Assay of Picric Acid by Coulometry at Controlled Potential - Analytical

Organic Polarography. Stanley. Wawzonek. Analytical Chemistry 1958 30 (4), 661-674. Abstract | PDF | PDF w/ Links. Cover Image ...
0 downloads 0 Views 517KB Size
V O L U M E 2 8 , N O . 1, J A N U A R Y 1 9 5 6 Table 111.

103

Comparison of Vapor-Liquid Ratio Results

Obtained in duplicate with similar apparatus a t different laboratories motor fuel sample 8 I< L G Laboratory a b a b a b Duplicates Temperature, 1: Vapor-Liquid Ratios 110 0.6 0.7 0.6 0.6 0.6 O.R 115 1.0 1.0 0.0 0.9 1.1 0.9 120 125 130 135 140 145

2.2 5.1 10.2 19.2 28.7 38.7

2.2 5.1 10.2 19.2 29.2 38.7

1.8 3.6 8.1 16.0 26 2 37.3

1.9 3.8 8.7 16.9 27.7 39.4

1.5 3.0 7.6 15.6 27.6 40.7

1.6 3.5 7.7 15,6 27 2 38 0

130 135.5 140.5 146

130 135.5 140.5 145.5

131.5 137 141.5 I46

131 136 5 141 145

132 137 141 144.5

131 137 141.5 14.5.5

for

6

0 0 1.2 1.9 3.5 7.2 15.8 26.0 36.0

Vapor-Liquid Ratios 10 20 30 40

13% 137 142 147

of vapor locking tendency (6). The results are considered to offer better criteria for the suitability of motor gasolines than the more usual volatility characteristics such as vapor pressure or distillation range. From the results obtained a t the Sunbury laboratories for the blend of 30 parts of n-pentane to 70 parts by volume of iso-octane (Table I ) it will be seen that the variations in temperature from the mean values corresponding to the vapor-liquid ratios chosen are: Vapor-liquid ratio

10

~ ~ a r i a t i o n f r o m m e a n t e m p e r a t u rO eF .,}

t:::

20

0":

30

40

::: ::;

from which the average deviation from the mean is 50.7' F. Furthermore the data from the CRC apparatus in Table I ('ompare favorably with the results obtained on the Siiiibury ap-

par:ttus. The deviatioti from the tiican Sunbury values varies between $0.4" and -0.8' F. When considering the duplicate results on motor fuel blends (Table 11), the maximum temperature differences between any pair of results for vapor-liquid ratios of 10 to 40 are 1O F. which may be expressed as k0.5' F. This order of repeatability (i0.5"F.)is satisfactory and is confirmed by the duplicate results reported from three other laboratories (Table 111). The data at present available for the assessment of the reproducibility of the inetliod we somewhat meager and are limited to those presented in Table 111. The maximum temperature difference a t any of the vapor-liquid ratios quoted-viz., 10, 20, 30, and 4 0 4 s 2" F. and may be int,erpreted tentatively as f 1 ° F. However, further trials of the procedure a t different, laboratories are necessary before this figure for reproducibility may be substantiated. The satisfactory agreement of the experimental with the calculated theoretical data on the blend of 30 parts of n-pentane to 'io parts by volume of iso-octanc (Table I ) further indicates the suitability of the apparatus for the determination of vapor-liquid ratio data. LITERATURE

crrm

(1) lm.Pctroleuui Inst., API

Project 44, "Selected Values of Properties of Hydrocarbons and Related Compounds," data sheets &E, k,k-E. ( 2 ) ('itmpbell. J. l l . , Lovell, W. G., and Mulligan, hl. J., J . SOC. Automotive Engrs. Trans. 39, KO.3, 356 (September 1936). (3) Coordinating Research Council, "Evaluation of the CRC 1-ILApparatus," report issued January 1951, rev. March 1953. (4) Inst. Petroleum (London), "Standard Methods for Testing Petroleum and Its Products," 14th ed., method 59/55, Table \-, 1955.

( 5 ) Sheppard, G. F., and Withers, J. G., "ilssessment of Car and I'iiel Factors ilffecting Vapor Locking Tendency," Inst. llech. Iltigrs. (London), preprint (January 1955). RECEIVED for review July

14, 105:.

