782 Analyst, November, 1973, Vol. 98, pp. 782-791 Spectrophotometric Determination of Low Levels of Mono-, Di- and Triethylene Glycols in Surface Waters BY W. H. EVANS AND A. DENNIS (Department of Trade and Industry, Laboratory of the Government Chemist, Cornwall House, Stamford Street, London, SE1 9NQ) A method is proposed for the determination of mono-, di- and triethylene glycols in surface waters, based on the oxidation of the glycols to aldehydes. These are made to react with 3-methylbenzothiazol-%one hydrazone hydro- chloride to give green cationic chromogens, which are then measured spectro- photometrically a t 630 nm. Sample blank values, to compensate for natural interferences, are obtained by omitting the oxidation stage. The method enables glycol levels of upwards of 0.5 mg 1-1 to be determined, satisfactory recoveries of each glycol being obtained for concentrations of 1 to 6 mg l-l, with a precision of 7 per cent., for a range of water samples.A sensitive variation of the method, for the determination of monoethylene glycol alone, is described separately. MONO-, di- and triethylene glycols, frequently mixed with volatile aliphatic alcohols, have been used as de-icing agents for aircraft and airfield runways. Airfield drainage may subse- quently contaminate surface waters with low levels of these glycols. It has been recommended that mono- and diethylene glycol concentrations should not exceed 1 mg 1-1 in water reser- voirs,l but no methods for measurement at these levels appear in the literature. Because of the polarity of the glycols it is not possible to concentrate them quantitatively by solvent extraction, therefore their determination must be accomplished in the aqueous phase.This factor, associated with low sensitivity, excludes many instrumental techniques and also thin-layer and gas - liquid chromatography. For similar reasons, spectrophotometric methods involving the specific reaction of glycols with l-naphthol in sulphuric acid,2 or methods involving general reactions for aliphatic hydroxyl groups, such as dichromate ~ x i d a t i o n , ~ ~ ~ the formation of vanadium(V) hydroxyquinolinates5 or reaction with ammonium cerium( IV) Accordingly, the conversion of the glycols into other compounds has to be considered. This conversion could be achieved by oxidation of the primary alcohol groups to aldehyde or carboxyl groups, but methods for the determination of these functional groups are, in general, equally insensitive.The determination of aldehydes has been reviewed by Alt- shuller, Cohen, Meyer and Wartburg.8 The formation of Schiff’s bases9 or 2,4-dinitrophenyl- hydrazoneslO is not specific to aldehydes and methods based on their formation have low sensitivity, while the reaction with phenylhydrazine and potassium hexacyanoferrate(1II)ll requires the presence of a minimal amount of water. Cleavage of the vicinal glycol group to fonnaldehyde,12 followed by reaction with 1,8-dihydroxynaphthalene-3,6-disulphonic acid (chromotropic acid),13 would enable only one of the glycols to be determined down to the level of 1 mg 1-l.Three reagents that give cationic chromogens with aldehydes are 2-hydrazino- benzothiazole,14 2-hydrazinobenzothiazole with 9-nitrobenzenediazonium tetraflu~roboratel~ and 3-methylbenzothiazol-2-one hydrazone hydrochloride (MBTH) ,16 The first two of these reagents require alkaline conditions and the third requires acidic conditions. The reagents have shown superior sensitivity to aldehydes when compared with chromotropic acid and related compounds.17 2-Hydrazinobenzothiazole and MBTH react only with aliphatic aldehydes, and the latter is more sensitive to a wider range of these compounds. More recently, the sensitivity of MBTH has been greatly improved,l* and the reaction has been applied to the determination of olefins after oxidation,19 to compounds containing 2-aminoethanol and ethylenediamine fragments20 and to sugars containing aldehyde groups or polyhydroxyalde- hyde precursors.21 The sensitivity of MBTH to a wide range of aliphatic aldehydes suggested that, provided suitable conditions for the oxidation of the glycols to aldehydes can be achieved, this reagent @ SAC; Crown Copyright Reserved. are unsuitable.EVANS AND DENNIS 783 might be the basis for a sensitive method for the determination of glycols in surface waters. A method is proposed; based on the investigations described, whereby the three glycols can be monitored in surface waters at concentrations of upwards of 0.5 mg 1-l. METHOD REAGENTS- with de-ionised water. These should be of analytical-reagent grade when available; solutions can be prepared Sulphuric acid, 4 N.Potassium