THE NITROGEN CONTENTS OF COMMERCIAL DISTILLED WATER BY WILLIAM C U L L E N UHLIG, P H . B. DISSERTATION SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY IN T H E FACULTY OF PURE SCIENCE OF COLUMBIA UNIVERSITY. N E W YORK 1904 CITY PRESS OF GEO. G PECK 17 CHAMBERS ST NEW YORK 663. 5 Uh6n INTRODUCTION. The subject of this Dissertation was suggested by a paragraph in the Report for the Connecticut Agricultural Experiment Station for the year 1900. The Station had examined samples of all waters sold in Connecticut and had published the results of the analyses, with comments. Among the samples were three distilled waters and after tabulating the results the station commented on them as follows: i i The variations shown . . . indicated a difference in the skill with which these waters were prepared, as well as a difference in the nature of the water used for distillation, and show that there is a great variation in commercial distilled water." One reason why this work was undertaken was to try to discover what an analysis of distilled water means (following the usual methods for sanitary water analysis).. The growing importance of distilled water as a commercial product, and the increasing number of distilled waters on the market, will soon make it necessary to have some method of analysis or interpretation of analysis in order to compare different distilled waters. In the second place a number of States have pure food and drug laws in which the standard for drugs is the United States Pharmacopoeia; distilled water in this connection is classed as a drug. The writer realizes that a water answering to a strict interpretation of all the tests of the Pharmacopoeia is a commercial impossibility, and almost an impossibility even on a small scale; so, in this work, a comparison is made between commercial distilled water and distilled water prepared according to the Pharmacopoeia. In considering the constituents, as determined in a sanitary water analysis, we may assume at the beginning that the water must have no color, taste nor odor. Chlorine must be absent as its presence argues carelessness in distillation, or in other words some of the original water has been carried over, which takes the product out of the class, Distilled Water. iii IV INTRODUCTION. Solids will be present invariably. These come from the glass vessels in which the water is kept and transported, and also from dust gathered by the water in passing from one receptacle to another. They might be eliminated by substituting platinum or silver for glass and by taking extreme precautions in filling receptacles, but either would be prohibitory commercially. We may allow a distilled water to have not more than 10 parts per million total solids; of these solids not more than one half should be " Organic and Volatile." The amount of solids will be influenced by the age of the water, that is, the length of time it has been in glass. * This leaves only the Nitrogen contents by which to judge the water. It is understood that in a distilled water Nitrogen as, ammonia, nitrites or nitrates has not the same significance as the same constituent in a natural water. In fact it has no significance in respect to the heathftimess or deleteriousness of the water, but it is the only variable and consequently gives the only figures by which we can judge two waters comparatively. The writer has tried to discover under what conditions these Nitrogen containing c ompounds are produced and the possibility of eliminating them. The water operated on was Croton water. The four stills used differed either in size or in position of heating surface; an outline description of each still is given before giving the results. Four runs were made with each still. The first run was with water alone, corresponding to the method of the Pharmacopoeia. The second run was with Potassium Hydroxide in order to liberate the free ammonia quickly somewhat parallel to the analytical determination of free ammonia. The third run was with Potassium Hydroxide aud Potassium Permanganate to parallel the determination of total (free and albumenoid) ammonia. These runs were all made from clean stills. The fourth run was made after running approximately 4000 Gallons from the Still after Run 3, in order to see the effect of the accumulated residues on the product. The stills, which are of varying capacities, have a general resemblance, consisting of two cylinders, a larger below and a smaller above. The lower contains the heating coils and was •Action of Water on Glass.—F. Foerster, Ber. 26. 2915-2922. Chem. 33. 322-335. E. Hoyer, Chem. Zeit. 22. 1033. Zeit. anal. INTRODUCTION. V filled with the water to be distilled, while the upper contains two or three reboiling pans. The first steam condensed in the lowest pan, was reboiled by the steam passing into it, condensed again in the next pan, volatilized again and so on into the condenser. The coils were heated by steam at a gauge pressure of 70 lbs. corresponding to a temperature of 157°C. In every case 20% of the contents remained in the still at the close of the operation. The differences among the stills will be noted in the discussion of results. The methods of analysis employed were the usual ones for sanitary water analysis. For the determination of Nitrites the Sulphanilic acid aiid Naphthylamine hydrochloride method was used; for Nitrates the Phenolsulphonic acid method. Determinations of free ammonia and of nitrites were made within three hours after the distillation; albumenoid ammonia and nitrates within twenty-four hours. The results are expressed as Nitrogen in parts per million. ACKNOWLEDGMENT. This work was undertaken with the approval of Professor Charles E. Pellew. It is due solely to his consideration that it could be carried on and I take this opportunity to express my gratefulness for his kindness and interest. I wish also to extend my thanks to Mr. Arthur B. Park, M. E., for his invaluable assistance with parts of the work, and for his many suggestions. WILLIAM COLUMBIA UNIVERSITY, April, 1904. VI C. UHLIG. TABLE OF CONTENTS. Acknowledgement iii Introduction iv Operations 1 Still No. 1 1 Still No. 2 3 Still No. 3 5 Still No. 4 8 Discussion 10 Free Ammonia 10 Albumenoid Ammonia 14 Nitrites and Nitrates 19 Notes on the Analytical Operations 20 The Condition of Free Ammonia 21 The Influence of Time on the Nitrogen Contents of Distilled Water 23 Commercial Distilled Water of Various Types 26 Aqua Destillata, U, S. P 28 Conclusions 30 Biographical , 33 OPERATIONS. STILL N O . 