Chapter II: Part 2

=29. Estimation of Lime.=—One hundred cubic centimeters of the solution (containing one gram of the original substance) are evaporated in a beaker to about fifty cubic centimeters; ten cubic centimeters of dilute sulfuric acid (one to five) are added; and the evaporation is continued on the water-bath until a considerable crop of crystals of gypsum has formed[20]. The solution is then allowed to cool, when it generally becomes pasty, owing to the separation of additional gypsum. When it is cold, 150 cubic centimeters of ninety-five per cent alcohol are slowly added, with continual stirring, and the whole is allowed to stand for three hours, being stirred from time to time. After three hours, it is filtered, with the aid of a filter-pump, into a distillation flask, and the beautifully crystalline precipitate, which does not adhere to the beaker, is washed with ninety-five per cent alcohol. The filter, with the precipitate, is gently removed from the funnel and inverted into a platinum crucible, so that, by squeezing the point of the filter, the precipitate is made to fall into the crucible, and the paper can be pressed down smoothly upon it. On gentle heating of the crucible, the remaining alcohol burns off, and when the paper has been completely destroyed, the heat is raised to the full power of a bunsen for about five minutes. After cooling in a desiccator the crucible containing the calcium sulfate, is weighed. The filtration may also be accomplished on asbestos felt.

=30. The Ammonium Oxalate Method.=—This method has been extensively used in this country in commercial work, and is best carried out as described by Wyatt.[21] The total filtrates from the iron and alumina precipitates, secured as described in paragraph =33=, are well mixed and concentrated to a volume of about 100 cubic centimeters. There are added about twenty cubic centimeters of a saturated solution of ammonium oxalate, and after stirring, the mixture is allowed to cool and remain at rest for six hours. The supernatant liquid is poured through a filter, the residue washed three times by decantation with hot water and brought upon the filter. The beaker and precipitate are washed at least three times. The precipitate is dried and ignited at low redness for ten minutes. The temperature is then raised by a blast and the ignition continued for five minutes longer, or until the lime is obtained as oxid. The precipitate is likely to contain magnesia. The magnesia is estimated in the filtrates from the lime determination by first mixing them and concentrating to 100 cubic centimeters, which, after cooling, are made strongly alkaline with ammonia. After allowing to stand for twelve hours the ammonium magnesium phosphate is collected and reduced to magnesium pyrophosphate by the usual processes. If one gram of the original material has been used the pyrophosphate obtained, multiplied by 0.36, will give the weight of magnesia contained therein.

=31. Lime Method of Immendorff.=—The tedious processes required to determine the lime in the presence of iron, alumina, and large quantities of phosphoric acid are well known to analysts. Immendorff has published a method, accompanied by the necessary experimental data, based on the comparative insolubility of calcium oxalate in very dilute solution of hydrochloric acid. He has shown in the data given that the lime is all precipitated in the conditions named and that the precipitate, when properly prepared, is not contaminated with weighable amounts of the other substances found in the original solution[22]. The ease with which oxalic acid can be determined volumetrically with potassium permanganate solution aids greatly in the time-saving advantages of the process.

In a hydrochloric acid solution of a mineral phosphate an aliquot part of the filtrate representing about 250 milligrams of calcium oxid, usually about twenty-five cubic centimeters, should be taken for the analysis. Ammonia is added in slight excess and then the acid reaction restored with hydrochloric until shown plainly by litmus. The solution is then heated and the lime thrown down by adding a solution of ammonium oxalate in excess. In order to secure a greater dilution of the hydrochloric acid after the precipitation has been made, water should be added until the volume is half a liter. Before filtering, the whole should be cooled to room temperature. The precipitate should be washed first with cold and afterwards with warm water. The well-washed precipitate is dissolved in hot dilute sulfuric acid and the solution, while hot, titrated with a standard solution of potassium permanganate set by a solution of ammonio-ferrous sulfate.

If one cubic centimeter of the permanganate represent 0.005 gram of iron it will correspond almost exactly to 0.0035 gram of calcium oxid.

_Example._—Sample of rather poor mineral phosphate, five
grams in half a liter. Strength of potassium permanganate, one
cubic centimeter equivalent to 0.00697 gram of iron and to
0.003484 gram of calcium oxid.

Twenty-five cubic centimeters of the solution, representing one
quarter of a gram, in which the lime was precipitated as above
described, required 9.6 cubic centimeters of the potassium
permanganate to saturate the oxalic acid. Then

9.6 × 0.003484 = 0.0334464 gram,

or 13.38 per cent of calcium oxid. The method is also applicable
to basic slags.

=32. Estimation of Iron and Alumina in Mineral Phosphates.=—When mineral phosphates are to be used for the manufacture of superphosphates by treatment with sulfuric acid their content of iron and alumina becomes a matter of importance. By reason of the poor drying qualities of the sulfates of these bases their presence in any considerable excess of a few per cent becomes exceedingly objectionable. The accurate estimation of these ingredients is not only then a matter of scientific interest but one of great commercial significance to the manufacturer.

The conventional methods so long in use depending on the precipitation of the iron and alumina as phosphates in the presence of acetic acid have been proved to be somewhat unreliable. Not only does the acetic acid fail to prevent the precipitation of some of the lime, but it also dissolves more or less of the iron and aluminum phosphates. The solution of the precipitate and its reprecipitation by the addition of ammonia, may free the second precipitate from lime, but it increases the error due to the solubility of the aluminum salt. The methods recently introduced for the estimation of iron and alumina in presence of excess of lime and phosphoric acid are not entirely satisfactory, but are the best which can now be offered.

