摘要:

JULY 1905. Vol. XXX. NO. 352. ON THE ACTION OF SLIGHTLY ALKALINE WATERS ON IRON. BY CECIL H. CRIBB B.Sc. (LoND.) F.I.C. AND F. W. F. ARNAUD F.T.C. (Read at the Meeting April 5 1905.) IT has been the practice of one of us during the last ten or fifteen years when examining waters intended for use in steam boilers to make a trial experiment with the water using a slipof thin steel 4 inches long by 1 inch wide. A hundred C.C. of the water are placed in a boiling-tube the steel slip-brightly polished with emery not cleaned by chemical means-inserted and the whole kept at 100" C. (or as near to that temperature as a water-bath will bring it) for twenty-four hours after which the nature of the action is examined and its extent measured. In most cases a similar experiment is made at the air temperature.At the end of the time the deposit if any is washed off the surface of the metal and the total amount of iron in solution and in suspension in the water is estimated. Of course the test is a very imperfect one. I t does not reproduce in any way th THE ANALYST. conditions obtaining in a steam boiler; but some of those conditions such as the high temperature and pressure the rapid circulation and the constant influx of fresh feed water could not possibly be imitated without altogether prohibitive expense and trouble. It is simple and cheap ; requires but little of the water ; the conditions are under control and can always be exactly reproduced ; any peculiarities in the action are visible and its extent is easily and rapidly estimated.I n the somewhat limited number of cases in which such a method of testing is likely to be of use it has been found very fairly satisfactory. As a means of investigating the problems connected with the action of water upon iron and steel, there is a good deal to be said in its favour. The quantitative results obtained respond to almost every change in the conditions. They are affected by the chemical composition of the saline residue of the water by its gaseous contents by the temperature by the duration of the experiment by the concentration of the saline contents of the water by the physical or chemical condition of the surface of the metal by the electrical condition of the metal and by the presence or absence of light. It is not unreasonable therefore to conclude that in the majority of cases in which a natural water has a corrosive action on the metal of a boiler some indication will be given of the fact by this simple method of trial.In actual practice this conclusion has been fully borne out though opportunities for comparison between the behaviour of the water in the laboratory and that in the boiler are not so very frequent inasmuch as the majority of waters cause incrustation and do not corrode, Having had occasion to apply the method to artificially softened waters in which as is almost invariably the case the liquid was slightly alkaline owing to the presence of lime caustic soda or sodium carbonate it was found that when the alkalinity exceeded a certain amount no action whatever occurred and the surface of the metal was as bright at the end as at the beginning of the experiment.This inhibitive action is we believe generally known; but we have nowhere seen any recognition of the fact that it has its limits below which action does occur although the water is still alkaline much less any statement as to what those limits are. There seems to be a widespread impression that alkaline waters do not act upon iron at all and the fact is usually regarded as one of the great advantages of the chemical treatment of hard waters. We find however that with an alkalinity not exceeding a certain amount-dependent on the nature of the alkali present and to a less extent on the temperature -not only does action occur equal in intensity to what would occur in the absence of any alkali but that this action takes place in a very curious and interesting manner for which a t present we are unable to offer any altogether satisfactory explanation.The following table (I.) shows the effect of adding to New River water small proportions of sodium hydroxide lime-water and sodium carbonate. The quantities added are expressed in parts of the alkali or carbonate per 100,000 of water and also in cubic centimetres of decinormal solution per 100 c.c. and the amount of the action in milligramines of Fe,O,. The plan has obvious advantages THE ANALYST. 227 23.0 18.3 23-0 19.3 0 TABLE I.-TWENTY-FOUR-HOUR EXPERIMENTS WITH NEW RIVER WATER TO WHICH HAD BEEN ADDED SMALL QUANTITIES OF ALKALINE SALTS. General but on one side Broadill-defined patches.P r o j e c t i n g growths, No action. only. well defined. 15 9 9 A T 100" c. 11 AT AIlt TEMPERATURE. 1.0 2.5 5.0 10.0 15.0 0.5 1.25 2.5 5.0 7.5 10.0 15.0 1.0 2.5 5.0 10.0 15.0 20.0 30.0 40.0 4.0 10.0 20.0 40.0 60.0 1-85 4.6 9.25 18.5 27.7 37.0 55.5 5.3 13.25 26.5 53.0 79.5 106.0 159.0 212.0 Series 1.-Sodium Hydroxide. General. Local ; large patches. Local ; smaller ; well defined. Ditto. No action. Series 2.- Calcizinz Hydroxide. General. 25.5 Many small rounded 25.3 Long ill-defined streaks. 18.0 Smaller well - defined 21-0 Round sharply-defined 18-5 Three small patches. 24.0 No action. 0 patches. patches. growths. General. Few rounded patches.Streaks ill defined. Patches well defined. Very little action. No action. 9 9 , General. Many small spots. Well-defined streaks. Well-defined projecting patches. Large ill-defined patches. Few narrow streaks, sharply defined. Few small spots and one narrow streak. No action. 22.5 21.3 14.0 15.0 14.5 20.2 15.0 0 General. General but more pro-nounced in parts. Long patches ill defined. Ill-defined patches. Well-defined patches. Round well - defined spots. No action. 9.0 7.3 7.0 0 0 9.0 6.5 7.0 7.0 7.0 0 0 9.0 10.0 7.5 5.0 7.0 6.0 0 0 The peculiarity of the action consists in this-that whereas with quantities below those given in the table the iron slips are acted upon fairly uniformly all over rust being formed on the surface of the plates and then owing to their nearly vertical position settling slowly down to the bottom of the vessel; with larger but still very small proportions of alkali the action on the iron assumes a local character.At firs 228 THE ANALYST. the localization is extremely slight but with increasing alkalinity not only do the areas acted upon get smaller but they get to be more and more sharply defined ; finally when a certain strength of alkalinity is reached all action ceases and at this stage no deposit is observable in the liquid and the surface of the metal remains as bright and untarnished after twenty-four hours as at the beginning of the experiment, only a trace of iron being found in solution in the liquid and none in suspension, after several days.If in those cases in which action has taken place the amount of iron removed from the metal slips be estimated it will be found to be-with the smaller proportions of alkali that produce any degree of localized action-not markedly different from that produced by the water alone. As the alkalinity increases the quantity diminishes slightly but not at all in proportion to the area of action. Frequently when the area, acted upon is not part of the total surface of the plate it will be found that the iron removed has only diminished by about 5 per cent. of the maximum amount. When the higher limit of alkalinity is reached the quantity of iron removed diminishes with a sudden drop almost to nil-ie. to quantities which give no visible precipitate with ammonia.Coincident with the increased localization of the action is a tendency for the oxide of iron (or carbonate?) formed to adhere to the surface of the metal instead of slowly subsiding to the bottom of the liquid and this tendency goes on increasing until the rust projects from the surface like the growth of a mould, and these growths get firmer and more closely adherent as they get fewer in numbers and smaller. I n most cases they are of a dark-green or even blue colour beneath the surface but if examined in s i t z h without shaking they are generally found to be covered with a loose layer of bright-red oxide. A qualitative examination is sufkient to show that the greater part of the iron they contain is in the ferrous state and it seems fairly certain that a ferrous salt is first produced to be afterwards oxidized to ferric when it ceased to be in contact with the metallic surface.When rubbed off the growths leave evidence of their action on the metal in the form of dark-coloured patches which can be felt as distinct and generally sharply-defined depressions OD the surface of the plates. For the purpose of studying the phenomena under somewhat simpler conditions, a number of experiments have been made with distilled water instead of New River water. The dissolved salts in the latter are more or less precipitated by the alkaline hydroxides and salts which formed the subject of the experiments recorded in Table I. and the dissolved gases are greater and more variable in amount so that the alkali added does not represent the alkalinity of the solutions.With distilled water the dissolved gases art? the only uncontrolled constituents. We have determined for all the ordinary alkaline hydroxides and carbonates used or likely to be used in water softening the degree of alkalinity at which this local action begins and finishes and we have noted certain points in connection with the nature of the action which appears to differ both in kind and in intensity with different bases (see photographs). Table 11. shows the results of experiments made in the cold and at or just below the boiling-point with potash soda ammonia lime and baryta ; and Table 111. with sodium carbonate THE ANALYST. 229 17.8 13.7 13.7 12.8 0 0 0 TABLE 11.-TWENTY-FOUR-HOUR EXPERIMENTS WITH DISTILLED WATER TO WHIUH ALKALINE HYDROXIDES HAD BEEN ADDED.General but tendency to streaks. Long thick streaks well defined. Few short thin streaks. No action. 9 J 9 > f f 9 ) 9 8 , 0.25 0.50 1.0 2.5 5.0 10.0 15.0 0.25 0.5 1.0 2-5 5.0 10.0 15.0 0.125 0.25 0.50 1.25 2-5 5.0 7.5 1 2 4 10 20 40 60 1.4 2.8 5.6 14.0 28.0 56.0 84.0 0.46 0-92 1 -85 4.6 9.25 18.5 27-75 A 1 7 100" c. 11 AT Ar~t TEMPERATURE. Area and Natnre of Action. Series l.-Sodizm Hydroxide. Liquid turbid ; action general but slightly localized in streaks. Liquid turbid ; action general but more pronounced in parts. Long narrow patches. Narrow patches. Two small streaks.No action. 20.3 14.5 18-5 19-3 5.8 0 0 Series 2.-Potassium Hydroxide. Ill-defined streaks one side; general the other. Fairly general tending to streaks. Distinctly local ; irre-gular patches. One small projecting growth. No action. 5 > I 9 9 9 , Long thin streaks over Thin streaks less than Thin streaks still less. One large patch. No action. whole plate. above. 9 ) 9 9 )) Y 9 19.0 17-0 20-8 22.0 20.5 3.5 0 General but tending to streaks on one side. General but more pro-nounced in parts. Long patches. A very few patches. Little apparent action. No action. $ 9 ? ? ,? Series 3.-Calciz~m Hplroxide. General on one side; In narrow wavy streaks. patchy the other. In long patches.In long streaks. Well-defined streaks. Very little action. No action, 5.3 6.7 7.0 2.8 0 0 0 5.0 6.2 5.8 0 0 0 0 7.5 8-5 6.5 9.5 7.5 5.0 230 0.25 0.50 1.0 2.5 6-0 10.0 15.0 0.25 0-50 1.0 2.5 5.0 10.0 15-0 Almost general but 19.7 Streaks along whole 17.0 Well-defined streaks. 11.0 Very thin streaks. 11.5 Few dots only. 11.5 No action. 0 some streaks. length. I $ 9 9 0 2-13 4.27 8-55 21.37 42 -74 85-48 128-22 General but streaks Long patches ill defined Long streaks also spots Few thin streaks and spots (projecting). Few spots. No action. just apparent. 0.85 1.7 3-4 8-5 17.0 34.0 51.0 THE ANALYST. Series 4.-Barizm Hydroxide. Long narrow fern-like streaks.Broad zebra-like mark-ings. Thick streaks projecting One broad patch on one side; one spot on the other. No action. 9 ) I $ 9 ,, 15.7 20.0 14.4 11.3 0 0 0 General; one broad, ill-defined patch. Streaks and spots; form projecting growth. One thin streak. No action. No action. 5.3 7.3 5.2 0 0 0 0 7.5 7.5 7.0 6-0 5.0 0 TABLE III.-TWENTY-FOUR-HOUR EXPERIMENTS WITH DISTILLED WATER TO WHICH SODIUM CARBONATE HAD BEEN ADDED. 0.25 0.50 1-0 2.5 5.0 10.0 15.0 20.0 30.0 40.0 50.0 1.32 2.65 5.3 13.25 26.5 53.0 79-5 106.0 159.0 212.0 265.0 A r 100" c. Extent and Kature of Action. General ; darker streaks visible. Nany long thin streaks, bright in between.Long patches. Few broad patches. Broad patches of smaller Few patches. Broad patch one side, one spot on the other. Long broad patch one side spots on the other. No action. $ 9 9 3 area. 9 9 9 , 22.0 15.0 16.3 14.0 14-5 16.5 17.5 16.5 18.5 0 0 A r AIR TEMPERATURE. Extent and Rature of Action. Ill-defined patches. Long broad patches. Long patches. Streaks. Large well-defined spots. 9 s 5 > Very sharply - defined Few spots. spots. No action. 8 -5 7.5 7.0 5.5 6 -0 6.5 7.0 6 -0 THE ANALYST. 231 Though each alkali has a more or less marked individuality as regards the pattern formed on the slips (shown fairly well in the photographs) and even the colour and tenacity of the growths of oxide the same sequence of changes occurs with eaoh as the proportion of alkali rises.The figures in the tables show that the amount of iron removed is practically independent of the area attacked so that the more restricted the area the deeper the excavation that occurs there. We have not yet made any definite experiments on the point but from one or two isolated observations it would seem probable that as the area of action gets less and the products of the action assume the form of growth-like projections from the metal the proportion of ferrous iron increases. At all events with a water of neutral reaction we could get no evidence of the presence of ferrous iron at all while with the projecting ‘( growths ’’ ferrous iron was in greater quantity than ferric.The colour of the oxide supports this view inasmuch as the rust from neutral distilled water is always reddish-brown while with the dilute alkaline solutions and more especially with baryta it is of a dark blue-green colour. I n one case a growth of a light blue colour was observed-a not very surprising result inasmuch as Nicolardot (Journ. Chem. SOC. 1905) has shown that a white FesOs can be prepared. The fact that alkaline solutions are capable of acting in the way described is of considerable theoretical interest as well as of possible practical importance. To deal with the purely scientific aspect of the matter it is necessary to consider the explanations that are usually given of the rusting of iron. According to the view originally proposed by Traube (Berislzte xviii.1877-1887) and more recently advocated by Dunstan (Proc. Chem. SOC. 1903 p. 150) oxygen and water react with the metal both being necessary and ferrous hydrate and hydroxyl are formed. The latter reacts again with iron forming a hydroxide of the formula Fe20,(0H),, which according to Dunstan represents the ordinary composition of iron rust. Neither of these workers has been able to detect the presence of hydrogen peroxide, but this is (to them) explained by the rapidity with which that substance is acted upon and destroyed by metallic iron. Traube asserts as a well-known fact that iron is not oxidized when put into alkaline solutions and states that it decomposes hydrogen peroxide when in such a solution without being itself oxidized. Moody more recently (Proc.Chem. SOC., 1903 p. 240) has found that pure freshly-distilled hydrogen peroxide does not; act on iron unless an acid such as carbonic acid is also present. Dunstan (Zoc. cit.) explains this passivity of iron towards alkaline solutions by saying that all the substances which inhibit rusting do so by virtue of their power of decomposing the peroxide. If therefore dilute solutions of alkalies do act on iron either hydrogen peroxide is formed in spite of the presence of the alkali (in which case according to Traube it should not act on the iron) or the hydrogen peroxide is not formed and the hydroxylation theory becomes untenable. It must be borne in mind however that Dunstan was dealing chiefly with the atmospheric corrosion of iron and that he used apparently highly purified iron in his experiments.To ascertain whether small quantities of hydroxpl could survive the action of the slight alkalinity of the solutions employed we made up two solutions containin 