2 - C~LORON11R08tNISHt
/
3
- CHLORON11ROBSNZlNI
MIXIUPS
c
n
wzo13n'
1-1
1 1
lH201NO~ iH,O12HA,
INITIAL
I
6
10
I
v
(
~~
18
strument was also studied. A 50:50 mixture of 2- and 3chloronitrobenzene was scanned in the negative mode; the results are illustrated in Figure 5. These show the initial predominance of the 3-chloronitrobenzene isomer ionic 2 indicates each isomer has a characteristic and different pattern which is replaced by that of the less volatile 2pattern. chloronitrobenzene. This technique may thus prove useful All these compounds produce strong positive mobility in detecting an impurity of one isomer in a sample of the spectra, reacting with the ( H z O ) ~ H +and ( H z O ) ~ N O + other. reactant ions to form protonated ions, MH+ , and additive ions (M NO)+ (12). The ;educed mobility data listed in CONCLUSIONS Table I for these MH+ ions show the correspondence of Data obtained with PC show that halonitrobenzene their K O values to those of the respective negative molecucompounds can undergo simple electron attachment, dislar ions. sociative capture, or both, and that the reactions are a Figure 4 shows the uniformity of the mobility spectra of function of temperature. These qualitative studies reveal the pure compound after a sample of 2-chloronitrobenzene the various reactions involved for each compound after had been introduced and allowed to deplete slowly in conelectron attachment, and the possibility of using such difcentration. This constancy of pattern with changing conferences as an identification method for isomers. centration appears to be characteristic of the mobility
+
spectra of a pure compound. Any impurities present usually give peaks which either disappear quickly or shift in mobility with successive spectral scans. The effect of introducing a simple mixture into the in(12) F W. Karasek and D. W Denny, Anal. Chern. 46,633 (1974)
Received for review July 9, 1973. Accepted January 14, 1974. The research for this paper was supported by the Defence Research Board of Canada, Grant Number 9530116, and the National Research Council of Canada, Grant Number A5433.
Direct Determination of Ferrous Iron in Silicate Rocks and Minerals by Iodine Monochloride Subrata Banerjeel
Department of Geology, University of Georgia, Athens, Ga. 30602 The determination of ferrous iron in rocks and minerals presents a number of difficulties-namely, the method of bringing the sample into solution, the prevention of oxidation in the process, and also during sample preparation. A comprehensive discussion of the methods of a large number of workers has been presented by Schafer ( I ) . Present address, General Refractories Company, Research Center, P.O. Box 1673, Baltimore, Md. 21203. (1) H. N. S. Schafer, Analyst, (London),91, 1089, 755-762 (1966).
782
A N A L Y T I C A L C H E M I S T R Y , VOL. 46, NO. 6, M A Y 1 9 7 4