=33. The Acetate Method.=—The principle of this process is based on the fact that in a solution containing iron, alumina, lime, and phosphoric acid the iron and aluminum phosphates can be thrown down in a slightly acid solution by ammonium acetate while the calcium phosphate remains in solution. The acidity in the older methods is due to acetic and can be secured by making the solution slightly alkaline with ammonia and adding acetic to slight acidity. One of the best methods of conducting the operation is that of C. Glaser[23]. Glaser’s modification of the older processes is based on the assumption that at 70° the aluminum phosphate is quantitatively precipitated by ammonium acetate in a dilute hydrochloric acid solution and that the mixed precipitates of iron and aluminum phosphates obtained at this temperature are free of lime. The operation is conducted in the following manner:

The hydrochloric acid solution of the phosphate must contain no free chlorin and is treated with a few drops of a methyl orange solution. Ammonia is added until nearly neutral, but the acid reaction is retained as shown by the indicator. A few cubic centimeters of ammonium acetate are added, which produce a yellow coloration of the liquid and also a complete precipitation of the iron and aluminum phosphates when warmed to 70°. At this temperature the precipitation of any calcium phosphate is avoided. A small quantity of the lime may be carried down mechanically and therefore the precipitate should be dissolved in hydrochloric acid and the precipitation again made as above after the addition of some sodium phosphate. If the original solution contain any free chlorin, as may be the case when aqua regia is employed as solvent, before beginning the separation, ammonia should be added in slight excess and the acidity restored by hydrochloric after adding the indicator. In washing the precipitates, water of not over 70° must be used. As has been shown by Hess in the work cited in the next paragraph, the statement of C. Glaser to the effect that the precipitates obtained as above are free of lime has not been proved to be strictly correct. The process, however, is a distinct improvement over the older methods and forms the basis of the amended process given below, which appears to be sufficiently accurate to entitle the acetate method to favorable consideration.

=34. Method of Hess.=—Hess has lately made a thorough investigation of the standard methods of determining the iron and aluminum oxids in the presence of phosphoric acid and has shown that the assumption that the composition of the precipitate is represented by the formula Al₂(PO₄)₂ + Fe₂(PO₄)₂ is erroneous[24].

In the washing of the precipitated iron and aluminum phosphates there is a progressive decomposition of the compound with the production of a basic salt. The composition of the precipitate at the end is dependent chiefly upon the way in which the washing takes place. It is quite difficult to always secure a washing in exactly the same way and the final composition of the precipitate varies with almost every determination. It is not, therefore, an accurate proceeding to take half the weight of the precipitate as phosphoric acid or as iron oxid and alumina. In every case it is necessary to dissolve the precipitate and determine the phosphoric acid in the regular way. Hess proposes the following method for carrying out the acetate process of separation:

The mineral phosphate should be dissolved in hydrochloric acid and the solution made up to such a volume as shall contain in each fifty cubic centimeters, one gram of the original substance. This quantity of the solution is diluted with two or three times its volume of water to which a drop of methyl orange solution (1-100) is added, and ammonia added with constant stirring until the solution is just colored and still reacts slightly acid. Without taking any account of the precipitate which is produced by this approximate neutralization of the solution, there are added fifty cubic centimeters of acid ammonium acetate which in one liter contains 250 grams of commercial ammonium acetate. The acidity of the solution is due to an excess of acetic in the commercial salt. The temperature is then carried to 70° and the precipitate produced immediately separated by filtration, washed four times with water below 70°, and again dissolved in dilute hydrochloric acid. The dissolved precipitate is treated with ten cubic centimeters of a ten per cent ammonium phosphate solution and again almost neutralized as described above, twenty-five cubic centimeters of the ammonium acetate solution added and warmed to 70°.

The precipitate obtained is once more dissolved and precipitated as above described, and is then collected upon a filter, washed, ignited, and weighed. The residue after ignition is dissolved in the crucible by heating with a little concentrated hydrochloric acid, and washed into a beaker. Any silicic acid present is separated by filtration, ignited, and weighed, and subtracted from the total weight of the precipitate. To the filtrate is added ammonia to diminish the acidity, but not sufficient to produce a precipitate and the clear solution is treated with thirty cubic centimeters of the ordinary ammoniacal citrate solution and fifteen cubic centimeters of magnesium mixture, and the precipitation of the ammonium magnesium phosphate hastened by stirring with a glass rod.

It is advisable to always make the filtrate from the third precipitation slightly ammoniacal and to boil it for a long time. If the operation have been carried on correctly, there occurs only a slight precipitate of Ca₃P₂O₈ amounting only to a few milligrams. In some cases it may be necessary to dissolve the precipitate and reprecipitate the iron and aluminum phosphates a fourth time.

The whole time required for the triple precipitation, according to Hess, if all the operations be properly conducted, is from three to four hours. It is therefore possible by this variation of the acetate method to secure a determination of the iron and alumina as phosphates in the same time which is occupied by the Glaser-Jones method when the separation of lime is taken into account.

If the solution of the mineral phosphate employed contain any notable quantity of organic material, it must be destroyed by boiling with bromin or some other oxidation agent, before the precipitation by the acetate method is commenced.

The presence of silicic acid need not be taken into special consideration since this can be detected and determined in the phosphate precipitates after they have been ignited and weighed. While the determinations of the phosphoric acid in Hess’ method were made by precipitation in the presence of citrate, he found that they agree perfectly with the previous precipitations with molybdic solution.