Permangunate solution, 0.0126 M-Dissolve 0.2 g of potassium permanganate in water and dilute the solution to 100 ml. Sodium arsenite solution, 0.07 M-Dissolve 9.1 g of sodium arsenite in 100 ml of water. For use, dilute 10 ml of this solution to 100 ml. 3-Methylbenxothiaxol-2-one hydraxone hydrochloride (MBTH), 2.0 per cent. m/V solution- Dissolve 2 g of reagent in 100 ml of water. Iron(Il1) chloride - sulphamic acid solution-Dissolve 2 g of iron( 111) chloride hexahydrate and 3 g of sulphamic acid in water and dilute the mixture to 100 ml. Standard glycol solutions-Mix 10 g of each glycol with water in separate 1-litre calibrated flasks, and dilute 50 ml of each solution to 1 litre to give solutions containing 500 mg 1-1 of glycol.Immediately prior to use, dilute 5 ml of each of these standards to 500 ml to give standard solutions containing 5 mg 1-1 of glycol. PROCEDURE- Measure 5ml of sample (after allowing it to settle or filtering it through pre-washed cotton-wool), or a suitable aliquot diluted to 5m1, into a 10-ml calibrated flask, and also prepare a reagent blank with 5 ml of de-ionised water. To each flask add 0.5 ml of sulphuric acid followed by 1 ml of potassium permanganate solution, mix well, and immerse the flasks in a boiling water bath for exactly 5 minutes, Withdraw them from the water-bath and remove any unreacted permanganate with 1 ml of sodium arsenite solution. Add 1 ml of MBTH reagent solution and immerse the flasks in the water-bath for a further 6 minutes. Remove them, cool to room temperature, transfer 1 ml of iron(II1) chloride - sulphamic acid reagent by pipette into each flask and dilute to 10 ml in each instance.Allow to stand for 20 minutes and read the optical density of each solution at 630 nm in a clean 2-cm cell with water in the reference cell. The net optical density of the sample is obtained by subtracting the reagent blank value. To compensate for natural interferences in surface waters, a sample blank value should be obtained with the permanganate oxidation stage omitted. Measure 5 ml of filtered or settled sample, or a suitable aliquot diluted to 5 ml, into a 10-ml calibrated flask; also prepare a reagent blank with 5 ml of de-ionised water. Add in order 0.5 ml of sulphuric acid, 1 ml of sodium arsenite solution and 1 ml of MBTH reagent solution, mixing after each addition, and immerse the mixture for 6 minutes in a boiling water bath. Ascertain the optical density of the sample blank, less that due to the reagent blank, by the procedure described previously after the addition of 1 ml of iron(II1) chloride - sulphamic acid reagent.The optical density due to glycols, obtained by subtracting this natural blank value from the net sample optical density, can be expressed as the concentration of the glycol, if known, or in terms of a specific glycol from the calibration graphs. CALIBRATION- To a series of 10-ml calibrated flasks add, with a pipette, 0, 1, 2, 3, 4 and 5 ml of the standard solution containing 5mg1-1 of glycol, and dilute to 5ml with de-ionised water when necessary.Ascertain the optical densities, less that due to the reagent blank, by the procedure described. The resulting calibration graphs of optical density against concentration of glycol are linear in the range from 0 to 5 mg 1-1 for each glycol. DISCUSSION The oxidants used most frequently in organic chemistry include permanganate, chromic acid or systems involving metal ions such as cerium(IV), manganese(III), cobalt(II1) or vana-784 EVANS AND DENNIS : SPECTROPHOTOMETRIC DETERMINATION OF [Analyst, VOl. 98 dium(1V). The magnitudes of the electrode potentials of various oxidising media indicate that cerium(1V) in perchloric acid (E", 1.71 V) or nitric acid (E", 1.61 V) and acidified per- manganate (E", 1-52 V) are more effective for oxidation than acidified dichromate (E", 1.33 V) and other systems.These figures may not reflect, however, the effectiveness of their oxidation properties when applied to organic substrates. Chromium(V1) has been extensively used to oxidise primary alcohols but invariably further oxidation to carboxylic acids, or side-reactions yielding esters via the hemiacetals, occur. Non-substituted 1,2-diols, such as monoethylene glycol, are normally oxidised to the dialdehyde or dicarboxylic acid without cleavage.Z2 Cerium(IV), in an acidic medium, oxidises alcohols via complex intermediates that may not easily be oxidised further, while lJ2-diols may be cleaved22 and would yield formaldehyde from monoethylene glycol with a resulting disproportionately high molar extinctionlS compared with the extinctions obtained with di- and triethylene glycols.I t has been