The use of iodine monochloride in ferrous iron determination was first used by Heisig (2) and later by Hey ( 3 ) . Nichols ( 4 ) applied it to carbonaceous shales where 4% carbon can be tolerated. But the method again calls for decomposition of the sample and then transfer of the contents to a bottle where the oxidation and iodometric titration is carried out. Hence, this suffers from the possibility of aerial oxidation like other methods. (2) G. B. Heisig. J . Amer. Chem. SOC.. 50, 1687 (1928) (3) M. H. Hey, Amer. Mineral., 34, 769 (1949). (4) G. D. Nichols, J. Sediment. Petrology, 30, 603 (1960)
Table I. In Polyethylene Bottles Sample
G-2
Total, ml
0.95 1.oo
FeO %
M1 KIOi Average
Net consumption
Calculated
Actual (9-11)
Deviation, Calculated - Actual Actual
(
1.04
0.94
1.35
1.44
-6.2
1.35
1.25
1.79
2.30
-22.2
1.15
1.05
1.51
2.04
-26.0
2.98
2.88
4.14
4.94
-16.2
4.25
4.15
5.96
6.79
-12.2
5.56
5.46
7.84
8.91
-12.0
5.28
5.18
7.44
8.72
-14.7
1.oo
0.90
1.29
1.35
-4.6
0.59
0.49
0.70
0.83
-15.6
4.11
4.01
5.76
6.61
-12.9
12.06
11.96
17.18
19.80
-13.2
x
100)
1.10
GSP-1
AGV-1
PCC-1
DTS-1
BCR-1
w-1 GA
GH
BR
Fe-MICA
Blank
1.10 1.35 1.40 1.30 1.35 1.20 1.20 1.10 1.10 2.90 2.95 3 .OO 3.05 4.30 4.20 4.25 4.25 5.65 5.50 5.55 5.55 5.25 5.35 5.30 5.20 1.oo 0.95 1.05 1.oo 0.60 0.55 0.55 0.65 4.20 4.15 4.05 4.05 12.10 12.10 12 .oo 12.05 0.05 0.05 0.10 0.15 0.15
0.10
The present investigation is carried out by simultaneous decomposition of the sample, the oxidation of the ferrous iron, a n d evolution of iodine which is absorbed by carbon tetrachloride followed b y titration in presence of boric acid with potassium iodate solution. Similar methods have been successfully carried o u t i n the determination of arsenious ion in silicate glass ( 5 ) and of combined ferrous iron and sulfide ion in amber-colored silicate glass (6). The determination of both metallic copper a n d cuprous copper has also been carried out in borate glasses (7) very precisely.
EXPERIMENTAL Apparatus. Fifty-milliliter Nalgene Tefzel centrifuge tubes with Teflon caps were used for the decomposition of the samples. (5) A. Paul, Glass Tech. 6 , 1, 22 (1965). (6) M. S.Zarnan and A. Paul, Glass Techno/., 10 ( 4 ) , 93 (1969). (7) S.Banerjee and A. Paul, Anal. Chim. Acta, 68, 226-230 (1974).
A rotator driven by a KO-H.P. motor at 60 rpm was used for shaking the samples overnight. It had two vertical perforated plates attached to the main shaft for supporting the tubes containing the samples. Reagents. All the chemicals used were of analytical reagent quality unless otherwise stated. Hydrofluoric acid used was 40% strength (low in iron), w/w. Hydrochloric acid B N was prepared by diluting concd HCl (Sp. gr.1.18)with an equal volume of deionized distilled water. Carbon tetrachloride was free of oxidizing and reducing materials. Iodine monochloride consisted of 10 grams of KI and 6.44 grams of KI03 dissolved slowly in 100 ml of 9A: HC1 (3 parts of concd HC1 to one part of dionized distilled water) until all the solid particles went in solution. The solution was just freed from iodine by (8) W. M. Latirner, "Oxidation Potentials," Prentice Hall, New York, N . Y . , 1959. (9) F. J. Flanagan, Geochim. Cosmochim. Acta. 33, 81 (1969). (10) M. Fleischer, Geochim. Cosmochim. Acta, 33, 65 (1969). (11) P. M. Rowbault. H . LaRoche, and K. Govinda Raju. Sci. Terre. 13 ( 4 ) , 379 (1968).
