Determination of Sulfur in Volatile Hydrocarbon Mixtures by Lamp

Determination of Sulfur in Volatile Hydrocarbon Mixtures by Lamp-Conductometric Method. E. R. Quiram. Anal. Chem. , 1955, 27 (2), pp 274–277. DOI: 1...
1 downloads 0 Views 482KB Size
ANALYTICAL CHEMISTRY

274 sample of crude progesterone with sodium borohydride. sample contained 7.2y0 3-ketobisnor-4-cholene-22-al.

The

ACKNOWLEDGMENT

The authors wish to thank J. L. Johnson for infrared measurements and interpretations and N. H. Knight for other valuable chemical measurements. LITERATURE CITED (1) Bryant, W. M. D., and Smith, D. AI., J . Am. Chem.SOC.,57, 57 (1935). (2) Him, li,E., and Heyl, F. W., Ibid., 74, 3627 (1952). (3) Heyl, F. W., and Herr, 31. E., Ibid., 72, 2617 (1950).

Heyl, F. W., and Herr, M. E. (to Upjohn Co.), U. S. Patent 2,601,282 (June 24, 1952). Iddles, H. A,, Low, A. W., Rosen, B. D., and Hart, R. T., IND. ENG.CHEX.,ANAL.ED., 11, 102 (1939). Lyttle, D. A., Jensen, E. H., and Struck, W. A, . ~ N A L .CHEM., 24, 1843 (1952).

Mitchell, J., Jr., and Smith, D. AI., Ibid., 22, 746 (1950). Schoniger, W., and Lieb, H., Mikrochemie Ter. Mikrochim. Acta, 38, 165 (1951).

Schoniger, W.,Lieb, H., and Gassner, K., Mikrochin. Acta, 1953, 434.

Shepherd, D. A . , Donia, R. A . . Campbell, J. h.,Johnson, B. A., HOly.32, R. P., Slomp, G., Stafford, J. E., Pederson, R. L , and Ott, .4.C., J . Am. Chem. SOC.,in press. Siggia, S.,and Maxcy, W., ANAL.CHEM.,19, 1023 (1947). RECEIVED for review June 1, 1954. Accepted Soiember 11, 1954

Determination of Sulfur in Volatile Hydrocarbon Mixtures by a lamp-Conductometric Method E. R. QUIRAM Esso Laboratories, Research Division, Standard O i l Development Co., Linden, N . 1, A conductometric method has been developed for the determination of sulfur in the products of combustion resulting from lamp burning of volatile hydrocarbon mixtures. It is simpler than conventional procedures, and has eliminated many of the weaknesses encountered with turbidimetric, nephelometric, volumetric, and gravimetric methods. A new type of absorber has been designed which permits the determination to be completed in one vessel. Means are provided for eliminating errors caused by temperature fluctuations and dissolved carbon dioxide. The method has been found suitable for all ranges of sulfur concentration. The presence of aldehydes, halogens, nitrogen, and other acid-forming substances will interfere.

r

HE lamp method for the determination of sulfur in volatile r h y d r o c a r b o n mixtures has been used for many years in the petroleum industry. The sample is allowed to burn from a wick-fed lamp and the products of combustion are absorbed in an aqueous solution of hydrogen peroxide which oxidizes the oxides of sulfur produced t o sulfuric acid. This acid is then determined either volumetrically ( 1 ) or gravimetrically (2). The volumetric procedure has been used to a greater extent than the gravimetric procedure, because it is simple and rapid. An objectionable feature of the volumetric method is that it is not satisfactory for measuring extremely low concentrations of sulfur. The turbidimetric procedure described by Zahn ( 5 )permits the detection and estimation of as little as 0.0001% of sulfur; however, most refinery control laboratories are reluctant to use turbidimetric or nephelometric methods, because they are too time-consuming and require a considerable amount of shill and care on the part of the analyst. Another feature of the volumetric procedure which has been a problem to refinery laboratories is the manner of predicting within reasonable tolerance the amount of sulfuric acid in the absorbing solution. If the material tested has a high sulfur content and a large sample is burned, the excessive amount of sulfuric acid formed will necessitate a long tedious titration. Brown (3) has overcome this difficulty by measuring the conductivity of the solution during burning. When a predetermined conductivity value is reached, the burning is discontinued. The conductivity reading by Brown's procedure indicates the approximate sulfur content of the sample, which then enables the operator to com-

plete the analysis by conventional methods. Rae ( 4 )investigated the possibility of determining sulfur in isa-octyl alcohol by means of a direct conductivity method, but the results were not entirely satisfactory. He believed that incomplete removal of sulfuric acid from within the fritted disk of the .UT11 absorber was the cause of low results. This paper describes a rapid conductivity method for the detcmnination of sulfur in volatile hydrocarbon mixtures which eliminates many of the weaknesses encountered within turbidimetric, volumetric, and gravimetric procedures. APPARATUS

