trichloroethane. The extracted hydro- gen cyanide is then transferred into sodium pyrophosphate solution, and the determination is completed colorimetrically by a modification of the method of Bark and Higson.The effect of temperature on the extraction has been studied. Extractions can be carried out in the field, and conditions have been estab- lished for bringing extracted samples back to the laboratory for colorimetric analysis. The method can be used in the range 0 to 2 mg of hydrogen cyanide 1-1 and less than 0.01 mg 1-1 is detectable. Satisfactory tests have been carried out on solutions of known hydrogen cyanide content, on solutions of metallic complex cyanides and on river waters containing added cyanide. INVESTIGATIONS into the toxicity of cyanides to aquatic organisms have shown that the substance usually responsible for the acute toxicity of cyanide solutions to fish is free hydrogen cyanide, and not cyanide ion. Evidence supporting this view has been published by various authors, including Wuhrmann and Wokerl for simple cyanides, and Doudoroff, Leduc and Schneider2 for complex cyanides. Consequently there is a need in fish toxicology for an analytical method for determining hydrogen cyanide.At least two river authorities in England are now setting upper limits for the concentrations of “uncomplexed”3 or “simple ~oluble”~ cyanide in effluents discharging into rivers. It appears that cyanide ion is toxicologically important only because of the hydrogen cyanide with which it is in equilibrium, according to the equation CN- + H20 + HCN + OH-. The ratio of hydrogen cyanide ion in a sample varies with pH, temperature and ionic strength. By using 6.2 x 10-lo as the dissociation constant of hydrocyanic acid at 25” C,5 the ratio of hydrogen cyanide to cyanide ion at infinite dilution in distilled water is 161 at pH 7, 16 at pH 8, 1.6 at pH 9 and 0.16 at pH 10.As far as complex cyanides are concerned the work of Doudoroff, Leduc and Schneider,2 and others, has shown that concentrations of hydrogen cyanide toxic to fish are not normally formed by ferro- and ferricyanide (except in direct sunlight), whereas the complex cyanides of zinc and cadmium are almost completely dis- sociated in dilute solution and, therefore, exhibit the same toxicity as cyanide itself; nickelo- cyanide is of intermediate stability (and toxicity), and the degree of dissociation is strongly dependent on the pH value. The case of copper is more complicated, as cuprocyanides and cyanogen can be formed as well as cupricyanide.If colorimetric methods for cyanide are applied directly to polluted waters there is an indeterminate contribution from complex cyanides, and other methods in common use give a value for “total” cyanide (k, hydrogen cyanide PZUS cyanide ion ~ Z U S complexed cyanide) ; in either instance the result may bear no relation to the toxicity of the sample to fish. Various methods have been published for the determination of hydrogen cyanide or “free cyanide” in aqueous solution. Worley and BrowneG compared the amounts of hydrogen cyanide evolved on passing the same volume of air through the sample and through a standard solution of hydrogen cyanide ; the proportion of hydrogen cyanide evolved was small, therefore the equilibrium was not significantly affected.Schneider and Freund’ used a similar principle to .that of Worley and Browne in a sensitive gas-chromatographic method, and a static “partial” aeration method has been developed at the Water Pollution Research Laboratory.8 * College-based sandwich course student from Woolwich Polytechnic. 8 SAC; Crown Copyright Reserved.MONTGOMERY, GARDINER AND GREGORY 285 Most other methods involve the more-or-less complete removal, or reaction, of cyanide (some- times with the addition of reagents, or adjustment of pH) and, therefore, seem to be most useful in situations in which the possible disturbance, in the method of analysis, of the equili- bria between hydrogen cyanide, cyanide ion and complex cyanides would be unimportant. Such methods have been based on solvent extractionJg distillation,1° aeration,ll J2 polaro- graphyll ,13 and ion exchange.14 In the proposed method a small proportion of the hydrogen cyanide is extracted for subsequent colorimetric analysis.The proportion extracted (less than 4 per cent.) is such that the equilibria will scarcely be affected, and the dissociation of complexes to form more hydrogen cyanide will not occur to a significant extent. A modification of Bark and Higson's methodl6 is used for the final colorimetric determination. The extraction procedure is designed for field use. CHOICE OF SOLVENT AND CONDITIONS OF EXTRACTION- DenigW6 has published distribution coefficients for hydrogen cyanide between water and various organic solvents. Of the solvents heavier than water, chloroform appeared to be most suitable for the present application, as the distribution coefficient ([HCN] in chloro- form) to ([HCN] in water) is 0.28 at 18" C.When tested, chloroform was found to be suitable for the extraction of hydrogen cyanide, with a sample-to-solvent ratio of 4 : 1, but apparently carbon dioxide was extracted as well as hydrogen cyanide because the pH value of river waters rose, typically by about 0.3 unit. Changes in pH value must be avoided as far as possible because of the effect on the ionisation of hydrogen cyanide and on the dissociation of the complex cyanides of nickel and, probably, copper. Various solvents and solvent mixtures were tested and 1,1, l-trichloroethane (methylchloroform : "Genklene") was found to give satisfactory extraction of hydrogen cyanide, with a relatively small effect on the pH value.For example, when a sample of stream water of initial pH 7.70 was shaken with a mixture of equal volumes of chloroform and toluene the final pH value was 7-92, but with l,l,l-trichloroethane the final pH value was 7.71. The conditions finally chosen were to shake 100 ml of sample with 35 ml of l,l,l-trichloroethane in a conical separating funnel of nominal capacity 100 ml and actual capacity 145 ml; the head space is thereby kept small to limit the loss of gaseous carbon dioxide. Changes in pH value under these conditions are shown in Table I; all river-water samples, except the first two, were examined in the field with a battery-operated pH meter of limited accuracy.EXPERIMENTAL TABLE I EFFECT OF EXTRACTION ON pH VALUE pH value River waters . . .. Sewage effluents . . .. before extraction after extraction 7.88 8.70 7-50 7.40 7.70 8.10 8.50 6.74 6.72 7.85 8-64 7.65 7.40 7-40 8.00 8.70 7.02 7-10 It is seen from Table I that changes in pH value are not completely avoidable, and it is possible that substances other than carbon dioxide are extracted on some occasions. Equilibrium in the distribution of hydrogen cyanide between the phases was Virtually reached after shaking for 15 seconds, and the time of shaking was standardised at 30 seconds. The proportion of hydrogen cyanide extracted is 3.7 per cent. at 22" C, giving a distribution coefficient of 0-11. Emulsions may be formed at the interface with samples of polluted water.Sometimes the emulsion can be broken by shaking with a glass bead, but if the emulsion persists it is permissible to take a measured volume of the clear organic phase for analysis.286 MONTGOMERY et al. : DETERMINATION OF FREE HYDROGEN [Analyst, Vol. 94 The results are affected by the purity of the 1,1,l-trichloroethane, hence the best results were obtained when the solvent was distilled before use through an 0.4-m Vigreux fractionating column. The distillate is passed through a bed of anhydrous sodium sulphate to remove water droplets. Comparison of distilled l,l,l-trichloroethane with solvent from a freshly opened bottle showed that the distilled solvent gave a more rapid separation of the phases and higher final optical densities than did the solvent that had not been distilled.Used 1 ,l,l-trichloroethane was successfully recovered for re-use by distillation in the same appara- tus. Recovery of used solvent is recommended because l,l, 1-trichloroethane is extremely toxic to the bacteria responsible for the anaerobic digestion of sewage sludge, and hence it should not be discharged into a sewage system. In the absence of evidence that the solvent is harmless to aquatic life, disposal to effluent channels and rivers is also inadvisable. Re- distilled solvent should be used within 2 months, as acids are formed on prolonged storage. Before applying the colorimetric procedure of Bark and Higson it is necessary to restore the extracted hydrogen cyanide to aqueous solution. Emulsions are formed when l,l,l-tri- chloroethane is shaken with sodium hydroxide solution, and solutions of various multivalent alkaline salts were therefore tried.A 2 per cent. solution (10 ml) of tetrasodium pyrophos- phate did not give emulsions and was subsequently used. On extracting the organic layer with two successive volumes of pyrophosphate solution (pH 10.25) it was found that 94-8 per cent. of the hydrogen cyanide was extracted by the first batch of reagent and 5-0 per cent. by the second. A single extraction is used in the proposed method. COLORIMETRIC PROCEDURE- The determination is completed by a modification of the sensitive method of Bark and Higsonfi (for samples of limited size). The amounts recommended by Bark and Higson were increased so that the whole of the pyrophosphate extract could be taken, and the determination is carried out in 20-ml graduated flasks.The isosbestic point was found to be at 508 nm under the conditions used, and readings were, therefore, made at this wavelength instead of at 515nm. STORAGE OF EXTRACTED CYANIDE- It is inconvenient to carry out the whole determination in the field, and inadvisable to bring samples to the laboratory for analysis because of the possibility of changes in temperature, loss of carbon dioxide, or loss of hydrogen cyanide by physical, chemical or biological means. The conditions for bringing extracted cyanide back to the laboratory were therefore investigated. Storage in pyrophosphate is unsatisfactory because the blank value increases on standing, and storage in l,l,l-trichloroethane is also unsatisfactory because the hydrogen cyanide content falls on standing.However, numerous tests have shown that the hydrogen cyanide can be stored overnight (18 hours) in a stoppered vessel without signifi- cant change if the pyrophosphate extract is acidified with hydrochloric acid. In practice the pyrophosphate extract is run into a 20-ml graduated flask containing 0*3ml of con- centrated hydrochloric acid, which is the first reagent required in the colorimetric procedure. A single experiment was carried out to investigate the effect of prolonged storage. Low results were obtained when solutions were kept for 42 hours at room temperature, but there was little change at 5" C (although the blank value increased slightly). CALIBRATION- The method is calibrated by the extraction of acidified standard solutions of potassium cyanide in the range 0 to 2 mg of hydrogen cyanide 1-l.An almost straight line is obtained. For consistent results it is necessary to make up solutions in water distilled from glass; de- ionised water is not satisfactory. Jenkins, Hey and Cooperll also found that de-ionised water was unsatisfactory for the detection of small amounts of cyanide. EFFECT OF TEMPERATURE- Over sixty determinations in the range 9" to 22" C showed that the amount of hydrogen cyanide extracted by l,l,l-trichloroethane is reduced by an average of 2 per cent. for each deg C by which the temperature falls below 22" C. Although the temperature coefficient is probably not a linear function, a correction of 2 per cent. per deg C is sufficiently accurate.April, 19691 INTERFERENCES- As Bark and Higson reported that their colorimetric method was subject to few inter- ferences, no systematic study of interferences was made, although phenol ( 6; mg1-l) was tested and found not to interfere. All glassware should be cleaned frequently with chromic acid, or high blank values and erratic results will be obtained.CYANIDE IN RIVER WATER BY A SOLVENT-EXTRACTION METHOD 287 METHOD REAGENTS- All reagents must be of the best quality available, and all water used distilled from glass. 1,l ,1-Trichloroethane-Re-distilled or recovered as described previously. Store in a dark bottle and use within 2 months. Sodium pyrophosphate solution, 2 per cent.-Dissolve 10 g of sodium pyrophosphate, Na,P,O,.lOH,O, in water and dilute to 500 ml.Check the pH value from time to time and discard the reagent if the pH falls to 10.1. Hydrochloric acid, comcentrated. Hydrochloric acid, 0.5 M. Bromine water, satwated. Arsenous acid solution-Dissolve 2 g of arsenous oxide in 100ml of water by heating under reflux. Pyridine solution-Mix 150 ml of pyridine with 100 ml of water and add 25 ml of con- centrated hydrochloric acid. p-Phenylenediamine solutio-Dissolve 1 07 g of $-phenylenediamine dihydrochloride in 500 ml of 0.5 M hydrochloric acid. Mixed reagennt-Add 3 volumes of pyridine solution to 1 volume of fi-phenylenediamine solution immediately before use, and mix; do not prepare more than is necessary (8 ml per sample). Standard cyanide solzltion-Dissolve 1-65 g of potassium cyanide in water, add 25 ml of M sodium hydroxide solution, and dilute to 250 ml.Standardise by titration against silver nitrate solution.17 The solution is equivalent to a solution of hydrogen cyanide of concentration 2700 x 2 x volume of O*lOOO N saver nitrate required volume of cyanide titrated mg of hydrogen cyanide 1-l. Immediately before use, dilute 5 ml to 250 ml and then dilute 20 ml of the resulting solution to 500 ml to give a working solution equivalent to about 2 mg of hydrogen cyanide 1-l. PREPARATION OF CALIBRATION GRAPH- Introduce 35 ml of l,l,l-trichloroethane into a separating funnel, capacity 140 to 150 ml, then add 100 ml of water (for determination of the blank value) and 0.1 ml of 0-5 M hydro- chloric acid. Stopper the funnel and shake it vigorously for 30 seconds.Allow the layers to separate and transfer the organic layer to a 50-ml separating funnel containing 1 O m l of sodium pyrophosphate solution; discard the original aqueous layer. Shake the 50-ml funnel vigorously for 30 seconds, allow the layers to separate and remove the organic layer for subsequent recovery. Transfer the aqueous layer to a 20-ml calibrated flask containing 0.3 ml of concentrated hydrochloric acid. Rinse the funnel with a few drops of water, add the washings to the flask and stopper the flask tightly. Repeat the procedure, extracting measured amounts of the working cyanide solution diluted to 100ml in each case with water. Add 0-2mI of bromine water to each flask, stopper the flask and set aside for 5 minutes. Add 0.4 ml of arsenous acid, stopper and shake vigorously to de-colorise not only the solution but also any drops of l,l,l-trichloroethane that may be present, then add 8 ml of freshly prepared mixed reagent.Dilute to the mark, mix well and set aside for 30 minutes. Within the next 20 minutes, measure the optical density at 508 nm, with water in the reference cell, mixing if necessary to avoid striations. Use 4-cm cells for concentrations up to 0.2 mg of hydrogen cyanide -l, thereafter use 1-cm cells. Correct for the blank value and construct calibration