=35. Method of Glaser.=—The principle on which this method rests depends on the preliminary removal of the lime by conversion into calcium sulfate and its precipitation in the presence of strong alcohol.[25] It is conducted as follows:

Five grams of the phosphate are dissolved in a mixture of twenty-five cubic centimeters of nitric acid of 1.2 specific gravity and about 12.5 cubic centimeters of hydrochloric acid of 1.12 specific gravity, and made up to a volume of half a liter, and filtered. One hundred cubic centimeters of the filtrate, equivalent to one gram of the substance, are placed in a quarter liter flask and twenty-five cubic centimeters of sulfuric acid of 1.84 specific gravity added. The flask is allowed to stand for about five minutes and meanwhile shaken a few times. About 100 cubic centimeters of alcohol of ninety-five per cent are then added and the flask filled with alcohol to the mark and well shaken. A certain degree of concentration takes place and this is compensated for by lifting the stopper and filling again with alcohol to the mark and shaking a second time. After allowing to stand for half an hour the contents of the flask are filtered, 100 cubic centimeters of the filtrate being equal to four-tenths gram of the substance. This volume, filtered, is evaporated in a platinum dish until the alcohol is driven off. The alcohol-free residue is heated to boiling in a beaker with about fifty cubic centimeters of water. Ammonia is added to alkaline reaction, but in order to avoid strong effervescence it is not added during the boiling. The excess of ammonia is evaporated, the flask allowed to cool, the contents filtered, precipitate and filter washed with warm water, ignited, and the phosphates of iron and alumina weighed. Half of the weight of the precipitate represents the weight of Fe₂O₃ + Al₂O₃. The estimation, as before indicated, should be carried on without delay, the whole time required not exceeding from one and a half to two hours.

=36. Jones’ Variation.=—The method of Glaser described above, as practiced by the German chemists, has been found by Jones to be inaccurate on account of the alcohol not being added in sufficient quantity in the precipitation of calcium sulfate and for the additional reason that the amount of sulfuric acid added is more than is actually necessary[26]. Jones modifies the method as follows: Ten grams of the material are dissolved in nitro-hydrochloric acid and the solution made up to 500 cubic centimeters and filtered. Fifty cubic centimeters of this solution, representing one gram, are evaporated to twenty-five cubic centimeters and, while still hot, ten cubic centimeters of dilute sulfuric acid (one to five) added. The mixture is then well stirred and cooled. One hundred and fifty cubic centimeters of ninety-five per cent alcohol are next added and after stirring, the solution is allowed to stand three hours. The calcium sulfate is collected on a filter, washed with alcohol, and the filtrate and washings collected in an erlenmeyer. The washing is completed when the last ten drops, after dilution with an equal volume of water, are not colored with a drop of methyl orange.

The moist calcium sulfate is transferred to a platinum crucible, the filter placed on it, the alcohol burned off, the filter incinerated, and the calcium sulfate ignited and weighed. The contents of the flask are heated to expel the alcohol, the residue washed into a beaker, made slightly alkaline with ammonia, and again heated till all the ammonia is driven off. This treatment is necessary to prevent the precipitate from being contaminated with magnesia. The precipitate is collected on a filter, washed four times with hot water, or water containing ammonium nitrate, dried, ignited, and weighed. One-half of the weight of the precipitate represents the weight of the ferric and aluminic oxids.

=37. Estimation of Iron and Alumina in Phosphates by Crispo’s Method.=—The phosphate of ferric iron is subject to a slight decomposition in presence of both hot and cold water with a tendency to the production of basic compounds. It is soluble to a slight extent in hot and cold acetic acid, almost insoluble in ammonium acetate, and quite insoluble in ammonium chlorid and nitrate. Aluminum phosphate is likewise soluble, to a slight degree, in acetic acid and ammonium acetate, and insoluble in ammonium chlorid and nitrate. The method of Crispo for the separation of iron and alumina in phosphates is based on the above properties.[27] Five grams of the mineral phosphate are dissolved in fifty cubic centimeters of aqua regia, composed of forty cubic centimeters of hydrochloric acid of 1.10, and ten of nitric acid of 1.20 specific gravity, and this solution is diluted to half a liter. To fifty cubic centimeters of the filtered solution are added two of ammonia (0.96) and fifty of a half saturated solution of ammonium chlorid, and the whole boiled. The liquid should remain clear, but if it become cloudy add a little dilute nitric acid, drop by drop, until the turbidity is removed, and then ten cubic centimeters of a saturated solution of ammonium acetate, and boil for three minutes, cool, and filter. The precipitate is washed twice with a ten per cent solution of ammonium chlorid and redissolved with two cubic centimeters of nitric acid, and the filter washed with hot water. The phosphoric acid is separated by forty cubic centimeters of molybdate solution, and the precipitate washed three or four times with a one per cent nitric acid solution.

To the filtrate are added fifty cubic centimeters of a one-half saturated ammonium chlorid solution, ammonia is added in slight excess to produce precipitation and the mixture boiled for a few minutes. After filtering, the precipitate is washed with hot water three or four times, dissolved in two cubic centimeters of nitric acid, and the filter washed with hot water. Again, fifty cubic centimeters of half saturated ammonium chlorid are added and the precipitate thrown down once more by ammonia in slight excess. The precipitate is washed with hot water and finally ignited and weighed as iron and aluminum oxids.

According to Crispo, the original Glaser method, with its various modifications, is not to be considered reliable, and the choice lies between the molybdic method as usually practiced, and his own for the accurate estimation of iron and alumina. Manganese disturbs the accuracy of the results unless the directions given are carefully followed. Manganese phosphate is soluble at all temperatures below fifty. If then the mixture of the phosphates be allowed to cool before filtering, the iron and aluminum salts are not contaminated with manganese. This method of Crispo is somewhat tedious, but it is claimed that these variations of the molybdic method render it exact in respect of the determination of iron and alumina.

=38. Method Employed in Geological Survey.=—Chatard gives the following directions for conducting the Glaser-Jones process[28]: The distillation flask containing the alcoholic filtrate is connected with its condenser and heated on a water-bath until no more alcohol comes over. This distillate, if mixed with a little sodium carbonate and redistilled over quicklime, can be used over and over again, so that the expense for alcohol is really very slight, while in the use of the Glaser method, with its large amount of sulfuric acid, all the alcohol is lost.