1. Holding contents 400 gallons. Output 320 gallons. Type of still: Two cylinders, the upper very little less in diameter than the lower, and containing two reboiling pans. The heating surface consists of three spiral coils one above the other, each provided with valves allowing them to be shut off independently of one another. The top coil was uncovered after 181 gallons had been distilled; the second was uncovered after 266 gallons. The third was still covered when the total 320 gallons had been removed. During the runs each coil was shut off as soon as the water reached that level. This still was fitted with inferior valves and the coil would remain hot, however, due to the valves not being absolutely tight. As will be seen this had an appreciable effect on the nitrogen contents of the output. RUN NO. 1. The still was filled with croton water having the following nitrogen contents:* Nitrogen as " < < " " Free Ammonia, Alb. " Nitrites, Nitrates, Sample 1- after 2- " " 3- " 4- " " 5- " 6- " .021 .117 0 .15 58 2 gallons or 14.55% of holding capacity. 116.4 " 29.1 % 174.6 " 43.65% " 232.8 " 58.20% 291 " 72.75% 320 " 80.00% * The figures given for croton water were taken from the weekly reports of the New York Department of Health. 2 WILLIAM CULLEN UHLIG. Nitrogen as Free Ammonia. Alb. Ammonia. Nitrites. Nitrates. tnple 1 .061 .028 0 ..025 .041 2 .015 0 .028 .013 3 .026 0 .027 4 .018 .025 .023 .002 .024 .018 .021 .001 5 6 0 .025 .036 .051 * R U N No. 2. The clean still was filled with croton water and 60 grammes of Potassium hydroxide added. Nitrogen contents of water used : Nitrogen as Free Ammonia, .021 " Alb. " .128 " Nitrites, .0 " Nitrates, .15 The samples were taken at the intervals as in Run No. 1. Nitrogen as Free Ammonia. Alb. Ammonia. Nitrites. Nitrates. tnple 1 .061 .025 0 trace < < .048 2 0 0 3 .016 .038 0 .02 4 .046 .003 0 trace i i .041 5 .006 0 a 6 .069 .013 0 *• R U N No. 3. The clean still was filled with water and 60 grammes of Potassium hydroxide and 120 grammes of Potassium permanganate added. Nitrogen contents of water used : Nitrogen as Free Ammonia, .021 " Alb. " .148 " " Nitrites, 0 " Nitrates, .175 The samples were taken at the intervals as in Run No. 1. Nitrogen as Free Ammonia. Alb. Ammonia. Nitrites. Nitrates. Sample 1 .148 .029 0 .01 2 .118 .018 0 .01 3 .084 .02 0 .02 NITROGEN CONTENTS OF COMMERCIAL D I S T I L L E D WATER. Sample 4 5 6 .084 .066 .184 .023 .018 .043 trace .001 3 .02 .02 .03 RUN N O . 4 After Run No. 3 the still, without having been cleaned, was filled with water, KOH and K M n 0 4 added and the water distilled off. This process was repeated until the still contained the concentrated residue from 4000 gallons. It was then filled with water and 40 grammes KOH and 80 grammes K M n 0 4 added; the water was then distilled and samples collected. Nitrogen contents of water used : Nitrogen as Free Ammonia, .021 " Alb. " .169 " Nitrites, 0 " " Nitrates, .205 Sample 1- after 53.3 gallons or 13.32% of holding capacity. " 2 3 4 5 6 " " " " " 106.6 160 213.3 266.6 320 " . 26.65% 40 % 53.32% 66. 6 5 % 80 % a i i i i a a i N i t r o g e n as F r e e A m m o n i a A l b . A m m o n i a . Nitrites. Nitrates. .234 Sample 1 .049 0 0 .163 .021 0 0 ^ " 2 .026 .151 0 0 " 3 .128 .041 0 0 " 4 .168 .049 0 0 " 5 .006 .332 0 .107 " 6 STILL N O . 2. Holding capacity 375 gallons. Output 300 gallons. Type of still: Two cylinders, the upper one having a diameter one half that of the lower and containing three reboiling pans. The heating surface consists, as before, of three coils with shut off valves. The top coil was uncovered after 225 gallons had been distilled; the second after 279 gallons; the third remained covered at the end of the operation. The procedure was the same as in still No. 1, that is the coils were shut off as soon as the water reached their level. WILLIAM CULLEN UHLIG. 4 RUN N O . 1. The still was filled with croton water having the following Nitrogen contents: Nitrogen as Free Ammonia, .016 " Alb. " .154 " "* Nitrites, 0 " " Nitrates, .225 Sample 1 - after 43 gallons or 11.46% of holding capacity. " " 2 " 86 22.93% 3 "129 34.40% 4 45.86% " 172 5 " 215 57.33% " 6 "258 -68 8 0 % < < 7 "300 80. % N i t r o g e n as F r e e A m m o n i a . A l b . A m m o n i a . Nitrites. Sample 1 0 .015 .117 .008 0 " 2 .077 " 3 .054 0 .006 a 4 0 .043 .008 0 .063 0 " 5 .038 .016 .001 " 6 i< 7 .002 .010 .067 Nitrates 0 0 0 0 0 0 • 0 RUN N O . 2. The still was filled with croton water and 60 grammes of K O B added. Nitrogen contents of water used: Nitrogen as Free Ammonia, .012 " Alb. " .140 " " Nitrites, 0 " " Nitrates, .25 Samples were taken at the same intervals as in Run No. 1. Nitrogen as Free Ammonia. Alb. Ammonia. Nitrites. Nitrates. imple 1 " 2 " 3 4 " 5 " 6 " 7 .076 .013 .038 .034 .031 .028 .030 .025 .008 .003 .007 0 .001 .001 0 0 0 0 0 0 0 0 0 0 0 0 0 .04 NITROGEN CONTENTS OF COMMERCIAL DISTILLED WATER. 5 RUN N O . 3. The clean still was filled with croton water and 60 grammes of KOH and 120 grammes of K M n 0 4 added. Nitrogen contents of water used: Nitrogen as Free Ammonia, .012 " Alb. " .115 " " Nitrites, 0 " " Nitrates, .225 Samples were taken at the same intervals as in Run No. 1. Nitrogen as Free Ammonia. Alb. Ammonia. Nitrites. Nitrates. Sample 1 2 3 4 5 6 " 7 .197 .123 .095 .092 .087 .107 .161 .051 .021 .030 .013 .023 .038 .031 0 trace i i .001 .001 .002 .004 trace 0 0 .01 .015 .02 .03 R U N N O . 