graphs. Note the temperature at which the extractions are carried out.288 MONTGOMERY et al. : DETERMINATION OF FREE HYDROGEN [Analyst, Vol. 94 PROCEDURE FOR SAMPLES- Adjust the temperature of the bottle of 1,1,l-trichloroethane to about that of the river, then obtain the sample with the least possible disturbance (metallic sampling devices are not recommended). Immediately measure 100 ml of sample into a separating funnel containing 35 ml of l,l,l-trichloroethane.Add a glass bead (3 to 4 mm) if experience suggests that an emulsion is likely to be formed. Note the temperature, stopper the funnel and shake it vigorously for 30 seconds. Continue as described under Preparation of calibration graph and, with each batch of samples, carry out a blank determination on 100 ml of distilled water. If a persistent interfacial emulsion is formed on extracting the sample, run off as much as possible of the clear organic layer and measure the volume before extraction with the pyro- phosphate reagent.If the extraction is carried out in the field, continue until the pyrophos- phate solutions are in the stoppered flasks, in the presence of hydrochloric acid, before bringing them back to the laboratory. Then, if t, is the temperature at which the calibration graph was prepared, t, is the temperature of sample, ZI is the volume in millilitres of organic phase taken and C is the concentration of hydrogen cyanide read from calibration graph (after correcting for the blank)- c x 35 Concentration of hydrogen cyanide in sample = z, [I - 0.02 (tl - t,)] * RESULTS REPRODUCIBILITY AND SENSITIVITY- Replicate determinations on standard solutions gave standard deviations of &0-006 mg 1-1 for 0.028 mg of hydrogen cyanide I-l, and + O * O l l mg 1-1 for 0.28 mg of hydrogen cyanide 1-1.The standard deviation of forty-three determinations of the blank value in 4-cm cells was f0.005 mg l-l, suggesting a limit of detection of 0.01 mg 1-l; however, lower standard devi- ations are possible when working with a single batch of reagents, and it was confirmed that less than 0.01 mg 1-1 could be detected. EFFECT OF pH VALUE- To' test the method, the concentrations of hydrogen cyanide found in solutions of known pH value and known total cyanide content were compared with the calculated concentrations. Solutions of cyanide containing 0.005 M diethanolamine were prepared and varying amounts of hydrochloric acid were added to give pH values between 8 and 10. The total concentration of cyanide was 0.90 mg 1-1 (as hydrogen cyanide). The solutions were analysed in duplicate by the proposed method and the pH values were determined after extraction.It was shown that the diethanolamine and diethanolamine hydrochloride had no effect on the blank value at the concentrations used. All extractions and pH measurements were made at 25" C. The theoretical hydrogen cyanide concentrations were calculated from the equation aH+ x acN- ~ H C N Ka = where a denotes activity and K , is the thermodynamic dissociation constant of hydrocyanic acid (6.17 x 10-10 at 25" C).s The pH measurements were assumed to give the hydrogen-ion activities, and the activity coefficient of un-ionised hydrocyanic acid was taken to be unity. Cyanide-ion activity coefficients were calculated by using the simple form of the Debye- Hiickel equation as the ionic strengths were low.The results (Table 11) show satisfactory TABLE I1 PREDICTED AND OBSERVED CONCENTRATIONS OF HYDROGEN CYANIDE AT DIFFERENT pH VALUES Hydrogen cyanide -- Predicted, Found, pH value Ionic strength mg 1-1 mg 1-1 9.90 2 x 10-4 0.15 0-16 9.61 7 x 10-4 0.25 0.30 9.04 2 x 10-3 0.53 0-50 8.85 3 x 10-3 0.61 0.68 8.39 4 x 10-3 0.78 0.79April, 19691 289 agreement between the predicted and observed concentrations. The observed concentrations have been corrected when necessary for the slight disturbance of the equilibrium caused by the extraction of a small proportion of the hydrogen cyanide; in no case did the correction exceed 0-01 mg 1-l. EFFECT OF COMPLEXING METALS- Determinations of hydrogen cyanide were carried out in solutions of the complex cyanides of iron, which are virtually undissociatedJ2J1 and in those of zinc and cadmium, in which dissociation is practically complete at concentrations of 10" M and lower.The results were in accordance with expectation (Table 111). Also included in Table I11 are results for mixtures of nickel sulphate and copper(I1) sulphate with four equivalents of potassium cyanide. Each result in Table I11 is the average of either two or three determinations. The zinc, cadmium, nickel and copper cyanide solutions were prepared by mixing dilute solutions of potassium cyanide and of the metallic sulphates; the solutions were then used within 4 hours of mixing and the pH values were determined before extraction. For nickel, the hydrogen cyanide content of acidified solutions depended on whether acid was added before or after the cyanide and nickel sulphate solutims were mixed.By using 5.4 x lod1 as the dissociation constant for nickelocyanide,2 it can be shown that the results in Table I11 are in reasonable agreement with the expected values, with the exception of the four results in brackets, which are all low. The reason for the discrepancies was not investigated, although the virtual absence of buffering ions may have permitted pH changes during extraction, In the case of copper it was not possible to predict the equilibrium concentration of hydrogen cyanide, as the extent of decomposition of cupricyanide to cuprocyanide and cyanogen was unknown. The results in Table I11 for solutions containing nickel and copper are included mainly to illustrate the use of the analytical method; in solutions containing nickel, the concentration of hydrogen cyanide present appears to depend to some extent on the previous treatment of the sample.TABLE I11 DETERMINATION OF HYDROGEN CYANIDE IN THE PRESENCE OF CYANIDE IN RIVER WATER BY A SOLVENT-EXTRACTION METHOD COMPLEXING METALS Test solution pH value K,Fe(CN), . . .. .. .. ZnSO, plus 4KCN . . .. .. NiSO, plus 4KCN . . .. .. .. .. .. .. .. K3Fe(CN)6 CdSO, plus 4KCN . . (pH value not adjusted) NiSO, plus 4KCN . . ., NiSO, plus 4KCN . . .. .. (KCN acidified before addition of NiSO,) (acid or alkali added after mixing) - - - - 6.2 6.5 6.5 6-9 5.5 5.8 6-2 3.1 5.0 5.3 5.6 7.1 9.4 CuSO, plus 4KCN . . .. .. 6.8 (no pH adjustment) (acid or alkali added after mixing) CuSO, plus 4KCN .. .. .. 3.5 4.8 5-6 9.0 Total cyanide present (expressed as hydrogen cyanide), mg 1-1 0.81 0.81 0.56 0.56 1-35 1.35 1.35 1-36 1-34 1.34 1.34 1.35 1.34 1-34 1-36 1.35 1-35 1.34 1.34 1.34 1.34 1-34 Hydrogen cyanide found, mg 1-1 <0-01 0.01 0.56 0.56 1.00 (0.26) 0.58 0.43 1.16 1.17 1.00 1-18 (0.33) (0-35) (0.30) 0.25 0-02 0.60 1.27 0.83 0-95 0.35 RIVER WATERS- Known amounts of potassium cyanide were added to samples of river water. Hydrogen cyanide was then determined in triplicate by the proposed method. The average results290 MONTGOMERY et al. : DETERMINATION OF FREE HYDROGEN [Artahst, Vol. 94 are shown in Table IV together with the expected values, which were calculated by allowing for the ionisation of hydrocyanic acid, but ignoring any effect of ionic strength on the dissociation constant.TABLE IV RECOVERY OF ADDED CYANIDE Sample River water A . . . . A .. .. B .. .. B .. .. c .. .. D .. .. D .. .. E .. .. E .. .. pH value after addition of cyanide - 7-6 8.05 7.2 7.9 8.0 8.0 8.0 - Cyanide added (as hydrogen cyanide), mg 1-l 0.04 1-07 0.04 1.07 1.03 0.03 1.10 0.04 1.10 Calculated hydrogen cyanide content allowing only for pH value, mg 1-1 0.04 1.05 0.04 1.02 1-02 0.03 1-05 0.04 1.05 Hydrogen cyanide found, mg 1-1 0-04 0.99 0.05 0.98 0-87 0.02 0.80 0-03 0.83 Samples A and B gave almost quantitative recoveries of hydrogen cyanide. Recoverieswere somewhat low with samples C, D and E, all of which were from polluted rivers, and it is likely that complexing metals were present.The proposed method was compared with this laboratory's diffusion method* on a biologically treated sewage effluent and on two samples of river water (Table V). The river and the sewage works both receive discharges from plating works. Results for total cyanide ( i e . , free plus complex cyanide) are included. The two methods for hydrogen cyanide were in agreement, within experimental error. TABLE V COMPARISON OF METHODS FOR DETERMINING HYDROGEN CYANIDE Hydrogen cyanide Hydrogen cyanide SampIe Total cyanide, by diffusion method, by proposed method, mg 1-l mg 1-1 mg 1-1 River water . . .. .. 0.01 - 0.005, 0.009, 0-009 River water . . .. . . 0.025 0.015 0.01, 0.01 Sewage effluent . . . . 0.21 0.036 0.04, 0.05, 0.06 DISCUSSION The criteria used in developing the proposed method were that only a small proportion of the hydrogen cyanide should be extracted (thus avoiding disturbance of the equilibria), that the temperature and pH value of the sample should remain constant during the extrac- tion and that the method should be convenient for field use.The first criterion is fulfilled by the use of conditions whereby less than 4 per cent. of the hydrogen cyanide is extracted, and maintaining a reasonably constant temperature is merely a matter of technique. Changes in pH value are not completely avoided (Table I) but are likely to be less severe than in aeration methods, as unpublished work at the Water Pollution Research Laboratory has shown that carbon dioxide is removed from aqueous solution by aeration more rapidly than hydrogen cyanide.The method is convenient for field use. The evaluation of the proposed method is difficult because it was not convenient to use the method of Schneider and Freund' for comparison, and no other method of proved accuracy was available. Measurement of the recovery of cyanide added to polluted river waters is not a good means of studying the accuracy of the method because such waters will normally contain complexing metals and, even if the concentrations of iron, nickel and copper were measured, it would not be possible to predict the equilibrium concentration of hydrogen cyanide accurately. However, from the fact that a reproducible, almost straight- line, calibration graph is obtained, it appears that hydrogen cyanide obeys the Distribution law in the system and concentration range (0 to 2 mg 1-l) studied.The only factor likely to affect the distribution coefficient is the salting-out effect of electrolytes in the sample, resulting in an increase in the amount of hydrogen cyanide extracted, but no such increase was noted in the results of Table IV, which were obtained with hard waters, even with samples A and B, which were unpolluted and probably low in complexing metals; in any case, no large increaseApril, 19691 CYANIDE IN RIVER WATER BY A SOLVENT-EXTRACTION METHOD 291 is to be expected. The fact that the apparent recoveries shown in Table IV for samples A and B are slightly low is less likely to be caused by a change in the distribution coefficient than by an increase in the dissociation constant of hydrocyanic acid, due to the hard-water salts present ; the predicted hydrogen cyanide concentrations would, therefore, be high.In a similar experiment with sea water (pH 7.8, chlorinity 17.7 parts per lOOO), the apparent recovery of hydrogen cyanide was 71 per cent., the discrepancy possibly resulting from hydro- cyanic acid being ionised to a greater extent in sea water than in distilled water at the same pH value. However, such effects would not influence the distribution coefficient or the validity of the proposed method, although they affect the evaluation of the method. It is clear that nickelocyanide is insoluble in l,I,l-trichloroethane (Table IIT), and there seems no reason to suppose that other complex cyanides would be extracted (except that cationic detergents and related compounds, if present in the sample, might conceivably form extractable salts with complex cyanides).Further indirect evidence as to the validity of the method is provided by the results in Table 11, by the good results obtained with the cyanides of iron, zinc and cadmium, and by the fact that, with four inexplicable exceptions, the results for nickelocyanide were in accordance with theory (Table 111). A comparison with another distribution method,* in which a small proportion of the hydrogen cyanide is allowed to diffuse from the sample into alkali, is shown in Table V. The diffusion method has various drawbacks and the agreement between the methods must be considered remarkably good. Thus it appears that the theoretical basis of the proposed method is sound, and there is considerable indirect evidence that the method is valid.The extraction procedure is convenient for use on the bank of a river, and the only difficulty so far encountered has been slight emulsion formation at the interface. Effluents are of less interest to the fish toxicologist than river waters, as the concentrations of hydrogen cyanide resulting from the discharge of such effluents into rivers cannot be calculated accurately from the dilution ratio, especially if complex cyanides are present, because the equilibrium might be affected by the dilution and by changes in the pH value on mixing. Nevertheless, apparently satisfactory determinations of hydrogen cyanide have been carried out on a biologically treated sewage efluent and on a treated industrial effluent.The proposed method appears to be of general applicability but for concentrations of hydrogen cyanide greater than 2 mg 1-1 it would be necessary to confirm that the Distribution law was applicable and to take less than 10ml of the pyrophosphate extract for analysis. We thank Mr. D. E. Caddy and Mr. B. R. Dashwood for their assistance in the experi- mental work. Mr. T. F. Walshe for his work on the calculations and Mr. R. G. Glanville of Ipsley 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. Sewage W&ks for providing facilities. Wuhrmann, K., and Woker, H., Schweiz. 2. Hydrol., 1948, 11, 210. Doudoroff, P., Leduc, G., and Schneider, C. R., Trans. Amer. Fish SOC., 1966, 6. Annual Report for year ended 31st March, 1966, Trent River Authority, 206 Derby Road, Not- Annual Report for year ended 31st March, 1967, Severn River Authority, Church Street, Malvern, Izatt, R. M., Christensen, J. J., Pack, R. T., and Bench, R., Inorg. Chem., 1962, 1, 828. Worley, F. P., and Browne, V. R., J . Chem. SOC., 1917, 1057. Schneider, C. R., and Freund, H., Analyt. Chem., 1962, 34, 69. “Water Pollution Research 1966,” Ministry of Technology, H.M. Stationery Office, London, 1967, Kruse, J. M., and Mellon, M. G., Sewage Ind. Wastes, 1951, 23, 1402. Fleps, W., Gesundheits-Ingenieur, 1963, 84, 209. Jenkins, S. H., Hey, A. E., and Cooper, J. S., Air Wat. Poll. Int. J., 1966, 10, 496. Boye, E., Gas- u. WassFach, 1964, 105, 1393. Karchmer, J. H., and Walker, M. T., Analyt. Chem., 1955, 27, 37. Hissel, J., and Cadot-Dethier, M., Trib. Cebedeau, 1965, 18, 272. Bark, L. S., and Higson, H. G., Talanta, 1964, 11, 621. Denigks, G., Bull. Trav. SOC. Pharm. Bordeaux, 1940, 78, 61. Joint Committee of the Association of British Chemical Manufacturers and the Society for Analyti- cal Chemistry, “Recommended Methods for the Analysis of Trade Effluents,” W. Heffer &Sons Ltd., Cambridge, 1958, p. 88. Received July 16th, 1968 Accepted October 16th, 1968 REFERENCES tingham, p. 60. Worcs., p. 39. p. 152.