suggested that alkaline conditions are more effective than neutral or acidic conditions for the permanganate oxidation of primary alcohols, while 1,2-diols are not noticeably cleaved to aldehydes with permanganate.22 An oxidising system, involving the use of acidic or neutral conditions, was required that would convert the hydroxyl groups of mono-, di- and triethylene glycol into alde- hyde groups, while, at the same time, minimising further oxidation to the corresponding carboxylic acid compounds. Of several oxidising systems initially investigated, dichromate in dilute acetic or sulphuric acid and neutral permanganate produced no measurable aldehyde at temperatures up to 100 "C; permanganate in dilute sulphuric acid gave a good response for aldehydes for oxidation at 100 "C.One of the factors normally determined in water examination is the 4-hour permanganate value, measured in acidic conditions. For river waters this value has been related empirically to the permanganate consumed on heating the sample for 30 minutes at 100 "C, the consumption being doubled for the latter condi- t i o n ~ . ~ ~ The retention of acidified permanganate as oxidant would therefore be an advantage for this method, an allowance for the normal permanganate consumption of river waters being incorporated in the amount of permanganate used. The total determination involved a number of stages that required careful investigation in order to obtain a reproducible response.For the oxidation stages, the concentration of permanganate, time of reaction required for oxidation, acidity and concentration of reductant for removal of excess of permanganate required particular attention. For the spectrophoto- metric stages the factors that required investigation were the concentration of MBTH reagent, the time of reaction to form the intermediate azine, the acidity and, finally, the conditions for oxidation, with iron(II1) chloride, of the reagent MBTH to a reactive cation and subsequent formation of the final blue or green cationic chromogen. SPECTROPHOTOMETRIC DETERMINATION WITH 3-METHYLBENZOTHIAZOL-2-ONE HYDRAZONE HYDROCHLORIDE (MBTH) An interfering opalescence in the colour development of the MBTH procedure was originally controlled by dilution with acetone, the resulting method losing sensitivity.16 Subsequently, Hauser and Cumminsl* substituted sulphamic acid for acetone for controlling this opalescence and achieved a six-fold increase in sensitivity for formaldehyde, the final reaction volume being reduced to 12 ml.Preliminary work indicated that, as the reagent concentration decreased with increasing volume of solution, so the relative response to the aldehyde precursors decreased; this effect was only partially overcome by increasing the concentration of MBTH. To obtain the necessary sensitivity the final solution volume was therefore limited to 10 ml. Because the addition of several reagents involved up to 5 ml of solution, sample volumes used in this determination could not exceed 5 ml.Each glycol oxidation product on reaction with MBTH produced a green chromogen that gave an absorbance maximum in the wavelength region of 630nm. The effects of variations in the reaction conditions are exemplified by the net optical densities in Table I. Values in italic type are the averages (per mgl-l) of eight series of readings obtained for 1, 2, 3,4, 5 and 6 mg 1-1 of each glycol by a single operator who used the procedure described under Method. The average coefficient of variation from the average response at each level for each glycol for this series of readings was 3 per cent. In no instance did the averages of other operators differ from the average values by more than 3t3 per cent. The sensitivityNovember, 19731 LOW LEVELS OF GLYCOLS I N SURFACE WATERS 785 of each aldehyde precursor to MBTH concentration reached a maximum with 1 ml of 1.5 per cent.of reagent, theoretically a many-fold excess, and declined as the concentration increased further. One millilitre of 2 per cent. MBTH was chosen as being the optimum concentration of reagent, thereby avoiding excessively high optical densities with the higher concentrations of monoethylene glycol, which might prove difficult to read accurately on some spectro- photometers. The extent of the reaction a t 100 "C did not alter over the time span 2 to 10 minutes, although response to the product of oxidation of diethylene glycol declined slightly for a reaction time of 10 minutes; a reaction time with MBTH of 6 minutes was therefore selected. Oxidation of the MBTH reagent to a reactive cation and subsequent formation of the green cationic chromogen was achieved with iron( 