Analytical Control of the Ammonia Oxidation Process. - Industrial

Citation data is made available by participants in Crossref's Cited-by Linking ... Die Anwendung der thermischen Dissoziation der Ammoniumhalogenide i...
0 downloads 0 Views 663KB Size
1106

T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y

ratios may be calculated for the materials designated "A,,, ((B"and (IC.>> A-o:p:m = 6.5:12.8.. B-o:p:m = 6.0:13.0:1 C-o:p:m = 8.3:11.5:1 AS meta-cresol boils only 3" higher t h a n p-cresol, it seems certain t h a t in t h e original materials dealt with t h e amount of p-cresol must much exceed t h e amount of m-cresol. Weiss a n d Downs in t h e preparation of their graphs made u p synthetic mixtures in which t h e ratio of para t o meta was fixed a t about I : I and they s t a t e t h a t t h a t is t h e approximate ratio in which they exist in their oils. The question arises whether their method would give reliable results when applied t o t h e t a r acids dealt with b y t h e author. I n t h e present method such variation in t h e ratios of t h e cresols present would be eliminated, a t a n y rate t o a considerable extent, although t h e ratio of cresols worked with in t h e preparations of t h e synthetic mixtures was not varied as far as t o include a ratio, m : p = I : I , so t h a t definite figures cannot be given. It would be interesting t o follow up this a n d t h e other points raised a t greater length b u t t h e author has not now access t o t h e necessary crude materials. CHEXICAL DEPARTMENT, MCGILLUNIVERSITY MONTREAL. CANADA

ANALYTICAL CONTROL OF THE AMMONIA OXIDATION PROCESS BY GUY B. TAYLOR A N D Jos. D. DAVIS Received October 17, 1917

I n t h e manufacture of nitric acid from ammonia, a mixture of air a n d ammonia gas is passed over a suitable catalytic material heated t o temperatures above red heat. Two reactions occur: 4 NH3+5 oz=4 N 0 + 6 HzO (1) 4 NH3+3 0 2 = 2 N2+6 H20 (2) It has been suggested t h a t reaction ( 2 ) may come about by means of t h e intermediate reaction,* 4 NH3+6 N O = j N2+6 H20. (3) I n testing out a particular t y p e of commercial oxidizer, t h e authors believe they have found evidence t h a t under certain conditions this reaction does occur. T h e oxidizer in question was so constructed t h a t the burned gases remained hot f o r Some time' Further' there was irregular local cooling of t h e catalyst due t o eddy currents in t h e burned gases. Consequently, there must have been considerable ammonia passing such points unburned. I n fact, analysis of samples taken from points near t h e cold spots showed ammonia while in samples taken outside the catalyst chamber t h e ammonia content was very low. T h e authors believe t h a t most of t h e NH, passing through t h e cold spots was subsequently "burnedj, b y the hot NO. ~~~~l~~ with this apparatus were uniformly about I O per cent lower t h a n those obtained with t h e same lot of catalytic material in a differently constructed oxidizer. The nitric oxide is converted into nitric acid b y passing t h e gases through suitable absorption towers, 1 2

Published by permission of Director of U.S. Bureau of Mines. Reinders and Cats, Chem. Weekblad, 9 (1912), 47-58.

Vol. 9, No.

12

much t h e same as in t h e arc process of direct fixation of atmospheric nitrogen. T h e efficiency of a n ammonia converter depends upon establishing conditions favorable t o reaction (I) a n d suppressing as far as possible reaction ( 2 ) . If t h e ultimate product desired is nitric acid, little or no free ammonia should be allowed t o pass t h e converter unchanged. I n order t o test t h e efficiency of conversion, analyses of t h e entering ammonia-air mixture and t h e exit nitrose gases must be made. The former offers no special difficulties b u t t h e nature of t h e acid gas creates a special problem. As soon as t h e gas sample cools, t h e nitric oxide begins t o react with t h e excess oxygen present t o form NOz, which partially dissolves in t h e condensed water, so t h a t t h e gas taken into a n y sampling device consists of a mixture of nitrogen, oxygen, and nitrogen oxides of indeterminate molecular species. The principle upon which t h e efficiency calculation is based is as follows: t h e ratio of t h e nitrogen combined as ammonia t o total nitrogen in t h e intake gas is equal t o t h e ratio of t h e nitrogen derived from t h e ammonia t o total nitrogen in t h e exit gas. Let a =nitrogen combined as N H 3 in t h e air-ammonia b =free nitrogen sample and c =nitrogen combined as nitrogen oxides in t h e d = free nitrogen } exit gas f = nitrogen combined as ammonia escap) sample ing oxidation T h e nitrogen in each sample must of course be expressed in terms of the same unit. Then