4. T h e method of procedure was the same as in Still No. 1; the still containing the residue from 4000 gallons before being filled. The charge was 40 gms. of KOH and 80 gms. K M 0 4 . Nitrogen contents of water used: " as Free Ammonia, .012 " " Alb. « .090 " Nitrites, 0 " " Nitrates, .2 The samples were taken at the same intervals as in Run No. 1. Nitrogen as Free Ammonia. Alb. Ammonia. Nitrites. Nitrates. Sample 1 0 0 0 .306 trace 2 0 .005 189 i i 3 0 .008 165 ii 4 0 .025 148 5 .049 .001 0 153 6 << w ,201 ,255 .031 .008 trace .003 0 .035 STILL N O . 3. Holding capacity 700 gallons. Output 560 gallons. Type of still similar to still No. 1. Upper coil uncovered at 6 WILLIAM CULLEN UHLIG. 375 gallons; second coil at 498 gallons; third coil still covered at the end of the operation. The procedure was the same as in Still No. 1. RUN NO. 1. The still was filled with Croton water having the following Nitrogen contents: Nitrogen as Free Ammonia, .012 " Alb. " .100 " " Nitrites, 0 " Nitrates, .175 Sample No. 1 - after 70 gallons or 10% of holding capacity. i i a a 2140 20% iC a a 3210 30% a a 4- a 40% 280 a a 5- < i 350 50% a < 6- < < < 420 60% a a 7- < < 490 70% a a n 8560 80% Nitrogen as Free Ammonia Alb. Ammonia. Nitrites. ]STitral trace Sample 1 0 .092 0 2 .063 .001 0 0 0 .056 3 .007 0 4 .044 .003 0 0 0 .048 0 0 5 0 0 6 .067 .0 0 .04 .053 0 7 trace .04 8 .059 0 RUN N O . 2. The clean still was filled with Croton water and 100 gms. KOH added. Nitrogen contents of water Nitrogen as Free Ammonia, " Alb. " t( " Nitrites, " " Nitrates, Samples were taken at the used: .012 .111 0 .150 same intervals as in Run No. 1, NITROGEN CONTENTS OF COMMERCIAL DISTILLED WATER. 7 Nitrogen as Free Ammonia. Alb. Ammonia. Nitrites. Nitrates. Sample 1 .044 0 0 .082 2 0 .061 .012 0 3 .041 .016 0 0 4 0 .038 .015 0 0 0 .036 0 " ,5 6 0 0 .039 0 a ty 0 .03 .038 0 .04 0 0 .038 " ? RUN N O . 3. The clean still was filled with Croton water and 100 gms. KOH and 200 gms. KMnO 4 added. Nitrogen contents of water used: Nitrogen as Free Ammonia, .014 " Alb. " .104 " Nitrites, 0 " Nitrates, .175 Samples were taken at the same intervals as in Run No. 1. Nitrogen as Free Ammonia. Alb. Ammonia. Nitrites. Nitrates. Sample 1 .204 trace 0 0 < < 0 2 .130 0 0 .099 .016 0 3 4 .001 0 0 .100 .001 0 .016 5 .100 .002 0 6 .015 .095 a ty .002 .02 0 .102 .224 .025 0 .007 8 RUN N O . 4. The still contained the residue from 4500 gallons. It was filled with Croton water and 60 gms. KOH and 120 gms K M n 0 4 added. Nitrogen contents of water used: Nitrogen as Free Ammonia, .016 " Alb. «' .098 " " Nitrites, 0 " " Nitrates, .2 The samples were taken at the same intervals as in Run No. 1. WILLIAM CULLEN UHLIG. Nitrogen as Free Ammonia. Alb. Ammonia. Nitrites. Nitrates Sample 1 .049 0 .201 0 2 .161 .026 0 0 .118 .035 3 0 'o. .041 4 0 0 .107 .033 0 5 .117 0 .049 004 6 .168 0 .041 .178 .003 .03 0 .188 .018 8 .025 << IV STILL N O . 4. This still is a duplicate of Still No. 2 in every respect except heating surface. Instead of being heated by three separate coils it has one coil entering the side of the still, making three series of convolutions and leaving at a lower point; this arrangement necessitates heating the entire coil even after it's upper portion is uncovered. The highest point of the coil was uncovered after 218 gallons had been distilled. The charges were the same as in Still No. 2, so they will not be repeated. RUN N O . Nitrogen contents of water Nitrogen as Free Ammonia, u " Alb. " " " Nitrites, " Nitrates, Sample No, 1 after a a 2 " a < < 3 "' a a 4 " u a 5 " a u 6 " 1. used: .014 .08 0 .2 50 gallons or 13.33% of holding capacity. a 100 26. 67% a 0/ 40 150 i i 53.33% 200 a 250 66. 67% < < 300 80 % Nitrogen as Fr<3e Ammonia. Alb. Ammonia. Nitrites. Nitr; Sample 1 .031 .002 .067 0 2 .051 .030 .003 0 3 .038 .020 0 0 4 .031 0 .020 0 5 .026 .008 0 0 6 .056 0 .030 0 NITROGEN CONTENTS OF COMMERCIAL DISLILLED WATER. 9 RUN N O . 2. Nitrogen contents of water used: Nitrogen as Free Ammonia, .012 '• Alb. "• .061 " Nitrites, 0 " Nitrates, .2 Samples taken at the same intervals as in Run No. 1. Nitrogen as Free Ammonia. Alb. Ammonia. Nitrites. Nitrates. nple " " " " 1 2 3 4 5 6 .082 .044 .038 .023 .021 .041 .035 .007 .012 .001 .013 .016 .001 trace 0 0 0 0 0 0 0 0 0 0 RUN N O . 3. Nitrogen contents of water used: Nitrogen as Free Ammonia, .012 " Alb. " .08 " " Nitrites, 0 " Nitrates, .2 Samples taken at the same intervals as in Run No. 1. Nitrogen as Free Ammonia. Alb Ammonia. Nitrites. Nitrates. Sample 1 " 2 3 " 4 " 5 6 .193 .084 .056 .051 .122 .189 0 .015 .018 .023 .026 .064 .002 .001 .001 .001 .005 .007 RUN N O . 4. Nitrogen contents of water used: Nitrogen as Free Ammonia, .012 " Alb. " .08 " Nitrites, 0 " Nitrates, .175 Samples taken at the same intervals as in Run No. 1. 0 0 0 0 trace .015 10 WILLIAM CULLEN UHLIG. Nitrogen as Free Ammonia. Alb. Ammonia Nitrites. Nitrates. Sample 1 .158 0 0 0 " 2 .120 .026 0 0 " 3 .105 .038 0 0 " 4 .102 .048 0 0 " 5 .283 .084 .006 0 " 6 .444 .133 .013 trace DISCUSSION. Free Ammonia. RUN N O . 