FEATU
I hor Lysyj Rocketdyne Div., Rock well International Corp., Canoga Park, Calif. 9 1304
One of the more critical problem areas in the control and abatement of environmental pollution is the current status of its measuring technology. Traditionally, rneasurements on the quality of water and air have been made largely in gross terms. For example, the inorganic and organic compositions of water have been characterized by "hardness" and "Chemical Oxygen Demand'' (COD), respectively. Such information often has been supplemented by measurements of selected anions and cations present in the sample, and by a number of other physical and chemical measurements such as temperature, pti, and redox potential. More recently, advanced analytical and instrumental methods have been applied to problems of environmental pollution control a r d have led to some significant advances in the field. In most cases, this has involved adaptation of existing and proved analytical techniques, such as gas and liquid chromatography, mass spectroscopy, and optical and electrochemical methods for the detection, characterization, and quantification of molecular species present in environmental matrices. Despite some remarkable successes, this activity has shown some weaknesses. The extreme complexity of environmental processes and the very large number of chemical and biological species participating iri such processes have precluded, in some cases, effective utilization of analytical techniques based on the characterization of specific chemical entities. I t has become evident that. even if it were possible to characterize these thousands of individual chemical compounds, it would be extremely difficult to organize and make valid deductions from such an avalanche of data. There is a need for conceptually novel approaches for characterization and study of the dynamic and complex nature of environmental processes. At the present time, it is possible. usina state-of-the-art methods,' to characterize compositions of environmental matrices in gross terms-BOD, hardness: or in molecular t e r m s---. i ro n , Ie ad, acet o ne, I act ose , benze ne. Ne ithe r alternative provides an effective means for dealing with practical problems of environmental pollution control and abatement.
FIGURE 1
Typical pyrograms
Albumin heat
a1 (CH,]
+ b, (C,H,) + c, (C,HJ + d, (C,H,) + e, (C,HB) + Other gases
___F
-
Starch heat
a, (CH,)
+ b, (C,HI) + c, (C,HJ + d, (C,H,) + e2 (C,H,) + Other gases Volurre 8, Numbsr I , January 1974
31
FIGURE 2
Pyrographic system
Carrier gas
~Pyrolysischamber
Teletype Telephone
Computer
I t may be instructive to examine a common problem in pollution control, such as tracing a waste to its source and determining the relative contribution of a given waste source to the overall pollution load of a waterway. To solve this problem, the identity of the contributing waste source must be established and the quantity of its contribution determined. This must be done in the presence of waste compositions from different sources and in the presence of the natural chemical background of a waterway. The gross characterization methods do not possess differentiating capabilities for such a task. For example, the BOD value from one source cannot be differentiated from the BOD value of a mixture containing waste from two sources. If carried out in sufficient detail, the molecular specific methods could identify sources of pollution. Such would be the case when a specific chemical compound is found only in one waste discharge and thus becomes a tracer for that waste. These methods, however, could not cope with the overall quantitative assessment of waste load contributions from a number of waste discharges to a common waterway because various waste discharges will contain many common chemical components.
Pyrographic method To deal with the problem of water pollution monitoring and other related environmental measurements, a methodology is needed that can define multicomponent chemical compositions as identifiable and differentiable entities which could be handled quantitatively. A new analytical approach which combines physical means and mathematical logic was therefore proposed to the predecessor of the EPA in 1968. A feasibility study was undertaken in the same year at Rocketdyne. Based on the theory of multicomponent pattern recognition and differentiation, this approach represents a fundamental departure from the traditional working concept of analytical chemistry-i.e., resolution of a mixture into the molecular species, followed by identification and quantification of specific compounds. Multicomponent pattern recognition permits definition of a complex chemical composition as a separate entity, and provides means for studying its interactions with other compositions found in environmental matrices. 32
Environmental Science & Technology
The pyrographic method was selected as a physical vehicle for application of this theory to pollution monitoring. In this method, organic matter is subjected to elevated temperatures in the absence of oxygen, leading to thermal decomposition of organic molecules into several preferential fragments. The nature and quantity of such fragments reflect the elemental and structural character of the parent material. Through identification of the pyrolytically produced derivative composition by means of gas chromatography, the nature of the parent material can be defined. By comparing pyrograms of unknown matter with that of known materials, the qualitative nature of the substance in question could be determined. The possibility of quantitative analysis of mixed organic composition by pyrography was first demonstrated in 196869 by Lysyj and coworkers ( 7 ) . By use of an instrumental arrangement consisting of a gas chromatograph, a pyrolysis tube, and a specially designed injection valve, pyrograms for aqueous solutions of glycerol, oL-valine, and hexanoic acid were obtained. Peak intensities for each material were determined and used to generate a series of linear equations. Mixed solutions of these materials were then prepared and also analyzed pyrographically. The concentrations of the individual components in a mixture were determined by relating the intensities of all observed pyrographic peaks to the concentration of each component. The computations were performed by solving least squares linear equations and showed good agreement with known compositions of samples. This experiment demonstrated that, when pyrolytic fragmentation is carried out in the presence of a large excess of water, processes are linear with concentration and independent for each organic compound in a mixture. Pyrographic pattern produced by a mixture of organic compounds is a simple arithmetic summation of the contributing patterns of each compound present. When a single organic compound is pyrolyzed, a number of derivative molecular fragments are produced. A series of peaks with varying retention times and peak intensities result (Figure 1 ) . Many organic compounds on pyrolysis will produce similar products that will manifest themselves on a pyrogram as peaks with identical retention times but varying intensities. These variations produce patterns specific to those of the parent materials.