232 THE ANALYST. respectively 0.15 and 0.7 parts of hydrogen peroxide per 100,000 of distilled water ; 100 C.C. of each strength were heated in a water-bath for twelve hours and two similar quantities to which 1 C.C. of TG sodium hydroxide had been added were treated in the same way. At the end of the time some of the original solutions which had not been heated and the four solutions that had been heated were tested for hydrogen peroxide and abundant evidence of the presence of that substance-or at all events, of a peroxide-was obtained in every case.The peroxide is not destroyed by even 5 per cent. soda after heating in the water-bath for four hours. From the way in which the experiments were carried out as little as 0.015 per 100,000 could have been detected. The action on iron (as measured by Fe,O removed) was not materially increased by the presence of the hydroxyl and when the added alkali was raised to 15 C.C. it was entirely inhibited just as if no peroxide had been present, although the latter could be easily detected at the close of the experiment (twenty-four hours). I t seems quite plain therefore that the hydrogen peroxide theory as far as this particular form of rusting is concerned cannot be sustained unless as a mere pious belief unsupported by any sort of proof other than the indirect one of analogy for it cannot be proved to be present under ordinary circumstances and when added it neither increases the action nor gives rise to it when it would otherwise not occur.The other theory of the rusting of iron has its most recent exponent in Moody (Proc. Chem. Soc. 1903 pp. 157 and 239) and is to the effect that the primary action is due to the interaction of iron and carbonic acid and that rust is formed by the subsequent oxidation of ferrous salt. He shows that when iron is exposed to water and oxygen-previously freed as far as possible from carbon dioxide-the volume of oxygen remains practically unchanged but that on admitting carbon dioxide it diminishes rapidly and the iron rusts. I t is as difficult to believe that carbonic acid is any more necessary to the reaction than hydrogen peroxide in view of the experiments recorded in Table II., more especially those in Series 3 and 4.To make quite sure that the amount of carbonic acid gas dissolved in the distilled water was not greater than the quantity of baryta added a special experiment was made in which 0-5 C.C. of baryta water was added to 100 C.C. of the water and a steel slip immersed in the liquid At the end of the twenty-four hours when the usual characteristic action had taken place, titration of the clear solution using phenolphthalein as indicator showed alkalinity still present corresponding to 0.4 C.C. & Ba(HO), a loss of only 0.1 C.O. In view of the extreme insolubility of barium carbonate in hot water it is not unfair to assume that neither free nor combined carbonic acid took part in the reaction.Further it would appear from another experiment that in the presence of what must be regarded as a large excess of free carbonic acid iron is not acted upon for we find that sodium bicarbonate acts in a similar way to the normal carbonate if only enough of the salt is present-ie. a stage is reached when the iron is entirely untouched after being immersed in the solution either hot or cold for twenty-four hours. We have distinct evidence that localized action occurs with this salt but in a far less striking manner than is the case with the normal carbonate. With 10 C.C. o THE ANALYST. 233 8 normal solution the action is general but at the same time small; with 15 C.C. oE normal bicarbonate solution per 100 of distilled water the plate remained absolutely bright and untarnished after being in the cold for twenty-four hours and it gave no evidence (judging by the mercuric chloride test) of the presence of any normal carbonate at all-Le.the whole of the bicarbonate remained unchanged. A hot solution of the same strength was also without action although much normal carbonate was formed-ie. a large amount of carbonic acid must have been present, or at all events must have passed out of the liquid. TABLE IV. C.C. of Bicarbonate (Normal Solution). 3icarl)onate per 100,000 Vols. of Water. 1 5 7.5 10 15 84 420 630 840 1,260 Mgms. of Iron Removed. 8.5 action general. {Kid 12.5 {Gid i) marked. hot 13.5 slibht tel;hency to hot l2*5} local action.hot 0 ditto more According to Moody oxygen and water without carbonic (or other) acid will not cause rusting. According to Dunstan carbonic acid and water without oxygen is also without effect while the various experiments we have made certainly indicate that carbonic acid is not necessary though there can be no doubt that it plays an important part under the ordinary conditions of rust formation. Some further experiments go to show that the actual guantity of the dissolved gases has very little to do with the matter for we find that after twelve hours at water-bath temperature no further gases are evolved from water ; and yet when distilled water was first boiled and then heated for four hours on the water-bath in an open vessel the action on iron after twenty-four hours was 25.3 mgms.as compared with 22.2 with the same water not previously heated. Similar experiments with water containing 2.5 C.C. :G NaHO gave when heated beforehand 13.2 mgms. and when not so heated 17.8. As water entirely free from gases will not act on iron it seerng probable that a minute quantity of gas is necessary to start the reaction which then goes on continuously. It is obvious however that under the term ‘‘ rusting ” a number of different reactions are included and much more work is required before a satisfac-tory explanation of them all is likely to be forthcoming. I t will of course at once be objected that the experiments were made with impure iron-in this case steel We have not yet been able to get in the form of slips any iron of known purity but we have found that sheet-iron and Siemens steel and also steel wire alike give the characteristic action within the limits of alkalinity we have established.In summing up the result of his work (Zoc. cit.) Dunstan makes the statement (whether as the result of experiment or as a commonly-accepted opinion is not clear) that when the iron is impure or when another metal is present electroljrti 234 THE ANALYST. action occurs ; the electro-positive metal suffers oxidation and hydrogen gas is evolved. If the latter is a necessary condition of electrolysis then no electrolysis takes place in connection with the phenomena under discussion ; for if a boiling tube, prolonged at what would be its closed end by a narrow tube of .& inch internal diameter and about 4 inches long closed at its distal extremity be inverted over a steel slip immersed in water of slightly alkaline or neutral reaction and heated in the water-bath for twenty-four hours the gas that collects in the narrow portion of the tube does not differ in volume by & C.C.from that given off by the water alone in the absence of the steel slip. To turn to the practical aspect of the question it is quite obvious that assuming that these curious phenomena we describe really take place under the conditions which obtain in boilers feed-water tanks etc. they would constitute a very serious objection to the use of chemical processes of water-softening and also to the employ-ment of alkaline waters of any sort for steam-raising purposes for even the smallest degree of alkalinity in a feed-water will by the concentration that takes place in the boiler be raised to thirty or more times its original amount (Cribb ANALYST July, 1900)) and therefore the whole range of alkalinity with which action occurs is well within the conditions that would occur in practice.A glance at some of the plates is sufficient to show that even in the space of twentyfour hours (and especially where the area of attack is small) the metal is eaten away to quite an appreciable depth. Now it is generally admitted that (leaving on one side the question of incrustation) the rusting action is a continuous one although it slackens down considerably after the first onset and the following figures (Table V.) show that this is so and therefore either hot or cold it is quite plain that given the right strength of solution most serious pitting might be caused.TABLE V.-ACTION OF DISTILLED AND NEW RIVER WATER DURING VARIOUS PERIODS AT THE AIR TEMPERATUBE. Period. 1 2 hours. 24 9 , 48 9 , 168 9 , 240 $ 9 Distilled Water. Mgms. of Fe,O,. New River Water. Mgms. of Fe,O,. 6.5 8.0 25.8 39.5 --9.1 14.4 56.0 -It is important therefore to inquire how far if at all this could take place in practice. As regards the metal we find that sheet-iron Siemens-Martin steel with the 6 ‘ skin ” on as actually used for making boilers and the steel with which most of the experiments in connection with the paper were made act practically alike. Even the 1 L skin ” does not prevent the local action in any way and the action goes on at the same rate or even more energetically therefore the nature of the metal is no bar to the action THE ANALYST 235 The action of the dissolved salts likely to be met with in natural waters other than alkaline salts is similarly without any inhibiting effect.As shown in Table VI. the figures represent mgms. of Fe,O removed from the steel plates after immersion in solutions of the salts named and of the strengths specified for twenty-four hours at the water-bath temperature. TABLE VI. Strength of Solution. Sodium chloride . . Sodium sulphate . . . Sodium nitrate . . Sodium phosphate . . . Potassium nitrate . Calcium chloride . . . Calcium sulphate . . Magnesium chloride . . . Magnesium sulphate .. . Ammonium chloride . . . Ammonium carbonate . . . . . . Distilled water alone . New River water alone . 10 Parts per 100,000. 19.0 18-5 18.0 18-0 22.0 17.0 25.5 28-0 25.5 12.0 30.0 19.0 25.0 100 Parts per 100,000. 35.0 28.0 61.5 16.5 31.5 54.5 23.5 26.5 47-5 22.0 22.0 30.0 Of course only the salts of the alkalies and alkaline earths could be present under the conditions which give rise to local action. In the case of boilers especially high-pressure boilers the high temperature prevailing increases the amount of decomposition undergone by practically all the saline constituents of natural waters (Cribb ANALYST 1900) and therefore it would naturally be expected that the possibility of pitting would increase rather than diminish as the temperature rises.To ascertain whether the local action of alkalies actually does take place at temperatures above 100" C. we have so far only made one experiment. I t consisted in heating a steel slip immersed in distilled water con-taining 4 parts per 100,000 of sodium hydroxide to a temperature of 110" C. for some hours. We hope however to have the opportunity before long of employing a still higher temperature; but there is no reason for doubting that the action at temperatures above 100" C. will be more and probably considerably more than it is at the normal boiling-point . Even if the higher temperatures occurring in steam boilers brought about no increase in action it is quite obvious that the rate of corrosion occurring in our laboratory experiments might have the most serious consequences even after a com-paratively short period of working.Take for instance a 100 horse-power boiler with GOO square feet of water surface Assume that it is fed with New River water which according to Table VI. removes a little over 25 mgms. of Fe,03 in twenty-four hours. This would mean 6.7 ounces of metallic iron removed in a single day and night. Each square inch of +inch boiler plate only weighs 1-22 ounces. The usual action characteristic of that degree of alkalinity took place 236 THE ANALYST. There are on the other hand some facts which would tend to lessen if not altogether to prevent pitting in a boiler and there can be no doubt that these are largely operative otherwise pitting would be universal with all softened waters from the cause under discussion.The first is that most softened waters give a deposit when heated which is generally sufficient to cover the plates and tubes with a more or less protective coating. If it is not steps are taken to produce such a coating by what may be called artificial means. Secondly the rapid circulation especially in the case of tubular boilers would almost certainly prevent anything in the nature of the growths seen in the laboratory experiments. There are however in most boilers and especially in the non-tubular kinds certain places where the circulation is very much slower and it is in the experience of most engineers that it is in these spots that pitting most frequently occurs. It would be unnecessary to refer to this point at such length but for the fact that in using the steel slips we have found that when local action has started in certain areas and the growths attached are rubbed off the surface of the slip action does not always commence again in the same place.When the slip was polished with emery before replacing it in the liquid the action never commenced at the same spots. Finally there is another very interesting reason why action is less likely to occur in boilers and closed vessels and that is that it is unquestionably less energetic in the dark. The figures we bring forward in support of this are neither so numerous nor SO striking as we should like but we have never made a single experiment in which the stimulating effect of light is not apparent-that is to say in which the amount of corrosion has not been less in the dark than in the light.The following table shows side by side the result of exposing the same kind of steel to the action of the same liquid for the same time in ordinary daylight and in absolute darkness. The figures represent as usual mgms. of Fe,O removed. TABLE VII. Distilled water . . . . . Tax - w a t er . . . Di&illed water MgCI, 20 per 100,000 . 9 1 , NaHCb, 84 . I , Na,CO, 53 . Y , Ca(HO), 9-2 . 9 , NaHO 4 . Distilled water . . . . . New River water Distilled water + 4NaHO per 100,000 . . . Duration. 