J u l y , 1915
T H E J O U R N A L O F I N D U S T R I A L A N D EA’GINEERING C H E M I S T R Y
T H E SPECIFIC ABSORPTION OF REAGENTS FOR GAS ANALYSIS
587
For example, t h e specific absorption a t room temperature of a certain solution of alkaline pyrogallol B y R. P. ANDERSON f o r oxygen, obtained b y shaking a mixture of gases Received February 10, 1915 containing 2 1 per cent oxygen with t h e reagent for 3 A knowledge of t h e absorbing power of t h e reagents minutes in t h e Hempel double pipette, was found t o t h a t are commonly used in gas analysis for t h e gases be j . By this is meant t h a t successive samples of a absorbed b y t h e m a n d for t h e conditions under which gas mixture of t h e above composition were analyzed they are ordinarily employed would be of considera- for oxygen by pyrogallol in t h e manner above indible value t o the gas analyst. Definite information cated until t h e reagent was no longer capable of removof this n a t u r e is for the most p a r t lacking. Hempell ing all of t h e oxygen in t h e t i m e allotted. T h e total has proposed t h e t e r m “analytical absorbing power,” a m o u n t of oxygen absorbed u p t o t h a t point was tl suitable expression for this property, b u t his method found t o be j times t h e volume of t h e reagent t h a t of determining i t seems open t o objection. &4ccord- was employed. ing t o his method, a n excess of pure gas was placed in Determinations of specific absorption are being made contact with I cc. of reagent over mercury a n d shaken in this laboratory in connection with investigations with t h e reagent a s long “als noch schnelle Absorpon t h e most desirable concentrations of the various tion erfolgte, so dass in Laufe einer Minute mindereagents a n d t h e most desirable conditions for their stens mehrere Kubikzentimeter \-erschwanden.” T h e use. T h e results on alkaline pyrogallol are given in decrease in volume t h u s obtained was divided b y four the following article. a n d this value termed t h e analytical absorbing CORNELLUNIVERSITY,I T H A C A , N E W YORK power. S o a t t e m p t was made t o determine t h e absolute absorbing power of t h e reagent, nor was t h e REAGENTS FOR USE I N GAS ANALYSIS reagent required t o remove t h e gas completely as i t I. ALKALINE PYROGALLOL must in actual service. T h e rapidity with which a B y R . P. ANDERSOX reagent absorbs a pure gas t h a t is present in excess Received February 10, 1915 is quite different from t h a t with which i t completely HISTORICAL R f i S U M k A N D I N T R O D U C T I O N removes t h e gas from its mixture with other gases. T o Dobereiner’ is attributed the discovery of the I n t h e first case: t h e pressure t h a t is exerted b y the gas is presumably t h a t of t h e atmosphere, while in absorption of oxygen b y an alkaline solution of pyrothe second case, t h e partial pressure of t h e gas becomes gallol. Liebig* was t h e first t o utilize this reaction less a n d less until i t equals zero or is so small a s t o for t h e quantitative removal of oxygen from its mixbe negligible for technical purposes. In general, t u r e with other gases. He employed a solution of as t h e pressure becomes less, t h e rapidity of t h e ab- potassium hydroxide of I . 4 specific gravity and added sorption decreases, t h e last traces of t h e gas being t o i t one-half its volume of a solution of pyrogallol absorbed with difficulty, or not a t all. It is evident, containing I g. t o j-6 cc. of water. Much later therefore. t h a t t h e values t h a t are obtained b y H e m - Weyl a n d Zeitler3 studied t h e effect upon t h e absorbpel’s method have little bearing upon t h e efficiency of ing power of the reagent of varying t h e concentrathe various reagents for quantitative absorptions, tion of t h e alkali. I n their work. a definite volume a n d t h a t t h e y afford t h e analyst no certainty t h a t his of air was aspirated through a U-tube t h a t contained reagents m a y be used until t h e y have absorbed t h e broken glass a n d a measured volume of the alkaline q u a n t i t y of gas indicated b y t h e analytical absorbing solution of pyrogallol. T h e a m o u n t of oxygen t h a t remained in t h e air after this t r e a t m e n t was deterpower now assigned t o t h e m without possibility of mined by titration with a solution of sodium hypothe incomplete removal of t h e constituents. T h e determination of t h e volume of a gas which a sulfite, using indigo as t h e indicator. Three solucertain reagent will absorb up t o the point a t which tions of potassium hydroxide, of specific gravities the gas is n o t completely removed from its mixture ~ . o z j ,1 . 0 j o and I . joo, were employed, and t h e rewith other gases is needed b y gas analysts. I t is agent w a s prepared b y adding o . 2 j g. of pyrogallol here proposed t o express this value in cc. of gas per t o I O cc. of one of these solutions. T h e oxygen t h a t cc. of reagent a n d t e r m i t t h e s p e c i f i c a b s o r b i n g p o w e r , was not absorbed under t h e arbitrary conditions of or t h e speci$c absorptiofL, of the reagent. T h e state- the experiments amounted t o I . j 6 , 0 . 8 8 a n d 2 . 9 ment of t h e specific absorption of a reagent for a gas per cent for t h e solutions in which t h e specific gravity must contain t h e name of t h e a p p a r a t u s t h a t is em- of t h e alkali was ~ . o z j ,I . Cjo a n d I . 5 0 0 , respecployed, t h e method of i t s manipulation, a n d t h e tively. From these d a t a t h e authors conclude t h a t temperature of t h e reagent, since these conditions af- t h e solution of potassium hydroxide of specific gravity fect t h e results. Also t h e volume of gas t h a t is ab- I , o j o gives t h e reagent with the highest absorbing sorbed a t a time m a y have a n effect upon t h e specific power a n d t h e y s t a t e also t h a t the pyrogallol probably absorption and for t h i s reason t h e mixture t o be em- is quickly destroyed b y t h e solution of potassium ployed in t h e determination of this value should con- hydroxide of specific gravity I . j . Wehl a n d Goth‘ tain a n amount of the gas t o be removed such a s would performed similar experiments substituting sodium 1 Gilbert’s A n n . , 74 (1823), 410. most frequently be met with, a n d this amount should 2 Liebig’s Ann.. 77 (1851). 107. tie given in stating t h e specific absorption of a reagent. 3
1
“Gasanalytische Methoden.” 4th E d . , p. 128.
4
I b i d . , 206 (ISSO), 2 5 5 . B e y . . 14 (1881). 2659.
