Any wavelength passing through center of exit slit of monochromator, nm = Wavelength passing through center of exit slit when refractor plate is perpendicular to the optical axis of the monochromator, nm = Wavelength at center of absorption line profile, nm = Apparent half-width of absorption line, nrn = Half-width of absorption line, nrn = Doppler half-width of absorption line, nrn = Refractive index of refractor plate, no units = 3.1416. . . =
h A,
A0
Ah AXA Aho Ir lr
nez mc
=
P
=
2.65 X
cmz sec-l Coefficient to correct ko for instrumental broadening, no units
+n‘v
= Thickness of refractor plate, cm = Flow rate of solution into nebulizer, cma min-1 = Solid angle of radiation collected by the monochroma-
WI
= Frequency of mechanical chopper (source modulation),
7
tor, sr
Hz WZ
Frequency of refractor plate oscillation (wavelength modulation), Hz
RECEIVED for review August 5 , 1971. Accepted December 7, 1971. One of the authors (R. C. E.) thanks the Graduate School of the University of Florida for financial support in the form of a Graduate School Fellowship. This work was supported by AFOSR (AFSC), U.S.A.F. Grant 70-1880B.
Flame Detection Method for Determining Organic Carbon in Water F. T. Eggertsen and F. H. Stross Shell Development Company, Emeryville, Calif.
Trace organic carbon in water can be determined by means of a thermal analysis-flame detection system previously devised for the thermal stability and volatility characterization of organic materials. In the method developed, a small sample is heated in nitrogen carrier gas in two stages to determine volatile ( 4 5 0 “C) and nonvolatile (150-550 “C) organic carbon with the hydrogen flame ionization detector. The water evaporated in the first stage changes the detector response to some extent but this is taken care of by calibration. Performance of the method is illustrated in tests with 1 to 300 ppm of various types of organic material. The lower limit of detection is about 0.2 ppm. Advantages of the niethod are its simplicity and sensitivity and a differentiation of the organic material according to volatility. The analysis time i s about 15 minutes.
on clean environment has led t o great interest in the development of apparatus for determining organic contaminants in water. A type of analysis frequently desired is the measurement of total organic carbon. One of the best known commercial instruments developed for this purpose is the Beckman total organic carbon (TOC) analyzer, which utilizes a n analysis scheme developed by Van Hall and coworkers (1). In this scheme of analysis the organic carbon in a water sample is oxidized t o COz which is swept through a n infrared cell for measurement. Another instrument, introduced recently by the Precision Scientific Company, and based on the work of Stenger and Van Hall (2), is the “A,quarator”; in this approach the organic carbon is oxidized t o CO by C 0 2 carrier gas and the CO is determined by infrared. A more sensitive method, which utilizes a hydrogen flame ionization detector, was introduced by Lysyj et a/. (3, 4). In this method the water sample is THE CURRENT EMPHASIS
(1) C. E. Van Hall, J. Safranko, and V. Stenger, ANAL.CHEW,35,
315(1963). (2) V. Stenger and C. E. Van Hall, ibid., 39, 206 (1969). (3) I. Lysyj. K . H. Nelson, and P. R. Newton, “Methods for the Determination of Trace Organic Materials in Water,” U.S. Dept. of the Interior, Office of Saline Water, Research and Development Progress Report No. 239, February, 1967; J . Water Pollut. C o w . Fed., 41 ( 5 ) , Pt. 1, 831 (1969). (4) K. H. Nelson, I. Lysyj, and J. Nagano, Wafer Sewage Works, 117, 14 (1970).