Manifold System. The manifold system employed has been described in detail (2). A simplified technique for heating carbon dioxide is shown in Figure 1. This type of heating device is desirable when steam is not available. Lamp ilssembly. The lamp assembly is similar to that described in an ASTM method (1). Cotton Wicking. Clean, uniform, unused cotton wicking in 7-inch lengths. Conductance Bridge. Conductivity measurements were taken using the Serfass No. 3965 conductance bridge. Conductivity Cell. A micro dipping type No. 3997 supplied by -4rthur H. Thomas Co., Philadelphia, Pa. This type of cell can be inserted in a cylinder with a 0.5-inch inside diameter. The cell constant is approximately 1.0 reciprocal centimeter. Temperature Bath. Any temperature bath is acceptable, as long as it will maintain a temperature of 25" =t0.5' C. Absorption Assembly, as shown in Figure 2. REAGENTS A 3 D MATERIALS

Carbon Dioxide and Oxygen. Each gas having a purity of not less than 99.575, supplied in metal cylinders. Run a blank determination by bubbling the gases through the standard peroxide solution for a period of 1 hour and then determine the sulfur content by conductivity. Hydrogen Peroxide, 3%. C . P . 30% hydrogen peroxide diluted with distilled water. Twenty milliliters of the diluted hydrogen peroxide solution should not contain over 0.015 mg. of sulfur. Standard Sulfuric Acid Solution. Prepare 0.1249.V sulfuric acid. This solution may be prepared as follows: Pipet 4.1 ml. of concentrated sulfuric acid into a 1-liter volumetric flask and dilute to the mark R'ith distilled water. Standardize this solution against standard sodium hydroxide using methyl red indicator. This standard sulfuric acid solution should be approximately 0.15N. Remove exactly 500 ml. of the standard sulfuric acid and calculate the amount of distilled water which must be added to bring the final normality of the acid to 0.1249,V. Add this calculated amount of water from a buret and mix thoroughly.

V O L U M E 27, NO. 2, F E B R U A R Y 1 9 5 5 G.ASS

ADAPTER

L

HEATIFIG TAPE- I / Z " ' & l D E

ALD AFPRQX 6 FT LONG I15 VOLTS AND 275 W b T T S WRbPPED AROUND S 5 TUBING

1/4'00 STAINLESS S X g L TUBING

CALROO H E A T E R AND S S TUBING

I

LL

VAR14BLE TRANSFORMER

BOTM HEATERS CONNECTED TO ONE CIRCUIT

-

TO A C SUPPLY

.

, , -~-

(9

-0 Mbh'iFEcD

Figure 1.

Details of heating and mixing apparatus

PROCEDURE

Preparation of Conductivity Standardization Curve. One milliliter of the standard 0.1249S sulfuric acid solution is equivalent to 2.0 mg. of sulfur. From aliquots of this solution, other standard solutions are prepared so that 1 ml. of each solution will

