EPA-660/2 74 021
MARCH 1974
                       Environmental Protection Technology Series
   Analysis of Coprostanol, an  Indicator
   of Fecal Contamination
                               Office of Research and Development

                               U.S. Environmental Protection Agency
                               Washington, O.C. 20460

-------
            RESEARCH REPORTING SERIES
Research reports of the  Office  of  Research  and
Monitoring,  Environmental Protection Agency, have
been grouped into five series.  These  five  broad
categories  were established to facilitate further
development  and  application   of   environmental
technology.   Elimination  of traditional grouping
was  consciously  planned  to  foster   technology
transfer   and  a  maximum  interface  in  related
fields.  The five series are:

   1.  Environmental Health Effects Research
   2.  Environmental Protection Technology
   3.  Ecological Research
   4.  Environmental Monitoring
   5.  Socioeconomic Environmental Studies

This report has been assigned to the ENVIRONMENTAL
PROTECTION   TECHNOLOGY   series.    This   series
describes   research   performed  to  develop  and
demonstrate   instrumentation,    equipment    and
methodology  to  repair  or  prevent environmental
degradation from point and  -non-point  sources  of
pollution.  This work provides the new or improved
technology  required for the control and treatment
of pollution sources to meet environmental quality
standards.

-------
                                          EPA 660/2-74-021
                                          March 1974
          ANALYSIS OF COPROSTANOL,

    AN INDICATOR OF FECAL CONTAMINATION
                     BY

             J.  Edward Singley
             Cliff J.  Kirchmer
               Ryosuke Miura
           College of Engineering
           University of Florida
        Gainesville, Florida 32611
            Project  16020 EVG
          Program Element  1BA027
               Project Officer

           WILLIAM T. DONALDSON
Southeast Environmental Research Laboratory
           College Station Road
          Athens,  Georgia  30601
                Prepared for

     Office  of  Research and  Development
   U.S. Environmental Protection Agency
         Washington, D.C.  20460
   For sale by the Superintendent of Documents, U.S. Government Printing Office
              Washington, P.O. 20402 - Price $1.60

-------
            EPA Review Notice
This report has been reviewed by the Water
Quality Office, EPA, and approved for publication.
Approval does not signify that the contents
necesarily reflect the views and policies of
the Environmental Protection Agency, nor does
mention of trade names or commercial products
constitute endorsement or recommendation for
use.
                   ii

-------
                             ABSTRACT
The gas chromatographic analysis of coprostanol was improved by show-
ing that it was not necessary to esterify samples prior to injection.
A method of sample preservation was developed using one ml of concen-
trated H2SO^ per liter of sample.  Field surveys compared coprostanol
analyses with total and fecal coliform and confirmed the predicted ad-
vantages of a chemical method over a biological method.

The gas chromatographic method was used in extensive field surveys, in
degradation studies, in treatment plant efficiency studies and as the
standard for evaluation of a colorimetric method.  The method was cap-
able of determining twenty nanograms per liter, which was estimated to
be equivalent to approximately two coliforms per 100 ml.

An attempt was made to correlate coprostanol concentration with treat-
ment plant efficiency.  There was a reasonably good correlation between
coprostanol and BOD, COD and TOC.  At least enough so that further mea-
surements should be made in a large number of treatment plants.

A colorimetric method has been developed that can be used to determine
coprostanol at levels of one yg/1 in polluted water.  The time requir-
ed was estimated to be approximately two hours.  Several samples could
be run simultaneously in the same time.  Analysis of field samples gave
higher results than GLC but it was shown that the most probable inter-
ferants were other fecal steroids.
                                iii

-------
                             CONTENTS


ABSTRACT                                                            ill

CONCLUSIONS                                                           1

RECOMMENDATIONS                                                       3

INTRODUCTION                                                          4
Several Potential Chemical Indicators                                 4

GAS-LIQUID CHROMATOGRAPHY                                             8
Materials                                                            10
Procedures                                                           11
Chromatography of Free Sterols                                       12
Resolution of Coprostanol and Cholesterol                            13
Extraction and Cleanup                                               13
Clean-up with Acetonitrile                                           13
Clean-up with 70 Percent Ethanol and Acetonitrile                    16
Clean-up with N,N-dimethylformamide                                  17
Clean-up with Thin-Layer Chromatography        .                      17
Saponification                                                       18
Wastewater Effluents                                                 18
Activated Sludge Effluent                                            24
Raw Sewage                                                           25
Trickling Filter Effluent                                            25
Stability of coprostanol toward Microbial Degradation                27
Procedure                                                            28
Results                                                              28
Activated Sludge                                                     28
Preservation of Samples                                              36
Field Surveys                                                        38

CORRELATION OF COPROSTANOL CONCENTRATION WITH OTHER
MEASURES OF DOMESTIC WASTEWATER TREATMENT PLANT EFFICIENCY           44
Procedure                                                            45
Results                                                              46
Statistical Analysis of Results                                      52
Discussion of Results                                                62

DEVELOPMENT OF A COLORIMETRIC METHOD                                 63
Introduction                                                         63
Recovery of Coprostanol                                              64
Oxidation of Cholesterol and Coprostanol                             67
Formation of Coprostanone-2,4-Dinitrophenylhydrazine and
Removal of Unreacted 2,4-DNPH                                        77
Color Development                                                    79
Summary of Colorimetric Method                                       96
Applications of Colorimetric Method                                  97

-------
                             CONTENTS






Summary of Field Studies                                           107




ACKNOWLEDGEMENTS                                                   108




REFERENCES                                                         109




APPENDIX                                                           113
                              vi

-------
                              FIGURES


 1  Planar and  Perspective Representations  of Cholesterol             6
    and Coprostanol

 2  Structural  Differences Between Coprostanol,  Cholesterol           7
    and Epicoprostanol

 3  Chromatogram of Free Coprostanol                                 14

 4  Chromatograms Showing Resolution of Cholesterol and              15
    Coprostanol

 5  Environmental Sample—No Clean-up  by Thin-Layer                  19
    Chromatography

 6  Environmental Sample—Clean-up by  Thin  Layer                     20
    Chromatography

 7  Chromatogram—Extract of Effluent  from City of Gaines-           22
    ville Trickling Filter Plant Before Saponification

 8  Chromatogram—Extract of Effluent  from City of Gaines-           23
    ville Trickling Filter Plant After Saponification

 9  Rate of Change of Coprostanol Concentration in Sample            31
    1 and 2

10  Rate of Change of Coprostanol Concentration in Sample            32
    3 and 4

11  Total Coliform—Rate of Change in  Concentration                  33

12  Fecal Coliform—Rate of Change in  Concentration                  34

13  48-Hour Plate Count—Rate of Change in Concentration             35

14  Chromatogram of Extract of Return  Activated Sludge               37

15  Sampling Stations, Sweetwater Creek-Payne's Prairie              40

16  Sampling Stations, University of Florida Sewage Treatment        41
    Plant—Lake Alice

17  Coprostanol vs COD Campus Plant—Primary Treatment               53

18  Coprostanol vs BOD Campus Plant—Primary Treatment               54

19  Coprostanol vs COD Overall                                       55
                                vii

-------
                              FIGURES
20  Coprostanol vs BOD Overall                                       56

21  Coprostanol vs TOG Special Study on Municipal Contact            57
    Stabilization Plant

22  Scheme for the Fractionation                                     65

23  Oxidation of Sterols and Formation of Coprostanone with          70
    the Sarett Reagent at 20°C

24  Gas Chromatograms of Oxidation Products with the Sarett          72
    Reagent

25  Yields of the Oxidation Products of Cholesterol with the         74
    Sarett Reagent at 20°C

26  Yields of the First Oxidation Product of Cholesterol at          74
    Various Temperatures

27  Yields of the Second Oxidation Product of Cholesterol at         76
    Various Temperatures

28  Effect of Cr03 Dosage on Coprostanone in the Oxidation           76
    of Sterols with the Sarett Reagent at 80°C

29  Hydrolysis of Coprostanone-2,4-DNPH' with Acetic Acid—          81
    Concentrated HC1 (1:1)

30  Effect of Temperature and Time on the Hydrolysis of              82
    Copros tanone-2,4-DNPH'

31  Absorption Spectra of the 2,4-dinitrophenylazo—Deriva-          84
    tive of l-(N-naphthyl)-ethylenediamine

32  Absorption Spectra of the 2,4-dinitrophenylazo—Deriva-          85
    tive of a-Naphthylamine

33  Effect of the Periodate Dosage on the Stability of the           87
    Azo Compound

34  Calibration of 2,4Dinitrophenylhydrazine                         88

35  Calibration of Coprostanol  (3-5 jig Range)                        90

36  Calibration of Coprostanol  (0.25-3 yg Rang^                      91
                               viii

-------
                              FIGURES
37  Recovery of Coprostanone by the Combined Procedure of            94
    Saponification and Oxidation

38  Comparison of the Colorimetric Method Values and GLC             99
    Values

39  Typical GLC Pattern of the 5 Percent Ethylacetate Frac-         100
    tion of Sewage Samples

40  Gas Chromatograms of the 5 Percent Ethylacetate and             104
    Ethylacetate Fractions from an Alumina Column

41  Gas Chromatograms of Hexane Fraction of Sewage Samples          105
                                ix

-------
                              TABLES
 1  Survey of Effect of Saponification on the Analysis of            21
    Wastewaters

 2  Effect of Saponification on the Analysis of Effluent             24
    from the Activated Sludge Process

 3  Effect of Saponification on the Analysis of Raw Sewage           25
    Influent

 4  Effect of Saponification on the Analysis of Trickling            26
    Filter Effluent

 5  Data for Sample 1—Trickling Filter Effluent Before              29
    Chlorination

 6  Data for Sample 2—Trickling Filter Effluent Before              29
    Chlorination

 7  Data for Sample 3—University Sewage Plant Effluent              30
    After Chlorination

 8  Data for Sample 4—University Sewage Plant Effluent and          30
    Surface Water

 9  Preservation Study Results                                       38

10  Survey Data for Sweetwater Creek—Payne's Prairie                42

11  Survey Data for University of Florida Sewage Treatment           43
    Plant—Lake Alice

12  Data from Campus Plant—Primary Treatment                        47

13  Data from Campus Trickling Filter Plant                          48

14  Data from Campus Contact Stabilization Plant                     49

15  Data from Campus Plant Tertiary Treatment                        50

16  Data from Municipal Contact Stabilization Plant                  50

17  Data from Municipal Trickling Filter Plant                       51

18  Data from the Special Study on the Municipal Contact             51
    Stabilization Plant

19  Precision Test Results                                           52

-------
                              TABLES
20  Summary of Statistical Analyses                                  58

21  Estimation of Error in Recovery by Solvent Extraction-           66
    Washing

22  Effect of Saponification                                         68

23  Effect of Saponificatnon on Preserved Samples                    68

24  Elution Pattern on an Alumina Column                             75

25  Effect of Extraction Conditions for Coprostanone-2,4-            78
    DNPH

26  Time and Temperature Effects on Reagent Blanks                   78

27  Absorption Intensities and Reagent Blanks of Color               80
    Development Techniques in Alkali-Basic Solvent Systems

28  Absorption Intensities of the Azo Compounds and the              86
    Reagent  Backgrounds

29  Effect of Pyridine Base  Solution on  the Recovery of the          92
    Azo Compound by  Ethylacetate Extraction

30  Estimate of Analytical Time                                      92

31  Color Background of  Oxidation  Products of Cholesterol            95

32  Separation of  Coprostanone and Cholesterol  on  an Alumina        96
    Column

33  Coprostanol Concentration of Field  Samples

34  Statistics of  the Overestimation of  Coprostanol by the
    Colorimetric Method
                                xi

-------
                            CONCLUSIONS
                        Gas Chromatography

A procedure for analysis of coprostanol was developed that simplified
the existing procedures by eliminating both the preparation of the
trimethylsilyl ether (IMS) derivative and the saponification step.  It
was not necessary to prepare the IMS derivative as adequate resolution
was accomplished on a six foot glass column packed with Gas Chrome Q
using the free sterol.  The saponification step could be eliminated as
the 10-15 percent increase in quantity did not justify this lengthy
step.  It was found that acetonitrile could be used to clean-up the
hexane extract.

The degradation of coprostanol in sewage plant effluent was more rapid
in unchlorinated than in chlorinated samples leading to the conclusion
that the degradation was biologically mediated.  No correlation was
found between coprostanol degradation and either total or fecal coli-
form counts.

A method of sample preservation was developed using either sulfuric
acid or low temperature storage.

Field surveys showed that coprostanol degrades slowly in the environ-
ment and does not correlate with either fecal or total coliform levels.

In a study of the correlation of coprostanol concentration with other
measures of domestic wastewater treatment plant efficiency a good cor-
relation for a linear relationship was shown between coprostanol and
TOG, with a coefficient of determination of 0.977; however, COD and BOD
did not correlate very well.

A simple colorimetric method for the analysis of coprostanol was devel-
oped.  The method eliminated interferences from cholesterol and other
constituents by oxidation with Cr03~pyridine and by simple chromatogra-
phy on an alumina column.  Although the reagent also oxidized coprosta-
nol, the resultant coprostanone formed quantitatively and was separated
on the alumina column.

The color was developed as follows:  Coprostanone was reacted with 2,4-
dinitrophenylhydrazine  (2,4-DNPH) to form the hydrazone which was hy-
drolyzed after the excess hydrazine was removed.  The liberated 2,4-
DNPH was then oxidized with periodate to the diazonium salt and coupled
with a-naphthylamine.  The azo compound had intense absorption, with
£=35,000, at 500 nm in an acidic solution of ethylacetate-acetic acid-
concentrated HC1.

The complete analysis required about two hours.  The method was applied
to several field samples and over-estimated the coprostanol by a factor

-------
of 1.6, as compound to results obtained from GLC.  The compounds that
interfered were shown to have a high probability of being of fecal or
urinal origin.

-------
                          RECOMMENDATIONS
The recommendations can be considered under three basic headings, as
they relate to the areas of study.

Gas chromatography—-It is recommended that (1) additional studies fo-
cus on the preparation of halogenated derivatives that could be detec-
ted by electron capture and thus increase the instrumental sensitivity;
(2) the ratio of free coprostanol to total coprostanol be determined
for a variety of environmental samples in order to evaluate the need
for a saponification step, and (3) the correlation of coprostanol with
some conservative biological parameter needs to be developed.

Treatment plant efficiency—It is recommended that sufficient studies
be conducted to determine the correlation between coprostanol and TOG,
COD and BOD for a wide variety of waste treatment plants.

Colorimetry—It is recommended that the method developed in this study
be used as widely as possible in field study in order to ascertain its
application to a variety of environmental regimes.

-------
                           INTRODUCTION


Total coliform and fecal coliform have been used as measures of water
quality for uses ranging from shellfish  waters to public water sup-
plies.  Unfortunately, coliform organisms may originate from sources
other than feces.  Even the membrane filter procedure for fecal coli-
form gives only 93 percent accuracy for differentiation between coli-
forms of warm-blooded animals and coliforms from other sources [1].
To obtain good results, one must analyze for coliform within 30 hours
after sampling.  A minimum of 24 hours are needed for incubation, so
there is a delay between sampling and knowledge of fecal pollution.
Chlorinated sewage effluents often show low coliform counts even though
water quality is poor.  Because o£ the weaknesses of the coliform tests,
research has continued in the search for indicators of fecal pollution.

               Several Potential Chemical Indicators

1.  Uric Acid—Uric Acid has been used as a Pollution indicator  [2,3].
The amount of uric acid in a specimen is measured by observing the re-
duction in absorbance at 292 nm wavelength resulting from its oxidation
by the enzyme uricase.  Uric acid is the principal vehicle  of nitrogen
excretion for reptiles, birds and insects, and  is the end-point  of pu-
rine metabolism  in man, the higher apes and Dalmatian dogs  [4].  How-
ever, 95 percent or more of the uric acid is degraded by conventional
treatment processes and no uric acid could be detected in the Ohio Ri-
ver even in samples taken near the outfall of the Little Miami River
Sewage Plant [2].  This indicates that only recent pollution of  a
stream with raw waste water could be detected by use of uric acid.

2.  Cholesterol—Cholesterol has been suggested as an indicator  of fe-
cal pollution, and the concentration of cholesterol in natural waters
has been measured  [5].  Since it is known that  cholesterol  in surface
waters could come not only from excreta, but also from eggs, milk,
lard, wool, grease, etc., cholesterol can not be considered a specific
indicator of fecal pollution.  In seawater, cholesterol is  evidently
present naturally at  concentrations close to its water solubility  [6,7],
The origin of  this cholesterol is unknown.  Nonetheless, the use of
cholesterol as an indicator  in fresh water may  be useful; although it
would give false positive results, a more serious problem in assess-
ing water  safety.

3.  Coprostanol—Coprostanol, or  5B-cholestan-33-ol, has  shown  promise
as an indicator  of fecal pollution  [5,8,9].  This  sterol  is a  charac-
teristic  fecal organic compound which  is  found  in the  feces of  higher
animals,  including man [10].  Recently it has been claimed  that copro-
 stanol  is  present in whale  oil  [11].   For  all  practical  purposes,  the
only apparent  source of coprostanol  is the feces  of the  higher animals.
Under normal  conditions,  coprostanol is the major fecal  neutral ste-
 roid.  In addition,  the fecal neutral  steroid  fraction contains vary*-

-------
ing but usually small amounts of the following steroids of endogenous
orgin:  cholesta-5,7-dien-3B-ol (7-dehydrocholesterol), cholesterol,
5a-cholest-7-en-3(3-ol (lathosterol), 53-cholest-7-en-3$-ol, 4a-methyl-
5a-cholest<-7'-en-33-ol (metaosterol) , 5a-cholestan-33-ol (cholestanol) ,
5$-cholestan-3a-ol, 5a-cholestan-3-one, and 5g-cholestan-3-one [12].
It has been claimed that a small part of the sterols is present in the
form of esters [13,14,15].

The microbial conversion of cholesterol to coprostanol in the large
intestines of mammals has been demonstrated [16,17] and the mechanism
of the conversion process has been investigated [18,19,20],  The major
reaction  mechanism has been found to be a direct saturation of the
double bond mediated by intestinal microoganisms.  The pathway for
formation of coprostanol esters has been investigated and the conclu-
sion was reached that the main pathway was an esterification of free
53-cholestan-38-ol and not a hydrogenation of cholesteryl esters [21].
Figure 1 shows the structures of cholesterol and coprostanol.

Studies have shown that the fecal excretion of neutral fecal steroids
in human subjects fed standardized diets containing either saturated
or unsaturated fat ranges from 500-700 mg per day [22-27], the major-
ity of which consists of coprostanol, cholesterol and their esters.
However, studies of the concentration of coprostanol in sewage indicate
concentrations of coprostanol corresponding to about two grams of
coprostanol per capita per day [5],

Coprostanol was first isolated from human feces by Austin Flint, Jr.,
in 1862.  The pharmacologist von Bondzynski characterized the sterol
as an alcohol of the formula CoyH^gO and gave it the name coprosterol
(from the Greek "Kopros," dung).  The name was later changed to copro-
stanol for uniformity of nomenclature.

Coprostanol is the C5~epimer of cholestanol.  The typical saturated
sterols found in nature are A/B trans or 5 a compounds, and the only
exception to this rule is coprostanol.  It is said that epicoprostanol
(53f-cholestan-3a-ol) is thermodynamically more stable than coprostanol
because of steric strain from a axial hydrogens  [28].  Structures of
cholestanol, coprostanol and epicoprostanol are shown in Figure 2.