1. The discussion will be carried on in connection with the plates, the ordinates showing the amount of nitrogen in parts per million, and the abcissas showing percentage of water figured on the total capacity of the still. The nitrogen contents of the water used varied so little that its influence can be neglected in discussing results. We note the similarity of the curves of stills Nos. 1 and 4, and also of Nos. 2 and 3. In fact there is great similarity among all the stills until about 40% of the water has been distilled off, but from this point the individuality of each still affects the result. Nos. 1 and 4 are the consistent ones, having a gradual decrease in the amount of nitrogen, and then a rise due, most probably, to decomposition of nitrogen containing bodies in the concentrated residue; this is of course accentuated in the case of No 4, owing to the exposure of superheated coil and the baking of residue on this. In No. 2 we have a rise at about 50% followed by a sharp drop as soon as the top coil is shut off; the nitrogen falling to a minimum to rise sharply as soon as the second coil nears the surface. In No 3 the mininum is reached at 40% the rise continuing after the first coil is shut off, falling again when the second coil is shut off, with a final " concentration " rise not as abrupt as the others, this latter fact due most probably to the greater volume of water in this still. It is difficult to discuss the production of Nitrogen as Free Ammonia in Run No. 1, as our knowledge of the composition of the original water is so limited. We know that solutions of NITROGEN CONTENTS OF COMMERCIAL DISTILLED WATER. II Ammonia salts on simple distillation yield varying quantities of N H 3 . H C. Dibbits* has reported some figures in regard to the " Dissociation of Ammonia Salts in Aqueous Solution," he having experimented with Ammonium Chloride, Nitrate, Sulphate, Oxalate and Acetate. Dr. G. Goref in a paper entitled " Analysis of Drinking Water for Ammonia " says in his "Conclusions" :—"Water containing even a small amount of Ammonia or its salts yielded much of its Ammonia by simple distillation, and did not require the aid of Carbonate of Sodium to liberate it unless salts of Alumina were present. The production of Ammonia by distillation with carbonate of sodium appeared to be due to the carbonate acting on Nitrogenous matter." In his summary he gives figures very much higher than those given by Dibbets. The solutions he used were Ammonium Hydroxide, Carbonate, Chloride, Sulphate and Phosphate, and his figures varied from 20% N H 3 from Ammonium Hydroxide to 66% from Ammonium Carbonate. Ammonium Sulphate and Phosphate yielded each 50% of their N H 3 to the distillate. H. C. DibbitsJ also published the results of his experiments on the " Decomposition of Ammonium Salts by Potassium and Sodium Salts," showing the loss of N H 3 from a solution of Ammonium Sulphate when heated with Potassium Chloride. J. W. Mallet || says: " I n the case of waters containing Urea and other aminated bodies such as the leucine and tyrosine occurring among the products of putrefective decay, some ammonia is so easily formed from these substances by boiling with sodium carbonate, or even without this addition, that it is impossible to distinguish sharply between pre-existing " f r e e " ammonia (of ammoniacal salt), and that formed by the action of alkaline permanganate, the so called " albumenoid " ammonia. H. Bliicher § gives a long list of compounds which might be present in the natural water, and we can only infer that there must be interaction between some of these with the production of ammonia, or of compounds, readily decomposable, liberating ammonia. *Zeit, anal. Chem. 1874- 395-408. f Chem. News 1884- 182-186. t Zeits. anal. Chem,1876- 245-250. || Report of the National Board of Health, 1882-page 198. § Das Wasser, pages 120-124. See also Tiemann-Gartner's Handbuch der Untersuchung der Wasser pp. 20-24. 12 WILLIAM CULLEN UHLIG. We .have in Run No. 1 the production of ammonia by dissociation of ammonium salts, by the action of metallic salts on ammonium salts, and lastly, by changes in nitrogenous compounds of organic nature. Ammonia from this last source we cannot estimate even roughly as we do not know the compounds which may be present. The figures for free ammonia from Run No. 1 are of interest as showing the large amount of am-' monia produced and the rate of production. RUN N O . 2. In discussing Run No. 2 (with KOH alone) if we contrast the figures with those from Run No. 1 we will notice in Still No. 1 that the Nitrogen in the first sample was the same in each case, but in the other samples the quantity in Run No. 2 was higher and the mininum point was just after the first coil was exposed, then there was a rise followed by a slight fall with a final abrupt rise after uncovering the second coil. In the other stills the results in Run No. 2 are all lower than in Run No. 1 which is contrary to what we would expect. From the figures we must infer that Stills Nos. 2, 3 and 4 represent normal conditions and that Still No. 1 is abnormal. From this we may conclude that on distilling water alone the greater bulk of the ammonia existing as ammonium compounds is given off and also a quantity of free ammonia or free ammonia producing substance is formed by interaction and decomposition of Nitrogen containing bodies existing in the water. The reaction of the distillate is acid due to Carbon di-oxide and large quantities of this gas are given off. In the case of the KOH distillation the alkalinity of the solution prevents or diminishes this decomposition of nitrogen containing bodies and most of the ammonia we get is that from already existing ammonium compounds. We notice in Run No. 2 in Stills Nos. 2, 3 and 4 a gradually diminishing ammonia contents until the last sample when there is a slight rise; this rise is more pronounced in Still No. 4 than in either of the others, and this can be explained by a resemblance between Stills Nos. 1 and 4. In Still No. 1 there is a diminishing curve for about one half the run, and from this point the curve becomes irregular. The steam coils in this still were fitted with inferior valves, and when the valves were shut off considerable steam still leaked through thus keeping the exposed coil at a high temperature. NITROGEN CONTENTS OF COMMERCIAL DISTILLED WATER. 