ISSN:0003-2654    

DOI:10.1039/AN9699400284

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

292 Analyst, April, 1969, Vol. 94, $$. 292-299 The Determination of Very Small Amounts of Selenium in Plant Samples BY R. J. HALL AND P. L. GUPTA (Ministry of Agriculture, Fisheries and Food, National Agriculturai Advisory Service, Government Buildings, Kenton Bar, Newcastle upon Tyne, NE1 2 Y A ) The determination of very small amounts of selenium in plant samples with 2,3-diaminonaphthalene has been investigated. Losses of selenium are prevented by a controlled wet-oxidation procedure in a silicone bath. A method is proposed for the determination of selenium in plant materials at levels as low as 0-005 p.p.m. in dried material. SINCE the work of Schwarz and Foltz,l and of Patterson, Milstrey and Stokstad,2 showing that a selenium-containing factor prevented exudative diathesis in chicks, the r61e of selenium as an essential trace element has assumed considerable importance for some animals.Muth, Oldfield, Remmert and Schubert3 have demonstrated that a relationship exists between the selenium content of the diet of the pregnant ewe and white muscle disease in lambs. From the reports of these and other worker^,^ it is evident that levels of selenium well below 0.1 p.p.m. in the dried fodder need to be determined with reliability and that differences between 0-01 and 0-03 p.p.m. may be significant. However, most of the analytical procedures for the determination of selenium in plant samples have been developed to assist in the diagnosis of those conditions involving selenium toxicity when amounts greatly in excess of 1 p.p.m. had been expected.It is only in recent years that reagents have been introduced that are sufficiently sensitive to assess the selenium content of foodstuffs at the very low levels mentioned above, and few papers give results with selenium values below 0.1 p.p.m., notable exceptions being those of Allaway and car^,^ Watkinson,6,7 Ewan, Baumann and Pope* and Lindbe~-g.~ Of the newer techniques, those involving the use of 3,3’-diamino- bemidinelo and 2,3-diaminonaphtha1ene1l have been the most widely adopted by laboratories engaged in studies of animal nutrition. The more sophisticated methods, such as neutron- activation analysisl2 and isotope dilution,13 have been used primarily for reference purposes. In the present investigation it was necessary to distinguish between the selenium content of what were considered to be normal pastures and those suspected of being deficient.It was felt, therefore, that a method of analysis was required that was capable of determining levels of selenium in animal feeds as low as 0.01 p.p.m. with a reasonable degree of accuracy and reproducibility. As the facilities of the oxygen-flask technique with pelleted samples used by Allaway and Cary,6 and more recently by Lane,14 were not available to us and, moreover, appear to present certain difficulties and limitations, particularly in respect of large numbers, it was decided to examine wet-digestion procedures of the kind used by Stanton and McDonald15 and by Molloy16 for soils. In our hands poor recoveries were obtained with these methods, and it was clear that modifications were necessary to determine selenium at the low concentrations mentioned.Parker and Harvey,ll who have made detailed studies of the reactions between selenous acid and aromatic o-diamines, showed that by far the most sensitive reagent for the deter- mination of selenium is 2,3-diaminophthaleneJ with which it forms 4,5-benzopiazselenol, I. I 0 SAC; Crown Copyright Reserved.HALL AND GUPTA 293 This compound displays a brilliant lime-green fluorescence when excited at 366 nm in either cyclohexane or dekalin, and has marked advantages over the formation of the 3',4'-diamho- phenylpiazselenol, 11, formed by the reaction of selenium with excess of 3,3'-diaminobenzidine, first suggested by H0ste.l' As a result of a close appraisal of wet-oxidation procedures, a method is described in which perchloric acid is used only in the terminal stage of the digestion, for a minimum period of time and at a controlled temperature, by using a silicone bath.The selenium is subse- quently determined by fluorimetry after the formation and extraction of the 4,5-benzopiaz- selenol. The technique is capable of determining 0.005 p.p.m. of selenium in 5-g samples. METHOD REAGENTS- an all-glass apparatus. Alperton, Middlesex. Use micro-analytical reagent grade chemicals when possible and water re-distilled in Detergent R.B.S. 25-Obtainable from Raymond Lamb, 12, The Viaduct, Ealing Road, Nitric acid, sp.gr. 1.42. A nti-bumping granules. Hydrogen peroxide, 100 volumes. Perchloric acid (E. Merck), 70 per cent.w/w. SiZicone j u i d F. 190-Obtainable from Imperial Chemical Industries Ltd., Silicones Ammonia solution, 7 M-Prepare by diluting ammonia solution (sp.gr. 0.88). Hydrochloric acid, 6 M. Formic acid, 50 per cent. v / v . StabiZising solzdion-Dissolve 9-306 g of the disodium salt of ethylenediaminetetra-acetic acid (EDTA) and 25 g of hydroxylammonium chloride in 1 litre of water to give a 0-025 M solution of EDTA in 2.5 per cent. solution of hydroxylammonium chloride. 2,3-Diaminonap~thaZene reagent (DAN)-Obtainable from Koch-Light Laboratories, but needs to be purified as follows. Dissolve 1 g of 2,3-diaminonaphthalene (DAN) in 20 ml of ethanol in a 50-ml beaker. Boil to dissolve and filter the solution through a small Whatman No. 1 filter-paper. Cool the filtrate to -20" C.Filter off the resulting crystals on a small sintered-glass funnel, which has also been cooled to -20" C, and wash the crystals with ice-cold 50 per cent. v/v ethanol. Repeat the procedure with the washed crystals to give a doubly recrystallised product consisting of pale fawn coloured, silky crystals. Dry the crystals in an incubator at 40" C and store at -20" C, protected from light. Prepare 200 ml of a 0-1 per cent. solution of the recrystallised DAN containing 0-5 per cent. w/v of hydroxylammonium chloride in 0.1 M hydrochloric acid. Warm the solution at 50" C for 20 minutes, then cool and extract three times with 10 ml of dekalin. Discard the dekalin layer each time. Freeze the extracted DAN solution, contained in a 250-ml Erlenmeyer flask at -20" C.Allow to melt, then remove, with a Pasteur pipette, the residual dekalin that separates and filter the DAN solution through a sintered-glass funnel of No. 4 porosity. The solution should now be colourless. Dekalin-Re-distil and use the fraction collected at 188" to 190OC. Selenium standard-Dissolve, with warming, 50 mg of pure selenium in 5 ml of perchloric acid (sp.gr. 1.70). Dilute to 50 ml with water. Dilute 1 ml of this solution to 1 litre with water (1 ml of the dilute solution contains 1.0 pg of selenium). Department, Sunley Buildings, Piccadilly Plaza, Manchester 1.294 HALL AND GUPTA: DETERMINATION OF VERY [Autalyst, Vol. 94 APPARATUS- A Locarte fluorimeter, Model LFM/4, with a primary filter LF/2, transmission 340 to 380 nm, and a secondary filter LF/9, transmission 575 nm, was used in this work. Quartz tubes of 2-ml capacity were used in the instrument.Digestion flask avtd rejux adapter-The digestion of the samples and the subsequent formation and extraction of the DAN - selenium complex (4,5-benzopiazselenol) are all completed in an apparatus that consists of a modified Kjeldahl-type digestion flask (Fig. l), with a bowl capacity of 40 ml and a small, ground, flat base of 26-cm diameter. The neck is 10 cm in length (including the B19/26 joint), with an internal diameter of 2 cm. During the digestion, an adapter, 15 to 16 cm in length (including the B19/26 socket), is fitted into the neck to act as a reflux condenser. The whole apparatus is made of Pyrex glass and can be constructed by a competent glass blower.Those used in this laboratory were supplied by Messrs. Anderson and Butler, 10 Cotter Riggs Place, West Denton, Newcastle upon Tyne. F!g. 1. Modified Kj eldahl-type digestion flask and adapter Pastew $i@ettes-These are obtainable from A. G. Gallenkamp Ltd. Thermostatically corntrolled hot-plate (12 x 8 ivtches)-Obtainable from Townson and Mercer Ltd. PREPARATION OF SAMPLES- Spread the samples (100 g) on 10 x 8-inch shallow dishes lined with polythene sheeting and dry at 40" to 60" C for 48 hours in an oven with a forced-air draught. In this laboratory the dried material is milled in a Cassella grain mill to pass a 1-mm sieve. PROCEDURE- The whole of the digestion procedure must be carried out in a fume chamber with an efficient extractor fan. It is advisable to keep all of the beakers and flasks containing digestion mixtures, together with the bottles of acid, on a special tray with an acid and heat-resisting surface.In this laboratory, disposable plastic gloves are used to protect the hands from the acids and the corrosive vapours. Protectors are worn on the thumbs and forefingers to allow easy and safe handling of the hot flasks. These protectors are simply 2-inch lengths of thin-walled rubber tubing of +inch diameter cut down one side so that the tubing will open out and close over the inside of the thumb or finger; they have proved effective. Clean all glassware by immersing it for several hours in dilute R.B.S. 25 detergent. Rinse with tap water and then with glass-distilled water. Weigh up to 5 g of dried sample into a 250-ml Pyrex beaker.Add 30 ml of nitric acid (sp.gr. 1-42) and gently warm on a hot-plate in a fume chamber, constantly stirring with a glass rod, until the plant material has disintegrated in the acid. As soon as the reaction becomes vigorous (emitting a copious brown vapour) remove the beaker from the hot-plate. After a few minutes, when the main reaction has ceased and no more frothing takes place, transfer the contents of the beaker to the special digestion flask and wash out the beaker twice with 5 ml of nitric acid (sp.gr. 1-42), adding the washings to the flask. Add a fewApril, 19691 SMALL AMOUNTS OF SELENIUM IN PLANT SAMPLES 295 anti-bumping granules. Fit the adapter and place the flask on the hot-plate. Continue the digestion by refluxing for 30 minutes.Remove the adapter and rapidly boil off the nitric acid until the volume is reduced to between 5 and 10ml. Allow the flask and its contents to cool for about 3 minutes. Add 5 ml of 100-volume hydrogen peroxide, 1 mi at a time. This step must be carried out carefully, away from the hot-plate, as a vigorous reaction occurs, which is delayed for a few seconds after the addition of each 1 ml of peroxide. When the reaction of the last addition has subsided, replace the flask on the hot-plate and again reduce the volume to between 5 and 10 ml. Allow to cool. Repeat the addition of 5 ml of hydrogen peroxide in 1-ml aliquots, as before, and then reduce the volume to about 5 ml. Add 5 ml of 70 per cent. perchloric acid and place the flask in a 9-cm Pyrex crystallising dish containing silicone fluid F.190 to a depth of 1.5 cm and maintained at 195" to 200" C on the hot-plate.In practice, it has been found convenient to heat no more than six flasks at a time at this stage, three in each of two crystallising dishes. Allow the remaining nitric acid to boil off until the volume of liquid in the flask is reduced to between 5 and 6 ml. At this point a new reaction can be seen to take place, partially caused by the evolution of the residual nitric acid but, more particularly, by the decomposition of nitrates, as shown by the production of the brown vapour of nitrogen dioxide. A more vigorous reaction now ensues, sometimes with charring of the small amount of organic material still remaining, and the perchloric acid starts to boil. From time to time, gently swirl the contents of the flask to dislodge carbon particles.The digest will rapidly clear and become pale green to colourless, the obvious effervescence subsiding quite suddenly to a gentle simmer. The digestion is now complete, and it is most important not to continue heating after this stage has been reached, otherwise selenium will be lost. The duration of heating from the com- mencement of the boiling of the perchloric acid to the completion of the digestion should not exceed 10 minutes. Remove the flask and place it in an empty 1-litre beaker to allow excess of silicone fluid to drain off. When cool enough to handle comfortably, wipe the outside of the flask with paper tissues to remove the adhering silicone. When the flask and its contents are cold, add 5 ml of 6 M hydrochloric acid, 5 ml of 50 per cent.formic acid and 10 ml of stabilising solution. Mix and add 14 ml of 7 M ammonia solution. Again mix and add more 7 M ammonia solution, dropwise, until the mixture has a pH of 2 [demonstrated by a test strip of Whatman - B.D.H. narrow-range pH paper (range 1 to a)]. Transfer the flask to a water-bath maintained at 50" C. After 5 minutes, add 5 ml of DAN reagent and dilute to about 50 ml with water. Maintain the mixture at 50" C for 20 minutes. Remove the flask from the water-bath, cool to room temperature and add 3 ml of dekalin. Stopper the flask and shake vigorously by hand, or on a shaker, for 1 minute. With a Pasteur pipette, remove the dekalin layer as completely as possible and introduce it into a 10-ml graduated glass tube, with a glass stopper.Repeat the extraction twice more with 2ml of dekalin, adding these extracts to the first. Dilute the volume of the pooled extracts to 7ml with dekalin. Wash the extracts by shaking vigorously for 30 seconds with 4 ml of 0.1 M hydrochloric acid. Remove the acidic layer with a Pasteur pipette and discard. Repeat the washing procedure. Centrifuge the tubes at 3000 r.p.m. for 2 to 3 minutes to clarify the dekalin extract. Measure the green fluorescence of the extracts of the selenium - DAN complex, after setting the instrument to give full-scale deflection of the galvanometer with a standard prepared by complexing and extracting 0-5 pg of selenium after the addition of 5 ml of 70 per cent.perchloric acid, which is neutralised with ammonia solution, and treated with formic acid and stabilising reagent as for a determination. Prepare extracts of intermediate standards of 0-05 and 0.2 pg of selenium. All determinations should be carried out in triplicate, with reagent blanks and a recovery test of a suitable amount of selenium added to the digestion flask containing one of the samples to be analysed. NOTE- It has not been found essential to work in amber light, but direct sunlight or bright daylight should be avoided by using Venetian blinds at the windows, so that the procedures are performed in diffused light. All observations were made at least in triplicate. Most workers have reported that, with their particular experimental conditions, a single extraction with 5 ml of cyclohexane or dekalin removed all of the 4,5-benzopiazselenol from EXPERIMENTAL AND RESULTS EXTRACTION OF 4,5-BENZOPIAZSELENOL INTO CYCLOHEXANE AND DEKALIN-296 HALL AND GUPTA: DETERMINATION OF VERY [Analyst, VOl.94 the aqueous complexing medium. This was not our experience, and we found it necessary to extract three times with either solvent when dealing with plant digests. Surprisingly, complete recovery of the 4,5-benzopiazselenol with a single extraction of cyclohexane was achieved when 0.5 pg of selenium [as Se(IV)] alone was complexed (in an appropriate