the presence of copper, cobalt, and nickel. It is not possible t o mask zinc and cadmium adequately with cyanide, however, and if these metals are present the rare earth titration fails. Several other masking agents were tried. Copper is efficiently masked by thiourea, mercury(I1) interferes a t the pH used, and other divalent metals are not masked. Iodide is a useful and selective masking agent for mercury(I1). Triethylenetetramine, 2,3-dimercaptopropanol, and sodium sulfide were tried without success as masking agents for zinc. Some limited success was obtained in masking lead with 2,3-dimercaptopropanol, but only small amounts of lead can apparently be tolerated (Table IV). Diethyldithiocarbamate is very useful for avoiding interference by metals that react with hydrogen sulfide. I n many cases rare earths can be titrated with EDTA directly in the presence of the metal dithiocarbamates. If the metal dithiocarbamates are highly colored or a large amount of precipitate is present, the interfering metal dithiocarbamates are removed by eutraction (occasionally a double extraction is required) into chloroform. Sulfosalicylate is also added to prevent precipitation of rare earth hydrokides during the extraction. The method is particularly important, in that it permits the simple and accurate determination of yttrium and rare earths in samples containing uranium. Interferences by iron, cobalt,
nickel, copper, zinc, silver, cadmium, mercury, and lead are also avoided (Table V). Metal ions studied for which interferences could not be avoided were thorium, zirconium, and chromium. However, methods are in the literature for separating these metal ions from the rare earths, and it is probable that in at least some cases the EDTA titration with arsenazo could be used following the separation. DISCUSSION
Because of the similarity in chemical structure of arsenazo and p-sulfophenylazochromotropic acid (SPADNS), their indicator properties were compared. The absorption spectra of the thorium complexes are very similar, but thorium forms a strong complex with arsenazo over a much longer p H range than with SPADNS. The thorium and rare earth complexes with arsenazo appear to be definitely stronger than the corresponding complexes with SPADSS. One evidence of this is that thorium can be titrated with EDTA a t a lower p H using arsenazo. This is not conclusive in itself, but suggests greater complex stability. Stronger evidence is that many metal ions interfere in the titration of thorium using arsenazo, but both thorium and rare earths can be titrated in the presence of conipleving anions such as sulfosalicylate and sulfate. On the other hand, by use of SPADSS, thorium can be titrated in
the presence of metal ions such as zinc, aluminum, cerium, and lanthanum (6), but sulfosalicylate and other complexing anions interfere. This indicates that the thorium SPADKS complex breaks up a t a higher thorium concentration (lower pTh value) than the thoriumarsenazo complex. Although virtually all of the titrations reported in this paper were carried out with 0.05M EDTA, good end points are obtained even when much more dilute EDTA is employed. The possibility of using arsenazo as a n indicator for microtitrations and as a reagent for the quantitative spectrophotometric determination of rare earths and uranium is now being investigated. LITERATURE CITED
(1) ilnderegg, G., Flaschka, H., Sallmann, R., Schwarzenbach, G., Helti. Chini. Acta 37. 113 11953). Banerjee, ' G . , A4nal. Chin[. A c t a 16, 56, 62 (1957). Baneriee, G., Z. anal. Cheni. 146, 417
(1955).
Ihid., .147, 105 (1955). Ibid., 148, 349 (1955). Datta, S. K., Ihid., 149, 328, 333
(1956); 150, 347 (1956). Fritz, J. S., Ford, J. J., A s a ~ CHEAI. . 2 5 , 1610 (1953). Fritz, J. S., Lane, W. J., Bystroff, A. S., Ihid., 29, 821 (1957). Kuznetsov, V. I., Zhur. .Inal. Khzni. 7, 226 (1952). RECEIVED for review August 12, 1957. Accepted January 24, 1958. Contribution 631. n'ork performed in Ames Laboratory, L-. S. Atomic Energy Commission.
Determination of Acids and Basic Nitrogen Compounds in Petroleum Products IRA KUKIN' Gulf Research & Developmenf Co., Piffsburgh, Pa.
b Nonaqueous titrations with visual indicators were used to determine acids and weakly basic nitrogen compounds, including oils which were too dark to b e readily titrated by existing color-indicator methods.
C
oLoa-indicator titrations, where applicable, are rapid and simple. When naphthenic acids were titrated by a potentiometric method (ASTlI D 644) ( I ) , sharp inflection points could not be obtained when the acid number was less than 0.1; the indicator end point had to be approximated from a pre1 Present Address, L. Sonneborn Sons, Inc., 300 Fourth Ave., New York 10, X. Y.