When the distillation is ended the residue in the flask is washed into a platinum dish and evaporated to a small bulk on the water-bath. The dark brown color produced is due to the presence of organic matter and this must be destroyed, as it prevents the complete precipitation of the phosphate in the subsequent operation.

The organic matter is best destroyed by removing the dish from the bath, adding a small quantity of pure sodium nitrate, and heating very carefully over the naked flame, keeping the dish well covered with a watch-glass to avoid spattering. The mass fuses to a colorless, viscous liquid, becoming glassy when cooled and is readily soluble in a hot very dilute solution of nitric acid. The solution transferred to a beaker is made distinctly alkaline with ammonia and carefully neutralized with acetic acid, diluted with hot water, boiled, and the precipitate allowed to settle, after which it is separated by filtration.

After the precipitate has been completely transferred to the filter, the washing is completed with a dilute solution of ammonium nitrate. The precipitate is dried, ignited, cooled, and weighed.

The determinations should be made in pairs, one portion being used for the estimation of the phosphoric acid by fusing with a little sodium carbonate, and the other, after fusion with sodium carbonate, is dissolved with sulfuric acid and the iron reduced and titrated with potassium permanganate solution. The filtrate from the iron and alumina determination is evaporated to a small bulk, made strongly ammoniacal and allowed to stand for some time when the magnesia present separates as ammonium magnesium phosphate which is determined in the usual way.

If, during the evaporation of the filtrate, any flocculent matter separate, it should be removed by filtration and examined before precipitating the magnesia.

=39. Variation of Marioni and Fasselli.=—Glaser’s method has been shown to be subject to errors by Marioni and Fasselli[29] in the following respects:

1. The precipitation of a small quantity of calcium phosphate with the ferric and aluminum phosphates.

2. The possible precipitation of basic phosphates if all the iron and alumina are not combined with phosphoric acid in the mineral.

3. The partial solubility of ferric and aluminum phosphates in dilute acetic acid.

4. The decomposition of ferric orthophosphate into soluble acid phosphate and insoluble basic salt by boiling.

To avoid these errors the following procedure is proposed: From one to five grams of the phosphate are boiled in a flask for ten minutes with fifteen cubic centimeters of strong hydrochloric acid, and afterwards diluted with a double volume of water. A few crystals of potassium chlorate are added, and several drops of nitric acid, and the liquid boiled to expel chlorin. It is then filtered and washed until the volume of the filtered liquid amounts to 150 cubic centimeters. After cooling, a half gram of ammonium phosphate in solution is added, and two cubic centimeters of glacial acetic acid, followed by dilute ammonia, drop by drop, until a slight precipitate persists on stirring. Again the same quantity of acetic acid is added as above, well shaken, and left for two hours. The precipitate is collected on a filter and washed with a one per cent ammonium phosphate solution. The precipitate is dissolved by a minimum quantity of hydrochloric acid and the solution collected in the same vessel in which the precipitation took place. A second precipitation is conducted just as described above. The precipitate is washed as above described and ignited at a dull red heat. Half the weight obtained represents the ferric oxid and alumina.

=40. Method of Ogilvie.=—For the separation of alumina from phosphoric acid Ogilvie recommends that the filtrate from the phosphomolybdate precipitate be neutralized with ammonia, the precipitate thus formed redissolved in nitric acid, again precipitated with ammonia, filtered, ignited, and weighed as aluminum oxid.[30] If iron be present it will, of course, appear in the product. For use in the examination of mineral phosphates the method can not have a wide application without amendment.

=41. Method of Krug and McElroy.=—Krug and McElroy show that when sufficient alcohol is added to precipitate all of the calcium sulfate in the Glaser method, it will also cause a precipitation of a considerable quantity of iron, by means of which the calcium sulfate will be colored.[31] The presence of potassium and ammonium salts also affects very notably the precipitation of calcium. The method employed by them, in order to avoid these sources of error, is as follows:

One hundred cubic centimeters, equivalent to one gram of the substance, in a nitric acid solution, are placed in a half liter flask and a solution of ammonium molybdate added until all the phosphoric acid has been precipitated. The addition of ammonium nitrate will hasten the separation of the ammonium phosphomolybdate. The liquid should be allowed to stand for twelve hours. The flask is then filled to the mark, the contents well shaken, filtered through a dry filter, and duplicate samples of 200 cubic centimeters each of the filtrate taken for analysis.

A small quantity of ammonium nitrate is dissolved in the liquid, and ammonia cautiously added, keeping the solution as cool as possible. The iron and alumina are precipitated as hydroxids. The mixed hydroxids are collected on a filter, washed with water, the filtrate and washings being collected in a beaker. The precipitate should be dissolved with a small quantity of a solution of ammonium nitrate and nitric acid, again precipitated with ammonia, filtered, washed, ignited, and weighed. This treatment is for the purpose of excluding all possibility of error from the presence of molybdic anhydrid. After weighing, the mixed oxids should be fused with sodium bisulfate, the magma dissolved in water, and the iron determined volumetrically with potassium permanganate after reduction to the ferrous state.

McElroy has further shown by experiments in this laboratory that even the molybdate method of separating the iron and alumina from phosphoric acid with the improvements as first suggested by Krug and himself, may not always give reliable results.[32] In a solution containing ferrous iron equivalent to 56.4 milligrams of ferric oxid, were placed enough of a solution of sodium phosphate to correspond to 100 milligrams of phosphorus pentoxid. The precipitate was dissolved by adding nitric acid, oxidized with bromin water, and the phosphoric acid thrown out with ammonium molybdate. The precipitate was washed with weak nitric acid and the combined filtrate and washings neutralized with ammonia. The resultant precipitate was dissolved in a solution of ammonium nitrate and nitric acid, filtered, and again precipitated with ammonia. In two instances the quantities of material recovered after ignition were 56.9 and 57.3 milligrams, respectively, instead of the theoretical amount, _viz._, 56.4 milligrams.