12 hours 24 hours 9 ) 9 , ? ) 12 i;7urs 240 hours 9 9 360 hours In the Dark. 4.5 7.5 5.2 6.0 5-0 7.2 2.0 9.5 51.7 16.0 In the Light.6-5 13.5 7.2 8 5 7 -0 11.5 4.2 39.5 56.0 48. THE ANALYST. a37 The difference is considerably greater than would appear from the above inas-much as during somewhat less than half the total duration of the experiments (except of course the first) all the liquids were necessarily in the dark. The further investigations we propose to undertake will we hope throw more light on this interesting point. DISCUSSION. The PRESIDENT (Mr. Bevan) said he should like to know more definitely how the authors accounted for the corrosion taking place in patches except on the electroo lytic theory which although their experiments seemed conclusive he did not regard as being entirely disposed of. It would be interesting to know the actual composition of the iron because if the iron were chemically pure it would be difficult to see how the action could take place in patches ; while if it were not chemically pure it was pretty obvious that in certain places there might be a little more or a little less of one constituent or another which of course would give rise to electrolytic action.I n the case in which the corrosion took place on one side only was the plate exactly verticd? Mr. CRIBB said that he did not mean to imply that the action took place neces-sarily on one side only. It was curious however that this did sometimes happen, He did not think the difference with regard to light would be sufficient to account for it. Mr. ARCHBUTT thought that it would have been better if the authors had started with steel boiler plate of known analytical composition and had used strips from the same plate throughout the experiments.He also thought that it would have been better if the water had been freed from gases as would be the case in a steam boiler. Although the authors had found corrosion to be so general in their experiments with alkali whether in the caustic form or in the form of carbonate he did not think that this was the case in actual practice for if it were one would expect the boilers in the country generally to be in a much worse condition than they really were in. In certain cases however serious corrosion did occur from causes which at present could not be explained and therefore any investigation which would help to throw light on these obscure causes would be useful.I n a case within his knowledge in which pitting had occurred in boilers using Manchester water it had been remedied to a large extent by scraping out the magnetic oxide that had been formed in the pits, brushing these places over with petroleum and using in the boilers about 1 pound of sodium carbonate per 1,000 gallons of water evaporated. The fact that this prac-tically stopped the corrosion seemed to indicate that the action of the alkali in a boiler was not quite the same as under the conditions of the author’s experiments. I t might be mentioned that pitting and corrosion were nearly always local pitting being sometimes so serious in boiler tubes as to result in holes right through the tubes. As a rule in boiler plates also corrosion took the form of local pitting.He had at present under observation however an extraordinary case in which the barrels of locomotive boilers were corroded all over below the water-line. The boilers took water from several different sources and all of these had been investi-gated but no reason for the corrosion was a t first apparent. No mineral or fatty Possibly the degree of exposure to light had some influence 238 THE ANALYST. acid could be traced as the cause. None of the waters were artificially softened and no chemical or disincrustant was used in the boilers. The water of which the largest quantity was used came out of the Derbyshire limestone and contained about 21 grains per gallon of total dissolved matter including 7.55 grains of calcium carbonate 3.99 grains of magnesium carbonate 1.94 grains of magnesium sulphate, 4.74 grains of sodium sulphate 1-19 grains of sodium chloride and 0.95 grain of silica.The free carbon dioxide amounted to only 0-52 grain and the dissolved oxygen to 0.66 grain. I t would not be thought that such a water would seriously corrode boilers He was however forced to conclude that such water was liable to corrode bare boiler plate and that unless the plates are protected by a scale, corrosion was more or less bound to occur. The scale which this water formed was particularly soft and porous and did not protect the plates. I t might be also that there was some organic acid in the water. Both of these causes would be met by softening the water with lime which would remove the greater part of the car-bonates but would leave the sulphates behind thus encouraging the formation of a small quantity of thin hard scale whilst the excess of lime would neutralize any organic acid that might be present.An experiment on these lines now had been in progress for several months and perhaps at some future time he might be able to communicate the result to the Society. Meantime he should be interested to hear if the authors could suggest any other explanation of the corrosion. Mr. DIBDIN said that in connection with the boilers at the pumping-stations of the London County Council it was at first thought that magnesium chloride seemed to facilitate pitting as much as anything. This was particularly the case with some deep well waters and especially with that from a trial well sunk at Crossness.The water there was brackish and contained a good deal of magnesium chloride derived, no doubt from tidal water entering through a fault in the strata. The most effective preventive of this action on the boilers was found to be the use of an alkaline solution of tannin made by boiling bark with caustic soda. The pitting seemed to take place with almost every kind of water even when distilled in the ordinary process of condensation-New River water tidal water or water supplied say by the West Middlesex or East London Water Companies. H e had often thought over the matter but must confess that the only conclusion he had been able to arrive at was that the corrosion was due entirely to electrolytic action and that it would therefore, vary with the character of the iron.It seemed to him that if the authors’ experiments could be made in the reverse direction-namely by taking varying qualities of iron and water of constant composition-a good deal of light might be thrown on the subject. Mr JULIAN L. BAKER suggested that bearing in mind the medium in which the protuberant growths were developed these possibly consisted largely of an anaerobic form of mould or bacterial growth. H e should like to hear whether any of the experiments had been made under aseptic conditions. Mr. CRIBB said that in every case where these growths developed in the cold they also occurred in the corresponding experiment at 10Go C. Mr JOHN WHITE said that a short time ago he had submitted to him a piece of an iron branch water-main 3 inches in diameter.It had been used for about si THE ANALYST. 239 months for a fairly pure public supply which however contained some vegetable matter. After that time the pipe had become so “made u p ” that a lead pencil would not pass through it. The only conclusion he could come to was that as the water contained free oxygen and carbonic acid oxide of iron was formed which attached itself to the walls of the pipe and this caused the deposition of organic matter as well. The deposit was not an ordinary incrustation ; it contained about 60 per cent. of oxide of iron the remainder being water and organic matter. Mr. J. H. B. JENKINS inquired whether the authors stated in their paper the extent to which corrosion took place with New River water and distilled water respectively to which no alkali had been added.The addition of caustic soda to London water would form carbonate with the bicarbonate present in the water so that at a certain stage there would be present only the normal carbonates of soda and lime. Perhaps the authors could indicate in the tables when that stage was passed and caustic alkali could be assumed to be present in the water. He had noticed in examining polished steel sections microscopically that the sulphide dots or threads always found in commercial mild steels became nuclei for corrosion by which the sulphide was decomposed. After rubbing such corroded sections afresh there would probabIy be no longer any sulphide spots left in the earlier positions, but others would be exposed at the new surface so that the next points of corrosion would be expected to appear at fresh places.I n connection with the experiments in the light and in the dark he should like to hear whether the temperature was about the same in each case because generally one would expect the temperature to be somewhat higher in the daytime. Mr. CRIBB said that the results of experiments with New River and distilled water alone were given in Tables VI. and VII. He thought that the point when all the bicarbonate in the New River water was converted into carbonate would probably have been either reached or passed when 5 C.C. of i\ alkali had been added. With regard to the experiments in light and darkness various methods had been adopted to insure equal temperature or nearly so.Either the dark experiment was made in an incubator at 20” C. (ie. was slightly warmer than the other) or the two cylinders were placed side by side one being under a cardboard cover and the other exposed. He thought that the differences of temperature would in most cases be far too small to produce any appreciable effect on the results. Mr. JENKINS continuing said that in the boilers of which he saw most-namely locomotive boilers-the possibilities of galvanic action were undoubtedly great because the fire-box inside the boiler was of copper. In nearly all cases, however unless the water was quite soft a protective scale was formed which kept the water from contact with the metal surfaces. Very often they met with cases in which if the scale had been broken or was not sufticiently hard to be impervious, galvanic action occurred.When pitting and corrosion took place it was generally at those parts of a boiler where the circulation was least vigorous. H e thought that this was due to the fact that in those places the gases which came out of solution when the water was heated remained in contact with the metal for the longest time. In a locomotive boiler this was generally below the level of the fire near the foundation ring or along the belly of the boiler 240 THE ANALYST. Mr. HERBERT E. BURGESS suggested that the local character of the action might be accounted for by the presence in certain places of a thin coating of grease. If the plate happened to be touched with the fingers the grease therefrom would form a protective coating and he suggested that it would be well to wash the metal with chloroform before the experiments were made.The PRESIDENT thought that the peculiar pattern oE the corrosion was against Mr. Burgess’s theory. Mr. CRIBB in reply said that he himself should be inclined to consider that these phenomena were due to electrolytic action were it not for the difficulties in the way of accepting that explanation. With a material like Siemens-Martin steel it must be due to the distribution of the different impurities (i.e.? substances other than iron) throughout the mass of the metal and therefore assuming that one alkali would have the same chemical action as another there was no particular reason why different alkalies should produce different electrolytic effects ; and therefore the patterns should be the same.They were however very different. As far as he knew the distribution of the different constituents of steel was a microscopic and not a macroscopic matter and therefore any patterns due to electrolysis might be expected to be microscopic and not macroscopic. Mr. JENKINS remarked that lines of impurity such as sulphide threads would be found drawn out in the direction of rolling of the steel plate. This happened to be the same direction as the corrosion streaks developed in the author’s experiments. Supposing that other strips were cut so that their long dimensions were transverse to the direction of rolling would the direction of corrosion streaks be altered ? Mr. CRIBB said he had noticed with Siemens-Martin steels in the case of which the ‘‘ skin ” was allowed to remain on the front and the sides had been cut through, the local action took place on the sides as well.There was no connection between the pattern on the sides and that on the front. That was one of the difficulties standing in the way of the electrolytic theory. Practically the same character of pattern was obtained with different pieces of metal when the same alkali was used. With regard to the analytical composition of the metal many of these experiments had been made a long time ago when testing the value of the method as a purely empirical process for getting at the corrosive powers of different waters. The slips of metal used however were in all cases of the same composition and they might be able to malyse one later on.Mr. Dibdin’s suggestion would be rather dificult to carry out as it was not easy to get in a convenient form samples of iron of varying and known composition. They had however been able to show that with actual steel boiler-plate the same phenomena occurred. They did not of course contend that every time an alkaline water was used in a boiler this local action would necessarily take place and they had pointed out the causes which would be likely to prevent it. I t was curious that in spite of the widely-accepted opinion as to the corrosive effect of magnesium chloride there was scarcely any evidence in its favour except that derived from the analysis of boiler waters themselves. A 1 per cent. solution of magnesium chloride when heated had no more action on metal than plain ‘distilled or tap water.As a matter of fact it did not decompose when boiled. He admitted that it apparently did decompose under high pressure but he had no THE ANALYST. 241 been able to make any laboratory experiments under those conditions except one or two conducted in an autoclave which however failed to yield any conclusive results. Mr. Jenkins’s suggestion that the impurities formed nuclei which were removed by rubbing afforded a very reasonable and probable explanation. The rubbing however was very slight and moreover the action especially when it was beginning was hardly of a nature suggesting the existence of nuclei,:: inasmuch as it commenced in the form of patches without anything in the nature of a nucleus.It was only when the alkalinity was reaching its higher limits that small dots were produced. He did not think that grease could have had any influence in these experiments because all the plates were systematically and carefully cleaned with emery and were never touched afterwards except at the edges. The strength of the alkali too would in some cases be sufficient to saponify any slight traces of grease. He was afraid he could not suggest any explanation of the curious case of corrosion mentioned by Mr. Archbutt. ADDENDUM. Mr. Archbutt writes that having by the courtesy of the authors been permitted to read the paper more carefully since the meeting he does not feel convinced that the chief corrosive agent was not after all carbonic acid the action of which ceased when the amount of alkali added was sufficient.The authors do not appear to have deter-mined the amount of free CO originally present in the water and therefore it is impossible to say what relation the amount of alkali which was found to stop corrosion bore to the free CO,. Is it not possible that in highly-dilute solutions bicarbonates may be mildly corrosive to iron and steel and need the presence of a certain amount of free caustic or nionocarbonated alkali to restrain their action? In the experiment with baryta the amount added was just enough to neutralize 1.1 parts of free CO per 100,000 parts of distilled water. Supposing more than this were originally present in the water some of the barium carbonate may have been held in solution and in the subsequent titration have neutralized acid.Therefore it seems that the experiment as described in the paper does not necessarily prove that the baryta removed all the carbonic acid. To prove their case conclusively the authors would need to show that distilled water which has been freed from every trace of carbon dioxide and oxygen by boiling, and has been prevented from reabsorbing any of these gases is more corrosive to iron when slightly alkaline than when no alkali is added. Referring again to the case of locomotive boiler corrosion which had formed the subject of some of his previous remarks he had within the last few days had an opportunity of inspecting two locomotive boilers one of which had been supplied during the last nine months exclusively with the water of which the analysis had been stated and the other with the same water after softening with lime. The difference in the appearance of the two boilers was striking. The boiler which had used the untreated water was already extensively corroded whilst the boiler which had used the treated water was in very good condition being only pitted in a few * Since reading the paper experiments have been made Kith steel wire under the microscope and most distinct evidence of nuclei round which the action seemed t o centre has been obtained 242 THE ANALYST. places here and there. Both boilers at the commencement of the experiment were nearly new and quite free from corrosion. Mr. Cribb in answer to the first point in Mr. Archbutt’s ‘( addendum,” writes that the solution in the experiment referred to was titrated directly iising phenolphthalein, which does not react with barium carbonate. Also that in Table II. Series 4 an experiment is recorded in which local action occurred the barium hydroxide added being equivalent to no less than 5.5 parts of CO per 100,000. The boilers referred to in the final paragraph obviously confirm the author’s contention that local action and pitting may occur in actual practice when artificially softened waters have to be employed. The action of the water of which Mr. Archbutt gave the analysis before treatment was probably due solely to the small proportion of dissolved gases. The water in boilers is always being reinforced with fresh feed-water and consequently there is always enough gas in solution to cause action either general or local. I n this case there was not enough of the scale-forming compounds to form a protective coating. Water really free from CO and 0 does not act on iron at all