if 2 or 2000 organic compounds are pyrolyzed simultaneously, the pyrographic representation will be a summation of pyrograms produced by individual compounds. i n this case, each common peak for each component will be a summation of contributions from each component. Specific peaks will, of course, appear independently. Normally, it is possible to solve a series of simultaneous linear equations for the unknown concentrations of organic materials represented by the pyrogram. The maximum number of components for which a solution is possible equals the number of peaks observed on a pyrogram. The pyrogram, of course, can be used for a lesser number of components, and a least squares approach to the solution can be used in this case. The organic compositions into which pyrographic information can be differentiated are defined by calibrations that provide the coefficients for a series of simultaneous equations. The system can be calibrated in terms of molecular entities, classes of organic materia!% or any other arbitrarily defined combination of organic molecules. As a result of these development efforts, the new analytical technique was evolved. Combining instrumental hardware, mathematical logic, and computer programs provided physical means for the application of multicomponent pattern recognition and differentiation theory to problems of environmental pollution control. The instrumental system (Figure 2) was fabricated as a functionally utilized package that can be used under field conditions with a minimum of laboratory support. Analysis of the water samples is performed automatically. No sample preparation, such as extraction or separation, is required. Details of the analytical procedure and operating conditions were published (2). Interpretation of data is aided by two computer programs in basic language-CALIF and BAJA. The CALIF program reduces the raw data produced by a series of pyrographic runs on a given standard solution. The BAJA program calculates the actual concentrations of individual organic components present in an aqueous solution. Test cases The analytical technique has been studied and evaluated in waste source identification and differentiation, industrial effluent analysis, control of waste treatment process. characterization of organic content of natural wa-
TABLE 1
Three-component mixture of industrial wastes Run #
Poultry
Textile
Papel
1 2 3 4 5
0.77 1.08 1.44 0.8 1.03 1.02 1.0%
0.13 0.19 0.12 0.17 0.16 0.15
0.1%
0.51 0.43 0.47 0.44 0.41 0.45 0.5%
1.34 1.89 2.53 1.4 1.79
3.8 5.53 3.29 4.95 4.45
5.27 4.47 4.86 4.55 4.29
1.79
4.4
4.69
Av Present
1 2 3 4 5 Av
ters, and process control in beer brewing and other fields. Its use for stream monitoring, industrial effluent analysis, and waste treatment control are discussed below. To comply with the provisions of current and pending water improvement acts: the discharge of wastes into natural water bodies must be controlled. Practical and economical means for the monitoring and surveillance of many pollution sources contributing to a typical water body are needed. The monitoring system must deal with identities of pollution sources and must be capable of determining quantitative contributions of each source. Monitoring today, however, is directed mainly toward the measurements of water quality parameters and generates information not always relatable in quantitative terms to specific sources of pollution, especially in waterways with multiple waste discharges. Consequently, while there are approximately 500 multiparameter water quality monitoring instruments in use in the United States, they
FIGURE 3
Pyrograms for industrial waste effluents 2x
I Volume 8, Number 1. January 1974
33
TABLE 2
Mixed industrial wastes in natural water matrix Natural background: proteins 1.22 mg/l., carbohydrates 7.47 mg/l., lipids 0.46 mg/l. Volume percent from each source source
Run1
Chemical Co. Textile Finishing
0.48 0.66
ChemicalCo. Textile Finishing
0.91 2.85
2
3
4
0.42 0.44 0.46 0.66 0.87 0.91 Concentration of 0.8 0.87 0.86 2.67 2.7 2.82
5
monitor the impact of pollution on water quality rather than pollution itself. Surveillance of pollution, for it to be an effective tool for maintaining water quality, must provide answers to the following questions: What are the volumes of waste effluents discharged into a particular water system by ail the waste discharges, and how much organic waste . , , does each source contribute to the waterway'! sucn Inrormation should be obtainable by objective physical means. An analytical methodology that can treat and characterize the sum of chemicals-such as found in a typical industrial waste discharge-in an aqueous solution as a single and unique entity must be used. The method of quantitative pyrographic analysis has such a capability. By use of this technique it is possible to define a pyrogram as a standard for a specific waste (Figure 3), and when a number of such wastes are present in a solution, to determine each, both qualitatively and quantitatively. ^