amount of neutralised perchloric acid), but the fluorescence was only about two thirds of that in dekalin. TABLE I EXTRACTION OF 4,5-BENZOPIAZSELENOL INTO DEKALIN Fluorimeter readings Selenium, pg i s t extract (4 ml) 2nd extract (3 ml) 3rd extract (3 mi) 0.1 27-5 4.5 1.5 0.2 48.5 6 2.5 0.5 107 22 8 In Table I are shown the results of single and multiple extractions of standard amounts of selenium under our conditions.The 4,5-benzopiazselenol was extracted with 4, 3 and 3-ml volumes of dekalin and each separate extract made up to 10ml with the solvent. The fluorescence of each extract was compared with that of 0.5 pg of added selenium extracted in the same way but with the three extracts pooled to a volume of 10 ml. The fluorimeter was set with this extract to give a maximum galvanometer deflection of 140. TABLE I1 EXTRACTION OF 4,5-BENZOPIAZSELENOL INTO DEKALIN Fluorimeter readings Selenium, pg '1st extract (5 ml) 2nd extract (2 mlj 0 3 2.5 0.2 65 14 0.4 106 27 0.5 140 32 The results in Table I1 record a similar experiment but, in this instance, two extracts with 5 and 2 ml of dekalin were made and measured separately, the readings being related to the first 5-ml extract of 0.5 pg of selenium, which was used to set the instrument to give a maximum reading of 140.It is evident from these results that a single extraction is insufficient for the complete separation of the 4,5-benzopiazselenol before fluorimetric measurement. In every instance the second extract exhibited the typical green fluorescence of the 4,5-benzopiazselenol. RECOVERY OF SELENIUM FROM PLANT DIGESTS- It rapidly became apparent that at the very low levels of selenium in which we were interested, results were variable when the plant samples were digested by several methods described in the literature, and attempts to recover 0.05 to 0.5 pg of selenium from 5 g of grass by these methods gave rise to low and inconsistent values.It was felt that the losses were associated with high digestion temperatures, or prolonged digestion , or a combin- ation of these factors. Tests showed the recovery of 0.2 to 1.Opg of selenium added to 5 and l o g of plant material with low selenium content to vary from 40 to 91 per cent. when digested with nitric and perchloric acids. A study of the changes in the digest characteristics, as the temperature rose following the removal of nitric acid, indicated that the digestion with perchloric acid itself must be minimal if losses of selenium are to be prevented A digestion procedure was, therefore, adopted, which relies mainly on the destruction of organic material with nitric acid and hydrogen peroxide, leaving only a little plant wax.The digestion is completed with a relatively small volume of perchloric acid for a short time at the controlled temperature range of 195" to 200" C. In Table I11 are shown the differences in the selenium levels of grass samples after controlled and uncontrolled digestion with perchloric acid. For the uncontrolled digestions the procedure was the same as in the described method, except that 5 or 10ml of perchloric acid were added after the peroxide stage, and the volume of the digest reduced to 2 ml by boiling on the hot-plate without the use of a silicone bath. It can be seen that the differences are considerable in some samples.April, 19691 SMALL AMOUNTS OF SELENIUM IN PLANT SAMPLES 297 TABLE I11 SELENIUM LEVELS (P.P.M.) OF DRY HERBAGES: MEASURED AFTER PERCHLORIC ACID DIGESTION Uncontrolled boiling Controlled perchloric Sample perchloric acid acid at 195" to 200" C J.1370 0.025 0.039 5.1371 0.012 0.026 5.1372 0.009 0.035 J.1373 0.025 0.036 5.1374 0.025 0.041 J.1375 0.017 0.025 J.1413 0.027 0.044 J.1414 0.024 0.030 J .4195 0.012 0.041 K.729 0-046 0-078 K.730 0.044 0.069 K.731 0.053 0.100 Mixed grass 0.047 0.057 Kale 0.099 0.127 Casein 0.203 0.309 RECOVERY TESTS- Recoveries of 0.05 to 0-5 pg of selenium added to 5 g of barley grain, which had a very low selenium content, are shown in Table IV.It is clear that the losses from the entire procedure finally adopted are small.Similar results have been obtained with digested grass samples. TABLE IV RECOVERY OF SELENIUM ADDED TO BARLEY GRAIN Selenium determined, pg Mean Percentage recovery Barley grain (5 g) . . . . 0.059, 0.063, 0.063 0.062 - + Selenium, pg 0.05 . . 0.111, 0.118, 0.118 0.116 108 0.10 . . 0.146, 0.146, 0.150 0.147 86 0.20 . . 0.263, 0.246, 0.242 0.250 94 0.30 . . 0.325, 0-334, 0.380 0.336 91 0.40 . . 0,446, 0-438, 0.434 0-439 94 0.50 . . 0,555, 0-534, 0.558 0.549 97 The replication of the method with a mixed grass sample and Bowen's standard kale,18 when ten simultaneous determinations were carried out on each sample, is shown in Table V. TABLE V REPLICATION OF SELENIUM DETERMINATION Fluonmeter readings 1. r \ 84 70 86 69 85 69 84 65 83 66 85 71 87 71 89 69 92 71 100 73 Mixed dried grass Dried kale Mean ... . 87.4 Mean .. . . 69.4 Standard deviation 5.2 Standard deviation 2.4 DISCUSSION Of the problems associated with the determination of small amounts of selenium, that of loss during digestion of biological samples has probably attracted most attention, resulting in a multiplicity of digestion reagents and procedures being proposed. Several of the digestion methods suggested in the more recent publications were tried in this laboratory, including those incorporating catalysts such as di~hromate,~ molybdate8Jg and vanadate.20 Despite the most careful adherence to the recommended procedures, we continued to obtain low recoveries for 0.05 to 1.0 pg of selenium, either alone or with 5 or 10 g of plant sample. It is also worth noting that, in our hands, the use of molybdate and vanadate as catalysts298 HALL AND GUPTA: DETERMINATION OF VERY [Analyst, Vol.94 produced unexpected and potentially hazardous effects. With these reagents the digestions appeared to progress smoothly, when suddenly a violent reaction occurred, frequently with ejection of the digest from the flask. (Ewan, Baumann and Pope8 mention the possibility of explosions in the use of molybdate.) We have been reluctant to use the oxygen flask because of the large amount of organic matter to be combusted. Nor is the use of this particular procedure conducive to the analysis of large numbers of samples. In his detailed study of digestion methods for trace-element analysis, Gorsuch21 ex- perienced serious losses of selenium at the 1 p.p.m.level when using a mixture of nitric and sulphuric acids, but obtained 100 per cent. recovery with a mixture of nitric and perchloric acids. We formed the opinion that the losses of selenium from our samples were caused by prolonged boiling with perchloric acid at elevated temperatures. We studied the temperatures as the digestion progressed, in much the same way as Smith,22 although his observations did not relate to the determination of selenium. In the first stage, after the destruction of most of the organic material by nitric acid and hydrogen peroxide, the temperature remains at 110" to 120" C, when most of the nitric acid boils away. The temperature then rises rapidly to about 150" C when nitrate is decomposed, as seen by the production of nitrogen dioxide.Another rise in temperature to between 170" and 180' C follows, when charring of the small amount of organic material (mainly wax), which is not destroyed by the nitric acid and peroxide, may take place. After 1 or 2 minutes, the digest clears and perchloric acid boils off at 195" to 200°C. Often there is no charring at all. Immediately after the charring has cleared, the digest still has a slight straw-coloured appearance. When the digest was neutralised with ammonia solution at this point and complexed with DAN, we obtained highly coloured dekalin extracts that seriously interfered with the fluorimetry. To clear the digest completely and reduce its volume to about 2 ml often required as long as 1 hour, the temperature rising as high as 230" C, with subsequent losses of selenium.It seemed, therefore, that the terminal stage of the digestion with per- chloric acid should be controlled within a maximum fixed temperature range and in the shortest time. We have found that with a 5-g sample and 5 ml of perchloric acid the time required from the beginning of the charring to complete clarification of the digest vanes from 4 to 10 minutes, with an average of 6 to 7 minutes, no attempt being made to reduce markedly the volume of the digest. Silicone oil with a high viscosity and high flashpoint was found suitable for the bath; the use of ordinary high boiling-point oils would be dangerous. No difficulties have been experienced with the silicone chosen, but for complete control of the digestion it is prudent to work with small numbers at this stage.The levels of selenium shown in Table I11 clearly demonstrate the advantage of the controlled digestion. The use of the modified Kjeldahl-type flask and adapter described above enables digestion of the samples, complexing of the selenium and extraction of the 4,5-benzopiaz- selenol to be completed with the minimum of exposure to glass-surface area and the smallest number of transfer operations. The numbers of samples digested at any one time depends on the size of the hot-plate; we normally handle up to thirty flasks in the preliminary stages of the digestion. Like Watkinson' we found it unnecessary to separate the selenium from the digests of our plant samples by co-precipitation with arsenic.Such a step was required when forming the piazselenol with 3,3'-diaminobenzidine for spectrophotometry, but serious interference in the measurement of fluorescence was experienced with the arsenic co-precipitation and was presumably caused by impurities. In the course of this investigation, attempts were made to complex the selenium with DAN, with a more precise control of the pH using a conventional buffer system. In an acetate - hydrochloric acid buffer (0.33 M), formation of the 4,5-benzopiazselenol was found to be optimal at pH 1-1 for a standard of selenium in 5 ml of perchloric acid neutralised with ammonia solution and containing the EDTA stabilising reagent. However, when this technique of pH adjustment was applied to plant digests, low recoveries were obtained, and it was established that for these the optimum pH was 2.3 to 2.5.No detailed studies have been found in the literature on the 4,5-benzopiazselenol formation in various buffer systems. Parker and Harveyll concluded that maximum formation of the complex takes place a t pH 2.0 in 0.01 M hydrochloric acid, while Lott, Cukor, Moriber and S01ga~~ measured maxi- mum fluorescence when complexing was effected at pH 2.2, which approximates closely to ourApril, 19691 SMALL AMOUNTS OF SELENIUM IN PLANT SAMPLES 299 own observations for actual samples. Further efforts to control the pH at which the 4,5-benzo- piazselenol is formed were unsuccessful, and we have adopted the usual buffer containing formic acid and adjusting the pH with the help of narrow-range indicator papers.When an internal indicator, such as cresol red, was used we found that the pH was not close enough to the optimum for complexation. Further studies on this aspect of the use of DAN for the determination of selenium are worthy of attention. With the exception of one sample, it was not possible to relate our findings to those which would have been obtained by one of the accepted reference techniques in which radio- nuclides are used. Throughout all the work described in this paper Bowen's standard kalels was used as a reference sample. Bowen18 reported that this material had been subjected to multiple analyses by fluorimetric and neutron-activation procedures, the results for selenium being 0.139 More than sixty selenium determinations with this material have now been carried out by the procedure described in this paper, the range of results being 0.117 to 0.146 p.p.m., with a mean of 0.127.The value of 0.146 was obtained once only, the next highest being 0.136. The mean of 0.127 p.p.m. is somewhat lower than that reported by Bowen,l* by both neutron activation and fluorimetry, and it is difficult to explain the differences or the extent of the range in which more than 90 per cent. of the results were below 0.136 p.p.m. The recovery of selenium added to dried barley grain and grass was, however, encouragingly high. It may be that some selenium had been lost during storage (the material has been kept for the past 3 years in a polythene screw-capped jar at -20" C, but with periods at room temperature). Several worker^^^^^ have reported losses of selenium from stored samples, and Lewis, Johnson and Delwiche26 have shown that volatile selenium compounds can be released from both selenium accumulator and non-accumulator plants. It has also been established by Challenger, Lisle and Drans- and by Dransfield and Challenger,28 that dimethyl selenide is a volatile metabolite of certain fungi, including Aspergillus niger.The method proposed above is now being used to evaluate the selenium content of a wide range of pastures and animal feeds, the details of which we hope to publish in due course. We are indebted to Dr. H. J. M. Bowen, of Reading University, for his generous gift of standard kale. We are grateful to Mr. N. Trinder, Regional Nutrition Chemist, National Agricultural Advisory Service, Newcastle upon Tyne, for much helpful discussion and interest throughout this work.0.004 and 0-155 & 0.0143 p.p.m., respectively. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. REFERENCES Schwarz, K., and Foltz, C. M., J. Amev. Chem. Soc., 1957, 79, 3292. Patterson, E. L., Milstrey, R., and Stokstad, E. L. R., PYOC. SOC. EX$. Bid. Med., 1957, 95, 617. Muth, 0. H., Oldfield, G. E., Rernmert, L. F., and Schubert. J. R., Science. 1958, 128, 1090. Kubota, J., Allaway, W. H., Carter, D. L., Cary, E. E., and Lazar, V. A., J. Agric. Fd Chem. Allaway, W. H., and Cary, E. E., Analyt. Chem., 1964, 36, 1359. Watkinson, J . H., Ibid., 1960, 32, 981. Ewan, R. C., Baumann, C. A., and Pope, A. L., J. Agric. Fd Chem., 1968, 16, 212. Lindberg, P., Acta Vet. Scand., 1968, Supp. 23. Cheng, K. L., Analyt. Chem., 1956, 28, 1738. Parker, C. A., and Harvey, L. G., Analyst, 1962, 87, 558. Bowen, H. J. M., and Cawse, P. A., Ibid., 1963, 88, 721. Kelleher, W. J., and Johnson, M. J., Analyt. Chem., 1961, 33, 1429. Lane, J . C., Ir. J. Agric. Res., 1966, 5, 177. Stanton, R. E., and McDonald, J., Analyst, 1965, 90, 497. Molloy, J. P., Ir. J . Agric. Res., 1967, 6, 133. Hoste, J., Analytica Chim. Acta, 1948, 2, 402. Bowen, H. J. M., Analyst, 1967, 92, 124. Cummins, L. M., Martin, J. L., Maag, G. W., and Maag, D. D., Analyt. Chem., 1964, 36, 382. McNulty, J. S., Ibid., 1947, 19, 809. Gorsuch, T. T., Analyst, 1959, 84, 135. Smith, G. F., Analytica Chim. Acta, 1957, 17, 175. Lott, P. F., Cukor, P., Moriber, G., and Solga, J., Analyt. Chem., 1963, 35, 1169. Monon, A. L., and Rhian, M., Proc. S. Dak. Acad. Sci., 1938, 18, 20. Beath, 0. A., Eppson, H. F., and Gilbert, C. S., Wyo. Agr. Sta. Bull., 1935, 206, 1. Lewis, B. G., Johnson, 0. M., and Delwiche, C. C., J. Agric. Fd Chem., 1966, 14, 638. Challenger, F., Lisle, D. B., and Dransfield, P. B., J. Chem. Soc., 1954, 1760. Dransfield, P. B., and Challenger, F., Ibid., 1955, 1153. 1967, 15, 448. -, Ibid., 1966, 38, 92. Received November 18th, 1968 Accepted January 2nd, 1969