,

(3)

where x is t h e free nitrogen derived from t h e reaction expressed by Equation ( 2 ) . The efficiency is then expressed by t h e relation -C- - Yield

c + f + x or, substituting t h e value of x from (3), a(c

b, +d+ I,X

IOO

= Per cent Yield.

(4)

(j)

EXPERIMENTAL METHODS

I n t h e course of some experiments on ammonia oxidation, several methods of procedure were developed: M E T H O D I-The apparatus is sketched in Fig. I. Air and was passed through the meter A , the water pressure gauge B , and t h e bottles c containing ammonia liquor ( 7 per cent "3). The ammoniaair mixture passed over t h e catalyzer in D a n d t h e acid gases Out through E' The tubes H H',capacity about 1 2 0 0 cc., were filled with water. K contained 2 0 cc. N / s and K" '5 cc* N / 2 NaoH plus 3 cc. Of 3 per cent hydrogen peroxide' The ammonia-air sample was drawn through K b y allowing a measured volume of water t o flow from H . Before drawing t h e acid sample t h e air in K" was displaced b y pure oxygen passing in at G and out at F. T h e two samples were then drawn simultaneously SO

''

I

T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERIlVG C H E M I S T R Y

Dec., 1917

t h a t about one liter of gas was taken into each sample in about one hour. The excess acid in K \\-as titrated with N / j alkali for t h e determination of t h e value of a. T h e gas in K ” was displaced with oxygen into H ‘ , noting t h e total volume of water drawn therefrom. The two tubes H H‘ were then detached and immersed in a large t a n k of water, a manometer attached and t h e temperature a n d pressure noted. T h e gas in H’ was analyzed for nitrogen by absorbing t h e oxygen in alkaline pyrogallol. T h e composition of t h e air in H was taken as 79 per cent nitrogen. This furnished the d a t a necessary for t h e calculation of t h e values b and d . For determination of t h e value c t h e contents of K ” were washed into a small Erlenmeyer flask and boiled for about j minutes, then thoroughly cooled a n d t h e excess alkali titrated with N / j NaOH, methyl orange indicator a n d a companion flask. T h e nitrite present tended t o destroy t h e indicator in acid solution, turning t h e color a permanent yellow, unaffected by either acids o r alkalies. By taking care not t o run over t h e endpoint, noting t h e first change in shade from t h e color of t h e companion flask, this titration was accurate a n d satisfactory. T h e boiling and cooling was found t o be absolutely necessary in order t o recognize t h e end-point a c a l l . Carpenter’ condemned t h e alkali HzOz reagent for absorbing nitrous gases on account of t h e difficulty

A

B

C

FIG.I-ANALYRCAL

C

CONTROL

OF

AMMONIA OXIDATION

of:the end-point a n d proposed t h e use of a solution of hydrogen peroxide alone, b u t we have not found i t as efficient as t h e alkali, when t h e gas is bubbled through it. When methyl red or sodium alizarine sulfonate is used as a n indicator, t h e boiling is not necessary, both indicators working fairly well in t h e presence of nitrites. Sodium alizarine sulfonate is rather t h e better indicator t o use, being more sensitive. nlethyl red is destroyed t o some extent by t h e nitrite b u t t h e end-point is sufficiently sharp even when N/IO solutions are used. Since no appreciable ammonia escaped oxidation in tests where this method was used, t h e value of j is zero. T h e absorption of t h e acid in K ” was practically complete. The water drawn from H‘ was never acid and t h a t remaining in t h e tube was titrated after shaking when especial accuracy was required. This rarely increased t h e value of c more t h a n one per cent. METHOD 11-This method was essentially t h e same as Method I, except t h a t t h e nitrogen-oxygen mixture, after absorption of t h e acid, was drawn over hot copper t o remove t h e oxygen, t h u s eliminating t h e analysis of t h e gas collected in H’. A small Meyer sulfur bulb was used as a n absorber a n d a 4-liter aspirator bottle substituted for H’,Fig. I. “Note on Some Condition Affecting the Oxidation of Nitrous Acid t o Nitric Acid,” J . SOL.Chem. Ind.. 6 (1886). 287.