^
Test results Waste effluents were obtained f ,-,,. ,.._,_. ,....-strial operations in Alabama, Florida, Georgia, Mississippi, and South Carolina, including industrial effluents-poultry, chemical, food, textile, paper and pulp, oil refinery, brewh g , pharmaceutical, and service industries. Samples of waste effluents from poultry and paper and textile plants were used in the initial test. Total organic content and calibration constants were obtained and stored in the BAJA computer program for each waste composition. Then the mixtures of the three wastes were prepared in various concentrations and analyzed (Table 1). Results were obtained for a test solution containing 1% waste effluent from a poultry plant, 0.1% waste effluent from a textile plant, and 0.5% waste effluent from a paper mill-l.02% for poultry, 0.15% for textile, and 0.45% for the paper plant. The results also indicate, as might be expected, that the waste load contribution of each waste source is not necessarily proportional to the volumes in the solution. The poultry waste contribution is 1.79 mg/l. of organic waste; the textile, 4.4 mg/l.; and the paper mill, 4.60 mg/l. The textile waste, which constitutes only 9% of all waste volume present in the s o b tion. comprises approximately 40% of the organic waste load of the solution. To perform analysis for waste volumes and organic waste load contribution in natural waters, the magnitude of the natural organic background must be compensated for in the calculation of results. To do so, the computer program BAJA was modified to differentiate the organic background from the total pyrographic pattern for a water sample before differentiating the remaining component'of the pattern in terms of waste load contributions. To test the performance of this program, a number of industrial waste mixtures were prepared in water from the Oconee River, collected in the town of Athens, Ga.. and analyzed (Table 2). Practical implementation of this technique for stream surveillance would require collection of waste discharge samples from all principal operations in the water basin. In actual operation, pyrographic patterns for each source
_-
34
Environmental Science &Technology
Av
Actual
6
7
8
0.5 0.92 1 0.95 0.46 0.5 0.98 0.86 1.0 0.67 0.8 0.91 organic carbon from each source 0.95 0.87 1.75 1.9 1.8 3.84 2.67 2.0 2.48 2.82
9
10
Av
Actual
0.97 0.84
1 0.87
0.97 0.82
1.0 1.0
1.84 2.6
1.9 2.7
1.84 2.54
would be developed and stored i n a computer. Surveiilance of waste load.contributions could then be implemented by collecting and analyzing downstream water samples. The overall patterns of river water can be interpreted, in terms of specific contributing waste patterns, by the computer program BAJA. Results will provide information on the identity of polluters and their relative .~ ~I.L '.._ .L_ _^,,..*:-- ,-,.A ^' wasie c o n i r i w m i i L V me u v e r a i puituuuri I U ~ Y V I L I I C river. The prototype pyrographic analyzer was fabricated in 1968 and underwent extensive field testing. The automated Mark I I pyrographic analyzer was designed and tested ~~. . . . instruments . . . (Mark - . Ill) . . were in 1969. Two production-type fabricate?d in 1972. One IS being used by HocKerayne on a variet:f of research programs dealing with water pollution an(j one was delivered in June 1973 to the U S . Army at Ft. Belvoir, Va.
~~~~.~
..
~~
Pluses osf method The (:onventional qualitative and quantitative analysis for organ i c content o i water samples using instrumental method: 3 , such as gas chromatography, mass spectroscopy or infrared analysis, requires separation of organic matter from the water matrix. Because no sample pretreatment is involved in pyrographic analysis, substantial reduction in man-hours required for analysis can be expected Past fieid experience indicates that operational use of the instrument can be managed by personnel with limited technical training. The training of operating personnel must include some instruction in the use of a teletype for transfer of data from instrument to computer. The instrument can generate qualitative and quantitative information on organic composition of an aqueous sample in 15-30 min. The results of the completed study indicate that pyrography is ready to take its place as a major tool in pollution control by providing a means for rapid and effective characterization of organic composition in natural and waste waters. Additional reading (1) Lysyj, I.. Nelson, K., Webb, S. R., Water .Res., 4, 157 (19701. (2) Lysyj, I., Newton, P. R . , Taylor, W. J.. Anal. Chem. 43, 1271-81 (1971).
lhor Lysyj, principal scientist at Rockefdyne, i s responsible For the development of advanced instrumentation and sensing devices under a number of government contracts. He is the author of more than 40 publications dealing with various aspects of analyfical chemistry.
