Tải bản đầy đủ - 0trang
IV. Methods of Chemical Analysis
DALE E. BAKER AND LEON CHESNIN
50% error. From results for six sewage treatment plants of Pennsylvania
being sampled every 2 weeks, prepared and analyzed in duplicate, a 50%
coefficient of variation over time appears to be a realistic goal for a composite sample from a treatment plant.
In a discussion of the criteria for judging acceptability of analytical
methods, McFarren et al. (1970) point out that a method must be sufficiently precise (measured by coefficient of variation within one laboratory)
and sufficiently accurate (mean error from collaborative studies) if the
results are to be sensible and unbiased. Generally the results for trace elements are biased on the positive side especially when their concentrations
approach the detection limit of the procedure. The total error is defined
as the sum of two standard deviations plus the mean error expressed as
a percentage of the “true” value. Excellent methods have total errors of
25% or less; acceptable methods have total errors of 50% or less; and
unacceptable methods have total errors greater than 50%. McFarren et
al. (1970) concluded that atmic absorption spectrometry was acceptable
for the determination of Zn, Cr, Cu, Mg, Mn, Fe, and Ag but unacceptable
for the determination of Pb and Cd.
Adequate definitions of precision and accuracy are difficult (Murphy,
1961), especially when applied to an overall process or a “system of
causes” including the material, operator, instrument, laboratory and day.
Verification of the precision or accuracy is another measurement process
distinct from the one existing for the purpose of testing materials on a
routine basis. Chow et al. (1974) report a study in which prepared unlabeled samples of sea water were standardized for Pb at one university
by isotope dilution and circulated among participating oceanographic laboratories at seven United States universities and one in the United Kingdom. None of the laboratories obtained reliable values by either atomic
absorption or anodic stripping voltammetry.
Whitney and Risby (1975) suggested that methods of analysis should
be judged on the basis of seven factors: (1) required sensitivity, ( 2 ) accuracy of the method, (3) presence of interferences, (4) time per sample,
( 5 ) number and technical skill of laboratory personnel required, ( 6 ) required use of standard or reference methods, (7) cost per sample. Their review included 224 references providing an excellent summary of the current
status of optical, electrochemical, neutron activation, and chromatographic
methods. For optical methods, theoretical considerations are presented
for colorimetry, spectrophotometry, atomic fluorescence specrtrometry,
X-ray fluorescence spectrometry, and atomic absorption spectrometry.
Electrochemistry techniques are discussed for polarography, anodic strip-
CHEMICAL MONITORING OF SOILS
ping voltammetry, and ion selective electrodes. No one technique will enable the analysis of all desired elements and simultaneously solve problems
associated with all factors above. The choice of a method involves a series
of compromises. Whitney and Risby developed a system of factor weightings to compare methods for first row transition metals. With respect to
Zn, for example, on a scale of 0 for poor to 100 for excellent methods,
colorimetry was given a value of 90 for cost per sample, 62 for sensitivity,
and only 55 for technical skill and number of personnel; while Zn by
atomic absorption spectrometry rated 92, 98, and 75 for the respective
factors. When sensitivity and accuracy weightings were readjusted to three
for sensitivity and accuracy and five for interferences, the results presented
in Table VII were obtained. Morrison and Pierce (1974) and Lisk ( 1974)
have reviewed methods of analyses of trace elements. The review of Lisk
was especially valuable in that specific references are included for several
instrumental methods and their applications for different elements. Because
of the valuable contributions by Walsh (1971), Whitney and Risby
( 1975), and Lisk (1974), a discussion here of the theoretical aspects of
the various methods seems unnecessary.