1107

The above methods were especially useful in small scale experiments where sampling into evacuated bottles as described below would have seriously interfered with t h e constancy of t h e gas flow through t h e catalyzer tube. M E T H O D 111-For larger scale experiments samples taken into evacuated bottles have been found convenient. This is t h e only method permitting determination of t h e ammonia escaping oxidation, f. Bottles of known volume, I to 2 . j liters capacity, were provided with single capillary stopcocks and evacuated t o less t h a n z mm. pressure. Ground glass stoppers were more satisfactory b u t soft rubber stoppers served well enough when t h e samples were drawn immediately after evacuating. Sufficient N / j H2S04 was admitted t o t h e ammonia sample and a measured volume of water, after absorbing t h e ammonia, until t h e residual air was a t atmospheric pressure. Enough condensation occurred in t h e acid sample t o permit sucking in a measured volume of water containing hydrogen peroxide t o absorb t h e acid fumes. The volume of t h e bottle minus t h a t of solution introduced gave t h e volume of t h e oxygen-nitrogen gas which, together with its analysis, furnished t h e necessary d a t a for calculating t h e free nitrogen. The acid solution was titrated with N/5 NaOH for free acid and t h e n distilled with caustic into A r / j HzS04 for determination of ammonia f. Since this ammonia mas present as nitrate, its equivalent was added t o t h e free acid already found for t h e calculation of c. While this method requires rather more manipulative skill t h a n t h e average works chemist possesses, the authors consider it t h e best available and have used i t quite extensively for control work with very satisfactory results. Duplicate determinations on samples taken a t t h e same instant should agree within I per cent, b u t in practice agreement so close will hardly be consistently obtained unless t h e gas flow is very uniform. I t is desirable t o check t h e analysis method and t h e standard solution used by the analysis of gas samples containing a known amount of nitric oxide. The NO required may be conveniently prepared by means of a nitrometer. I t is often desirable t o know whether or not there has been a leak between t h e point where t h e NH3 sample was taken and t h e point where t h e nitrous gases were sampled. T o determine this it is only necessary t o measure t h e amount of oxygen admitted t o t h e sample of nitrous gases a n d use water t o which a known amount of standard H202 has been added for absorption. All t h e combined nitrogen in t h e sample will now be present as nitric acid. After t h e excess oxygen gas has been determined t h e liquid is divided into t w o parts. I n one part t h e acid is determined by titration with standard alkali a n d in t h e other part t h e excess HzOz is determined by titration with permanganate. T h e oxygen used is now known a n d t h e amount required for t h e oxidation of t h e ammonia converted may be calculated, a n y excess being due t o leakage. Obviously, t h e yield may be calculated from a knowledge of t h e relation of combined nitrogen t o oxygen in both the NH, air sample a n d t h e sample of nitrous gas. A few such calculations