Coprostanol is a white crystalline solid at room temperature, showing
a melting point of 101°C, A variety of acids form esters with copro-
stanol, including the acetate, propionate and benzoate which exhibits
melting points of 89-90°, 99-100° and 124-5°C respectively.  It is
very soluble in ethanol, ether, benzene,and chloroform, and slightly
soluble in methanol [29].  It is only very slightly soluble in water.
No values for its solubility in water are given in the literature, but
values for a structurally similar sterol, cholesterol, have been given
as 2.6 x 10~8 g/ml at 30.0°C, which corresponds to 26 yg/1 [30].  Co-
prostanol would be expected to show a similar solubility in distilled
water.  However, in primary sewage effluent, values have been reported

-------
HO
                                                      HO
HO
                                                                                                        vo
(   Cholesterol
(Choles-5-en-3£-ol)
    Coprostanol
(5B-cholestan-33-ol)
         FIGURE 1.  PLANAR AND PERSPECTIVE REPRESENTATIONS OF CHOLESTEROL AND COPROSTANOL

-------
as high as 260 yg/1 for cholesterol and 750 yg/1 for coprostanol [5],
which indicate that the amounts of cholesterol and coprostanol found
in sewage are not limited by their solubilities and that much of the
cholesterol and coprostanol is present combined with particulate matter,

Coprostanol has been analyzed by colorimetric methods as well as by
gas-liquid chromatography [31-34], but the colorimetric methods previ-
ously used suffered from a lack of sensitivity as well as the ever-
present possibility of spectral interferences.  Gerson [33] considered
a colorimetric method, which employed a color reagent consisting of
ferric chloride—acetic acid—concentrated sulphuric acid.  It was
not suitable for the determination of less than 75 yg of coprostanol.
This amount would suffice for the analysis of coprostanol in 1-2 liters
of primary treated sewage, but, in many cases, secondary treated sew-
               CH3
                                                       CH3
                                    HO
                                                     H
     Coprostanol
 (5P-cholestan-3$-ol)
                                         Cholestanol
                                      (5a-cholestan-30-ol)
                                  CH3
                                       Epicoprostanol
                                    (5 3-choles tan-3a-ol)
FIGURE 2.  STRUCTURAL DIFFERENCES  BETWEEN COPROSTANOL,  CHOLESTANOL,
                         AND  EPICOPROSTANOL

-------
age contains coprostanol in concentrations too low to be detected by
Gerson's method.  The absorption peak, measured at 440 nm, may ex-
hibit interferences, since sewage is a complex mixture containing
many compounds that may absorb at 440 nm.  Thus, a clean-up step in-
volving digitonide precipitation and/or thin-layer chromatography
(TLC) would be necessary.

Despite the difficulties Involved in developing a colorimetric method
for detecting coprostanol in the environment, it would be well worth
the effort.  A colorimetric method for cholesterol that is capable of
detecting amounts as small as 2 yg has been developed [31],  A similar
method for coprostanol may be feasible.  Advantages of a colorimetric
method would be more rapid analysis, less expensive equipment and less
technical training.  It may be possible to develop a colorimetric test
for total sterols, which could be used to measure the efficiency of
sewage treatment processes, since coprostanol is removed in a properly
designed and operated activated sludge plant.

Another possibility is the use of liquid-liquid chromatography (LLC)
for the separation and analysis of coprostanol.  LLC would seem to be
a good separation method, but the detectors used in LLC are not as
sensitive as those available for use with 6LC,

This project has centered on 6LC as the principal analytical tool for
the analysis of coprostanol because of the excellent separation that
is possible, as well as for the sensitivity of the method, which can
detect concentrations of coprostanol less than 1 yg/1 in the environment.

Samples have been taken in both glass [28] and plastic [9] containers.
Glass containers would seem to be preferable since adsorption onto glass
is less likely than adsorption onto plastic.  There have been no report-
ed studies on the preservation of water samples for the determination
of coprostanol.  Indeed, no mention has been made that preservation may
be necessary.

The purpose of the present study was; (1) to simplify the gas-liquid
chromatographic method for the analysis of coprostanol and thus reduce
the time necessary for an analysis, (2) to develop a colorimetric me-
thod of analysis, (3) to evaluate the use of coprostanol concentration
as a measure of treatment plant efficiency and (4) to evaluate copro-
stanol as measure of fecal contamination in the environment.

                    GAS-LIQUID CHROMATOGRAPHY

The previous studies on the analysis of coprostanol in the environment
have used slightly different techniques for extraction and clean-up,
prior to gas-liquid chromatography (GLC)[5,8,9].

Murtaugh and Bunch were the first to analyze for coprostanol in the
                                 8

-------
environment and used a procedure which consisted of extraction with hex-
ane and clean-up by thin<-layer chromatography (TLC) [5].  Smith and
Gouron also extracted water samples with hexane and used a clean-up by
TLC [8].

Black, Singley and Nordstrand [9] found it only necessary to extract
water samples with chloroform, evaporate the solvent, and redissolve
the residue in chloroform prior to analysis by gas-liquid chromatogra-
phy.  No clean-up by TLC was found necessary for the samples they ana-
lyzed .

Rosenfeld's experimental technique involved a preliminary separation by
column chromatography on alumina, followed by a saponification of the
esters and quantification by GLC.

On the basis of the evidence for the existence of coprostanol esters,
Murtaugh and Bunch included a saponification step in the analytical
procedure for coprostanol.  Samples were refluxed for three hours in
a solution of 7.5 percent potassium hydroxide in 70 percent ethanol
after initial extraction with hexane and prior to clean-up by TLC.

Other investigators have omitted the saponification step [8,9].  No
reasons were given for the omission.  Evidently, either the step was
considered too  time-consuming or the fraction present as esters was
considered too  small  to need recovery.

The basic technique for gas-liquid chromatographic sepatation of the
steroids was established by Vandeheuvel et al. in 1960  [35].  The
technique used  a  column of 2-3  percent SE 30 on 80-100 mesh Chromo-
sorb  W. A numberofsteroid compounds including hydrocarbons, ketones,
alcohols, ether and acetyl esters were found to separate at 222°C
with  no sign of decomposition.

The technique of  Vandenheuvel was applied to the analysis of fecal
neutral steroids  [36].  Some  investigators found that the use of tri-
methylsilyl ether derivatives of the sterols gave  sharper peaks and
increased resolution  [36,37].   The method of Vandenheuvel has been
used, with certain modifications, by all of the investigators of copro-
stanol  in environmental water samples.

Murtaugh and Bunch used a one-eighth inch by five  foot  stainless steel
column packed with 60-80 mesh Chromosorb W coated with  5 percent SE 30.
Oven  temperature  was  235°C and  injector port and detector were both
kept  at 260°C.  Helium carrier  gas was used at a flow rate of 40 ml/min-
utes.  Trimethylsilyl ether derivatives of all samples  were prepared
prior to injection onto the column.  Quantitative  determinations were
based on peak height  compared to a standard curve  prepared from known
reference standards.

Smith and Gouron  used a 6mm by  1,83 m glass tube packed with 3 percent

-------
SE 30 on 80-100 mesh Gas-Chrom Q,  Analyses were conducted using Hew-
lett-Packard Corp. F and M Models 400 and 402 gas chromatographs equip-
ped with hydrogen flame detector systems.  Column temperature was 230°C
and injection port was maintained 260°C.  Nitrogen gas was used as car-
rier at a flow rate of 20 ml/minute.  Free sterols were chromatographed,
but no quantitative determinations were made.

Black et al, used a 4 mm by 4 foot glass U tube packed with 3 percent
OV17 on 60-80 mesh Gas-Chrom Q.  A Research Specialties Company Model
600 gas chromatograph equipped with a Sr ionization detector was used
for all analyses.  Argon was used as carrier gas with inlet pressure
being maintained at 30 Ib/inch.  Column temperature was maintained at
240°C or programmed from 150-240°C, and the injection port was preheat-
ed to 270°C.  Peak heights obtained from the chromatography of standard
solutions were used for quantitative analysis.  Free sterols were chro-
matographed and found satisfactory.  However, for verification of co-
prostanol, silyl ether derivatives were prepared on column by bis-(tri-
methylsily1)acetamide.

                             Materials

All solvents used in his study were reagent grade or its equivalent.
Pesticide quality hexane (Matheson, Coleman and Bell) was used as the
extracting solvent.  Reagent grade 95 percent ethanol as well as ace-
tonitrile (Mallinckrodt Chemical) were used in the clean-up procedure.

Coprostanol was supplied in pure form by K and K Laboratories, Inc.,
Plainview, New York, as well as by Applied Sciences, Inc., State Col-
lege, Pennsylvania.  Cholesterol was obtained from K and K Laborator-
ies, Inc., and cholestane was purchased from Applied Sciences, Inc.
Coprostanol was judged pure on the basis of melting point determination
as well as gas chromatographic data.

Membrane filter media (Bacto-m Endo Broth MF, Bacto-m FC Broth Base
and Bacto-Rosolic Acid) were purchased from Difco Laboratories, Detroit,
Michigan.  Disposable plastic petri dishes (48 x 8.5 mm) were obtained
from Millipore Corporation, Bedford, Massachusetts (see Figure 3).

Three percent SE-30 on Gas-Chrom Q served as liquid and support phase
for gas chromatography*  Three percent QF1 on Gas Chrom-Q was used for
confirmatory analyses*  Gas-Chrom Q is an acid washed, silanzied, di-
atomaceous earth.

Glass sampling bottles were used for all samples.  Extraction was done
in two-liter separatory funnels equipped with Teflon stopcorks to avoid
contamination from stopcork grease.  Either a Bilchi Evaporator (Rinco
Instruments Company) or a Buchler Flash Evaporator (Buchler Instruments)
were used to reduce the volume of hexane extracts.  A 500 ml Kuderna-
Danish Evaporative Concentrator equipped with a 5 ml receiver was used
to collect the reduced volume of hexane (Ace Glass Company)»  With this
                                10

-------
apparatus it was possible to remove the 5 ml receiver with the concen-
trated hexane extract and thus avoid loss of extract on transfer,  A
10 yl syringe (#701 syringe Hamilton Company, whittier, California)
was used for gas chromatography.  Six feet by one-fourth inch coiled
glass columns (Applied Sciences) were employged in gas chromatography.

A. F. & M, Corporation Model 810 gas chromatograph, equipped with both
flame ionization and Ni electim capture detectors was used.  The re-
corder was a Sargen Model SR, reading one millivolt full scale.

                            Procedures

Most extractions of water samples were done in the following manner.
Exceptions are noted in discussion of specific procedures.  One liter
water samples were taken, and 2 ml HC1 (cone.) and 5 ml 20 percent
NaCL were added.  The solution was extracted with two 50 ml portions
of hexane.  The hexane extracts were combined and washed with two 25
ml portions of 70 percent ethanol followed by two 25 ml of acetoni-
trile saturated with hexane.  The hexane was reduced to near dryness
on a flash evaporator.  The Kuderna Danish Evaporative Concentrator
was then washed with a small amount of hexane and the concentrated ex-
tract was colleted  in the 5 ml receiver.  The receiver was removed and
the hexane was evaporated to dryness with a  stream of dry air.  After
redissolving  in a known amount of hexane the sample was ready for gas
chromatography,

To correct  injection  errors,  the following technique of filling the mi-
crosyringe was used.  First  a small amount of air was drawn into the
syringe;  The syringe was held  down and  then the sample liquid was
drawn  iti  followed by  more air*  This  gave a  sandwich of liquid sample
between two  slices  of air.   The volume of the liquid sample could be
read accurately by  means of  the gradations of the barrel.  After the
needle was  thrust through the septum,  the plunger  could be pushed all
the way into  the  barrel.  The air ahead  of the  sample permitted accu-
rate measurement  of the volume  of contained  liquid and the air behind
the liquid permitted  all of  the sample to be injected  into the chroma-
tograph.

General conditions  for gas chromatographic analyses were  as follows:
Any variations on these general conditions are  specified  under experi-
mental  procedures.

Column  temperature  was kept  at  250°C, and injector port and flame
ionization detector were maintained at 270°C.   Nitrogen gas flow rate
was approximately 40 ml/minute.  A  6  foot by 1/4 inch coiled glass
column  of three percent  SE 30 on Gas-Chrom Q was for all  analyses
except  those  noted  otherwise,

Coliform analyses were done  using the Membrane  Filter  technique.
Standard procedures,  as outline in  the 13th  edition of Standard Methods
                                 11

-------
for the Examination of Water and Wastewater [1], were used on all ana-
lyses.

The procedure used for analysis of coprostanol to date was somewhat
long and complicated.  Therefore, a series of studies were initiated
on modification or elimination of some of the steps.

Either hexane or chloroform may be used to extract coprostanol from
water samples [5,8,9].  Chloroform will extract more non-lipid materi-
als from sewage than will hexane [38,39,40].  Hexane is a more selec-
tive solvent than chloroform, since it will extract mostly non-polar
material while chloroform also extracts many polar substances,

Acetonitrile, as well as dimethylformamide, was used in pesticide ana-
lyses . to remove the chlorinated hydrocarbons from lipid-like material
[41,42].  The lipid-like material remains in the hexane layer, while
most pesticides, including the chlorinated hydrocarbons, are extracted
into the acetonitrile or dimethylformamide.  It may be possible to a-
dapt the procedure for coprostanol by analyzing for the lipids in the
hexane layer, and discarding the acetonitrile or dimethylformamide con-
taining undesired substances.

There is disagreement over the necessity of preparing the trimethyl-
silyl (TMS) ether derivative of coprostanol prior to gas chromatogra-
phy.  Hurtaugh and Bunch [5] prepared the TMS derivative, while others
[8,9] found that the free sterol gave good results.

The limit of detection of coprostanol using a flame ionization detector
was approximately 20 nanograms [5].  In environmental samples, this li-
mit of detection is difficult to achieve because of interferences from
other substances present in the water.

Picogram amounts of steroids have been analyzed by formation of the
heptafluorobutyrate derivatives and detection using an electron cap-
ture  (EC) detector [43,44,45].

                  Chromatography of Free Sterols

Summary of Results

No extraneous peaks were found dn any of the chromatograms of free co-
prostanol, and since the peaks obtained were sharp with only a small
amount of tailing, it was not necessary to prepare the trimethylsilyl
ether   (TMS) derivatives of coprostanol prior to gas chromatography.

Two conditions must be met in order to chromatograph, free sterols.
First, an all-glass column should be used, Since metal columns have
been known to cause breakdown of steroids.  Copper, in particular,
will catalyze the decomposition.  Deactivation is achieved by silani*-
zing the solid support.  Gas ChronrQ is an acid-washed, silanized
                                 12

-------
support of diatomaceous earth.

In order to use peak height as a quantitative measure of coprostanol,
there should be sufficient resolution of coprostanol from cholesterol,
since cholesterol is invariably found with coprostanol in the environ-
ment.  Complete resolution was not necessary.  Figure 4 showed that
the required resolution of cholesterol was acheived by use of a 6 foot
column.  If one wishes to increase resolution of cholesterol and copro-
stanol, it can be done by increasing the length of the column, lowering
the column temperature, lowering the percent liquid phase on the col-
umn, adjusting the flow rate of the carrier gas, or changing to a more
selective liquid phase.

The formation of the heptafluorobutyrate derivative of coprostanol evi-
dently does not take place as easily as reported in the literature for
plasma estrogen and plasma testosterone.  Klyne stated that for reac-
tions dependent on accessibility of x (esterification and hydrolysis
of esters) equatorial  reacts more rapidly than axial.  Since the con-
figuration of the OH group in coprostanol is axial, the reaction may
need more drastic conditions than those reported in the literature
for the estrogens and  testosterone.

Several injections of  free coprostanol, dissolved in hexane, were made.
Column conditions were as described previously.  A typical chromatogram
of  coprostanol  is shown  in Figure 3.

             Resolution  of Coprostanol  and Cholesterol

Chromatographs  were obtained  for coprostanol, cholesterol and combined
coprostanol and cholesterol.  Cholestane was used as an internal stand-
ard to compare  locations of peaks.  Figure A shows that the coprostanol
and cholesterol peaks  overlap somewhat  but the peak height for coprosta-
nol was the true peak  height when cholesterol was also present.

                       Extraction and  Clean-up

Summary of Results

N,N-dimethylformamide  cannot be used  to clean-up the hexane extract,
since some of the coprostanol was extracted  into the dimethylformamide.
Acetonitrile was much  more efficient  in cleaning up the coprostanol ex-
tract.  Indications were that very  little of the coprostanol was removed
by  the acetonitrile.   Since acetonitrile has a potential for extracting
some of the interferents in the GLC analysis for coprostanol, a washing
with acetonitrile was  used in all subsequent analyses.

                     Clean-up with  Acetonitrile

Fifty yg of coprostanol  was dissolved in one liter of distilled water,
and the water was then extracted with two 50 ml portions of hexane
                                13

-------
f?
o
o

a
(D
H


ff
CO
•d
o
3
CD
(D
I
CD
O
                                                CONDITIONS

                                                3%  SE  30  on

                                               100-120 mesh

                                                Gas Chrom Q

                                           Column temp. - 250°C

                                          Carrier gas  - Nitrogen

                                             at 40  ml/minute
 0
I

4
i

6
12345678



FIGURE 3.  CHROMATOGRAM OF FREE COPROSTANOL
9 Minutes
                            14

-------
              CONDITIONS

3% SE 30 on 100-120 mesh Gas Chrom Q

6 foot glass column

Column temp.  - 250°C

Carrier gas - Nitrogen at 40 ml/minute
                  1
                  *j
                  to
                  o
                  n
                  ex
                  o
                  u
                 0)

                 4J

                 CO

                 0)
     (U
     CO

     

CO

01

iH

O

43

U
     Chromatogram #1                Chromatogram #2     Chromatograms'ffl & 2             Chromatogram #3

                                                         Superimposed




                 FIGURE 4.  CHROMATOGRAMS SHOWING RESOLUTION OF CHOLESTEROL AND COPROSTANOL

-------
(shaken for 2 minutes).  The hexane extracts were combined and washed
with two 25 ml portions of acetonitrile (shaking for 2 minutes),  The
acetonitrile washings were combined and reduced by flash evaporation
to one ml.  The hexane extracts were similarly reduced to one ml.  Hex-
ane extract, acetonitrile washings and standard coprostanol were run on
a column of 3 percent QF  1  on Gas Chrom Q.  Column temperature was
250°C and both inlet and flame ionization detector were at 270°C.

                          Hexane Extract

          yl Injected                       Peak Height

              3.5                             4.78 cm
              3.5                             4.55
              3.5                             4.20
                                              4.51 Avg.
                       Acetonitrile Washings
              3.5                             0.14
                       Reference Coprostanol
              3.5                             4.66
              3.5                             4.54
                                              4.60 Avg.
          % Recovery of coprostanol in hexane = 4.51 x 100 = 98.1%
          % Recovery of coprostanol in acetonitrile * 0.14 x 100 = 3.04%
                                                      3760
            Total Recovery = 98.1 + 3.04 = 101.1%

         Clean-up with 70 Percent Ethanol and Acetonitrile

Forty yg coprostanol was added to one liter of distilled water. Two ml
HC1 (cone.) and 5 ml 20 percent NaCl were added to the solution.  The
water sample was extracted with two 50 ml portions of hexane (shaking
for 2 minutes) and then washed with two 25 ml portions of 70 percent
ethanol, followed by two 25 ml portions of acetonitrile.  The combined
hexane extracts were evaporated to dryness and then redissolved in 0,5
ml benzene.

The above procedure was repeated on two more samples and GLC was run on
the three extracts, using a column of 3 percent SE 30 on Gas Chrom Q.
Column temperature was 250°C, inlet temperature 270°C, and flame ioni-
zation detector temperature was 270°C.  A standard sample of coprosta-
nol was prepared by dissolving 40 yg coprostanol in 0.5 ml benzene.
Three-aliquot samples of the standard solution were run on GLC.

Data are given in terms of peak height obtained per 1 yl injected into
the column from a total sample of 0.5 ml.
                                16

-------
      Peak Height/yl Injected
        Standard Solution

          3.93 cm/Ml
          4.06 cm/yl
          3.74 cm/pi
                                 Peak Height/yl Injected
                                 	Samples	

                                        3.63 cm/Ml
                                        3.82 cm/yl
                                        3.63 cm/yl
          3.91 cm/pi Avg.

      % Recovered
                                        3.69  cm/yl Avg.
                           94.4% ,
                    3.69 x 100
                    3.91
      The 95 percent C.I. can be calculated as y = x + tn-l (a/2) 5/yra.