ISSN:0003-2654    

DOI:10.1039/AN9699400292

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

300 Analyst, April, 1969, Vol. 94, @. 300-305 The Determination of Fluorides in Preservative-treated Spruce and Redwood BY A. I. WILLIAMS (Forest Products Research Laboratory, Princes Risborough, Aylesbury, Bucks.) A method is described for the rapid determination of fluoride in wood that has been treated with preservatives containing sodium fluoride or potassium hydrogen difluoride. The fluoride is leached from thin sections of wood with sodium hydroxide solution and determined by spectrophoto- metric measurement of the blue complex formed between fluoride ions and the red lanthanum chelate of alizarin fluorine blue. The procedure is particularly useful for the study of the distribution of fluoride-containing preservatives in wood. SODIUM fluoride and potassium hydrogen difluoride are used in some commercial preservative formulations for pre-treating timber or for remedial treatment of telephone or power trans- mission poles in which there is internal decay.In this country, remedial preservation treatments are used to increase the service life of poles that have been poorly creosoted and attacked prematurely by wood-destroying fungi in the untreated interior at or near the ground-line. The preservation process consists either of spreading a fluoride-containing paste over the surface of the wood in the area of attack, and covering with a waterproof bandage, or of the mechanical injection of the paste into the w0od.l In the course of time the preserva- tive diffuses into the uncreosoted interior of the pole and then, when the loading reaches the toxic threshold, further destruction of the wood by the fungi is stopped.There are various formulations of the preservative available commercially, and they contain either sodium fluoride or potassium hydrogen difluoride and some or all of the following compounds: sodium arsenate, sodium chromate, sodium borate and 2,4-&nitrophenol. Also, new timber is preserved by vacuum and pressure processes or by dip treatments involving the use of solutions containing these chemicals .2 To evaluate the effect that remedial treatment with fluoride-containing preservatives has on telephone and power transmission poles, many sample borings need to be analysed to obtain statistically significant results. In the past, the time taken for analysis by the existing slow and tedious methods for determining fluoride, involving ashing and distillation procedures,2s3~4 has limited the number of samples that could be examined in surveys of this type of treatment.To enable extensive investigations to be carried out on the loading achieved by treatment, and the permanence and distribution of the leachable fluoride in the remedially treated poles, a rapid and accurate method of analysis is required for determining 0.005 per cent. upwards of fluoride. Previous work in this laboratory has shown that unfixed preservation chemicals can be rapidly and quantitatively leached from thin sections of Sitka spruce, Picea sitchensis, a relatively impermeable species, was used for the experimental work. The low permeability did not cause any apparent difficulties during the leaching out of fluoride ions when sections about 0.2 to 0.3mm thick were used.The lanthanon chelates of alizarin fluorine blue are possibly the only reagents that give a direct colour reaction with fluoride ions suitable for the spectrophotometric determination of fluoride. These reagents were first described by Belcher, Leonard and Although they are specific for fluoride ions, other ions interfere with the colour reaction by forming complexes or precipitates with the lanthanon ions, or by competing with the lanthanon ions for the available fluoride and the alizarin fluorine blue. Chromium is the only element present in the preservatives that can cause serious interference in the colour-forming reaction. 0 SAC; Crown Copyright Reserved.WILLIAMS 301 Chromium is applied to the wood as chromate, but it is reduced to tervalent chromium, which causes interference.Oxidation of the tervalent chromium ions by leaching the preserva- tive out of the wood with sodium hydroxide and hydrogen peroxide, and the action of chromate ions on the colour-forming reaction, were examined. The action of arsenate ions, which is reported to cause minor interference, was also examined. EXPERIMENTAL To examine techniques for determining fluoride in wood it was necessary to prepare standard samples containing a known amount of preservative. This was achieved by im- pregnating wood, by the full cell process,2 with solutions of known concentrations of two formulations of preservative. Preservative A contained 33 per cent.of sodium fluoride, 25 per cent. of sodium arsenate and 42 per cent. of sodium chromate. Preservative B contained 25 per cent. of sodium fluoride, 25 per cent. of sodium arsenate, 37.5 per cent. of sodium chromate and 12.5 per cent. of 2,4-dinitrophenol. The impregnated wood was freeze-dried8 to prevent re-distribution of preservative during drying. PREPARATION OF STANDARD SAMPLES- Weighed quarter-sawn blocks of Sitka spruce of known moisture content, with over-all dimensions of 30 x 20 x 10 mm, cross-section 30 x 10 mm, radial face 30 x 20 mm and tangential face 20 x 10 mm, were submerged in solutions of known concentrations of either preservative A or B and subjected to a vacuum of 94.65 kNm-2 for half an hour to evacuate air from the wood cells.The vacuum was released, and a pressure of 686-5 kNm-2 was applied for half an hour to the solution containing the specimens. The pressure was released, and the specimens removed from the solution , their surfaces being superficially dried on filter-paper to remove excess of solution, and weighed. The wood absorbed about three times its own weight of solution. To prevent or minimise re-distribution of preservative the specimens were freeze-dried to a moisture content of about 7 per cent. From the increase in weight the percentage of fluoride, based on oven-dry wood, dried at a temperature of 105" C, was found by calculation to range from 0.013 to 0.361 per cent., the range in which the amount of fluoride is normally found in preserved wood. The structure of the wood in the small blocks used for the experiment varies.In pre- served timber, concentration gradients of deposited preservative can occur across the growth rings; more will be found in the spring or early wood, as the void space is greater than in summer or late wood. Therefore, for development work on the procedure, it was decided to use thin radial or cross-sections cut over several growth rings to obtain samples containing a representative amount of preservative in the bulk of the wood. DEVELOPMENT OF THE PROCEDURE- The recovery of fluoride from solutions containing sodium arsenate, sodium borate, sodium chromate and 2,4-dinitrophenol in concentrations in excess of those normally present in treating solutions and preserved wood was examined. Synthetic treating solutions contain- ing an accurately measured amount of fluoride were allowed to react with lanthanum - alizarin fluorine blue reagent solution.Comparison of the optical densities of these solutions with a calibration graph showed that complete recovery of the fluoride was achieved. Following this observation, a leaching reagent was sought that would leach out the preservative from the wood and at the same time oxidise chromium(II1) ions to chromate and, therefore, eliminate interference from this source. Previous work on chromium-containing preservatives showed that chromium(II1) ions in preserved wood were oxidised by hydrogen peroxide in 2 N sodium hydroxide solution , and unfixed chemicals were rapidly leached out of thin sections of wood by this solution.When sodium hydroxide solutions containing known amounts of fluoride were heated in Pyrex beakers , low and erratic recoveries of fluoride resulted, and when the solutions were not heated the correct amount of fluoride was found. The low recoveries probably resulted from interference by aluminium and calcium dissolved from the glass by hot sodium hydroxide solution. Because of this the leaching stage of the procedure was carried out in nickel crucibles. Although nickel ions are known to interfere with the formation of the fluoride complex, insufficient were taken into solution to cause any interference. After302 WILLIAMS : DETERMINATION OF FLUORIDES I N [Analyst, Vol. 94 leaching, the wood was removed from the solution and the solution neutralised with a pre- determined volume of 2 N acetic acid solution.As the optical density of the fluoride complex is dependent on the concentration of sodium acetate in the test solution, the amount of sodium acetate in the reagent solution was adjusted to allow for that in the neutralised leaching solution. Measurements of optical densities of the test solutions were made at wavelength 620 nm with a Unicam SP600, with 10-mm cells. The spectrophotometric calibration graph was constructed in the range 0 to 40 pg of fluoride and a straight-line relationship was obtained. It was found necessary to prepare a new calibration graph for each new batch of reagents. The reagent solution showed no deterioration when allowed to stand in shaded daylight for periods up to 2 weeks. This was not the maximum life of the reagent solution, but the longest time any one batch lasted.It was observed that the presence of wood in the final solution caused low fluoride recoveries. Therefore, the leaching solution was decanted off from the wood before the formation of the fluoride complex. Telephone and power transmission poles are usually pressure impregnated with creosote before being put into service. If, during creosoting, pockets of sapwood remain untreated, fungal attack may take place and then the wood will need remedial treatment to slow down the decay and increase the service life of the pole. The samples taken for the determination of fluoride may, therefore, contain some creosote. The effect of creosote on the recovery of fluoride was investigated by adding solutions of known fluoride concentration to thin sections cut from 4.5 mm diameter borings of creosoted Scots pine, Pinus syZvestris, and the fluoride determined by the proposed procedure.Although the leaching solution was coloured pale yellow, no interference in the fluoride recovery was experienced. Following this, the effect of creosote on the extraction of fluoride ions from creosoted wood was investigated. Two adjacent thin sections were cut from borings taken from different, remedially treated poles. One section from each boring was leached with chloroform to extract the creosote, dried and analysed, together with the remaining sections containing creosote, by the proposed procedure. The results given below showed that the presence of creosote in the wood did not interfere with the recovery of fluoride ions.Fluoride, per cent. r h t Boring Creosote extracted Creosote present 1 0.163 0.164 2 0-128 0.132 3 0.134 0.126 After the method had been developed, with samples 0.1 mm thick, the maximum thick- ness of wood from which the total preservative content could rapidly be recovered was investigated. Adjacent radial or cross-sections of increasing thickness, up to a maximum of 0.3 mm, were cut on a microtome and analysed by the proposed procedure. The results showed that complete recovery of fluoride in preservative was attained with radial sections up to 0.2 mm and cross-sections up to 0.3 mm thick. SAMPLING- Radial or cross-sections, cut across several growth rings, are taken for analysis to obtain samples representative of the whole piece of wood.To prepare samples from pressure-impregnated or dip-treated bulk wood, small speci- mens are sawn, and sections up to a maximum of 0.3 mm cut on a microtome. To determine the lateral distribution in remedially treated poles, Mattson borings, 4.5 mm in diameter, are taken and specimens for sectioning are cut from the borings corresponding to different depths in the poles. RESULTS The procedure described was applied to thin radial and cross-sections of the standard samples, and the results are given in Table I. Six thin radial sections up to 0.2 mm thick, or cross-sections up to 03mm thick, were taken at evenly spaced intervals through the standard samples. The results showed some small variation through the specimens, but the average of the six determinations was in agreement with the calculated over-all fluoride content. The standard deviation, based on twelve determinations at the 0.1 per cent. fluoride level, was ~ 0 .0 0 4 per cent.April, 19691 PRESERVATIVE-TREATED SPRUCE AND REDWOOD 303 B A B A B A 0.061 0-060 0.091 0.093 0.117 0.171 0.262 0.361 TABLE I DISTRIBUTION AND LOADING OF FLUORIDE IN STANDARD SAMPLES COMPARED WITH LOADING CALCULATED FROM SOLUTION RETENTIONS Preservative Calculated fluoride content Fluoride formulation based on oven-dry weight, found, Average, per cent. per cent. per cent. B 0.013 0.012 0-011 0.01 1 0.01 1 0.010 0.01 1 0.01 1 0.066 0-060 0-058 0.053 0.062 0-059 0-095 0.091 0.095 0-091 0-094 0.093 0.116 0-113 0.1 12 0.116 0.111 0.111 0.1 13 0.179 0-169 0.176 0.173 0-175 0.160 0.152 0.27 1 0-265 0.260 0-269 0.262 0-272 0.256 0.374 0.359 0.360 0.362 0.362 0.342 0.351 The results obtained by analysing sections cut from borings taken from creosoted, remedially treated poles with the proposed leaching procedure were confirmed by ashing the remains of the samples with calcium oxide, distillation of fluoride as fluorosilicic acid (H,SiF,) from perchloric acid solution and colorimetric determination with lanthanum - alizarin fluorine blue reagent.The results given in Table 11, obtained by the two methods, were in good agreement. TABLE I1 COMPARISON OF RESULTS OBTAINED BY USING THE LEACHING PROCEDURE WITH THOSE OBTAINED BY ASHING AND DISTILLATION Fluoride content, Pole No. Procedure per cent. 1 Leaching . ... .. .. 0.056 Ashing and distillation . . .. 0-054 2 Leaching . . .. .. .. 0.076 Ashing and distillation . . .. 0.068 3 Leaching . . .. .. .. 0.081 Ashing and distillation . . .. 0.083304 WILLIAMS : DETERMINATION OF FLUORIDES IN [ANfZlySt, VOl. 94 To demonstrate the usefulness of the proposed method, the distribution and depth of penetration of fluoride in western hemlock, Tsuga heterophylla, that had been commercially preserved by spray or dip treatment with a fluoride - chromate solution were investigated. The preservative does not penetrate very deeply into the wood with this technique and, therefore, it was necessary to take samples for analysis at close intervals, until no fluoride was detected, to form a distribution pattern. In the outer 2.5 mm of the radial face of the specimen, samples, 0.1 mm thick, were cut on a microtome from the surface at a depth of 0-3 mm, and then at intervals of 0.1 mm.The curve obtained by plotting the fluoride content against depth is shown in Fig. 1. The whole analysis, including sample cutting, twelve deter- minations and plotting the distribution curve, was completed within 2 hours. I Depth from surface, mm Fig. 1. Distribution of fluoride in dip-treated western hemlock METHOD REAGENTS- Use reagents of analytical-reagent grade when possible. Lanthanum - alizarin juorine blue solution-Dissolve 0-2395 g of alizarin fluorine blue in 5 ml of a solution containing 20 per cent. w/v of ammonium acetate and 10 per cent. v/v of-ammonia solution (sp.gr. 0.88) and 5 ml of water. Filter the solution through a well washed, small Whatman No.1 paper (to ensure that all of the reagent is in solution) into a 1-litre graduated flask containing 40.80 g of hyrated sodium acetate (CH3.COONa.3H,0), 30 ml of glacial acetic acid and sufficient water to dissolve the solids. Wash the filter-paper with a small volume of water. Add 300 ml of isopropyl alcohol and mix the contents of the flask. Dissolve 0.2040 g of lanthanum oxide in 12.5 ml of 2 N hydrochloric acid, warming gently to aid dissolution, and add to the flask, with mixing. Dilute to the calibration mark with water and mix. Sodium hydroxide, 2 N-Dissolve 16.0 g of sodium hydroxide in water and dilute to 200 ml with water. Standardise and adjust the solution to exactly 2 N. Store in a poly- thene bottle.Acetic acid, 2 N-Ddute 23.0 ml of glacial acetic acid to 200 ml with water. Determine, by titration, the volume of this solution required to neutralise 1 ml of the 2 N sodium hydroxide solution, with phenolphthalein as indicator. Hydrogen peroxide, 60 volumes.April, 19691 PRESERVATIVE-TREATED SPRUCE AND REDWOOD 305 PROCEDURE- Transfer the weighed sample to a nickel crucible; from a pipette, add 1 ml of 2 N sodium hydroxide solution and then 1 drop of hydrogen peroxide (50 volumes) solution. Warm the solution, at about 130°C, on a hot-plate until the hydrogen peroxide has decomposed (2 to 3 minutes) and continue heating for 5 minutes, but do not allow the solution to evaporate to dryness; maintain the volume of solution at about 1.Oml by small additions of water.Cool, rinse the crucible wall down with 5 ml of water, allow to stand for 5 minutes, decant the solution into a 25-ml graduated flask, wash the wood and crucible with small volumes of water and decant the washings into the flask. Neutralise the solution with the pre- determined volume of 2 N acetic acid solution, add, with mixing, 10ml of lanthanum- alizarin fluorine blue reagent solution, dilute to the calibration mark with water and mix. Allow the solution to stand for half an hour and then measure the optical density of the blue fluorine complex against a reagent blank, prepared in a similar way, in 10-mm cells at wave- length 620nm on a Unicam SPW spectrophotometer. To obtain the fluoride content in the test solution compare the spectrophotometer reading with a calibration graph.CALIBRATION- Preparation of standard solution A-Dissolve 0.1105 g of sodium fluoride in water. Transfer the solution to a 100-ml graduated flask, dilute to the mark with water and mix. Store in a polythene bottle. Prefiaration of standard solution &-Transfer, by pipette, 10 ml of standard solution A into a 100-ml graduated flask, dilute to the mark with water and mix. Store in a polythene bottle. Transfer, by pipette, aliquots of 0.1, 0.2, 0.3, 0.4, 0.6 and 0.8 ml of standard solution B to nickel crucibles. The aliquots taken contain 5, 10, 15, 20, 30 and 40 pg of fluoride, respectively. Continue as described under Procedure. Plot the optical densities against micrograms of fluoride to obtain the calibration graph. 1 ml of standard solution A = 500 pg of fluoride. 1 ml of standard solution B = 50 pg of fluoride. This paper is published by permission of the Ministry of Technology. REFERENCES 1. Van Groenou, H. B., Rischen, H. W. L., and Van Den Berge, J., “Wood Preservation During the Last 50 Years,” Second Edition, A. W. Sijthoff’s Uitgeversmaatschappij N.V., Leiden, Holland, 1952, pp. 40 and 35. 2. British Standard 3463 : 1962. 3. Oelschlager, W., 2. analyt. Chem., 1962, 191, 408. 4. Gersonde, M., and Kattlors, C., HoZz Roh- u. Werkstoff, 1961, 19, 346. 5. Williams, A. I., Analyst, 1968, 93, 111. 6. Belcher, R., Leonard, M. A., and West, T. S., J . Chem. SOL, 1952, 2390. 7 . - - - , Talanta, 1959, 2, 92. 8. Smith, D. N. R., and Cockcroft, R., Nature, 1961, 189, 163. Received August 27th, 1968 Accepted September 13th, 1968