13 The spray thrown on the coil was immediately evaporated leaving a residue of the solids from the water mixed with Potassium Hydroxide; this mixture at this temperature would liberate ammonia containing compounds. * In Still No. 4 the whole coil is heated all the time, but the coil is placed lower in the still so the rise of the ammonia curve does not take place until further along in the operation, and the upper part of the coil being nearer the surface of the boiling water does not have the same opportunity to get so thoroughly dry as to accentuate the baking process to the same degree as in Still No. 1. RUN N O . 3. In this run we get, for the first time, results in every case similar to those which we could have predicted; that is, a large quantity of nitrogen in the first sample, then a quick diminishing to the second and then a gradual falling off till we reach a minimum. At this point concentration begins to have an effect and we get a rapid rise of nitrogen in the final samples. This is especially noticeable in still No. 4, but contrary to expectation it does not take place in still No. 1 until after the second coil is exposed. There is nothing in the free ammonia in this run to merit special discussion. RUN N O . 4. In this run the conditions are as follows : The stills contained in addition to the water, fresh Potassium hydroxide and fresh potassium permanganate, residual hydroxide and permanganate, oxides of manganese in various states of reduction, and the innumerable products resulting from the partial oxidation of the nitrogen containing compounds originally in the water. In this run we expect an increased yield of ammonia, as we have compounds more or less oxidized by the previous treatment, and the sum of their ammonia plus the free and albumenoid ammonia in the fresh water give the high figures recorded. The curves are exceedingly abrupt, descending in the first samples and ascending in the last. Again the parallel between stills Nos. 1 and 4 can be seen, the only discrepancy being the low figure for the first sample from still No. 4. This the writer cannot explain. Another noticeable *Wanklyn, Water Analysis, 10th Edition, page 195. 14 WILLIAM CULLEN UHLTG. feature is the comparatively low figures for the nitrogen in the first five samples from still No. 3, the amounts being not very much higher than those from run No. 3 of this still. This is due, in all probability, to the larger capacity of the still, and the greater bulk of liquid seems to affect the result, the dilution being the same. When the contents became more concentrated the greater part of the albumenoid ammonia was liberated suddenly (sample 6), and in the remainder of the run the curve is less abrupt than in any of the other stills. A Ibumenoid A mnionia. The term "Albumenoid Ammonia," as used in water analysis, is only a convenience; here it has the same significance, meaning ammonia obtained by distillation with alkaline permanganate. Any ammonia obtained in this manner from distilled water must be due to the decomposition of nitrogen containing compounds which are themselves volatile or volatile with steam. In run No. 1 wTe can assume that any " Albumenoid Ammonia" found already, existed as a volatile compound or such compound is formed by the heat of the operation. In run No. 2 we have the same causes for the production of " Albumenoid Ammonia" as in run No. 1, but the results ought to be affected by the potassium hydroxide. We should have either an increased amount of ''Albumenoid Ammonia," due to the above causes, plus other volatile "Albumenoid Ammonia" compounds formed by the action of KOH on other nitrogen containing bodies in the water, or less "Albumenoid Ammonia," due to the KOH acting on the previously existing volatile "Albumenoid Ammonia" compounds and changing them to some other forms. We find that the latter is the case with the single exception of still No. 3, and the greater bulk of water in the still may have some influence on this. In run No. 3 we eliminate all previously existing "Albumenoid Ammonia," and all results we obtain must be from compounds formed during the reaction. In run No. 4 we may expect the same results as from run No. 3, but with the quantities increased by the presence of partly decomposed and partly formed bodies, due to the accumulation of residues. Rtin No. 1.—In this run in still No. 1 we have the nitrogen NITROGEN CONTENTS OF COMMERCIAL DISTILLED WATER. 15 figure gradually falling until the first coil is uncovered and rising from this point with a final concentration rise. In still Xo. 2 the quantity is very small and the curve comparatively flat until about 50% of the water has been distilled. After this point the Albumenoid Ammonia disappears to reappear after the first coil has been uncovered. The concentration rise is larger in amount but it falls again before the greatest concentration is reached. In still No. 3 we have peculiar results. The first sample contains no Albumenoid Ammonia; there is a slight amount in the second and the maximum is reached in the third, the down curve, not as abrupt as the up curve, disappears by the time 50% of the water has been distilled; there is no further production, the concentration rise being entirely absent. The state of dilution was the same in all the stills but the amount of water treated was greater in this still and this must be the factor affecting the result. Still No 4 shows results similar to still No. 1 except that the curve is flatter in the earlier stages of the run> and the minimum reached is lower; this last result is due, no doubt, to the different relative positions of the heating surfaces iii each still. The similarity is due to the same causes mentioned in the discussion of free ammonia in these two stills. Run No. 2.