                                                 • 94-4 + 4.3 (2.05)
                                                              3
                                                 = 94.4 + 5

                Clean-up with N,N-dimethylformamide

Five hundred ug of coprostanol was added to 50 ml hexane and the hexane
was washed with two 25 ml portions of N,N-dimethylformamide.  The hex-
ane was washed with one 25 ml portion  of distilled water to remove all
traces of N,N-dimethylformamide,  The hexane was evaporated to dryness
by flash evaporation and the dried residue was redissolved in one ml.
The procedure was  repeated on another sample.  A column of 3 percent
SE 30 on Gas Chrom Q was used.  Column temperature was 250°C and inlet
and detector were 270°C.
      Peak Height/yl
         Injected
         Standard

        2.19 cm/yl
        2.55 cm/yl
        2.24 cm/yl
        2.09 cm/yl
        2.31 cm/yl
        2.28 cm/yl Avg.
                    Peak Height/yl
                       Injected
                       Sample 1

                      0.754 cm/yl
                      0.705 cm/yl
                      0.808 cm/yl
                      0.756 cm/yl Avg,
                                               Peak Height/yl
                                                  Injected
                                                  Sample 2

                                                 0.753 cm/yl
                                                 0.748 cm/yl
                                                 0.810 cm/yl
                                                 0.770 cm/yl
% Recovery
Sample 1 = 0.756 x 100
       ;    2.28
      Sample 2
           0.770 x 100
           2.2
                               33.2%

                               33.8%
           Clean-up with Thin-Layer Chromatography (TLC)

A number of environmental water samples were run to determine the neces-
sity of TLC as a preliminary clean-up step.  It was found that TLC was
necessary for samples with concentrations below approximately 2 yg/1,
using one-liter samples, but TLC may be omitted on samples with higher
                                 17

-------
coprostanol concentrations.  Figure 5 shows a chromatogram of an environ-
mental sample with no prior clean-up by TLC while Figure 6 shows a chro-
matogram prior to clean*-up of a sample which required clean-up by TLC.

                          Saponification

Saponification is a time-consuming step in the analytical scheme, re-
quiring three hours out of a total analysis time of four hours.  The
Saponification step could be omitted if the amount of coprostanol in
the form of esters was small.

                        Wastewater Effluents

Summary of Results

The initial survey of the effect of Saponification on the analysis of
wastewaters showed no statistical difference between the saponified and
unsaponified samples.  Since only one. sample was taken from each source,
three more experiments were performed, refining the experimental design.
In these three experiments, 10 repetitions from each source were analyz-
ed before and after Saponification.  It was found that Saponification
had a slight effect on the quantitative determination of coprostanol.
However, the difference was slight, on the order of 10-15 percent.  Such
a slight difference does not seem to justify a three-hour Saponification
step in a procedure that takes four hours to complete the total analysis.
Without Saponification, analysis time was only one hour.  Since the pur-
pose of the analysis is to detect fecal pollution, it would seem that
free coprostanol could do this as well as total coprostanol, including
the esters of coprostanol.  The ratio of free coprostanol to total co-
prostanol needs further investigation for a variety of samples.

The esterification of sterols in feces requires an enzyme system, pro-
bably of bacterial origin.  It is highly unlikely that extensive esteri-
fication of coprostanol could occur in the natural environment.

                             Procedure

Samples were obtained from three wastewater treatment plants in the
Gainesville, Florida area.  The three plants sampled were the City
Plant (2,7 million gallons per day [MGD], contact stabilization and
2.2 MGD, trickling filter), the University Plant (1.3-1,8 MGD, contact
stabilization and 0,7 MGD, trickling filter), and the Sunland Training
Center Plant (0,3 MGD, trickling filter).

Sample  volumes varied from 125 ml to 1000 ml.  Two samples were taken
simultaneously from each location.  A 500 ml and a 1000 ml sample were
extracted with two 50 ml portions of hexane, and the combined hexane
extracts were then washed with two 25 ml portions of 70 percent ethanol
and one 25 ml portion of acetonitrile.
                                 18

-------
                                                                                 CONDITIONS

                                                                     3% SE 30 on Gas Chrom Q,  100-120 mesh

                                                                            Column temp.  - 250°C

                                                                     Carrier gas - Nitrogen  at 40 ml/minute
                                                                               3.42 yl injected
0)
CO
e
o
(X
CO
d)
0)
O
CJ
0)
o
0)
                                                                                                            ffi
 0
6       7

Minutes
T

10
11
 I

12
                  FIGURE 5.  ENVIRONMENTAL SAMPLE—NO  CLEANUP BY THIN LAYER CHROMATOGRAPHY

                                        (Calculated  Cone.  24.3 ug/1)

-------
                        2
i
3
8
                                     Minutes
FIGURE 6.  ENVIRONMENTAL SAMPLE—CLEANUP BY THIN LAYER CHROMATOGRAPHY
                    (Calculated Cone. 0.22 yg/1)
                                 20

-------
The hexane extracts were evaporated on a rotary evaporator using a 500
ml Kuderna-Danish evaporative concentration with a 5 ml receiver.  Re-
sults of these analyses are shown in Table 1.  Figures 7 and 8 show
typical chromato grams from samples before and after saponification.

                             TABLE  1

             Survey of Effect of Saponification on the
                      Analysis of Wastewaters
                        yg/1 Coprostanol
                        No Saponification
                                                 yg/1 Coprostanol
                                                    Saponified
City Influent
University Influent
Sunland Influent
City Effluent
CTrickling Filter)
University Effluent
(Trickling Filter)
Sunland Effluent
(Trickling Filter)
City Effluent
(Contact Stabilization)
University Effluent
(Contact Stabilization)
448
340
755

91.5

72.6

181

16.1

1.8
493
412
741

84.6

70.0

168

9.1

1.0
 Statistical Analysis
         y  %n- o
Ho ~=
HA E
         y  %D
         0
                                       a =  0.2
Sample   No Sapon.   Sapon
                                  % D
  1
  2
  3
  4
  5
  6
  7
  8
_
% D = 9.6
              448
              340
              755
               91
               72
              181
               16



5
6

1
8

493
412
741
84.6
70.0
168
9.1
1.0

-10.04
-21.17
+ 1.85
+ 7.54
+ 3.58
+ 7.18
+43.47
+44.44
76.85
                                                 (% D)2

                                                 100.8
                                                 448.2
                                                  3.42
                                                 56.9
                                                 12.8
                                                 51.5
                                              1,890
                                              1.975
                                              4,538.62
              .        -
      Therefore 0«/D =/i-[Z(%D)2 -  (Z%D)2]
                 MJ v 7            >  £
                                                 23.3
      Then t = %D - 0
   From standard statistical tables, two-sided t2Q%
   Not significant therefore  HQ accepted
                                                          1.415
                                 21

-------
CO
a
o
cu
(A
M
0)
o
0)
      CONDITIONS
3% SE 30 on 100-120 mesh
      Gas Chrom Q
 Column temp, - 250°C
Carrier Gas'- Nitrogen
   at 40 ml/minute
    1.97 yl  injected
                                                                                                        (M
          I         I         f          I          I          I          I          I          1
         123456789
                                           Minutes
 FIGURE 7.  CHROMATOGRAM EXTRACT  OF  EFFLUENT FROM CITY OF GAINESVILLE TRICKLING FILTER PLANT
                                     BEFORE SAPONIFJCATION
                                                                                           10

-------
03
tt>
C
O
P-
CO
5
Pi
M
Q)
O
U
U
0)
                                                                        r-l
                                                                        O
                                                                        H
                                                                        (U
                                                                        4J
                                                                        CO
                                                                        01
    CONDITIONS
3% SE 30 on 100-120 mesh
mesh Gas Chrom Q
Column temp, - 250°C
Carrier gas - Nitrogen
at 40 ml/minute
2.0 yl injected
                                                                                                         OJ
                                         4         5
                                          Minutes
              <
             9
10
                                     FIGURE 8.  CHROMATOGRAM
   EXTRACT  OF  EFFLUENT FROM CITY OF GAINESVILLE TRICKLING FILTER PLANT AFTER SAPONIFICATION

-------
                      Activated Sludge Effluent

A 10-liter sample was taken from the effluent of the contact stabili-
zation unit of the University Treatment Plant,  This sample was divid-
ed into one-liter samples, each of which was preserved by the addition
of 1 ml concentrated ^SO*.  Five ml of 20 percent NaCl solution was
added to each sample, and the samples were extracted in the usual man-
ner with hexane.  The hexane was reduced to dryness, the residue re-
dissolved in 0.5 ml hexane, and chromatographed by GLC.  The concen-
tration of coprostanol in each sample was estimated from peak heights.
The ten samples were then each dissolved in 50 ml of 7,5 percent KOH
in 70 percent ethanol, and then refluxed for three hours.

After this saponification step, each sample was diluted with 50 ml dis-
tilled water, and coporstanol was re-extracted with two 25 ml portions
of hexane.  The hexane was washed with two 15 ml portions of 70 percent
ethanol, evaporated to dryness, and the residue redissolved in 0.5 ml
hexane. The concentration of coprostanol in these extracts was then
determined from the peak heights after gas-liquid chromatography.  Re-
sults of these analyses are given in Table 2,

                              TABLE  2

       Effect of Saponification on the Analysis of Effluent
                 from the Activated Sludge Process

                          yg /I Coprostanol       yg/1 Coprostanol
   Sample Number	Before Saponification    After Saponification
1
2
3
4
5
6
7
8
9
10
4.55
4.75
4.90
4.30
4.30
4.20
4.20
4.65
4.45
4,25
4,92
5.77
5.15
5.49
5.79
6.18
4,74
5.95
5.49
4,28
   Statistical analysis
      Ho E XT - x2                    t = xl " X2
                                            2     2  = 4.38

                                          nl    n2

   From standard statistical tables, ta_Q.i=1.734; result is significant
   Ho can be rejected and it can be concluded that there is a real dif-
   ference between the two methods.  Calculating the 95 percent confi-
   dence interval for sample means we get:  Before saponification 4.48
   + 0.14; After saponification 5.38 + 0.30.
                                 24

-------
                            Raw Sewage

The same procedure as in experiment 2 was used except that the source
of the sample was raw influent from the University Sewage Treatment
Plant.  The five-liter sample was divided into ten samples of 500 ml
each.  Results of these analyses are given in Table 3.

                             TABLE  3

             Effect of Saponification on the Analysis
                      of Raw Sewage Influent

                         yg/1 Coprostanol         Mg/1 Coprostanol
    Sample Number	Before Saponification     After Saponification
1
2
3
4
5
6
7
8
9
10
554
522
576
520
548
536
522
504
502
492
600
551
*
584
590
599
585
549
576
560
*  Sample was lost.

    Statistical analysis
    Ho = x^ = X2                      t = 4.81

    From standard statistical tables, ta=,Qi]_=1.74 result is
    cant, therefore Ho can be rejected and it can be concluded that
    there is a real difference between the two methods.  The 95 per-
    cent confidence interval can be calculated:  Before Saponification
    528 + 15; After Saponification 577 + 12

                     Trickling Filter Effluent

A ten-liter sample was taken from the trickling filter effluent of the
University Sewage Treatment Plant, and then divided into 20 samples of
0.5 liters each.  Each sample was preserved with 1 ml of concentrated
H2S04 and 3 ml of 20 percent NaCL was added.

Ten samples were extracted in the usual manner.  They were not subjected
to Saponification.  After normal extraction the other ten samples were
saponified for three  hours in 50 ml of 7.5 percent KOH in 70 percent
ethanol.  The saponified alcohol solutions were diluted with 50 ml of
water, followed by a re-extraction with two 25 ml portions of hexane,
and washed with two 15 ml portions of 70 percent ethanol.
                                 25

-------
Both unsaponified and saponified samples were analyzed by 6LC.  Results
of these analyses are given in Table 4,

                              TABLE  4

              Effect of Saponification on the Analysis
                    of Trickling Filter Effluent

                         yg/1 Coprostanol           yg/1 Coprostanol
    Sample Number      Before Saponification      After Saponification
1
2
3
4
5
6
7
8
9
10

67.3
65.7
66.9
72.2
84.3
59.9
64.7
69.2
70.2
70.2
Average 69 . 1
67.6
77.8
76.5
82.9
85.9
71.8
82.2
75.9
77.1
67.0
76.4
   Statistical analysis
   HQ = 5E1 = x2                      t - 2.61

   From standard statistical tables ta=0.1SBl'734.  Result is significant,
   therefore, HQ can be rejected and it can be concluded that there is a
   real difference between the two methods.  The 95 percent confidence
   intervals can be calculated:  After Saponification 76,40 + 1.54; Be-
   fore Saponification 69.06 + 1.46

   Precision of Analysis
   For those experiments in whichseveral samples were taken for analyses,
   the precision of the analyses can be calculated.
                          Average Error         Standard Deviation
                          (Expressed as           (Expressed as
                         Percent of Mean)        Percent of Mean)
Raw Sewage
Before Saponification
After Saponification
Trickling Filter
Before Saponification
After Saponification
Activated Sludge
Before Saponification
After Saponification

3.94
2.79

6.02
6.18

4.61
9.23

4,96
3.40

9.23
9.93

4,45
11,2
                                 26

-------
       Stability of Coprostanol toward Microbial Degradation

It has been shown that species of Proactionomyces present in soils
are capable of utilizing coprostanol as their sole source of carbon
for growth [46].  In addition more than 1,000 bacterial strains are
capable of degrading cholesterol which is structurally almost identi-
cal to coprostanol [47].

If the disappearance were due to biodegradation, then the rate of dis-
appearance should be affected by chlorination of sewage effluent.  In
addition, the rate of disappearance of coprostanol is of interest in
determining the usefulness of coprostanol as an indicator of fecal
pollution.

Smith and Gouron have stated that the removal of coprostanol by the
activated sludge process may be due to a physical removal by the floe
rather than biodegradation [8].

Laboratory experiments were performed to determine the rate of disap-
pearance of coprostanol in both chlorinated and unchlorinated sewage
effluents.  At  the time microbiological analyses were made to see
if the disappearance of coprostanol could be correlated with the de-
crease in indicator organisms.  An experiment of exploratory nature
was also performed to determine the mechanism of coprostanol removal
in the activated sludge process.  Experiments were performed to deter-
mine  the necessity of preserving samples.

Summary of Results

It is obvious from the degradation studies that coprostanol decreased
at a  more rapid rate in sewage effluent taken before chlorination than
in samples treated by chlorination.  One explanation is that the disap-
pearance of coprostanol is due to biological degradation.

The apparent decrease in rate of disappearance  of coprostanol with time
can be explained as a mass of action effect due to the decreasing con-
centration of coprostanol with time.

The curve of coprostanol versus time obtained in sample 1 did not fit
a first order equation.  Two factors may determine the rate change of
coprostanol, coprostanol concentration and concentration of organisms
capable of biodegrading coprostanol, since both are necessary for the
biodegradation  to occur.

Results obtained for indicator microorganisms can be divided into two
categories, those obtained from chlorinated samples and those obtained
from  non-chlorinated samples.  In general, the  results obtained for
chlorinated samples were more erratic, showing  decreases after chlo-
rination, followed by regrowth of organisms.  Non-chlorinated effluents,
on the other hand, snowed a general decrease in numbers of organisms with
time.
                                27

-------
Both  indicator microorganisms and coprostanol showed the same general
trend for non-chlorinated samples, decreasing with time,

Chlorination affected both indicator microorganisms and coprostanol
concentration, but in different ways.  While indicator microorganisms
showed an erattic pattern of decrease followed by regrowth, coprostanol
concentration decreased slowly with time.  It would appear from these
experiments that correlation between indicator microorganisms and copro-
stanol may be found where the sewage effluent is not chlorinated.  Where
chlorination is practiced, since the effect of chlorine on microorgan-
isms  and coprostanol is different, no correlation may be expected.

                             Procedure

Four  samples of approximately 20 liters each were taken.  Samples 1
and 2 were trickling effluent taken from the University Sewage Treat-
ment  Plant before chlorination.  The third sample was chlorinated ef-
fluent from the University Plant and the fourth sample was a 1:2 mix-
ture  of chlorinated sewage effluent from the University Plant and a
surface water from a small creek entering Payne's Prairie.  All samples
were  taken to the laboratory and placed in large glass cylinders open
to the atmosphere, where they were held and stirred constantly during
the experiment.

Analyses for coprostanol were made on all samples over a number of days.
Simultaneously with coprostanol analyses on samples 2, 3 and 4 total
coliform, fecal coliform and standard plate counts were taken.  Micro-
biological tests were also made on the creek water before mixing with
sewage effluent.  In order to determine if there was any physical re-
moval of coprostanol due to adsorption and deposition onto the glass
cylinders, they were washed down with 50-75 ml hexane after termination
of the experiments.   The hexane was evaporated down to approximately
5 ml and aliquots were analyzed gas chromatographically.  Results of
the analyses of samples 1-4 are given in Tables 5-8.

                              Results

No coprostanol was detected in the hexane washings of the glass cylin-
der after termination of the experiment.  Figure 9 shows a plot of the
data for change in coprostanol concentration with time found in samples
1 and 2.

Figure 10 shows plots of the rate of change of coprostanol with time
as found in samples 3 and 4.   Figures 11-13 show the rate of change
of microbiological indicators with time as found in sample 2-4,

                    Activated Sludge—Procedure

One liter of return activated sludge from the University of Florida sew-
age treatment plant was taken and extracted with two 75 ml portions of
                                28

-------
                             TABLE  5

 Data for Sample 1—Trickling Filter Effluent Before Chlorination
                  (Sample Taken January 26, 1971)
Time (Hours)
0
3
6
12
24
48
100
144
216
244
Coprostanol (yg/1)
95.5
78,7
86.8
76.7
47.9
21.8
11.4
5.71
5.38
4.0
hexane.  It was found necessary to add approximately 100 ml, 95 percent
ethanol, to break the emulsion that formed.  The hexane was separated,
washed with two 25 ml portions of ethanol and a gas chromatographic
analysis was run on the hexane after reducing the volume to 2 ml by
flash evaporation.

                             TABLE  6

 Data for Sample 2—Trickling Filter Effluent Before Chlorination
                 (Sample Taken February 23, 1971)
Time  (Hours)
Coprostanol
 lie/liter
   Total
  Coliform
Number/100 ml
   Fecal      Standard Plate
  Coliform        Counts
Number/100 ml  Number/100 ml
0
3
6
12
24
48
96
192
70.2
55.8
67.8
54.2
32.7
17.6
_
—
9.7 x 106
4.7 x 106
-
_
4.2 x 105
4.8 x 105
6.2 x 105
4.1 x 105
4.2 x 105
2.9 x 105
-
_
3.4 x 104
2.95 x lO4
1.41 x 10H
5.9 x 103
7.4 x 108
1.04 x 109
-
_
7.3 x 107
4,56 x 107
5.8 x 107
2.15 x 107
No Coprostanol was detected in the hexane washings of the glass cylinder
after termination of the experiment.
                                29

-------
                             TABLE  7

     Data for Sample 3—University Sewage Plant Effluent After
           Chlorillation (Sample Taken February 23, 1971)

Time
(Hours)
0
3
6
12
24
48
96
192

Coprostanol
yg/liter
42.6
39.4
41.1
40.9
47.1
35.2
41.7
29.7
Total
Coliform
Number/100 ml
3.2 x 102
1.7 x 103
-
—
2.08 x 101*
6.6 x 103
TNC(>3xl06)*
2.9 x 107
Fecal
Coliform
Number/100 ml
2
9
-
—
0
2
270
684
Standard Plate
Counts
Number/100 ml
1.4 x 104
2.04 x 106
-
-
5.9 x 105
TNC(>3xl06)*
1.5 x 108
2.9 x 108
 *  TNC—Too Numerous to Count
No coprostanol was detected in the hexane washings of the glass cylinder
after termination of the experiment.