The results presented in Table VII and developments in flameless atomic
Results of an Unequal Weighting of Factors Considered Important in the Analysis
of First Row Transition Metals”
- 65.0 70.0 - 72.4 72.4 - 69.6 75.0 72.4
67.5 71.5 75.1 76.5 78.9 76.5 71.5 73.9 80.5 75.3
- 81.7 81.7 89.7 89.3 88.7 89.1 84.1 89.7 90.7
- 76.3 78.7 74.9 78.3
80.3 80.3 80.3
81.8 87.0 87.2 91.8 91.8 89.2 89.0 86.6 91.0 91.4
83.7 83.7 83.7 86.1 86.1 86.7 87.7 87.5 83.7 88.7
89.3 89.3 89.3 89.3 89.3 89.3 89.3 89.3 89.3 89.3
74.8 72.2 79.8 79.8 79.8 76.4 79.8 62.2 79.8 19.8
81.3 81.3 85.9 81.3 81.3 81.3 81.3 86.3 86.3 81.3
a R. G. Whitney and T. H. Risby, “Selected Methods in the Determination of First Row
Transition Metals in Natural Fresh Water,” Pennsylvania State Univ. Press, University
Park, Pennsylvania, 1975.
bZero rating = poor; 100 = excellent. Methods of equal or maximum ratings are
DALE E. BAKER AND LEON CHESNIN
absorption spectrometry explain the popularity of this method for trace
element analysis. Flameless atomic absorption has been useful for about
60 elements and is especially useful where sample size is limiting. For example, Cd may be determined on the kidney of a single chick or mouse.
Spectrographic methods have been used extensively for plant analysis
work involving several essential elements, Mitchell ( 1956) described arc
emission spectrographic methods, and Jones and Warner (1968) and
Baker et al. ( 1964) described procedures for direct-reading spark emission
spectrographs. The direct-reading spark emission spectrograph has the advantage that it is possible to determine concentrations of several elements
simultaneously. The precision of the method using rotating disk electrodes
is acceptable when the matrix of the samples remains relatively uniform
as with plant samples. Several of the principles of soil testing and plant
analysis with respect to essential macro and trace elements (Walsh and
Beaton, 1973) are applicable to other metals and compounds. Kopp and
Kroner (1965) determined 19 trace elements in natural water with a
direct-reading spectrochemical procedure. The elements determined were
Ag, Al, As, B, Ba, Be, Cd, Co, Cr, Cu, Fe, Mo, Mn, Ni, Pt, Pb, Sr, V,
and Zn. Concentrations in processed samples were in the order of 0.01-100
ppm. While the sensitivity for most of the elements was satisfactory for
detecting potentially toxic elements in agricultural chemicals added to cropland it would not be adequate for most of the trace elements in soil testing
solutions. The echelle grating spectrometer (Matz, 1973) with the argon
plasma jet for atomization and excitation of elements or the use of the
argon plasma jet with conventional emission spectrographs may prove satisfactory for multielement analyses where sensitivity is important and freedom from interferences and wide concentration ranges are encountered.
Neutron activation analysis received relatively low ratings in Table VII
because of technical skill and laboratory personnel requirements. However,
when these services are available, the method is extremely sensitive for
several elements (Haskin and Ziege, 1971 ; Koloczkowski and Jester,
1973). Morrison and Potter (1972) obtained quantitative and qualitative
results for 3 1 elements using chemical group separations and high-resolution gamma spectrometry. The method was used at The Pennsylvania State
University to determine the relative amounts of elements in several samples
to help with the decision regarding which elements present in sewage sludge
should be monitored routinely.
Although gas chromatography is used extensively for the determination
of organochlorine and other pesticides, techniques are being developed for
its use in metal analysis (Chesters et al., 1971; Serravallo and Risby,
1974). With respect to soil monitoring, however, the technique has been
most useful in pesticide residue analysis. Taylor (1970) described the theory and design of columns for the separation of pesticides by order of
CHEMICAL MONITORING OF SOILS
their vapor pressures. Identification techniques and principles of gas
chromatography have been published by Leathard and Shurlock ( 1970).