588
T H E JOC'RN-4L O F I S D L 7 S T R I d L - A S D E - V G I N E E R I N G C H E M I S T R Y
hydroxide a n d sodium carbonate for t h e potassium hydroxide. Soon after Liebig's experiments, it was found' t h a t carbon monoxide was formed under certain conditions on t h e oxidation of alkaline pyrogallol. Lewest states t h a t a solution prepared b y adding I O g. of pyrogallol t o I 50 cc. of a 2 0 per cent solution of sodium hydroxide must not be employed for t h e absorption of oxygen more t h a n four or five times on account of t h e formation of carbon monoxide. Clowes3 studied t h e behavior of alkaline pyrogallol with especial reference t o t h e formation of carbon monoxide. H e found t h a t although a reagent containing I O g. of pyrogallol a n d 24 g. of potassium hydroxide t o I O O cc. of solution c a n be used i n determining oxygen when i t does n o t exceed 28 per cent of t h e t o t a l volume, carbon monoxide is formed on t h e analysis of richer mixtures. H e recommends t h a t t h e a m o u n t of potassium hydroxide be increased t o 1 2 0 g. whenever gases t h a t contain more t h a n 28 per cent of oxygen are t o be analyzed. An a t t e m p t t o determine oxygen in a gas mixture containing 90 per cent of i t , with t h e reagent containing only 2.4 g . of potassium hydroxide t o IOO cc. of solution, resulted in t h e formation of 6 per cent carbon monoxide, while no more t h a n a trace was formed when t h e stronger solution of alkali was employed. Berthelot' found, from a s t u d y of t h e various factors t h a t might affect t h e f o r m a t i o n of carbon monoxide, t h a t alkali i n large excess a n d pyrogallol sufficient t o absorb four or five times t h e a m o u n t of oxygen present are necessary i n order t o prevent t h e formation of more t h a n a negligible a m o u n t of t h i s undesirable product of t h e reaction. He also investigated t h e behavior of t h e reagent when t h e potassium hydroxide was replaced b y sodium hydroxide, barium hydroxide a n d a m monium hydroxide. T h e t o t a l a m o u n t of oxygen t h a t was absorbed b y t h e solution containing sodium hydroxide was f o u n d t o be practically t h e same a s with t h e use of potassium hydroxide, while t h e solution i n which barium hydroxide was used gave a smaller absorption, a n d t h e one i n which ammonium hydroxide was used gave a larger absorption. T h e effect of these substitutions upon t h e formation of carbon monoxide was small. Benedictj found, from precise determinations of t h e oxygen content of t h e atmosphere, t h a t t h e results t h a t were obtained depended upon t h e concent r a t i o n of t h e alkali t h a t was employed. T h e average of 14 determinations made with Haldane's6 solution containing I g. of pyrogallol t o I O cc. of potassium hydroxide, I . j j specific gravity. was 2 0 . 9 j 6 ; t h e average of similar determinations made with a slightly modified Haldane reagent which contained somewhat less alkali t h a n t h e original. was 2 0 . 9 3 8 ;
YO^. 7 . XO. j
a n d other solutions containing less alkali gave still lower results. Benedict suggests t h a t t h e discrepancy in t h e results obtained from t h e various solutions m a y be due t o t h e formation of carbon monoxide. Haldane states t h a t t h e solution prepared according t o his directions yields no carbon monoxide on t h e a b sorption of oxygen. d n inspection of t h e proportions of pyrogallol, potassium hydroxide, a n d water t h a t have been recommended b y various gas analysts for t h e preparation of alkaline pyrogallol shows t h a t there is little uniformity as regards either t h e composition of t h e reagents or t h e t e r m s i n which their descriptions are couched. I n t h e accompanying list, for purposes of comparison, a n a t t e m p t h a s been made t o standardize a s far a s possible t h e s t a t e m e n t of t h e composition of t h e reagent b y expressing with sufficient accuracy t h e a m o u n t of pyrogallol a n d of potassium hydroxide in I O O cc. of t h e solution: 4UTKOR
Grams per 100 c c . of solution of PYROGALLOL KOH 3i* 24 I20 22" 84 2.5 70 43
Liebigl.. . . . . . . . . . . . . . . . . . . . 6 t o 7 Clowesl . . . . . . . . . . . . . . . . . . . a 10 b 10 Winkler? . . . . . . . . . . . . . . . . . . . 5 Hempel3 . . . . . . . . . . . . . . . . . . 1,5 Berthelot'. . . . . . . . . . . . . . . . 1.5 8 Franzenj . . . . . . . . . . . . . . . . . . Gill6 . . . . . . . . . . . . . . . . . . . . . . . Q 4.9 b 4.9 Haldane'. . . . . . . . . . . . . . . . 9.4 Benedict'. . . . . . . . . . . . . . . . . . . 9,4
$:* 66'
~~
1 See Calvert, C o m p t . vend.. 5 1 (1863). 8 7 3 : Cloez, I b i d . , p. 8 7 5 ; Boussingault, p. 885; a n d Poleck, Z. a n d . Chem., 8 (18691, 451. 2 J. Soc. C h e m . Ind., 10 ( 1 8 9 1 ) . 40i. 3 C h e m . News, I1 (1895). 2 8 8 ; J . S o l . C h e m . I n d . , 15 (1896), li0; C h e m . News, 14 (1896). 199. 4 Ann. c h i m . p h y s . , [ i ] 16 (1898). 294, C o m p f . r e n d . , 146 (1898), 1066.
1459. 5
5
" T h e Composition of the Atmosphere." p. 113. "Methods of Air Analysis." 1912 D d . , p . I ? .
1 LOC.
Cif.
"Lehrbuch der technischen Gasanalyse," 1901 Ed., p . 81. "Methods of Gas Analysis," 1902 E d . , p. 149. I n t h e fourth edition of this work (1913). Hempel recommends practically t h e same proportions a s are given b y Benedict. "Trait6 Pratique de L'Analyse des Gas," 1906 E d . , p. 185. 5 "Gasanalytische iibungen," 1907 Ed., p. 4. 6 "Gas and Fuel Analysis," 1912 Ed., p. 5 4 . 7 "The Composition of t h e Atmosphere." 1913 E d . , p. 80. ?
Four of t h e solutions were prepared b y using potassium hydroxide of a definite specific gravity. T h e a m o u n t of wlater-.free potassium hydroxide is given in these cases a n d is marked b y a n asterisk t o distinguish it from t h e amounts used i n t h e other solutions where ordinary potassium hydroxide of unknown water content was employed. T h e water content of potassium hydroxide t h a t is sold in stick form varies between rather wide limits a n d t h e a m o u n t of alkali in Haldane's solution would ordinarily be found t o exceed t h a t recommended by a n y other authority with t h e possible exception of Clowes' stronger solution. I n view of t h e conflicting results t h a t have been obtained b y earlier investigators of alkaline pyrogallol, a n d because of t h e a p p a r e n t lack of a n y effort t o determine what constitutes t h e most desirable reagent from a consideration of t h e various determining factors, i t was decided t o a t t e m p t a systematic investigation of t h e reagent, ii s t u d y of t h e effects upon t h e specific absorption' of t h e reagent of variations i n t h e a m o u n t s of potassium hydroxide a n d pyrogallol was made t h e starting point of t h e experimental work a n d . i n order t o eliminate a s much unnecessary work as possible. a n a p p a r a t u s was constructed in which t h e specific absorption of various solutions of alkaline pyrogallol could be easily determined. 1
See preceding article, TIIISJ O U R K A L , I (19151, 587.
S PE C I A I. d PP .-IR .-I T 1 - S F 0 K E S P E R I 11E. 1 - TA L If'0 R K
artificial mixtures of oxygen a n d nitrogen were shaken with the alkaline pyrogallol over mercury and t h e readings were made under constant 1-olume (approximately) i. e . , t h e amount of oxygen nbsorlied I\-as determined by t h e change in t h e pressure t h a t was exerted by the gas a s indicated b y t h e change in n manometer rending, during thc absorption. -4 nen- sample was not t a k e n after each determination. nearly t h e same result being obtained l-iy t h e addition of oxygen t o t a k e t h e placc of t h a t which had been previously absorbed. T h e contact between t h c gas mixture containing oxygen a n d the alkaline pyrogallol as effected in the pipettc P which connects b y m e a n s of n long, enameled ruhber t u b e with t h e leveling bulb L . T h e pipette is closed a t t h e top by a tn-o-way stopcock, -A, by means of which connection m a y be nrade either
In the a p p a r e t u s in Fig.