ISSN:0003-2654    

DOI:10.1039/AN905300225b

出版商:RSC    

年代:1905

数据来源: RSC

摘要:

242 THE ANALYST. ABSTRACTS OF PAPERS PUBLISHED IN OTHER JOURNALS. FOODS AND DRUGS ANALYSIS. The Determination of Ammonia in Milk. W. N. Berg and H. C. Sherman. (Jouriz. Amer. C7zem. SOC., 1905, xxvii., 124).-The authors believe that the determination of ammonia affords a valuable indication of the extent of the decomposition of the proteids in milk, and have conseqaently elaborated a method for its determination. Distillation with even exceedingly dilute sodium carbonate under atmospheric pressure yields erroneous results, due to the breaking-down of the organic nitrogenous compounds during the boiling, and hence recourse was had to a modified Eoussingault-Schaffer method (Amer. Journ. PhysioZ., 1903, viii., 330). I n this method 50 C.C. of the milk and 50 C.C. of methyl-alcohol are distilled with 10 grams of sodium chloride and 0.5 gram of sodium carbonate in a two-litre flask under reduced pressure.The flask is closed by a two-holed rubber stopper fitted with an inlet-tube, through which air may be admitted, and a distilling head connected to two absorption cylinders, placed in ice-water and connected with a suction-pump by an empty wash-bottle. I n the cylinders are placed 25 C.C. of sulphuric acid diluted to 40 C.C. with water. Boiling is started by bringing the flask into a water-bath at a temperature of 60" to 65" C., after which the temperature may be allowed to fall a little. With a pressure of 50 millimetres boiling should continue at from 56" to 62" C. After distilling for fifteen minutes, boiling is stopped; the volume of the distillate is noted, and this is then diluted to 250 c.c., and titrated with & sodium hydroxide, using congo red as indicator.As the presence of methyl-alcohol affectsTHE ANALYST. 243 the end-point, a correction must be made for it. This varies with the quantity of alcohol distilled over : for 30 to 39 C.C. of distillate it amounts to 0.4 C.C. of alkali ; for 40 to 49 c.c., to 0.5 C.C. ; and for 50 to 60 c.c., to 0.6 C.C. The volume of distillate obtained is usually between 45 and 55 C.C. A number of check experiments on water and milk containing known quantities of ammonium chloride showed the method to give exceedingly accurate results, the error generally being less than 0-1 C.C. of & alkali. I n applying the method, it was found that if the addition of sodium chloride, which is used to diminish the hydrolytic dissociation of the sodium carbonate, is omitted, the results for fresh milk remain practically unaffected, whilst those given by stale milk are considerably higher.I t is suggested, therefore, that by making a distillation both with and without the addition of sodium chloride, an indication of the easily decomposable proteids, as well as of the ammonia present, may be obtained in the case of stale milk. A number of tests in which the above ammonia method was compared with the '' alkalinity '' method of Richards and Woodman Air, Water, and Food," p. 152), showed that after making certain corrections for the alkalinity figure, it would be brought into comparative agreement with the ammonia figure in the case of fresh milk, yet for stale milk the ammonia figure increased much more than did the alkalinity figure, and hence the first afforded a truer indication of the state of the milk.Preliminary experiments on the development on standing of acidity and ammonia in milk showed that there is no necessary connection between the two, since one may increase at a much greater rate than the other. A. G. L. On the Specific Rotation of Salts of Casein. J. H. Long. (Jozmz. Amer. Chem. SOC., 1905, xxvii., 363.)-The specific rotation of very thoroughly purified casein from cow's milk when dissolved in different alkali solutions was determined, I t was found that for 5 grams of dry casein 45 C.C. of $; alkali hydroxide were needed to give a solution neutral to phenolphthalein; but half this quantity of alkali sufficed to give a practically clear solution.All readings were made at 20" C., using 5 grams of dry casein dissolved in 100 C.C. of the alkali solution. The results obtained are given in the following table : Alkali Hydroxide. C.C. of Solution Used. Sodium hydroxide ... ... ... 2 7 9 , 9 ) , l 1 1 1 1 ... ... ... . . . ... . . . ... ... ... Potassium hydroxide ... ... Lithium hydroxide ... ... ... Ammonium hydroxide ... ... 9 , ... ... ... 7, 45.0 22.5 67.5 90.0 45-0 45.0 22.5 45.0 - 103.5" - 95.2" - 107.6" - 111.8" - 104.4' - 100*8" - 94-8O - 97.8" I n connection with these results, it may be noted that BQchamp (Bull. Soc.244 THE ANALYST, Chim. [3], iv. 18 ; ii. 152) obtained for casein dissolved in water [uIj = - 117*7O, is approximately equal to [a], = - 115".A. G. L. The Baudoin Reaction in Human Milk. Dr. Engel. (Chem. Zeit., 1905, xxvii, 363.)-As it has been clearly shown that in the case of cows and goats this reaction cannot always be obtained in the milk of animals which have been fed with sesame oil cake and even with the oil itself, experiments have been made on the same lines with the milk of wet nurses. After a dose of sesame oil had been given in salad, the furfurol reaction could be observed in the milk within a short time. After six hours it disappeared, but re- turned after ten, was more or less distinct for another four or five, at the end of which it finally disappeared. These results confirm the uncertainty of the reaction as a test for the feeding.A. N. C. The Analytical Values of Goat's Butter. L. Hardy. (BUZZ. SOC. Chim. Belg., 1905, xix., 13.)-Two samples of goat's butter examined by the author gave the following results, who calls attention to the very low Reichert value of this butter: a t looo C. Reading a t 40" C. Value. per Cent. Meissl Value. Specific Gravity Butyro-refractometer Crismer's Insoluble Fatty Acids Reichert - 1. 0.8659 42 50 87.46 20.6 2. 0.8650 42 52.5 87.77 21-4 C. A. M. Detection of Artificial Colowring Matters in Fats. H. Sprinkmeyer and H. Wagner. (Zeit. Untersuch. Nuhr. Genussmittel, 1905, ix., 598, 599.)-Ten C.C. of the melted fat are placed in a separating funnel and dissolved in 10 C.C. of light petroleum. The presence of an artificial colour is shown by the lower-acetic acid-layer becoming yellow or red in colour, the latter coloration being given by alizarine and aniline yellows and by tropeolin. The ordinary colours employed for dyeing butter and margarine, such as Orleans, saffron, aniline, and alizarine yellows, turmeric, tropeolin, picric acid, and commercial butter colours, can all be detected by this method.w. P. s. The solution is then shaken with 15 C.C. of glacial acetic acid. Examination of Lard from Pigs fed on Cottonseed Meal. L. M. Tolman. (Journ. Amer. Chem. SOC., 1905, xxvii., 589-596.)-Working on a, large number of samples of lard obtained from all parts of the bodies of pigs fed on cottonseed meal, the author was unable to detect the presence of phytosterol, even in those samples which gave a coloration with Halphen's test corresponding with an admixture of 15 per cent.of cotton oil (see ANALYST, 1901, xxvi., 292). I t was found that the addition of 2 to 3 per cent. of cottonseed oil to lard could be detected by means of the phytosterol acetate method, and that heated cottonseed oil which does not give Ralphen's reaction can be detected by this method. A. D. Emmett and H. S. Grindley, who also investigated this subject (Jouurrt. Amer. Chem. SOC., 1905, xxvii., 263-270), found that these lards, besides giving coloration withTHE ANALYST. 245 Halphen's test, yielded cryatals resembling phytosterol, showing that at least i~ part of the oil existing in cottonseed meal is absorbed by the animal body and trans- mitted in an unaltered condition to the fat cells.Tolman considers that a very large percentage of phytoeterol must have been present in the residues obtained by Emmett and Grindley, as a mixture of 25 parts of phytosterol with 75 parts of cholesterol still gives the peculiar telescopic form of crystals mainly due to cholesterol. Before crystals are obtained which cannot be distinguished from phytosterol, at least 50 per cent. of the latter must be present. With such an amount present the acetate must have had a melting-point of 123" to 125" C., but w. P. s. this melting-point Emmett and Grindley were unable to obtain. The Composition of American '' Noodles " (Nudeln) and Methods for the Analysis of the Same. A. L. Winton and E. M. Bailey. (Joz~nz. Anzer. Chem. Soc., 1905, xxvii., 137-142).-Egg-noodles are prepared from flour, with the addition of eggs or egg-yolks and salt.As manufacturers have recently placed on the market products dyed a golden yellow colour and devoid of eggs, the authors have examined a number of samples by Juckenack's method, with the object of ascertaining whether the presence or absence of eggs could be determined analytically. Juckenack's method (Zeit. Unterszuh. Nahr. Genussrn., 1900, iii., 13) is based on the determination of the lecithin-phosphoric acid in the sample, the details being as follows : Thirty grams of the finely-ground material are extracted for ten hours with absolute alcohol in a Soxhlet apparatus at a temperature, inside the extractor, of not less than 55" to 60°C. When the extraction is completed, 5 C.C. of 4 per cent. alcoholic potash are added to the contents of the extraction flask, and the alcohol is distilled off.The residue is transferred to a platinum basin by means of hot water, evaporated, dried, and charred. The charred mass is extracted with nitric acid, washed, burnt to a white ash, again extracted with nitric acid, and the phosphoric acid determined in the united solutions as usual. A determination of the amount of fat in the original sample is also of importance, as will be seen from the following table compiled from figures given by Juckenack and Pasternack (Zeit. U?zterszuh. Nahr. Genussna., 1904, viii., 94) : Number of Eggs per Pound of Flour. * 0 1 2 3 12 Ash. Per Cent. 0.460 0-565 0.664 0.758 1.426 Total Phos- phoric Acid. Per Cent. 0.2300 0.2716 0.3110 0.3482 0.6123 Lecithin-phos- phoric Acid.Per Cent. 0.0225 0.0513 0.0786 0.1044 0.2875 Ether Extract. Per Cent. 0-66 1.56 2.42 3.24 7.94 Nitrogen x 6-25, Per Cent. 12.00 12-99 13.92 14.81 21-09 Only five samples contained appreciable amounts of eggs ; and but two contained All the results are calculated on more than one egg, or the yolk of one egg, per pound of flour.:" w. P. s. * The German pound here referred to is approximately 468 grams. the dry material.246 THE ANALYST. Sophistication of Pimento. P. Suss. (Pharnz. Centr., xlvi., 159 ; through Pharm. Journ., 1905, lxxiv., 589.)