When the work was repeated after the addition of 400 milligrams of calcium oxid the weight of the precipitate recovered was 62.3 and 63.1 milligrams in duplicate determinations. Similar determinations were made with a known weight, _viz._, 35.6 milligrams of alumina. The treatment of the mixture was precisely as indicated above for iron. The quantity of alumina finally obtained was 28.9 and 29.3 milligrams, respectively, in duplicate determinations. When the lime was added, however, the weights of alumina, recovered, fell to 19.8 and 20.6 milligrams, respectively. These results show that the molybdate method for the separation of iron and alumina in the presence of a large excess of lime and phosphoric acid is subject to widely varying results, but that the error due to the excess of iron in the weighed product is partly corrected by the one due to deficiency of alumina.

=42. Method of Wyatt.=—A method largely used in this country, both in private laboratories and by fertilizer factories, for determining iron and alumina, is described by Wyatt[33]. It is claimed for this method that, while it may not be strictly accurate, yet it is rapid and easy, and in the hands of trained analysts yields concordant results. Fifty cubic centimeters of the first solution of the sample in aqua regia, or an amount thereof equivalent to one gram of the phosphate, in a beaker, are rendered alkaline by ammonia. The resulting precipitate is first redissolved by hydrochloric acid, and a slight alkalinity is again produced with ammonia. Fifty cubic centimeters of strong acetic acid are next added, the mixture stirred and placed in a cool place and left until cold. The precipitate is then separated by filtration and washed twice with boiling water. The vessel holding the filtrate is replaced by the beaker in which the precipitation was made. The precipitate is dissolved in a little fifty per cent hot hydrochloric acid and the filter washed with hot water. After rendering slightly alkaline, as in the first instance, the treatment with acetic acid is repeated as described. The precipitate is washed this time, twice with cold water containing a little acetic acid and three times with hot water. The precipitate is dried, ignited, and weighed as iron and aluminum phosphate. Half of this weight may be taken to represent the quantity of iron and aluminum oxids.

To separate the iron and alumina the precipitate just described is dissolved in hot hydrochloric acid, filtered into a 100 cubic centimeter flask, and made up to the mark by hot wash-water.

The phosphoric acid is determined in one-half of the filtrate and in the remaining half the iron is reduced with zinc and determined with potassium permanganate in the usual way. The phosphoric acid and iron having been thus determined the alumina is estimated by difference. The chief objection to this process is in the excessive quantity of acetic acid used and the danger of solution of the precipitated phosphates caused thereby.

=43. Estimation of the Lime and Magnesia.=—The filtrate and washings from the first precipitation, (paragraph =41=,) of iron and alumina in the method of Krug and McElroy, above described, are collected and sufficient ammonium oxalate is added to precipitate the calcium. The precipitated calcium is very fine and should be collected on a gooch, under pressure. The filtrate and washings from the calcium precipitate are again collected, and a solution of sodium phosphate added to precipitate the magnesia. The solution must be kept cool and slightly alkaline with ammonia during the above operations in order to prevent the separation of molybdic anhydrid.

=44. Estimation of Sulfuric Acid.=—As a rule, sulfates are not abundant in mineral phosphates. In case the samples are pyritiferous, however, considerable quantities of sulfuric acid may be found after treatment with aqua regia.

The acid is precipitated with barium chlorid, in the usual way, in an aliquot portion of the filtrate first obtained. The precipitate of barium sulfate is washed with hot water until clean, dried, ignited, and weighed. If the portion of the filtrate taken represent half a gram of the original material then the weight of barium sulfate obtained multiplied by 0.6858 will give the quantity of sulfur trioxid in one gram.

=45. Estimation of Fluorin by the Method of Berzelius as Modified by Chatard.=—The method of estimating fluorin as proposed by Berzelius, has been found quite satisfactory in the laboratory of the Geological Survey, with the modifications given below.[34]

Two grams of the phosphate are intimately mixed in a large platinum crucible with three grams of precipitated silica and twelve grams of pure sodium carbonate, and the mixture is gradually brought to clear fusion over the blast-lamp. When the fusion is complete the melt is spread over the walls of the crucible, which is then rapidly cooled (preferably by a blast of air). If this have been properly done, the mass separates easily from the crucible, and the subsequent leaching is hastened. The mass, detached from the crucible, is put into a platinum dish into which whatever remains adhering to the crucible, or its lid, is also washed with hot water. A reasonable amount of hot water is now put into the dish, which is covered and digested on the water-bath until the mass is thoroughly disintegrated. To hasten this, the supernatant liquid may, after awhile, be poured off, the residue being washed into a small porcelain mortar, ground up, returned to the dish and boiled with fresh water until no hard grains are left. The total liquid is then filtered, and the residue is washed with hot water. The filtrate (which should amount to about half a liter) is nearly neutralized with nitric acid (methyl orange being used as indicator), some pure sodium bicarbonate is at once added, and the solution (in a platinum dish, if one large enough is at disposal, otherwise in a beaker) is placed on the water-bath, when it speedily becomes turbid through separation of silica. As soon as the solution is warm it is removed from the bath, stirred, allowed to stand for two or three hours, and then filtered by means of the filter-pump and washed with cold water.

The filtrate is concentrated to about a quarter of a liter and nearly neutralized, as before, some sodium carbonate is added, and the phosphoric acid is precipitated with silver nitrate in excess. The precipitate is separated by filtration and washed with hot water, and the excess of silver in the filtrate is removed with sodium chlorid.