111) chloride, the opalescence being con- trolled by sulphamic acid.The response to each aldehyde precursor increased as the volume of 2 per cent. iron(II1) chloride - 3 per cent. sulphamic acid reagent increased. A volume in excess of 1 ml of this reagent produced increases in the reagent blank that were dispro- portionate to the increase in response obtained for both di- and triethylene glycols, and hence 1 ml of the composite reagent was used in this method. It was necessary to cool the reaction mixture to room temperature prior to this addition. The final colour formed reached a maximum response after 15 minutes and remained stable for a further 15 minutes.Measure- ment 20 minutes after addition of the reagent was therefore recommended. CONDITIONS FOR OXIDATION OF GLYCOLS TO ALDEHYDES- The effects of varying the reagent conditions for oxidation are summarised in Table I1 by the net optical densities shown; values in italic type are averages, as in Table I. A constant response was obtained for mono- and triethylene glycols, at the levels indicated, with 1 to 2 mg TABLE I EFFECT OF VARYING REAGENT CONDITIONS ON SPECTROPHOTOMETRIC DETERMINATION WITH MBTH Optical density ( x 1000)* Monoeihylene glycol/mg 1-1 Diethylene glycol/mg 1-1 Triethylene glyc&/mg 1-1 a' 1 1 Variable 2 4 2 4 2 MBTH concentration, t per cent. 0.6 1-0 1.6 2.0 3-0 Time of reaction: / minutes 2 4 6 8 10 Volume of reagent added§/ml 0.5 0.76 1.0 1-25 1-50 220 436 60 138 91 292 580 90 190 115 308 620 104 212 127 260 520 106 212 118 186 374 93 173 98 26 1 528 102 200 121 257 537 97 212 118 260 520 106 212 118 262 624 92 208 117 256 628 96 196 118 142 277 51 115 76 206 442 72 166 98 260 520 106 212 118 274 600 116 234 130 325 670 112 242 136 * Values in italic type are averages.t Concentration in 1 ml of reagent. $ Reaction a t 100 O C with 1 ml of 2.0 per cent. MBTH. 8 2 per cent. iron(II1) chloride - 3 per cent. sulphamic acid reagent. 198 242 259 236 193 246 247 236 234 240 138 200 236 268 274786 [A nalysf, Vol. 98 of permanganate, while a constant level was obtained for diethylene glycol for the range 1.5 to 2 mg of permanganate. The maximum permissible sample volume employed in this procedure is 5 ml and it can be calculated that a water sample with a- 4-hour permanganate value of 10, representing a badly contaminated river water, would consume 200pg of per- manganate.If the empirical relationship of a doubled consumption when heated at 100 "C for 30 minutes is accepted,23 400 pg would be the maximum amount of permanganate used for a 5-ml volume of such a river water. The use of 2 mg of permanganate therefore leaves sufficient reagent to ensure reproducible oxidation of the glycols to their corresponding aldehydes. An increase in the time of oxidation at 100 "C indicated a decrease in response to each aldehyde precursor. This effect can be attributed to the progressive loss of the aldehydes formed as a consequence of their volatility at elevated temperatures.The change in response to di- and triethylene glycols was small between 3 and 7 minutes' reaction time and hence an optimum time of 5 minutes was chosen for oxidation a t 100 "C. A single addition of sulphuric acid and permanganate gave inconsistent results in pre- liminary trials. This inconsistency was partly caused by the instability of permanganate EVANS AND DENNIS : SPECTROPHOTOMETRIC DETERMINATION OF TABLE I1 EFFECT OF VARYING CONDITIONS FOR OXIDATION OF GLYCOLS TO ALDEHYDES Optical density ( x 1000)* A r Monoethylene glycol/mg 1-1 Diethylene glycol/mg 1-1 Triethylene glycol/mg 1-1 Variable Amount of permanganatelmg 0.50 1.00 1.60 2.00 Time of oxidation?/ minutes 1 3 5 7 10 Sulphuric acid concen- t r a t i o n $ / ~ 2 3 4 5 6 Sulphuric acid concentration (effect on MBTH reaction)$/N 4 6 Volume of 0.50 0.75 1-0 1.25 1-50 0.07 M arseniteqlml I - 2 ------? 2 2 4 232 472 73 166 102 217 257 511 96 188 120 233 254 509 103 202 117 236 260 520 106 212 118 236 303 605 95 184 132 259 280 566 109 218 129 254 260 520 106 212 118 236 248 486 108 214 114 246 219 44 1 100 204 111 220 298 596 110 220 124 260 27 1 528 102 206 120 242 260 520 10G 212 118 236 222 466 87 180 100 218 220 458 87 186 104 210 269 547 105 203 120 237 220 463 94 174 111 216 261 534 107 23 1 120 247 263 538 108 216 124 248 260 520 106 212 118 236 250 520 106 218 119 239 261 530 99 203 114 235 * Values in italic type are averages.t Oxidation a t 100 "C with 2 mg of permanganate. $ Volume = 0.5 ml. 9 Volume = 0.5 ml; reaction after oxidation with 0.5 ml of 2 N acid.Effect of use of various volumes of 0.07 M arsenite to reduce the excess of permanganate.November, 19731 LOW LEVELS OF GLYCOLS I N SURFACE WATERS 