Accepted September 10, l!hj:.

Assay of Picric Acid by Coulometry at Controlled LOUIS MEITES'

and

T H E L M A MEITES

Sterling Chemistry Laboratory, Y a l e University, N e w H a v e n , Conn.

i t a mercury cathode whose potential is kept constant at -0.40 volt us. S.C.E., the reduction of picric acid from hydrochloric acid solutions proceeds rapidly and quantitatively under the proper conditions to 2,4,6triaminophenol. Integrating the current which flows during such an electrolysis is recomniended for the assay of picric acid in conjunction with a conventional alltalimetric titration.

N ACCURATE evaliintion of t,he purity of picric acid by classical techniques is so far from being a simple matter that cntirely satisfactory procedure for the assay of this reagent appcws ever to have been proposed. Since lower nitrated phenols are certain to be present as cont,aminants, it is evident that the common alkalimetric "assay" procedure necessarily gives a fictiously high value of the picric acid content. Moreover, the fact' that thoroughly dry picric acid is a somewhat hazardous subst,ancc, so that the assay is preferably carried out on a wet sample, makes it difficult to set up a standard by which the result can be judged. In this paper there is proposed a procedure which involves two I10

' Present address, Department of Chemistry, Polytecl~nic Institute of Brooklyn, 00 Livingston S i . , Brooklyn 1, N.Y.

independent analyses of an uiidried saiiiple. One consists of the measurement of the amount of current consumed in the quantitative electroreduction of the organic nitro groups by the electrolysis of a hydrochloric acid solution of the sample a t a mercury cathode, whose potential is kept constant a t a suitable value. This gives the number of equivalents of reducible material per gram of sample. The other analysis consistg of a conventional alkdimetric titration, which gives the t,otal number of equivalents of replaceable hydrogen per gram of sample. As is shown, the reduction of one mole of picric acid, containing one equivalent of replaceable hydrogen, consumes 18 faradays of electricity. Conseqnent,ly, the ratio between the coillometric and alkalimetric values is exactly 18 for picric acid containing water as the only impurity. .Is this rat,io is 12 for any dinitrophenol present, it.s value serves as a measure of the relative amounts of tri- and dinitrophenol present. The electrolytic reduction of picric acid from hydrochloric acid solutions a t a mercury cathode was first studied by Lingane ( 3 , 4 ) . Employing a 0.4mM solution of picric aci'd in O.lJ4 hydrochloric acid, Lingane found n = li.1, a value which Bergmnn and James, using identical conditions, later duplicated ( 1 ) . On this basis Lingane ( 4 ) suggested that the reduction prodwt is bis(3,5-diamino-4-hydroxyphen~l)-hydrazine:

104

A N A L Y T I C A L CHEMISTRY OH

desired volume of picric acid solution. The solution in the working electrode compartment was then completely deaerated by bubbling a rapid stream of prepurified rlit,rogeri through it for not less than 15 minutes.

NH:!

Kolthoff and Lingane in collaboration with Wawzonek ( 2 ) discussed, the possibility that this assumed product might undergo the benzidine rearrangement, while Muller (6) attempted to explain the polarograms of picrate solutions at various p H values on the basis of an over-all 17-electron reduction. Pearson (8) and Neiman, Kuznetsov, Rabinovitch, and Ryabov ( 7 ) also studied the polarographic characteristics of picric acid a t various pH values. Pearson stated that the total diffusion current corresponded approximately to a 16-electron reduction, assuming that picrate and benzoate ions have equal diffusion coefficients. In view of the rather large differences betweell the sizes and formula weights of these ions, this is very unlikely to be the case. Herein it is shown that the reduction of picric acid consume8 exactly 18 faradays per mole (and must therefore give 2,4,6-triaminophmol) provided that the concentration of picric acid is very low and the acidity is relatively high. Otherwise-e.g.. above 0.2m.M picric acid in 0.1M hydrochloric acid, above lmAl picric acid in 1M hydrochloric acid, or above 1.5mM picric acid in 3 M hydrochloric acid-the value of n falls below 18. This is evidently the result of a side reaction which involves some int'ermediate reduction product: This reaction is favored by a high concentration of the intermediate (or, naturally, of the original picric acid), and is retarded by a high concentration of hydrogen ion. With 0.4mM picric acid in 0.1M hydrochloric acid, the authors find, in agreement with Lingane and with Bergman and James, that the apparent value of n is close t o 17. At still higher concentrations of picric acid, however, n becomes distinctly less than 17. This may be interpreted to mean that the reduction of picric acid, like that of nit,robenzene, proceeds via a substituted hydroxylamine, which can t,hen rearrange to form an aminophenol (in this case, a dihydroxytrianiinobenzene). Values of n then can be found lying between 16 and 18, depending on the relative fractions of the hydroxylamine undergoing rearrangement and reduction. EXPERIRIENTAL