The fundamental principles of retention and column selectivity are
References to other analysis methods will be discussed with the specific
elements. The tendency among analytical laboratories is to seek greater
and greater sensitivity for each analysis. Although greater sensitivity is desirable when the existing sensitivity is inadequate €or detecting significant
levels of biological activity, the preparation of a larger sample could often
be a realistic alternative to the use of more sensitive methods. However,
it must be realized that interference problems are greater for more concentrated samples. For methods, such as flameless atomic absorption, anodic stripping voltammetry, neutron activation analysis, X-ray fluorescence,
and gas chromatography, which allow the use of relatively small samples,
the accuracy is determined largely by the use of “clean-room” techniques
and finally the competence of the analyst with respect to preparation of
solutions, use of instruments, and evaluation of the data. Once methods
have been made operational, quality control must be an essential part of
the routine laboratory operations; and periodic reevaluation of methods
and techniques should be provided for. Standard reference materials including orchard leaves and bovine liver, available from the National Bureau
of Standards (Table VIII), are most useful in determining the overall accuracy of a method. For trace element analysis such as that for Mo by
some spectrographic methods (Baker et al., 1964) and for Cd in plants
and sewage sludge by flame atomic absorption without background correction, the precision can be very good and yet the accuracy may be in error
by 100 to 1000% of the true value. These are the errors of concern to
Chow et al. (1974). Standards or limits set on the basis of erroneous results are not very useful. For example, if biological material from control
treatments are reported to contain 10 times their actual concentrations of
a toxic metal and the treatment causes an increase of 10 times the control,
then the biological change being induced into the food chain would be
up to 100 times the true value for the control.
Monitoring of Macroelements
Elements present in macro amounts in soils and considered essential
for plants are N, K, Ca, Mg, P, S, and Fe, whereas sodium chloride is
required in macro amounts only by animals. Because of this requirement,
the relative abundance of Na in animal wastes is greater than in most soils.
DALE E. BAKER AND LEON CHESNIN
Biological Standards of the U.S.National
Bureau of Standards"
0.11 f 0.02
25 k 3
45 k 3
0.08 k 0.01
0.28 f 0.04 IDSSMS
0.27 f 0.04 ATA
0.25 f 0.06 POL
126 k 1 I IDSSMS
127 k 6 NAA
133 f 2 ATA
0.36 f 0.08 IDSSMS
0.31 k 0.08 POL
1 . I 2 k 0.04 NAA
1 . I 1 k 0.04 IDSSMS
k 12 IDSSMS
193 k 8 ATA
193 k 3 IDMS
3.23 f 0.26 IDSSMS
SRM, standard reference material; IDSSMS, isotope dilution-spark-source spectrometry; ATA, atomic absorption ;
POL, polarography; NAA, neutron-activation analysis; IDMS,
isotope dilution-mass spectrometry.
* Values in parentheses are tentative.
The monitoring of sodium chloride and other soluble salts in soils has been
adequately developed. An excellent bibliography for the period 1965 to
1970 is available as Serial Number 1425 from the Commonwealth Bureau
of Soils, Harpenden, England.
Chesnin et al. (1 975) stated that salt is the most serious pollutant of
animal manure in western states where irrigation and dry-land farming is
practiced extensively and where salt balances in the soil are critical to plant
growth and economic agricultural food and fiber production. High levels
of salt are fed to cattle to stimulate appetite, increase water intake to prevent the formation of urinary calculi, and to save labor through self-feeding. Rations for self-feeding involving salt additions to control the amount
of feed consumed by dairy and beef cattle contain as much as 5-9% salt.
In Colorado some cattle are being fed 10-100 times more salt than is
CHEMICAL MONITORING OF SOILS
needed to achieve optimum gains. Most of the salt is eliminated in the
urine of the animals, and a smaller amount is in the feces. Under feedlot
or confined feeding conditions, the solid or slurry wastes can be relatively
high in salts. Samples of feedlot manures (Table IX) from the desert
Southwest confirm the high salt and Na content of these wastes. The soluble
salts in feedlot wastes range from 4.2 to 14.3% (Chesnin et al., 1975).