787 in an acidic medium, with decomposition to manganese(IV), and sulphuric acid was therefore added before the permanganate. As the sulphuric acid concentration increased so the response of each aldehyde precursor decreased. Subsequent variation of this acid. concentration in the reaction stage with MBTH reagent, while maintaining the acid concentration at a constant level during oxidation, indicated that acidity is a factor relevant only to the final reaction stages. The change in response was small for the addition of 0.5 ml of 3 or 4 N sulphuric acid and the latter concentration was selected.The removal of excess of permanganate with arsenite was possible over a relatively wide range of sodium arsenite concentrations and 1 ml of 0.07 M arsenite was finally used in the procedure. This feature was singular, and the need for careful control of reagent additions, and, in particular, the time allowed for oxidation with permanganate must be emphasised in view of previous comments. Different batches of MBTH reagent showed no significant difference in response to each aldehyde precursor, and similarly no difference in sensitivity was apparent when reagents were prepared or dilutions made with demineralised or distilled water. Ageing of reagent solutions (over a period of 7 days) produced no change in response but the reagent blank increased with the age of the reagent; for freshly prepared reagent solutions, the reagent blank should have an optical density not exceeding 0.080.Calibration graphs of optical density against concentration of glycol were linear for the range 0 to 6 mg 1-1 for each glycol, enabling measurements to be made in the range 0 to 5 mgl-l. Typical optical density responses for each glycol, measured by the procedure described, were : monoethylene glycol, 0-130; diethylene glycol, 0.053; and triethylene glycol, 0.059 per mg 1-I. PRODUCTS OF OXIDATION- The molar extinctions of the final chromogenic solutions, obtained by the procedure outlined, were 8060, 5620 and 8850 for mono-, di- and triethylene glycols, respectively. From results obtained for formaldehyde,ls a molar extinction of 1 x lo5 would be expected if complete cleavage of monoethylene glycol to formaldehyde had occurred.Normal oxidation of the glycols would be expected to yield dialdehydes, e.g., glyoxal (biformyl) would be the expected product from monoethylene glycol. Other possible products would be glyoxylic acid (oxoacetic acid), glycollic aldehyde (hydroxyacetaldeh yde) and glycollic acid (hydroxy- acetic acid) or oxalic acid; for the latter two, no aldehyde response would be obtained. A molar extinction of 28000 has been recorded for glyoxal.16 We have found that the reaction of MBTH reagent with glyoxal solutions, prepared from both the polymerised monohydrate and from a 40 per cent. rn/m solution, gave a response, at concentrations of less than 1 mg l-l, corresponding to that obtained by this procedure for monoethylene glycol.Reaction of MBTH with glyoxylic acid monohydrate, however, gave a molar extinction of 59 000, while glycollic aldehyde gave a molar extinction of 36 500. It is emphasised that organic reactions that involve oxidation will invariably give several reaction products, but this evidence suggests that the dialdehyde is the major product of oxidation from mono- ethylene glycol for the conditions defined in this method. It is not certain whether both aldehyde groups react with 2 mol of MBTH initially to give a diazine. A maximum response is, however, obtained with 1 ml of 1.5 per cent. reagent (Table I), representing a very large excess of reagent, whereas previous applications of MBTH to a wide range of aldehydes and aldehyde precursors have involved the use of 1 ml of 0.8 per cent.reagent1g$21; it is not clear whether the latter was the over-all optimum concentration. The molar extinction obtained for diethylene glycol is lower than that obtained for the other two glycols. This effect could arise because of the formation of a monoazine with MBTH reagent, or the presence, after permanganate oxidation, of one aldehyde group only. The relatively slow increase to a maximum response with increasing amounts of permanganate (Table 11) suggests that labile oxidation of one hydroxyl group via the aldehyde to a carboxylic acid group is unlikely. Experimental evidence is inadequate to decide which of the pair of alternatives is the cause of this low molar extinction; it is sufficient that a constant level of response is attainable. SURFACE-WATER SAMPLE BLANKS- Typical water sample blanks, expressed in terms of optical densities, including and excluding the permanganate oxidation stage, are