r" i

CHIMNEY

1SSEMBLY

Figure 2. Absorption assembly for sulfur by conductivity

contain 0.2, 0.02, and 0.005 mg. of sulfur, respectively. Suitable volumes of these solutions (0.005 to 12.0 mg. of sulfur) are added to 50-ml. glass-stoppered graduates and made up to a volume of 25 ml. with distilled water. If samples of extremely high sulfur content are anticipated, the standardization curve is extended beyond the 12.0-mg. limit. The graduates are stoppered and the contents are thoroughly shaken and the vessels are placed in a constant temperature bath, maintained a t 25" & 0.5" C. The conductivity cell is inserted in the graduate and the specific conductance of the solution is recorded. The concentration of sulfur is plotted against specific conductance of the solution on linear graph paper. Burning Operation. Turn on the heater of the heating apparatus a t least 10 minutes before using carbon dioxide. Record the weight of a clean, dry, empty absorber and bubbler to the nearest 0.05 gram. Pipet approximately 20 ml. of freshly prepared 3% hydrogen peroxide into the absorber. The original absorber and bubbler weight can be recorded on the apparatus, so that subsequent weighings are not neressary. ilttach a chimney to the bubbler and connect it to the manifold by means of rubber tubing. Prepare the burner by threading two strands of cotton wicking through the burner from bottom to top, cut them off evenly, and draw them down flush with the top of the burner. Add from 5 to 10 grams of sample to the flask, insert the wick and burner, and allow the flask to stand until the wick has become primed with the sample. Weigh the assembled flask and burner to the nearest milligram. Connect the side tube of the burner to the burner manifold by means of rubber tubing. Open the primary air needle valve to pass 1.0 to 2.0 cu. feet per hour of combustible gas through the burner. A combustion gas mixture of 65% carbon dioxide and 35% oxygen is suitable for most samples. Open the secondary needle valve to raise the total flow of gas to about 5 cu. feet per hour. Light the burner with a sulfur-free alcohol lamp, insert the burner into the chimney, and fasten it in place with a small spring. Continue the combustion for approximately 1 hour. Cut off the combustion gas supply to the lamp, remove the burner and flask, and immediately reweigh. The difference in weight before and after the burning step represents the weight of the sample. Lift the bubbler attachment slightly and add 4 to 5 ml. of distilled water to the absorbing solution in order to bring the total volume of solution to approximately 25 ml., which is the amount used in making the standardization curve. Weigh absorber plus bubbler and its contents to fO.05 gram. The difference between this weight and the weight of the dry empty absorber plus bubbler indicates the weight of the solution in the absorber. Place the absorber plus bubbler and its contents in a constant

ANALYTICAL CHEMISTRY

276 temperature bath niaintaiiied at 25' C. Bubble air through the :tbsorbing solution for 5 minutes. This air should be scrubbed through 1N sodium hydroxide and bubbled in at a rate of 3 to 4 cu. feet per hour. Insert the conductivity cell into the solution and record the conductivity reading. Whenever a new peroxide solution is prepared or when a new cylinder of carboii dioside or oxygen is used, run a blank determination.

Table 111.

Test

Effect of Dissolved Carbon Dioxide on Conductivity of Peroxide Solution

Condition of Test Solution saturated with carbon dioxide Solution saturated with carbon dioxide Above saturated solution air blown Above saturated solution air blown

NO.

1

2 3 4

Conductivity, hlicromhos 38 37 6.3 8 0

Equivalent Sulfur Content,

11g 0 . 049 0.047

0.009 0.012

CALCULATIONS

The sulfur content of the sample is obtained from the following equation: Per c e n t s = (.\I8 -

M b )

x

1 10

x

w ' 25

1-

I -

x W.

where milligrams oi sulfur corresponding to the conductivity reading of the sample dl* = milligrams of sulfur corresponding to the conductivity reading of the blank = weight of absorbing solution in scrubber after burning = weight of sample in grams

M,

=

w, w,

RESULTS AND DISCUSSION

A dcsirablc feature of the conductivity method is the simplicity and rapidity with which a n analysis can be performrd. When burning in a combustion atmosphere of 65% carbon dioxide and 35% oxygen approximately 3 t o 5 grams of sample can be burned in 1 hour. The ratio of carbon dioxide to oxygen is varied, however, depending upon the components present in the fuel. An aliphatic type sample such as iso-octane will burn Iwst n-hen a ratio of 65% carbon dioxide and 35% oxygen is used, while an aromatic-type sample such as benzene burns best 17 hc>nthe ratio is 77% carbon dioxidr and 23% osygen. The lower tho oxygen content of the combustion mixture, the slower will be the burning rate. The results shown in Tables I and I1 were obtained under varying carbon dioxide to oxygen ratios, and indicate extremely

Comparison of Conductivity and Nephelometric Lamp Sulfur Procedures

Table I.

Sample

Per Cent Sulfur Nephelometric Conductix; 0.0002 0.0002 0.0007 0.0025 0.0039 0.0051 0.0046 0.0055 0.0104 0.011a 0.0107 0.0106 a ASTM cooperative sample, tested by six cooperating laboratories by the proposed ASTM carbon dioxide-oxvgen procedure ($1 T h e average result of t h e laboratories wa5 0 011% of sulfur. SO.

Table 11. Comparison of Conductivity and Gravimetric Lamp Sulfur Procedures Sample ?*TO.