A N A L Y T I C A L C H E M I S T R Y , VOL. 46,
NO. 6,
M A Y 1974
783
Table 11. In Tefzel Tubes with Polycarbonate Caps M1 KIOs Sample
G-2
GSP-1
AGV-1
PCC-1
DTS-1
BCR-1
w-1
GA
GH
BR
Fe-MICA
Blank
Net consumption
Actual (9-21)
(
Total, ml
Average
1.15 1.20 1.15 1.20 1.75 1.75 1.80 1.75 1.55 1.55 1.50 1.50 3.60 3.55 3.60 3.60 4.80 4.85 4.80 4.80 6.25 6.25 6.20 6.20 6.10 6.15 6.10 6.10 1.10 1.10 1.05 1.05 0.75 0.75 0.70 0.65 4.70 4.70 4.65 4.75 13.85 13.85 13.80 13.80 0.15 0.20 0.20 0.25 0.20
1.18
0.98
1.41
1.44
-2.0
1.76
1.56
2.24
2.30
-2.6
1.53
1.33
1.90
2.04
-7.0
3.58
3.38
4.81
4.94
-2.6
4.81
4.61
6.62
6.79
-2.5
6.23
6.03
8.66
8.91
-2.8
6 .ll
5.91
8.48
8.72
-2.7
1.08
0.88
1.26
1.35
-6.7
0.74
0.54
0.77
0.83
-7.2
4.70
4.50
6.46
6.61
-2.3
13.83
13.63
19.56
19.80
-1.20
x
100)
0.20
dropwise addition of KI03 or KI solution. Five milliliters of this iodine free solution was diluted to 100 ml with 9N HC1. The standard K I 0 3 solution consisted of 0.535 gram of previously dried (120 "C for 2 hr.) and ground KIOJ dissolved in 1000 ml deionized distilled water (M/400) ( 1 ml of M/400 KI03 = 0.0005583 gram Fe). Procedure. Fifty milligrams of the rock powder (-200 mesh) was weighed into a Tefzel 50-ml Nalgene centrifuge tube. Twenty milliliters of 6N HC1, 3 ml of diluted IC1, 2ml HF, and 3 ml cc14 were taken in a 50-ml polypropylene measuring cylinder. This solution was slowly poured down the side of the centrifuge tube. The tube was then stoppered by a Teflon cap (made from a Teflon rod), hand shaken a few times, and set on the rotator overnight. By allowing the reaction to proceed overnight, all the rock particles were decomposed. The solution containing the pink CC14 layer was transferred into a glass-stoppered borosilicate glass tube (70-ml capacity) containing 10 ml of 6N HC1 and 1 gram of boric acid and then titrated with M/400 K I 0 3 solution until the CCld layer was clear. During the titration, and specially near the end, the solution was vigorously shaken. In case of the presence of 784
Calculated
70Deviation, Calculated - Actual Actual
A N A L Y T I C A L C H E M I S T R Y , VOL. 46, NO. 6, M A Y 1974
any organic particle, a blackish layer appeared at the interface of the two liquids, in which case it caused a little difficulty in getting the end point. But addition of another 2 ml of CC14 seems to produce good comparison for the end point. Blanks were run everytime with samples.
RESULTS AND DISCUSSION Seven USGS standard rocks (G-2, AGV-1, GSP-1, DTS-1, PCC-1, BCR-1, and W-1) and four French standard rocks (GH, GA, BR, and MICA-Fe) were chosen for analysis to see the validity of the method. Results using different containers are given in Tables I, 11, and III. The main problem in getting an accurate result was to obtain complete decomposition of the rock. Since H F was not added in large excess and it gets fairly diluted when added (2 ml in 25 ml) the rate of decomposition was slow. In previous studies with iodine monochloride (6, 7), the glasses dealt with were simple binary systems and
Table 111. In Tefzel Tube with Teflon C a p M1 KIOI Sample
G-2
GSP-1
AGV-1
PCC-1
DTS-1
BCR-1
w-1 GA
GH
BR
Fe-MICA
Blank
% Deviation,
FeO%
(Calculated Actual - Actual
Total, ml
Average
Net consumption
Calculated
Actual (9-11)
1.20 1.15 1.15 1.20 1.75 1..80 1.75 1.80 1.55 1.55 1.60 1.50 3.60 3.60 3.65 3.60 4.85 4.75 4.85 4.80 6.35 6.30 6.35 6.30 6.20 6.25 6.15 6.15 1.15 1.10 1.05 1.10 0.80 0.75 0.70 0.70 4.75 4.70 4.65 4.70 13.90 13.90 13.80 13.85 0.20 0.10 0.15 0.15 0.15
1.17
1.02
1.47
1.44
t 2 .o
1.78
1.63
2.34
2.30
$1.7
1.55
1.40
2 .oo
2.04
-2
3.61
3.46
4.97
4.94
+O .6
4.81
4.66
6.70
6.79
-1.3
6.32
6.17
8.86
8.91
-0.6
6.19
6.04
8.68
8.72
-0.5
1.10
0.95
1.38
1.35
+2.2
0.74
0.59
0.85
0.83
+2.4
4.70
4.55
6.55
6.61
-0.9
13.86
13.71
19.69
19.80
-0.5
x
100)
.o
0.15
decomposed readily in HCl or HF. In these cases the reaction was carried out in stoppered Pyrex glass tube for 1h hr, which was sufficient for the complete decomposition of the sample. In the present investigation, it was observed that complete decomposition was not achieved in 'j hr. Results were low by as much as 40%. Furthermore, H F will attack the glass tube considerably with longer period of time. In an attempt to avoid this, polypropylene bottles were tried. But there were two disadvantages: (1) the bottles were not transparent enough to see the color of the carbon tetrachloride layer and ( 2 ) the polypropylene, polycarbonate bottles all absorb iodine to a considerable extent and almost in proportion to the amount of iodine liberated (as is evident from the results in Table 1)-resulting in lower values. The only alternative was to find a tube which will not be attacked by H F and also not absorb iodine.