                             TABLE  8

  Data for Sample 4—University Sewage Effluent and Surface Water
                  (Sample Taken February 23, 1971)
                         Total
 Time    Coprostanol    Coliform
(Hours)   yg/liter    Number/100 ml
         Fecal
        Coliform
      Number/100 ml
    Standard Plate
        Counts
     Number/100 ml
0
3
6
12
24
48
96
192
17.0
17.8
15.0
15.7
12.2
13.9
17.0
7.7
1.38 x 101*
2.12 x 105
-
_
2.12 x 10H
3.8 x 105
2.3 x 107
3.4 x 106
10
1300
-
—
„
36
-
2
<—
8 x 105
—
-
1.5 x 106
TNC(>3xl06)
8.9 x 108
3.5 x 107
Creek Water - Total Coliform
                1.55 x
Fecal Coliform
      10
Standard Plate Count
    49 x 103
No coprostanol was detected in the hexane washings of the  glass  cylin-r
der after termination of the experiment.
                                30

-------
        90
                                        Sample  1
                                        Sample  2
           0    20   40   60   80   100  120  140  160  180   200  220  240
                                          Hours

FIGURE  9.  RATE OF CHANGE OF COPROSTANOL CONCENTRATION IN SAMPLES 1 AND 2 (TRICKLING FILTER
                                  BEFORE CHLORINATION)

-------
  50 -
£40
o


3  30
CO
o
M
cu

3  20 H
  10 -
                                                             Sample 3 University Sewage Plant

                                                             Effluent After Chlorination



                                                             Sample 4 Chlorinated Sewage

                                                             Effluent and Surface Water

                                                             Mixture
                                                                                            OJ
                           I
                                               I
                   40    60      80    100    120     140    160     180    200

                                    Hours
       0      20




  FIGURE  10.   RATE OF CHANGE OF COPROSTANOL CONCENTRATION IN SAMPLES 3 AND 4

-------
                                             2nd Experiment
                                             (Before Chlorination
                                     	 3rd Experiment
                                             ^.fter Chlorination)
                                             4th Experiment
                                             (Environmental)
              /\
     10*-
     10 3-=
                   V
         i
     102-
              r    T    i    i     i    i    i    i    t    i
         0   20   40  60  80  100 120 140 160 180 200
                            Hours
FIGURE 11.  TOTAL COLIFORM - RATE OF CHANGE IN CONCENTRATION
                             33

-------
  105 -
   10" -
  1000 —
(D
O
o
   100-
    10

   1

!\
I     \
         I \
         I  \        /
                          2nd Experiment (Before Chlorina-
                          tion)
                      _  3rd Experiment (After Chlorina-
                          tion)
                          4th Experiment (Environmental)
 ,   \  /
             I   I   1   I    I    I    I    I    I    I
        0   20  40  60  80  100 120 140  160 180  200
                           Hours
 FIGURE 12.  FECAL COLIFOBM - RATE OF CHANGE IN CONCENTRATION

-------
      10b .
      107-
       106 -
   s
   o
                                          2nd Experiment
                                           (Before Chlorination)
                                      	 3rd Experiment
                                           (After Chlorination)
                                          4th Experiment
                                          (Environmental)
       105
              —i	r—i—i	r—T	1—i	1—r
             0    20    40    60    80   100  120   140   160   180  200
                               Hours
FIGURE 13.   48-HOUR TOTAL PLATE COUNT KATE OF CHANGE IN CONCENTRATION
                                35

-------
                             Results

A very complicated chromatogram was obtained with evidence of copro-
stanol, as well as cholesterol (Figure 14),  Large amounts of cho-
lesterol were found to be present, with lesser amounts of coprostanol.
No exact quantitative value for the amount of coprostanol could be de-
termined due to the overlap of another, unknown peak but approximately
100 yg/1 of coprostanol was estimated to be present.  This value was
considerably less than the concentration found in incoming sewage (ap-
proximately 500 yg/1) but more than that found in effluent from the ac-
tivated sludge process (approximately 10 yg/1).  Thus it appears that
a combination of physical removal and biological degradation may account
for the removal of coprostanol, since physical removal would require
that coprostanol concentration be higher in the concentrated biological
floe than in the sewage influent, while complete biological floe be
no higher than that in the effluent.

                     Preservation of Samples

Summary of Results

Preservation of stored samples is absolutely necessary.  The studies
show that only 33.4 percent of coprostanol  was recovered after 8
days' storage without preservation.  The results also indicate that
preservation may be accomplished either by the addition of 1 ml of
I^SO, per liter of water or by storing the samples at 4°C.

                            Procedure

A four-liter sample of trickling filter effluent was taken from the
Gainesville Sewage Treatment Plant.  The sample was divided into
eight 500 ml samples.  Two samples were preserved with 0.5 ml I^SO.
each and analyzed immediately upon return to the laboratory.  Two sam-
ples were preserved by refrigerating at 4°C and analyzed after 8 days.
Two samples were preserved with 0.5 ml I^SO, each and preserved at room
temperature.  They were analyzed after 8 days in storage.  Two samples
were stored at room temperature for 8 days and then analyzed.  Analysis
was done in all cases by extracting with two 50 ml portions of hexane
and washing with two 25 ml portions of 70 percent ethanol and two 25 ml
portions of acetonitrile saturated with hexane.  The hexane extracts
were evaporated to dryness and redissolved in 0.5 ml hexane.  Five yl
were injected for gas chromatography.  Results of this experiment are
given in Table 9.
                                36

-------
(D
o
rt
       1234      5     6

                            Minutes
8
10
FIGURE 14.  CHROMATOGRAM OF EXTRACT OF RETURN ACTIVATED  SLUDGE
                              37

-------
                             TABLE 9

                   Preservation Study Results
Sample
                                                   Coprostanol
#1
#2
#3
#4
#5

#6

#7
#8
Preserved with
Preserved with
Stored at 4°C
Stored at 4°C
Preserved with
temperature
Preserved with
temperature
Not preserved
Not preserved
H2S04 -
H2SO -
*• H

H2S04 -

H2S04 -

- Stored
- Stored
No storage
No storage


Stored

Stored



at room

at room

at room temperature
at room temperature
79
90
73
84

82

111
30
26
.4
.6
.0
.0

.0


.9
.0
Assuming the average value of samples //I and #2 represents the true
value of coprostanol concentration in the original sample, we can
calculate recoveries for the various experiments.

Preserved by refrigeration

        % Recovery - ?8.5 x 10Q = g2.6%
                     85.0
Preserved by addition of

        % Recovery - liil x 100 = 113%
                     85.0
No preservation

        % Recovery = ||iA x IQQ = 33.4%
                     o_>. U

                          Field Surveys

The field studies reported here were undertaken in order to determine
the applicability of coprostanol as a useful indicator of water quality,
and to determine any possible relationship between coprostanol and bio-
logical indicators in water samples.

Surveys were made of two areas in the city of Gainesville, Florida.  The
first area surveyed was Sweetwater Creek—Payne's Prairie, and the sec-
ond was the University of Florida Sewage Treatment Plant-^Lake Alice,
                                38

-------
Samples were taken at locations shown on the maps (Figures 15 and 16),
Numbers on the maps correspond to station numbers In Tables 10 and 11,

Summary of Results

The concentration of coprostanol decreased with distance from the
sewage plant outfall but did not correlate with total or fecal coliform
analyses because of the effect of chlorination on coliforms and
because of a large bird population in Payne's Prairie.  Samples taken
from Sweetwater Creek above the sewage plant outfall had very low
coprostanol concentrations.

Every sample for chemical analyses was preserved with 1 ml cone. H^SO,
per liter.  Extraction with hexane was done as soon as possible after
sampling.  Extraction and analysis were done in the manner detailed
previously.  None of the extracts were saponified to free the esters.
It was found necessary to clean up by thin-layer chromatography all
samples with a coprostanol contant less than 1 ug/1.  Thin-layer chro-
matography was performed using Eastman Chromagram Sheets.  Development
of the plates was with benzene.

Standard coprostanol and cholesterol were identified by spraying with a
saturated solution of antimony trichloride and chloroform, as well as
with ultraviolet light.  Unknown locations corresponding to those of
standards were cut out using scissors and redissolved in a small amount
of hexane.  Samples for microbiological tests were also taken at the
same time as the chemical samples, and were analyzed the same day for
total coliform, fecal coliform and standard plate count.

The concentration of coprostanol in Sweetwater Creek and Payne's
Prairie was found to decrease with increase in distance from the
source of pollution, the City of Gainesville Sewage Treatment Plant.

High concentrations of coprostanol were found in all samples taken in
Sweetwater Creek downstream from the treatment plant.  Very low concen-
trations of coprostanol were found in samples taken upstream from the
treatment plant as well as for samples taken well inside Payne's Prairie.

No apparent relationship between coprostanol and any of the microbio-
logical indicators could be found.  Chlorination had the effect of low-
ering appreciably the concentration of microorganisms, without apparent-
ly affecting the coprostanol concentration.  The result was that low
counts of microorganisms were found near the plant, where the water
was hardly of good quality, while high counts of microorganisms were
found at distant points where one would have expected water quality
to have been improved.  High counts of coliform organisms found in
Payne's Prairie may have been due to the very high aquatic bird
population (estimated 15,000-20,000 ducks, 1,000 Sand Hill Cranes
and 3,000 egrets and herons) [48] found there during the winter months.
One duck is said to contribute 11,000 Fecal Coliform per 24 hours
                                39

-------
                                        Scale, feet

                                       i     i     i   	1

                                 0   1000  2000   3000  4000
                                                 N
 ~~ Pane *s Prairie

              V
                                             »V- ^  Alaohua
                                             « i  __«^   _. ,
                                                .. -   Sink
                        •^W*   1
       (-
                            —         \i
IGURE 15.  SAMPLING STATIONS, SWEETWATER CREEK-PAYNE'S PRAIRIE
                        40

-------
                                                                          JL
                                                                         ewage
                                                                       Disposai
FIGURE 16.   SAMPLING STATIONS,  UNIVERSITY  OF  FLORIDA  SEWAGE TREATMENT PLANT
                                LAKE ALICE

-------
                      TABLE 10

v Survey for Sweetwater Creek—Payne's Prairie
                   (March 9, 1971)
Station
1

2
3

4

5

6

7

8

9
Volume of
Location Sample, liters
100 feet upstream from
sewage outfall
Sewage effluent
300 feet downstream
from sewage outfall
Highway 331 and Sweet-
water Creek
Termination • of Sweet-
water Creek
120 feet inside Payne's
Prairie
Intersection of North-South
and East-Vest Canals
250 yards south of
Station 7
Alachua Sinkhole
1.0

0.5

0.5

1,0

1.0

1.0

1,0

1.0
1.0
Coprostanol
Ug/1
0.027

65.0.

50.9

39.2

24.3

24.9

0.033

0.059
0.061
Total
Coliform
Number/100 ml
6.4 x 104

26

12

72

1.12 x 103

3.8 x 102

1.47 x 101*

1.78 x 101*
3.38 x 101*
Fecal 48 Er. Standard
Coliform Plate Count
Number/100 ml Number/ICO nl
80

0

1

0

0

0

61

57
15
3.8 x 106

7.4 x 103

6.2 x 103

2.8 x 101*

1.12 x 106

1.03 x 107

6.9 x 105

4.95 x 106
3.35 x 106
                                                                                             CM

-------
                                TABLE 11

Survey Data for University of Florida Sewage Treatment Plant-Lake Alice
                            (March 23, 1971)
1 1 "
Station
1
2
3

4
5
6
7

Location
Effluent after
chlorination
After oxidation pond
100 yards downstream
from $2
Near underpass of canals
in North-South Road
East end of Lake Alice
Middle of Lake Alice
West end of Lake Alice
near out flow

Volume of
Sample, liters
0.5
0.5
i n
J-i U
1.0
1.5
1.5
1.5

Coprostanol
ug/1
54
15
9.9

2,4
0,22
0.014
0,022
'Total
Coliform
Number/100 ml
3.22 x 101*
5.42 x 10s
5,55 x 10s

1.85 x 10s
7,3 x 101*
1.37 x 10*
1,17 x 102
Fecal 48 Hr. Standard
Coliform
Number/100 ml
9
960
1.37 x 103

250
96
66
Plate Count
Number/ 100 ml
1,87 x 107
1.06 x 107
4.0 x 107

2.03 x 107
1.68 x 108
1.46 x 108
1.34 x 107

-------
 while a man contributes 2,000 Fecal Coliform [49].   Thus, a 15,000-
 20,000 duck population alone would correspond to the Fecal Coliform
 production of a human population of 82,500 to 110,000,

 As in the first survey, the concentration of coprostanol in the Uni-
 versity of Florida Sewage Treatment Plant—Lake Alice survey was found
 to decrease with increasing distance from the source of pollution.
 There was a significant decrease in coprostanol as  the  sewage effluent
 passed through the oxidation pond (54 yg/1 to 15 yg/1),  Fairly high
 concentrations of coprostanol were found in all samples taken from the
 stream leading into Lake Alice.   However, very low  concentrations of
 coprostanol were found in Lake Alice itself.   Only  the  east end of Lake
 Alice could be said to contain significant amounts  of coprostanol.   No
 difficulty was encountered in analyzing for free coprostanol in the
 environment.

 Once again chlorination had the  effect of decreasing significantly the
 number of microorganisms to be found in the water in samples taken near
 the sewage outfall.   There was an apparent regrowth followed by a de-
 crease in numbers of coliform organisms.  This effect of regrowth is
 apparently a common phenomenon.   As a result of the changing character
 of the coliform organisms, no relationship could be found between
 these microbiological indicators and coprostanol levels.

 It would appear from the above study that no  simple relationship bewteen
 microbiological indicator organisms  and coprostanol exists,  especially
 if the sewage effluent is highly chlorinated.   Coprostanol would appear
 to be a better indicator of fecal pollution than coliform in those waters
 which receive highly  chlorinated effluent.  Coprostanol concentrations
 in such cases would be high, where coliform organisms may be quite  low
 and a danger  to  either resistant pathogenic bacteria or viruses  may  be
 present.

 High aquatic  bird populations  may cause high  coliform counts in  water
 [49],  while no  real danger  exists.   Although  some diseases may be
 transmitted from birds  to man  [50],  none of these diseases are normally
 transmitted by the water  route.   None of the major  water  borne diseases
 listed in Control of Communicable Diseases  in Man [51]  are given as
 being caused  by birds  as  reservoirs  of  the  disease,

 CORRELATION OF COPROSTANOL  CONCENTRATION WITH OTHER MEASURES  OF  DOMESTIC
              WASTEWATER TREATMENT PLANT EFFICIENCY

 There  is a growing need  for evaluation  of wastewater treatment plant
 efficiency in order to  improve design and operation of  such plants as
well  as  to improve the  quality of receiving waters.

Presently, the biochemical oxygen demand, BOD, test  has the widest
application in measuring waste loadings  to  treatment plants and  in
                                44

-------
evaluating the treatment efficiency.  The test is extremely time
consuming.  Comparison of BOD values cannot be made unless the results
have been obtained under identical test conditions.  Another popular
parameter used is COD (chemical oxygen demand) .  The method needs
close control and fails to include some organic compounds which are
biologically available to the organisms.  Unpolluted waters may have
anomolously high BOD and COD values.  Both tests suffer from the fact
that they are neither specific nor precise,

A specific indicator which could be used is coprostanol.  As reported
earlier, coprostanol loading has been estimated to be approximately
2 grams per capita per day [5],  Coprostanol was found to occur at
detectable levels in sewage treatment effluents and is removable by
adequate secondary sewage treatment [5].  Investigations on the
Missouri River showed that coprostanol level varied with the degree
of domestic waste pollution [52].

It was the purpose of this phase of the study to determine the possi-
bility of using coprostanol as an indicator of domestic waste treat-
ment plant efficiency.  Efforts were made to correlate coprostanol
with BOD and COD as well as TOG (total organic carbon).  Evaluation
and control of domestic wastewater treatment efficiency could be
greatly facilitated if such relationships could be found.

Summary of Results

There was a good correlation between TOG and coprostanol concentration
with a coefficient of determination (R2) of 0.977 for a linear fit.
COD and coprostanol could be related by a parabolic curve but R2 is
only 0.630.  BOD and coprostanol could be related linearly but R2 is
even lower, only 0.429.

Many more data points are needed to confirm these relationships.

                            Procedure

The University of Florida Campus Sewage Treatment Pland and Gaines-
ville Municipal Sewage Treatment Plant (both in Gainesville, Florida)
were selected for sampling sites due to their domestic wastewater
source, close control and availability.  Samples were collected at
different times from different locations in order to obtain a wide
range of variations of the wastewater characteristics.  In order to
obtain representative wastewater samples and to have maximum accuracy
on analysis, three-hour composite sampling was adopted,  BOD, COD and
TOC analyses were performed immediately after sampling to prevent
preservation error.  Coprostanol samples were put in one-liter glass
bottles and preserved by acidification (3 ml concentrated sulfuric
acid per liter sample) and refrigeration.  Analysis was done within
5 hours after sampling.
                                45

-------
     The analysis was done according to Standard Methods for the
Examination of Water and Wastewater [1].  Three dilutions were made
of each sample.  Contact stabilization plant raw influent was used
for seeding material.  Modified Winkler Method (Aside Modification)
with Full-Bottle Technique developed by the Environmental Protection
Agency was used for the dissolved oxygen measurements [53].

COD—The dichromate reflux method described in Standard Methods [1]
was used for high level COD analysis.  Low level COD analysis was
carried out following procedures outlined in the EPA Manual [53].

TOC—Samples for TOG analysis were put in 150 ml glass bottles to
prevent the interference of extractable carbonaceous substances.
Precautions were taken to avoid decomposition from exposure to light.
Analyses were performed immediately after sampling on the Dow-Beckman
Carbonaceous Analyzer Model 915.

Coprostanol—Samples were analyzed by GLC as discussed earlier.

                             Results

Eight samples were taken from primary treatment step of the University
of Florida Campus Sewage Treatment Plant (the results are shown in
Table 12), 22 samples from the trickling filter plant (Table 13), 18
samples from the contact stabilization plant (Table 14), and 6 samples
from the tertiary treatment (aerated lagoon, Table 15),  Ten samples
were taken from the contact stabilization plant of the Gainesville
Municipal Sewage Plant (Table 16) and 7 samples from trickling filter
plant Table 17).  For the special experiment on TOC analysis, 10
samples were taken from the municipal contact stabilization plant
(Table 18).

In order to confirm the possible relationships among coprostanol and
BOD, COD and TOC, the authors tried to fit the data to equations.  The
data were plotted.  By comparison with figures of various functions,
the linear form of polynomial functions was chosen as the model.
A general form of the model can be represented by:  Y = 3Q + 3,X, +
&2X2 + ^3X3 + •" + fyc^k + e» wnere v ^s fche dependent variable re-
lating to a set of independent variables, X represents unknown
parameters, e is the random error.  No interaction terms were included
and the second order terms were the highest order used.

