necessary to chromatograph the standards if the flow rate is known. Also, information in the time domain is retained. Projections. The most serious limitation of the vidicon detector system used in this work is the photometric drift which results from the single beam operation and we now have work under way designed to provide double beam operation with a stabilized source. This can be achieved using a single detector and one set of dispersion optics and thus shows an advantage over linear diode array detectors where two matched detectors and two sets of dispersion optics are necessary for double beam operation. Another less critical limitation is that some tradeoff must be made between spectral range and resolution and we are working on optical systems which will provide both broad coverage and good resolution. Recent work in this laboratory has demonstrated that the use of integrated area under an absorption profile can improve photometric precision in comparison with single wavelength data (9). We have work under way to determine if this technique offers any significant advantage for chromatographic processes. The vidicon detector offers a convenient approach for derivative spectroscopy (15,16) and there may be useful applications of derivative spectra for liquid chromatography. Finally, although this paper has emphasized the absorption process, vidicon detectors can also be used as fluorescence detectors ( I 7), and multiwavelength fluorescence detection should offer real advantages in liquid chromatography. We conclude from this work that the silicon target vidicon is a viable multiwavelength detector for liquid chromatog-
raphy, that it offers several potential advantages over single wavelength detectors, and that there still are several aspects of the concept which merit attention.
LITERATURE CITED L. R. Snyder and J. J. Kirkiand, “Introduction to Modern Liquid Chromatography”, John Wiley and Sons, New York, N.Y., 1974. H. M. McNair, J . Chromatogr. Sci., 14, 477 (1976). M. S.Denton, T. P. DeAngells, A. M. Yacynych, W. R. Helneman, and 1.W. Gilbert, Anal. Chem., 48, 20 (1976). R. E. Dessy, W. G. Nunn, C. A. Titus, and W. R. Reynolds, J. Cbromatcgr. Scl., 14, 195 (1976). R. E. Dessy, W. D. Reynolds, W. G. Nunn, C. A. Titus, and G. F. Molar, Clin. Chem. (Winston-Salem, N.G.), 22, 1472 (1976). M.J. Milano, S.Lam, and E. Grushka, J. Chromatogr., 125, 315 (1976). A. E. McDowell and H. L. Pardue Anal. Cbem., 48, 1815 (1976). M.J. Milano, H. L. Pardue, T. E. Cook, R. E. Santini, D. W. Margerum, and J. M. T. Raycheba, Anal. Cbem., 46, 374 (1974). H. L. Felkel and H. L. Pardue, Anal. Cbem., in press. A. Savitzky and M. J. E. Golay, Anal. Chem., 38, 1627 (1964). T. A. Nleman and C. G. Enke, Anal. Chem., 48, 619 (1976). W. H. Lawton and E. A. Sylvestre, Tecbnometrics, 13, 617 (1976). R. G. Berg, C. Y. KO, J. M. Clemons, and H. M. McNair, Anal. Chem., 47, 2480 (1975). A. E. McDowell, R. S. Harner, and H. L. Pardue, Clln. Cbem. (Winston-Salem, N.C.), 22, 1862 (1976). T. E. Cook, R. E. Santini, and H. L. Pardue, Anal. Chem., 48, 451 (1976). T. E. Cook, R. E. Santini, and H. L. Pardue, Anal. Chem., In press. I. M.Warner, J. B. Caliis, E. R. Davidson, and G. D. Christian, Clin. Cbem. ( Winston-Salem, N.C.), 22, 1483 (1976).
RECEIVED for review March 3,1977. Accepted April 29,1977. This investigation was supported in part by P H S Research Grant No. GM 13326-10 and 11from the National Institutes of Health.
Determination of Anions in Boiler Blow-Down Water with Ion Chromatography Timothy S. Stevens” and Virgil T. Turkelson Dow Chemical, U.S.A. Michigan Division Analytical Labs, 574 Bldg., Midland, Michigan 48640
William R. Albe Dow Chemical, U.S.A., Instrument Applications and Communications, 1603 Bldg., Midland, Michigan 48640
An application of ion chromatography to the analysls of boiler blow-down water is described. Glycolate (produced by decomposition of EDTA), chloride, sulfite, sulfate, and orthophosphate are separated and quantified. The data are used to monitor boiler feedwater treatment. The analytical columns required have a life expectancy of more than one month when employed In an on-line analyzer.