-Pimento berries which have been collected when nearly ripe become almost black in colour on drying, are deficient in aroma, and are covered with a saccharine exudation. Such unsightly berries are oft en artificially coated by means of a ferruginous substance, probably a clay or brown ochre.This may be detected by boiling the fruits with hydrochloric acid for half a minute, filtering, and testing the filtratR with potassium ferrocyanide solution. If the fruits be genuine only a, bluish-green coloration is obtained, whilst a precipitate of Prussian w. P. s. blue is given in the case of artificially-coated berries. The Composition of the Tubers of Edible Labia-. Balland. (Journ- Pharm. Chim., 1905, xxi., 491-496.)-Numerous species of Coleus are cultivated in different parts of tropical Africa, and the tubers of two species of Plectranthus are also used for food. The specimens of tubers examined by the author resembled the potato in composition. They contained no sugar, and the ash was free from manganese.Tubers of- Coleus Dazo ... C. l n n g o u a s - sieizsis ... c. ?-otuizdifolius (var. cdba) ... C. rotundifolius (var. nigra) . . . C. rotundifolius (var. rubra) , . . Plec tiant hus ter- natus . . . .. P. tuberosus ... Water. Per Cent. 77.30 87.10 7640 72.90 78.26 71-30 77.00 Nitro- gen ou s stances. Per Cent. Sub- 1.72 1.59 2.08 1-46 1.31 2-74 1.52 Fat. Per Cent. 0-54 0.09 0.33 0.30 0.20 0-39 0.26 Starch. Per Cent. 18-29 10.0'7 19.45 23.40 18.57 22 *69 18-90 Cellu- lose. Per Cent. 1.34 0.52 0-83 0.87 0.85 1.26 1 -05 Ash. Per Cent. 0.81 0.63 0.91 1-07 0.89 1.62 1-27 Phosphoric Acid. Per Cent. 0.134 0.134 0.165 0.133 0-157 - - C. A. M. Characteristics of Apricot-tree Gum. P. Lemeland. (Journ. .Pharm. Chim., 1905, xxi., 443 - 448.) - Two samples examined by the author contained (1) 16.15 per cent, and (2) 16.50 per cent.of moisture, and yielded on ignition (1) 2.85 per cent. and (2) 3.4 per cent, of ash. One of the samples dissolved in water to the extent of 76-61 per cent., and the soluble portion had a specific rotation of [uID = - 1.93". Oxydases were identified in both. The organic matter consisted chiefly of gelactans and arabans, one sample yielding 19.8 per cent. of galactose and 38.87 per cent. of arabinose calculated on the original substance. C. A. M.THE ANALYST. 217 Colour Reactions produced by Vanillin with Hydrochloric Acid. L. Rosenthaler. (Zeit. anal. Chenz., 1905, xliv., 292-301.)-A tabular summary is given of different colorations recorded by previous observers as produced by this reagent.The reaction with acetone is particularly sensitive, and will detect OpO1 per cent. of that substance in the distillate from urine. The test is applied by mixing the 1 per cent. solution of vanillin in hydrochloric acid with an equalvolume of sulphuric acid, adding one or two drops of the acetone solution, and heating the whole for fifteen minutes on the water-bath. The violet coloration can be easily dis- tinguished from the red colour produced by the reagent alone under the same con- ditions. Laevulose and cane-sugar give a red coloration, whilst dextrose and lactose produce only a brown coloration. Most of the aromatic ketones tried gave negative results, and no coloration was obtained in cases in which a C6H5 group was inimme- diate combination with a GO group.Thus benzal acetone (C,R,.CH = CH.CO.CH,) gave a coloration, but not benzalacetophenone (C,H,.CH = CH.CO.C,H,). Some of the most frequently adulterated essential oils, such as clove oil, cinnamon oil, and aniseed oil, give little or no coloration with a 1 per cent. solution of vanillin in hydrochloric acid, whereas marked reactions are given by some of their adulterants, such as turpentine oil, and copaiba and gurjun balsams, and the addition of 10 per cent., and sometimes less, of these substances can be detected in this way. The reaction can also be used in testing Peru balsam, which gives no reaction either in the cold or on heating, whilst copaiba balsam gives a faint violet colour which gradually fades, and gurjun balsam gives an immediate purple-red colour, changing to violet.found possible to detect 5 per cent. of gurjun balsam and 10 per cent of copaiba balsam in Peru balsam, and also to detect the presence of 10 per cent. of gurjun balsam in copaiba balsam, the faint coloration produced by the latter in the cold disappearing within fifteen minutes. A comparison of the results led to the following conclusions : (1) Essential oils containing pinene or limonene (CioHls) give a green coloration with the reagent on heating, as is also the case with turpentine oil and other conifer oils. Sesquiterpenes such as cedrene, caryophyllene, and cadinene do not give the green coloration. (2) Oils containing linalool or geraniol or their esters give a violet coloration on heating.(3) The rose and violet colorations given by certain oils in the cold must be attributed to the presence of ketones or phenols. (4) Eucalypt01 or Cineol give a blue coloration on heating, especially after the cooled liquid has been extracted with ether. I n the case of oils containing eucalyptol the colour may be masked by the green colour due to terpenes. Morphine and codeine give a reddish-violet colour on heating, whilst if antipyrin be present the liquid becomes orange, and green in the presence of picrotoxin, but in the latter case only after boiling for three or four minutes. In each case the colour after heating is violet. I t C. A. M. The Assay of Opium and its Preparations. W. H. Lenton. (Pharm. Jourrt., 1905, Ixxiv., 652, 653.)-In the process described the precipitated.morphine is filtered of€ and titrated without previous weighing. For the assay of tincture of opium the details are as follows : The British Pharmacopceia process is followed up248 THE ANALYST. to the filtration of the alkaline liquid. Fifty C.C. are transferred by means of a pipette to a separating funnel, having a small plug of cotton-wool placed in the constriction above the tap. It is also well to deposit a thin layer of disintegrated filter-paper or asbestos fibre on the cotton-wool, and moisten the whole plug with morphinated water immediately before use. The usual quantities of ether, alcohol, and ammonium chloride are now added and the whole well shaken, care being taken not to dislodge the plug. After standing for the specified time the liquid is forced through the plug, air-pressure from an ordinary bellows being applied for this purpose.Ten C.C. each of ether and morphinated water are then added, shaken gently, and forced through the plug. This is repeated once again, after which the contents of the separator are washed with morphinated water until free from chloride, and finally with 2 C.C. of water. Twenty-five C.C. (or any convenient known quantity) of TG acid are now introduced into the separator, which is shaken thoroughly to dislodge the plug. The excess of acid is titrated back with & sodium hydroxide solution, using methyl orange as indicator, 8s suggested by Dowzard (see ANALYST, 1904, xxix., 91). w. P. s. A New Reaction for Aconitine. E. P. Alvarez. (Chenz. News, 1905, xci., 179-181.)-The following colour reaction, formed by the successive action of bromine, nitric acid, and potassium hydroxide on the alkaloid, is described.From 0-2 to 0.5 mgm. of the aconitine is treated in a small porcelain crucible with 5 drops of pure bromine, slightly heating the mixture on a brine-bath. One C.C. of fuming nitric acid is then added and the whole evaporated to dryness, a little more bromine being added when the acid loses its colour. About 0-5 C.C. of a saturated alcoholic solution of potassium hydroxide is now added to the yellow residue and again evaporated to dryness, when a red- or brown-coloured mass is obtained, according to the quantity of aconitine present, After allowing the crucible and its contents to cool, 5 or 6 drops of a 10 per cent.solution of copper sulphate may be poured over the residue. On moving the copper solution about it assumes a deep-green colour. w. P. s. The Detection of Antipyrin in Dimethylamido-antipyrin. P. Bowrcet. (Bull. SOC. Chim., 1905, xxxiii., 572-573.)-Dimethylamido-antipyrin, known com- mercially as pyramidon or amidopyrin, has ten times the selling value of antipyrin, and is, therefore, frequently adulterated with the latter, sometimes to the extent of a third of its weight. To detect this from 0.01 to 0.02 gram of the sample is dissolved in 4 to 5 C.C. of cold water, and the solution treated with 2 drops of sulphuric acid (66" BB., S.G. 1.84) and 2 drops of a saturated solution of sodium nitrate or a pinch of the crystalline salt. On shaking the tube an immediate violet-blue coloration is produced, which rapidly disappears, leaving a colourless liquid when the dimethyl- amido-antipyrin is pure, If, on the other hand, antipyrin is present, the violet coloration gradually disappears, especially after the addition of more sodium nitrate, and is replaced by a pronounced bluish-green stable coloration, the intensity of which increases with the proportion of antipyrin in the mixture.In this way it is easy to detect 2 per cent, of antipyrin, and the nitroso compound formed is so slightly soluble in water that an approximate gravimetric determination can be made by modifying the conditions of the test. C. A. M.THE ANALYST. 249 Valuation of Aloes. A. Tschirch and R. Hoffbauer. (Schweiz. Wochen- schr. Chem. Pharnz., 1905, xliii,, 153; through Chem. Zeit. Rep., 1905, xxix., 106,)- Several colour tests, both positive and negative, by which a good Cape aloe can be recognised are described, The active constituent may be quantitatively estimated in the following way : Five grams of the sample are digested in a, small flask with 5 C.C. of methyl alcohol for two hours. The liquid is then warmed to 50" or 60" C., 30 C.C. of chloroform cautiously added with continued shaking, and the mixture allowed to stand for half an hour. The solution is filtered off into a n Erlenmeyer flask, and the chloroform distilled. The total yellow residue finally obtained is dried at 100" C., and should not weigh less than 4 grams. Detection of Potassium Hydrogen Tartrate in Potassium Antimony Tartrate. Schwartz. (Repertoire, 1905, xvii., 127 ; through Pharm. Jozirn., 1905, lxxiv., 721.)-The presence of uncombined potassium hydrogen tartrate in tartar emetic may be detected by its action on sodium thiosulphate. A solution of the latter is decomposed by potassium hydrogen tartrate, sulphur being set free, whilst tartar emetic has no action. No turbidity should, therefore, be formed when a saturated solution of tartar emetic is mixed with an equal volume of thiosulphate solution and the mixture allowed to stand for five minutes. w. P. s. This extraction with chloroform is repeated four times. A. N. C.