The Na content of feedlot manure usually ranges from 0.3 to 2.8%. Usually the soluble salt fraction of feedlot manure is dominated by K, which
exceeds Na content by about 10-fold. Manure content of K ranging from
1.2 to 10.7% on a dry matter basis is also dependent on the amount of
urine present in the waste. According to Miner (1971), K in dairy cattle
manure ranged from 0.34 to 3.0% on a dry basis, which is lower than
the range indicated above for beef cattle feedlot manure. Since less salt
is fed to dairy cattle and self-feeding of high salt rations is not a common
practice, Na in dairy cattle manure is generally lower than in beef cattle
In a study of the composition of chicken manure collected in southern
California, Bell ( 197 1 ) found that the average salt content of 40 samples
was 6.15% . The salt content of poultry manure is generally less than that
of beef cattle manures in a feedlot environment. Poultry rations tend to
be low in salt content, and this is evident in the manure (Table X). K
content of poultry manure is higher than that of Na (Chesnin et al., 1975).
Some Chemical Characteristics of Feedlot Manure
from the Phoenix, Arizona Area0.b
4 Feedlots X
Data for feedlots 1 , 2, and X are from Stubblefield and Smith (1964).
Data for feedlots A through E are from Abbott (1968).
All figures reported on an oven-dry basis.
DALE E. BAKER AND LEON CHESNIN
Composition of Fresh-Dried Poultry Excreta
12 Samples of dried
Uric acid and salts
Mujor minerals ( %)
Truce elements (ppm)
Lowman and Knight (1970).
Sheppard (1970). Data for Mg, Fe, Mn, Cu, Co, and Zn from 2 samples only.
On an oven-dry basis, K in poultry manure ranges from 1.0 to 4.5%. The
average of a large number of samples from Georgia was 1.70 and 1.88 %
K for broiler and hen manures, respectively.
The monovalent cation K is generally not viewed with alarm, although
present in much higher amounts in manures than is Na. The presence of
large amounts of K in manures may be an additional negative factor deserving of further consideration.
In a field study (Cross et al., 1973) of the influence of heavy applications
of manure from beef cattle feedlots to an irrigated soil in eastern Nebraska,
3.9 pounds of Na and 13.4 pounds of K were removed per inch of runoff
water. The plots had received 260 tons (dry weight basis) of manure per
acre. The hydraulic conductivity of disturbed samples, measured 4 months
after the application of manure, decreased. This decrease was related to
the high levels of Na and K in the percolate.
Murphy et al. (1972) found large increases in exchangeable Na and
CHEMICAL MONITORING OF SOILS
K contents of a Kansas soil after the application of 324 tons of manure
per acre. The electrical conductivity of the soil saturation extract increased
linearly with rate of application of manure. Crop yield reduction occurred
when more than 140 tons of manure per acre was applied probably because
of the soluble salts in the manure.
Weeks et al. (1972) found that application of 193 tons of manure (wet
weight) during the winter increased the NaCl content of a sandy loam
soil to 1920 ppm. While leaching tended to reduce the salt content, additional applications of manure increased it.
Mathers and Stewart ( 1971) applied 242 tons of manure per acre per
year in western Texas. After one year, the electrical conductivity of the
saturated paste increased to 11.7 mmho. At the end of the second year,
samples from plots that received a second manure application had a conductivity of 10.6 mmho, while samples from plots that did not receive the
second application of manure had a conductivity of 3.0 mmho.
Mathers et al. (1 973) indicated that high rates of application of manure
may increase salinity sufficiently to reduce crop growth. During the early
season, salt injury to plants, especially during germination, is a problem
with application rates above 10 tons of manure per acre, except where
sufficient rainfall or irrigation water is applied.
The problem of salt in animal wastes can be alleviated by reducing the
amount of salt added to the rations. Klett (1973) found that decreasing
the levels of salt added to a cattle ration resulted in a linear decline in
the Na content of the waste produced with no effect on animal
When high rates of manure are applied, the salt content of the soil should
be monitored periodically to determine the level of pollution hazard and
the amelioration procedures needed. Where pastures are grown on acid
sandy soils, a high K status in the soil suggests the possibility of grass
tetany (Baker, 1972).