Industrial boilers used for process steam and electrical power generation may require periodic and costly overhaul because of corrosion and scale buildup in boiler tubes. These problems are controlled by chemical treatment of the demineralized and deaerated boiler feedwater and by continuous “blow-down” of the boilers, Le., a small fraction of the feedwater is not boiled to steam but is continuously drained from the boiler as blow-down water. Monitoring the concentration of the treatment chemicals is needed and this can be done by analyzing the blow-down water. 1170
ANALYTICAL CHEMISTRY, VOL. 49, NO. 8, JULY 1977
Ethylenediaminetetraacetic acid (EDTA) and orthophosphate are added to sequester scale-forming calcium and magnesium ions. The orthophosphate added has the additional effect of passivating the boiler tube surface against general overall corrosion. Sodium sulfite treatment scavanges residual oxygen from the boiler water, preventing pitting corrosion. Finally, corrosion caused by hydrogen ion is controlled by adding sodium hydroxide. The reader is referred to (1)for a more detailed discussion of boiler water treatment. In our boiler blow-down water, sulfite concentration is maintained between 5 and 10 ppm as Na2S03. Orthophosphate concentration is maintained between 10 and 20 ppm as Na2HP04. Hydrogen ion concentration is controlled by maintaining the pH at 10. EDTA treatment is accomplished by continuous feed at the same rate from day to day. Although EDTA is added to the feedwater, it is not detected in blow-down water. However, glycolic acid is detected (confirmed using NMR) and its presence suggests that in the high temperature and pressure of the boiler ( 1100 psig) the EDTA hydrolyzes to glycolic N
Table I. Chromatographic Conditions Eluent 0.005 M sodium carbonate 0.004 M sodium hydroxide in deionized water Flow rate 105 mL/h Analytical column Chromex anion exchanger, 2.8 X 1000 mm (a 50-pm surface layer active type) Dionex Corp. Stripper column Dowex 50-W X 16, 2.8 X 300 mm Detector sensitivity 10 umho full scale Injection volume 500 p L Regenerant 1 M sulfuric acid in deionized water Pump pressure 200 psig
10 r
Glycolate
5
0
10
15
20
25
Minuter
Flgure 2. Typical boiler blow-down sample
I
0
n
5
7.5 ppm as Na,SOo
10
15
20
Minutes
Figure 1. Standard of “10 ppm” Na2S03
acid and ethylenediamine. EDTA treatment is nevertheless effective in controlling scale. The subject of this contribution is the determination of glycolate, chloride, sulfite, sulfate, and orthophosphate in boiler blow-down water using ion chromatography. The work was undertaken to indicate the feasibility of incorporating an ion chromatograph in an automated on-stream analyzer. Ion chromatography is a new analytical technique using two ion-exchange columns in series followed by a flow-through electrical conductivity detector. The first column separates the ions in the injected sample, while the second column suppresses the conductance of the electrolyte in the eluent but not that of the separated ions (2). Analyses obtained using this procedure were compared with the previously used iodometric technique for sulfite determination (3),and the heteropoly blue method for orthophosphate determination (4). The ability to monitor EDTA treatment by determining glycolate provided previously undetermined information. Finally, analytical column stability was studied in long term use.
EXPERIMENTAL The ion chromatograph used was a pre-production prototype of the Model 10 unit available from Dionex Corp., 1228Titan Way, Sunnyvale, Calif. 94086. Chromatographic conditions are given in Table I. Procedure. The system was standardized for glycolate, chloride, sulfate, and orthophosphate using reagent grade salts dissolved in deionized water. Calibration curves were then drawn using peak height vs. concentration from the chromatographed standards. Standardization for sulfite was complicated as a fraction of the sulfite reacted with oxygen in the dilution water, forming sulfate, even though the water was pretreated using a sparge of nitrogen. Thus, the actual sulfite concentration had to be appropriately corrected for the sulfate level determined in each sulfite standard (see Figure 1). Boiler blow-down water samples from each boiler were chromatographed as received. If the method was to be used in an automated on-stream analyzer,
Minutes
Figure 3. Sample when feedwater deaerator failed
it had to provide usable analytical column performance for at least a month of continuous use. The operation was tailored to provide one analysis in 30 min. Thus, in one day, 24 mL of sample (48 injections of 0.50 mL each) would pass through the analytical column, totaling 672 mL in four weeks. Column stability was studied by the following procedure, simulating actual use: column performance was determined when new and after directly pumping sample through the analytical column equivalent to 7, 14,and 28 days of operation. The injection volume selected (0.5 mL) was larger than normal for this technique (2), but it produced chromatograms with excellent signal-to-noise ratio, an important consideration for an on-line analyzer. Finally, the eluent used (solution of Na2C03and NaOH) has been found to be widely applicable to the separation of anions using ion chromatography (5).