1 1 14
ANALYTICAL CHEMISTRY
determined e.m.f. value. In such cases, it was more convenient to titrate to a n indicator end point, using a titration flask (6) which facilitated the detection of sharp visual end points. As no ASTlI method exists for determining weakly basic nitrogen conipounds in distillate fuels, a nonaqueous titration with methyl violet as the indicator ivas developed. ACID NEUTRALIZATION METHOD
Apparatus, Reagents, and Method.
BURETASSEMBLY.Two drying tubes containing Ascarite were sealed t o a 5-ml. microburet of the Koch type, graduated in 0.01-ml. subdivisions (Figure 1).
ALCOHOLIC SODIUM HYDROXIDE. Diesolve about 16 grams of sodium hydroxide (analytical grade) in 100 ml. of methanol-ethanol (1 to l), and filter off the small amount of insoluble carbonate. As required, prepare approximately 0.03.V titrant solution by diluting with absolute ethanol. KO discoloration or floc formed when the alcoholic solutions of sodium hydroxide were stored for about a year. hIETHOD. Weigh 20 grams of the Oil into the titration flask. Add benzeneisopropyl alcohol (1to 1) to theetchmark (120 ml.) on the flask. Add 5 drops of p-naphtholbenzein indicator (Q), 0.1% in methanol. Stir the solution with a magnetic stirrer, and titrate with 0.035 alcoholic sodium hydroxide. The rate of addition of sodium hydroxide can be
rapid and continuous, hut he sure that each descending drop becomes mixed with the bulk solution before the next drop enters the neck of the flask. The end point is the first change from orange to green, as seen in the narrow neck of the flask. One really watches the rate of color change and not the color itself; this is helpful in detecting the end point in dark colored oils. Results. COMPARISON WITH ASTM METHODS. Tables I and I1 show random results for distillate and synthetic oils. The proposed method could be used with dark colored solutions where the detection of the end point would be difficult b y the ASTM D 974 color-indicator ( 1 ) .
Certain synthetic oils which contained zinc weakly acidic metal salts-e.g., salt of thiophosphoric acid, nentralization number of 8 h h o w discrepancies between color-indicator and potenticmetric methods (Table 11). This a p pears to result from pnaphtholbenzein changing color at a pH which is slightly helow the equivalence point of the weakly acidic zinc salts. Discussion. The oils containing naphthenic acid have a steeper inflection t h a n the synthetic oils which
1 Oil A
contained weakly acidic zinc salts (Figure 2). Thus, unless a n indicator is used which turns color at the exact equivalence point of the acid, a larger titration error would be expected for the synthetic oils. p-Naphtholbenzein turned green at p H 10 or slightly before the potentiometric equivalence points for the oils shown in Figure 2, curves B and C. If the indicator turns green at point 1, whereas the equivalence point of naphthenic acid is at 2. a maximum titration
Table I. Acid Numbers of No. 2 Fuel Oils Acid Neutralization Number, Mg. KOH per Gram ASTM ASTM No. 2 Fuel Oil Naph- D97455T D664 Proposed Color thenic colorpotentic- colorASTM acid indicator metric indicator Description Union added method method method 0.04* 0.002 1... 0.01
2 OilB
2-
3 Oil B, heated 40 hr. at 210' F.
8-
,..
4 Oil B, plus nitrogen-containing dispersant
2-
...
0.01
5 No. 4, heated 40 hr. at 210" F.
8-
0.90
0.91 n 1-
6 Naphthenic acid added to oil A
(bdilute) 1-
0.01
0.04-
0.01
0.08.
Toodark
0.07*
7 1: 1 blend of 6 and 3
5-
n ,E
8 2:5 blend of 6 and 3
7-
0.22
Too dark
9 1:5 blend of 6 and 3
8.
