2,4,6-Tripyridyl-s-triazine as a Reagent for Iron Determination of Iron in Limestone, Silicates, and Refractories PETER F. COLLINS' and HARVEY DIEHL Department of Chemistry, Iowa State University, Ames, Iowa
G. FREDERICK SMITH Department of Chemistry, University of Illinois, Urbana, 111.
b A new reagent, 2,4,6-tripyridyl-striazine (TPTZ), is proposed for the colorimetric determination of iron. The intense violet color formed by TPTZ with iron(l1)is due to the ion Fe(TPTZ),+2, with only two of the three pyridyl groups of each TPTZ ligand attached to the iron atom. In the presence of perchlorate, Fe(TPTZ)*+* can be extracted into nitrobenzene, thus providing a concentration method and a means of removing iron from reagent solutions. The molar extinction coefficient is slightly greater in nitrobenzene (24,100 at 595 mp) than in water (22,600 at 593 mp). The reagent, which is highly specific for iron, has been applied to the determination of iron in silicates, burnt refractories, and limestone following fusion with sodium carbonate-sodium borate in silver crucibles. The results on standard samples are generally lower than the certificate values, possibly because of the elimination of iron from the reagents used.
T
HE colorimetric determination of iron has been greatly facilitated by the advent of the various ferroin reagents, l ,lo-phenanthroline, 2,2:-bipyridine, and analogous compounds, which react with iron(I1) to yield highly colored, water-soluble derivatives. The ferroin reaction is excellent in its sensitivity and selectivity toward iron(I1) and in the stability of the colored derivatives, and has been further improved in the last few years by the introduction of phenyl groups para to the ring nitrogen atoms. A notable example is bathophenanthroline, 4,7diphenyl-1,lO-phenanthroline (1, ?'), which forms an Iron(I1) derivative which has a molar extinction coefficient of 22,400 (11,100 for the 1,lO-phenanthroline iron(I1) derivative) and which can be extracted into isoamyl alcohol. Certain real advantages accrue
Present address, Lithium Corp. of America, 2400 Dakota .4ve., Minneapolis 16, Minn. 1862
0
ANALYTICAL CHEMISTRY
ffom the extraction feature and it is unfortunate that the various phenyl substituted ferroin reagents are difficult to prepare and expensive. A new series of compounds, the amino and Zpyridyl substituted symmetrical triazines, has recently been synthesized by Case and Koft ( 2 ) . The compounds contain the ferroin or terroin functional groups which are responsible for the reaction of the l,l0-phenanthrolines and polypyridines with various metals. The work described here is a n investigation of the reaction of iron(I1) with one of the Case and Koft compounds, 2,4,6tris(Y-pyridyl)-s-triazine (TPTZ), and its application to the determination of iron in siliceous materials.
nium chloride] and 1gram of sodium perchlorate, and extracting the solution with nitrobenzene. FERROIN REAGENTS. The various pyridyltriazines titrated in nonaqueous solution were the original materials of Case and Koft (2). DEIONIZED WATER. Distilled water was passed through a column of Amberlite MB-3 ion exchange resin. Apparatus. SPECTROPHOTOMETRIC MEASUREMENTS. All absorption spectra were recorded using a Cary Model 12 spectrophotometer with 1- or 2-cm. silica cells. All individual absorbancy measurements were obtained using a Beckman Model D U spectrophotometer with 1-cm. Corex cells. PH MEASUREMENTS.A Beckman
REACTION OF IRON(I1) WITH 2,4,6-TRIS(2'-PYRIDYL)-s-TRlAZINE EXPERIMENTAL
Special Reagents. 2,4,6-T~1s(2'-PYRIDYL)-S-TRIAZINE (TPTZ), 0.001M. Initial work was done on a small quantity of the original T P T Z of Case and Koft (2). Subsequently, two preparations were carried out using essentially the Case and Koft procedure with yields of 29 and 11% (melting point 242-3.5" C., reported 244-5' C.). The reagent is available from the G. Frederick Smith Chemical Co., Columbus, Ohio. Solutions of TPTZ, 0.001M, were prepared by dissolving 0.312 gram of the compound in a few drops of hydrochloric acid and diluting to 1 liter with water. IRON STANDARDS. Standard iron solutions were prepared by dissolving electrolytic iron in hydrochloric acid and diluting to the desired volume. IIYDROXYLAMMONIUM CHLORIDE. A 10% solution was prepared by dissolving 100 grams of the salt in 900 ml. of water. Iron was removed by adding 10 ml. of 0.001N TPTZ and 1 gram of sodium perchlorate, and extracting the solution with nitrobenzene. SODIUM ACETATE-.kCETIC ACID BUFFER. A solution 2M in sodium acetate and 2M in acetic acid was prepared by dissolving 164 grams of sodium acetate and 115 ml. of acetic acid in water, and diluting to 1 liter. Iron was removed by adding 10 ml. of 0.00151 TPTZ, 10 ml. of 10% hydroxylanimo-