A Linear Least Square Curve Fitting Program was used to determine
whether the equation chosen could be made to fit the data by a suit-
able choice of coefficients.  By computer analysis with APL (A Pro-
gramming Language) on an IBM 360 computer 2741 terminal, the follow-
ing major information was obtained on each set of data [54]:
1,  Coefficient of determination (R2), i,e., how well the equation
fitted.  If R2 = 0.700 for example, it means that 70 percent of the
                                46

-------
                      TABLE 12

      DATA FROM CAMPUS PLANT-PRIMARY TREATMENT
Sample
No.
1
2
3
4
5
6
7
8
Sampling
Station
P,I.
P.E.
P.I.
P.E.
P.I.
P.E.
P.I.
P.E.
BOD
(mg/1)
267.50
47,50
62.50
100.00
137.50
120.00
245.00
182.50
COD
(mg/1)
364.14
77.60
85.36
234.74
122.00
135.00
263.00
213.00
Coprostanol
(yg/1)
200.00
10.00
13.75
50.00
36.25
41.66
82.50
92.50
P.I.;   Primary Influent
P.E.:   Primary Effluent
                          47

-------
                       TABLE 13

       DATA FROM CAMPUS TRICKLING FILTER PLANT

 Sample    Sampling      BOD       COD     Coprostanol
  No.      Station      (mg/1)    (mg/1)     (pg/1)


   1         TF-R      267,50    364.14      200.00
   2          P-l      106.00    180.14       76.00
   3         TF-1       35,03    128.36       35.00
   4          S-l       29.50    103.46       29.00
   5         TF-1        8.13     52.38       65.00
   6          S-l       20.50     73.72       56.25
   7         TF-R       47.50     77.60       10.00
   8          S-l       16.25     42.68        3.33
   9         TF-1        3.60      8.00        3.33
  10          P-l       26.25     27.16       43.75
  11          S-4         *       83.42        6.66
  12          S-3         *       60.14       12.50
  13          S-l         *       79.54       71.25
  14         TF-1         *       87.30        6.25
  15         TF-2         *       56.26       56.25
  16          P-4         *      230.00      110.00
  17          P-l         *      217.28       40.00
  18         TF-R         *      234.74       50.00
  19         TF-R      137.50    122.00       36.25
  20         TF-R      120.00    135.00       41.66
  21         TF-R      245.00    263.00       82.50
  22         TF-R      182.50    213.00       92.50
TF^-R:  Raw Influent
 P-l:  Undiluted Primary Clarifier Effluent
 P-4:  Diluted Primary Clarifier Effluent
TF-1:  High Rate Trickling Filter Effluent
TF-2:  Standard Rate Trickling Filter Effluent
 S-l:  No. 1 Secondary Clarifier Effluent
 S-3:  No. 2 Secondary Clarifier Effluent
 S-4:  Combined Effluent of Secondary Clarifier
*Data unavailable due to missing BOD bottles
                           48

-------
                        TABLE  14

      DATA FROM  CAMPUS  CONTACT STABILIZATION PLANT
Sample
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Sampling
Station
CSP2
CSP1
CSPR
CSPR
CSPR
CSPR
CSPR
CSPR
CSP2
CSP2
CSP2
CSP2
CSP2
CSP2
CCE
CSPR
CSP2
CSP1
BOD
(mg/1)
28.25
53.00
237.50
46.66
105.00
24.16
43.33
19.16
14.00
7.00
10.00
3.00
4.00
5,50
*
*
*
*
COD
(mg/1)
295.00
201.00
380.45
407.59
337,93
409.21
72.35
72.25
41.47
400.00
180.00
104.75
36.70
33.37
0.10
153.26
64.02
40.74
Coprostanol
(yg/l)
141.00
33.75
162.00
140.00
142.50
159.00
5.00
4.50
3.00
10.50
25.00
3.50
1.66
0.10
1.00
67.50
12.50
1.75
CSPR:   Contact Stabilization Plant Raw Influent
CSP1:   Effluent from Contact Plant No. 1
CSP2:   Effluent from Contact Plant No, 2
CCE:   Combined Effluent of Contact Plant No. 1 and No.
*Data unavailable due to missing BOD bottles
                            49

-------
                    TABLE 15




       DATA FROM CAMPUS TERTIARY TREATMENT
Sample
No.
1
2
3
4
5
6
L.I.:
Li • EA • *
Sampling
Station
L.I.
L.E.
L.I.
L.E.
L.I.
L.E.
Lagoon Influent
Lagoon Effluent
BOD
(mg/D
21.00
19.00
16.00
21.00
6.00
6.50


COD
(mg/1)
19.00
7.60
34.30
19.00
27.16
42.88


Coprostanol
(yg/D
3.93
5.00
1.88
4.17
18.75
0.17


                    TABLE 16




DATA FROM MUNICIPAL CONTACT STABILIZATION PLANT
Sample Sampling BOD
No. Station (mg/1)
1 A 263.25
2 C 56.00
3 D 8.00
4 E 6.67
5 B 220.00
6 A 167.50
7 C 38.00
8 E 8.15
9 D 4.46
10 B 145,00
A: Contact Plant Raw Influent
E: Primary Treatment Effluent
B: Contact Basin Effluent
C: Clarifier Effluent
D: Final Effluent of Municipal
COD
(mg/1)
300.00
7.60
7.60
5.32
231.00
427,00
49.40
22.80
38,80
53.20




Plant
Coprostanol
(yg/D
115,00
0.01
0.01
13.75
33.33
100.00
15.00
0,62
3.33
80.00





                        50

-------
                            TABLE 17
          DATA FROM MUNICIPAL TRICKLING FILTER PLANT
Sample Sampling BOD COD
No. Station (mg/1) (mg/1)
1 G 110.00 133,00
2 H 30.00 40.00
3 K 16.00 10.25
4 G 96.50 202.00
5 H 29.00 86.50
6 K 112.50 108.00
7 F 145.00 407.00
K: Trickling Filter Raw Influent
F: Primary Clarifier Influent
G: Trickling Filter Effluent
H: Secondary Clarifier Effluent
Coprostanol
(yg/D
53.33
14.00
54,16
133.75
31.66
177.50
116,66




                            TABLE 18

                   DATA FROM THE SPECIAL STUDY
          ON THE MUNICIPAL CONTACT STABILIZATION PLANT
Sample
No.
1
2*
3
4
5*
6
7
8
9
10
Sampling
Station
A
B
C
D
E
F
G
H
I
J
BOD
(mg/1)
121.00
— —
10.00
2.87
	
120.00
12.50
5.00
2.00
2.40
COD
(mg/1)
255.36
	
17.28
5.76
	
211.20
53.76
92.16
34.56
80.64
TOG
(mg/1)
126.00
	
12.00
12.00
	
8.70
8.00
6.00
2.50
2.50
Coprostanol
(yg/D
48.00
—
11.07
11.43
—
10.00
3.76
5.83
6.43
4.06
*Results on Sample No.  2 and No,  5 were discarded due to
 extremely high TOC values compared to low values of BOD,
 COD and Coprostanol
                                51

-------
sum squared of deviations of the Yi about their mean can be account-
ed for by the variables X^, X_ ••• X^.
2.  The magnitudes of the intercept QBO) and of the other coefficients
(e±rs).

Precision—For each method, one sample was portioned into ten equal
parts and analyzed simultaneously.  The results are shown in Table 19.

                            TABLE 19

                     Precision Test Results
Test
Coprostanol
COD
BOD
TOC
Mean
Value
57.50 ug/1
137,50 mg/1
82.50 mg/1
*
Standard
Deviation
0.72 vg/1
3.00 mg/1
4.28 mg/1
*
Relative Standard
Deviation with Respect
to Mean
1.25%
2.18%
5.18%
5-10%**
Unavailable
**Data from Standard Methods, 13th edition
                 Statistical Analysis of Results

An example of the technique used for analysis of the results of a
given set of data is shown below for Experiment 1 - Campus Plant—
Primary Treatment.  Model I in each case was a parabolic function and
Model II linear.  The curves of best fit are shown in Figures 17-21
and a summary of the statistical results are shown in Table 20.

         Experiment 1 - Campus Plant—Primary Treatment

a.  Coprostanol vs. COD
                                          Y:  Coprostanol
                                          X:  COD
Model I:  Y
30 +
                       + $
                                  + e
Prediction equation:  y - 23.5 - 0.197X + 0.00182X2
n - 8        R2 - 0.878
Test on Curvature by two-sided student's t test:
HQ:  B2 - 0
         A
TS:  t - 6? - 0 . 2.02
         MSEC22
RR:  |t| > £1 , d.f.

From standard statistical tables,
to,025,5 " 2-57           to,05,5 = 2<01
                                                                CD
                               52

-------
  200 -
60
3.
§100
•u
CO
o
M
CX
O
fj
Model I.


Model II
                                            COD (mg/1)


                  FIGURE 17.  COPROSTANOL vs COD CAMPUS PLANT—PRIMARY TREATMENT
                                                                                                            •on

-------
200 -
 100 -
                                                                                300
                                         BOD (mg/1)



                FIGURE 18.  COPROSTANOL vs BOD CAMPUS PLANT—PRIMARY TREATMENT

-------
  200-
60

\_x

H
O

§
4->
P.
O
0
      0
100
200
300
                                         COD  (mg/1)


                                 FIGURE 19.   COPROSTANOL vs COD OVERALL

-------
         200-
-
O


I
W

8
                     o   o
100 -
                                       t

                                       100
                                                        200
300
                                   FIGURE 20.   COPROSTANOL vs BOD OVERALL

-------
  40 -
  30 -
rH

60
CO
O

Cu
O
O
  20
  10 -
   0
                                                                                                         r--
                                                                                                         u-i
                                                         100
                                            TOC  (rag/1)
      FIGURE  21.   COPROSTANOL vs TOC SPECIAL  STUDY  ON MUNICIPAL CONTACT STABILIZATION  PLANT

-------
           TABLE 20



SUMMARY OF STATISTICAL ANALYSES




Experiment
No. Data Source
1

2
Campus Primary
Treatment

Campus

Contact
Stabilization

3

4
Plant
Campus
Filter
Campus

Trickling
Plant
Secondary
Treatment

5

Campus

Tertiary
(1)
(2)

y -

R2 =
y •
R2 -
y "
R2 „



Y: Coprostanol
Xx: COD
Coprostanol vs. COD
y - 23.5 - 0.197X + 0.00182X2
R2 - 0.878, or
y - -40.8 + 0.571X
R2 - 0
-32.5

'0.878
33.0 -
| 0,706
-0.186
0.719


.851 (FiK. )
- 0.702X - 0.000666X

(Fie, )
0.183X + 0.00167X2
(FiR. )
+ 0.625X - 0.000594X2
(Fig. )

(Fig. >
X2: BOD
X3: TOC


Coprostanol vs. BOD
(1) y " 1.63 -
R2 - 0.805,
(2) y - -0.313
R2 - 0.
y - 30.

R2 - 0.
y - 16.
R2 » 0.
y - 26.
R2 • 0.


76">
1 +

W
8 +
S9S
9 +
374

(Fig



Cocrostanol vs. COD. BOD
0.110X + 0.00245X2 Y - 00 + 0^
or
+ 0.668X
(FiR. )
0.682X

(Fig. )
0.432X
(Fig. )
O.A61X
(Fig. )

. )

Y -


y •
R2
y m
9
R2
Y -
Treatment
6

Campus
Overall
Plant

y --
R2-
0.1195
0.731
+ 0.623X - 0.000587X2
(Fig. )
y - 16,
R2 - 0,
7 +
709
0.472X
(Fie, )
Y -


8 + 0,X,
o 1 1

-4.29 + 0.
• 0.859
12.46 + 0.
0.403X2
• 0.470
30 + B^
+ 33X2 + c
00 + 8j_Xj_

^

+ BoXi2 + e
H2 1
-
58X! - 1.77X2

09S9X, +
1

+ e2x 2

•f c

                                                                                         oo
                                                                                         10

-------
                                                   TABLE 20  (Cont.)
Experiment
No. Data Source
7
8
9
10
11
12
Municipal Plant
Overall
Municipal
Trickling Filter
Plant
Municipal Plant
Overall
Contact Sta-
bilization Plant
Overall
Trickling Filter
Plant Overall
Secondary Treat-
ment Overall
y •
R2

y "
R2
y •
R2
y •
R2
y •
R2
Coprostanol vs. COD
9.31 + 0.228X
- 0.640 (Fig, )
(Fig. ) '
32.8 + 0.151X
•» 0.311
-0.144 + 0.572X - 0.000550X2
- 0.776
7.89 + 0.437X - 0.000415X2
= 0.405
-5.26 + 0.55X - 0.000528X
- 0.600
Coprostanol vs. BOD
y «• 0.950 + 0
R2 - 0.709
y - 18.0 + 0.
(1) y - -3.19 + 1
R2 - 0.527
(2) y - 19.3 + 0.
R2 - 0.377
y - 24.1 - 0.
R2 - 0.291
y - 24.6 - 0.
R2 - 0.459
y = 24.4 + 0.
R2 - 0.366
.383X
84 IX
.24X - 0.00354X2
421X
404X
47 OX
438X
Coprostanol vs. COD, BOD
i-«. + V, + «'
y « 11.6 - 0.694X7
+ 1.10X,
R2 - 0.552
**
y = 8.23 + O.lOlXj^
+ 0.333X2
R2 - 0.425
Y— Q J_ ft Y -L f
P— • PrtA« T fc
y - 15 + 0.087 + 0.357X2
R2 - 0.446
**No relationships were found by the model given.

-------
TABLE 20  (Cont.)
Experiment
No.
13
14
Data Source
Campus and
Municipal
Plants Combined
Special Study On
Municipal Contact
Stabilization
Plant
Coprostanol vs. COD
y - -8.08 + 0.563X - 0.00535X2
R2 - 0.630 (Pig. )
Coprostanol vs, TOG
y - 4.94 + 0.343X
R2 - 0.977 (Fig. )
Coprostanol vs. BOD Coprostanol vs. COD, .BOD
y - 17.3 + 0.457X y - 8.85 + O.lOOXj^
R2 - 0.429 (Fig-. ) + 0.362X2
Coprostanol vs. TOC, COD, BOD
Y - B0 + B3X3 + e
                                                                                     o
                                                                                     v£>

-------
    There is  insufficient evidence to reject H  at = - 0.05.

    Model II:  Y  -  Bo  + BjX + e

    Prediction equation:   y « -40.8 + 0.571X                  (2)
    n = 8      R2  = 0,851
    Test on dependency of y by two-sided student's t test:

    V  *i-J>
    TS:  t »  §! - 0
             MSEC^=  5'88


    '0.025,6  = 2<44     Ho

Conclusion

    Whether there is curvature or not will need more data points
    to verify.

    Possibly, Coprostanol (Y) could be related to X(COD) by equa-
    tion (2)  as shown  in Figure  17.  Equation (1) was also plotted
    for reference.

b.  Coprostanol vs. BOD                    Y:  Coprostanol
                                           X:  BOD

    Model I;   Y - BQ + BjX2 + e
    Prediction equation:   y - 1.63 - 0.110X + 0.00245X2      (3)
    n - 8        R2 -  0.805
    HQ:  $2 * °     Calculated t - 2.55
    Critical  t-values:  tn M, - = 2.57, tn nR - = 2.01
                         U t U£3)3          W•U3 j J
    There is  insufficient evidenct to reject HQ at X - 0.05.

    Model II; Y  -  Bo  + B-jX + e
    Prediction equation:  y - -0.313 + 0.668X                (4)
    n - 8      R2  » 0.765
    HQ:  &i - 0        Calculated t • 4.45

    Critical  t0(025,6  "2.44        HO rejected

Conclusion

    Coprostanol  could  be related to BOD as shown in prediction
    C4) (Fig. 18).  More data are needed to determine whether
    there is  curvature or not.
                               61

-------
c.  Coprostanol vs. COD & BOD             Y;   Coprostanol
                                         X,:   COD
                                         X2:   BOD

    Complete model;  Y = 3Q +  3-^ +  fyX^2 +  33X2 + 34X£2 + e
    Prediction equation:  y - 7.70 - I.QX^ + 0.00350X-L2       (5)

                              + 1,31X2 -  3.54X22
    n = 8     R2 = 0.970
    Reduced model;  Y « 3  + 3iX., + 30X0  + e
    -        O    J- J-    2 £.
    Prediction equation:  y = -43.7 + 0.410X-L + 0.225X2       (6)
    n = 8      R2 = 0.871
    Analysis ;  Test on curvature  (complete model) :  H :   62  = PA  =  0

    TS:  F - ss Dr°P _ = 3.36
             SSE (complete model)
    RR:       j>

    Test on dependency  (reduced model):   H :   3-^-0      t  =  6700

    HQ rejected       Critical t0>025 5  = 2.57
    HQ:  32 ~ 0    t =  0.900    HQ accepted.

Conclusion

    Coprostanol could be related to COD  by Y = 3  + 3-,X1  +  e.

                      Discussion of Results

1,  Coprostanol vs. COD:  The best fit was found  in Experiment  1
(Campus Plant — Primary  Treatment, R2 - 0.851)  and Experiment  2
(Campus Contact Stabilization Plant, R2  - 0.878).  Experiment 9
(Municipal plant, Overall, R2 = 0.311) and Experiment  11  (Trickling
Filter Plant, Overall,  R2 = 0.405) show  lack of fit.   No  relation-
ships were found in Experiment 5 (Campus Tertiary Treatment)  and
Experiment 8  (Municipal Trickling Filter Plant).  Most of the models
fitted show a parabolic relationship between Coprostanol  and  COD.
2.  Coprostanol vs. BOD:  Good results were found only in Experiment  1
(Campus Plant — Primary  Treatment, R2 = 0.765)  and Experiment  7
(Municipal Plant, Overall, R2 » 0.709).   No relationship  was  found
in Experiment 5 (Campus Tertiary Treatment)..   Besides  Experiment 5
and the uncertainty of  Experiments 1 and 9, a  straight line relation-
ship was found to exist between coprostanol and BOD.
3.  Coprostanol vs. COD and BOD:  The best fit of the  combined  models
was found in Experiment 3 (Campus Trickling Filter Plant, R2  -  0.595).
No relationship was shown in Experiment  9 (Municipal Plant , Overall) .
4.  Coprostanol vs. TOC:  The coefficient of determination  (R2) in this
analysis was found to be 97 percent which is a good fit.  The model
                               62

-------
showed that coprostanol and TOG might be correlated by straight line.
5,  Coprostanol vs. TOG, COD and BOD;  The results on the analysis of
a combined model revealed that variation in the coprostanol value is
accounted for mostly by TOG,  COD and BOD contribute less information
compared with TOC.

              DEVELOPMENT OF A COLORIMETRIC METHOD

Summary of Results

A simple colorJbaetric method was devised for the analysis of co-
prostanol in sewage and polluted surface waters.  The method
eliminated interferences from cholesterol and other constituents by
oxidation with CrO^-pyridine and by simple chromatography on an
alumina column.  Although the reagent also oxidized coprostanol, the
resultant coprostanone formed quantitatively and was separated by
the alumina column.

The color was developed as follows:  Coprostanone was reacted with
2,4-DNPH* to form the hydrazone.  The hydrazone was hydrolyzed after
excess 2,4-DNPH was removed.  The liberated 2,4-DNPH was then oxidized
with periodate to the diazonium salt and coupled with ct-naphthylamine.
The azo compound had intense absorption, with e = 35,000, at 500 run
in an acidic solution of ethylacetate-acetic acid-concentrated HC1.
The complete analysis required about 2 hours.

The method was applied to several field samples, with over-estimation
of the coprostanol concentration by a factor of 1.6 as compared to
the results obtained by GLC.  Although some natural waters contained
colorimetry-sensitive materials which were not coprostanol, the
amount was insignificant.  The compounds that interfere were shown
to have a high probability of being of fecal or urinal origin.

                          Introduction

The gas-liquid-chromatographic (GLC) method for coprostanol analysis
has major disadvantages for actual surveys of this indicator:  GLC
instrumentation is expensive and not always available to sewage treat-
ment plants or field observers.  The object of this part of the study
was to develop a less expensive and simple, yet sensitive technique
for the determination of coprostanol in sewage and surface waters.

Colorimetric technique's have been reported in the literature.  Hartel
and Dam [55] reported that coprostanol obtained from feces gave a
"faint" reaction with a molar absorbence of 1066-930 at 528 nm the
Tschugaeff reaction (with acetylchloride and zinc chloride in chloro-
form) .  Wells and Morres [56] evaluated the use of digitonide which
forms insoluble complexes with sterols.  They concluded that the tech-
nique was less applicable than GLC, because of the solubility of the
*2,4»-Dinitrophenylhydrazine
                                63

-------
 coprostanyl digitonide and of the difficulties in removal of interfering
 cholesterol.  They further reported that the detection limit was  several
 hundred micrograms.   This high level is found only in raw sewage  or
 solid fecal material.   The concentration of coprostanol in polluted
 water is usually much less than one hundred micrograms per liter.
 Gerson [57] applied the Zlatkis-Zak color development method to the
 analysis of coprostanol in fecal lipids.  He stated that the detection
 limit was about 75 yg and that cholesterol interfered.

 The principal  difficulties in the colorimetric techniques for copros-
 tanol analysis in sewage and  surface waters are:
 1.   Removal of interfering materials, principally cholesterol, and
 2.   Discovery  of a sensitive  color reaction which enables the de-
 termination of less  than one  microgram of coprostanol.
 To  overcome the first  problem, the oxidative decomposition of choles-
 terol followed by simple chromatographic removal  of its oxidation
 products,  together with the elimination of other  interfering con-
 stituents were considered. Several oxidation methods were evaluated.
 To  overcome the second problem a wide variety of  color reactions were
 studied.

 The following  sections describe  the optimization  of the above steps,
 resulting in an analytical procedure that can quantitatively determine
 less than one  microgram of coprostanol in sewage  and  surface waters.

                     Recovery of  Coprostanol

 Summary of Results

 It was found necessary to  saponify  the samples because of  high ratios
 of  esterified  to unesterified  coprostanol in  many samples.   A simple
 saponification  procedure was  developed.