0.15
Too dark
".*a
".%'
0.004 0.W7 0.008 0.02 0.02
0.040
0.009
0.07O
0.008
0.04-
0.03
0.94
0.89 0.90
='
0.48
0.48 0.49
0.26 0.26 0.15 0.20
0.22 0.22 0.16 0.16
".*I
No inflectionpoint obtained; e.m.f. value, CG.of n.fi5 volt used to determine end mint. Table II. Acid Numbers of Synthetic Diester Oils Acid Neutralization Number, Mg. KOH per Gram 1LSTM Pro csed D107455T nnlnrColor L 1.". Sample ASTM indicator Potentiometric indicator No. Description Union method method method 0 Base oil (esters of di4'/= 0.03 0.03 0.03 basic acids) 1 Base oil plus Series A 4'/r 0.19 0.15 additives. 0.18; 0.18 0.17;0.18 2 Base oil plus Series B 4'11 0.80 1.04 0.88;0.83 additives1.02; 1.02 0.89; 0.89 0.91 0.87 1.45; 1.43 1.49 Too dark No, 1 after oxidation >8 3 1.50; 1.44 test 1.99 Too dark No. 2 after oxidation >8 4 1.89; 1.88 2.01; 1.98 test Too dark 3.04 5 Oil 0 after oxi dation >8 3.50; 3.59 2.62; 2.63 test 6 Oil De after oxidation >8 1.84; 1 . w L.10; J.,.J test Std. dev. 0.122* 0.008 a Containing various combinations of carboa[lic, phcnulic, and phosphoric acid wits of zinc, calcium, and barium. The zinc salta had iwid numbrm of about $3. b If one excludes sample 5, or the value of3.04, l h v ahndard dwiation would la 0.023.
."Y"r-
Where the acid number is greater than 0.1, the ASTM D 974 colorindicator, the ASTM D 664 potentiometric, or the proposed color-indicator method can be used (Tables I and II), except in dark colored oils. I n such cases, either a potentiometric method or an indicator method can be used with a titration flask. Where the acid number is less than about 0.1, no inflection was obtained by the ASTM D 664 method; in these cases, the proposed color-indicator method gave results in better agreement with the expected values.
VOL. 30, NO. 6, JUNE 1958
11 15
error of about 0.2 ml. of 0.1N sodium hydroxide would result (about 9%). For the synthetic oil (curve C), approximately 0.7 ml. of 0.1N sodium hydroxide was required between points 1 and 2; this would be equivalent to an error of 0.4 acid number, or 20% for the particular oil having an acid number of 2.0. Thus, p-naphtholbenzein is convenient for determining carboxylic acids in oils. In titrating used lubricating oils by the proposed color-indicator method, it would be desirable to use an indicator whose transition point is somet\-hat higher than p-naphtholbenzein, because lubricating oils usually contain organozinc compounds, added as antiwear agents.
Table 111.
Iiumber 1
2 3 4 5 6
BASIC NITROGEN DETERMINATION
7
iill nitrogen compounds having a base dissociation constant greater than that of pyrrole could be titrated by using a solvent containing anhydrous acetic acid. These pyridine and quinoline compounds, and their salts, \ w e titrated potentiometrically ( 3 , 7 ) , Clark also used the indicator, methyl violet, to titrate amine picrates ( 2 ) .
8
and Method. PERACID, 0.08N. Mix 7 nil. of
Reagents CHLORIC
72% perchloric acid with 250 ml. of glacial acetic acid, and add 20 ml. of acetic anhydride. Dilute t o 1 liter with glacial acetic acid and allow to stand overnight (6). Standardize with potassium hydrogen phthalate, 0.05 to 0.07 gram, add 50 ml. of glacial acetic acid, and bring the solution to a gentle boil to effect solution. After cooling, add 50 nil. of benzene and 5 drops of indicator (0.25% methyl violet in glacial acetic acid). Titrate rrith perchloric acid to the first complete disappearance of the violet color. METHOD. Weigh 20 grams of the oil into the titration flask. Add benzene-glacial acetic acid (1 to 1) to the etch mark (120 ml.) and 5 drops of indicator. Titrate with 0.08N perchloric acid. Results. C O M P S R I S O N W I T H A P O TENTIOMETRIC LfETHOD. As the exact nature of the basic nitrogen compounds in distillate oils is obscure (8), a k n o m compound was not titrated by the proposed color-indicator method. The accuracy of the method was tested by comparing the results with a potentiometric method developed for distillate oils (4). ST7ith either method, excellent agreement was obtained by prior standardization against known quantities of pyridine. I n the potentiometric method, a 75-gram sample diluted with 400 ml. of glacial acetic acid was titrated with approximately 0.05N perchloric acid (0.05-ml. increments) in glacial acetic acid. The end point was determined from the maximum change in potential following each incremental addition of perchloric acid,
1 1 16
ANALYTICAL CHEMISTRY
9
Color-Indicator and Potentiometric Methods for Total Basicity (in Glacial Acetic Acid) of Distillate Fuels Sample Base Number Mg. of KOH per &am Color, Size, .4SThI Potentio- Colorgrams L-nion Description metric indicator 20 11,'*Blend -4 0.081 0.081 0.076 0.083 20 2Blend B 0.112 0.106 0.116 0.105 20 3Catalytic cracked fuel oil 0.542 0.557 0,559 20 5Blend C, in storage 1year 0.272 0.272 0.281 0.268 20 4Sample 4, treated with adsorb0.099 0.096 ent 0.099 20 1Blend D, inhibited, in storage 0.171 0.172 b months 0.174 0.162 20 5Blend D, inhibited with a metal 0.134 0.106 soap, in storage 6 months 0.134 0.122 '2' 220 Sample 6, treated with silica gel 0.032 0.042 0.022 0.038 0.204.25 .. . Sitrogen compounds desorbed 56.0 56.0 from silica gel extract in 54.2 sample 8 ~~~~
Table IV.