1 2 3 4

5 R

7 8 Q 10 11 12 13

14 15 16 17

Per Cent Sulfur Gravimetric Conductivity 0.023 0.024 0.028 0.029 0.037 0.057 0,068 0,066 0.068 0,073 0.074 0,075 0.077 0.084 0.084 0.089 0.088 0.092 0.087 0.097 0.104 0.097 0.092 0.100 0.090 0.479 0.472

-_0

IO

20

30

4c

T E M P E R A T U R E , *C

Figure 3.

Effect of temperature on conductivity reading

good accuracy when compared with results obtained by conventional nephelometric and gravimetric procedures. Eight detcrininations can be completed in about 3 hours using eight apparatus assemblies, whereas coiiventional methods require a considerably longer time. Xephelometriv and turbidimetric methods may take as long as 2 days. Conductivity methods eliminatr guesdworlr, as they do not require a previous knowledge of the amount of sulfuric acid formed. The method has been found eminrntly suitable for all ranges of sulfur concentration. Absorber. Appreciable errors in conductivity measurements ran be made with solutions having extremely small hydrogen ion concentrations. T o minimize such errors it is desirable to work with a more concentrated aimid solut'on. One means of accomplishing this end is by caollecting the acid gases in an :tbsorber containing a minimum amount of hydrogen peroxide solution. Selecting the propel' type of absorber, however, presented a major problem. With the .\SThZ type absorber ( 1 ) a large volume of solution is required to remove all combustible gases efficiently. Evaporation of the absorbing solution to effect concentration was also considered, but was rejected because inhcrent errors are normally encounterrd when transferring solutions from one vessel to another. Figure 2 shows the type of ahsorber finally select'ed. The advantages of this type of absorber are: it is of simple design and can easily be made by a glassblower, the vessel is light in neight and readily adaptable for weighing on a balance, efficient recovery of combustible gases can be obtained when using a small volume of absorbing solution, means are provided for efficiently removing dissolved carbon dioxide, and conductivity electrodes fit conveniently into the absorption cylinder. Rinsing the chimney is unnecessary. l'vperiments have shown that a quantity of sulfur too sinall to nicitaure is retained in the chimney after combustion of nonleaded fuels. Effect of Dissolved Carbon Dioxide. During the burning operation, the hydrogen peroxide solution becomes saturated with carbon dioxide, which is reflected in a marked increase in conductivity. Since any increase in conductivity resulting from

V O L U M E 27, NO. 2, F E B R U A R Y 1 9 5 5

277

dissolved carbon dioxide will be expressed in terms of sulfur, it is essential that excess c:trt)on dioxide be removed. This is best accomplished by aerating the absorbing solution. Table 111 indicates the noticeable decrease in conductivity that can be obtained when a peroxide solution saturated with carbon dioxide is air blown for 5 minutes. Effect of Temperature. Temperature is an important variable to be considered when making arcurate conductivity measurements. Figure 3 indic:tt,es that :tn error of approximately 2.07; ran he obtained with rJwh l o C. rhange in temperature. Obviously this is an appreci:ihlo source of error; therefore, all readings should be made in ii c:ircsfully controlled t,emperature bath. Interferences. The presence of halogens and nitrogen will interfere with this conduc.tometric procedure. Purified air cannot be used in place of carbon dioside*xygen because oxides of nitrogen will form during the combustion of the sample. The most likely source of these interferences will be found in yasolinw containing tetraethyllead (halogens present in scavenging agents), fuels from shale oil, and samples after hypochlorite treating. Tlie presence of aldehydes, formed b y the incomplete combustion of the gasoline, mill a180 interfere and tend t o inereixe conductivity. This may be caused hy the formation of acids resnlting from the oxidgt,ion of the aldehydcs in a hydrogen peroxide solution.