The filtrate from the silver chlorid (after addition of some sodium bicarbonate) is evaporated to its crystallizing point, then cooled and diluted with cold water; still more sodium bicarbonate is added, and the whole is allowed to stand, when additional silica will separate, and this is to be removed by filtration.

This final solution is nearly neutralized, as before; a little sodium carbonate solution is added; it is heated to boiling and an excess of solution of calcium chlorid is added. The precipitate of calcium fluorid and carbonate must be boiled for a few minutes, when it can be easily filtered and washed with hot water. The precipitate is then washed from the filter into a small platinum dish and evaporated to dryness, while the filter, after being partially dried and used to wipe off any particles of the precipitate adhering to the dish in which it was formed, is burned, and the ash is added to the main precipitate. This, when dry, is ignited, and allowed to cool; dilute acetic acid is added in excess, and the whole is evaporated to dryness, being kept on the water-bath until all odor of acetic acid has disappeared. The residue is then treated with hot water, digested, filtered on a small filter, washed with hot water, partially dried, put into a crucible, carefully ignited, and weighed as calcium fluorid. The calcium fluorid is then dissolved in sulfuric acid by gentle heating and agitation, evaporated to dryness on a radiator, ignited at full red heat, and weighed as calcium sulfate. From this weight the equivalent weight of calcium fluorid should be calculated, and this should be very close to that actually found as above, but should never exceed it. The difference, which is generally about a milligram (sometimes more), is due to silica precipitated with the fluorid. The percentage of fluorin is, therefore, always calculated from the weight of the sulfate, and not from that of the fluorid obtained.

The main improvements in this method are the use of sodium bicarbonate to separate the silica, and the keeping of the earlier solutions as dilute as possible, which can not be done, if ammonium carbonate be used for the separation of the silica. These changes make the fluorin estimation, although still tedious, far more rapid than before, and the results are very satisfactory.

=46. Modification of Wyatt.=—By reason of the tediousness of the method of Chatard given above, Wyatt has sought to shorten the process by the following modification:[35]

Five grams of the finely ground phosphate are fused in a platinum dish with fifteen grams of the mixed carbonates of sodium and potassium and three grams of fine sand. After fusing very thoroughly with a strong heat for a quarter of an hour, the dish is removed from the fire and cooled. Its contents, dissolved in hot water, are then put into a half liter flask, and a considerable excess of ammonium carbonate is added to the liquid. All the soluble silica falls out of solution, and the flask, after cooling, is made up to the mark with distilled water, well shaken, and then set aside for twenty-four hours to settle. At the end of this time 200 cubic centimeters are carefully decanted through a filter; the filter is well washed, and the filtrate, after being nearly neutralized with hydrochloric acid, is treated with an excess of calcium chlorid solution.

The precipitate, consisting of phosphate, fluorid, and some calcium carbonate, is allowed to settle, and is then carefully washed with boiling water, first by decantation several times, and finally on the filter. After being properly dried in the gas-oven, calcined, and cooled, the residue is treated with acetic acid, placed upon the water-bath, and evaporated to complete dryness.

The calcium acetate is now well washed out by several treatments with boiling water, and the residue is brought upon a filter, dried, calcined, and weighed. The weight represents the calcium phosphate and fluorid contained in two grams of the original sample; and if the calcium phosphate in the residue be determined, according to the usual methods, the difference will be calcium fluorid and may be thus estimated.

_Example._—Assuming the calcined residue of calcium phosphate and fluorid in two grains of the original sample to have amounted to one and six-tenths gram and the calcium phosphate in this quantity to have been determined as 1.540 gram, the calcium fluorid is thus proved to be 0.060 gram, and, therefore, 2: 0.60::100: x = 3 per cent calcium fluorid which, multiplied by 0.4897, gives 1.46 per cent of fluorin.

The above method, while shorter, is not to be preferred to the Chatard process when great accuracy is desired. All the soluble silica may not fall out of the solution as Wyatt says. Finally the fluorin is calculated from small differences in the weight of very heavy precipitates and all the error of the process may be found affecting the numbers for fluorin. For commercial purposes, however, the method is to be recommended for its comparative brevity.

GENERAL METHODS FOR ESTIMATING PHOSPHORIC ACID IN FERTILIZERS.

=47. Preliminary Considerations.=—The chief sources of the phosphoric acid in commercial fertilizers are the mineral phosphates and bones. In respect of the analyses of mineral phosphates detailed directions have been given in the preceding pages. Bones are valuable for fertilizing materials, both because of their content of phosphoric acid and of their organic nitrogen. The methods of treating bones for their phosphoric acid will be found in the general methods for fertilizing materials, and their nitrogen content can be determined by the processes to be described hereafter. Other fertilizing materials also contain phosphorus, as ashes, tankage, oil cakes, and other organic products. In general, the methods for determining the phosphoric acid is the same in all cases, but the means of destroying the organic matter precedent to the analysis vary in different cases. In most cases a simple ignition is sufficient, while, if the phosphorus be found in certain organic products, the oxidation must be accomplished by one of the methods described in the processes adopted by the official chemists, or by the means described in volume first, paragraph =378= or =382=. In all cases of acid phosphates and superphosphates, the water and ammonium citrate soluble phosphoric acid is to be determined as well as the total. In basic slags the amount soluble in ammonium citrate or dilute citric acid is also to be ascertained.

In all cases where soluble or so-called reverted acid is to be considered, the analysis must be performed without previous desiccation or ignition. If water content or loss on ignition are to be considered, the operation to determine them must be conducted on a separate part of the sample.