Model G or Model H-2 pH meter equipped with a glass-calomel electrode system was used for all p H measurements and nonaqueous titrations. The calomel electrode used in the nonaqueous titrations was the sleeve type. Preliminary experiments showed that TPTZ reacts with iron(I1) to form a violet, water-soluble derivative. Attempts to extract this compound into isoamyl alcohol, n-hexyl alcohol, benzene, chloroform, and ethyl acetate in the presence of chloride, iodide, acetate, or perchlorate anions were unsuccessful, but the compound was extractable into nitrobenzene in the presence of perchlorate or iodide. Absorption Spectra. The absorption spectra of the iron(I1)-TPTZ derivative in water and in nitrobenzene are shown in Figure 1. The total iron concentration in each solution was 4.97 x lO-SM, and 1-cm. cells were used. To obtain the absorption spectrum in nitrobenzene, an aqueous solution of the compound was prepared. perchloratr added, and the solution extracted with nitrobenzene. The extract was then diluted to the desired volume with ethyl alcohol. Sensitivity and Stability. A series of solutions containing excess T P T Z and an i r o n concentration varying from 0 t o 6 X 10-6M was prepared by pipetting aliquots of a 1.000 X 10-4M iron solution into 50-ml. volumet'ric
Table. 1. Absorbancy of Fe(TPTZ)*+* as a Function of Concentration
,-\
Figure 1. Absorption spectra of Fe(TPTZ)*
-. ---.
,o
,
-
\
A ueous
Aqueous solutions
Nlkobenzene
450
flasks and adding 5.0 to 10.0 ml. of 0.001M T P T Z to each solution. The molar ratio of TPTZ t o iron was greater than three in each solution. Two milliliters of 10% hydroxylammonium chloride was added to reduce any iron (111) and each solution was buffered with 10 ml. of sodium aretate-acetic acid solution. The absorbancy of each solution was determined at 593 mp using 1-cm. cells. After 32 hours, the absorbancy of each solution was again measured and no significant cl- znge observed. From the data given in Table I and the method of least squares, the molar extinction coefficient WBB 22,600.
T o determine the molar extinction coefficient of the perchlorate salt in nitrobenzene, solutions containing known amounts of iron, excess TPTZ, hydroxylammonium chloride, sodium perchlorate, and acetate buffer were prepared. Each solution was extracted three times with nitrobenzene and the extracts were combined and diluted to the desired volume with ethyl alcohol. The absorbancy of each nitrobenzene solution was then determined a t 595 mp using 1-cm. cells. After 12 hours, the absorbancy of each solution was again determined and no significant change observed. From the data given in Table I and the method of least squares, the molar extinction coefficient was 24,100. Effect of pH. Solutions 2.00 X 10-5M in iron and containing excess TPTZ, hydroxylammonium chloride, and varying amounts of hydrochloric acid or ammonium hydroxide were prepared. The p H of each solution was measured and its absorbancy determined at 593 m p using 1-cm. cells. To determine thc: effect of pH on extraction, a series of aqueous solutions of the iron-TPTZ complex was prepared. The p H of each solution was adjusted to some definite value with hydrochloric acid or ammonium hydroxide, and the solution was then extracted three times with nitrobenzene. The combined extracts were diluted to the desired volume with ethyl alcohol and the absorbancy was determined at 595 mp. The data obtained are shown eraDhicallv in Figure 2. Nature of IronlII) Derivative of TPTZ. SPECTROPHOTOMETRIC TITRATION. A spectrophotometric titration of T P T Z with iron(I1) was "
I
500
700
600 WAVE LENGTH (MILLIMICRONS)
2 0
30
40
50
60
SoPution Fe concn., AbsorbM X ancy at 1W 593 mp 0.20 0,043 1.00 0.219 2.00 0.440 3.00 0.665 1.00 0.889 1.111 5.00 6.00 1.333 6 = 22,600
Nitrobenzene Solution Fe concn., AbsorbM X ancy a t 1W 595 mp 0.80 0.189 1.60 0.389 2.40 0.577 3.20 0.769 4.00 0.967 6.00 1.447 e =
24,100
70
PH
Figure 2. Effect of pH on formation of Fe(TPTZ)*
-. ---.
Aqueous solution Extraction into nitrobs;.zenc
MOLE FRACTION
OF T P T Z
Figure 4. Continuous variations study of iron(l1)-TPTZ system
ML
OF 2
50
X
10.'