ISSN:0003-2654    

DOI:10.1039/AN9699400300

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

306 Analyst, April, 1969, Vol. 94,++. 306-307 Interference in the Determination of Mercury in Plant Materials with Dithizone Due to 2-Chloro-2-nitrosopropane BY D. F. LEE AND J. A. ROUGHAN (Ministry of Agriculture, Fisheries and Food, Plant Pathology Laboratory, Hatchhg Green, Har$enden, Herts.) 2-Chloro-2-nitrosopropane, formed from acetone used to clean glassware, has been identified as a cause of interference in the determination of mercury with dithizone. THE determination of mercury in plant materials by the method recommended by the Joint Mercury Residues Panell depends on acid digestion of the plant material, followed by extraction of the mercury(I1) ions formed with a solution of dithizone in chloroform. The mercury is separated from co-extracted cations by back-extracting the dithizone solution with aqueous sodium thiosulphate solution. The mercury - thiosulphate complex is destroyed with sodium hypochlorite, the excess of hypochlorite and free chlorine allowed to react with hydroxylamine, and the mercury(I1) ions extracted into a standard volume of a standard solution of dithizone in chloroform.The net increase in absorbance of the standard dithizone solution at 490nm, caused by the formation of mercury dithizonate and the removal of dithizone, is then measured. EXPERIMENTAL In the Joint Mercury Residues Panel’s paper it is emphasised that all apparatus must be scrupulously cleaned, and it is recommended that the digestion apparatus should be pre- extracted with boiling acids, but no recommendations are given for the cleaning of other glassware used.We adopted a regime of soaking all glassware not in use in acidified potassium permanganate solution, followed by a brief soaking in dilute hydrogen peroxide solution and rinsing with distilled water just before use. It was difficult to wet the surface of vessels that had contained chloroform solutions when rinsing with distilled water between processing successive samples, so they were rinsed first with acetone. When using this method we experienced difficulty in obtaining reproducible recoveries in the 1 to 5-pg range, and sometimes obtained negative absorbance changes when no mercury(I1) ions had been added to the plant material. We observed that the low recoveries and negative blank values were associated with the occurrence of a blue colour in the chloro- form wash solution following the hypochlorite - hydroxylamine stage of the reversion process.The formation, but no explanation, of this blue colour was reported in the description of the method (ref. 1, page 614). We found that the blue colour was formed only when acetone was used to rinse the glassware. The compound responsible for the blue colour was identified by gas - liquid chromato- graphy and mass spectrometry as 2-chloro-2-nitrosopropane, formed by the reaction of chlorine evolved from the acidified hypochlorite with acetoxime2 derived from acetone used to rinse the glassware and hydroxylamine added to the acid digest of the plant material. When the blue compound is formed, even when the aqueous phase is given additional washings with chloroform, the absorbance due to dithizone is reduced, so that in the absence of mercury(I1) ions a negative net gain in absorbance is observed.In the presence of mercury(I1) ions the net gain in absorbance is less than expected because of the diminution of the dithizone component and, in extreme instances, because there is insufficient dithizone to form a complex with all of the mercury(I1) ions. 0 SAC; Crown Copyright Reserved.LEE AND ROUGHAN 307 2-Chloro-2-nitrosopropane bleaches the dark green colour of dithizone solution to a pale yellow. As an example of this effect, 1 O r n l of a chloroform solution of dithizone diluted with 1 ml of chloroform had an absorbance against chloroform of 0.465 at 490 nm; the same dithizone solution diluted with 1 ml of a 0.1 per cent. solution of 2-ehloro-2-nitrosopropane in chloroform had an absorbance against chloroform of 0-350, 1 minute after mixing, and 0.210, 12 minutes after.Simple carbonyl compounds, other than acetone,3 also readily form chloronitroso de- rivatives by the reaction of chlorine with solutions containing the carbonyl compound and hydroxylamine, so that it is essential when using the method of the Joint Mercury Residues Panel to ensure that all digests, reagents and apparatus are free from ketonic and aldehydic materials. The mass-spectrographic investigation was undertaken by Dr. D. H. Calam of the Rothamsted Experimental Station, to whom the authors express their thanks. REFERENCES 1 . Report by the Joint Mercury Residues Panel of the Advisory Committee on Poisonous Substances Used in Agriculture and Food Storage, the Analytical Methods Committee and the Association of British Manufacturers of Agricultural Chemicals, Analyst, 1961, 86, 608. Piloty, O., Ber. dt. chem. Ges., 1898, 31, 462. Piloty, O., and Stock, A., Ibid., 1902, 35, 3093. 2. 3. Received June 14th, 1968 Accepted October Blst, 1968

ISSN:0003-2654    

DOI:10.1039/AN9699400306

出版商:RSC    

年代:1969

数据来源: RSC

摘要:

308 Analyst, April, 1969, Vol. 94, #q5. 308-312 A Nephelometric Method for the Determination of Sulphate in the Presence of Barium by Using Chloroaminobiphenyl: Application to Organic Compounds and Hydrocarbon Oils BY A. E. MENDES-BEZERRA" AND P. C. UDEN (Department of Cltemistry, The University, P.O. Box 363, Birmingham 15) The method described for the determination of sulphate in the presence of barium can also be extended to the determination of sulphur in organic compounds and hydrocarbon oils containing barium. After oxygen-flask combustion of the sample, ammoniacal EDTA can be added directly to the solu- tion and the total sulphate then determined nephelometrically after precipitation with 4-chloro-4'-aminobiphenyl. Alternatively, free sulphate can be determined separately by direct precipitation as chloroaminobiphenyl sulphate after filtering off precipitated barium sulphate ; the latter can then be dissolved in ammoniacal EDTA solution and the sulphate also determined as chloro- aminobiphenyl sulphate.THE determination of sulphur in organic materials containing barium presents no problem when the sample is decomposed in a tube in a stream of oxygen, by using any of the variants of this well established method of combustion analysis. When the oxygen-flask combustion procedure is used, however, the absorbing solution contains a mixture of barium sulphate and sulphate ion (if sulphur is present in excess) and the determination then becomes more complicated. For moderate amounts of barium and sulphur, for example, those arising from the decomposition of a pure organic compound, it might be sufficient to filter off the barium sulphate and to determine the two components separately. Other possibilities are also readily apparent.The main problem arises when only trace amounts of sulphur and barium are present, and this occurs frequently in the petroleum industry where lubricating oils may contain barium in addition to sulphur. A simple method for the determination of trace amounts of sulphate ion has recently been described by this department .I92 Sulphate is precipitated with chloroaminobiphenyl (CAD) hydrochloride, followed by nephelometric measurement ; the reaction is more sensitive than with barium. This method of measurement should be suitable for the determination of traces of sulphate after flask combustion, but it was not known if it was applicable in the presence of barium.We thought it might be possible to determine the amount of barium sulphate by dissolu- tion in alkaline EDTA solution3 followed by the addition of chloroaminobiphenyl reagent, but certain preliminary tests were necessary, and two problems had to be considered. It is known that barium sulphate re-precipitates from its solution in EDTA on acidifica- tion and chloroaminobiphenyl hydrochloride cannot be added to an alkaline solution because the free base precipitates. As it was not known at which limiting pH barium sulphate remains in solution this had to be established. Provided that this was not more than pH 7, it should be possible to add chloroaminobiphenyl hydrochloride without risk of precipitating the free base.If these conditions could be fulfilled, it would then be necessary to establish whether or not the precipitate formed would be entirely chloroaminobiphenyl sulphate. In the presence of competing barium ions, there was the possibility that some barium sulphate might also be formed. * On study leave from the Institute of Chemistry, Federal University of Ceara, Brazil. 0 SAC and the authors.MENDES-BEZERRA AND UDEN 309 EXPERIMENTAL APPARATUS- Nephelometric measurements were made with the EEL Nephelometer head in conjunc- tion with the EEL Unigalvo, Type 20. Test-tubes ($ inch diameter, volume about 25 ml) provided with the instrument were used. REAGENTS- CAD* hydrochloride sohtiod-(Obtainable from Burdick and Jackson Laboratories Inc., Muskegon, Michigan 49442, U .S.A.) Two grams of 4-chloro-4'-aminobiphenyl hydrochloride were dissolved in 800 ml of hot (70" C) 0.05 N hydrochloric acid. The solution was cooled and filtered through a Whatman No.40 paper into a dark bottle. Gum ghatti solutiod-This was prepared by dissolving 0.3 g of finely ground gum ghatti in 200 ml of 0-05 N hydrochloric acid by warming to 70" C. The solution was cooled and undissolved material filtered off on a double Whatman No. 40 paper. The solution was stored at 0" to 5" C and discarded at the first signs of the formation of small particles. OPTIMAL CONDITIONS FOR PRECIPITATION- A suspension of barium sulphate was prepared by mixing equivalent volumes of 0.1 M potassium sulphate and barium chloride.The resulting barium sulphate suspension was dissolved in 10 ml of 0.1 M EDTA plus 2 to 3 drops of 0.88 ammonia solution, dissolution occurring on standing for a few minutes. The solution was warmed to expel the excess of ammonia and the pH was adjusted to 6.4 to 6.5 by the addition of N hydrochloric acid. At this pH, precipitation of barium sulphate was delayed indefinitely. (At pH 6.2 the solution became turbid after 2 hours.) COMPETITION BETWEEN Ba2+ AND CAD FOR SOa2--- When 10 ml (excess) of CAD hydrochloride solution, obtained as described previously, were added to the solution at pH 6, a white precipitate was immediately produced. Ignition of this precipitate at 800" C in a porcelain crucible, after filtration through a Whatman No. 42 filter-paper, gave no residue, and at least 99 per cent. of the original barium present was recovered from the filtrate by re-precipitation as barium sulphate.Thus there is an optimal pH at which barium does not precipitate from EDTA solution, but which is not sufficiently alkaline to cause precipitation of the CAD base. Moreover, when CAD hydro- chloride is added to a "neutral" solution containing barium sulphate, all the sulphate is precipitated as CAD sulphate; there appears to be no induced precipitation of barium sulphate. NEPHELOMETRIC PROCEDURE- The method used for the determination of sulphate as the CAD sulphate suspension was a simplified version of that used by Martin and Stephen,2 as the use of a colloidal stabiliser such as peptone was not found advantageous. Ten millilitres of CAD hydrochloride solution were added to 10 ml of the sulphate solution, followed by 2 ml of gum ghatti solution, and the mixture shaken for 20 seconds, then allowed to stand for 20 minutes. After transferring to a nephelometer tube, which was inverted three times to produce a uniform suspension, the galvanometer reading was taken after the disappearance of air bubbles.A linear calibration graph, passing through the origin, was obtained of gradient 5 divisions per 10 pg of sulphate in the 10-ml sample (over a range 10 to 200 pg per 10 ml of sulphate). Calibration was made against a blank solution, with no sulphate present, subjected to the same procedure. Readings obtained were completely reproducible after 24 hours. For an individual determination of an unknown sample it was sufficient to reconstruct the calibration scale by preparing a standard suspension corresponding to the maximal (100) deflection on the galvanometer scale.This calibration was confirmed for several standard sulphate samples. DETERMINATION OF PRECIPITATED BARIUM SULPHATE- One millilitre each of 0.01 M barium chloride and 0.01 M potassium sulphate solutions were mixed to give a barium sulphate precipitate (SOZ- present 0.96 mg). The barium * The abbreviation CAD, derived from the former nomenclature, chloraminodiphenyl, is retained to avoid confusion.310 MENDES-BEZERRA AND UDEN : A NEPHELOMETRIC METHOD FOR [ATZaZySt, VOL 94 sulphate was dissolved in 5 ml of 0-1 M EDTA plus 2 drops of 0.88 ammonia solution, warmed to expel the ammonia and the pH adjusted to 6.5 with N hydrochloric acid. The solution was diluted to 50 ml with distilled water and aliquots determined by the standard nephelo- metric procedure, the total volume of the sulphate sample being made up to 10ml in each instance with distilled water before the addition of 10 ml of CAD hydrochloride and 2 ml of gum ghatti.To avoid possible interferences from an excess of EDTA, its final concentration should not exceed 0.01 M. Good correlation with the calibration graph was obtained within the +5 per cent. limit of instrumental accuracy. DETERMINATION OF PRECIPITATED BARIUM SULPHATE IN THE PRESENCE OF EXCESS OF FREE One millilitre of 0-1 M barium chloride and 2 ml of 0-1 M potassium sulphate were mixed in a 25-ml beaker and left to stand for 24 hours.The solution was then filtered under suction with a Pregl No. 4 micro filter. The precipitated barium sulphate was dissolved in 10 ml of 0.1 M EDTA plus 2 drops of 0.88 ammonia solution, the filter being shaken in a test-tube closed by a rubber stopper through which passed the stem of the filter. After standing for 30 minutes the ammonia was expelled by warming, the pH adjusted to 6-5 and the sulphate determined nephelometrically. The filtrate containing free sulphate was diluted to 50ml and the sulphate determined directly with CAD. Values obtained for both free sulphate and barium sulphate were within 3 per cent. of those expected. SULPHATE- FLASK-COMBUSTION DETERMINATIONS PURE ORGANIC COMPOUNDS- The nephelometric method was applied to the determination of sulphur in pure (micro- analytical standard) S-benzylthiuronium chloride containing 15.8 per cent.w/w of sulphur. The standard flask-combustion procedure with a 500-ml flask was used. Different amounts of barium chloride were introduced into the absorbing solution to produce differing ratios of barium sulphate to free sulphate in the final absorbate, thus simulating the conditions for samples containing both barium and sulphur. When no barium chloride was added, the absorbing solution consisted of 5 ml of distilled water, 0.25 ml of N hydrochloric acid and 8 to 10 drops of 100-volume hydrogen peroxide; 2 to 3 mg of benzoic acid were added to aid combustion. After combustion the flask was shaken for 10 minutes and the absorbing solution then heated to boiling before determining the free sulphate nephelometrically.Alternatively, when barium chloride was added to the absorbing solution, it was found necessary to use larger samples of the sulphur standard to ensure complete precipitation of barium ~ulphate.