I 108

T H E J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y

were made with results agreeing fairly well with those b y t h e regular method of calculation. T h e vacuum method offers a convenient means of following t h e reactions in a n absorption system; for example, i t is a simple matter t o determine t h e efficiency of a given tower b y sampling a n d analyzing t h e gas on both sides of it. Ordinarily t h e leakage will not be great enough t o interfere, O T H E R METHODS-Analysis of t h e ammonia-air intake gas may be made by ordinary gas analysis methods by taking proper precautions. I n common gas analysis practice i t is customary t o make all volume readings with t h e gas saturated with water vapor. When mercury is used as a confining liquid, t h e walls of t h e burette are always kept wet. This precaution is essential because of t h e very different vapor pressures of t h e solutions used as selective absorbents. I n dealing with ammonia gas i t is obviously necessary t o have t h e walls of t h e burette dry a n d t o absorb t h e ammonia in a reagent whose vapor tension is equal t o t h e partial pressure of t h e water vapor in t h e sampIe. Now this result is accomplished if t h e gas is thoroughly dried a n d t h e absorption made in a reagent whose vapor tension is practically zero. T h e apparatus employed consisted of a special bulb form of compensated burette with graduations extending from 140 t o 165 cc., easily readable t o 1 0 . 0 2 cc. Clean, dry mercury was used as a confining liquid a n d t h e ammonia was absorbed in concentrated sulfuric acid. T h e sample was introduced into t h e burette through a U-tube filled with pieces of solid potassium hydroxide, which thoroughly dried it. This method proved very satisfactory in practice, a n d checked the vacuum bottle method. Its advantages are ease of manipulation a n d simplifying of t h e calculation, since no temperature or barometer observations are required a n d no titrations are made. T h e calculation follows: a =contraction in volume of t h e sample b =volume after contraction X 0.79 X 2 . An equally simple method suggested itself for analysis of t h e nitrous gases. It was thought t h a t a sample of t h e nitrous gases drawn into a nitrometer containing mercury a n d sulfuric acid, a n d then shaken, would convert all t h e nitrogen oxides t o NO and cause t h e excess oxygen t o disappear. Analysis of t h e remaining mixture of nitrogen a n d nitric oxide would give t h e values for c a n d d in volumes as in t h e burette method for t h e "3-air mixture. I n experiments with this method i t was found t h a t a concentrated solution of chromic acid' rapidly absorbed N O and effected quantitative separation of this gas from artificially prepared mixtures of nitrogen a n d nitric oxide. When t h e method was applied t o t h e nitrous gases t h e results were always low. A series of experiments was then made by introducing a measured volume of air into t h e decomposition bulb of a d u P o n t nitrometer and adding a known quantity of nitric acid dissolved in concentrated sulfuric acid. After shaking several minutes t h e residual gas was measured, transferred t o a gas analysis burette a n d t h e N O absorbed 1

Bohmer. 2.anal. Chem., 2 1 (1882). 212.

Vol. 9, No.

12

by passing into a pipette containing t h e chromic acid reagent. Mercury was used as a confining liquid a n d t h e walls of t h e burette were kept wet. The results in every case showed t h a t t h e residual gas measured corresponded t o t h a t calculated, assuming disappearance of t h e oxygen in t h e volume of air used and evolution of nitric oxide from t h e nitric acid introduced. However, analysis of this gas always showed a considerably less percentage of N O t h a n t h a t calculated as well as a small amount of oxygen as determined by absorption in alkaline pyrogallol. Evolution of t h e N O in t h e nitrometer bulb in t h e presence of nitrogen instead of air gave more nearly t h e correct percentage of N O in t h e residual gas. I n view of these results t h e nitrometer method appeared unavailable. A second method for eliminating titrations was devised. Fig I1 shows t h e essential details of t h e apparatus. T h e ammonia air sample is drawn into t h e gas burette ( 2 ) through t h e drying t u b e (I) containing pieces of solid potassium hydroxide. Burette ( 2 ) is provided with a right angle two-way stopcock at t h e t o p and a leveling t u b e below not shown i n figure. It is filled with

U

U

no. 11-ANALYTICALCONTROL

OF

AMMONIAOXIDATION

dry mercury. Bulb (3) is detached from t h e apparatus a n d completely filled with water containing hydrogen peroxide and indicator.' A small leveling bottle is attached by means of a flexible rubber tube at t h e bottom. About I O cc. of pure oxygen, which need not be measured, are introduced into t h e bulb through one branch of t h e T tube. The nitrous gases are t h e n drawn from t h e oxidizer into t h e bulb through t h e same t u b e by allowing 7 0 t o 80 cc. of water to run out into t h e leveling bottle. This can be done without loss of acid since t h e indicator shows plainly t h e diffusion of any acid through t h e water while t h e sample is being taken. T h e bulb is t h e n shaken until all t h e nitrogen oxides are absorbed. It is then restored t o its position (Fig. 11) and t h e residual gas transferred t o t h e mercuryfilled burette ( 5 ) . T h e walls of this burette are kept wet. T h e gas is now passed into alkaline pyrogallol in t h e Williams pipette (4),the residual nitrogen brought back into t h e burette a n d measured. This volume is called d , and is discarded. T h e ammonia sample in ( 2 ) is 1 Methyl

red or sodium alizarine sulfonate.