Sample Size by Proposed Methods for Neutralization Numbers Seutralization Size of Sample, Estimated Titrant Solution Sumber Grams Repeatability -4cids 0.03N alcoholic sodium hg20.0 i 2.0 0.015 droxide 0 . 0 to 0.35 2.0 f O . l 0.05 0.35 to 3 . 5 3 . 5 to 35 0 . 2 f 0.01 0.5 35 to 100 0.065 =t0.01 2 100 t'0 250 0.025 f 0.005 5 Basic Sitrogen Compounds 0.08N perchloric acid in glacial acetic acid 0 . 0 t o 1.0 20.0 i 2.0 0.02 1 . 0 t o 5.0 5.0 f 0 . 5 0.03 1.0 =t0.1 5 t'0 20.0 0.2 20 to 100 0.2 2k 0.01 2 0.05 xb 0.01 4 100 to 250 as determined with a pH meter using glass and calomel electrodes. Frequently, several minutes elapsed before steady potentials were obtained. Table I11 compares this potentiometric method with the proposed, rapid color-indicator method for basic nitrogen. The indicator end points (disappearance of violet) were sharp; they could be matched easily to within 0.015 ml. of 0.08.Yperchloric acid, even in the darkest oils.
.80
3
- 12 -11
-
- IO r-2 W
-9
ff
-8
5
-7
E
e
-6
CONCLUSION
A speciaI titration flask permitted sharp visual color end points with oils of dark color, as in the case of aged distillate fuels or used oils. Where only carboxylic acids are present, pnaphtholbenzein is a suitable indicator. However, where a weakly acidic zinc salt may also be present, an indicator with a higher transition point is preferable. For pyridine-type nitrogen compounds as well as amines, a nonaqueous
2 6 10 14 18 22 M I L L I L I T E R S OF O.IN ALCOHOLIC KOH
Figure 2. Titration curves, ASTM D 664 method A.
Aqueous buffer, potassium hydrogen phthal-
ate
E. Naphthenic acid diluted with oil until acid number was 0.68 mg. of potassium hydroxide per gram C. Oxidized synthetic oil, 10-gram sample, contoining on oil-soluble zinc salt
titration with methyl violet as indicator is well suited. Table IV gives suitable sample sizes to cor’er the range of acids and nitrogen compounds in using the titration flask shoTTn in ~i~~~~ 1, ~h~ flask can be
(2) (3) (4)
used for other color-indicator methods, provided the free space in the neck (about 10 nil.) is not exceeded. LITERATURE CITED
(1) Am. Soc. Testing hIateriale, Standards on Petroleum Products and
(5)
Lubricants, D 974-55T; D 664 (1955). Clark, J. R., ]Tang, S. M., ANAL. CHEW26, 1230 (1954). Deal, V. Z., Weiss, T. T., White, T. T., Zbid., 25,426 (1953). Fanale, D. T., Fricioni, R. B., Hutton, D. IT., Snyder, R. E., Clark, R. O., Twentieth Mid-Year Meeting, Division of Refining, American Petroleum Institute, May 10, 1955, St. Louis, &lo. Fritz, J. S., “Acid-Base Titrations in Nonaqueous Solvents, G. Frederick Smith Chemical Co., Columbus, Ohio, 1952.
(6) Kukin, I., ANAL. CHEM. 29, 461 (1957). (7) Moore, R. T., McCutchan, P., Young, D. A., Zbid., 23, 1639 (1951). (8) Pozefsky, A,, Kukin, I., Zbid., 27, 1466 (1955). (9) Tucker, E. B., Section A, RD VI, ASTM Committee D-2, “Indicators for Neutralization Numbers,” D 974, Houston, Tex., February 1955.
RECEIVED for review December 20, 1956. Accepted February 13, 1958. Seventh Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy, February 28, 1956.