The methods of analysis which have been adopted by associations of chemists should be given the preference in the conduct of the work, although it must be admitted that they may contain sources of error, and may be in no respect superior to processes employed by chemists in their private capacity. In this country the methods adopted by the Association of Official Agricultural Chemists should be followed as closely as possible. The great merit of other methods, however, must not be denied. Especially those methods which shorten the time required or diminish the labor and expense of the analysis are worthy of careful consideration. In factory work, for instance, it is often far more important for the chemist to be able to rapidly determine the phosphoric acid in a great number of samples with approximate accuracy than to confine his work to one with absolute precision. Some of the shorter methods, moreover, notably the citrate process, appear to be quite, if not altogether, as reliable as the molybdate method, while in the case of the uranium volumetric process, it must not be forgotten that it is almost the only one practiced in France. Other volumetric processes are given in full, as, for instance, the one perfected by Pemberton, but data are still lacking to justify their strong recommendation. It should be remembered that this manual is not written for the beginner but rather for the chemist already acquainted with the principles and practice of general chemical analysis, and it is, therefore, expected that each analyst will make intelligent use of the data placed at his disposal.

=48. Determination of Phosphoric Acid with Preliminary Precipitation as Stannic Phosphate.=—This method, once much in use and highly recommended, is now almost unknown among the processes of fertilizer control. It was first proposed and described by Girard, and rests on the precipitation of the phosphoric acid in a nitric acid solution by means of metallic tin.[36] The stannic acid formed by the oxidation of the tin unites with the phosphoric acid held in a free state by the nitric acid. The precipitation of the phosphoric acid is said to be complete, but a trace of it has been found in the iron and alumina subsequently separated from the solution. The precipitate obtained is dissolved in caustic potash, whereby soluble potassium metastannate and phosphate are obtained. Following is the method of conducting the analysis as described by Crookes:[37]

The phosphate should be dissolved in nitric acid, and any chlorin present be expelled by repeated evaporations with the solvent. Finally, to the evaporated mass the strongest nitric acid is added. Pure tin-foil is added and heat applied. The phosphoric acid is precipitated by the stannic acid formed. The quantity of tin used should be from six to eight times as great as that of the phosphoric acid present.

The precipitate is collected on a filter, washed, and dissolved in caustic potash. The solution is saturated with hydrogen sulfid, and on adding acetic acid in slight excess the tin sulfid is separated and removed by filtration. The whole of the phosphoric acid, supposed to be almost free of tin, is now found in the filtrate. The filtrate is concentrated to small bulk and any tin sulfid present separated by filtering, and the phosphoric acid finally removed from the ammoniacal filtrate by precipitation with magnesia mixture. The chief difficulties of this method are to be found, on the one hand, in the retention of some of the phosphoric acid by the iron and alumina which may be present, and on the other, in the presence of some tin in the final magnesium pyrophosphate. If the tin be all removed as sulfid, the latter source of error will be avoided. It is difficult to secure pure metallic tin, and this is another disturbing element in the process. It can not be recommended for the work which agricultural analysts are usually called on to perform.[38]

=49. Water-Soluble Phosphoric Acid.=—The method of procedure recommended by the Association of Official Chemists is as follows:[39] Place two grams of the sample in a nine centimeter filter; wash with successive small portions of cold water, allowing each portion to pass through before adding more, until the filtrate measures about 250 cubic centimeters. If the filtrate be turbid, add a little nitric acid. Make up to any convenient definite volume; mix well; take any convenient portion and proceed as under total phosphoric acid.

=50. Citrate-Insoluble Phosphoric Acid.=—Heat 100 cubic centimeters of strictly neutral ammonium citrate solution of 1.09 specific gravity to 65° in a flask placed in a bath of warm water, keeping the flask loosely stoppered to prevent evaporation. When the citrate solution in the flask has reached 65°, drop into it the filter containing the washed residue from the water-soluble phosphoric acid determination, close tightly with a smooth rubber stopper, and shake violently until the filter paper is reduced to a pulp. Place the flask back into the bath and maintain the water in the bath at such a temperature that the contents of the flask will stand at exactly 65°. Shake the flask every five minutes. At the expiration of exactly thirty minutes from the time the filter and residue were introduced, remove the flask from the bath and immediately filter as rapidly as possible. It has been shown by Sanborn, in this laboratory, that the filtration is greatly facilitated by adding asbestos pulp. Wash thoroughly with water at 65°. Transfer the filter and its contents to a crucible, ignite until all organic matter is destroyed, add from ten to fifteen cubic centimeters of strong hydrochloric acid, and digest until all phosphate is dissolved; or return the filter with contents to the digestion flask, add from thirty to thirty-five cubic centimeters of strong nitric, and from five to ten cubic centimeters of strong hydrochloric acid, and boil until all the phosphate is dissolved. Dilute the solution to 200 cubic centimeters. If desired, the filter and its contents can be treated according to methods (1), (2), or (3), under total phosphoric acid. Mix well; filter through a dry filter; take a definite portion of the filtrate and proceed as under total phosphoric acid.

In case a determination of citrate-insoluble phosphoric acid be required in non-acidulated goods it is to be made by treating two grams of the phosphatic material, without previous washing with water, precisely in the way above described, except that in case the substance contain much animal matter (bone, fish, etc.), the residue insoluble in ammonium citrate is to be treated by one of the processes described below under total phosphoric acid, (1), (2), or (3).