M IRON
Figure 3. Spectrophotometric titration of TPTZ with iron(l1)
carried out to determine the combining ratio of the reactants. Solutions containing 5.00 ml. of 1.OOO X 10-3M TPTZ, 1.0 ml. of 10% hydroxylammonium chloride, 5.0 ml. of 10% sodium acetate, and varying quantities of 2.50 X lO-'M iron solution were prepared in 50-ml. volumetric flasks. After dilution to volume, absorption spectra for the solutions were recorded and the absorbancy values a t 593 mp were used for the titration curve shown in Figure 3. The union of TPTZ and iron(I1) takes place in the ratio of two to one. CONTINUOUS VARIATIONS STUDY.The combining ratio of TPTZ and iron(I1) was also determined by the method of continuous variations. A series of solutions was prepared in which the con-
centration of TPTZ was varied from 0 to 2 X 10-4M. The molar concentretion of iron(I1) plus TPTZ was also held constant a t 2 X 10-4. Hydroxylammonium chloride was added to reduce any iron(III), and sodium acetate served as a buffer. Absorption spectra were recorded for the various solutions and the values at 593 mp were used for the plot of ahmrbancy us. mole fraction of TPTZ (Figure 4). The peak which occurs at 67 mole yo of TPTZ indicates a combining ratio of TPTZ with iron(I1) of two to one. PREPARATION A N D ASaYiIS OF IRON(II)-TPTZ-TODIDE COMPOCSD.Equivalent amounts of TPTZ 10.001 mole) and ferrous ammonium sulfate (0.0005 mole) were dissolved in 300 ml. of watcr containing 1 gram of hydroxylammonium chloride and 1 mi. of hydrochloric acid by heating. The solution was then cooled and ammonium hydroxide added to adjust the p H to 4.8. Unreacted TPTZ was removed by filtration and 10 grams of potassium iodide added to the heated filtrate. On cooling, thr, iron(11)-TPTZ-iodide compound wab precipitated as small crystals. These crystals were removed by filtration, dried, and extracted with ethyl ether in a Soxhlet extractor for 30 hours to reniove any potassium iodide. T h r compound, which is not soluble in ethyl rthcr, !vas then dried a t 115' C. for 6 hours. The compound was analyzed for iron colorimetrically using I , I 0-phenanVOL. 31, NO. 1 1 , NOVEMBER 1959
* 1863
Table 11.
Nonaqueous Titrations
Compound Titrated 1,10-Phenanthroline
Solvent Acetonitrile
TI%(1,1O-~henanthroline)iron(II1 perchlorate
of Various
Acetonitrile
."
Acetic acid-nitromethane (1:lO)
2-Amino-l,6bis(2'-pyridyl>
Acetic acid-nitromethane (1:12)
2-Amino-4,6bis( 4'-ethyl2'-pyridyl )+triazine
Acetic acid-nitromethane (1:12)
ZAmino-l,6bis(4'-phenyl2 '-pyridyl )+triazine
Acetic acid-nitromethane (5:2)
2,4,6Tris(2'-pyridyl)-~-tri-
Nitromethane
s-triazine
azine
Bia [2,4,6tns(2'-pyridyl> s-triazine ]iron(11) perchlorate
Table 111.
Nitromethane
+f
CO + f
4.8
2.7 5.3 10.6 99.4 110 10.4
Ni +*
Zn Mn +-I Cr +a +f
20.8
73.0 100 100
*$a:
Mg+Y Ca Sr +-I Ba +) Cd +*
100
99 101
100 100 100
m+:' Bi
Sn +I Pb +a Th +4
100
uo, +a 1,i + K+
NH4 Na + AR CN PO1-'
+
+
FCZHSOZBrINOS-
NOzSO,-'
Clod(310,-
s,os - 2 SCNszos --I
BO?-8 BrOJ-
MOO&-'
1864
One end mint correamnding to -1 equivalen't per mole One end point correaponding to 2 equivalents per mole One end point corresponding to 2 equivalents per mole One end point corresponding to 2 equivalents per mole Two end points, the first corresponding to 1equivdent per mole, the second to a totalof 3 eouivalents per mole One end point corresponding to 2 equivalents per mole
Effect of Various Ions on Color Formation
Concn., P.P.M. 1.3 2.5 6.3 1.2 2.4
Ion cu
€&?Enllt
One end point correaponding to 1 equivalent per mole No end mint
.
4,6Diamino-2(2'-~vridvlb . - .
s-triazine
Ferroin Reagents
101 120 115 1,020 1,067 1,033 5,600 102 500 528 502 14,400 556 497 504 500 512 524 548 538 507 528 545 499 34
ANALYTICAL CHEMISTRY
Ni( C104)t Ni( C104), Ni( ClO,), ZnClt MnSO, K~Cr207-SO* KICr20,-S@t Be(ClO')-I AlCl, MgSO4 CaCOa-HCl Sr(C101)2
Relative Error, oI +0.7" +1.4 +4.8 +0.9 +1.8 +3.6 1-0.4 +1.5 +2.7 +0.2 +0.2 +0.6 +2.4 0.0 0.0 0.0 0.0
+0.2 +0.2 -0.7 Ppt.