ISSN:0003-2654    

DOI:10.1039/AN9053000242

出版商:RSC    

年代:1905

数据来源: RSC

摘要:

THE ANALYST. 249 ORGANIC ANALYSIS. A Study of Methods for the Determination of Formaldehyde. R. H. Williams. (Journ. Amer. Chenz. Soc., 1905, xxvii., 596-601.)-The following con- clusions, resulting from a critical examination of certain methods, practically confirm those previously arrived at by B. H. Smith (ANALYST, 1904, xxix., 5). The iodometric method is rapid and accurate, and is to be preferred for pure dilute solutions of formaldehyde, For concentrated impure solutions the peroxide method is the most satisfactory, but the time necessary for complete oxidation varies considerably, depending upon the concentration and temperature. The potassium cyanide method, which yields lower results than the oxidation methods, is to be recommended for dilute impure solutions. The end point of Legler’s method is not satisfactory, which fact, as well as the low results, must be attributed to causes other than the influence of strong acids on hexamethylenetetramine.The discrepancy in the results yielded by the two different types of methods is due to conditions inherent in the methods themselves, and not to the presence of impurities. Apparently the con- densation reactions are not complete, or a small part of the formic acid produced in the oxidation methods is further oxidized, giving high results. Paraforinaldehyde when present behaves as formaldehyde. w. P. s. A Colorimetric Method for the Deteetion and Determination of For- maldehyde. F. Bonnet, Junr. (Jozmz. Anzer. Chenz. Xoc., 1905, xxvii., 601-605.) -The method is based on the coloration obtainsd by the action of formaldehyde on morphine dissolved in sulphuric acid, first pointed out by Griinhut (Zeit.anal. Chem., 1900, 329). To detect the presence of formaldehyde in milk, 60 C.C. of the latter are250 THE ANALYST. placed in a porcelain basin, and 1 C.C. of a solution containing 0.35 gram of morphine sulphate in 100 C.C. of concentrated sulphuric acid (specific gravity 1.84) is floated on the surface. In a few minutes a coloration, varying from pink to blue, appears in the morphine solution should formaldehyde be present. In the case of minute traces (1 : 4,000,000) from one to two hours may elapse before the solution becomes slightly coloured. The test is applicable to curdled or sour milks, and also to butter, the latter being flattened over the bottom of the dish before adding the reagent.By comparing the coloration obtained with those yielded by the same volume of milks containing known quantities of formaldehyde the test may be made quantitative. After ascertaining the approximate amount of formaldehyde present in the milk under examination, the determination is repeated, using as nearly as possible the correct quantity of formaldehyde in the control as is considered to be present in the sample, care being taken that the tests are started at the same time and made under like conditions. Over a range extending from 8 : 1,000 to 8 : 1,000,000, solutions con- taining the same amounts of formaldehyde give the same characteristic coloration in the same period of time. The other ordinary milk preservatives give no coloration The dish is then immediately covered with a glass plate.with this reagent. w. P. s. On the Volatility of Lactic Acid in Steam. F. Utz. (Chew. Zeit., 1905, xxvii., 363.)-Many authorities hold that there is no appreciable loss of lactic acid on boiling an aqueous solution, but in some experiments on milk the author has been led to disagree with this opinion. The results of his experiments show that lactic acid is volatile in steam, and to a, greater degree in concentrated than in dilute solutions, but that it is not sovolatile that the lactic acid can be completely distilled from an aqueous solution, though, according to A. Partheil and others, this is possible in a concentrated solution by the use of superheated steam (cf. Bevan, ANALYST, 1-01.xix., 242). A. N. C. Detection of Palm Oil when used as a Colouring Material in Fats and Oils. C. A. Crampton and F. D. Simons. (JOZLYIZ. Aines.. Che17~. Xoc., 1905, xxvii., 270-274.)-The high natural colour of palm oil has led to its use as a means of imparting a butter-like colour to oleomargarine, Small quantities of this oil may be detected in margarine and butter by the following tests: 100 C.C. of the fat are dissolved in 300 C.C. of light petroleum, and shaken out with 50 C.C. of 0.5 per cent. potassium hydroxide solution. The aqueous layer is separated, acidified, and shaken out with 10 C.C. of carbon tetrachloride. The carbon tetrachloride solution is treated with 2 C.C. of a solution of 1 part of cryetallized phenol in 2 parts of carbon disulphide, and 5 drops of hydrobromic acid (specific gravity 1-19).The almost immediate development of a bluish-green colour is indicative of palm oil. Ten C.C. of the melted and filtered fat when shaken with 10 C.C. of acetic anhydride and 1 drop of sulphuric acid (specific gravity 1.53) impart a blue coloration to the lower layers of the mixture on settling if the fat coiltltin palm oil. Sesame oil gives a similar coloration, but not after previous extraction with alcohol. Extraction of palm oil with alcohol has no effect on the formation of the colour.THE ANALYST. 251 I n each test the coloration is transient; changes which occur after the lapse of a The above reactions will detect the few minutes are therefore to be disregarded. presence of about 0.5 per cent.of palm oil in other oils or fats. w. P. s. The Characteristics of Stearin Pitches. E. Donatti. (Chew. Rev. Fett-zc. Ham-Ind., 1905, xii., 42-44, 73-75.)-The commercial products sold as ‘( stearin pitch” are of various origin. Those obtained from the distillation of crude fatty acids in the candle industry are viscid black pitches, which when distilled yield fluorescent inflammable oils. Other “ stearin pitches,” prepared from the fatty acids separated from wool-washings are brownish or black residues, differing considerably in characteristics from true stearin pitch and from wool pitch, and the author there- fore proposes to term them (‘ stearin wool pitch.” Wool pitch and stearin wool pitch are characterised by containing dark neutral compounds resembling asphaltum.These are partially soluble in ether, and dark cohesive precipitates are formed on adding alcohol to the solution. Stearin wool pitch yields much more ash than fat pitch, and the presence of calcium sulphate in large proportion indicates its prepara- tion from wool washings. The undecomposed fatty acids, etc., can be extracted in a, Soxhlet’s apparatus by means of ether or petroleum spirit, and the solution on evaporation leaves a brownish fat-like residue. Further extraction with benzene yields dark solutions, from which on evaporation brilliant black solid or viscid residues are obtained, very similar to those yielded by asphaltum. Only a small residue is left after the two extractions, and this dissolves in carbon bisulphide, with the exception of the mineral matter, in which respect it resembles that from some natural asphaltums and other pitches.As regards the valuation of these pitches, the melting-point and proportion of neutral fats and fatty acids are important criteria, whilst the proportion of con- stituents of an asphaltum nature is of primary importance when the pitch is required for the manufacture of varnishes for cables and insulating materials. Ash or decom- position products insoluble in ether or benzene can never be regarded as an advantage. Two samples of stearin wool pitch examined by the author had acid values of 14-07 and 16.07, and saponification values of 46.56 and 70.32 respectively. On treatment with ether they yielded 52.84 and 77.56 per cent. of insoluble matter respectively, whereas true stearin pitches yielded from 0.97 to 8.67 per cent.C. A. M. Determination of the Melting-points of Pitch, Asphalt, etc. M. Wendriner. (Zeds. f. angew. Chenz., 1905, xviii., 622.)-The method is a modification of one due to Kraenier and Sarnow, in which the melting-point of the pitch wa8 taken to be the temperature at which mercury broke through a stopper of the pitch in a tube heated in a water-bath. The author found, however, that the process gave varying results for the same sample of pitch ; he therefore suggests the modification that, in order to avoid differences in the rate of heating, the pitch should be put into a bath heated a few degrees above the melting-point, and not into a bath which is gradually heated up after the tube of pitch has been placed i n .The use of an air- jacket is then desirable.252 THE ANALYST. The author gives figures to show that the melting-point of a sample of pitch can be consistently and accurately determined by his process, and he also gives some figures showing the effect of varying the size of the pitch-stopper, the quantity of mercury, and the temperature of the water-bat h. The author describes the apparatus he uses, which can be obtained from C. Gerhardt, Bonn. For details the reader is referred to the original paper. E. K. H. The Detection of Sodium Silicate in Soap. Ahmed Hussein. (J0w-n. Pharnz. Chim., 1905, xxi., 496, 497.)-In the ordinary method the soap is dissolved in hot alcohol, the insoluble residue dried in the oven, dissolved in water, and the solution treated with hydrochloric acid, which should give a precipitate of silica. The author, however, finds that sodium silicate after being separated by means of alcohol and dried becomes almost insoluble in water, and so may escape detection. He therefore recommends heating the residue with a little water to which has been added a trace of sodium hydroxide, to bring the silicate into solution. The liquid is then filtered, and hydrochloric acid added until the reaction is acid, the silica still remaining in solution. Ammonia is now added until the liquid is alkaline and the silica is precipitated. C. A. M.