—In still No. 1 we have at first approximately the same amount of Albumenoid Ammonia as in run No. 1, but this is followed by a sample showing absence of Albumenoid Ammonia; then a rise followed by another fall and then a consistent rise to the end of the run. In still No. 2 we have a larger amount in the first sample than in the corresponding sample of run No. 1. The next three samples approximate the same in the two runs, the Albumenoid Ammonia disappearing in the fifth sample, but in this run the concentration rise is very minute. In still No. 3 we have the one case where the Albumenoid Ammonia is higher in run No. 2 than in run No. 1. There is a large amount in the first sample, then a sharp drop followed by a slight rise and then a drop to zero and no further appearance of albumenoid Ammonia. In still No. 4 we have figures which may be compared with those from still No. 1, that is a fall then a rise, then a fall to a i6 WILLIAM CULLEN UHLIG. minimum about the point of uncovering of heating surface and then a rise to the end of the run. Run No. 3.—Stills Nos. 1 and 2 show a resemblance in this run; the amount of Albumenoid Ammonia is high compared to the previous runs, but the noticeable point is the undulating character of the curves. It would appear from this that the volatile compound which produces ammonia, when distilled with alkaline permanganate, is passing over with the steam continuously, but also a certain quantity is accumulating in the still and is liberated suddenly, and after an interval another quantity accumulates and is liberated; in other words some of the formation of volatile Albumenoid Ammonia producing compound is intermittent. In still No. 3 we have very little Albumenoid Ammonia but that which is produced passes over in this intermittent manner. Still No. 4 acts differently. In this run there is a very smooth curve from zero in the first sample through a flat curve to the fifth sample and then a sharp rise a-t the end. Run No. 4.—In the beginning of this run we see a resemblance between Stills Nos. 1 and 3, and between Stills Nos. 2 and 4. Judging by the construction of the stills as long as all the heating surface is covered Stills Nos. 1 and 3 are identical in every respect except size, while Stills Nos. 2 and 4 are exactly identical, and the writer had expected to find resemblances between the members of each group and contrasts between the two groups, but here in the beginning of this run is the only place where this occurs. Stills Nos 2 and 4 have the upper cylinders much smaller in proportion to the lower (areas have the ratio 1:4) than have Stills Nos. 1 and 3 (No. 1 areas as 1:1.75. No. 3 areas as 1:1.9). In this run the water operated on had in solution a number of decomposition products and foamed much more than any previous run. In Stills Nos. 2 and 4 the foaming, when the water commenced to boil, carried to nearly the top of the still filling the reboiling pans with the solution in the still. This pair of stills have each three reboiling pans, and the amount of solution carried into each pan varied inversely as the height. When the water settled down to quiet boiling the volatile albumenoid ammonia compound was passed from one pan to the next, and in each pan it was acted on by the reagents and broken up am- NITROGEN CONTENTS OF COMMERCIAL DISTILLEfj WATER. 17 monia being liberated. This eliminated all albumenoid ammonia from the first sample but it also exhausted some of the permanganate in the reboiling pans. As the permanganate became exhausted more of the volatile compound passed over into the distillate until we reach a maximum, in Still No. 2, in the fifth sample; we may assume that the amount of the compounds, capable of producing this volatile body, in solution was becoming less, and so, in Still No. 2, there is a decrease from this point. A parallel point in Still No. 4 is about the fourth sample but here the effect of the exposed heated coil is apparent and we get a rapid rise due to the baking effect of the uncovered coil. In Stills Nos. 1 and 3 we have the beginnings similar. There is no chemical reaction in the reboiling pans as the foaming not being constricted did not rise so high in the upper cylinder, and there being but two pans the lower is not as close to the surface of the water as in the other stills, so the first albumenoid ammonia compound formed or pre-existing in the water was volatilized or carried over by the steam before it could be attacked by the permanganate. As the operation went on the amount of this compound diminished showing, in Still No. 3, the undulations of the curve previously noticed, and finally in this still sinking to a minimum. In Still No. 1 we have a diminishing amount of albumenoid ammonia compound in the second sample, a slight rise in the third sample, but then when we ought to get another drop the effect of the heated coil is felt and we get a continued rise, though the undulating curve is still seen as the rise from the third sample to the fourth is much more abrupt than that from the fourth to the fifth and the final one is again abrupt. From the time of Wanklyn's first paper on "Albumenoid Ammonia" in 1867 the identity of the compounds which give rise to this substance has been unknown. Wanklyn's first impressions of the ''all-siifficeing" qualities of this reaction have been modified until now the amount of "Albumenoid Ammonia " is only one of several factors used in judgeing a water. Wanklyn and his co-laborators tried the reaction on various nitrogen containing bodies and obtained varying results with different compounds. They found that some compounds gave off volatile substances which on redistilling with alkaline permanganate gave i8 WILLIAM CULLEN UHLIG. more albumenoid ammonia. To quote Wanklyn's words:— *' The conversion into ammonia of only half the nitrogen of so many of the natural alkaloids is an interesting fact. t Some light is thrown on it by the example of narcotine, which, although it gives up all its nitrogen as ammonia, gives it up slowly, and by dint of putting back