Ppt. -0.2
+0.2
+0.2 +0.4
0.0
+0.4
0.0 -0.2
Ppt. Very large +0.2 +0.2 -0.2
+0.2
0.0
+0.2 Large
0.0 0.0
+0.2
0.0
+0.2 +0.4
0.0 0.0
Very large
TPTZ Added, Moles x 1 W 0.5 0.7 0.9 0.5 0.7 1.1
0.8 1.3
2.3 10.0 2.0 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 5.0 0.5 0.5 0.5 2.0 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5
0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5
2.0
throline after wet ashing with nitric and perchloric acids. The TPTZ content was determined by adding exceas iron(I1) to a solution of the compound and then measuring the absorbancy at 593 mp. T o obtain the calibration curve, a series of standards containing known amounts of TPTZ and excess iron(I1) was prepared and the absorbancy of each solution determined at 593 mp. Iodide was determined by passing a solution of the compound through a column of Amberlite IR-120 cation exchange resin in the hydrogen form to give a colorless solution containing hydriodic acid. This solution was neutralized with dilute sodium hydroxide and titrated with silver nitrate using eosin as the indicator. (Found: Fe, 6.11; TPTZ, 66.5; I, 26.95%. Calculated for Fe(TFTZ)&: Fe. 5.98: TPTZ. 66.86: I. 27.17%) NONAQUEOUS 'TITRATION
OF
TPTz,
IRON(II) DERIVATIVE, AND RELATED COMPOUNDS. Nonaqueous titrations of various ferroin reagents and iron(I1) derivatives 'were carried out to gain some insight into the structure of-the iron(I1) derivative of TPTZ. The titrant' was 0.00968M perchloric acid (standardized against potsssium acid phthalate) in glacis1 acetic acid to which a small quantity of acetic anhydride had been added. Nitromethane, acetonitrile, and acetic acid were used as solvents. The bis [2,4,&tris(2'-pyridyl) s-triazine]iron(II) perchlorate was prepared by passing a solution of the iodide salt previously described through a column of Amberlite IRA400 in the perchlorate form and evaporating the solution so obtained to dryness. The crystals obtained were dried at 115" C. for 24 hours. The compounds titrated, solvents used, and results are given in Table 11. Study of Interferences. Various ions were tested for possible interference by preparing a series of solutions in the following way:
A 10.00-ml. portion of a solution 1.o00 x 10-'M in iron was pipetted into a 50-ml. volumetric flask containing a solution of the ion to be tested. T P T Z was added in various amounts (Table 111), followed by 1.0 ml. of 10% hydroxylammonium chloride and 5.0 ml. of 20% sodium acetate. The solution was then diluted to volume and its absorbancy determined a t 593 mp. The solutions containing the various ions to be tested were prepared from the reagent grade compounds listed in Table 111. In testing the possible interference of silver and chloride ions, hydroxylammonium sulfate WAS substituted for the chloride salt. T o test the possible interference of sodium or acetate ions, a solution containing sodium acetate was compared with a solution containing no sodium acetate, the pH being adjusted with dilute ammonium hydroxide. RESULTS AND DISCUSSION
The molar extinction coefficients for
the iron(I1) derivative of TPTZ in aqueous and nitrobenzene solutions a t the wave length of maximum absorption were 22,600 at 593 mp and 24,100 a t 595 mp, respectively; thus, even without considering the concentration effected by extraction, there is some gain in sensitivity by making the spectrophotometric measurement in nitrobenzene solution. The color conforms to Beer’s law in both water and nitrobenzene solution. No fading is observed for at least 32 hours in aqueous solution and 12 hours in nitrobenzene. The p H range over which the color is completely formed is from 3.4 to 5.8 in aqueous solution. With extraction into nitrobenzene the p H range is slightly greater, 2.7 to 7.0. While this range is not as wide as for many of the 1,lO-phenanthroline and polypyridine reagents, a suitable p H can be eaaily obtained by use of an acetate buffer. A combining ratio of two TPTZ to one iron(I1) is shown by both the spectrophotometric titration and the continuous variations study. The empirical formula calculated from the analysis of the iodide salt of the complex is Fe (TPTZ)1. aaI1 .M. In the titration of the amino and pyridyl substituted symmetrical triazines, the number of equivalents of acid consumed corresponds to the number of pyridyl groups in every case. Therefore, it appears reasonable to attribute the basicity of the compounds to the pyridyl nitrogens. I n the case of 1,lO-phenanthroline, when the nitrogen atoms become bonded to iron(II), they can no longer be titrated. If this is also true with the symmetrical amino a n d pyridyl substituted triazines, it follows from the results presented in Table I1 that, in the iron(I1) derivative of TPTZ, two pyridyl groups of each ligand are united to the iron atom, as only one of the three can be titrated. On the basis of this evidence, it is evident that the ion is Fe(TPTZ)*+l and that TPTZ acts aa a tridentate ligand, presumably with two pyridyl nitrogen and one ring nitrogen attached to the iron atom and the third pyridyl group free. It is generally assumed that the organic molecule is planar in the ferroin metal compounds and, therefore, in this ion the two TPTZ ligands must lie a t right angles to each other, the nitrogen atoms of the respective ligands occcupying the 2, 3, and 4, and 1, 5, and 6 positions.