The effect of land applications of municipal sewage sludge and especially
effluents on the ionic balance and soluble salt content of soils (Kardos
and Sopper, 1973) is critical, especially when it involves possible salt accumulation in arid regions and significant changes in the cationic ratios
within the plant rooting zone of soils which can impair plant growth in
both arid and humid areas. For manure effluent, the concentrations of K
were found excessive compared with other macroelements (Baker et al.,
1975). For municipal effluent, Kardos and Sopper ( 1973) concluded that
only small changes in soil chemical quality occurred after 6 years of sewage
effluent treatment and that these changes posed no problems for the future.
Soluble salt buildup in soils from applications of animal wastes and municipal effluent is more easily estimated and controlled than is that resulting
DALE E. BAKER AND LEON CHESNIN
from the use of NaCl and CaCl, on highways. Westing (1969) estimated
that six million tons of salt were used on the highways of the northern
states in 1969, and the rate was expected to increase to 10 or 12 million
tons. Thus, the amount of salt used to keep ice off the highways in the
northern United States is about equal to the amounts of commercial fertilizers used for crop production by all 48 states. A typical highway in
New England receives up to 20 tons of salt per mile or 4 pounds per linear
foot along each side of the road. Although winter salting is essential for
safe, uninterrupted use of highways, these amounts of salt can be harmful
to vegetation for some distance away from the roadside. In addition, the
excess salt can be poisonous to wildlife and adds to the pollution of
streams, lakes, and groundwater. The use of snow fencing or windbreaks,
more plowing and less salting, highway engineering modifications to prevent drifting and provide drainage systems to keep the salt off cropland,
and more use of CaCl, and less use of NaCl could prevent the salt damage
to vegetation and reduce pollution of streams and lakes. Soluble salts are
determined routinely in soil testing where soluble salt problems are expected. Methods of analysis and their interpretation are adequate and will
not be discussed. The presentation by Bower and Wilcox ( 1965) includes
adequate detail regarding procedures and instrumentation.
Even though some algae fix atmospheric N, environmental problems associated with soil nitrogen, N, include its contribution to eutrophication.
In addition, toxic levels of nitrate (NO,) and nitrite (NO,) in water and
more recently carcinogenic nitrosamines have been under investigation. Nitrogen is usually the most deficient of the essential plant nutrients in soils
for nonlegume crops. Almost all the manufactured N and about an equal
amount in the form of animal manure reach United States soils each year
(Lathwell et al., 1970). The agricultural N requirements for the United
States are estimated to be 16.8 million metric tons per year (Alexander
et al., 1972). The role that agriculture plays in the N pollution of groundwater, streams, and lakes is being studied. Results of studies by Bower
and Wilcox (1965) indicate that soils vary greatly with respect to the
downward movement of NO,- and N losses through denitrification.
The biochemistry of nitrogen oxidation and the production of nitrosamines and related compounds which are reported to be carcinogenic, teratogenic, and/or mutagenic when present in food products, are under investigation (Ayanaba and Alexander, 1973, 1974). An excellent review on
N in the environment has been prepared by a committee of the National
Research Council (Alexander et al., 1972).
CHEMICAL MONITORING OF SOILS
The current situation with respect to N fertilization of crops is somewhat
analogous to the “soil acidity Merry-Go-Round” (Jenny, 1961 ) . Systems
of management to make the most efficient use of available N either from
crop residues, manure, or commercial fertilizers were developed during
most of the first half of the twentieth century when the supply of available
N was the dominant limiting factor for crop production. With declining
costs for N after World War 11, the question of efficient recovery became
a declining issue. The objective was to fertilize until the value of the expected yield increase no longer compensated for the cost of the last increment of fertilizer.
Public concern regarding possible N and P pollution of streams, lakes,
and groundwater required a reassessment of the present practices of N
fertilization of crops. The recovery of fertilizer N by cultivated crops often
represents only 3 0 4 0 % . Allison (1955) and Cooke (1964) reported recoveries in the range of 70-100% for pastures. Field studies of Owens
( 1960) using 15N showed 15-25 % recovery in corn stover and grain from
150 pounds of N per acre from NH,NO,. Leaching losses accounted for
5-20%, 38% remained in the soil, and 33% was presumably lost by denitrification. Broadbent and Clark (1965) showed nitrogen losses of 1 4 0 %
for greenhouse studies and cited results for field studies of N losses in excess of 50%.