illustrated in Table 111.The good agreement788 EVANS AND DENNIS SPECTROPHOTOMETRIC DETERMINATION OF [AfiPzalyd, VOl. 98 obtained indicates the general absence of a natural level of aldehydes, which would be further oxidised by permanganate, and equally, natural levels of aldehyde precursors (Le., hydroxyl compounds containing primary alcoholic groups). Exceptions to this agreement were noted for domestic effluents. While this agreement holds within our experience, there is a possi- bility that a surface water could prove anomalous because of polluting species.In such an instance, a sample of the surface water immediately before contamination would be required in order to ascertain the surf ace-water sample blank with permanganate oxidation. During the monitoring, over a period of 6 months, of river G with a high constant water flow, this sample blank was consistent (Table 111), but for other river systems subject to large fluctuations in river flow it might vary with the extent of floodwater, e.g., rivers C and D (Table 111). In our experience, volatile alcohols associated with de-icing glycols are seldom encountered in airfield run-offs at water temperatures of greater than 0 "C, because of their volatility, and would not be expected in receiving waters. TABLE I11 COMPARISON OF WATER SAMPLE BLANK VALUES WITH AND WITHOUT THE PERMANGANATE OXIDATION STAGE Optical density ( x 1000) Sample Spring water .. .. .. Well water .. .. .. Treated swimming-pool water . . River water: A . . .. .. B .. .. .. C, October, 1972 . . C, November, 1972 D, April, 1972 . . D, October, 1972 . . Estuary water E . . .. .. River water: F . . .. .. G, June, 1972 . . G, August, 1972 . . G, November, 1972 G, December, 1972 Airfield ditch H . . .. .. Domestic sewage effluent . . Partially treated domestic sewage Airfield ditch J . . .. .. .. .. .. .. .. .. * . .. .. .. .. .. .. .. .. .. * . . . .. Wiih oxidation 0 21 35 7 36 32 75 40 112 77 107 90 96 96 40 140 127 107 a4 Without oxidlation 0 29 37 0 29 37 66 46 104 69 105 80 89 - 34 132 76 55 A SENSITIVE VARIATION FOR DETERMINATION OF MONOETHYLENE GLYCOL INVOLVING CLEAVAGE TO FORMALDEHYDE Cleavage with periodic acid of vicinal diols followed by reaction with chromotropic acid of the aldehydes formed,13 and similarly, cleavage of molecules containing the 2-aminoethanol and ethylenediamine fragments to aldehydes and subsequent measurement with MBTH reagent, have been described.20 A similar oxidation of monoethylene glycol would yield formaldehyde, which would give a coloured cation with a high molar extinction with MBTH reagent.Calculation from values of optical densities obtained previously for forma1dehydel8 indicates that an optical density of 0.660 per mg 1-1, measured at 630 nm in P-cm cells, would be expected from the splitting of monoethylene glycol and subsequent measurement in a final 25-ml volume of solution.In practice, water sample blank values with and without periodate oxidative cleavage disagreed (Table IV) . This disagreement could be explained by the extreme sensitivity of the reaction conditions to polyhydroxyaldehyde precursors, involving both primary and secondary alcoholic groups, which cleaved to form formaldehyde or aldehydes of low relative molecular mass of similar sensitivity. This effect must be com- pared with the absence of aldehyde precursors for acidified permanganate as the oxidation medium, when only primary alcoholic groups are involved. Allowing for this inherent weakness, which implies that monitoring a surface water would require sample blank values before contamination, the extreme sensitivity of this variation commends it for attention.A brief discussion and an outlined procedure is therefore included below for information.November, 19731 LOW LEVELS OF GLYCOLS IN SURFACE WATERS TABLE IV WATER SAMPLE BLANK VALUES WITH AND WITHOUT PERIODATE OXIDATION Optical density ( x 1000) r A I River water . . . . .. c D F G With periodate. . .. .. 19 153 254 96 Without periodate . . .. 