=51. Total Phosphoric Acid.=—In case of ignition the residual material is to be dissolved in hydrochloric acid. The following methods of treating the raw material, using two grams in each case, may be employed: (1) Evaporate with five cubic centimeters of magnesium nitrate, ignite, and dissolve in hydrochloric acid. (2) Boil in a Kjeldahl flask graduated to 250 cubic centimeters, with from twenty to thirty cubic centimeters of strong sulfuric acid, adding from two to four grams of sodium or potassium nitrate at the beginning of the digestion and a small quantity after the solution has become nearly colorless; or adding the nitrate in small portions from time to time. After the solution is colorless, add 150 cubic centimeters of water and boil for a few minutes, cool, and make up to volume. (3) Digest with strong sulfuric acid and such other reagents as are used in either the plain or modified Kjeldahl or Gunning methods for estimating nitrogen. Do not add any potassium permanganate, but after the solution has become colorless add about 100 cubic centimeters of water and boil for a few minutes, cool, and make up to a convenient volume; two and five-tenths grams of substance and a digestion flask graduated to 250 cubic centimeters are recommended. This method will be found convenient when both the nitrogen and the total phosphoric acid are to be determined in a fertilizer. In this case, after diluting the volume and mixing, a part for the estimation of nitrogen, may be removed with a pipette and the remainder then filtered through a dry filter and a portion taken for the determination of the total phosphoric acid. (4) Dissolve in thirty cubic centimeters of concentrated nitric acid and a small quantity of hydrochloric acid. (5) Add thirty cubic centimeters of concentrated hydrochloric acid, heat, and add cautiously, in small quantities at a time, about five-tenths gram of finely-pulverized potassium chlorate. (6) Dissolve in from fifteen to thirty cubic centimeters of strong hydrochloric and from three to ten cubic centimeters of nitric acid. This method is recommended for fertilizers containing much iron or aluminum phosphate. Boil until all phosphates are dissolved and all organic matter is destroyed; cool and dilute to 200 or 250 cubic centimeters; mix and pass through a dry filter; take an aliquot part of the filtrate corresponding to a quarter, half, or one gram, neutralize with ammonia, and clear with a few drops of nitric acid. In case hydrochloric or sulfuric acid have been used as a solvent, add about fifteen grams of dry ammonium nitrate.

To the hot solutions, for every decigram of phosphorus pentoxid that is present, add fifty cubic centimeters of molybdic solution. Digest at about 65° for an hour, filter, and wash with water or ammonium nitrate solution. Test the filtrate by renewed digestion and the addition of more molybdic solution. Dissolve the precipitate on the filter with ammonia and hot water and wash into a beaker to a bulk of not more than 100 cubic centimeters. Nearly neutralize with hydrochloric acid, cool, and add magnesia mixture from a burette; add slowly (about one drop per second), stirring vigorously. After fifteen minutes add thirty cubic centimeters of ammonia solution of 0.95 density. Let stand for some time; two hours are usually enough. Filter, wash with dilute ammonia, ignite gently at first and then at white heat for ten minutes, and weigh. For the quantity of magnesia mixture to be added see paragraph =21=.

=52. Citrate-Soluble Phosphoric Acid.=—The sum of the water-soluble and citrate-insoluble subtracted from the total gives the citrate-soluble phosphoric acid.

=53. Preparation of Reagents.=—(1) _Ammonium Citrate Solution._—(_a_) Mix 370 grams of commercial citric acid with 1,500 cubic centimeters of water, nearly neutralize with commercial ammonia, cool, add ammonia until exactly neutral (testing with saturated alcoholic solution of corallin) and bring to a volume of two liters. Test the specific gravity, which should be 1.09 at 20°, before using.

(_b_) _Alternate Method._—To 370 grams of commercial citric acid add commercial ammonia, of 0.96 specific gravity, until nearly neutral; reduce the specific gravity to nearly 1.09 and proceed as follows: Prepare a solution of fused calcium chlorid 200 grams to the liter, and add four volumes of strong alcohol. Make the mixture exactly neutral, using a small amount of freshly prepared corallin solution as a preliminary indicator, and test finally by withdrawing a portion, diluting with an equal volume of water, and testing with cochineal solution. Fifty cubic centimeters of this solution will precipitate the citric acid from ten cubic centimeters of the citrate solution. To ten cubic centimeters of the nearly neutral citrate solution add fifty cubic centimeters of the alcoholic calcium chlorid solution, stir well, filter at once through a folded filter, dilute with an equal volume of water, and test the reaction with neutral solution of cochineal. If acid or alkaline, add ammonia or citric acid, as the case may be, to the citrate solution, mix, and test again as before. Repeat this process until a neutral reaction of the citrate solution is obtained. At the end the specific gravity must be 1.09 at 20°.

(2) _Molybdic Solution._—See paragraph =22=.

(3) _Ammonium Nitrate Solution._—Dissolve 200 grams of commercial ammonium nitrate in water and bring to a volume of two liters.

(4) _Magnesia Mixture._—See paragraph =22=.

(5) _Dilute Ammonia for Washing._—See paragraph =22=.

(6) _Magnesium Nitrate._—Dissolve 320 grams of calcined magnesia in nitric acid, avoiding an excess of the latter; then add a little calcined magnesia in excess, boil, filter from the excess of magnesia, ferric oxid, etc., and bring to volume of two liters.

=54. Official Methods with Norwegian Fertilizers.=—The Director of the Chemical Control Station of Norway, expresses the opinion, that for Norwegian, Swedish, Danish, and German conditions, the American methods for the determination of phosphoric acid, notwithstanding their analytical exactness, are quite inapplicable.[40] In those countries are found many, in part, poorly pulverized and badly mixed manures, such as ammonium superphosphate, potassium superphosphate, and potassium ammonium superphosphate, and these can not usually be so well pulverized and mixed that one can take out a true average sample of from two to two and five-tenths grams. Care in the analysis is useless when the material employed does not represent the average condition of the materials investigated. Therefore, in the countries named, often from ten to twenty grams, and almost never less than five grams of substance are taken in the preparation of the solutions, except for instance, in the determination of nitrogen and reverted phosphoric acid.