ISSN:0003-2654    

DOI:10.1039/AN9053000249

出版商:RSC    

年代:1905

数据来源: RSC

摘要:

252 THE ANALYST. INORGANIC ANALYSIS. Estimation of the Halogen in Halogen Salts of Mercury. T. Fiseher. (Chenz. Zeit. 1905 xxvii. 361.)-The ordinary method of estimating the bromine by precipitation with silver nitrate after the removal of the mercury with sulphuretted hydrogen was found in some analyses of double salts to be both troublesome and inaccurate. After some unsuccessful experiments a distillation method was worked out which gives good results for either bromine or chlorine. To the mercury salt in a distilling flask 5 C.C. of 10 per cent. soda are added, and warmed for twenty minutes on the water-bath. After cooling 3 C.C. of (1 1) sulphuric acid are added and after another cooling 0.4 gram of potassium per-manganate suspended in 10 C.C. (1 1) sulphuric acid.The flask is quickly connected to a receiver containing a solution of potassium iodide and boiled for five minutes, when all the bromine will have been distilled over. The solution in the receiver is acidified with two drops of sulphuric acid and the freed iodine titrated in the usual way with sodium thiosulphate. For iodine another method must of course be used the separation of the mercury being accomplished by the use of magnesium powder. The iodine salt is mixed in a small flask with magnesium powder and 10 to 20 C.C. of water On shaking this mixture magnesium iodide and magnesium amalgam are formed the latter being quickly converted into magnesium hydroxide and mercury with evolution of hydrogen by the water present. The reaction lasts for one or two minutes and the solution is then filtered acidified with chlorine free nitric acid and the iodine precipitated as silver iodide.This is reduced to silver in a current of gas with the usual precautions and weighed. A. N. C THE ANALYST. 253 The Determination of Acetic Acid in White Lead. G. W. Thompson. (Jounz. Chem. Ind. 1905 p. 487.)-Having found the published methods for this estimation unsatisfactory the author has worked out a direct distillation method, which givea him good results. Eighteen grams of the dry white lead are mixed with 40 C.C. of syrupy phos-phoric acid 18 grams of zinc-dust and 50 C.C. of water in a 500 C.C. distilling-flask connected with a steam supply. The liquid is alternately distilled down to a small bulk and made up again to about 200 C.C.by passing in steam till 10 C.C. of the distillate require but 1 drop of & alkali to affect phenolphthalein. The total distillate is then titrated against & alkali. Vallety. ( A m . de Chim. afzal. 1905 x. 193.)-The method is based upon the behaviour of silver nitrate towards the different products (1) Copper displaces silver com-pletely from its solutious. (2) Oxide of copper on treatment with silver nitrate gives a precipitate of basic nitrate which dissolves in 5 per cent. nitric acid. (3) Silver nitrate reacts with the matte (ignited Cu,S) yielding silver silver sulphide, and copper nitrate. (4) Silver nitrate does not react to any appreciable extent with copper silicate. I n the analysis 5 grams of t,he scork (enclosing 0.75 per cent. of copper) are treated with 100 C.C.of a solution of silver nitrate (50 grams per litre), and 5 C.C. of nitric acid subsequently added. After twelve hours’ digestion the liquid is filtered the residue washed with water containing 3 per cent. of nitric acid and the copper in the filtrate and washings determined by the usual methods. A. N. C. The Determination of Copper and Free Matte in Scorize. C. A. M. Estimation of Arsenie in Fuels. G. McGowan and R. B. Floris. (Journ. Chem. Ind. 1905 xxiv. 265.)-When a precipitate of arsenious sulphide is boiled with water the sulphide is converted quantitatively into oxide which remains in solution. This process is used to shorten the authors’ method for the estimation of arsenic described in the final Report of the Royal Commission.In the method as modified the powdered fuel is ignited with its own weight of lime for three or four hours. The residue is digested with a little nitric acid dried and again ignited. This residue is dissolved in dilute hydrochloric acid the solution filtered the filtrate reduced with aqueous sulphurous acid boiled saturated with sulphuretted hydrogen, and allowed to stand for several days. The precipitate is then filtered in a Gooch crucible and boiled with its asbestos for three hours in 200 C.C. of water. A fourth part of this extract is (( marshed,” and the mirror obtained compared with a standard. The results compare well with those obtained by the original method. A. N. C. On the Estimation of Sulphur in Burnt Pyrites. W. Jene. (Cheijz. Zeit., 1905 xxvii.362.)-In many works-laboratories the sulphur is still estimated by digesting the sample with nitric acid the excess of nitric acid removed by repeated evaporation with hydrochloric acid and the sulphuric acid in the filtered solution directly precipitated with barium chloride. Analyses made by the author show that the results by this method are lower than those obtained by the fusion method of Fresenius sometimes by as much as 40 per cent. and that this difference is due not to a loss of sulphur as sulphurette 254 THE ANALYST. hydrogen but to the incomplete nature of the acid extraction. The author recom-mends fusion of the pyrites with sodium peroxide a method giving excellent results. A. N. C. On the Determination of Zirconium in Presence of Titanium especially in Rocks.M. Dittrich and R. Pohl. (Zeds. anorg. Chem. 1905 xliii. 236.)-The authors’ method is based on a direct weighing of the mixed oxides of zirconium and titanium followed by a colorimetric determination of titanium zirconium being then found by difference. I n the analysis of rocks the ignited ammonia precipitate obtained as usual is fused with caustic soda to remove alumina and phosphoric acid. The residue left on dissolving the melt in water is fused with potassium bisulphate and dissolved in cold water iron and manganese being then precipitated as sulphides by ammonium sulphide in the presence of tartaric acid. The filtrate from these sulphides, which contains only titanium and zirconium is evaporated to dryness the residue taken up in dilute sulphuric acid and the tartaric acid destroyed by the cautious addition of potassium persulphate followed by heating.The liquid is then evaporated until nearly all the sulphuric acid has been volatilized the residue is dissolved in hydrochloric acid and water and titanium and zirconium hydroxides precipitated by ammonia. The precipitate is washed re-dissolved re-precipitated ignited and weighed as TiO + ZrO,. The oxides are then fused with potassium bisulphate and titanium dioxide determined colorimetrically according to Weller (Ber. 1882 xv. 2592). Results obtained on pure solutions containing iron titanium and zirconium show a maximum error of about 3 mgms. on the oxides the quantities used being about 0.1 gram of Fe304 and 0.01 gram of each of the other two oxides.Better results are obtained if both the iron and the mixed titanium and zirconium oxides are finally purified from silica by a potassium bisulphate fusion. I n this case the error on each oxide only amomts to afew tenths of a mgm. A Reaction of the Compounds of Rhodium of Use in Chemical Analysis. E. Alvarez. (Chem. News 1905 216.)-The author describes a reaction by which the formation of the blue sodium per-rhodate (RhO,Na,) serves easily to distinguish this metal from the others of the same group. To an aqueous solution of any soluble rhodate an excess of soda is added and the liquid exposed to the gaseous mixture produced by the action of cold concen-trated hydrochloric acid on potassium chlorate. The dilute and almost colourless solution turns through yellow to red and as the current of gas continues a faint green precipitate appears dissolving again to a bright blue solution.Peroxide and persulphate of sodium decolorize the liquid with evolution of oxygen. A. N. C. No test analyses of rocks are given. A. G. L. This is the per-rhodate of sodium of Claus. A New Reagent for Potassium. Eugenio Piner~a Alvarez. (Chenz. News, 1905 xci. 146.)-The reagent described consists of a 5 per cent. solution of ‘‘ iconogene,” sodium amidonaphthol sulphonate - -NH (1) ‘S0,Na (6). C,,,H/OH (2 THE ANALYST. 255 The solution should be freshly prepared with cold boiled-out distilled water just before using. With all potassium compounds it gives in neutral solutions a pre-cipitate of the white very brilliant crystalline potassium amidonaphthol sulphonate, which is slightly soluble in water and insoluble in alcohol.The crystals consist of orthorhombic plates. The reaction requires some little time and if less than 2 per cent. of the potassium salt is present the crystals will not form for several hours. The delicacy of the test is stated to be as great as that of the platinum chloride method. An advantage of iconogene is that no precipitate is obtained with ammonium salts ; magnesium also does not interfere if sufficient ammonium chloride is present. Iron and manganese give no precipitate with iconogene ; nickel cobalt, and bismuth give a precipitate insoluble in excess ; copper gives a precipitate soluble in excess of the reagent with a green colour. A. G. L. The Solubility of Barium Sulphate in Chromic Chloride Solution.F. W. Kuster and Georg Dahmer. (Zeits. nnorg. Chem. 1905 xliii. 348.)-The authors found on boiling 1.4198 grams pure barium sulphate with 15 C.C. of a fairly concen-trated solution of chromic chloride that 0*0015 gram passed into solution. On treating 4.4063 grams barium sulphate in the same way 0.0051 gram was dissolved. On boiling 15 C.C. of chromic chloride solution acidulated with hydrochloric acid with 1.0262 grams barium sulphate 0.0176 gram of the latter was dissolved after ten days. A. G. L. The Estimation of Potash in Soils Plants and Fertilizers. F. P. Veitch. (Jozmz. Anzer. Chenz. SOC. xxvii. 56.)-The author has examined Moore’s method (Journ. Amer. Chem. SOC. xx. 342) for determining potash in soils plants and fertilizers.The method consists in dissolving the sample in acid destroying ammonium salts and organic matter by means of aqua regia and evaporating repeatedly with hydrochloric acid. The residue is extracted with water and a few drops of hydrochloric acid a slight excess of platinic chloride is added and the whole evaporated nearly to dryness on a water-bath. The evaporation must not be carried to complete dryness which would render the iron and aluminium chlorides difficultly soluble. The cold mass is treated with 15 C.C. to 25 C.C. of acidified 90 per cent. alcohol stirred and filtered after a few minutes being washed with alcohol and ammonium chloride Etolution. The potassium platinochloride obtained is dried, dissolved in hot water and weighed as usual.The alcohol used is prepared by passing dry hydrogen chloride into 90 per cent. alcohol until 1 C.C. of the alcohol neutralizes about 2.3 C.C. of normal alkali solution. For soils the author found that the results given by the method as described above and by a modification in which ordinary 90 per cent. alcohol was substituted for the acidified alcohol were practically identical with those given by the official method. For fertilizers on the other hand the method always gave somewhat higher results (maximum difference 6.19 per cent. K,O instead of 5~70)~ and it was found that whilst more concordant results could be obtained by destroying the organic matter by ignition in a porcelain dish instead of by aqua regia yet in this case the residue obtained required digestion with hydrochloric acid occasionally for some time in order to extract all the potash.The author recommends the Moor 256 THE ANALYST. method original or modified for soils but sees no advantage over the official method in using it for fertilizers. A. G. L. The Detection of Sodium Salts by Fre'my's Reagent. J. Bougault. (Jozwn. Pharm. Chim. 1905 xxi. 437-442.)-The following modification of Frkmy's reagent is easily prepared and is stated to give constant results A mixture of 10 C.C. of pure potassium hydroxide solution (33.3 per cent.) and of 45 C.C. of hydrogen peroxide (10 vol. solution) is gently heated with 1 gram of antimonous chloride until after five to ten minutes the solution is complete with the exception of a very slight amorphous sediment from which the liquid can be decanted.In using the reagent the following precautions should be observed (1) A very small amount (0.5 c.c.) should be used. (2) The solution of the sodium salt should be as con-centrated as possible and should be neutral or alkaline. (3) Boiling the mixed solutions for a few seconds facilitates the formation of the precipitate. The reagent gives an immediate precipitate with 0.4 mgm. of sodium chloride. Potassium salts interfere with the sensitiveness of the reaction and should be eliminated as far as possible preferably by treatment with tartaric acid. Lithium salts react like sodium salts and 0.05 gram of lithium nitrate in 1 C.C. of water gives an immediate precipitate of lithium pyroantimonate with 1 C.C. of the reagent. This crystallizes in microscopic hexagonal lamellz ; whereas the sodium compound if deposited slowly forms prisms approximating to cubes or octahedra or if deposited rapidly, is in the form of needles united together at the top or of small elongated prisms.C. A. M. A Gravimetric Method of Determining Nitric Acid. M. Busch. (Berichte, 1905 xxxviii. 861-866.)