Dec., 1917

T H E J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y

now bubbled slowly i n t o t h e acid liquid in (3) until neutral, noting t h e volume required. T h e air from (3) is measured in burette (5) a n d this volume multiplied b y 0.79 corresponds t o t h e nitrogen volume b. I n calculating t h e yield b y t h e general formula c(a b)

- + a(c

+ + n’

neglecting f, since b y t h e above procedure a a n d c are equal, we have a + b a d‘ T h e volume of dry gas introduced from ( 2 ) into (3) corrected t o moisture saturation, minus t h e volume of air measured, gives t h e volume of ammonia. This volume must be divided b y t w o t o obtain t h e corresponding volume of nitrogen, a, t o be applied in t h e formula. When t h e yield is known t o be in t h e neighborhood of 90 per cent approximate results may be obtained by simple division of t h e measured nitrogen volumes, i. e . , b / d . While t h e method has not been thoroughly tested experimentally, i t is presented here because t h e principle involved offers possibilities for development of a rapid method of works control.

+

DISCUSSION

It was realized at t h e beginning of this investigation t h a t no method of determining t h e efficiency of a n industrial process appeals t o a manufacturer more t h a n pounds p u t in t o pounds taken out. I n order t o p u t our analytical results upon a n unquestionable foundation, t h e method was checked by experiments in which t h e ammonia input was determined b y weight and t h e total acid produced absorbed. These experiments were carried out as follows: two tared flasks containing a known weight of carefully analyzed ammonia solution were connected in series so t h a t air could be blown through them, t h e air-ammonia mixture passing over a catalyzer in a glass tube. T h e acid gases were absorbed in a series of wash bottles containing water a n d finally in t h e last ones I O per cent N a O H solution. T h e ammonia was determined by weighing t h e flasks and re-analyzing; t h e acid, by titrating t h e HNOs absorbed in t h e water a n d analyzing t h e alkaline solutions for nitrate a n d nitrite. Although t h e absorption system was quite elaborate some acid passed through unabsorbed. This was estimated b y frequent samples of t h e issuing gas. Method I1 was used for checking t h e above outlined procedure because t h e time i n drawing t h e samples could be made t o coincide more nearly with t h e length of t h e run. The results follow: Liters GRAMS Air r H N O a RECOVEREERoom Grams Alkali Bottles Time Condi- N H I Water as as Min. tions Taken Bottles Nitrate Nitrite 144 200 10.82 24 40 1.40 4.90 180 270 1344 33.00 1.40 6.80 172 187 8.74 22.70 0.60 3.15

Grams HNO: End

Loss Grams

PER CENT -YIELD-

BY

By Analyfrom Total Train HNOs Weights sis 090 31.60 79.0 82.5 42.25 85.1 87.0 1.05 0.60 27.05 83.8 85.5

There is no doubt in our minds t h a t t h e analysis method gives t h e t r u e efficiency of conversion. The complete absorption of t h e acid gases was exceedingly difficult a n d losses undoubtedly occurred through leaks at rubber connections. These facts account for t h e