The potentiostat, and current integrator used in this work were manufactured by Analytical Instruments, Inc., Brist,ol, Conn., and have been described ( 6 ) . All electrolyses were carried out in a double diaphragm cell (6) filled t'hroughout with hydrochloric acid of the same concentration. An efficiently stirred mercury pool served as the working electrode; the auxiliary electrode was a helix of stout platinum wire. A few drops of satilrated hydrazine dihydrochloride mere added t o the ausiliary electrode compartment to serve m a depolarizer. A sample of pure picric acid was secured by recrystallizing the reagent grade chemical once from water, once from 95% ethyl alcohol, and finally again from water. A small portion of the product was dried for several weeks over anhydrous marnesium perchlorate, and used t80prepare an aqueous solution of exactly know-n concentration. -411 other chemicals used (excepting one sample of t.echnica1 grade picric acid) were ordinary reagent grade and were not further purified. The carbonate-free sodium hydroxide solution was standardized against primary standard grade potassium hydrogen phthalate. All weights and volumetric apparatus had been carefully cali1)rated by conventional techniques. To begin a coulometric determination, the central and auxiliary electrode compartments of the cell were filled with hydrochloric acid of the desired concentration. About 75 ml. of acid was then added to the rvorking electrode rompartment, followed by the

Complete removal of oxygen from the solution is more important here than in most coulometric analyses. Sormally, the interference of oxygen results from the fact that it is reduced a t the working electrode togethe], with the substances being determined. As this reduction proceeds a t a finite rate, one can usually be content with removing most, but not all, of the dissolved oxygen before beginning the electrolysis. Ordinarily only a small fraction of the oxygen left will be reduced, whereas most of i t will be removed by the stream of nitrogen. This is not true, however, in the picric acid reduction, for the reduction product, formed at the very beginning of the electrolysis is moxidized almost instantaneously by dissolved oxygen, the entire amount of which is thus rapidly reduced, giving a positive error. Since, as is shown, only very small amounts of picric :wid can be hendled in the coulometric procedure, the relative error thus incurred can be very serious indeed. Wlien the deaeration is voniplete, 25 ml.of mercury is added to the working electrode compartment, the poteritiostat is adjusted to maintain the working electrode potentid ( E = , e ,at ) -0.40 volt 21s.S.C.E., and the electrolysis is begun. I n these experiments, the total volume of the solution in t,he working electrode compartment was always close to 80 ml., and in consequence t.he electrolyses proceeded more rapidly than with the larger volumes of solution used in earlier work (6). Every electrolysis was complete-i.e., the register on the current integrator came to a full s t o p w i t h i n 25 t o 30 minutes. The solution must be discarded \\Then the electrolysis is complete; addition of another aliquot of the sample in an attempt t o run a check determination invariably leads to air oxidation and high results. All of the electrolyses xere c.ai.ried out n.ith the integrator set on its 1 microfaraday per count range, and the number of counts recorded was multiplied by the previously determined ( 5 )calibration factor 1.00082, t o secure the number of niiurofaradays used. DATA ANI) DISCUSSIOS

A polarogram of picric acid in 1JI hydrochloric acid consists of a single irreversible wave rising from zero applied electromotive force and having an extremely well defiocd plateau. F h e n dilute solutions of picric acid were electrolyzed a t -0.40 volt us. S.C.E., the data shown in Table I w r e secured. These data show conclusively that the reduction of picric wid under these conditions is an 18-electron process, and yields 2,~,ii-trianiiriophenolas the product:

Tile electrolysis of a dilute piwic. acid solution finally >,icldn :I perfectly colorless solution, which becomes yellow brown almost instantly when exposed to air, :is :i i,twilt of the reac*tion

0 OH H ? S \?;Et.

0 NHz

II

+ H1O

+ 1 20,NH

As might be expected on the basis of the large number of electrons involved, the course of the reaction is very complex. The initially yellow solution becomes successively green, brown, pink, and, a t about 98% completion, a light purple n-hich fades s10a.l~

105

V O L U M E 2 8 , N O . 1, J A N U A R Y 1 9 5 6 Table I.