~ After combustion the solution was heated to boiling, cooled and filtered on a Pregl No. 4 micro filter to remove barium sulphate. The free sulphate present in the filtrate was determined directly with CAD, while the precipitated barium sulphate was dissolved in ammoniacal EDTA and sulphate determined nephelometrically after adjusting the pH to 6.5. The results are given in Table I. OIL SAMPLES- Three oil samples containing different amounts of both barium and sulphur were analysed. To ensure complete combustion of the sample a modification in the standard flask-combustion method was used. The oil sample was introduced into a weighed methylcellulose capsule by means of a capillary tube, the capsule being rotated after capping to spread the oil evenly over the inner surface and thus facilitate combustion.The capsule was wrapped in a filter- paper holder and enclosed in platinum jaws, the flask was filled with oxygen and the wick of the holder ignited separately. Just before the wick burned down, the jaws were squeezed to crush the capsule and the whole plunged into the flask, which was quickly inverted and swirled at an angle of 45" to ensure maximal circulation of oxygen. For the sample low in both barium and sulphur the absorbing solution consisted of 10 nil of distilled water, 0.25 ml of N hydrochloric acid and 8 to 10 drops of 100-volume hydrogen peroxide. Although no precipitate was observed, analyses for both barium sulphate and free sulphate were carried out separately in the manner previously outlined, to confirm that all the sulphur was present as free sulphate (Table 11).April, 19691 THE DETERMINATION OF SULPHATE IN THE PRESENCE OF BARIUM 31 1 For the samples with higher concentrations of barium and sulphur the absorbing solu- tion consisted of 5 ml of distilled water and 5 to 6 ml of 100-volume hydrogen peroxide, the large volume of peroxide being necessary to ensure complete oxidation of sulphur to sulphate.For these samples it was at first thought that although no precipitation was observed, some barium sulphate might remain, adhering to the gauze.Therefore after com- bustion the gauze was placed in a beaker containing 10 ml of 0.1 M EDTA and 3 drops of 0.88 ammonia solution, and warmed for 20 minutes to dissolve any adhering barium sulphate. The EDTA solution was then added to the absorbing solution together with 2-5 ml of N hydro- chloric acid and the mixture heated to boiling. After cooling, the solution was diluted to 100 ml in a graduated flask, a 10-ml aliquot was taken and sulphate determined nephelo- metrically. Good results for total sulphate were obtained and no interference from EDTA was observed. Subsequently, however, the EDTA dissolution step was omitted, and the determination of free sulphate in the absorbing solution was shown to give the theoretical recovery for total sulphate (Table 111).RESULTS AND DISCUSSION TABLE I AMOUNTS OF FREE SULPHATE AND BARIUM SULPHATE FOUND FROM S-BENZYLTHIURONIUM CHLORIDE COMBUSTION Volume of 0.1 M Free SO,2- as barium sulphate Weight of barium chloride r - sample, mg added, ml expected, mg found, mg expected, mg found, mg 0.606 0 0.29 0.28 - - 0,422 0 0.20 0.20 - - 9.503 0-25 2.1 2.0 2.4 2.5 8.991 0.50 - - 4.3 4.15 TABLE I1 AMOUNTS OF FREE SULPHATE AND BARIUM SULPHATE FOUND FOR AN OIL SAMPLE OF LOW SULPHUR AND BARIUM CONTENT Oil sample A. Quoted analysis: sulphur 0.52 per cent. w/w; barium 0-28 per cent. w/w Free SOf- SO,'- as barium sulphate Weight of 1 & sample, mg expected, mg found, mg expected, mg found, mg 6.663 0.09 1 0.105 0.013 - 6.484 0.075 0.090 0.01 1 - TABLE I11 AMOUNTS OF SULPHATE FOUND FOR OIL SAMPLES OF HIGH SULPHUR AND BARIUM CONTENT Quoted analysis5 r > Oil sample C : A Oil sample B : barium 10.5 per cent.w/w sulphur 3-0 to 3.3 per cent. w/w sulphur 2.9 to 3.3 per cent. w/w barium 5.2 per cent. w/w & r Weight of sample, mg . . .. . . 9-830* 8.9lOt 6.246* 11.732t Free SO4!?- expected, mg . . .. 0.17 0-15 0.34 0.63 Free found, mg . . .. .. 0.90 - 0-57 - SO,2- as barium sulphate expected, mg . . 0.72 0.65 0.23 0.43 SO,2- as barium sulphate found, mg . . 0 Total SO,2- expected, mg . . . . 0.89+, 0*80# 0.57$ 1-06+, Total found after EDTA dissolu- - 0 - tion, mg . . . . . . . . - 0.83 - 1-15 * SOO2- determined without addition of EDTA. t SO4'- determined after EDTA treatment of gauze. # SO4'- calculated for S = 3.0 per cent.312 MENDES-BEZERRA AND UDEN These results show that sulphate can be determined nephelometrically as CAD sulphate when the sulphate is present wholly or partly as precipitated barium sulphate, provided that the latter is dissolved in EDTA- ammonia solution and the pH adjusted to 6.4 to 6.5.When an excess of free sulphate is present this can also be determined quantitatively within the instrumental accuracy. Provided that the concentration of barium ions in the solution prepared or derived from flask combustion is high enough to ensure precipitation of barium sulphate, the latter can then be filtered off, re-dissolved and determined. When there is doubt about the completeness of precipitation, the EDTA treatment can be applied and the determination of total sulphate carried out without interference.For the oil samples analysed no difficulty is observed when the ratio of sulphur to barium is high as, for the amount of sample used, ionic strengths are below those needed to induce barium sulphate precipitation. All sulphur is thus determined as free sulphate. For those samples in which the proportion of barium to sulphur is higher, it appears that at the con- centration levels occurring in flask combustion, no appreciable amount of barium sulphate is precipitated and all the sulphur is present as free sulphate, provided that a larger amount of hydrogen peroxide is added to ensure complete oxidation of the sulphur. If there is any indication of precipitation in an individual combustion, the EDTA procedure may be followed as a safeguard, but if there is no indication of precipitation or deposition on the gauze, the EDTA treatment can be omitted.For all oil samples, the combustion technique used is of prime importance in ensuring complete combustion and recovery of sulphate. RECOMMENDED PROCEDURE- Weigh into a methylcellulose capsule a sufficient amount of sample to yield about 1 mg of total sulphate. The volume of 100-volume hydrogen peroxide to be added is determined by the proportion of barium present. Without precipitate of barium sulphate-After combustion shake the flask for 10 minutes, bring the absorbing solution to the boil and then dilute to 50 ml with distilled water. Take an aliquot of this solution, make up to lOml, add 10ml of CAD hydrochloride solution followed by 2ml of gum ghatti solution, shake it for 20 seconds and allow to stand for 20 minutes before measuring nephelometrically and comparing with a standard calibration graph. With precipitate of barium sdphate-If precipitated barium sulphate is evident, add 5 ml of 0.1 M EDTApZus 2 drops of 0438 ammonia solution. Warm to complete dissolution and expel the ammonia, then adjust the pH to 6.5 with N hydrochloric acid. Dilute the solution to 50 ml with distilled water and measure the sulphate nephelometrically as CAD sulphate. We thank Professor R . Belcher for his interest and advice, Dr. W. I. Stephen for dis- cussions on the use of CAD, and Mr. G. Turton for practical assistance on flask-combustion methods. We also thank the Esso Petroleum Company Limited for providing suitable oil samples. One of us (A.E.M.-B.) gratefully acknowledges receipt of a grant from the Conselho Nacional de Pesquisas, Brasil. REFERENCES 1. Martin, J. M., and Stephen, W. I., Analytica Chim. Acta, 1967, 39, 175. 2. - - , Ibid., 1967, 39, 625. 3. Belcier, R., Gibbons, D.. and West, T. S., Chem. G. Ind., 1954, 127. 4. LaMer, V. K., and Dinegar, R. H., J . Amer. Chem. SOC., 1951, 73, 380. 5. Tentative Method of Test for Sulphur in Petroleum Products. High Temperature Method, A.S.T.M. Received May 23rd, 1968 Accepted October 14th, 1968 Designation, D1552-58T, 1958.

ISSN:0003-2654    

DOI:10.1039/AN9699400308

出版商:RSC    

年代:1969

数据来源: RSC