1109

slightly lower results of t h e total absorption method. H a d we cared t o elaborate t h e train, no doubt all t h e acid could have been absorbed and t h e agreement would have been better. I n a recently published article Fox1 has raised t h e question of possible oxidation of ammonia t o ammonium nitrite b y hydrogen peroxide. This point was carefully tested in a number of experiments a n d no evidence secured of any such oxidation in significant quantity under t h e conditions obtaining in t h e above outlined procedures. The following experiment seems t o be conclusive. A sample of t h e nitrose gases was drawn into a 2-liter evacuated bottle from a n experimental converter running a t low velocity so t h a t no unchanged ammonia passed t h e catalyzer. A solution containing 40 cc. N / 5 NaOH, 15 cc. 3 per cent hydrogen peroxide and 3.5 cc. N / 5 NHaOH was introduced into t h e bottle immediately, followed b y I O O cc. distilled water. T h e NaOH was in excess of t h a t needed t o neutralize t h e acid present. After vigorous shaking t h e bottle was allowed t o stand several minutes. Then a n excess of N / I HzSO4was added t o fix t h e ammonia. The contents of t h e bottle were transferred t o a Kjeldahl distillation apparatus, made alkaline with NaOH, and t h e ammonia distilled into N / 5 acid. Exactly t h a t amount of ammonia taken was recovered. No ammonia was oxidized. We have further evidence pointing t o t h e f a c t t h a t under t h e conditions of analysis t h e use of Hz02 in alkaline solution does not affect t h e accuracy of t h e results by oxidation of NHa t o nitrite. Two chemists using t h e vacuum method described above made a number of analyses of parallel samples, one using alkali with H 2 0 2 a n d t h e other using water with H 2 0 2for absorbing t h e n i t r o u s gases. The results agreed very well, showing t h a t there could not have been appreciable oxidation of N H a by H 2 0 in alkaline solutions. iVoreover, Hoppe-Seyler2 found no evidence of oxidation of ammonia in solutions of ammonia, or a m monium carbonate, or when a fixed alkali was present, b y dilute hydrogen peroxide on 24 hrs. standing. It was necessary t o concentrate such solutions by evaporation t o small volume before nitrite could be detected b y colorimetric methods. The statements of Schbnbein, Weith, and Weber3 are not t o be taken t o mean t h a t ammonia is oxidized by hydrogen peroxide “abundantly” under all conditions. FOX’S,* analytical method is based on t h e assumption t h a t N O in t h e presence of a n excess of oxygen reacts immediately t o NOz (or Nz04). It has been shown b y a number of investigations5 t h a t t h e reaction 2 NO 02 _r 2 NO2 occurs in measurable time a n d t h a t a mixture of S O 2 a n d N O behaves toward alkali as NzOa, being rapidly absorbed as nitrite. I n Fox’s method all t h e acid would be absorbed b u t it is

+

1

THISJOURNAL, 9 (1917). 737-43.

2

Be7 , 16 (1883), 1917-24. I b i d . , 7 (1874), 1745.

8



LOC. Cil.

Holwech. 2. angeu. Chem.. 21 (1908), 2131; Le Blanc, 2. Elck.. IS (1906), 541; Foerster and Koch, 2. angeu. Chem., 21 (1908).2161 and 2209; Foerster and Blich. Ibid.. 28 (1910), 2017. 6

T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y

I110

extremely doubtful t h a t t h e NO could go completely t o NO2 in his apparatus, in which case results calculated b y t h e formula given would be too low. A C K N 0 WLE D G M E N T S

The authors have h a d valuable assistance, in t h e work on which this paper is based, from Lieut. G. A. Perley, Mr. J. H. Capps, a n d Mr. L. R. Lenhart; M r , A. S. Coolidge performed many of t h e experiments in trying out t h e nitrometer method. BUREAUO F MINES WASHINGTON. D. C.

______

A VOLUMETRIC METHOD FOR THE DETERMINATION OF FORMIC ACID OR FORMATES IN THE PRESENCE OF HYDROXIDES, CARBQNATES, OXALATES AND ACETATES By F. TSIROPINAS Received August 16, 1917

This method is based upon t h e well-known fact t h a t formic acid is quantitatively oxidized t o carbon dioxide b y chromic acid in boiling solution, and t h a t therefore every 44 parts by weight of carbon dioxide correspond t o 46 parts of formic acid according t o t h e equation, 2Cr03 3HCOOH = Crz03 3 H 2 0 3C02. APPARATUS-The accompanying sketch shows t h e arrangement of t h e apparatus used: A is a n Erlenmeyer flask of resistant glass with a capacity of about 500 cc.; B is a thermometer for t h e control of t h e temperature of t h e contents of t h e flask; C is a 15-in. Liebig reflux condenser, for t h e condensation of t h e vapors produced during t h e boiling of t h e liquid and is connected b y means of a glass tube' and heavy rubber tubing t o D, t h e gas collector. T h e latter consists of a glass tube, about 1.5 in. in diameter, having a capacity of approximately 600 cc. a n d is graduated in cc.'s; connected t o this gas collector b y means of a rubber t u b e is t h e reservoir E , b y means of which t h e air in t h e collector may be discharged and t h e water level obtained, when t h e COZ gas has been collected in t h e graduated tube. T o prevent t h e absorption of gas by t h e water used for leveling purposes, a I-in. layer of paraffin oil is floated upon t h e surface. REAGENT-50 g. of C. P. sodium bichromate are dissolved in 500 cc. of distilled water, 80 cc. of C. P. concentrated sulfuric acid added a n d t h e liquid boiled for 5 minutes t o expel all dissolved gases; then i t is allowed t o cool t o room temperature. P R O C E D U R E (1)-In t h e absence of other substances which give CO2 when boiled with chromic acid, t h e procedure is as follows: A solution is prepared, so t h a t 5 0 cc. of t h e same will contain from 0.5 t o 1.0 g. of formic acid; 5 0 cc. of this solution are emptied b y means of a pipette into t h e clean Erlenmeyer flask and if alkaline, 2 t o 3 drops of methyl orange are added, a n d t h e solution made acid with dilute sulfuric acid ( I : I ) ; 400 cc. of t h e chromate solution are now