Determination of n for Picric Acid

(Known amounts of pirric acid w c r ~added to 7 5 - i d . portions of hydrochloric acid a n d the deaerated mixture6 were elertrolyzed a t a mercury cathod; whose potential w a s maintained a t -0.40 volt 13s. S.C.E.) Micromoles of n Faradays/ Picric Acid HCI, Microfaradays Bf Consumed ;Mole Taken 136 0 18,02 0 .1 7 58 18.00 272 8 0.1 1: 14 1 8.03 1 545 0 30 22 18.03 1 1079 6 .59 87 1 7.97 1 1353 R 75 34 18.00 1356 1 1356 1 18.00 1336 1 18.00 17.99 1 1624 R 90 33 18.01 1626 fi 18.00 90 R7 3 IR31 8 17.99 151 3 3 2722 4 17.99 2722 6

when the initial concentration of picric acid in a I N hydrochloric acid solution is Ieea than about 1mM; the electrolysis of a more concentrated solution yields an orange solution, and the quantity of electricity consumed is abnormally low. This orange product cannot be reduced a t any potential before the discharge of hydrogen ion begins. The orange product is no doubt formed by the interaction of the products of the primary reaction a t the electrode surface. Except for the picric acid concentration a t which the negative error becomes appreciable, the general trends in 0.1 and 3.W hydrochloric acid are very similar to that in 1M arid. The nonelectrolytic side reaction is evidently retarded by increasing acid concentration.

Table 111.

Mean 18 00 4~0.016 (std. d e l . )

Effect of Hydrochloric Acid Concentration on Electrnlytic Reduction of Picric Acid

ry1, iintil the electrolysis is complete.. It is not known whether these Intermediates result from a stepwise reduction or from reactions betwen the electrolysis product arid unreduced picric acid. Effect of Working Electrode Potential. Khen solutions of 31 88 micromoles of picric acid in 75 ml. of I M hydrochloric acid nete electrolyzed a t various potentials. the data shown in Table I1 v,c'ie scrured. The values secured a t potentials between -0 20 :tnd -0.60 volt are in good agreement nith the expected v:tluc.. hut a t more negative potentids hydrogen ion begins to he i~duc-ctldong Kith the picric acid, and high results are ohtained 1 moorking elertrode potential of -0.40 volt m S.C.K. IS thriefot P recommcndcd

Table TI.

-0

Effect of Working Electrode Potential on Electrolytic Reduction of Picric Acid

20

-0.40

627 8

627 7 628 2 627 8

Error,

%

+O

+o

07 14

0.5

55.91

-0.18

1.0

55.99 56.09 R R .03

-0 04 +0.14 +0.04

2.0

86.10 55.94

+o.

3.0

5 A 00

55.89

-0.21

6.0

56,39 56.57

$0.68 f1.00

16 -0.13 -0.02

Analyses of Commercial Samples of Picric Acid. TWOcommercial samples of picric acid were analyzed by the proposed procedure. The wet material was thoroughly mixed, and a 1.5gram sample was weighed out, transferred to a 250-ml. volumetric flask, dissolved in recently boiled water, and diluted to the mark. Two 100-ml. portions of this solution were titrated with standard sodium hydroxide. In the interest of securing greater

-0.02 f0.06

+o.oo

625 5 627.9

+0.03 -0.37 $0.02

- 0 60

627.6 627.6

-0.03 -0.03

-0.70

629.2 628.5

$0.22 fO.11

-0.80

636.5 634.5

+1.4 4-0.9

-0.90

659.5

+5.1

628.0

0 3

Micromoles of Picric Acid Taken Found 56 01 56 05 .56 09

Table IV. Coulometric Determination of Various Amounts of Picric Acid in Hydrochloric Acid Micromoles Picric Acid Taken