2 FS++
TPTZ is very sensitive toward iron and possesses many of the desirable properties of a colorhetrio reagent. The extractability of the iron(I1) derivative into nitrobenzene is advantageous because it provides a simple way to remove iron from various reagent solutions and serves m a concentration step, thus making possible a reliable determination of extremely small amounts of iron. As TPTZ is also relatively simple to prepare, it may become a preferred reagent for the colorimetric determination of iron. The results obtained in the study of interferences are shown in Table 1x1. The method of calculating relative error is essentially the same m that proposed by Fortune and Mellon (4). Of the ions tested only CI~+*,C O + ~ ,Ni+’, C F , Ag+, Hg+*, Bi+a, CN-, CIOI-~,and N02- interfere significantly. The interference of Co+2, Cu+*, and Ni+’ is due to the formation of colored compounds with TPTZ; however, 2.5 p.p.m. of Cuf2, 2.4 p.p.m. of Co+’, or 5.3 p.p.m. of Ni+2results in a relative error of less than 2% in the determination of iron. A precipitate is formed in the presence of Ag+, Hg+2, and Bi+a, and the other ions retard color development or interfere because of the color of the ion. I n the presence of most of the transition metals, the color development is retarded if an excess of TPTZ is not present. Several applications of this reaction to the colorimetric determination of iron have been studied. One of these, the determination of iron in silicates, is described below. Investigations of other applications will be published later.
All of the work reported here was done with silver crucibles. EXPERIMENTAL
Silver Crucibles. These were made by electroplating silver on 20- or 30-ml. nickel crucibles using 30 grams of silver nitrate, 36 grams of sodium cyanide, and 112 grams of potassium nitrate per liter of solution. The crucibles were rotated during the plating and a current of 0.5 ampere was used. Silver anodes in the form of rods were cast from silver which had been purified previously by electrolytic deposition as loose dendritic crystals from a neutral silver nitrate solution. A f t q 80. to 100 grams of silver had been deposited, the nickel crucible was dissolved away with hydrochloric acid. Several fusions of a mixture of sodium carbonate and sodium borate in the silver crucibles reduced the blank for iron determinations tc a negligible value. Recommended Procedures. BURNT REFRACTORIES, SILICATES, ARGILLACEOUS LIMESTONE. Weigh a samrle
containing 3 to 5 mg. of ferric oxide into a silver crucible and add 1.0 gram each of sodium carbonate and sodium borate decahydrate. Mix thoroughly and gently heat the crucible and contents over a Meker burner until the water from tht, sodium borate has been vaporized. Gradually increase the heat to-melt thr 0ux and continue heating until the sample is completely decomposed. ROtate the crucible while cooling to cause the melt to solidify on the sides of the crucible. When the crucible has cooled to room temperature, add 10 ml. of water and 5 ml. of hydrochloric acid, cover with a watch glass, and gently heat on a hot plate until the residue has
DETERMINATION OF IRON IN LIMESTONE, SILICATES, AND REFRACTORIES One of the main difficulties in determining the iron content of siliceous materials is complete decomposition of the sample without loss of iron. Treatment with hydrofluoric acid often leaves a residue that is insoluble in hydrochloric acid so that a fusion of the sample may be required. The loss of iron during fusions in platinum crucibles has been studied by Shell (6). During the fusion, iron is evidently reduced and it alloys with the platinum. This alloyed iron is difficult to recover completely and affects subsequent analyses inasmuch as it may be partially released a t a later time. To overcome this absorption of iron, Shell suggested the use of silver crucibles, with the flux consisting of a mixture of equal amounts of sodium carbonate and sodium borate. I n preliminary experiments involving sodium carbonate fusions in platinum crucibles, serious losses were encountered in some analyses. These losses were traced to the crucibles, confirming the observations of Shell and others.
dissolved. If necessary, add more hydrochloric acid to obtain complete dissolution of the residu? (a precipitate of silver chloride and silica will remain which is later removed). Transfer the contents of the crucible to a 250ml. volumetric flask and dilute to volume. Mix well and filter or centrifuge a portion of the solution to remove any suspended silver chloride and silica. Pipet a 5.00-ml. aliquot of this solution into a 50-ml. volumetric flask, and add 2.0 ml. of 10% hydroxylammonium chloride, 5.0 ml. of 0.001hf TPTZ, and 10 ml. of a 2M sodium acetate-2M acetic acid buffer. Dilute the solution to volume and determine the absorbancy at 593 mp using 1-cm. cells. Run a blank on the reagents and silver crucible in exactly the same manner. SILICATE, LOW-IRON(G-1 granite). Fuse a 3.0-gram samDle with 5.0 grams of sodium carbonate’and 5.0 graks of sodium borate decahydrate in a 50ml. silver crucible, and continue heating until a clear melt is obtained. Rotate the crucible while cooling and place it in VOL. 31, NO. 1 1 , NOVEMBER 1959
1865
Table IV.