-This is based upon the fact that dipbenyl-endanilo-dihydro-triazole-N=C, . C,H, which is sold by Merck under the name of Yzitron forms an insoluble salt with one molecule of nitric acid a crystalline precipitate being formed on treating 5 to 6 C.C. of the nitrate solution with a few drops of a 10 per cent. solution of the reagent in dilute acetic acid (5 per cent.). Insoluble compounds are also given by hydrobromic hydriodic, nitrous chloric perchloric chromic and thiocyanic acids so that the absence of these must be insured before testing for nitric acid.The reagent is capable of detecting 1 part of nitric acid in 60,000. For a quantitative determination the substance which should contain about 0.01 gram of nitric acid is dissolved in 80 to 100 C.C. of water, and the solution after the addition of a few drops of dilute sulphuric acid is heated to the boiling-point and treated with 10 to 12 C.C. of the acetic acid solution of the reagent. The flask is then kept for one and a half to two hours in ice water after which the contents are filtered with the aid of a pump and the precipitate washed with 10 to 12 C.C. of iced water and dried for forty-five minutes at 110" C. The weight of the precipitate multiplied by 0.168 gives the weight of the HNO,.When nitrous acid is present with nitric acid it may be destroyed by a preliminar THE ANALYST. 257 treatment with hydrazine sulphate but the results for nitric acid subsequently obtained by the above method will then be slightly too high. c. A. ni. The Determination of Nitrogen as Nitrites in Waters. R. S. Weston. (Jounz. Amar. Chem. Soc. 1905 xxvii. 281-287.)-The following modification of the naphthylamine method is recommended The two solutions required are prepared by dissolving 8 grams of sulphanilic acid in 1 litre of dilute acetic acid (specific gravity 1*044) and 8 grams of u-naphthylamine also in 1 litre of this dilute acid. I t may be necessary to warm the latter solution to dissolve the salt and afterwards to filter it.Two C.C. of each of these reagents will be found sufficient for 100 C.C. of the water. The coloration produced is compared with that given by known quantities of standard nitrite solution. Excess of acetic acid does not interfere with the delicacy of the reaction and is to be preferred to hydrochloric acid in making up the reagent. The water and reagents should be stirred after mixing to prevent the dense acid solution falling to the bottom of the test cylinder. w. P. s. Colorimetric Determination of Phosphorus. Thomas E. Hewitt. (Journ. Amer. Chenz. Xoc. 1905 xxvii. l2l.)-The method depends on the colour given with hydrogen sulphide by alkaline solutions of ammonium phosphomolybdate. To carry it out the phosphomolybdate precipitate is obtained as usual from 2 grams of the steel or iron The precipitate is washed with 2 per cent.nitric acid. The filter and precipitate are then placed over a 100 C.C. measuring-flask and the precipitate is treated first with hot water and then with sodium hydroxide solution (4 grams NaOH per litre) until just dissolved. Half a6 much again of the sodium hydroxide solution is then added and the flask is filled to the mark with water. An aliquot part of the solution is placed in a 50 C.C. Nessler tube diluted to 25 c.c. and hydrogen sulphide is passed through the liquid for five minutes. After allowing the tube to stand in boiling water for another five minutes it is filled to the mark with water and its colour compared with a similar tube containing 10 C.C.of standard phosphomolybdate solution similarly treated. The colours should be compared by looking downwards into the tubes light being reflected upwards through them. The treatment with boiling water renders the colours permanent for about two hours. The standard phosphomolybdate solution is made by dissolving 0.2737 gram of dry ammonium phosphomolybdate in a 50 per cent. excess of the sodium hydrate solution and diluting to 500 C.C. with water. Of this solution 10 c.c.=0-000009122 gram phosphorus. Several determinations by this method on each of four samples of iron and steel gave results in most cases within 0.001 per cent. of those obtained by titration with standard sodium hydrate. In the case of a Bessemer pig-iron how-ever containing 0-102 per cent.of phosphorus the results varied between 0,089 and 0.083 per cent. ; and another similar sample containing 0.088 per cent. of phosphorus, showed 0-083 to 0-079 by this method. A. G. L. Methods of Titration of Phosphoric Acid. 0. J. Hlavnicka. (Zeits. f. angew. Chem. 1905 xviii. 655.)-The author claims that Hundeshagen’s method as compared with other volumetric processes for the estimation of phosphoric acid is simple and practical. The method in detail used in the author’s laboratory is as follows The phosphate is precipitated as MgNH,PO by the citrate method an 258 THE ANALYST. the precipitate is collected with the aid of suction on a hardened filter-paper, washed first with 2g per cent. ammonia and then with alcohol till free from ammonia. The total volume of the washings need not be more than about 40 C.C.The pre-cipitate is then washed off with cold water into a porcelain dish (total volume required about 200 c.c.) and after the addition of a few drops of methyl orange solution (strength 0.1 per cent.) decomposed with a slight excess of i%-N HCl and the excess found by titration with ib-N NaOH (1 C.C. -5%-N = 0.01065 gram P,05). For this process the amount of substance taken should be such as to yield about 0.15 - 0-2 gram P,05. The estimation in the case o€ superphosphate takes from two to two and a half hours. This method the author finds gives results agreeing with those obtained by the gravimetric molybdate method but about 0.1 to 0.2 per cent. higher than the gravimetric citrate method figures. This result was also obtained by Hundeshagen (Cham.Zeit. xxv. 446). Although the method due to Raschig (ANALYST this vol. p. 221) where pure water is used for washing the MgNH,PO precipitate is simpler the author con-siders the above method better adapted for technical analyses as a greater volume of washing is permissible and the use of an ordinary hardened filter is also an advantage. E. K. H. The Determination of Free Phosphorie Acid in Superphosphates. (Chem. Zeit. 1905 xxix. 178.)-A process depending on the reaction Gerhardt. of phosphoric acid with calcium carbonate according to the equation-2H3P0 + CaCO = H,O + CaH,(P04) + GO,, and which consists in treating the superphosphate with excess of calcium carbonate, and estimating either the CO evolved or the unused CaCO remaining.The presence of small amounts of iron sulphate in the sample forms a source of error as this substance also reacts with calcium carbonate rendering the result low. This error can be avoided by precipitating the iron with potassium ferrocyanide. The analytical process is as follows 20 grams of the sample are shaken with water in a litre flask for half an hour. One gram of potassium ferrocyanide dissolved in water is added and the solution made up to the litre. One hundred C.C. are filtered, and shaken with an excess of calcium carbonate for one half-hour again filtered, washing the precipitate only slightly and the filtrate evaporated to dryness. The residue consisting of calciuni phosphate is carefully ignited and weighed. Or the process can be carried out in a Scheibler’s or similar apparatus and the resulting CO estimated.The author recommends however the volumetric estimation of the excess of carbonate. For this purpose the solution after the reaction has been completed is treated with 25 C.C. of hydrochloric acid made up to 200 c.c. and 100 C.C. filtered through a dry paper. NaOH using methyl orange as indicator. The colour change is sharp and this method is certainly the most accurate one. This is titrated with H. A. T. Estimation of Combined Sulphurie Acid by the Methods of Lunge and (Report 01 tlze lYiizth Sub- Cominittee of the Iizterizutioiznl Analysis of Silberberger THE ANALYST 259 Commission. Zeits. f. mzgezu. Cltem. 1905 xviii. 449; Berichte 1903 xxxvi. 2755.)-Silberberger criticised Lunge’s method and suggested a new method using strontium chloride in place of the barium salt used by Lunge.This new method has been investigated and found unreliable and therefore distinctly inferior to the original method devised by Lunge. Drawn ‘ ~ p by G. Lunge. E. K. H. The Titration of Sulphurous Acid in Alkaline Solution. 0. Ruff and W. Jeroch. (Berichte 1905 xxxviii. 409-419.)-Direct titration of sulphurous acid by means of iodine solution in the presence of sodium bicarbonate yields incorrect results owing to the oxidation of the sulphite through the agency of the iodine ions which act catalytically. On the other hand the method in which an excess of iodine is added and the excess titrated with thiosulphate solution is also shown to be unreliable and the authors therefore recommend the addition of 10 to 20 per cent.of mannitol to the liquid and carrying out the titration in an atmosphere of carbon dioxide to prevent oxidation. Direct titration is preferable but if iodine be added in excess arsenious acid should be used in place of thiosulphate for the back titration. C. A. M. On the Analysis of Sodium Hydrosulphite. A. Binz and H. Bertram. (Zeits. angew. Chem. 1905 xviii. 168.)-The authors have compared Ekker’s method (Rec. trav. chim. Pays-Bas 1894 xiii. 36) for the determination of hydrosulphite with other methods especially with the indigo method and find it convenient and reliable. I n a solution in which by the indigo method 2.938 grams Na,S,O per litre were found Ekker’s method gave values varying from 2-818 to 2.970.The method is carried out by dissolving the hydrosulphite in air-free water in an atmosphere of carbon dioxide adding a drop of ferrous sulphate solution and running in standard potassium ferricyanide solution until a blue colour appears air being carefully excluded all the time. Na,S,O + 2K,Fe(CN) = 2S0 e 2K,NaFe(CN),. Since Turnbull’s blue is decomposed by alkalies a little acetic acid should be added to the liquid if so much alkali is present in the sample that the sulphur dioxide formed during the reaction is insufficient to neutralize it. The oxidation proceeds according to the equation : A. G. L. A New Method of separating Chlorides Bromides and Iodides in Mixtures. (Zeits. f. angew. Chem. 1905 xviii. 696.)-This method depends upon the fact that when alkali iodides are shaken up with mercurous chloride or bromide double decomposition takes place and mercurous iodide is pre-cipitated whilst an equivalent quantity of chloride or bromide goes into solution ; in the same w~ty alkali bromides when shaken with mercurous chloride yield inercurous bromide and an equivalent of a soluble chloride.These reactions are quantitative. These facts have been applied to a separation of any combination of chlorides, bromides and iodides by the author. Examples of the four possible combinations are given but the method of working and of calculation will be clear if the most complex case only is here described-that namely of a mixture of the three. 0. Wentzki 260 THE ANALYST, I n an actual experiment a solution containing per litre 0.585 gram NaCl, 1.19 gram KBr and 1-66 gram KI was used and the following determinations were made : (1) One hundred C.C.were precipitated with silver nitrate ; the weight AgCl + AgBr + AgI = 0.5640 gram. (2) Three hundred C.C. were shaken in a stoppered flask with mercurous bromide until in a filtered small sample iodine could not be found by nitrous acid and chloroform. One hundred C.C. of the filtrate mere precipitated with silver nitrate ; weight of AgCl+ AgBr = 0.5175 gram. (3) Three hundred C.C. of the original solution were shaken with mercurous chloride until neither iodine nor bromine could be detected in a filtered sample. One hundred C.C. of the filtrate were then precipitated with silver nitrate and the weight AgCl = 0.4294 gram.As the iodide in (1) has been replaced in (2) by its equivalent of bromide the difference between (1) and (2) x the ratio AgI (AgI - AgBr) i.a. 5.007 gives the weight of AgI in the first precipitate = 0.2327. Similarly the difference between (2) and (3) x the ratio AgBr (AgBr - AgCl) i.~. 4.221 gives the weight of AgBr contained in the second precipitate = 0.3718. Then by subtraction of this value from the weight of (2) we get the weight of AgCl contained in (2) =0*1457. This quantity is also contained in the first precipitate, and we have already found the weight of AgI therein hence : AgCl+ AgBr + AgI = 0.5640 AgCl0*1457+AgI 0.2327 = 0.3784 From these figures we get AgRr = 0.1856 Chlorine. Bromine. Iodine. Found . . - . 0.0360 0.0790 0.1257 Calculated . . . . . . 0-0354 0.0799 0.1268 The author found the process hold equally for sodium potassium ammonium, magnesium and alkaline-earth salts. E. K. H. % 4 8 * * REVIEW. THE BREWING INDUSTRY. By JULIAN L. BAKER F.I.C. F.C.S. This little work is published by Methuen and Go. and constitutes one of their well-known series of “Books on Business.” As such it has not been written for the expert but is intended to convey to any member of the lay public who may consult its pages useful information in connection with the materials used in brewing and a clear notion of the more important principles underlying the operations of the brewery. Mr. Baker’s qualifications for the task he has undertaken are well recognised and this little book which is both accurate in matter and clear in arrangement should do much to dissipate the ignorance and misconception which prevail so extensively in connection with the manufacture of our national beverage. The book contains very few errors typographical or otherwise and has a number of well-executed illustrations. A. C. C

ISSN:0003-2654    

DOI:10.1039/AN9053000252

出版商:RSC    

年代:1905

数据来源: RSC

摘要:

260 THE ANALYST, ERRATuM.-Page 184, second line from top, for 50 c.c.” Tend 5 c.c.”

ISSN:0003-2654    

DOI:10.1039/AN9053000260

出版商:RSC    

年代:1905

数据来源: RSC