The Nalgene Tefzel 50-ml centrifuge tube was the ideal solution, but the caps supplied by Nalgene for these tubes were made of polycarbonate which absorbed iodine, causing slightly lower results (Table II). Hence Teflon caps were made from Teflon rods to fit in the tubes and all the problems were eliminated. The results are given in Table III. The blank readings in these studies were extremely low (0.10 ml to 0.20 ml of M/400 KI03 solution). The samples must be fine enough to pass through a 200-mesh screen in order to ensure complete decomposition. It's evident from the following values of the reduction potentials of the redox systems below that iodine monochloride, which has a lower reduction potential than either permanganate or dichromate, is preferable to use for iron. Mn0,-
+ 8H' + 5e
d
Mn'+
A N A L Y T I C A L C H E M I S T R Y , VOL. 46,
+
4H,O
NO. 6,
M A Y 1974
785
EO = +1.52 volts (1) 7H20 Cr20,2- 14H’ EO = +1.33 volts (2) 2e == Mn2+ 2Hz0 MnOp 4H’ Eo = +123 volts (3) I+(Cl-) e == %I2 C1- Eo = +1.19 volts@) (4) Fe3+ e e Fez+ Eo +0.77 volts (5) Thus, there is no problem of any interference from Mn2+
+
+ 6e
+
== Zr3+ +
+
+
+
+
+
5
or C r s which are in general present in rocks in small amounts (Mn from 0.05 to 0.2%, Cr from 1to 4000 ppm). The presence of an appreciable amount of “acid decomposable sulfide” invalidates the ferrous iron determination. Pyrite is not appreciably attacked by mixture of HF and HC1 but other sulfides, such as pyrrhotite, are more extensively decomposed liberating hydrogen sulfide which
will result in higher values of ferrous iron. Organic matter other than graphite will completely invalidate the meth-
od. The relative 70 deviation has been calculated on the amounts present.
CONCLUSIONS Ferrous iron can be determined by the iodine monochloride method without any possible aerial oxidation and without any interference from manganese or chromium. Ferrous iron in carbonate and other acid decomposable rocks (which are attacked by HCl or HCl and HF) can also be determined. Acid decomposable sulfides and organic matter other than graphite invalidate the method. Received for review October 6, 1973. Accepted January 7, 1974.
Automated Loading of Discrete, Microliter Volumes of Liquids into a Miniature Fast Analyzer C . A. Burtis, W. F. Johnson, and J. B. Overton Oak Ridge National Laboratory, Oak Ridge, Tenn. 37830
A miniature Fast Analyzer is under development at the
Oak Ridge National Laboratory (1-4). The analyzer combines the inherent advantages of the Anderson fast analyzer concept (5) with those of miniaturization. One of the advantages of the latter is a further decrease in sample and reagent volume requirements. For example, for each assay, the miniature analyzer requires only from 2 to 10 ~1 of sample, which is analyzed in a total reaction volume of only 120 to 130 rl. Consequently, to fully realize the advantages of the analyzer, it is necessary to have the capability of precisely and accurately introducing volumes of this magnitude into the system. Samples and reagents are introduced into the miniature Fast Analyzer via a 17-cuvet rotor into which aliquots of sample(s) and reagent(s) are loaded, transferred, and mixed (within their respective cuvets). The ensuing reactions are then photometrically monitored. One of the two modes in which the aliquots of sample(s) and reagent(s) can be introduced into the rotor is a discrete mode ( 3 ) , in which a dispensing device is used to obtain aliquots of the liquids and dispense them into their respective sample and reagent cavities in the rotor. Using two commercially available, automatic pipets and a unique carousel-tumtable assembly, a sample-reagent loader which automatically performs this discrete loading operation has been designed and fabricated.