Results on Determination of Iron in Siliceous Materials Sample No. and Re td Av. Ran e of Reptd. Description FerOl Found, % 6aGe &slues
NBS 76 burnt refractory NBS 77 burnt refractory
2 . 1 1 , 2.11, 2 . i i , 2 . 1 2 , 2.08, Av. 2.11 0.n 8 2 , 0 . 8 1 , 0.82, Av.
2.38
2.22 to 2 . 5 0
0.90
0 . 7 9 to 1.39
0 . 7 1 , 0 . 7 1 , 0 . 7 0 , Av. 0.71 1.59, 1.57, Av. 1.58
0.79
0 . 7 0 to 1.17
1.63
1.57 to 1.69
0.083, 0.083, 0.084, Av. 0.083 0 . 9 2 , 0.92, 0 . 9 3 , 0 . 9 2 ,
0.084
0.082 to 0.086
0.98
0 . 9 2 +a1.01
2.05
2 . 0 0 to 2.11
0.073
0.067to0.077
0.081
0.070 to 0.095
0.076
0 , 0 7 to 0.078
Seetext
1.29t02.99
See text
10.70to 12.19
0"
U.BL
NBS 78 burnt refractory NBS l a argillaceous limestone XBS 88 dolomite
NBS 97 flint clay NBS 98 plastic clay NBS 81 glass sand NBS 91 opal glass
A ~ n. 02 1 . 9 7 , 1.97, 1.99, 1.97, 2 . 0 0 , Av. 1.98 0.074, 0.076, 0.075, 0.074, Av. 0.075 0.073, 0.074, 0.073, 0.074. 0.073. 0.076. Av. 0.'074 0.078, 0.078, 0.079, Av. 0.078 1.85, 1 . 8 5 1 1 . 8 4 , 1.85, Av. 1.85 10.91, 10.94, 10.87, Av. 10.91 '
NBS 93 borosilicate glass G-1 granite W-1 diabaae
a 600-ml. beaker Add 100 ml. of hydrochloric acid and 200 ml. of water, and heat until the residue is completely dissolved. Cool, remove the crucible with washing, and dilute the solution to exactly 1 liter in a volumetric flask. Pipet a 25.0-ml. aliquot of this solution into a 250-ml. beaker, add 5 ml. of hydrochloric acid, and heat for several hours t o precipitate silica. Cool the solution, transfer to a 250-ml. volumetric flask, and dilute to volume. Filter a portion of this solution (no washing) to remove silica and silver chloride. Pipet a 15.00-ml. aliquot of the filtered solution into a 50-ml. volumetric flask and complete the determination as directed in the last paragraph of the procedure for burnt refractories. SILICATE, HIGH-IRON (W-1 diabase). Mix 0.22 gram of the sample with 1.0 gram of sodium carbonate and 1.0 gram of sodium borate decahydrate in a silver crucible and fuse. Continue heating until the decomposition of the sample is complete (about 15 minutes), cool, and add 20 ml. of water and 10 ml. of hydrochloric acid. HeaD the crucible on a hot plate until the residue is completely dissolved (a residue of silica and silver chloride will remain), cool, and dilute to exactly 500 ml. in a volumetric flask. Pipet a 50.0-ml. aliquot of this solution into a 500-ml. volumetric flask, dilute to volume, and centrifuge a portion of the final solution to remove silica and silver chloride. Place 15.00 ml. of this solution in a 50ml. volumetric flask, add 2.0 ml. of 10% hydroxylammonium chloride, 5.0 ml. of 0.001M TPTZ, and 10.0 ml. of acetate buffer, and dilute to volume. Mix well and determine the absorbancy of the solution a t 593 mw. Run a blank through the entire procedure. All volumetric flasks and pipets used in this determination were calibrated. GLASS AND GLASS SAND. Weigh a sample containing 0.5 to 1.0 mg. of ferric oxide into an iron-free platinum crucible. (Iron may be removed from a 1866
a
4NALYTICAL CHEMISTRY
platinum crucible by repeated heating t o 1OOO' to 1200' C. in a muffle furnace and leaching with hot hydrochloric acid.) Add 2 ml. of water and 4 ml. of hydrofluoric acid if the sample is glass, or 4 ml. of hydrofluoric acid if the sample is glass sand. After the reaction has subsided, add 1 ml. of perchloric acid and evaporate to dryness on a hot plate without boiling. Cool, add 2 ml. of hydrofluoric acid, and again evaporate to dryness. Place the crucible and contents in a 250ml. beaker, add 20 ml. of hydrochloric acid and 50 ml. of water, and heat. If complete solution is obtained. cool, transfer the solution to a 250-ml. volumetric flask, dilute to volume, and continue the determination as directed in the following paragraph. If an insoluble residue remains, filter the solution into a 250-ml. volumetric flask using medium-porosity filter paper. After washing first with dilute hydrochloric acid (1 to 100) and finally with water, ash the filter in a silver crucible. Add 1 gram of sodium carbonate and 1 gram of sodium borate decahydrate and heat until a clear melt is obtained. Cool to room temperature, add 5 ml. of hydrochloric acid and 10 ml. of water, cover with a watch glass and heat on a hot plate until the residue dissolves. 9 precipitate of silver chloride and silica will remain. Transfer the contents of the crucible to the 250-ml. volumetric flask containing the filtrate, and dilute to volume, After miuing, centrifuge or filter a portion of this solution to remove any silica and silver chloride. Pipet a 25.0-ml. aliquot of the solution into a 50-ml. volumetric flask and add 2.0 ml. of 10% hydroxylammonium chloride and 5.0 ml. of 0.001M TPTZ. Add ammonium hydroxide dropwise until the violet color of the iron derivative remains on mixing, add 10 ml. of the 2M sodium acetate-2.V acetic acid buffer, and dilute to volume. Determine the absorbancy at 593 m9 using 1-cm. cells. Run a blank on reagents