the distillate; it doubtless also yields part of the nitrogen as methylamine in the first instance. A careful examination of strychnine has disclosed a somewhat similar state of matters in the case of that alkaloid. Apparently, the missing half of the nitrogen in strychnine passes provisionally into the state of some volatile alkaloid." * S. Hoogenwerff and W. A. Vandorpf some years later made experiments similar to Wanklyn's with results agreeing more or less with his, but neither Wanklyn nor these later investigators mention the strong probability of similar reactions occuring when the process is used in water analysis. When J. W. Mallet made his investigation on "Methods for the Determination of Organic Matter in Potable W a t e r v in 1881, he mentioned this probability. Under his direction Dr. Smart made a number of experiments similar to Wanklyn's, and showed in some cases the production of a volatile compound which yielded more ammonia when portions of the distillate where returned to the retort.J Mallet in his "Special Conclusions as to the Albumenoid Ammonia Process," says:— ' i There is evidence that in some cases nitrogenous organic matter is volatilized during the distillation for free ammonia, which,, if it had been retained would have yielded up its nitrogen as albumenoid ammonia; not affecting the Nessler reagent, such nitrogenous matter escapes detection under either head." || Thus calling attention to this probability. Ira Remsen in a " R e p o r t on a Peculiar Condition of the Water of Boston in Nov. 1881," says:— " T h i s clearly indicated that in the first stage of the process as usually conducted, there passed over with the free ammonia some nitrogenous substance capable of yielding ammonia with permanganate of potash." *Water Analysis. 10th Edition, page 195. fBerichte 1877, 1936-1939, 1878,1202-1206, 1879,158. {Report of the National Board of Health, 1882, pp. 252-25?. I Report of the National Board of Health, 1882, page 198. NITROGEN CONTENTS OF COMMERCIAL DISTILLED WATER. 19 Chas. W. Marsh in " Notes on the Ammonia Process for Water Analysis," * calls attention to the same fact. We may conclude from this investigation that the " Albumenoid Ammonia" compounds belong to, at least, two classes, volatile and non-volatile. We have some vo'atile on simple distillation, which appear to be unaffected or only slightly affected by potassium hydroxide. Then we have others formed during distillation with alkaline permanganate similar to the results in the experiments cited above; as these latter are formed by oxidation in the still we are justified in concluding that the already existing volatile compounds in the water were formed, from compounds similar to the latter, by natural oxidative processes. / Nitrites and Nitrates. The discussion of the causes of the presence of nitrites and nitrates in distilled water presents no great difficulties The quantities are small and in nearly every case the greatest amounts are toward the close of the run. In Still No. 1, Runs Nos. 1 and 2, we have the only cases where the nitrates are persistent in approximately the same amounts throughout the entire run. We can suppose that this nitrate existed in the water, most probably as ammonium nitrate, and the salt was dissociated in the boiling and carried over with the steam. The smaller quantity in run No 2 of this still is due to the reduction of some of the nitrates by the action of potassium hydroxide on the metal of the still ;f and also to some of the nitrate combining to form potassium nitrate. In Stills Nos. 2 and 3 in these runs we have the nitrites and nitrates appearing only at the close of the run, and this would tend to show that some nitrogen is oxidized and liberated when the still contents get concentrated. In Still No. 4, Run No. 1, we have some nitrites in the first two samples due to dissociation, and in Run No. 2 a less quantity in the corresponding samples, some having been eliminated by reduction. In Run No. 3 in all the stills we have the greatest production of both forms of oxidized nitrogen, and we would expect this from the oxidation in this run. As the concentration in*Chem. News, 1883, 19-20. f Leeds, Zeit , Anal. Chem., 1879, 428-430. 20 WILLIAM CULLEN UHLIG. creases so does the amount of oxidized nitrogen; this is partly an oxidation of ammonium salts, and from the high ammonia content of this run we would expect a high nitrogen oxide content. In Run No. 4 we have a still more vigorous oxidation but a falling off in nitrous and nitric oxides. This may be explained by the quantity of permanganate present, a large amount of alkaline permanganate acting on ammonium compounds giving nitrogen gas, while a less amount gives partly nitrogen and partly nitric acid. * In these runs we again notice the increase of nitrogen with the increase of concentration. As dilution is favorable to dissociation we must conclude that this oxidized nitrogen, in most cases, does not exist as such in the water but is formed during the process, concentration evidently contributing to its formation. N O T E S ON T H E A N A L Y T I C A L O P E R A T I O N S . The quantity used for the estimation of the ammonias was 500 c.c. In a few cases all the free ammonia came over in three tubes (50 c c. each), while in three cases it required nine tubes to bring over all the total ammonia; these three cases were all from Run No. 4, being samples 1 and 6 from Still No. 1, and sample No. 6 from Still No. 4. The average was less than six tubes. The difficulty, noticed in determining total ammonia in some natural waters, of getting a sharp end point on account of the ammonia coming over indefinitely was entirely absent, as the color in each tube was noticeably less than that in the preceding one and the final tube was invariably without color. In several instances the total ammonia (the sum of the free and albumenoid ammonias) was less than the' free ammonia. In two instances the figures were rejected on this account and some more of the sample analized but with the same result. This is probably due to oxidation of the ammonia. This explanation has been suggested by Tidy f in his objections to the Wanklyn process * Wanklyn, Jour. Chem. Soc, 1868. f Jour. Chem. Soc. ; 1879, 57. NITROGEN CONTENTS OF COMMERCIAL DISTILLED WATER. 