0 35 40 27 789 DISCUSSION- The results for different oxidative reagent conditions are shown in Table V. Complete reaction at room temperature was uncertain and results indicated that no disadvantage accrued from heating, consistent response being achieved for 1 to 8 minutes' reaction time at 100 "C; a time of 2 minutes was chosen for this reaction stage. There were no restrictions on oxidation for initial volumes of 0 to 11 ml of solution, thus enabling the monoethylene glycol content of 10-ml samples of surface waters to be determined.Variation of the sulphuric acid concentration for the MBTH reagent stages showed that as acidity increased, sensitivity decreased. It was found that the acid used, 1 ml of 2 N sulphuric acid, was not essential for the initial oxidative cleavage, but it was convenient to use acidified periodate as it increased the solubility of potassium periodate in a cold aqueous solution and reduced the number of solutions that had to be added. Variation in the MBTH reagent concentration indicated that reaction with 1 ml of 2 per cent. MBTH reagent at 100 "C was required; constant response was obtainable fora reaction time of 4 to 8 minutes and 6 minutes was accepted as the optimum time.The response to the addition of the 2 per cent. iron(II1) chloride - 3 per cent. sulphamic acid reagent was a maximum for the addition of 1.0 to 1-25 ml of reagent, stable readings being obtained 15 to 20 minutes after addition. For smaller volumes of the composite reagent, a lower response was obtained, while for larger volumes, fading of the final blue cationic TABLE V EFFECT OF VARYING REAGENT CONDITIONS FOR PERIODATE OXIDATION OF MONOETHYLENE GLYCOL Optical density ( x 1000) A r \ Monoethylene glycol concentration/mg 1-1 . . 0.2 0.6 1.0 Reaction conditions with 1 ml of 0.04 M periodate r 1 Temperature Timelminutes Ambient 5 90 276 10 114 334 15 133 387 20 141 386 30 133 385 100 O C 1 144 414 2 140 412 4 144 417 8 139 417 2 129 404 3 134 394 4 123 370 Sulphuric acid concentration*/N MBTH concentration, per cent.Timelminutes 0.5 4 41 160 1.0 4 89 307 2.0 2 111 31 1 2.0 4 139 404 2-0 8 140 414 * Volume of acid = 1 ml. t Volume of reagent = 1 ml. 500 5 64 669 66 1 662 669 667 674 675 674 672 615 280 485 54 1 671 67 1790 EVANS AND DENNIS : SPECTROPHOTOMETRIC DETERMINATION OF [Analyst, Vol. 98 chromogen occurred within this time span. A volume of 1 ml of composite reagent was therefore selected and the final colour was read 20 minutes after its addition. By using the procedure described, a molar extinction of 1.05 x lo5 is obtained for mono- ethylene glycol, which compares favourably with the value of 7.3 x lo4 obtained previously in a general method for polyhydroxyaldehyde precursors.21 REAGENTS- with de-ionised water.These should be of analytical-reagent grade when available; solutions can be prepared Sulphuric acid, 2 N. Potassium Periodate solution, 0-04 M-Dissolve 0.92 g of potassium periodate in 100 ml of 2 N sulphuric acid. Sodium arsenite solution, 1 M-Dissolve 13 g of the reagent in 100 ml of water. 3-Methylbenzothiazol-%one hydrazone hydrochloride (MBTH), 2 per cent. m/V solution- Dissolve 2 g of the reagent in 100 ml of water. Iron(1II) chloride - sulphamic acid solution-Dissolve 2 g of iron(II1) chloride hexa- hydrate and 3 g of sulphamic acid in water and dilute the mixture to 100 ml. Standard monoethylene glycol solution-Mix 10 g of the glycol with 1 litre of water to give a solution containing 1000Omgl-l. Immediately prior to use, dilute 50ml of this standard solution to 1 litre to give a solution containing 500 mg 1-l; 2 ml of this solution diluted to 500 ml gives a working solution of 2 mg 1-l.PROCEDURE- Measure 10 ml of settled sample (or sample after filtration through pre-washed cotton- wool), of a suitable aliquot diluted to 10 ml, into a 25-ml calibrated flask, and also prepare a reagent blank with 10 ml of de-ionised water. Add 1 ml of periodate reagent to each flask, mix the contents well and immerse in a boiling water bath for 2 minutes. Withdraw the flasks from the water-bath and remove excess of periodate by reaction with 1 ml of sodium arsenite, then add 1 ml of MBTH reagent solution and immerse them in the water-bath for a further 6 minutes. Remove and cool to room temperature, add by pipette 1 ml of iron(II1) chloride - sulphamic acid reagent to each flask and dilute the contents to 25 ml.Stand for 20 minutes and read the optical density at 630 nm in a clean 1-cm cell with water in the reference cell. The net optical density for the sample is obtained by subtracting the reagent blank. To compensate for interference from the natural levels of oxidisable material in surface waters, a sample blank value before contamination must be obtained. The optical density due to monoethylene glycol, obtained by subtracting this natural blank value from the net sample optical density, can be expressed as concentration of glycol by reference to a cali- bration graph prepared by diluting 0, 1, 2, 3, 4 and 5 ml of standard solution containing 2mg1-1 of glycol to 10ml with de-ionised water in 25-ml calibrated flasks.A graph of the optical densities, less reagent blank, plotted against the concentration of glycol is linear for the range 0 to 1 mg l-l, with an optical density of 0.680 per mg 1-1 for monoethylene glycol. RESULTS OF RECOVERY