                       Procedure and  Results

 1.   Evaluation  of error  in the extraction and washing  steps.  Because
 of its hydrocarbon nature,  coprostanol  is  soluble  in many  organic sol-
 vents.  Hexane  has been  used as the  principal extraction solvent by
 most  authors [8, 13, 58].   Hexane, however, can extract not  only
 coprostanol but also many  other organic constituents in waste waters
 and natural waters in which they are present in solution and as sus-
 pended solids.   These organic materials interfere with the successful
 extraction of coprostanol  by concentrating at the interface between
 the hexane and water layers.  These materials also interfere with the
 spectrophotometric determination by generating color and turbidity.
 To eliminate these interfering materials, washing the hexane layer
with  70 percent ethylalcohol  [8] and with acetonitrile  [59] has been
 used.  The loss of coprostanol through the extraction-washing was
 evaluated.  An experiment was conducted for the estimation of this
 error; the organic constituents in raw and treated sewage were
                                64

-------
separated into four fractions by the procedure diagramed in Figure 22,
i.e., hexane, 70 percent ethylalcohol , acetonitrile and residual wa-
ter layers.  The coprostanol concentrations were determined with GLC,
after the extracts of each fraction were saponified, oxidized at 50°C,
and the coprostanone fraction was separated, as discussed later.  The
results obtained are summarized in Table 21, and show that 5 to 23
percent of the coprostanol was lost by the above extraction and wash-
ing technique.  Most of the error resulted from the solvent washing
steps.  The reduced recovery was unsatisfactory and could have been
caused by the solubility of free coprostanol in the washing solvents;
however, the error was too high.  A reasonable explanation for this
error i's based on the observation of higher levels of insoluble ma-
terials at the interface between the two solvent layers in the case
•of samples that had not been diluted.  This suggests that coprostanol
esters present in the water, are so soluble in more polar solvents
that they would be included in the solids at the interface.  This
would result in the high errors observed.  In the case where the con-
centration of coprostanol is low, for example, 6 yg/1 in the sample
after the oxidation pond, the low recovery seems to be attributable
both to extraction of free coprostanol with the solvents and to the
high ratio of esterified coprostanol to free coprostanol.  The amount
of free coprostanol, expected to be extracted by the solvents was
calculated as 0.7 yg/1, based on the assumption that the equilibrium
ratio in hexane and the washing solvents was 91.5 and 8.5 percent
respectively, by the following calculation:
               (6.0 + 1.8) yg/1 x  (   -) =  0.72  yg/1
                                   yi . j/o
 Although this estimation is very rough, it supports  the suggested  in-
 terpretation, i.e. , the existence  of esterified forms of coprostanol.
                      SAMPLE WATER,  1 liter

                  3 ml of  Cone'  H2S(>4
                  NaCl 40  g
                  in 2-1 Separatory  Funnel
                  Hexane Extraction,  50 ml, twice
         P                                      •       —j
       HEXANE LAYER                             WATER RESIDUE

            70% Ethanol Washing
                25 ml, twice
                Acetonitrile Washing          70% ETHANOL LAYER
                 (Saturated with Hexane)
                     25  ml, twice
        T
       HEXANE  LAYER                            ACETONITRILE LAYER

              FIGURE 22.   SCHEME FOR THE FRACTIONATION


                                65

-------
                            TABLE 21

               ESTIMATION OF ERROR IN RECOVERY BY
                 SOLVENT EXTRACTION AND WASHING3
                                  Recovery of Coprostanol .
Layers Raw Sewage I"
Raw Sewage IIC
Effluent
After
Oxidation
Pond
yg/i
%
yg/1
%
yg/1
%
yg/l
%
Hexane
70% Ethanol
Acetonitrile
Residual
   Water

Total loss
               371.0   86
39.7
19.5
 1.8
9.2
4.5
0.4
                61.0   14
17,1
                                <1.0d
 0.9
95
                                 15.1   91   6.0   77
                 <1.0d
        1.5
                  <1.0
                                             1.8   23
a.  The sampling stations were at the University of Florida, Campus
    Treatment Plant, Gainesville, Florida.
b.  One liter was extracted without dilution (week day samples).
c.  200 ml of raw sewage was diluted to one liter (weekend samples).
d.  Less than one microgram of coprostanol could not be determined by
    GLC with FID; the coprostanol content was measured by combining
    three samples from 70 percent ethanol, acetonitrile and
    residual water.
                                66

-------
2.  Effect of Saponification,  As was discussed in the previous
section,  coprostanol occurred in sewage and surface waters, princi-
pally in the free state.   The remainder is present in more complex
forms, probably as esters,  Rosenfeld [15] estimated that 30 percent
of the total coprostanol  in fecal sterols is in the form of esters.
One method used for saponification was refluxing for 3 hours with
7.5 percent potassium hydroxide in 70 percent ethanol [5].  This mild
saponification technique  seems to be satisfactory for the recovery of
esterified coprostanol, however, it takes too much time.  A more
rapid saponification method has been found which will saponify the
esterified coprostanol within a few minutes.  The procedure is simple,
i.e., the well-dried hexane extracts in a test tube were heated to
boiling on a flame or in  a water bath after 1 ml of 0.5N sodium hy-
droxide in absolute methanol had been added.  The volume was reduced
to about 0.2 ml.  This method was used to determine the amount of
esterified coprostanol in field samples.  The results are shown in
Table 22 for samples that were not preserved with H2SO,.

The saponifiable coprostanol in raw sewage and treatment plant efflu-
ent, in the case of unpreserved samples (Table 22), were 25 to 30
percent.  These values are the same as reported to Rosenfeld [15],
however the samples which were obtained from the receiving stream or
the samples which were acidified with concentrated H^SO,  (Table 23),
showed higher ratios, 45  to 55 percent.  These higher ratios could
have been due to esterification of the free coprostanol by the action
of the acid on other constituents in the sewage and natural waters
or to the formation of sulfate esters.  The high ratio present in the
stream sample might have  been due to the chlorination process in the
treatment plant, or to some unknown stream conditions such as bac-
terial actions or chemical reactions between free coprostanol and
other components of natural waters.

Because the acidification of sample water is a necessary step for
long-term preservation to prevent bacterial decomposition of co-
prostanol, it is obvious  that the saponification procedure is re-
quired in the analysis,  if exact quantitative results are desired.

            Oxidation of  Cholesterol and Coprostanol

Summary of Results

Cholesterol was separated from coprostanol by oxidation with the
Sarett reagent  (analydrous chromic acid—pyridine complex in glacial
acetic acid) followed by adsorption of the oxidation product of
cholesterol on an alumina column.

                     Procedures and Results

The removal of cholesterol was  one of  the major objectives of this
part of the study.   It is difficult to separate coprostanol and
                                67

-------
                            TABLE 22

                    EFFECT OF SAPONIFICATION
Sample3
Raw Sewage"
Treatment Plant
Effluentb
Receiving Stream0
Saponified
(yg/D
368
78
117
Unsaponified
(yg/1)
266
59
66
Saponifiable
Coprostanol
(%)
27
24
43
aThe samples were analyzed within 30 minutes after sampling,  with-
 out adding 3 ml concentrated t^SO^.

bThe City Sewage Treatment Plant, Gainesville, Florida.

cSweet Water Branch, 1/4 mile down stream from the Gainesville
 Sewage Treatment Plant outfall.
When the samples were preserved by adding 3 ml of concentrated H-SO^,
the amounts of the saponifiable coprostanol were about one half of the
total coprostanol, as shown in Table 23.
                            TABLE 23

         EFFECT OF SAPONIFICATION ON PRESERVED SAMPLES*
Sample
Raw Sewage
Raw Sewage
Mean
Saponified
(yg/D
490
478
484
Recovery of Coprostanol
Unsaponified Saponifiable
(ug/1) Coprostanol
(%)
247
212
230
49.7
55.6
52.4
*The samples were acidified with 3 ml of concentrated H2SO^, and
 analyzed after several hours.  The samples were obtained at the
 City Sewage Treatment Plant, Gainesville, Florida.
                               68

-------
cholesterol by liquid column chromatography because of their similar
reaction with the columns;  however,  cholesterol (I) has a double bond
at As, which will react differently with certain reagents than will
coprostanol.  It can be inferred that there should be an appropriate
agent which is selective to cholesterol.
1.  R » H:  Cholesterol
    R » OAc:  Cholesterylacetate
II.  R - H:  Coprostanol
     R » OAc:  Coprostanylacetate
For example, ozone,  osmate or permanganate will react only with the
double bond of cholesterol.  Preliminary experiments showed that
cholesterol was oxidized by the modified Remieu-von Rudloff reagent
(periodate with a catalytic amount of permanganate), [60, 61, 62]
and coprostanol (II) was unchanged when the alcoholic groups of the
sterols were protected by acetylation.  Chromic acid anhydride
oxidation in glacial acetic acid also showed the same results on the
acetates of these sterols.  These reagents were not useful for analysis,
because coprostanylacetate (II) has a low color intensity.  The molar
extinction coefficient was determined to be 1,150 when color develop-
ment by the Zlatkis-Zak method was used.  This was lower than that of
coprostanol itself.

On the other hand, both cholesterol and coprostanol have alcoholic
hydroxy group at C, as another reactive site.  Fieser [63] reported
that the initial oxidation product of cholesterol with chromic acid
is the ketone, A5-Cholesten-3-one (III), and Ellis and Petrow [64]
reported that they obtained the 3-keto-cholesterolepoxide from 'V-
epoxy-cholesterol by milder oxidation with the chromic oxide-pyridine
complex.  This evidence indicates that the alcoholic groups of the
sterols can be converted to the ketones by mild oxidation.  The
Sarett reagent, anhydrous chromic acid-pyridine complex in glacial
acetic acid, was studied with the following results; cholesterol was
oxidized rapidly along with the rapid conversion of coprostanol to
coprostanone (IV).  In Figure 23 the recoveries of cholesterol,
coprostanol and coprostanone are plotted with reaction time.
                               69

-------
    100-
     0
                10
20
I
80
                                                                                             90
                                   Reaction Time in Minutes

FIGURE 23.  OXIDATION OF STEROLS AND FORMATION OF COPROSTANONE WITH THE SARETT REAGENT AT 20°C

                    A:  Coprostanol   B:  Cholesterol   C;  Coprostanone

-------
III.  A5-Cholesten-3-one
                                   IV.  Coprostanone
While the formation of coprostanone was quantitative and coprostanone
itself was resistant to the oxidation reagent, the formation of A5-
cholesten-3-one was not quantitative.  At least two oxidation products
of cholesterol in the hexane extract were identified gas chromato-
graphically as shown in Figure 24.  One of these has been identified
as A5-cholesten-3-one (III), by comparing with the retention time of
the authentic compound on the gas-chromatograph.  This compound
appeared dominantly at the early stage of the oxidation, i.e., in a
few minutes.  This identification is supported by Fieser [63] as he
reported that the initial product of the oxidation of cholesterol was
A5-cholesten-3-one, which undergoes further oxidation to 63-hydroxy-A1*
cholesten-3-one (V) and then to Al|-cholesten-3,6'-dione  (VI):
Cholesterol
                      A^-cholesten-3-one
V.
                 in AcOH-Benzene
SB-hydroxy-A^-cholesten-S-one
VI.  A^-cholesten-3,6-dione
The other oxidation product may have been the  second product,  68-hydroxy-
A4-cholesten-3-one (V).  This conclusion is  supported by Ellis and
Petrow [64], who could isolate only eg-hydroxy-A^-cholesten-S-one in
                               71

-------
                                                                                                 (SI
                                                                                                 r>
FIGURE 24.  GAS CHROMATOGRAMS OF OXIDATION PRODUCTS
             WITH THE SARETT REAGENT

            a:  Cholesterol
            b:  Coprostanol

            A:  Coprostanol
            B:  First Product
            C:  Second Product from Cholesterol
0
                                 Reaction Time in Minutes

-------
 the oxidation of "e"-epoxy-cholestan-3-ol with  the chromic acid-pyri-
 dine complex   Because the oxidation with chromic acid-pyridine com-
 datioin f II f /Cet±C aC±d  ±S mlld' ±t:  dOSS not cause further oxi-
 dation of 66-hydroxy group to the ketone.   In Figure 25, the changes
 in these two products are plotted with reaction time.  After choles-
 terol disappeared  the amounts of these  two products increased, with
 the rate of formation of A Wholes ten-3-one greater than that of 6g-
 hydroxy-A^-cholesten-3-one.

 These oxidation products can form hydrazones with 2,4-dinitrophenyl
 hydrazine (DNPH)  as orange precipitates, and interfere with the colori-
 metric determination of coprostanol.  The hydrazone of A5-cholesten-
 3-one is less soluble in hexane than that of coprostanone, but it can
 be partially extracted with the solvent when the hydrazone of copro-
 stanol is separated from unreacted 2,4-DNPH reagent.  The problem
 then becomes how to remove these interfering oxidation products.

 It was discovered that the second oxidation produce, eg-hydroxy-A^-
 cholesten-3-one,  could be eliminated by passage through an alumina
 column,  i.e.,  the product was eluted into the free sterol fraction
 with pure ethylacetate as shown in Table 24.  On the other hand, the
 first product  and also pure A5-cholesten-3-one, behaved on the column
 similar  to coprostanone,  and could not be eliminated by the alumina
 column.

 The product, A5-cholesten-3-one,  was more reactive toward oxidation
 reagents  and should undergo  further oxidation to the second product.
 As is  seen in  Figure  25,  the first oxidation product reached the max-
 imum yield  in  20  minutes  of  reaction time and was present even after
 40 minutes  at  the reaction  temperature of 20°C.   From this evidence,
 it could be  expected  that  the yield of the first product would become
 negligible,  if the  reaction  proceeded longer or if the temperature
 were  increased.   As longer reaction times than 30 minutes are not suit-
 able  for practical  analysis,  the  effects  of  temperature on the fate
 of  the first product,  A5-cholesten-3-one  at  40°, 60°  and 80°C was stud-
 ied.

 The changes of the  oxidation  products  of  cholesterol  at various  tem-
 peratures are plotted against reaction  time  in Figure  26 and  27.   The
 first product, A5-cholesten-3-one,  in Figure 26, decomposed  completely
within ten minutes, when the  temperature  was over  60°C.   However,
at 40°C, it was present for over  twenty-five minutes with the maximum
yield at ten minutes.  When the temperature  was  high,  80°C, the  first
product was oxidized within five minutes.  Figure  27 indicates the  change
in yield of the second product with reaction time at various  temperatures.

The above experiment indicated that cholesterol  can be  eliminated by
oxidation with Sarett reagent at a temperature of 60°C with at least
ten minutes reaction time.  However, coprostanol, when  oxidized at  80°C,
decomposed and the recovery of coprostanone  decreased  to  85 percent
                                73

-------


£
a fl>
(D M
H* O*
09 CO
ST
rt t i
31"
M 
\
\
\
\C
\
\
\
'
.
1


y
/
                                                B
                            10             20             30
                             Reaction Time in Minutes
  FIGURE 25.  YIELDS  OF  THE OXIDATION PRODUCTS OF CHOLESTEROL WITH THE
                       SARETT REAGENT AT 20°C
 A:  The First Product    B:   The Second Product
          C:  Cholesterol
          3  -
         *
         H-
       33 (D
       H- O.
       TO tOj
       ft H-

       ?^2
       o o
       r
      2,Si -
                                10        15       20
                           Reaction Tine in Minutes
                   25
FIGURE 26.  YIELDS OF THE FIRST  OXIDATION PRODUCT OF CHOLESTEROL AT
                       VARIOUS TEMPERATURES
                  A:  40°C   B:   60°C
C:  80°C
                                74

-------
                             TABLE 24

              Elution Pattern on an Alumina Column**



Solvent








Hexane

5% Ethyl-
acetate



Elution
Volume
(ml)







5.0
5,0
2.0

2.0

'O
ti
n)
r-t rH
0 0
M G
o) cd
4J + '
CO CO
CO O
rH Vl
O ft
,C O
CJ U
-*
-
_

_



0)
C
0
C
CO
4-1
CO
o
$»4
ft
0
CJ
-
-
+

+
4-1 1
o a
3 CU
T3 4-1
O CO
M 0)
PH iH
O
4-1 rC
co o
t 1 1
•Hin
pn <3 (U
c
CD T3 o
X C! I
EH (0 c*1
-
-
+

+
O
TTJ
O
l-i
PL)
C
O
o

en

cu
r~j
H
-
-
_

_



CO
c
0
i-l

o
o

rjj
^>
ffi
-
-
_

_
^— «fc
OO
CM
n^
03
o
CN

*^s








              2,0

              2.0

              2.0

Ethylacetate  2.0

              3.0
 *  H- Positive; + Trace; - Negative.

**  The moisture content of alumina absorbent was critical.  When the
    alumina was moistened in the atmosphere overnight, the separation of
    coprostanone failed, and 5 percent ethylacetate eluted the sterols
    and the second oxidation product  of cholesterol.  The absorbent must
    be used fresh or the moisture content controlled by the procedure in
    Appendix-6.
                                75

-------
  FIGURE 27.  YIELDS
    8 100
   o
   Hi
 (6 o
 (t> l-(
 S3 O
 n m
   n
      50 .
   g
   §
   (6
  T

  10        15         20
Reaction Tine  in Minutes

                  PRODUCT OF

                C:  80°C
                                                           25
                     A:   40°C   B:  60°C
         5 ing per tube
         40 rag per tube
                          Reaction Time in Minutes
                                                         20
FIGURE 28.  EFFECT OF Cr03 DOSAGE ON  COPROSTANONE IN THE OXIDATION OF
               STEROLS WITH THE SARETT REAGENT AT 80°C
                               76

-------
after ten minutes as shown in Figure 28.  From this observation, it was
found that the reaction temperature was critical and must be optimized
so as to decompose  the first product but not to degrade the resultant
oxidation product of coprostanol, coprostanone*

     Formation of Coprostanone-2,4-Dinitrophenylhydrazone and
                   Removal of Unreacted 2,4-DNPH

2,4-Dinitrophenylhydrazine (2,4-DNPH)  (VII), which is widely used for
the characterization of carbonyl compounds, was employed in this study
in absolute methanol solution*[64,65].  The coprostanone-2,4-DNPH1
(VIII), formed as shown below, is a yellow precipitate that is sparing-
ly soluble in water, but very soluble  in many organic solvents such as
hexane, acetic acid and acetonitrile,  even though it has polar substi-
tuents in the phenylhydrazine moiety.
                NHNH,
                         (Coprostanone)
VII,  2,4-DNPH
VIII.  Coprostanone-2,4-DNPH'
The formation of coprostanone-2,4-DNPHl  seemed to be quantitative,
since the hexane extract  from the reaction solution did  not  show  the
presence of coprostanone,

2,4-DNPH is soluble  in acidic solution,  however,  hexane  partially
extracts the reagent,  and may cause a high reagent background in  the
measurement of  the developed color.  Techniques for removal  of the
reagent have been proposed in the literature by absorption chromato-
graphy  [66,67]  and oxidative decomposition [68] of the reagent.   Ac-
cording to Reich, et.  al., [69] 2,4-DNPH can be removed  by oxidation
with either Fehling's  solution or Benedict's reagent, producing m-dini-
 *  95  percent ethanol is commonly used as a hydrazonation solvent,
    but this solvent gives problems with the oxidative removal of un-
    reacted 2,4-DNPH reagent with acidic dichromate.  The 2,4-DNPH
    does not dissolve completely in 2N IL^SO^ solution, in which the
    coprostanone-2,4-r»DNPH is suspended, and results in incomplete der-
    composition of unreacted 2,4-DNPH reagent.
                                 77

-------
trobenzene.  A new approach was tried in this study, i.e,, oxidative
decomposition of the unreacted 2,4-DNPH.  Several oxidation reagents
were tried.  Preliminary experiments showed that both potassium dichro-
mate in acidic solution and hydrogen peroxide in alkaline solution gave
satisfactory results.  The conditions required to obtain a reasonably
low reagent background were determined for dichromate oxidation.

In Table 25, the effects of extraction conditions are summarized.  When
the unreacted 2,4-DNPH reagent was not oxidized with acidic dichromate,
the reagent blank was high, E^j] = 0.570, however, when it was oxidized
the blanks were greatly reduced.  In the case of extraction from alka-
line solution reagent blanks of zero were obtained.

The effect of temperature on the removal of unreacted 2,4-DNPH reagent
is shown in Table 26.  It shows measurable values for the reagent blanks
at temperatures below 60°C; however, when the temperature was 80°C, the
reagent blank became zero.

                             TABLE 25
     EFFECT OF EXTRACTION CONDITIONS FOR COPROSTANONE-2,4-DNPH

                                           Reagent Blank Value*
Dilution Solvent                        Esoonm vs. Distilled Water
Not oxidized with Dichromate:**
Distilled Water
Oxidized with Dichromate:
2N H2S04
12N H2S04
4N NaOH
Solvent***
0.570
0.570

0.032
0.010
0.000
0.000
  *  The absorbence in 10 ml of solution.  The  color  development pro-
     cedure followed was the method discussed in  the  next section.
 **  After the hydrazonation, the solvent was diluted with 10 ml of
     carbonyl-free water without adding dichromate  solution or 5 ml
     2N H2S04 solution.
***  Ethylacetate-acetic acid concentrated HC1  (6:3;1, by bolume).