I.
t u s might be rocked back a n d forth with t h e rod a s a n axis b y means of a hot-air engine and t h e proper connections. glass j a r containing v a t e r w a s brought up around t h e apparatus' until K : P a n d E were ~017ered, in order t o maintain a uniform temperature i n t h e separate compartments where gas w a s confined. T h e enameled rubber tubing passed from the pipette over t h e edge of t h e j a r t o t h e leveling bulb n-hich was supported b y a separate stand. T h e a p p a r a t u s lvas prepared for use by pouring mercury, sufficient in amount t o fill t h e pipette P , into t h e leveling bulb L , drawing water into t h e manometer -If through S by lox-ering L n-ith .,I and F in t h e proper positions. and filling the reservoir K through 2 and Ti- 11-ith oxygen under pressure. T h e air in R was previously displaced b y oxygen b y opening I' and allowing the oxygen t o bubble through t h e water in t h e cell. -4fter these operations were performed, t h e manipulation of the apparatus w a s carried out in the following fashion: 1 I h S I P I - L A T I O K OF T H E A P P A R A T r S
TS'itli F closed, mercury from t h e pipettc was driven u p t o H b y raising t h e leveling bulb. Connection was made a t this point x i t h a capillary tube, also filled with mercury. t o a Hempel burette containing nitrogen from a phosphorus pipette. a n d 2 0 cc. of this gas n-as drawn into t h e pipette. following it w i t h
with t h e burette B or with one a r m of manometer -If through stopcock F , a n d with t h e oxygen resen-oir R through stopcocks K a n d 2. Reversal of the position of K permits connection of t h e pipette with t h e atmosphere a t H . T h e other a r m of t h e manometer bends downward a n d terminates in t h e enlargement' E a n d stopcock T. T o support t h e a p p a r a t u s , a tall, narron-, rectangular iron frame was constructed of iron rods a n d clamps a n d t h e upper a n d lower ends of the pipette P were fastened t o horizontal cross pieces near t h e bottom of t h e frame. I n t u r n , the frame was suspended a t a point about midway between t h e top a n d bottom (corresponding approximately t o K on t h e apparatus) t o a horizontal iron rod in such fashion t h a t t h e appara1 This enlargement was made of such size t h a t a change of pressure atmosphere inside t h e pipette would produce as large ,a movement of of the water in the manometer a s was permissible.
mercury from t h e burette over t o K ,which was thercupon closed. The proper amounts of solutions of pyrogallol a n d potassium hydroxide were then introduced separately into t h e pipette through the burette B . This was accomplished without danger of a d mission of air providing the mercury in t h e leveling bulb was a t a slightly lower level t h a n t h a t in thc pipette a n d care w a s t a k e n in t h e manipulation of A . Connection with the manometer was next made b y turning 4: a n d P , and t h e leveling bulb was t h e n lowered until t h e water stood a t some definite point on t h e scale near t h e upper end of t h e left-hand a r m of the manometer. S was thereupon closed. At this point t h e engine t h a t rocks t h e pipette back a n d f o r t h was started in order t o obtain a thorough mixt u r e of t h e solutions of pyrogallol a n d potassium hydroxide in the pipette a n d t o facilitate t h e absorption of t h e small a m o u n t of oxygen t h a t was usually present in t h e space between A a n d t h e water in t h e manometer. This absorption of oxygen decreased t h e pressure in the pipette a n d necessitated t h e readjustment of t h e level of t h e water in t h e manometer. Oxygen was next admitted from t h e reservoir R into t h e pipette until t h e manometer reading indicated t h a t t h e desired quantity* had been added. Shaking was started a t this point a n d continued for j minutes. a t t h e end of which time t h e motor was stopped, t h e manometer reading t a k e n , oxygen added t o t h e same manometer reading as before, a n d t h e shaking 1 To make t h e apparatus more compact, the capillary connections u-ere bent s o t h a t R and E were situated behind 1'. 2 This quantity was usually 5 cc., thus giving an artificial mixture containing 2 0 per cent of oxygen. T h e manometer reading t h a t cor responded t o the admission of 5 cc. of oxygen was obtained b y running in 5 cc. of nitrogen without changing the level of the mercury in the pipette.
590
T H E JOURLVAL O F I - V D r S T R I A L AAVD E N G I N E E R I S G C H E M I S T R Y
continued. This was became exhausted.
repeated
until
the
reagent
EFFECT O F AMOUNT OF POTASSIUM HYDROXIDE ON T H E SPECIFIC ABSORPTION O F THE REAGEKT
Five solutions of potassium hydroxide were employed. These consisted of a solution (of specific gravity I . jj) containing a b o u t I . 5 p a r t s of potassium hydroxide in stick form t o I p a r t of water, a n d four other solutions t h a t were prepared from t h e first in t h e proportions of 4 volumes of alkali t o I volume of water, 3 volumes t o 2 volumes, 2 volumes t o 3 volumes, a n d I volume t o 3 volumes, respectively. T h e solution of pyrogallol was prepared b y dissolving I p a r t of pyrogallol i n I . j p a r t s of water: 0 . 4 3 cc. of this solution (containing 0 . 2 g. of pyrogallol) was employed i n each case a n d t o it was added j cc. of one of t h e solutions of alkali. T h e saturation of each reagent was carried o u t as previously outlined a n d t h e results are shown in Fig. 2 i n which manometer readings t h a t were obtained after t h e absorption of oxygen from i t s mixture with nitrogen are plotted against t h e number of samples from which t h e oxygen h a d been absorbed. Curve I was obtained from t h e
solution containing t h e most alkali, Curve I 1 from t h e next weaker, etc. Three or more determinations were made on each of t h e five solutions. Owing t o difficulties i n admitting t o t h e pipette exactly t h e proper a m o u n t of reagent a n d in maintaining a uniform r a t e of shaking, duplicate manometer readings did not always agree as closely a s seemed desirable, b u t sufficient determinations were m a d e t o establish with t h e necessary accuracy t h e relative positions of t h e curves corresponding t o t h e different solutions with t h e exception of t h e l a t t e r portion of Curves I V a n d V. T h e differe n t determinations failed t o show definitely t h e positions of t h e breaks i n these t w o curves. T h e break i n Curve V falls a t t h e left of t h e break i n Curve I V a n d b o t h lie between t h e breaks in Curves I a n d 111. I t was n o t deemed of sufficient importance t o locate t h e m more accurately. T h e downward slope of t h e first portion of each of t h e curves is due t o t h e presence of a b o u t one per cent of nitrogen i n t h e oxygen. This caused a gradual decrease i n t h e a m o u n t of oxygen t h a t was added a t a time, a n d a corresponding drop i n t h e manometer reading. This fact proved t o b e of little consequence i n t h e s t u d y of t h e various solutions since i t did not obscure t h e sudden break which was obtained when
v01, j
,
KO.j
t h e reagent became nearly exhausted, a n d consequently no effort was made t o free t h e gas from t h e impurity. I t will be noted t h a t t h e largest manometer readings were obtained with t h e solution t h a t contained t h e most potassium hydroxide; i n other words, t h e greater t h e a m o u n t of alkali, t h e larger t h e a p p a r e n t oxygen content of t h e gas sample. T o determine whether t h e a m o u n t of alkali present i n Solution I was sufficient f o r t h e removal of all of t h e oxygen from a mixture of t h i s nature, similar determinations were made with a solution containing still more potassium h y droxide' a n d t h e same a m o u n t of pyrogallol. A s far as could be determined from t h e a p p a r a t u s t h a t was employed, t h e curve obtained from t h i s solution occupied practically t h e same position as Curve I , t h u s showing t h a t practically complete absorption is obtained with Solution I . T h e s t a t e m e n t of Weyl a n d Zeitler* regarding t h e effect o n pyrogallol of a concentrated solution of potassium hydroxide does not appear t o be substantiated b y these results. F u r t h e r , t h e results which t h e y give in support of t h e use of a solution containing alkali of specific gravity I . o j show only t h a t more rapid absorption is obtained with this solution, not t h a t t h e absorption is complete or t h a t t h e solution has a higher absorbing power t h a n one containing more alkali. T h e break in Curve I occurs a t a b o u t t h e ninth absorption, in I 1 a t t h e eleventh, a n d i n I11 a t t h e thirteenth absorption. T h e reason for this apparent increase in specific absorption with decreasing cont e n t of alkali is no doubt due i n p a r t t o t h e failure of t h e solutions showing this increase t o absorb all of t h e oxygen from t h e mixtures with which t h e y were t r e a t e d as is shown b y lower manometer readings. With still lower concentrations of alkali, a s in Solutions I V a n d V, t h e break i n t h e curve occurs before t h e break i n 111, indicating t h a t in this case t h e break is due t o insufficient alkali r a t h e r t h a n t o exhaustion of t h e pyrogallol. T h e remainders of Curves I V and V are n o t shown o n account of t h e uncertainty of t h e positions of t h e breaks, b u t these portions are much more nearly vertical t h a n is t h e case with t h e other curves. EFFECT OF AMOUNT OF PYROGALLOL ON T H E SPECIFIC ABSORPTIOK OF T H E REAGENT
Five solutions of pyrogallol were employed. These consisted of a solution containing 0 . 6 2 5 g. of pyrogallol per cc., which was obtained b y dissolving pyrogallol i n a n equal weight of water, a n d four other solutions t h a t were prepared from t h e first, containing 0 .j, 0 . 3 7 j, 0 .2 j, a n d 0 .I 2 5 g. of pyrogallol per cc. I t was desired t o keep t h e concentration of potassium hydroxide i n t h i s series of experiments as high as i n Solution i of t h e previous series3 a n d , a t t h e same time, introduce as much a s possible of t h e solution of pyrogallol i n order t o obtain a wide variation of t h i s constituent. Accordingly, a solution of potas1 T h e a m o u n t of potassium hydroxide was sufficient t o place the solution in the zero position in the sequence 0, I, 11. 111, I V , V. 2 LOG. cit. 5 cc. of a solution of KOH, sp. gr. 1 . 5 5 (4 g. J Solution I contained KOH), t o a b o u t 5.4 cc. alkaline pyrogallol.