Denitrification may be one solution to N pollution problems. Tofflemire
and Van Alstyne (1974) reviewed several projects where systems were
being developed to achieve biological nitrification and subsequent denitrification where wastewater and sewage sludge were applied. With the current
energy crisis and rapidly increasing cost of commercial fertilizer, the problem has changed from one of obtaining highest possible economic yields
without pollution of water to one of obtaining greater efficiency from the
Stevenson and Wagner (1970) presented an excellent review on the
chemistry of N in soils. From the principles developed, crop rotations and
management system decisions are possible without additional experimentation (Lathwell et al., 1970; Bouldin et al., 1971). On the other hand,
monitoring techniques to relate soil properties to biological and chemical
parameters will become increasingly important. For example, the practice
of fall application of N for the production of corn the following year might
have been economically feasible at some locations and might have been
equal to spring plow-down when the N rates were maximum. However,
maximum yields of corn have been obtained in Pennsylvania for three years
with only 75 pounds of N per acre per year when applied as a side-dress
application with an application of 20 pounds per acre banded at planting
time (Shuford and Baker, 1974).
DALE E. BAKER AND LEON CHESNIN
The feasibility of using controlled-release fertilizers is being studied extensively (Lunt, 1971). The problem of concern is the development of
systems to supply N at desirable levels of availability at the time it is needed
by the crop fertilized and recover the remaining amounts in subsequent
crops. Such a system using dairy cattle manure effluent has been proposed
by Baker et al. (1975). Where leaching occurs during late fall, winter,
and early spring, cover crops are required to remove residual NOs--N and
the N released by nitrification. On many soils, small amounts of N banded
in the row will enable corn to grow normally and utilize the N released
by nitrification until silking time. During grain development, relatively large
amounts of N are utilized, so side-dress applications are important. Other
management systems involving the use of N-SERVE to retard nitrification
has been studied by Boswell et al. (1974). The merit of a system to increase the efficiency of N uptake to decrease water pollution and maximize
production has been demonstrated by Bouldin et al. ( 197 1) .
Chesnin et al. (1975) point out that the recent history of high cost and
shortages of commercial fertilizers has stimulated a new interest in animal
manures. In a short time, animal manures changed from a waste delivered
free by the producer to a resource picked up and paid for by the consumer.
A continuation of this trend could partially restore the position manures
enjoyed as suppliers of crop nutrients prior to the large-scale production
of chemical fertilizers. Manure is without question an important crop production resource. Recommendations for the application of animal wastes
are usually adjusted to the nitrogen content of the material and the N needs
of the crop.
Nitrogen in feedlot manure is mostly in an organic form. Chesnin et
al. (1975) found the N content of feedlot manure (dry weight basis) in
11 feedlots in the Phoenix, Arizona, area to range from 1.29 to 2.66%,
with an average of 1.67%. Petersen et al. ( 1971 ) reported on the distribution of N in slurry manure ( a mixture of feces and urine) of confined
beef cattle feeding in Nebraska. The total N, NH,+-N, and NO,--N contents
of the waste were 1.9%, 0.33%, and 0%, respectively (Table X I ) . Additional manure composition data are presented in Tables XI1 and XIII.
The N content of dairy cattle manure varies from less than 1 % to about
4% (dry weight basis); almost all this N is in the organic form (Chesnin
et al., 1975). Hart (1960) reported the N content of poultry and sheep
manures was 5.4%, whereas the N content of dairy, beef cattle and swine
manures as 3.5%, 3.1 %, and 3.3%, respectively. Management practices,
such as the use of bedding, method of handling manure, and the diet of
animals, will have a marked influence on the N content of the waste.
Research by one of the authors in Nebraska has shown that so-called
“untidy” feedlots, with accumulated manure and the soil surface being