injected into a high temperature pyrolyzer and the organic carbon is determined by sweeping the products to the detector using nitrogen or water vapor as a carrier gas. The present report describes a method, also utilizing a flame detector, which is simpler is some respects than that of Lysyj and supplies additional information. It is carried out by means of a thermal analysis-flame ionization detection (TAFID) instrument previously developed for thermal stability and volatility measurements of organic materials (5, 6). The analysis is made simply by rapidly vapoiizing the water sample under controlled conditions, followed by a second heating stage, while sweeping the vapors with nitrogen carrier gas directly into the flame detector. A particular advantage of the method is that it characterizes the sample according t o volatility. Volatile organic material is evolved while vaporizing the water; and the relatively nonvolatile material is obtained by a rapid high temperature pyrolysis step. The feasibility of using the TAFID type of instrument for this analysis has been reported by Butz et a/. (7). In the present study this approach was investigated in more detail t o develop a rapid quantitative method. EXPERIMENTAL
Apparatus. A schematic diagram of the system is shown in Figure 1. The basic components include the following. A sample furnace, of Vycor tubing, heated by resistance wire winding, FID detector, modified Aerograph, the flame jet of which is also of Vycor and sealed to the sample furnace, capable of operation at 500 “C;amplifier, Aerograph electrometer 500-D; furnace heating controls, manual control by suitable adjustment of voltage (potentiometer) to the heater windings, or a temperature programmer; integrator, digital (5) F. T. Eggertsen, H. M. Joki, and F. H. Stross, “Thermal Analysis,’’ Vol. 1, R. F. Schwenker, Jr., and P. D. Garn, Ed., Academic Press, New York, N.Y. 1969, p 341. (6) F. T. Eggertsen, E. E. Seibert, and F. H. Stross, ANAL.CHEM., 41, 1175 (1969). (7) W. H. Butz, A. C. Stapp, and Alan Di Stefano, “Total Organic Content (TOC) Analysis with a New FID Thermal Analyzer,”
presented at the Pittsburg Conference on Analytical Chemistry and Applied Spectroscopy, Cleveland, Ohio, March 1970. ANALYTICAL CHEMISTRY, VOL.. 44, NO. 4, APRIL 1972
709
1
Recorder
A.
91 ppm Heptanoic Acid
500' I
1t. 1
Mixing
T
N2
H2
Figure 1. Schematic diagram of the apparatus
Isopropyl Alcohol 2 . 8 ppm Temperature
0.5 mv
1 -
J
Heptanoic Acid
H,O Blank
I 0
1
/ I
I 4
I
I
a
1
1 12
I
16
Time, minutes
150'
H;at
Figure 3. Curves for heptanoic acid
to 550"
u
0
4 8 Time, minutes
12
Figure 2. Curves for, isopropyl alcohol and distilled water blank type, or a planimeter for measuring peak areas; recorder, F. L. Moseley Co. Model 7100A, 1-500 mV, two-channel. The sample probe was constructed of stainless steel, except for a tapered aluminum O-ring joint which fits into the furnace tube. The sample holder is attached to a sliding rod, permitting the sample to be located at the cool furnace inlet while the detector is ignited. The sample is then thrust into the furnace to start the analysis. A thermocouple is located near the sample pan to permit accurate temperature measurement. Also required are suitable flow controls for nitrogen carrier gas and hydrogen and air for the flame detector. The gases were purified by passage through Molecular Sieve 13X. The septum inlet and mixing chamber shown in Figure 1 are for the purpose of injecting standard gas samples to calibrate the detector response. Also the gas flow scheme permits in710
ANALYTICAL CHEMISTRY, VOL. 44, NO. 4, APRIL 1972
iecting air through the system to oxidize any accumulated residues and thus keep the system clean. The thermal analysis system Model 3000 of Carle Instruments, Inc., Fullerton, Calif. (7), is suitable for use in the method described here. Materials. Test solutions for this study were prepared using various reagent grade chemicals dissolved in distilled water. In addition the following materials were employed, also as solutions in distilled water: gelatin, Eastman Organic Chemicals, purified calfskin ; starch, Fisher Scientific potato starch, purified powder; and Neodol 25-12, Shell Chemical Company alcohol ethoxylate type detergent. The biotreater samples were feeds and effluents of a biotreater which contained largely alcohols, ketones, and phenolic compounds. The distilled water was taken directly from that supplied to the laboratory without further purification. Procedure. The conditions for the analysis were generally as follows : A 50-111 sample was contained in a platinum combustion boat (10 mm long x 4 mm wide x 4 mm deep). The furnace temperature was 150 OC for 5 minutes; then heated in about 5 minutes to 550 "C. FID temperature was 500 "C. An ampliA = 1 mV (range 1 attenuation X fier sensitivity of 6 x 2 on the Aerograph electrometer) was employed. The flow
A.
i
90 ppm G e l a t i n
f
0.1
f
i . -
B.