J L ~ ~1915 Y,
T H E JOl-RiY'.-IL O F I S D C ' S T R I A L -4 N D E S G I N E ERI N G C H E M I S 1 R Y
sium hydroxide was prepared of such strength t h a t 4 . 4 . j cc. contained as much potassium hydroside as j cc. of t h e alkali in Solution I , and was employed in this series of experiments in t h e proportion of 4 . 4 5 cc. of alkali t o I cc. of one of t h e solutions of pyrogallol. T h e curves t h a t \yere obtained from t h e treatment of these various solutions in t h e special apparat u s are shown in Fig. 3, Solution I containing t h e most pyrogallol a n d Solution Y t h e least. F r o m these curves it is evident t h a t , within t h e limitsof this series a n d with the a m o u n t of potassium hydroxide
t h a t was employed, t h e specific absorption increases with the a m o u n t of pyrogallol which t h e reagent contains. THE
MAXIMIU31
SPECIFIC
ABSORPTIOS
OF
ALKALIKE
PYROGALLOL
T h e two series of experiments t h a t have been described show t h a t a high concentration of potassium hydroxide in alkaline pyrogallol is necessary for t h e quantitative absorption of oxygen in gas-analytical work a n d a high concentration of pyrogallol is necessary t o obtain a high specific absorption. Any effort t o increase t h e specific absorption of t h e reagent by increasing t h e concentration of pyrogallol cannot be carried too far without danger of incomplete absorption as a result of t h e unavoidable decrease in the concentration of potassium hydroxide. I n this connection, therefore, water, t h e third constituent of t h e reagent, is obviously undesirable in a n a m o u n t larger t h a n is necessary t o prevent the reagent from crystallizing a t ordinary temperatures. To determine what proportions of pyrogallol, potassium hydroxide, a n d water give t h e maximum specific absorption for this reagent, a third series of experiments was performed. One set of solutions was prepared b y dissolving L O , 2 0 , 40 a n d j o grams of pyrogallol in r o o cc. portions of a solution of potass i u m hydroxide, specific gravity I . j j , a n d a similar s e t was prepared using alkali of specific gravity I . 6 c . the necessary precautions being t a k e n t o prevent their deterioration through contact with t h e oxygen of the air. A . definite a m o u n t of reagent (about 2 j cc.) was transferred t o a Hempel pipette for use with mercury without allowing oxygen t o come into cont a c t with i t a n d was treated a t room temperature' 1 Hcnipel ha5 shown t h a t t h e absorption of oxygen b y alkaline pyrogallol i, much less rapid a t temperatures below 15O C . t h a n a t room temperature. I t was found t h a t a temperature change from Z O O t o 2 4 O C. caused little, i f a n y , variation in t h e specific absorption of t h e reagent a n d no analyses were made if t h e room temperature fell outside these limits. Temperatures helow 2 1 " or above 23' were unusual.
.i9 1
with I O O cc. samples of air1 until t h e absorption of oxygen was no longer sufficiently complete for technical purposes* a t t h e end of three minutes. T h e absorption was effected b y gently rocking t h e pipette a n d s t a n d backward a n d forward a t t h e rate of t w o or three times a second, using t h e front edge of t h e base of t h e pipette s t a n d as a n axis. T h e total volume of oxygen t h a t had been absorbed, divided b y t h e volume of t h e solution, gave the specific absorption of t h e particular reagent for the conditions under which t h e determinations were made. The values of t h e specific absorptions for t h e various amounts of pyrogallol in I O O cc of alkali are shown in Fig. 4. Curve I was obtained from t h e solutions containing t h e alkali of specific gravity I . 60 a n d Curve I1 from those containing alkali of specific gravity 1 55. I n t h e case of Solution I , the specific absorption is approximately proportional t o t h e amounts of pyrogallol t o I O O cc. of alkali, a n d the lack of strict proportionality m a y be imputed t o two causes. First. owing t o t h e increase in volume t h a t results when pyrogallol dissolves in a solution of potassium hydroxide, t h e concentration of pyrogallol in t h e various solutions is not proportional t o t h e a m o u n t of pyrogallol t h a t was employed, e. g.. t h e solution t h a t was prepared b y dissolving 20 g. of pyrogallol in I O O cc. of potassium hydroxide does not contain twice as much pyrogallol as the one prepared b y dissolving I O g. of pyrogallol in the same amount of alkali. and consequcntly t h e value t h a t is obtained for t h e specific absorption of t h e stronger solution should he less t h a n twice t h a t of t h e weaker. Secondly, the concentration of t h e alkali is gradually lessened b y the forrna-
tion of carbon dioxide a n d m-ater3 during t h e oxidation of t h e pyrogallol, t h u s causing a n earlier lack of complete absorption t h a n would be expected from t h e amount of pyrogallol in the solution. T h e effect 1 \Vhen it was known t h a t t h e reagent was still capable of removing the oxygen from many s a m p M o f air before becoming exhausted, pure oxygen was added from a burette t o lighten t h e routine work of the experi. ment. Care was taken t h a t the heat of the reaction between the oxygen and t h e alkaline pyrogallol did not cause more t h a n a slight rise in t h e temperature of t h e reagent. 2 It was assumed t h a t the absorption was complete as long a s t h e a p parent oxygen content of t h e sample amounted t o 20.8 per cent or more. 3 See Berthelot, A n n . chim. p h y s . . [:I 16 (18981, Z Y 4 .