and crucible in exactly the same manner. LIMESTONE.Weigh a sample of limestone containing 0.5 to 1.0 mg. of ferric oxide into a 250-ml. beaker. Cover with a watch glass, add 20 ml. of n-ater and 10 ml. of hydrochloric aoid, and heat gently. After the reaction is completed, filter the solution into a 250-ml. volumetric flask using medium-porosity paper. After washing the filter with dilute hydrochloric acid (1 to 100) and water, place it in a silver crucible, ash, cool, and add 1.0 gram each of sodium carbonate and of sodium borate. Heat gently at first and then more strongly to melt the flux and decompose the residue. Rotate the crucible while cooling to cause the melt t o solidify on the sides of the crucible. After cooling, add 10 ml. of water and 5 ml. of hydrochloric acid, and warm to dissolve the residue. After complete dissolution (a precipitate of silica and silver chloride will remain), transfer the contents of the crucible to the 250-nil volumetric flask containing the original filtrate. Dilute the solution to volume, filter or centrifuge a portion of the solution, and conclude the determination exactly as directed in the last paragraph in the procedure for the determination of iron in glass and glass sand. RESULTS A N D DISCUSSION
Various National Bureau of Standards samples and two rock samples (G-1 granite and W-1 diabase) from the U. S. Geological Survey were analyzed and the results are shown in Table IV. No definite values for the iron content of the granite and diabase have as yet been established. In the preliminary report published in 1951 (3).the averages and standard deviations of the results from 24 laboratories are: G-1 granite, % Fe = 1.44, u = 0.231; 11'-1 diabase, yo Fe = 7.88, u = 0.18. These values for the iron content expressed as Yo FenO, are 2.06 and 11.27. Later work reported by Goldich and Oslund (6) gave a n average value of 1.S6% Fe208for the granite and 11.09% FeaOl for the diabase. Preliminary analyses of the granite were carried out using 150-nig. samples (Fez08found: 1.70, 1.93, 1.73, 1.82y0). The poor precision was attributed to inhomogeneity of the sample inasmuch as much more precise results had been obtained by a similar procedure on Bureau of Standards samples of burnt refractories, etc. On examination of the material, it was noted that black particles, which are attracted to a magnet, are scattered throughout the sample. Presumably these particles are magnetite and the distribution through the mass is not absolutely uniform. I n subsequent analyses of the granite, larger samples (3 grams) were taken; the results are given in Table IV. For many of the samplea analyzed, the iron content as determined by the various TPTZ procedures is lower than
the average reported by the bureau. The possible presence of iron in the precipitate of silica and silver chloride RBS checked in the case of the NBS 76 burnt refractory and no significant amount of iron was found. The range of values obtained for each sample by the various analyses as shown in the certificates is often quite large, indicating some inherent difficulty in older analyses. l h e procedures usually employed for dctermining the iron content of silicates involve lengthy separations and the use of re1at)ively large amounts of reagents and reagent grade chemicals which frequently contain appreciable amounts of iron. I n the proposed TPTZ methods, most of the reagents can be easily freed of iron by virtue of
the extractability of the iron derivative of TPTZ into nitrobenzene. This reduces the blank to almost zero. These methods also have the advantage that none of the tedious separations usually encountered in the analysis of silicates are necessary. ACKNOWLEDGMENT
The authors wish to express their appreciation to Francis Case of Temple University for the various pyridyltriazine compounds used in this work. One of the authors, Peter Collins, wishes to express his appreciation to the Armstrong Cork Co. for financial assistance in the form of a fellowship.