+

+

+

Vol. 9 , No.

12

added and t h e flask attached t o t h e condenser; t h e stopper a t t h e t o p of t h e condenser is removed and t h e reservoir raised until t h e graduated t u b e is filled t o t h e zero mark. The stopper is now firmly replaced a n d a reading is made upon t h e graduated t u b e for correction of t h e air displaced by t h e stopper. T h e reservoir is then lowered a n d observation is made t o assure t h a t t h e apparatus is air-tight. If no leakage is observed, t h e reservoir is again raised t o a level slightly below t h a t of t h e liquid in t h e collector, t h e cooling water turned on, and t h e contents of t h e flask are heated t o boiling. As t h e gas collects, t h e reservoir is lowered from time t o time so t h a t atmospheric pressure is maintained. T h e boiling is continued until t h e volume of t h e gas is practically constant, which requires from 15 t o 2 0 minutes. The flame is now removed and t h e flask placed in a water b a t h a n d cooled until t h e temperature is reduced t o t h a t a t which t h e operation was started; t h e reservoir is brought t o t h e same level as t h e liquid in t h e collector a n d t h e volume of t h e gas noted. A correction is made for t h e air displaced b y t h e stopper a t t h e beginning of t h e operation and atmospheric pressure and room temperature are recorded. T h e volume of t h e gas must now be reduced t o normal barometric pressure a n d a temperature of o o C. according t o t h e following well-known formula,

V ( C - h) 760 ( I at) where VI = reduced volume, V = volume actually found, C = barometric pressure, h = tension of water vapor at temperature t , a = 0.00366, and t = room temperature. Taking t h e results of Guye, t h e weight of I cc. of CO, gas at 760 mm. and a t o o C. is 0.0019768 g., Therefore, t h e number of cc. multiplied by 0.0019768 gives t h e weight of carbonic acid gas generated. This result calculated t o IOO g. of material a n d multiplied by 46/44, or 1.04545, will give t h e percentage of formic acid. Trial experiments with known quantities of C . P. sodium formate (recrystallized from alcohol) gave t h e following results:

+

v1 = -

(0.5 Gram Sodium Formate Used in Each Case) CC. OF cog Pressure GENERATED HCOONa Difference TemD. ' C. Mm. Actual Corrected Found Gram 0.4988 -0.0012 163.1 187 25°C. 747 21OC.

26'C. 27'C.

743 743 743

185 188 190

163.7 162.2 163.0

0.5001 0.4955 0,4979

SO.0001 -0.0045 -0.0021

PROCEDURE (?)-In t h e presence of carbonates, bicarbonates, oxalates a n d acetates, t h e following procedure is t o be followed: t a k e sufficient material t h a t t h e formic acid content will range between 2 . 5 a n d 5.0 g. Dissolve in 50 cc. of water, boil a few minutes, make alkaline with N a O H if bicarbonates are present a n d a d d a I O per cent solution of calcium chloride in sufficient quantity t o precipitate all of t h e carbonates a n d oxalates. Allow t h e precipitate t o settle, filter into a 2 5 0 cc. graduated flask, wash thoroughly with boiling water, cool t o room temperature a n d fill t o t h e mark. Take 50 cc. of t h e filtrate and proceed as directed under Procedure ( I ) . Trial experiments with known quantities of C . P. sodium formate mixed with