592
T H E JOL'RNAL OF I N D U S T R I A L A N D ENGINEERING CHEMISTRY
of this gradual diminution in t h e amount of potassium hydroxide is especially noticeable in Curve 11, where t h e specific absorption of a solution containing 40 g. of pyrogallol t o I O O cc. of alkali, specific gravity I . j j , is b u t little more t h a n t h a t obtained from a solution containing one-half of t h a t amount of pyrogallol. The highest specific absorption t h a t was obtained from t h e solutions of these two series is about 7j for t h e conditions t h a t have already been stated. This probably does not actually represent the maximum specific absorption t h a t might be obtained, b u t is no doubt a close approximation t o it. T h e solutions t h a t contained j o g. of pyrogallol t o IOO cc. of alkali failed t o give a n absorption sufficiently complete for technical use, even when new, and t h u s their specific absorption equals zero. This was no doubt due t o t h e increase in volume which accompanied t h e solution of t h e pyrogallol a n d lowereg the concentration of alkali beyond t h e lower limit for t h e production of complete absorption. It might be possible t o increase t h e specific absorption b y using a higher concentration of pyrogallol in a solution of potassium hydroxide of higher specific gravity or b y preparing t h e solution b y adding a minimum of water t o a mixt u r e of solid pyrogallol a n d solid potassium hydroxide. b u t such reagents would certainly never be actually employed on account of high viscosity and a tendency t o foam on shaking, a n d consequently no a t t e m p t was made t o determine more accurately t h e maximum specific absorption. EFFECT O F TIJIE OF SHAKING UPON T H E SPECIFIC ABSORPTION O F T H E REAGEliT
T h e question as t o t h e effect of t h e time of shaking upon t h e specific absorption of alkaline pyrogallol
next presented itself for consideration. Reagents were prepared containing I O , I j and zo grams of pyrogallol t o 1 0 0 cc. of a solution of potassium hydroxide of specific gravity I . jj, and t h e specific absorption of each was determined in the manner described under the previous heading except for t h e time of shaking. This was made one minute for each reagent a t the s t a r t , lengthened t o t w o minutes when t h e absorption of oxygen w a s n o longer practically complete for t h e shorter interval, and finally increased t o three minutes for similar lack of complete absorption for t h e txo-minute period. I n this fashion, three values of the specific absorption corresponding t o t h e three periods of shaking were obtained for each solution. I n Fig. j these values are plotted against the time of shaking, T h e numbers below the curves
Vol. 7 , NO. 7
represent t h e amounts of pyrogallol t h a t were used in t h e preparation of t h e various solutions. I n t h e case of t h e reagent t h a t contains the least amount of pyrogallol, t h e specific absorption for I-minute shaking is nearly as large a s t h a t for 3-minute shaking, t h e difference amounting t o only 3 units. The specific absorption for I-minute shaking does'not increase as r a i i d l y as t h e value for 3-minute shaking when t h e amount of pyrogallol is made larger, as shown b y t h e differences in t h e shapes of t h e curves f o r I O , I j a n d 2 0 grams of pyrogallol t o I O O cc. of a solution of potassium hydroxide. This behavior of t h e solutions indicates a less rapid absorption as t h e amount of pyrogallol is increased (at least for t h e condition of the reagent corresponding t o incomplete absorption a t t h e end of I-minute shaking) and is probably due t o the smaller excess of alkali. THE
BIOST
DESIRABLE
REAGEST
FOR
TECHKICAL
USE
AS-D I T S S P E C I F I C A B S O R P T I O N I N T H E H E H P E L
PIPETTE
FOR
USE
WITH MERCURY
I n ascertaining t h e desirability of a reagent for technical use, t h e ease of preparation, t h e convenience of manipulation, and t h e specific absorption must each be taken into consideration. The reagents cohtaining a solution of potassium hydroxide of 1 . 6 0 specific gravity are less simple t o prepare and less convenient t o manipulate t h a n those containing alkali of I . j j specific gravity. T h e greater difficulty of preparation is caused b y the fact 'chat a solution of potassium hydroxide of I .60 sp. gr. crystallizes a t ordinary temperatures ( z o o C.) and consequently must be heated for t h e preparation of t h e reagent.' This is not true of t h e solution of I . j5 specific gravity. T h e greater difficulty of manipulation is caused b y t h e high viscosity of t h e solution t h a t contains t h e greater amount of alkali. I n t h e case of t h e reagents prepared from alkali of I . 60 specific gravity, t h e viscosity causes a comparatively thick film of t h e reagent to remain on t h e walls of the capillary tubing when a sample of gas is passed into t h e pipette, t h e film being of such thickness as t o render opaque t h e bore of t h e capillary, even with only moderately large amounts of pyrogallol in t h e reagent. This opaque film makes i t difficult t o ascertain t h e position of t h e reagent in t h e capillary when t h e sample is drawn back into t h e burette. T h e reagents t h a t contain a high concentration of pyrogallol spow a greater tendency t o foam t h a n those of low concentration. This tendency first becomes noticeable.in regard t o its effect upon t h e ease of manipulation in h e case of those solutions t h a t contain zo grams of pyrogalioi t o IOO cc. of alkali a n d becomes decidedly objectionable in solutions containing twice as much pyrogallol. T h u s it i,s seen t h a t high concentrations of potassium hydroxide and pyrogallol, t h e important factors in obtaining a reagent with a high specific absorption, are responsible for undesirable features such as diffi1 T h e reagent itself is stable in liquid form a t ordinary temperatures on account of the increase in volume t h a t accompanies the solution of pyrogallol and also on account of the replacement of a p a r t of the potassium hydroxide by water in the formation of potassium pyrogallate.
July,
1915
T H E J O C R - V A L O F I S D C S T R I d L A.VD E S G I , Y E E R I Y G C H E M I S T R Y
593
culty of preparation a n d manipulation. T o eliminate of preparing t h e reagent a n d in transferring it t o a these undesirable features necessitates limiting t h e pipette. specific gravity of t h e solution of potassium hydroxide When t h e specific absorption 3 1 has been reached, t o I . j j a n d t h e a m o u n t of pyrogallol t o less t h a n 2 0 g. each molecule of pyrogallol in t h e reagent h a s t h e n t o I O O cc. of t h e alkali. This involves a sacrifice in t a k e n u p about 2 . 3 atoms of oxygen. Berthelot’obtains specific absorption of about 4 0 units (see Fig. 4), 3 atoms of oxygen for each molecule of pyrogallol b u t i t is felt t h a t t h e limits t h a t have been mentioned for complete saturation of t h e reagent. should not be exceeded in choosing t h e most desirable Of t h e various solutions t h a t have previously been reagent for technical use. employed for t h e absorption of oxygen (see IntroIt was found t h a t foaming, such as was noticed in duction), t h e one described b y Haldane most closely t h e solution containing 20 grams of pyrogallol t o IOO resembles t h e one t h a t has just been proposed. The C C . of alkali of specific gravity I . j j?during t h e analysis specific absorption of a reagent prepared according of samples of air, disappeared almost entirely when t o Haldane’s directions is gil-en 011 the lower curl-e only I j grams of pyrogallol were employed. Further, of Fig. j. t h e specific absorption of t h e latter reagent for I-minute shaking is nearly as large as t h a t of t h e more concen- T H E S P E C I F I C A B S O R P T I O X O P T H E PROPOSE11 R E A G E S T I S THE H E M P E L D O U B L E P I P E T T E F O R L I Q V I D t r a t e d solution (see Fig. j ) . I t t h u s appears t h a t the R E ..1G E S T S reagent prepared b y dissoll4ng I j grams of pyrogallol Thirty grams of pyrogallol was n-eighed into an in I O O cc. of a solution of potassium hydroxide of I . j j specific gravity has practically t h e highest specific Erlenmeyer flask of 3 0 0 cc. capacity. and was dissol\.ecl absorption t h a t can be obtained in a solution satisfac- in 2 0 0 cc. of a solution of potassium hydroxide of I . j j t o r y in other respects and t h a t it is therefore t h e most specific gravity: 1 8 j cc. of this solution was then desirable for technical use.l The volume of t h e solu- transferred t o a Hempel double pipette? for liquid tion obtained above is approximately I I O cc. and reagents. Unnecessary deterioration of the reagent n-as avoided b y stoppering t h e flask while t h e process there are t h e n 1 3 . 