EXPERIMENTS FOR MONO-, DI- AND TRIETHYLENE GLYCOLS Results of recovery experiments, involving the concentration of surf ace-water samples, on a water-bath, suggest that attempts to improve sensitivity by this means cannot be recom- mended. While satisfactory recovery and precision were obtained for uncontaminated water samples, river-water samples gave excessive recovery and low accuracy. This result could be explained by possible biodegradation, in particular of monoethylene glycol, and in part by the inconsistencies of the water sample blanks on concentration.Sample blanks measured after concentration and permanganate oxidation increased according to the degree of concentration. Similar blanks obtained without oxidation were disproportionate to the extent of concen- tration; this disparity could be due to aerial oxidation of the organic constituents to aldehydes during the evaporation. The results of recovery experiments for each glycol in the range 1 to 5 mg l-l, with use of the acidified permanganate procedure, are shown in Table VI. For each glycol, 5 ml of each of 20 and 100 mg 1-1 solutions were diluted to 100 ml with each of the filtered waters indicated. Allowance was made, when necessary, for the water sample blank to account for dilution of water sample with standard. The freshly collected river waters were of varyingNovember, 19731 LOW LEVELS OF GLYCOLS I N SURFACE WATERS TABLE VI RECOVERY OF GLYCOLS FROM FORTIFIED SAMPLES Per- manganate value/ mg 1-1 Spring water.. . . 0.4 Well water . . . . 1-8 Public water supply 1.4 River water: A . . 2.0 C . . 7.0 D . . 2.6 G . . 6.0 G . . 6.0 Monoethylene glycol/mg 1-1 * 1 5 0.90 5-05 1-03 5.23 0.94 4-89 1.02 5-19 0.84 4-87 0-98 5-11 1.19 5-08 0-97 5-04 Diethylene . glycol/mg 1-l i 5 1.05 5.10 0.85 4-63 1.17 4.63 1.07 5-20 1.09 4-78 0.9 1 4-97 0.92 4.70 1.06 5.34 79 1 Triethylene glycol/mg 1-1 * 1 5 1.11 4.94 1.00 4.89 1-05 5.02 1.03 4-84 0.98 5-32 1-11 4.80 1.00 5-20 1-02 6-18 composition from different topographical origins.Rivers A and C originated as upland streams, while rivers D and G were slow-flowing rivers passing through agricultural land into which sewage effluents discharged. The average recovery was 100-5 per cent. with an over-all coefficient of variation of 7 per cent. At the 1 mg 1-1 level this value was 9 per cent., while at the 5 mg 1-1 level it was 4 per cent. Based on these coefficients of variation and those of reagent and sample blanks, a limit of sensitivity of the order of 0.5mg1-1 was indicated. Recoveries from estuary water E, with dissolved solids at a concentration of 15 300 mg l-l, were low and are not included; this low recovery suggests that the method would be invalid for saline waters. The method has been used successfully in these laboratories to monitor the biodegradation of low levels of glycols in river waters. This paper is published with the permission of the Government Chemist. 1. 2. 3. 4. 6. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. REFERENCES Plugin, V. P., Gig. Sanit., 1968, 33 (3), 16; Chem. Abstr., 1968, 68, 117 070. Kaye, S., and Adams, A. C., Analyt. Chem., 1950, 22, 661. Sargent, R., and Rieman, W., Analytica Chim. Acta, 1956, 14, 381. Reese, H. D., and Williams, M. B., Analyt. Chem., 1954, 26, 568. Tanaka, M., and Kojima, I., Analytica Chim. Acta, 1968, 41, 75. Reid, V. W., and Truelove, R. K., Analyst, 1952, 77, 325. Dennis, A., Evans, W. H., and Patterson, S. J., Wat. Treat. Exam.. 1972, 21, 350. Altshuller, A. P., Cohen, I. R., Meyer, M. E., and Wartburg, A. F., Analytica Chim. Acta, 1961, Blaedel, W. J., and Blacet, F. E., Ind. Engng Chem., Analyt. Edn, 1941, 13, 449. Heistand, R. N., Analytica Chim. Acta, 1967, 39, 258. Tanenbaum, M., and Bricker, C. E., Analyt. Chem., 1951, 23, 354. Malaprade, L., Bull. Soc. Chim. Fr., 1928, 43, 683. Speck, J. C., and Forist, A. A., Analyt. Chem., 1954, 26, 1942. Sawicki, E., and Hauser, T. R., Ibid., 1960, 32, 1434. Sawicki, E., and Stanley, T. W., Mikrochim. Acta, 1960, 510. Sawicki, E., Hauser, T. R., Stanley, T. W., and Elbert, W., Analyt. Chem., 1961, 33, 93. Sawicki, E., Hauser, T. R., and McPherson, S., Ibid., 1962, 34, 1460. Hauser, T. R., and Cummins, R. L., Ibid., 1964, 36, 679. Sawicki, E., Engel, C. R., and Guyer, M., Analytica Chim. Acta, 1967, 39, 505. Sawicki, E.. and Engel, C. R., Chemist Analyst, 1967, 56, 7. Sawicki, E., Schumacher, R., and Engel, C . R., Microchem. J., 1967, 12, 377. Wiberg, K. B., “Oxidation in Organic Chemistry,” Academic Press, New York, 1965. Holden, W. S., “Water Treatment and Examination,” Longman Group Ltd., London, 1970, p. 163, Received March 28th, 1973 Accepted June Sth, 1973 25, 101.