RESULTS AND DISCUSSION Figure 2 shows an ion chromatogram of a typical sample. Figure 3 shows the analysis of a sample when the feedwater deaerator failed, resulting in more oxygen in the water than the sulfte treatment could remove. Note that the sulfate level in this sample has increased significantly. The sulfite-sulfate ratio in a sample is an indication of the effectiveness of feedwater deaeration. Figure 4 shows the analysis of a sample when the EDTA feed pump seal leaked enough to prevent any treatment, with the resulting absence of a glycolate response. Table I1 compares data obtained by this method and the previously used “wet chemical” methods for the determination of sulfite and orthophosphate, for the same samples. The data indicate that there is no significant statistical bias between the methods studied. Figure 5 shows the stability of the analytical column during one month of simulated continuous use. It has lost 32% of ANALYTICAL CHEMISTRY, VOL. 49, NO. 8, JULY 1977
1177
Table 11. Sulfite and Orthophosphate Determinations in Boiler Blow-Down Water-Comparative Data Orthophosphate as ppm Na,HPO, Sulfite as ppm Na,SO, IonIonIodometryb Chroma Bias ColorimetricC Chroma Bias 2.6 2.8 +0.2 9 8 -1 4-3 +0.6 10 13 0.8 0.2 15 18 +3 -0.1 4.0 4.1 3.3 3.2 -0.1 9 8 -1 +2 +0.5 11 13 1.5 1.0 11 13 t2 +0.1 2.6 2.7 15 14 -1 -0.4 3.7 4.1
3 CI-
P
so:
%
I
$ c
8
0 0
5
15
10
20
25
Minutes
Flgure 4. Sample when EDTA feed pump failed
1.7 0.2 4.9 3.3 2.6
1.4 0.3 3.7 2.2 1.7
-0.3 +0.1 -1.2 -1.1 -0.9
9 20 20
7 20
-2
18
10 20
21
-2 -2 +1
8
0
Average bias, - 0.2 ppm Average bias, 0.2 ppm a Repetitive analysis of a single sample indicates that data obtained using this method have a relative standard deviation of 3% for sulfite determination and 4% for orthophosphate determination, b,c The precision of these methods is discussed in references 3 and 4.
mi of Boiler Blow Down Water passed through the Analytical Column
Figure 5. Analytical column stability as a function of the amount of sample passed through, simulating 1, 2, and 4 weeks of continuous use. (Capacity determined from the elution volume of bromide, corrected for column void volume)
fraction for analytical use. The use of this method in an automated on-line analyzer would require periodic up-dating of the integrator program (about every week) to shift the peak “windows” to correspond to elution times. Alternatively, an integrator could be used that identified one peak as a “marker” and continuously adjusted itself. Recent experience with an automated on-line analyzer has shown that column life is extended to about three months if the sample streams are fitered. The Balston Model B 2~ filter cartridges required replacement at least every month.
CONCLUSION The use of the ion chromatograph has been extended to the monitoring of chemical treatment of boiler water. This instrument can be automated in an on-line analyzer to provide the frequent and reliable data needed for effective treatment control.
LITERATURE CITED
Figure 6. System performance after four-week sample has been passed through the analytical column
(1) “Chemical Analysis of IndustrialWater”, J. W. McCoy, Chemical Publishing Co., New York, 1969. (2) H. Small, T. S.Stevens, and W. C. Bauman, Anal. Chem., 47,1801-1809 (1975). (3) “Standard Methods for the Examination of Water and Wastewater”, 13th ed., American Public Health Associatlon, Washington, D.C., 1971, pp 337-338. (4) Ref. 3, pp 530-532. (5) H. Small and J. Solc, “Proceedings of An International Conference on the Theory and Practice of Ion Exchange”, University of Cambridge, U.K., July 1976.
its capacity in this time, but still resolves the components similarly (see Figure 6). Column degradation is thought to be caused by trace levels of “poisons” in the sample which irreversibly combine with the ion-exchanger, inactivating that
RECEIVED for review February 28,1977. Accepted April 15, 1977.
II
I
5
10
15
20
25
Minutes
1178
ANALYTICAL CHEMISTRY, VOL. 49,
NO. 8, JULY 1977