2
150'
Heat to 550° 4
0
8
12
16
T i me, minutes
Figure 5. Curve for a detergent
II
18 DDm G e l a t i n
Total C =
rnv
i
l4ppm C
1.0 mv
NEODOL 25-12, 4.3 pprn
f
A.
Undiluted
2.0 rnv
f
,3.4 ppm C
I?/
0.1 mv
150' Heat to 550' H,O B l a n L
150"
I,
\
Heat to 550°
B. Diluted 10: 1
5 0
4
8 Time, minutes
12
16
Figure 4. Curves for gelatin
rates were nitrogen carrier, 30 ml per min; hydrogen, 25 ml per min; and air to FID, 800 ml per min. To perform the analysis, the furnace is cleaned by directing air through it for a few minutes while at a temperature of 550 "C. (This cleaning needs to be done only once or twice per shift.) Nitrogen carrier gas is again directed through the furnace and the F I D flame is lighted. The furnace is cooled (air jet) and its temperature is set at 150 "C. The base line is established at 1 mV recorder sensitivity. The sample boat is cleaned by heating to red heat in a bunsen flame. The sample is added to the cooled boat from a hypodermic syringe and is placed in the probe at the cool inlet of the furnace. The flame detector is relighted. The sample is immediately pushed into the 150 "C furnace and allowed to evaporate. As the evaporation proceeds, the recorder is attenuated as necessary to keep the curve on scale. After about 5 minutes, the F I D signal suddenly returns to near the original base line, indicating that the evaporation is substantially complete. At this point or a few minutes thereafter, the furnace is rapidly heated by increasing the voltage to the heater windings so as to reach 550 "C in about 5 minutes. This produces a peak
0.2 mv
I
\
150'
I
0
I
I
4
Heat to 550' I
8
12
I
16
Time, minutes
Figure 6. Curves for oil in water Distilled H 2 0 shaken with 94 ppm cat.-cracked gas oil for organic material volatilized at 150 to 550 "C. Typical recorded curves are shown in Figures 2 to 7. A blank test is made in the same way using 50 p1 of distilled water. This test is ordinarily made once per shift. Response factors for evaluating peak areas are determined as follows. A standard sample of 1 % n-butane in helium, usually 40 pl, is injected into the system with a gas-tight ANALYTICAL CHEMISTRY, VOL. 44, NO. 4, APRIL 1972
711
Table I. Analysis of Aqueous Isopropyl Alcohol Solutions Present, ppm Foundb i-PrOH Total C “FID” C@ C, pprn Peak ht, rnV
15.9 17 3.5 3.2 3.0, 2 . 9 0.7 1.3 1.3 0.25 0.64 0.6, 0 . 6 0.16 0.32 0.4, 0 . 4 0.10 0.13 0.2 0.04 a Calculated from total carbon using the value of 2.2 as the effective carbon number for isopropyl alcohol in the FID, based on 4.0 for n-butane (9). *The values were measured with reference to the curve for a 36.4 7.3 2.8 1.4 0.73 0.30
21.8 4.3 1.7 0.87 0.44 0.18
water blank.
150’
L 0
a
4
12
16
Time, minutes
Figure 7. Curve for feed to a biotreater syringe and the response determined in terms of peak area (or integrator counts) per microgram of carbon. A response factor is determined both for the dry system and for the condition in which the butane is injected while a distilled water sample is evaporating. In the latter case the butane was injected so as to emerge at various times during evaporation of the water. The average response factor so determined was 25 f 5 lower than that determined in the absence of water. A similar lowering in F I D response to organic carbon was reported for tests in which water vapor was added to argon carrier gas (8). The organic carbon content of a sample is evaluated from the recorded curves by separately measuring the peak areas (planimeter or integrator) for the volatile (