Ketones, on the other hand, are not easily oxidized and do not seem t,o interfere. The incomplete combustion of the gasoline may be ascertained b y noting the appearance of the flame during the burning operation. If the flame tends to smoke or a change in t.he color of t,he absorbing solution is noted, the sample should be discarded. ACKNOWLEDGMENT

The author wishes to acknon-ledge t,he a b l ~sssistance of P. B. Gerhardt, G. H. Byrd, Jr., and E. R. Hartmann, who made many helpful suggestions and carried oiit much of the experimental work. LITERATURE CITED (1) .Sm. Soc. TesTing Materials. Committee D-2, “Proposed hlethod of Test for Sulfur in Petroleum Products and Liquefied Petroleum Gases by the Carbon Dioxide-Oxygen Lamp Method,”

D 1266-53T. ( 2 ) Am. SOC.Testing Materials, “Sulfur in Petroleum Products by

the Lamp-Gravimetric Method” (Tenhtive) , D 90-50T. (3) Brown, C . W., ASAL. CHEM.,23, 1659 (1951). (4) Rae, J. A., private communication. ( 5 ) Zahn. V., IND. ENG. CHEM.,ANAL.ED.,9, 543 (1937). R H O E I V Efor O review J u n e 8, 1954.

Accepted Sovember 12, 1934.

Volumetric Determination of Phosphorus DAVID M. ZALL, EDWARD WAGMAN’, and N A T H A N INGBER2

U.S.N. Engineering Experiment

Station, Annapolis,

Md.

This work was initiated to find a simple and fast method for the determination of phosphorus in boiler compound. Boiler compound consists of a mixture of sodium carbonate. disodium phosphate, and starch. The specification limits this compound to these three constituents. Practically all of the procedures for the determination of phosphorus depend upon its precipitation by ammonium mol?bdate and the subsequent treatment of the >ellow precipitate. The method described in this paper employs a simplified approach. The phosphate and carbonate are titrated i n the presence of each other using the proper indicator. In cast iron, the phosphorus is first separated from the bulk of the iron with strong alkali. The filtrate containing thc phosphate is neutralized and then titrated with standard alkali.

T

HE procedure deperibrd was developed during a search for

:i

simple method of analysis for boiler compound. Boiler compound is used to treat boiler water and consists of a mixture of sodium carbonate. disodium phosphate, and starch. The methods employed for t8he determination of either one of the constituents when separated are n-ell known and need no elaboration. However, in a mixture as found in boiler compound, thr method for the determination of the constituents is more complicated. Similarly, tlie method for the determination of phosphorus in cast iron and phospho-organic compounds is subject to the same complications. The present method for phosphate determination is long and tedious. It normally requires the init>ial separation of the phosphate as a phospho-molybdate precipitate. The phosphate in the precipitate may be det’ermined either volumetrically or 1 2

Present address, K a i y Department, Bureau of Ships. Washington, D . C. Present address, Army Chemical Center, Edgewood, M d .

gravinirtrically. I n contrast, direct titration of phosphates by either acid or alkali provides accurate results which ran be obtained very rapidly. The exact end points of these reactions and the reversible nature of the monosodium disodium phosphates are well known. For example, the p H of a solution of monosodium phosphate can be calculated from the K I and Kz of phosphoric acid. The proposed method employs a simple direct t’itra,tion of phosphate by either acid or alkali. The results are accurate and rapid. The method is based on the reversibility of the monosodium-disodium phosphate equilibrium which has been discussed ljy Quimby ( 7 ) . Helrich and Rieman (6) have determined phosphorus in phosphate rock by tit’ration to a definite pH. Dah, Ilal’perin, and Kolker ( 2 ) have determined disodium phosphate with the aid of a mixed indicator. Ooudie and Rieman ( 5 ) used ion exchange to isolate the phosphoric acid from phosphate rock. Boos and Conn ( 1 ) used the method of Kolthoff and Cohn for the microdetermination of phosphorus. The phosphate and carbonate in boiler compound are determined on the same sample. The carbonate is titrated with standa,rd acid in the presence of the phosphate. The carbon dioxide formed in the carbonate titration must be eliminated to prevent interference in the subsequent phosphate titration. Dijksman (3) recommended elaborate precautions for elimination of carbon dioxide to prevent int,erference with the titration of phosphate. These precautions apply equally to the determination of phosphorus in cast iron and phospho-organic compounds. RECOMMMEh-DED PROCEDURE

Reagents. Standard sodium hydroxide (carbonate-free) and sulfuric acid solutions (0.LV). Phenolpht,halein indicator ( I % in alcohol) and methyl purple solution (obhainable from Fleisher Chemical Co., Benjamin Franlrlin Station, Washington 4,D. C.). Procedure. Transfer a sample (aliquot) containing no more than 0.15 gram of sodium carbonate and no more than 0.15 gram of disodium phosphate t o a 250-ml. Erlenmeyer flask. If the sample volume is less than 100 ml., di1ut.e with distilled Rater