Picric Acid Concn., m iM

Micromoles Picric Acid Found

Error,

%

0 1MHCI

7.58 15.14 22.72 30.22 59,87

0.10 0.20 0.29 0.38 0.72

13.88

7.59 15.14 22.39 28,69 54.84

'ro.12

0.18

13.89

+0.10

28,03

0.36

28.09 28.04

+0.25 +0.07

56.01

0.71

56.01 55.89

-0.01 -0.21

69,68

0 87

69.79 69.89 69.85 69.71

f0.16 +0.30 -t-0.24 -I-0.04

2zo.00

-1.5 -5.1 -8.4

1MHCI

Effect of Hydrochloric Acid Concentration. A number of solutions of 56.01 micromoles of picric acid in hydrochloric acid media of various concentrations were electrolyzed a t -0.40 volt us. S (?.E., with the results shown in Table 111. These data show that accurate results are obtainable a t any hydrochloric acid concentration between 0.3 and 3 M . In 6 M acid the values are high, and reducing the n-orking electrode potential to -0.20 volt us. S.C.E. decreases but does not entirely eliminate the error. In any case, no advantage would seem to result from the use of a hydrochloric acid concentration differing from 1.M. As seen in the following paragraph, the hydrochloric acid concentration and picric acid Concentration both evert a profound influence on the results obtained a-ith all but the most dilute picric acid solutions. Effect of Picric Acid Concentration. The results shown in Table IV were secured by adding known volumes (10 ml. or less) of a stock solution of picric acid to 75 ml. of hydrochloric acid, and electrolyzing the mixture a t -0.40 volt us. S.C.E. According to these data, accurate results can be serured onlj-

101.1

1.23

100.8 100.7

-0.31 -0.39

139.6

1.64

138.5 138.7

-0.79 -0.64

172.7

2.16

160.9

-6.8

3M HCI 15.14 30.22 59.87 90.67 151.3 172.7 226.0

4-0.15 4-0. 13 -0.10 -0.10 -0.06 -0.05 -1.4 -2.7

ANALYTICAL CHEMISTRY

106 accuracy, a Beckman Model G pH meter was used to locate the equivalence point of the titration. Meanwhile a 25-ml. aliquot of the solution was diluted to 250 ml. with 1 M hydrochloric acid in another volumetric flask, and two 25-ml. aliquots of this were analyzed coulometrically. When a 1.5052-gram sample of reagent grade picric acid (containing, according to ita manufacturer, 10.7% water) was analyzed by this procedure, it was found that 5.866 & 0.004 meq. of base would have been required to neutralize the entire saniple, and that 105.42 f 0.09 niillifaraday would have been required to reduce it. Assuming that the sample contained x millimoles of picric acid, for which n = 18, and y millimoles of dinitrophenol, for which n = 12, these values give

+ y = 5.866

2

and

18z

+ 12y = 105.42

whence the weight of picric acid in the sample is 1.338 grams (88.9%), and that of dinitrophenol is 5.2 mg. (0.38%). If water is the only other constituent of the sample, it would therefore be present to the extent of 10.75%, in good agreement with the direct determination. -4 sample of technical grade picric acid analyzed in the same way was found to contain 83.4% picric acid and 3.1% dinitro-

phenols. I t is evident that a simple alkalimetric titration of this material would give a seriously erroneous estimate of its picric arid content. ACKYOWLEDGMEYT

It is a pleasure to express thanks to Analytical Instruments, Inc., Bristol, Conn., for the generous grant-in-aid which made this work possible. LITERATIJRE CITED

(1)

(2) (3)

(4) (5) (ti) (7)

(S)

Bergman, I., and James, J. C., Trans. Faraday Soc. 50, GO (1954). Kolthoff, I. M., and Lingane, J. J., "Polarography" vol 2, p. 627, Interscience, New York, 1952 Lingane, J. J., "Electroanalytical Chemistry," p 369, Interscience, New York, 1953. Lingane, J. J , J . Am. Chern. SOC.67, 1916 (1945). Meites, L., ANAL.CHEW27, 1116 (1955). Muller, 0. H., in "Physical Methods of Organic Chemistry" (A Weissberger, editor), vol. 2, pp. 1845-6, Interscience, New York, 1949. Neiman, M. B., Kuznetsov, L. I., Rabinovitch, I. B., and Ryabov, A. V., Zavodskaya Lab. 15, 1280 (1949). Pearson. J., Trans. Faraday Soc. 44, 683 (1948).