                             TABLE 26
	TIME AND TEMPERATURE EFFECTS ON REAGENT BLANKS	

Times in Minutes                Reagent Blanks    (E5oonm/10 ml)
                                20°C    40°C      60°C     80°C
5
i ft
J.U
20
0.017
Om (\
. UJ.O
0.014
0.010
Oftfl7
• UU/
0.009
0.007
OftftR
• UUO
0.006
0.000
	
   The color development method was the same as  in note  (*) of Table 25,
                               78

-------
                        Color Development

1,  The Color of Coprostanone-254-Dinitrophenylhydrazone in Basic
Solvent Systems — Coprostanone-2,4-DNPH is soluble in hexane and
shows a yellow color, which has no absorption maximum in the visible
region but does have a small shoulder near the ultra violet region,
at 415-390 nm, with a molar absorbence of 7,270,  As this shoulder is
too close to the UV region and the absorption intensity is too low,
it was difficult to use this band for analysis.

The Janovsky reaction [70] of m-dinitro compounds was considered, in
which an alkali-acetone solution  (5 ml of acetone and 2 ml of 5 percent
sodium hydroxide) was employed.  The resultant wine-red color was un-
stable and started to fade in a few minutes.  In order to find a suit-
able solvent to stabilize the color of coprostanone-2,4-DNPH, alkaline
solutions of acetonitrile and pyridine were prepared and evaluated.
The brown to violet colors were measured on a spectrophotometer at 465
nm in a 1 cm cell.

The results are summarized in Table 27 and indicate that the acetoni-
trile system was not suitable because of a low absorption intensity,
    = 0.001.*  The pyridine system resulted in a high absorption inten-
 y     ..
sity, E^Q = 0.014, but had  a high reagent background, E^g = 0.112.**
This hign reagent background was considered  to be due to m-dinitroben-
zene which was formed from  unreacted  2,4-DNPH reagent by oxidation.  In
spite of the high absorption intensity  of the colored product this
technique cannot be used for the determination of amounts of copro-
stanol of one microgram or  less because of the high reagent background.

2.  Acid Hydrolysis of Coprostanol-2,4-DNPH' , and Formation of Azo Com-
pounds — Arylhydrazines have been known  to be oxidized easily with many
oxidizing agents such as selenious acid and  arsenic acid, [71] forming
the corresponding diazonium salts, which undergo coupling with aromatic
amines and phenols forming  azo compounds.  The azo compounds have in-
tense colors leading to their extensive use  in analysis.  For example,
nitrites in the sea water  [72], and nitrogen oxides in the atmosphere
[73] are determined by use  of this class of  reaction,  Feigl [71] states
that arylhydrazones are detectable by coupling with oi-naphthylamine
after they are saponified with concentrated  hydrochloric acid and the
liberated hydrazines are oxidized with  selenious acid to the diazonium
salts.  The hydrolysis of coprostanone-2,4-DNPHr with concentrated hydro-
chloric acid was shown to result in poor recovery of the hydrazine due
*  Erj  is defined for convenience,  as  the absorbence of one microgram
   (coprostanol equivalent)  of  color materials  in 10 ml of the solu-
   tion, with 1 cm light path.
** E-'-P indicates the absorbence of  blanks in 10 ml solution at the
   maximum absorption wave  length of the  color  materials, with 1 cm
   light path, in this  case  at  465  mu.
                                79

-------
                             TABLE 27

        ABSORPTION INTENSITIES AND REAGNET BLANKS OF COLOR
          DEVELOPMENT TECHNIQUES IN BASIC SOLVENT SYSTEMS
Solvent System*
 Maximum
   Wave
Length, nm
Absorption  Intensity  Reagent
(E°Xmax)**    (Ej-g)     Blank
                MB
A.  Acetonitrile   4.0 ml
    Methanol 0.5N  1,0 ml
    KOH            1.0 ml     465

B.  Pyridine       3.0 ml
    IN KOH in MeOH 1.0 ml
    95% Ethanol    2.0 ml     465
  3,680
 53,000
                           0.001
                           0.014
0.050
0.112
 *  The solvent systems were  designed  to  maintain alkaline conditions
    and to dissolve the resultant m—dinitrobenzene from unreacted 2,4
    DNPH reagent.
**  Molar extinction coefficient.
to the stability of the hydrazone and the reduced solubility in concen-
trated hydrochloric acid.   The maximum recoveries were 90 percent.
                                                     NHNH+ +
Coprostanone-2,4-DNPH'
   2,4-DNPH
                                Coprostanone
Several solvents which might dissolve the hydrazone, were studied.
Equal volumes of these solvents were mixed with concentrated hydrochlo-
ric acid.  The solvents were:  acetonitrile, methanol, ethanol, butanol,
octanol, tetrahydrofurane, chloroform, acetic acid and propionic acid.
The best solvent system was a mixture of acetic acid and concentrated
hydrochloric acid (1:1) which led to quantitative hydrolysis as shown
in Figure 29.  Experiments were conducted to optimize the conditions,
i.e., time and temperature, in the hydrolysis of coprostanone-2,4-DNPH
with this solvent system.  The results are shown in Figure 30.
                                80

-------
              0.8  -
            §0.6
            cr
            5
            n
            n
            ^0,4
              0.2  -
            §
              0.0
                         I     III
                   0   0.5   1.0        2.0        3.0
                     Coprostanone-2,4-DNPH' Added (ml)
FIGURE 29.  HYDROLYSIS OF  COPROSTANONE-2,4-DNPHf WITH ACETIC ACID
                     CONCENTRATED HC1  (1:1)
                                81

-------
   0.20
 en
 o
 n
§00.10
   0.00-

        0
                    I

5        10        15        20

   Reaction Tine in Minutes
30
 FIGURE  30.   EFFECT OF TEMPERATURE  AND THIS ON THE HYDROLYSIS  OF

                     COPROSTANONE-2,4-DNPH'
          A:   20°C    B:  50°C    C:   90°C
                          D:100°C
                               82

-------
When the temperature was low, 20° and 50°C, the hydrolysis proceeded
slowly and was not complete even after 90 minutes; however, at higher
temperatures, 90° to 100°C, the hydrolysis was completed within 15
minutes.  From this study the optimum hydrolysis conditions can be
specified as:  100°C (boiling water) for ten minutes with 1:1 HCl-gla-
cial acetic acid.

As the oxidation reagent to convert hydrazine to the diazonium salt,
Feigl [71] recommended selenious oxide, a weak oxidizing agent.  Since
selenious oxide was not available, sodium periodate solution was used.
The periodate oxidized 2,4-DNPH to the diazonium salt almost instan-
taneously and quantitatively, as shown in Figure 29.
                               104
                N02

2,4-DNPH-hydrochloride
                           N02

                  Diazonium salt
Both a-naphthylamine  and  l-(N-naphthyl)-ethylenediamine were evaluated
as coupling  agents  for  the  analysis, by  comparing  their color intensi-
ties and reagent backgrounds.   The  absorption  spectra  are  shown in Fig-
ures 31 and  32.
         N02
Diazonium  Salt
R=H:  a-naphthylamine
R=CH2CH2NH2:

   1-(N-naphthyl)-ethylene
   diamine
N02

  Azo-Compounds
 In Table  28,  the results are summarized,  which indicate that the mea-
 surement  of  the color of the azo compounds of a-naphthylamine in acidic
 solution  was best,  because of the high absorption intensity and low re-
 agent blank.   With  this color development method, it is possible to de-
 termine one  microgram of coprostanol;  EJ*   = 0.046 with reagent blank
 £3, - 0.004.   The reagent blanks in basic solution gradually increased
 in contact with the atmosphere.   This  phenomenon seems to have been
 caused by the oxidation of the naphthylamines by oxygen.
                                83

-------
  0.60
  0.50
0)
o

g-
§0.40-
o
ID
 g
  0.30 '
  0.20-
  0.10
  0.00.
            460
500       540      580

     Wave Length  (nm)
700
    FIGURE 31.  ABSORPTION SPECTRA OF THE  2,4-DINlTROPHENYLAZO—

            DERIVATIVE OF 1-(N-NAPHTHYL)-ETHYLENEDIAMINE

    A:  In acidic solution; water 8.00 ml,  acetic acid 0.50 ml and con-

        centrated HC1 0.5Q ml.  50 yg as coprostanol equivalent in 10 ml
        of the solution X
                         max
         520 nm
    B:  In basic solution; pyridine-water.   45  wg as coprostanol equi-
        valent in 10 ml of the  solution.   A,
                                           •max
                           600 nm
                                 84

-------
  0.40
  0.30
g 0.20
i-t

-------
                             TABLE 28

ABSORPTION INTENSITIES OF THE AZO COMPOUNDS AND REAGENT BACKGROUNDS

Coupling Agent
•
l-(N-naphthyl)-
ethylenediamine


a-naphthylamine



Condition
of Color
Solution*
acidic

basic

acidic

basic


E°X max**

32,700

56,900

53,200

57,200

F10
Ue
ro
(Eug)
0.008
(0.025)
0.015
(0.050)
0.014
(0.046)
0.015
(0.051)
V10
Erb

0*8)
___—
( 	 )
0.009
(0.030)
0.001
(0.004)
0.013
( 	 )
 *  The solvent system,  acetic acid-concentrated HCl-water  (9:0.5:
    0.5) was used for  acidic conditions and pyridine-water  (vari-
    able ratios) for basic  conditions.
**  Molar extinction coefficient, as coprostanol equivalent.
The intense pink to orange color of the azo  compound,  2,4-dinitrophenly-
azo-ct-naphthylamine-hydrochloride, was stable  in  acidic  solution for at
least 30 minutes, however, the color gradually faded and finally turned
yellow.  It was found that the fading of the color was not  due to the
effect of periodate, but probably to autoxidation, as  shown in Figure
33, where the effect of time on the color intensity was  examined.  In
Figure 34, the absorption intensity was calibrated by  using pure 2,4-DNPH.

3.  Calibration—Known amounts of coprostanol  (0.25 to 50.00 ug) were
converted to coprostanone, hydrozonated, and the  resultant  hydrazone
was hydrolysed at 100'C for 10 minutes.  The liberated 2,4-DNPH was
converted to the azo compound by coupling with a-naphthylamine and the
absorption intensity determined.  Unfortunately,  the calibration test
resulted in high and random reagent backgrounds,  which were due to the
appearance of white turbidity from m-dinitrobenzene that was derived
from decomposition products of the 2,4-DNPH reagent.   m-Dinitrobenzene
is a stable compound, insoluble in water but soluble in  many organic
solvents.  The solutions in organic solvents do not show absorption
in the visible wave length range.  If the compound were  solubilized in
one of the above solvents, the turbidity could be eliminated. Reich,
et.al.,  [69] experienced the same problem, which  they  overcame by dis-
solving  the 2,4-DNPH of a keto-steroid in chloroform.

A brief  study was conducted to evaluate the  solvent extraction and to
                               86

-------
   0.30
   0.25
B
°§0.15
w
 
-------
    0.70 '
    0.60 -
    °-50 '
 (0
 o
 >t
 a*

 8
    0.40 "
Oil-1
o
3,.  0.30


 M
 O




 N^/


   0.20
   0.10
   0.00
                             20
   0        10         20         30       40        50

    2,4-DNPH Added  (ug as  Coprostanol Equivalent)


FIGURE 34.  CALIBRATION OF 2,4-DINITROPHENYLKYDRAZINE
                                 88

-------
optimize the extraction conditions by neutralizing the acidic solution
with pyridine.

In Table 29 the effects of the hydrolysis solvents and the pyridine base
solutions are summarized.  When the solution of the azo compound was not
acidified by the hydrolysis solvent, there were no differences in the
color intensity between the samples which were diluted with different
strengths of the pyridine base solution.  When a sample which was acidi-
fied with the hydrolysis solvent, was diluted with distilled water and
extracted with ethylacetate, poor recovery of the azo compound resulted
only 86 percent.  The acidified samples, which were diluted with the
pyridine base solutions, showed an improvement in recovery which indi-
cated that the pyridine content was critical, i.e., when the solution
of pyridine-acetic acid-water  (10:8:82) was used to neutralize the acids,
the recovery was still low—91 percent.  If the pyridine content was
increased to 15 percent pyridine, pyridine-acetic acid-water (15:5:80)
100 percent recovery was obtained.  In the latter case, the reagent
blanks became almost zero.

In summary, the optimum color development procedure was as follows:

1.  Hydrolysis of coprostanone-2,4-DNPH—Dissolved first in 0.5 ml of
    glacial acetic acid, then added 0.3 ml of concentrated hydrochloric
    acid, and heated to 100°C in a boiling water bath for ten minutes.

2.  Diazonium salts formation—The above solution was diluted with 5
    ml of distilled water and one ml of periodate solution added,
    shaken thoroughly, and allowed to stand for about one minute.

3.  Coupling reaction—To the above mixture, one ml of a-naphthylamine
    solution was added, shaken well and allowed to stand for a minute.

4.  Extraction of the azo compound—The solution was then diluted with
    5 ml of pyridine-acetic acid-water  (15:5:80, by volume) and extract-
    ed twice with ethylacetate  (3.0 ml first, 2.0 ml second).  The ethyl-
    acetate extracts were poured into a 10 ml flask containing 3.00 ml
    of glacial acetic acid.

5.  Stabilization of the color and  measurement—To the flask, 1.00 ml
    of concentrated hydrochloric acid was added and the sample diluted
    with pure ethylacetate exactly to 10.00 ml.  The color intensity
    was measured at 500 nm in  a 1 cm cell.

Using the above procedure, 0.25 to 50.00 yg  samples of coprostanol were
used to prepare a calibration  curve:  The free coprostanol was converted
to coprostanone and the resultant coprostanone was separated by an al-
umina column.  The coprostanone was hydrozonated and the excess 2,4-DNPH
reagent was removed at 80°C for 10 minutes.  The calibration of copro-
stanol, 3.0 to 50.0 yg range,  is shown in Figure 35 and the calibration,
0.25 to 3.00 yg, range, is shown in the Figure 36.
                                89

-------
   0.50  '

 o4
 I
 (D
 /•»»
 M

pi
    0.40
    0.20  -
   0.10  .
   0.00
          0  3 5   10       20        30
                  Coprostanol (yg in 10 ml)
AT)
50
   FIGURE 35.   CALIBRATION OF COPROSTANOL  (3-50  yg/Range)
                            90

-------
    0.070 -
    0.050
 CD
 o
 3-
 o
 (D
00
og
 UJ
    0.030
     0.0.0
     0.000
                              1.0       1.5       2.0      2.5
                              Coprostanol (yg in 3.5 ml)
        0  0.25  0.5


FIGURE 36,  CALIBRATION  OF COPROSTANOL (0,25-3  ug Range)
                               91

-------
                             TABLE 29

EFFECT OF PYRIDINE BASE SOLUTION ON THE RECOVERY OF THE AZO COMPOUND BY
                      ETHYLACETATE EXTRACTION
Hydrolysis Solvent*   10% Pyridine**  15% Pyridine***  Distilled Water
Absent

Present

0.513****
(100%)
0.465
(91%)
0.514
(100%)
0.516
(100%)


0.442
(86%)
                          Reagent Blank   0.000
                                          (0.000 - 0.003)
   *  Pyridine-acetic acid-water  (10:8:82).
  **  Pyridine-acetic acid-water  (15:5:80).
 ***  0.5 ml of glacial acetic acid and 0.5 ml of concentrated HC1.
****  Absorbence in 10 ml of ethylacetate-acetic acid-concentrated
      HC1 (6:3:1) mixed solvent, at 500 mu in one cm cell.
                   Evaluation of Analytical Time

The total analytical time required for a single  sample was estimated
roughly to be two hours and a half.  Table 30 shows the time required
for each step.

                             TABLE 30

                    ESTIMATE OF ANALYTICAL TIME


    Process                  Step               Time  (Minutes)
1.

2.


3.

4.


Recovery

Saponification


Oxidation

Chromatography


Hexane Extraction
Drying
Boiling
Extraction
Drying
Reaction
Extraction
Hexane Elution
5% Ethylacetate Elution
Drying
20
10
5
10
5
20
10
15
5
5

30


20

30


25
                               92

-------
                             TABLE 30

                   ESTIMATE  OF ANALYTICAL TIME  (Cont.)
    Process                  Step                Time (Minutes)
5.  Hydrazonation       Hydrazone Formation        5
                        Removal of the Reagent     5
                        Extraction                10
                        Drying                     5     25
6.  Color Development   From the Hydrolysis to
                        the Azo dye formation     15
                        Extraction                 5
                        Color Measurement          5     25
                               Total Analytical Time   2:35
The above time must be considered a minimum or an attainable goal when
performed by a well-trained analyst.  However, the time can be reduced
by modifying the procedure.  For example, if the saponification of the
hexane extracts in Step  2  is combined with the oxidation of the ex-
tracts, the analytical time will be reduced by 15 minutes by eliminat-
ing the extraction and drying  steps in the saponification process.  If
the elution rate is increased  in the chromatographic separation of co-
prostanone, the time will  be reduced, and so on.  Several attempts to
reduce the analytical time were studied.

                         Summary of Results

It was found that the saponification and oxidation steps could be com-
bined by increasing the  amount of chromic acid slightly and the elution
time could be reduced by 10 minutes; an overall reduction of about 25
minutes.

1.  Reduction of Analytical Time by Combination of Saponification and
Oxidation:  The feasibility of the combination was evaluated.  Choles-
terol was included in order to estimate the background derived from
100 pg of cholesterol.   The results are shown in Figure 37.  Choles-
terol and coprostanol were oxidized in about ten seconds and the re-
sultant coprostanone represented 100 percent recovery in ten minutes,
and the first oxidation  product of cholesterol, A5-cholesten-3-one,
disappeared completely after ten minutes, as shown in Figure 37 where
the recovery of coprostanone and A5-cholesten-3-one are plotted with
time.  The results obtained here were almost the same as those of the
uncombined oxidation except that additonal chromic acid of the Sarett
                                93

-------
                              4- 100%
                             -O	
                         20       30         40
                     Reaction Tine in Minutes
 I
50
60
FIGURE 37.  RECOVERY OF COPROSTANONE BY THE  COMBINED PROCEDURE OF SA-
                    PONIFICATION AND OXIDATION

            A.  Coprostanone
            B.  A^-Cholesten-3-one
                                 94

-------
reagent was consumed.  The dosage of chromic acid-pyridine complex is
critical in the combined procedure.  It was found that 40 mg of chromic
acid was sufficient for successful analysis.  It was found, further,
that the effect of the oxidation products of cholesterol or the color
background was negligible, as shown in Table 31.
                             TABLE 31

       COLOR BACKGROUND OF OXIDATION PRODUCTS OF CHOLESTEROL
Time of
Oxidation
(minutes)
5
10
20
30
40
60
Recovery

Cholesterol
0.00
it
it
M
it
it
Based on GLC Peak Height
(inches)
First Oxidation Product
0.07
0.00
it
it
it
it
Color
Intensity
(Elg)
0.010
0.000
0.004
0.001
0.001
0.002
2.  Rapid Chromatographic  Separation:  The time for the chromatographic
separation of  coprostanone on  an alumina column was 25 minutes as shown
in Table 30; 15 minutes  for dissolving the sample in hexane and a hexane
wash, 5 minutes for  elution and 5 minutes for drying the eluate.  The
column used was a  funnel tube, 3.5 mm in diameter, packed with 0.5 grams
of 80 to 200 mesh  alumina.   The elution rate of this column was 0.6 ml
per minute.  A study was conducted to determine whether the separation
of coprostanone could be effectively accomplished when the elution rate
was increased  to 1.5 ml  per minute by using a larger diameter column.
In Table 32, the recovery  of coprostanone and cholesterol are shown.
In the first 5 ml  of the 5 percent ethylacetate elution,v coprostanone
was recovered  in 100 percent yield.  The time for the separation was 15
minutes, 10 minutes  for  the elutions of the hexane extracts, washing
with hexane and 5  percent  ethylacetate, and 5 minutes for drying.  This
procedure reduced  the analytical  time by 10 minutes without losing ef-
ficiency.  These changes could reduce the overall analytical time to
just over two  hours.
    This  2N H2SO^  solution does not contain 10 percent  NaSO^   Sodium
    acetate in the reaction mixture forms soduim sulfate by reacting
    with  2N H2&04, which is sufficient to effect salting out,  and  if
    the 2N ^SO^ contains 10 percent Na2S04,  it causes  low recovery of
    coprostanone by precipitating soduim sulfate crystals when the co-
    prostanone is  extracted with hexane.
                                95

-------
                              TABLE 32

   SEPARATION OF COPROSTANONE AND CHOLESTEROL ON AN ALUMINA COLUMN
Solvent                 Elution                Recovery %
                      Volume (ml)     Coprostanone     Cholesterol
Extracting Hexane
Washing Hexane
5% Ethylacetate




Ethylacetate
6
5.0
1.0
1.0
1.0
2.0
5.0
5.0
0.0
0.0
0.0
39.3
56.0
4.7
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
100.0
                   Summary of Colorimetric Method

The following is a summary of the analytical method:

1.  Recovery of Coprostanol:  Place one liter of the sample water in a
    2-liter separatory funnel, and add 3 ml of concentrated sulfuric
    acid and 40 grains of sodium chloride  crystals.  The sample is ex-
    tracted with 50 ml of hexane by shaking for two minutes.  Transfer
    the water layer to another 2-liter separatory funnel and extract
    with another 50 ml of hexane.  Discard  the water layer and combine
    the two hexane layers.  The hexane is washed with 25 ml of 70 per-
    cent ethanol twice and 25 ml of acetonitrile saturated with hexane
    twice.  The hexane layer is evaporated to reduce the volume to a few
    milliliters under vacuum on a water bath at a temperature below 60°C.
    The concentrated solution is then transferred to a test tube with a
    transfer pipet and dried.