Determination of Trivalent Chromium in the Presence of Chromate R.#W. CLINE’, R. E. SIMMONS, and W. R. ROSSMASSLER Union Carbide Nuclear Co., A Division of Union Carbide Corp., Paducah Planf, Paducah, Ky.
,A method i s given for the determination of small amounts o f trivalent chromium in water containing approximately 300 p.p.m. o f chromate. The chromium(lll) i s separated from chromate b y precipitating with ammonium hydroxide using aluminum hydroxide as a carrier. The chromium(lll) i s then oxidized to chromate with bromine and determined spectrophotometrically. Reproducibility i s good.
A
DEVELOPED, the method determines chromium(II1) in the range from 0 to 0.5 p.p.m., but this range could be extended. The principal problem involves the separation of chromium(II1) from much larger quantities of chromate (300 p.p.m.). The chromium(II1) may then be oxidized and determined colorimetrically by well known methods. The method of separation \vas adapted from an analytical procedure ( J ) , which described the determination of trace quantities of chromium in lithium salts and utilized an aluminum hydroxide carrier for separation of the chromium from the lithium. As adapted by the authors, the separation consists of a double precipitation of chromium hydroxide with an aluminum hydroxide carrier. The hydroxide precipitate is dissolved in sulfuric acid and diluted to 25 ml. with water. Two 10-ml. portions of this solution are then aliquoted. The chromium in one portion is oxidized to chromate with bromihe. The excess bromine is destroyed with phenol. This portion of the
s
Present address, U. S. Army, Fort Detrich, Md.
sample along with the unoxidized portion is then colored with diphenylcarbazide. The chromium content. of each is determined spectrophotometrically a t a wave length of 540 mp by comparison with a standard curve. The difference in chromium content of the two portions is due to the chromium(II1) in the sample. The chromium present as chromate in the unoxidized portion contributes not more than 0.03 p.p.m. to the analysis. This amount is usually insignificant; however, this step serves as a check on the separation efficiency of the analysis. EXPERIMENTAL
A curve conforming to Beer’s law was obtained by carrying standard solutions of chromium(II1) through the entire procedure. The standard solutions contained 0.1, 0.2, 0.3, 0.4, and 0.5 p.p.m. of chromium(II1). Transmittance measurements of the colored solutions were made in 5-cm. cells, using the Beckman spectrophotometer a t a wave length of 540 mp and a slit m-idth of 0.03 mm. Apparatus and Reagents. Color reagent. Dissolve 0.15 gram of diphenylcarbazide in 25 ml. of acetone and dilute t o 50 ml. with water. Prepare daily. Spectrophotometer, Beckman Model DU, %em. cells. Procedure. Pipet a 10-ml. sample into a 15-ml. centrifuge tube. Add 1 ml. of 1% aluminum nitrate solution, and precipitate the aluminum hydroxide with ammonium hydroxide. Centrifuge, and then wash the precipitate with 1 ml. of water. Dissolve the
Table I. Determination o f Chromium(lll)
Cr(II1) Pre‘sent, P.P. M , 0.1 0.2 0.3 0.4
0.5 0.6
Cr(II1) Found, P.P.M. 0.13,O. 16 0.19,0.18,0.22 0.28,0.28,0.28 0.39,0.38,0.40,0.40,0.40 0.53,0.60,0.55
0.57,0.63
precipitate in 4 drops of sulfuric acid, dilute to 10 ml., and reprecipitate as before. Again dissolve the precipitate in acid and dilute t o 25 ml. Take two 10-ml. aliquots of this solution, reserve one, and proceed with the oxidation of the other. Add 30 drops of bromine water. Then fade the bromine color to a faint yellow by adding 10M sodium hydroxide dropwise, finally adding 5 drops in excess ( 3 ) . Add more bromine, if necessary, to obtain the faint yellow color. Heat to 90’ to 95’ C. for 15 minutes, then add 1 to 5 sulfuric acid until the bromine color reappears, adding 2 drops in excess. Boil to remove the bromine. Cool, and add 4 drops of 1 to 5 sulfuric acid and 1drop of phenol. To both the oxidized aliquot and the reserved aliquot, add 1 ml. of the color reagent, dilute t o 25 ml., and determine their transmittance against a water blank a t 540 mp. The difference in two aliquots represents the chromium(111) in the sample. RESULTS A N D DISCUSSION
Although this method was developed for the determination of chromium(II1) in high chromate water, it may be apVOL. 30, NO. 6, JUNE 1958
11 17