6 g. of pyrogallol and ; I . j g. of potassium hydroxide t o I O O cc. of solution. The of solution was taking place and b y transferring it specific absorption of this reagent is about 2 7 n-hen t o the pipette as expeditiously as possible. eniployed a t room temperature in z j - c c . portions in During t h e process of exhausting thh reagent and a Hempel pipette. over niercury and treated with I O O clctcrmining the specific absorption of it for various cc. samples of air x i t h I-minute shaking. T h e same periods of shaking with I O O cc. samples of air, the reagent m a y tie used still further t o effect t h e ahsorp: analysis of air n-as alternated with t h e admission of tion of oxygen from air if t h e analyst prefers a longer nearly pure oxJ-gen from a calibrated gas holder. In time of shaking t o discarding the old and preparing other v-ords, t h e greater portion of t h e oxygen necesa nevi reagent. T h e specific absorption for 2-minute sary t o exhaust t h e reagent was slowly admitted t o shaking is about 30. and t h a t for 3-minute shaking t h e reagent from a gas holder in orcli-r t o s a v e time a n t 1 about 31. N o material increase in the specific ab- labor, an occasional analysis of air heing performed t o sorption of this solution can be effected b y incrcasing make sure t h a t t h e reagent n-as still gix-ing complt9tc the time of shaking beyond three minutes. Experi- absorption for a certain period of shaking. T h e n ments have shown t h a t t h e specific absorption t h a t is ob- t h e reagent r a s nearly exhausted (judging from the tained with a reagent prepared four months before results obtained in the Hempel pipettc for use ,with the time when i t vias found t o be no longer able t o mercury) the successive additions of oxygen were give complete absorption is as great as t h a t obtained made small t o avoid more t h a n a negligible error in with one exhausted shortly after i t was preparetl. I t t h e determinations of specific absorption. Yntiue t h u s appears t h a t t h e proposed reagent is stable oi-er heating of the reagent was prevented b y playing a a n indefinite period of time, although conclusive re- blast of air on t h e outside of t h e pipette and by consults have not yet been obtained on this point. trolling the rate a t n-hich the oxygen r a s T h e reagent t h a t has been proposed has the disad The specific absorption of the reagent a t room temvantage over using separate solutions of pyrogallol perature for I O O cc. samples of air was found t o be a n d potassium hydroxide t h a t it must be prepared 2 j ,I for I-minute shaking. 2 9 . 4 for 2-minute shaking outside the pipette. t h u s allowing some oxidation, a n d 3 I , 3 for 3-minute shaking. (due t o contact with t h e oxygen of t h e air, unless special Uniformity of procedure can rearlilp be obtained pr e cautions are introduce d TT h i ch \TO ul d complicate with different pipettes except for the amount of reits preparation. Howel-er, t h e specific absorption agent t h a t is left in t h e first bulb of t h e pipette when t h a t may be obtained \\-hen pyrogallol is dissolved t h e gas sample is introduced. This ill vary with tlirectly in a solution of potassium hydroxide of I . j j t h e size of t h e hull) and there is frequently a considspecific gravity is greater t h a n is possible x h e n separate erable variation in this particular in different pipettes. solutions of the two constituents are employed. a n d It seems likely. howe\-er, t h a t a n y change in t h e specific this fact more t h a n ofisets t h e deterioration t h a t 1 L O G . Lit. cannot readily be avoided in using the former meth’od . 1 modified form $vas employed. See T H I S J O ~ H B . A L , 6 (19141, 237. 1 A t least, in the particular absorbing device in question. Other absorbing devices, xvith widely different methods of manipulation, might perhaps operate more successfully with a solution prepared in R different fashion.
3 T h e effect of the rapid admission of oxygen without proper cooling is taken up in a later section. I n each case after the admission of oxygen, the reagent v a s allowed to cool t o room temperature before t h e analysis of a sample of air was a t t e m p t e d .
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
594
absorption of t h e reagent due t o this cause would be so small as t o be negligible. THE
EFFECT
OF
HIGH
PERCENTAGES
OF
OXYGEX
ON
T H E SPECIFIC ABSORPTIOK
T h e specific absorption of t h e proposed solution of alkaline pyrogallol was determined for t h e Hempel pipette for use with mercury a n d for t h e . Hempel double pipette for liquid reagents, using mixtures containing a b o u t 90 per cent oxygen. I n t h e case of t h e Hempel pipette for use with mercury, 2 5 cc. of t h e proposed solution was employed a n d this was t r e a t e d with 100-cc. gas samples containing a b o u t 90 per cent oxygen until t h e absorption failed t o be complete for different times of shaking. T h e specific absorption under these conditions was found t o be a b o u t 2 1 for I-minute shaking a n d about 24 for 2- a n d 3-minute shakings. These results are only approximate since t h e a m o u n t of oxygen i n each sample represents nearly four times t h e volume of t h e reagent. I n t h e case of t h e Hempel double pipette for liquid reagents, 18; cc. of t h e reagent was employed a n d this was t r e a t e d with 100-cc. gas samples of a mixture containing about 90 per cent oxygen, alternated with admission of pure oxygen, until t h e absorption failed t o b e complete for different times of shaking. T h e specific absorption under these conditions was found t o be zero for I-minute shaking, 26.9 for a-minute shaking, 2 9 . 9 for 3-minute shaking, 31. I for 4-minute shaking, a n d 32. 5 for 5-minute shaking. T h e striking differences i n t h e specific absorptions for t h e t w o forms of a p p a r a t u s i n t h e case of I-minute ’ shaking m a y be due, i n a measure a t least. t o t h e following causes: ( I ) T h e heat of combination of oxygen a n d pyrogallol would cause a greater elevation of t e m p e r a t u r e when t h e volume of t h e reagent is 2 j cc. t h a n when i t is 185 cc., which in t u r n would t e n d t o increase t h e velocity of t h e reaction as Hempel has shown. ( 2 ) T h e shaking process would probably result i n a more i n t i m a t e contact oE t h e gas a n d reagent when a small volume of t h e reagent is confined over mercury t h a n when t h e reagent fills t h e pipette, due t o t h e m o m e n t u m of t h e mercury when t h e direction o f s i t s motion is suddenly changed. O T H E R C O N S I D E R A T I O N S T H A T MAY A F F E C T T H E S P E C I F I C ABSORPTION UNDER WORKIKG CONDITIOKS
Conditions which affect t h e specific absorption of alkaline pyrogallol, such as t h e composition of t h e reagent (concentrations of pyrogallol a n d potassium hydroxide) a n d t h e method of i t s manipulation ( t h e a p p a r a t u s employed, temperature of t h e reagent, time a n d character of shaking, a n d percentage of oxygen i n t h e sample), have been t a k e n into consideration in t h e previous discussion. Other factors t h a t m a y influence t h e specific absorption under exceptional circumstances are t h e deleterious action of high temperature on t h e reagent, a n d t h e frequency with which analyses are made. The Deleterious -4 ctioiz of H i g h Temperatures-The effect of t h e temperature of t h e reagent upon t h e speed of absorption of oxygen has already been mentioned.