21 The free ammonia distillates from Still No. 4, Run No. 2, gave a greenish color with Nessler reagent. Dr. Smart in the analysis of natural waters claims that this is an indication of decomposing organic matter of vegetable origin. As the organic matter of croton water is largely of vegetable origin we may infer that the compound producing this peculiar shade was carried from the still with the steam into the condensed wTater, and then again from the distilling flask into the Nessler tube. The quantity may have been increased or diminished in this double distillation but as we know nothing of the constitution of this body we cannot determine this. T H E CONDITION O F F R E E AMMONIA. In another series of experiments samples were collected and 50 c. c. Nesslerized direct, and then another portion of the sample analyzed for free ammonia. The following results were obtained: 4 RUN < A." Sample. Reaction with Nessler Reag. 3 4 5 6 7 8 9 10 A m m o n i a free i i < i t i i c 11 i i ( C i i i i i i a i i {i i •; .009 .0056 .001 .0007 .0016 .0038 .004 .0047 .008 .073 .112 .102 .128 i i <. i Ammonia in 50 c.c* .178 .112 .02 .014 .032 .076 .08 .094 i i u Free Ammonia. .0034* .0036 .0056 .0051 .0064 A66 .146 .116 .108 .0083 .0073 .0058 .0054 RUN 6 7 8 9 10 << < i c i i i i i i i RUN 4 5 6 7 i t < i i i i t present trace free i i * The figures in this column were obtained by dividing the amount of ammonia by 20. WILLIAM CULLEN UHLIG. 22 mple. 8 9 10 11 Reaction with Nessler R eag. Free Ammonia. Ammonia in 50 c.c, Ammonia free .102 .074 .094 .108 .0051 .0037 .0047 .0054 a a a a n a • RUN •174 .0087 124 .0062 n a .106 .0053 10 .162 .0081 a n 11 .112 .0056 These runs were made earlier than the preceding ones for the purpose of obtaining an ammonia free water, and no account was taken of the nitrogen contents of the water used and no attention was paid to the relative position of the heating surface to the level of the water in the still. The first 12 or 15% of the distillate was rejected before sampling, and the samples were taken at very short intervals, so, the conditions being different, there can be no comparison of curves to the curves of the runs already discussed. Run " A . " was with potassium hydroxide and Run " B." with potassium hydroxide and potassium permanganate, both runs being made in Still No. 1. The samples were Nesslerized and all rejected which showed a reaction; those which showed no reaction were analyzed for free ammonia. The results were so much higher than had been expected that two more runs were made on a different still (Still No 2), and in Run " C . " the last two samples which showed a reaction with Nessler reagent were analyzed in addition to those which reacted ammonia free; in Run " D . " every sample showed a reaction so the last five were taken for analysis. ' In Run " A." we have a set of tubes containing from .0007 to .009 part of nitrogen as ammonia; in Run " B . " from .0034 to .0064, and in Run " C . " from .0037 to .0058, all reacting ammonia free. In the analysis of a potable water we use a standard ammonia solution containing .01 part of nitrogen per cubic centimeter and we can see arid compare the color produced by Nessler reagent in 50 c.c. of water containing 0.1c c. of this solution, so all of the above amounts should be shown by the reagent; the figures for Run " A . " , Sample No. 6 (.0007) are 7 8 9 Ammonia present U (( << i < NITROGEN CONTENTS OF COMMERCIAL DISTILLED WATER. 23 less than our smallest standard, but the color produced by this amount of ammonia, when compared with an ammonia free water containing an equal amount cf the reagent, is easily discernable. From this we would infer that the ammonia in distilled water is there not all as ammonium hydroxide or ammonium salt but partly combined with some organic combination. This view is strengthened if we compare the figures of the first two samples of Run " C." and all the samples from Run " D." with those already mentioned. We have from .0053 to .0087 part all showing a reaction for ammonia; these are all less than the largest amount .(.009) not reacting, and the one showing a trace (Run " C.", sample No. 5) contains less ammonia than some samples showing no ammonia reaction, and more ammonia than three samples in Run " D.", which gave a decided reaction. Run " D . " never reacting ammonia free contains less ammonia in its minimum sample than do two of the samples which did not react, and approximately the same amount •(within . 0005) as seven of the nonreacting ones. The figures obtained in this series would indicate that '' Free Ammonia" exists in distilled water in two forms, one in inorganic combination reacting with Nessler reagent, and the other in organic combination not reacting with the reagent. The water in Runs u A . " a n d " B . " and the "ammonia free" samples of Run " C" contains only the " organic combination" ammonia, while that of the first two samples of Run " C." and all of Run " D . " contains ammonia, most probably, in both combinations. THE I N F L U E N C E O F T I M E ON T H E N I T R O G E N CONTENTS OF DISTILLED W A T E R A sample of water was drawn directly from the condenser and the nitrogen determined. The demijohn was corked and determinations were made at intervals of one week. Nitrogen as Free Ammonia. Alb. Ammonia. Nitrites. Nitrates. Water freshly distilled, .174 .096 .007 .09 '' at end of 1 st week, .198 .216 .002 .03 .002 .028 .06 "