RECEIVED for review J u n e 6, 1955. Accepted October 17, 1955. Division of Analytical Chemistry, 128th Meeting, ACS, Minneapolis, Minn , September 1955. Contribution No. 1306 froni the Department of Chemistry of Tale LTniversity.

Determination of Hydrogen in Titanium and Titanium Alloys MAURICE CODELL and GEORGE NORWITZ Pitman-Dunn Laboratories, Frankford Arsenal, Philadelphia 37, Pa.

As hydrogen in titanium and titanium alloys is considered an undesirable impurity, its determination is a matter of importance. In this paper an ignition method is proposed. The sample is ignited in oxygen in the presence of lead as a flux, the gases are passed through hot copper oxide to ensure complete oxidation of the hydrogen to water, and the water is collected in an Anhydrone collection tube. An important probleni in developing the method was to find means for preventing the large amount of heat generated by the reaction between the titanium and oxygen from cracking the reaction tube. This problem was solved by weighing the sample in a piece of 96% silica tubing and resting this tubing in the reaction tirhe on clay supports.

T

HE effect of hydrogen

011 titanium is not as marked as the effect of oxygen and nitrogen on titanium, in so far as such properties as hardness, ductility, and electrical conductivity are concerned (22, 28). Hydrogen, however, can change the grain structure of titanium ( 5 , 12, 29) and has a profound effect on notch bar toughness (SO, 54). The effect of hydrogen that might be picked up during welding, forging, or pickling operations is controversial (26, 35, 69). K O beneficial effect of hydrogen on titanium has been noted; therefore hydrogen must be classified as a contaminant and closely controlled (34). For proper understanding of the problems involved in the determination of hydrogen iri a metal it is essential to consider the metallurgical relationships between hydrogen and the metal. For no other element is there such a close tie-in between metallurgy and analysis. Titanium belongs to a group of metals, which includes zirconium, cerium, tantalum, and niobium, that is capable of absorbing considerable hydrogen in the solid state (15, 21, 53, 5 4 ) . The solution of hydrogen in titanium is

exothermic, and the solubility decreases as the temperature increases, the decrease becoming rapid a t about 600" C. I t is not possible to absorb much hydrogen in titanium a t room temperature. However, if titanium is heated to about 300" to 400" C. and cooled to room temperature, the hydrogen that has been absorbed is retained. When hydrogen dissolves in titanium there is eo film formation, as with oxygen or nitrogen (24, 25). Hydrogen in titanium, when present in moderate amounts, is stable under a vacuum of 10-6 mm. of mercury to approximatel) 275" C. ( 2 5 ) . The properties of titanium saturated with hydrogen (titanium hydride, TiH,) (1-3, 20) have bearing on the determination of hydrogen in titanium. Titanium hydride contains about 4% hydrogen, is metallic in appearance, and is indefinitely stable a t room temperature. l17hen titanium hydride is heated in air, no change takes place until its temperature of dissociation is reached. The hydrogen is gradually expelled and burns with a quiet flame, and when most of the hydrogen has been expelled, the titanium starts to burn with considerable heat. It is important to contrast the behavior of hydrogen in titanium with its behavior in iron, nickel, copper, aluminum, silver, molybdenum, and tungsten (16, 55, 54) in considering application of published methods for the determination of hydrogen in these metals to its determination in titanium. These metals ran absorb only small amounts of hydrogen, and the hydride formed is not stable under ordinary conditions of temperatuIcx and pressure. The hydrogen is absorbed interstitially as atoms without specific interaction with the metal atoms. The process of solution is endothermic, and the solubility increases as the temperature is raised. When iron is cooled from a high tempernture, the hydrogen is slowly evolved over a period of weeks and finally approaches the solubility of hydrogen in iron a t room temperature. I n determining the hydrogen content of freshly cast iron and steel specimens, such treatments as covering with