Vol. 7 , No. 7
At temperatures much above t h a t of t h e room, this effect is also accompanied b y a decrease in t h e specific absorption of t h e reagent, i. e . , t h e reagent becomes exhausted earlier t h a n one t h a t is employed only a t ordinary temperatures. This decrease in t h e specific absorption of t h e reagent was noticed particularly in t h e case of a reagent which was placed in a Hempel double pipette for liquid reagents a n d which m-as t r e a t e d writh oxygen in such fashion t h a t its temperature rose t o a b o u t 5 0 ’ C. After t h e absorption of 1500 cc. of oxygen, t h i s reagent removed t h e oxygen completely from a sample of air with I-minute shaking, b u t after t h e absorption of a n additional 2 0 0 0 cc., i t failed t o remove t h e oxygen completely from a sample of air, even with 3-minute shaking. T h e t o t a l volume of oxygen u p t o this point would correspond t o a specific absorption of a b o u t 19 while t h e specific absorption of another reagent prepared in t h e same manner a n d exhausted without allowing t h e temperature t o rise much above t h a t of t h e room, T’i’as, as before s t a t e d , 3 1 . 3 for 3-minute shaking. It seems probable t h a t this decrease in specific absorption is t h e result of a change in t h e character of t h e reactions t h a t t a k e place between oxygen a n d pyrogallol a s t h e reagent is heated. rather t h a n t h e result of t h e direct action of heat upon t h e reagent. Whatever m a y be t h e cause, t h e drop i n specific absorption caused b y heating of t h e reagent would not be of importance under ordinary conditions unless m a n y samples containing high percentages of oxygen were being analyzed i n rapid succession. The Frequency with which Analyses Arc Made-It was found t h a t t h e rapidity of t h e absorption of oxygen from a given sample depended upon whether or no t h e reagent h a d been used a short t i m e before, t h e absorption taking place more rapidly when t h e reagent h a d been just previously used. This s t a t e of affairs was found t o exist i n t h e analysis of t h e gas samples containing a b o u t 90 per cent oxygen. As has previously been s t a t e d , t h e specific absorption of t h e proposed reagent i n t h e Hempel double pipette was found t o be zero for samples of t h e above composition for 1-minute shaking. After t h e reagent had been employed for a few analyses, however, i t was found t h a t complete absorption of t h e oxygen was obtained i n one minute a n d this continued for a few a b sorptions, after which 2-minute shaking was again necessary for complete absorption. F u r t h e r , when t h e reagent h a d been so far exhausted t h a t 3-minute shaking was necessary for t h e complete absorption of oxygen from a I O O cc. sample, t h e reagent was allowed t o s t a n d for a b o u t 24 hours without being used; when t h e t r e a t m e n t of t h e reagent was resumed, i t was found t h a t complete absorption was not obtained with 3-minute shaking of t h e first sample, although i t was i n t h e second. This occurred a second time with t h e same reagent under similar conditions. The cause of t h i s behavior of t h e reagent m a y be due t o a slight rise in temperature caused b y t h e first analysis or i t m a y be d u e t o t h e presence of some intermediate product of t h e reaction which hastens t h e absorption a n d which disappears on standing. At any r a t e , it
J u l y . 191j
T H E J O C R , Y d L O F I N D U S T R I A L AiYD E,VGI,VEERING C H E L V I S T R I -
is ( i t importance only when the reagent has nearly lost its ability t o absorb oxygen completely from a sample under the condition of its use, a n d t h e n only when i t has not been used previously for several 11our 5. THE. F O R M A T I O X O F C A R B O S Y O S O X I D E
OK THE OXIDA-
T I O S OF ALKALIXE PYROGALLOL
S o systematic s t u d y of t h e conditions t h a t affect t h e formation of carbon monoxide was a t t e m p t e d . However, samples of air t h a t were placed in contact with t h e proposed reagent after i t was no longer capable of removing t h e oxygen completely during sei-era1 minutes' shaking were examined for carbon monoxide a n d none was found. I n each case, shaking the same sample with a solution of alkaline pyrogallol. stili in good condition, resulted in t h e indication of the proper percentage of oxygen, a n d subsequent treatment with a n acid solution of cuprous chloride. freshly prepared, h a d no effect upon t h e volume of thc sample. It is therefore evident t h a t t h e formation of carbon monoxide a t t e n d a n t upon t h e use of t h e reagent t h a t has been proposed is not sufficient in xinount t o be objectionable, even when t h e reagent is almost exhausted. TIII:
1.Sb:
OF
POTASSIl'lI
HYDROXIDE
PURIFIED
BY
ALCOHOL
Hempel' has called attention t o t h e fact t h a t potass i u m hydroxide purified b y alcohol should not be employed in t h e preparation of alkaline pyrogallol, a n d many authors have mentioned this precaution in their description of t h e reagent. Hempel did not state wherein t h e unsuitability of this preparation lay, a n d in a n a t t e m p t t o explain why its use is undesirable. t h e following experiment was performed: Five grams2 of pyrogallol was dissolved in j o cc. of n solution of potassium hydroxide purified by alcohol of I , j j specific gravity; 2 j cc. of t h e resulting reagent was introduced i n t o a Hempel pipette for use with merdury a n d there treated with I O O cc. samples of air in t h e manner previously described. I t was found t h a t , within t h e limits of accuracy of t h e a p p a r a t u s , t h e absorption of oxygen was complete, a n d t h a t the only noticeable difference between this reagent a n d one prepared b y using potassium hydroxide not purified b y alcohol is in t h e specific absorption, which is about 1 8 in t h e first case a n d a b o u t 2 1 in the second. T h e apparent discrepancy between t h e results of this experiment a n d Hempel's statement m a y perhaps be explained b y t h e assumption either t h a t t h e nicthod of manufacture of potassium hydroxide purified b y alcohol has been changed in such a manner since t h e publication of Hempel's results t h a t t h e objection t o t h e use of this substance no longer holds: or t h a t Hempel's objection was made o n t h e ground of incomplete absorption, apparent in his a p p a r a t u s for exact analysis, b u t not noticeable in t h e use of t h e Hempel a p p a r a t u s for technical analysis. Ber., 20 (ISS;), 1865. This experiment was made before t h e question of t h e most desirable reagent for technical use hod been settled. 2
THE
USE
595
O F S O D I C M H Y D R O X I D E Ilr- P L A C E
OF POTAS-
SICM HYDROXIDE
F r o m t h e experiments of Weyl' and his co-workers and of Berthelot,' it is not clear b u t what sodium hy, droxide might be used t o replace potassium hydroxide in t h e preparation of alkaline pyrogallol, a substitution t h a t would result in a material decrease in t h e cost of t h e reagent. A4ccordingly, a solution w a s prepared b y dissolving j g. of pyrogallol in j o cc. of a solution of sodium hydroxide of 1 . 3 j j specific gravity prepared b y dissolving 8;. j g. of stick sodium hydroxide in 900 cc. of water: 2 j cc. of this reagent w a s transferred t o a Hempel pipette for use with mercury and it was t h e n treated with a I O O cc. sample of air. I t was found t h a t the absorption a t the end of 3-minute shaking amounted t o only I O per cent. b u t a t t h e end of I Z minutes absorption w a s complete. From this experiment i t seems t h a t such a reagent would scarcely be satisfactory for general use as an absorbent for oxygen. S L- 11M A R Y
I-After a consideration of t h e various factors t h a t determine t h e desirability of a solution of alkaline pyrogallol for general use, it was decided t h a t t h e solution prepared b y dissolving I j g. of pyrogallol in I O O cc. of a solution of potassium hydroxide of specific gravity I . j j is t h e most desirable. For t h e preparation of this solution of alkali there is needed from I . j t o 2 p a r t s of potassium hydroxide in stick form t o I p a r t of water, t h e a m o u n t depending upon the water content of t h e preparation. T h e volume of reagent t h a t is obtained is about I I O cc. a n d there are t h u s 1 3 . 6 g. of pyrogallol a n d 71. j g. of potassium h y droxide t o I O O cc. of solution. If a n y carbon monoxide is formed, the a m o u n t is too small t o have a n y bearing upon t h e use of t h e reagent for technical purposes. 11-Different values of t h e specific absorption of this reagent for different conditions are given in round numbers in t h e following tabulation. I t is understood t h a t t h e temperature of t h e reagent was between 2-24' C.. t h a t t h e method of shaking previously described w a s closely followed, t h a t t h e initial volume of t h e gas sample w a s always 1 0 0 cc., a n d t h a t t h e volume of the reagent w a s ? j cc. for the first pipette a n d 18j for t h e second. S P E C I F I C A B S O R P T I O S ~ I N HEMPEL P I P E T T E F O X USE WITH
. . . . .. ... . . . . . . . gas samples containing 20.9 per cent o x y g e n . . . .
h*o. of Minutes Samples %-ere Shaken..