LITERATURE CITED
(1) Case, F. H., J . Org. C h . 16, 1541 (1951). (2) Case, F. H., Koft, E., J . Am. Chem. SOC.81,905 (1959). (3) Fairbairn, H. W., et al., “Cooperative Investigation of Precision and Accuracy
In Chemical, Spectrochemical, and Modal Analysis of Silicate Rocka,” U. S. Geol. Surve Bull. No. 980 (1951). (4) Fortune, W. &., Mellon M. G., I N D . ENG. CHEM., ANAL.d D . 10, 60 (1938). \ - - - - I .
(5) Goldich, S. S., Oslund, E. H., Bull. Geol. Sac. Am. 67,811 (1956). (6) Shell, H. R., ANAL. CHEM.22, 326 (1950). (7) Smith, G. F., McCurdy, W. H., Jr., Diehl, Harvey, Analyst 77, 418 (1952).
RECEIVEDfor review July 14, 1959. Accepted September 4, 1959.
Zirconium Analysis by Production Control Quantometer C. L. EASTERDAY Mallory-Sharon Metals Corp., Ashtabula, Ohio
b A direct-reading spectrograph was used for a routine point-to-plane analysis of zirconium. The principles involved and a description of the preparation of standards are given. The results obtained are equal to, or more accurate than, those obtained by carrier-distillation methods. The same techniques, with appropriate modifications, should find application in the onalysis of zirconium alloys.
D
IREcT-reading niethods are estensively applied in the steel and aluminum industries, but very little application is evidenced in the zirconium i d u s t r y , because of lack of suitable methods and standards. Spect,rographic methods using photographic tcchniques have had a preclomiriant application in the analysis of higli purity zirconium. This paper d&tils the use of a direct-reading spectrograph for a point-to-plane analysis of zirconium. This technique greatly reduced or c,liminated photographic prcrcwsing, line density measurements, and smiple preparation. The man-hour savings per sample by this technique w r e approximately 30 minutes-consitterably greater when several were run a t a time. Each direct-reading instrument is designed for a specific type of alloy or other ninterial, usually to meet recurrent analytical requircmcnts. The Quantometer used was acquired for production control in the manufacturing of zirconium. The specifications upon which i t is designed covers the determination of aiuminum, iron, cobalt, titanium, tin.
vanadium, nickel, silicon, magnesium, manganese, boron, chromium, copper, cadmium, and calcium. I n designing a given instrument, the selection of spectral lines for the desired elements depends on several factors, including range of concentration and interferences. The selected spectral lines and ranges covered are listed in Table I. The instrument was first set up for the analysis of zirconium using the method that is utilized in most laboratories. It is based on converting the metal to the oxide, mixing the oxide with a carrier, and igniting it in a direct-current arc. I n working with the point-toplane technique it was found that niost lines used in the oxide analysis were also usable with this method. I n some instances other lines might be better for the spark proceduw, but it was necessary to mc t h r lines available in the instrument. APPARATUS
The instrument used for this work was a 2-meter production control Quantometer (Applied Research Laboratories, Glendale, Calif.). The operating conditions are listed in Table 11. PREPARATION OF STANDARDS
Samples of zirconium spongc n r r e chosen to give as wide a concentration range as possible for all the elenicnts listed in Table I. When sufficient ranges of concentfation were not available in the sponge, powdcred nwtal was blended with the sponge to give the desired concentration. This material was then compacted into 100-gram
Table 1.
Impurity A1 Fe
co Ti Sn
v
Ni
Si
B Cr cu Cd
Spectral Lines and Ranges
Concentration Range, P.P.M.
Line 3914 __
n.
2599.4 3453.5 x 3349.4 x 3262.3 X 3184.0 X 3002.5 X 2881.6 X 2795.5 X 2576.1 X 2497.7 X 4254.3 3247.5 2288.0
2 2 2 2 2
2 2 2 2
20-250 200-2000 5-100 20-100 1w20 20-50 20-1000 20-600 10-40
5-25 0.14.8 20-1500 5-8000 0.14.5
Table II. Operating Conditions Upper - - electrode, Sample anode Lower electrode, Graphite rod: ‘(4inch with 120” tlp cathode Positive Sample polarity Analytical gap, 2 mm. Excitation source 30-cycle overdamped 5 Pre-exposure, seconds 50 Exposure, seconds Output voltage, 940 Continuous, interDischarge rupted Inductance, ph. 50 Resistance, ohms 200 Capacitance, pf. 20
compacts and triple melted in a small laboratory button melter. The resulta n t buttons were approximately 1.5 inches in diameter and 0.5 inch thick. Chips from the standard buttons u’crc’ analyzed chemically for iron, chromium, VOL. 31, NO. 1 1, NOVEMBER 1959
1867