EXPERIMENTAL Instrument Description. The sample-reagent loader (Figure 1) sequentially and automatically obtains and dispenses aliquots of reagent(s) and sample(s) into the corresponding cavities in the (1) N. G. Anderson. C. A. Burtis. J. C. Mailen. C . D. Scott, and D. D. Willis, Anal. Lett. 5, 153 (1972). (2) C . A . Burtis, J . C. Mailen, W . F. Johnson, C. D. Scott, T . 0. Tiffany, and N. G.Anderson, Clin. Chem., 18,753 (1972). (3)C D . Scott and C . A . Burtis, Anal. Chem. 45,327A (1973). (4)C . A. Burtis, W . F. Johnson, J . C. Mailen, J . 8.Overton, T. 0.Tiffany, and M. B. Watsky, Clin. Chem. 19,895 (1973). (5) N. G.Anderson, Amer. J . Clin. Pathol., 53, 778 (1970).
786
A N A L Y T I C A L C H E M I S T R Y , VOL. 46, NO. 6, M A Y 1974
rotor. Primary components of the loader are two Micromedic automatic pipets (Model 25004, Micromedic Systems, Inc., Philadelphia, Pa.), which obtain and deliver aliquots of reagent and sample respectively. Both of these pipets are operated in a sample-diluting mode, which entails aspirating an accurately measured volume of sample or reagent into a small-diameter, flexible delivery tube and then dispensing and following it with a preset quantity of diluent (usually distilled water). As shown in the flow schematic in Figure 2, the loader utilizes one of the pipets to obtain aliquots of sample and reagent and the other to dispense and dilute the sample and reagent aliquots into their respective cavities in the rotor. This unique arrangement of pipets, pumps, and delivery lines is an important factor in minimizing the carryover of the loading system and will be discussed later in greater detail. The volumes aspirated and dispensed with these pipets are set by adjusting index counters on the control panel of the pipets. This adjustment allows the volumes to be varied from 5 to 100% of full pump capacity. To give a selectable volume range within that required by the miniature Fast Analyzer, 20-4 sample and reagent pumps and 50-pl diluent pumps have been utilized. Typically, the pumps of the pipets are set to obtain and deliver from 2 to 10 p1 of sample followed by 50 p1 of diluent, and 20 pl of concentrated reagent followed by 50 pl of diluent. To automate the sample- and reagent-loading process, a unique turntable and carousel assembly was developed for use with the Micromedic pipets. In an earlier evaluation, it was found that, during work with microliter volumes, maximum performance was obtained from the Micromedic pipet when its probe was kept stationary. Consequently, during routine operation of the automated sample-reagent loader, the sample probe remains stationary and the reagent probe moves only slightly, while the sample and reagent cups and the respective receiving cavities in the rotors are sequentially brought to the probes. Thus in a typical loading operation, (a) the sample and reagent cups are positioned and aligned below the sample and reagent probes; (b) the turntable is raised, the probes enter the liquids in the cups, and aliquots of sample and reagent are aspirated into the probes; (c) the turntable is lowered, thereby removing the probes, and then moved laterally to position and align the sample and reagent cavities of the rotor under the probes; (d) the turntable is raised, and the probes enter the rotor cavities, where the aliquots are dispensed and diluted with a preset quantity of diluent; (e) the turntable is lowered, thereby removing the probes, and then moved laterally to position and align the two cups of the wash station under the