1
MERCURY 2
3
4
For 27 30 31 . . . . ( b ) For gas samples containing a b o u t 90.0 per cent oxygen. 21 24 24 . . . . S P E C I F I C ABSORPTIONIS HEMPEL DOUBLEPIPETTE FOR LIQUIDREAGENTS 25 29 31 . . . . ( a ) F o r gas samples containing 20.9 per cent oxygen (a)
( h ) F o r gas samples containing about 90.0 per cent oxygen
0 27 30 3 I 1 2
I n t h e application of these figures t o routine gas analyses, t h e conditions t h a t have been specifictl should be followed accurately, otherwise proper allowance should be made for a n y deviation. Probably t h e greater portion of t h e samples t h a t are analyzed for oxygen contain less t h a n 2 0 . 9 per cent of i t , and t h e specific absorptions for samples of this composition are applicable t o all samples containing less oxygen.
2
1
LOC. cil.
596
T H E JOL-RN.1L OF I N D C S T R I A L .4-VD ENGISEERI.VG
T h e specific absorption for samples containing 9 0 . o per cent oxygen are applicable t o all samples containing less t h a n t h i s a m o u n t of oxygen a n d probably are applicable, for all practical purposes, t o t h e analysis of nearly pure oxygen. Due allowance should be made i n t h e use of solutions t h a t are nearly exhausted if t h e y are not used frequently, or if t h e y have previously been employed a t temperatures considerably above 24’ C. 111-The objection t o t h e use of potassium hydroxide purified b y alcohol does n o t appear t o be valid. IT.‘-The substitution of sodium hydroxide for potassium hydroxide is n o t practicable. Determinations of t h e specific absorption of t h e proposed reagent i n some of t h e other forms of absorbing a p p a r a t u s are being made a t t h e present time a n d i t is planned t o present t h e results t h a t are obt a i n e d as a subsequent article of t h i s series. C O R N E L L UNIVERSITY.
ITHACA. h-EW YDXK
THE INDICATOR IN PYROLIGNEOUS ACID By J. M.
1-01. 7 , No. 7
not t h e n be held accountable as t h e indicator prese n t in pyroligneous acid. T h e ethers of pyrogallic acid as well as those of its homologues, methyl-, a n d propyl-pyrogallic acid have been found‘ by Hofmann t o be present in wood t a r . As a rule these are compounds which are readily volatile with steam a n d give color reactions with ferric salts similar t o those referred t o . M a n y of these ethers have been synthetically prepared a n d similar color reactions were noted. T h e a u t h o r has synthetically prepared t h e dimethyl ether of pyrogallic acid a n d found t h a t even minute traces of t h i s ether dissolved i n water gave with milk of lime, known t o cont a i n iron, color reactions comparable t o those obtained with pyroligneous acid. It seems not a t all improbable t h e n t h a t t h e indicator present in pyroligneouq acid consists of t h e volatile ethers of pyrogallic acid 2nd its homologues. SYRACUSE UNIVERSITY,SYRACCPE, SEW TORIC
ON THE KAMBARA EARTH AND ITS BLEACHING ACTION ON OILS
JDHLIN
By SEIICHIU E X O
Received April 26, 1915
T h e fact t h a t pyroligneous acid contains its own indicator has long been a well-known fact employed in t h e wood distillation industries, a n d was recently referred t o i n THISJ O L - R K A L . ~ There seems t o be no theory or explanation offered in t h e literature regarding t h e n a t u r e of t h e indicator which, when t h e neutral point is reached, t u r n s pyroligneous acid ‘(a pronounced mrine-red.” T h e a u t h o r has observed t h a t when milk of lime prepared from chemically pure lime is added t o pyroligneous acid t h i s color reaction does not t a k e place. I t seemed evident t h e n t h a t this reaction must be due t o impurities i n t h e lime as well a s i n t h e pyroligneous acid. Trial experiments showed t h a t milk of lime, t o which traces of ferric salts h a d been added, gave with pyroligneous acid t h e color reactions noted. T h e color t h u s produced is not unlike t h a t caused b y t h e addition of ferric salts t o solutions containing tannins. Tests showed t h a t tannins, gallic acid a n d pyrogallic acid dissolved i n dilute acetic acid produced, with milk of lime containing traces of iron salts, color reactions comparable t o those produced between pyroligneous acid a n d t h e same sample of milk of lime. Since tannins are nonvolatile a n d cannot be distilled i t c’oes not seem likely t h a t a n y of these should be found i n pyroligneous acid. Pyrogallic acid boils close t o 3 0 0 ’ a n d is not volatile with s t e a m . I t is not probable t h e n t h a t t h i s compound should be prese n t in redistilled pyroligneous acid. Guaiacol a n d its homologues, methyl-, ethyl-, and propyl-guaiacol have been found2 present among t h e products of destructively distilled wood. Tests made with guaiacol a n d milk of lime containing iron failed t o give a n y color reactions, though t h e same sample of milk of lime gave decided color reactions with pyroligneous acid. Guaiacol or its homologues should
CHEMISTRY
Received January 4, 1915
K a m b a r a earth is called acid e a r t h ” or “acid ciay ” i n J a p a n a n d is widely used i n oil factories ai: t h e bleaching a n d refining agent of mineral a n d i a t t > - oils. On account of t h e simplicity of t h e decolorizing mcthod a n d t h e cheapness of t h e works, this use of t h e enrth has greatly increased. This e a r t h was recently studied b y K . Kobayashi2 b u t , as far as I knom-) no reports on t h e bleaching action of t h e earth on f a t t y oils have been published. ‘(
0CC U R R E 7 SC E , P R 0 D U C T I 0 S
, 31I S I S G
AS D P R E P .A R A TI 0 S
T h e so-called K a m b a r a ’ ! a n d similar earths are widely distributed in J a p a n . Although t h e acidic earths occur in several localities, t h e most efiective material for t h e bleaching a n d refining of oils is found a t Kawahigashi village, K i t a k a m b a r a district, Echigo province. Kawahigashi is a n inland village situated 7 miles southeast of t h e t o w n Shibata. T h e s t r a t a of t h e e a r t h mine are brownish orange, light yellow a n d bluish green. T h e e a r t h is worked much t h e same as a n y clay b a n k : first about 2 t o 3 f t . of t h e surface must be removed with pick a n d shovel, t h e n t h e e a r t h is dug in much t h e same manner. T h e massiT-e earth is brought t o a small factory b y hand-cars a n d is crushed into fist-sized lumps t h a t are t h e n dried i n a flat, iron pan with direct fire a n d ground t o bean-sized grains, causing t h e loss of about 40 per cent in weight on account of t h e evaporation of water. T h e dried e a r t h is brought l o a water-mill barn n h e r e i t is ground t o l-ery fine powder a n d passed through a silk sieve. T h e powdered a n d refined e a r t h is put in a paper sack covered with a straw bag. T h e weigh: of a bag is 2 0 Kwan ( 1 6 j . 6 lbs.) a n d t h e e a r t h factory can produce 130 bags per day. One bag of earth costs a b o u t 80. j o net. P R O P E R T I E S A S D CO1\IPOSITIOSS O F THE E A R T H S
T h e factory does not sort t h e earths according t o
1
THISJOCRNAL, 7 (1915). 47
1
Cf. Reilstein, “Organische Chemie.”
2
Cf. Beilstein, “Organische Chemie.”
e
THISJ O U R N A L , 4 (1912), 891.