2.  Saponification and Oxidation:  The dried residues are dissolved in
    1.0 ml of 0.5N NaOH in absolute methanol and heated to boiling until
    the volume is reduced to 0.2 ml or less, cooled and dried.  The dried
    residues are dissolved in 0.5 ml of glacial acetic acid, 1.0 ml of
    the Sarett reagent added, warmed for 20 minutes at 50° + 1°C in a
    water bath, and then cooled to room temperaturet The mixture is di-
    luted with 10 ml of 2N H2SO^ and extracted with 2 ml of hexane three
    times.

3.  Separation of Coprostanone:  The hexane solution is passed through
    an alumina column, and the column is washed with 5 ml of pure hexane.
    The Coprostanone fraction is obtained from the column by eluting with
    5 ml of 5 percent ethylacetate in hexane, and the eluate dried.
                                96

-------
4.  Hydrazonation:  The well-dried 5 percent ethylacetate eluate is dis-
    solved  in 0.5 ml of carbonyl-free methanol, 0,5 ml of 2,4-DNPH re-
    agent added and allowed to stand for 5 minutes.  The methanol solu-
    tion diluted with 5 ml of 2N H2S04 in carbonyl-free water and 1.0 ml
    of potassimum dichromate solution is added.  The mixture is then
    heated at 80°C for 5 minutes in a water bath, shaken several times
    and cooled to room temperature.  The solution is diluted with 5 ml
    of 4N NaOH and extracted with 2 ml of hexane three times.  The hexane
    is evaporated to dryness by blowing air over the surface.

5.  Color Development:  The well-dried, hydrazonated coprostanone is dis-
    solved in 0.5 ml of glacial acetic acid and 0,5 ml concentrated hy-
    drochloric acid added.  The mixture is heated for 10 minutes in boil-
    ing water (100°C) cooled in cold water and then diluted with 5 ml of
    distilled water.  To the diluted solution, 1.0 ml of sodium periodate
    solution is added and shaken well.  Then 1,0 ml of a-naphthylamine
    solution is added, shaken thoroughly and allowed to stand 5 minutes.
    The solution is then diluted with 5 ml of the pyridine base solution
    and extracted with pure ethylacetate twice, first 3 ml and second 2
    ml.  The ethylacetate is stored in a 10 ml flask containing 3.00 ml
    of glacial acetic acid, to which 1.00 ml of concentrated hydrochlo-
    ric acid is added.  The mixture solution is diluted with pure ethyl-
    acetate to exactly 10.00 ml.  The pink color of the azo compound
    is measured at 500 nm in a 1 cm cell using a spectrophotometer.

              Application of the Colorimetric Method

The colorimetric method was applied to actual field samples and the re-
sults obtained are discussed in this section.

The colorimetric method used to quantify coprostanol and its analogs is
based upon changing them to coprostanone and then eliminating interfer-
ing materials such as cholesterol, hydrocarbons, esters and other hex-
ane-soluble constituents which are more polar than coprostanone.  Sew-
age and natural waters contain many organic compounds; for example, a-
bout 50 mg of hexane extractables were present in one liter of raw sew-
age, the structures of most of which have not been elucidated.  These
could interfere with a colorimetric method, since many of these consti-
tuents have chemical natures similar to coprostanol (alcohol) or copro-
stanone (ketone).

To establish the colorimetric method as a measure of fecal contamination
and domestic wastewater pollution, the method has to detect only com-
pounds of fecal origin.  The optimum method would detect only coprosta-
nol.  A field trip was conducted to estimate the interference of unknown
constituents in the sample waters on the colorimetric determination of
coprostanol.

Sampling stations that were included were:

1.  Orlando City Sewage Treatment  (West and East Plants):  Raw sewage,
                                97

-------
    trickling filter, activated sludge effluents and finished waters.
    Fifty to 300 ml samples were diluted with distilled water to one
    liter.

2.  University of Florida Campus Treatment Plant:  Raw sewage, 200 ml
    was diluted with distilled water to one liter.  One liter of fin-
    ished water was collected.

3.  Lake waters:  (a) Newnan's Lake (located east of the City of Gaines-
    ville) - One liter of the water was collected at the west shore boat
    harbor.  The lake is characterized by high color concentration; (b)
    Lake Alice  (located on the University of Florida campus) - One liter
    of the water was sampled at the outfall.  The lake is eutrophied.

4.  Tap water:  City of Gainesville.  One liter sample.

5.  Control:  Carbonyl-free water, one liter sample.

The coprostanol concentrations were determined both by GLC and by the
colorimetric method as detailed in the summary.  Results of the two
methods are compared in Table 33.

Analyses of the lake samples by the colorimetric method detected small
amounts of color-sensitive materials as coprostanol equivalent, which
were not coprostanol.  The GLC method did not detect any trace of copro-
stanol in these samples.  Tap water and carbonyl-free water samples
did not show any traces of the color-sensitive materials.  This would
suggest that the lake waters gave the natural background of the colori-
metric method.  Since these lakes are characterized as organic rich
waters, this should lead to maximum natural background interference.

As expected for the sewage samples, the colorimetric method gave higher
results than the GLC method.  The coprostanol contents measured by the
colorimetric method are plotted versus those of GLC measurement in Figure
38.  The regression coefficient was computed to be 1.61 and the standard
error of the ratios of the colorimetric to the GLC results was calcu-
lated to be 0.34.

The above results indicate that sewage samples contained some colorime-
try-sensitive materials.  This overestimation should not lead to the
conclusion that the colorimetric method is not suitable for the evalu-
ation of domestic waste pollution, as discussed below.

The gas chromatograms of the coprostanol fractions (5 percent ethylace-
tate fraction) of sewage samples showed the existence of an unknown
compound between the two oxidation products of cholesterol resulting
from oxidation with the Sarett reagent, as shown in Figure 39.  The
following calculations were conducted to infer statistically the con-
tribution of the unidentified peak compound versus the heights of the
coprostanone peaks.  The correlation coefficient was computed between
                                98

-------
   2000
o
o
o
Cfi
rf


g  1000
o

o
03
O

3
n
o
00
    500
    200
    100
     50
     20-
     10
Y = 1.61 X -f 12
                  •         i           11            it

       10        20       40         100       200         500      1000

             Coprostaaol Concentration by GLC  (yg/l)-X-



 FIGURE 38.  COMPARISON OF THE COLORIMETRIC METHOD VALUES AND ON GLC

                               VALUES
                                 99

-------
                      FIGURE 39.   TYPICAL GLC PATTERN
                            OF THE 5 PERCENT ETHYLACE-
                            TATE FRACTION OF SEWAGE
                                     SAMPLES.
                          Coprostanone
                          Unidentified-II

                          Coprostanol
                          First Product of Cholesterol
                          or A5-Cholesten-3-one
                          Second Product of Cholesterol
                          Cholestanol
                          Cholestanone
o
o
Retention Time

-------
the two values.  The standard deviations:

     sj = (0.375)2 ;          s£ = (0.0240)2

The correlation coefficient:
                              r = 0.58

The confidence level of the value r was computed by using the variable
z - 1/2 In YJT"  with standard deviation approximately l/N-3.  The fol-
lowing range of the correlation coefficient p was obtained at the 5
percent level of significance:
                              - 0.14 < p < 0.90

As the value obtained (r - 0.58) was between the above values, the hy-
pothetical coefficient is accepted with 95 percent confidence.  The
above statistical inference indicates that the unidentified compound
might have contributed to the overestimation of coprostanol as deter-
mined by the colorimetric method.  Since the correlation  coefficient
was not large, there may be other factors which caused the higher
results in the colorimetry.


                             TABLE 33

            COPROSTANOL CONCENTRATION OF FIELD SAMPLES
Sample
Coprostanol
GLC
Sewage Plant
I


Sewage Plant
II

Sewage Plant
III
Lakes

Control


1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.

Finished
Trickling Filter
Activated Sludge
Raw Sewage
Finished
Trickling Filter
Raw Sewage
Finished
Raw Sewage
Newnan's Lake
Lake Alice
Tap Water
Carbonyl-Free
Water
19
76
218
292
272
285
680
15
88
u.
u.
u.
(0.

(yg/D
.8
.7





.1

d.*
d.*
d.*
0)

Concent rat ion
Colorimetry
31.
186
409
494
343
439
1,146
21.
169
2.
2.

0.

7






1

5
5

0

Ratio
Color. / GLC
1
2
1
1
1
1
1
1
1





.601
.425
.877
.692
.261
.540
.685
.397
.920





   *  Undetected by GLC,  lower than 0.05 yg/1 (shown as 0.022  in Table
      11  for Lake  Alice).
  I.  Orlando  City Sewage Treatment Plant (West).
  II.  Orlando  City Sewage Treatment Plant (East).
 III.  University of Florida,  Campus Treatment Plant (Weeken     aples).
                                101

-------
                             TABLE 34

        STATISTICS OF THE OVERESTIMATION OF COPROSTANOL BY
                       COLORIMETRIC METHOD
Sample
Number
(from Ta-
ble 33)
1
2
3
4
5
6
7
8
9
EX «
IY =
Ratio of
Overestimation

(X)
0.601
1.425
0.877
0.692
0.261
0.540
0.685
0.397
0.920
6.398
1.104
Ratio of
Unidentified
Peak
(Y=b/a)
0.109
0.178
0.154
0.119
0.113
0.117
0.111
0.130
0.109
IX2 = 5.472776
IY2 - 0.149002
Peak Height
(Inches)
Coprostanone Unidentified

(a)
1.01
1,01
2.80
7,65
3.90
2.73
3.96
3.22
3.76
.ZXY= 0.857163


(b)
0.11
0.18
0.48
0.91
0.44
0.32
0.44
0.42
0.41


An estimate of the neutral steroid levels in a domestic sewer system
would give an approximation which can explain the higher results of
the colorimetric method.

Danielsson and Tchen [74] reported that human excretion of neutral
steroids was 500 to 700 mg per capita per day.  If it is assumed that
the average is 600 mg and use of water is 400 liters per capita per day,
the concentration of neutral steroids in raw sewage is computed to be
600 =1.50 mg/1.  The concentrations of cholesterol and coprostanol in
raw sewage that were obtained in this study were approximately 600 yg/1
each.  When these two sterols are  subtracted from the neutral steroids,
the concentration of other neutral steroids is calculated to be 1.50 -
(0.60 + 0.60) = 0.30 mg/1.  Since the coprostanol concentrations, as de-
termined by the colorimetric method, did not include cholesterol, the
total neutral steroids are 0.30 + 0.60 = 0.90 mg/1, and the ratio of
this concentration to the coprostanol concentration is determined to be
0.90/0.60 = 1.50.  This is almost the same ratio a5 the regression coef-
ficient, which was calculated to be 1.61 from Figure 38.  It can be con-
cluded £rom the calculations, even with the approximations that were
made, that some neutral steroids from human intestinal tracts might have
contributed to the higher results of the colorimetric determination.

The chemical nature of the unidentified compound present in the gas
                                102

-------
chromatograms of sewage samples was examined by the following experi-
ment:  Two hundred milliliters of raw sewage was placed in a 2-liter
separatory funnel and diluted with 800 ml of distilled water, then 3
ml of concentrated H^SO/ was added, followed by 40 grams of NaCl, which
was dissolved.  The mixture was extracted with hexane and the hexane
was washed with 70 percent ethanol and acetonitrile and dried in a test
tube.  The dried residues were redissolved in 4.00 ml of hexane and 1.00
ml was pipeted into three tubes and dried.  They were then treated as
follows:

     Tube A.  Not treated.          Tube C.  Saponified and oxidized.
     Tube B.  Saponified.

The dried residues, after the above treatments, were dissolved in 6 ml
of hexane and passed through alumina columns.  The columns were eluted
first with 5 ml of pure hexane, then 5 ml of 5 percent ethylacetate in
hexane and then 5 ml of pure ethylacetate.  The eluates were evaporated
to dryness.  The residues from each fraction were gas chromatographed.
The results are shown in Figure 40.

The unidentified compound (Unidentified-II) appeared in the 5 percent
ethylacetate fraction (coprostanone fraction) only when the sample was
saponified and oxidized  (tube C).  The ethylacetate fraction (sterol
fraction) of unoxidized samples (tubes A and B) showed the existence
of another unknown peak  (unidentified I) with a retention time shorter
than that of Unidentified II.  The Unidentified I was not observed in
the 5 percent ethylacetate fraction of the oxidized sample (tube C).

The above evidence indicates that the unknown compound which appeared
in the coprostanone fraction and seemed to have caused higher results
by the colorimetric method, was a ketone formed by oxidation with the
Sarett reagent.   It  might have been a keto-sterol, i.e., an alcohol,
before oxidation.  That the unknown compound might have been present
in sewage in the form of a keto-sterol, was suggested by the behavior
of the second oxidation product of cholesterol, 68-hydroxy-cholesten-
3-one, which is a keto-sterol and appeared in the sterol fraction from
the alumina column.  Correspondingly, it is possible that the unknown
compound  (Unidentified II) was a polyketo-compound.

On the other hand, an experiment was performed where the hexane  extracts
of sewage samples were saponified, esterified, and then oxidized, both
with chromic acid anhydride  in acetic acid and with the Remieu-von Rud-
loff reagent  (periodate with catalytic amount of permanganate in approxi-
mately 80 percent acetic acid).  The gas chromatograms of the hexane
fraction  (ester fraction) from an alumina column indicated the exis-
tence of  an unidentified peak  (Unidentified IV) as shown in  Figure 41
(cholesterylacetate was  oxidized completely without coprostanylacetate
being affected).  The compound, Unidentified IV, was not found to be
the oxidation product of cholesterylacetate, and therefore,  might have
been derived  from sewage constituents.  The oxidation products of cho-
                                103

-------
   0
                T	T
      2468
                Retention Time in Minute

FIGURE 40.  GAS CHROMATOGRAMS OF THE 5% ETHYLACETATE AND
      ETHYLACETATE FRACTIONS FROM AS ALUMINA COLUMN
A:  Not Treated
B:  Saponified
C:  Saponified and
      Oxidized
                a:  Coprostanol                 e:
                b:  Cholesterol or Coprostanone
                c:  A5-Cholesten-3-one          f:
                d:  Second Oxidation Product of
                       Cholesterol
Unidentified-
  I
Unidentified-
  II
                                104

-------
FIGURE 41.  GAS CHROMATOGRAMS OF HEXANE FRACTION OF SEWAGE SAMPLES
    a:  Coprostanol; b:  Cholesterol; c:  Coprostanyl acetate; d:  Cho-
    lesteryl acetate; e:  Unidentified-Ill; f:  Unidentified-IV; g:
    Unidentified-V; h-i:  Oxidation products of cholesteryl acetate
    (not eluted with hexane).
                                105

-------
lesterylacetate were eliminated by the alumina chromatography.  As the
Remieu-von Rudloff reagent cleaves double bonds, Unidentified IV seems
to be a saturated compound.  Although there is not a direct proof to
show that Unidentified IV and Unidentified II have the identical pre-
cursor, the behavior of these two compounds on alumina columns and their
reactivities to the saponification and oxidation reagents, suggested
that the Unidentified IV was the ester and the Unidentified II was the
ketonic form of an identical compound.  The peak at the detention time
of 5.6 minutes of the saponified (but not esterified) sample, disap-
peared after it was esterified and two new peaks appeared, one of which
was decomposed by the oxidation and another was unchanged  (Unidentified
II), was the same as the evidence that was obtained in the other experi-
ment, i.e., Unidentified I appeared in the saponified (but not oxidized)
sample and it disappeared by oxidation with the formation of a new peak
(Unidentified II).

The above evidence indicated that Unidentified II might have been a satu-
rated polyketonic compound.

Cholestanol and cholic acid—the possibilities of cholestanol (IX)
and cholic acid (X) as contributors to the overestlmation by the colori-
metric method, were considered.

1.  Cholestanol (5
-------
colorimetry.  Two hundred ug of cholic acid was treated by saponifica-
tion, oxidation and alumina column separation.  The result showed that
cholic acid was not detected by the colorimetric method.  The acid was
not eluted by 5 percent ethylacetate in hexane from an alumina column
even after it formed a ketone structure, because it still possessed
polar carboxyl groups in the side chain.  If the carboxyl groups were
split off by microbial action to form 20 or 17-ketones, the products
would have contributed to the colorimetric overestimation.  Although
this degradation process is not confirmed, the cleavage of the side chain
would seem to be possible, forming 17-ketones, by analogy with the fact
that progestrone metabolites were reported to have formed 17-ketones by
splitting of the methoxy group at 17C by micro-organism [77].

It has not been possible to identify positively the exact structure
of the substances that gave high values for the colorimetric copro-
stanol analyses but it is almost certain that they are of fecal or
urinal orgin.

                   Summary of Field Studies

The colorimetric method was applied to field samples.  The method over-
estimated the coprostanol concentrations of sewage samples by a factor
of 1.6 as compared to the results obtained by. GLC.  Some natural waters
gave a trace of colorimetry-sensitive materials which were not copro-
stanol.  Tap water did not show any color sensitivity.

An unknown compound which seemed to have caused the overestimation of the
colorimetry, appeared on the gas chromatograms of sewage samples.  The
column characteristics and the reactivities of the unknown compound to
saponification, oxidation and esterification indicated that the unknown
compound was a saturated keto-steroid from fecal or urinal excretion.

Although the colorimetric method measured not only coprostanol but also
other materials, it can be used effectively for the evaluation of fecal
and domestic wastewater pollution for the following reasons:

1.  The color-sensitive materials seem  to originate in  fecal or urinal
    excretion, and
2.  The natural background of the colorimetric method  is  low.

The colorimetric method is simple, rapid  and  sensitive.   The results
can be obtained within two hours, and the method  is less  expensive
than GLC.  It could be used for routine analyses  by sewage treatment
plant operators and field observers.
                                107

-------
                         ACKNOWLEDGEMENTS
The study of correlation of coprostanol concentration with other
treatment plant efficiency parameters was conducted by Samuel
Yocha Wang.

The financial support of the project by the Environmental Protec-
tion Agency and personal support by Edmond P. Lomasne of the At-
lantic Regional Office is acknowledged with thanks.  The technical
advice and support of William T. Donaldson of the Southeast Environ-
mental Research Laboratory of Environmental Protection Agency is
particularly appreciated.
                                108

-------
                            REFERENCES


 1.   Standard Methods for the Examination of Water and Wastewater, 13th
     Edition, 684 (1971).

 2.   Kukchek, G,  J.  and Edwards,  G.  P.,  JWPCF, 34, 376-89 (April 1962).

 3.   O'Shea,  J.  and Bunch, R. L., JWPCF, 37, 1444-46 (1965).

 4.   Baldwin, E., Introduction to Comparative Biochemistry, 3rd Edition,
     Cambridge Univ. Press., (1949).

 5.   Murtaugh, J. J. and Bunch, R. L., JWPCF, 39, 404-409 (1967).

 6.   Smith, L. L. and Ma thews, W. S.,"7th Int'l Congr. Biochem.", Tokyo,
     Japan.

 7.   Jeffrey, L.  M., JAOCS, 43, 211-14  (1966).

 8.   Smith, L. L. and Gouron, R.  E., Water Research, 3, 141-148  (1969).

 9.   Black, A. P., Singley, J. E. and Nordstrand, E., Unpublished Work.

10.   Gould, R. G. and Cook, R. P.,  "The Metabolism of Cholesterol and
     Other Sterols in the Animal Organism," In Cholesterol,(Edited by
     R. P. Cook)  Academic Press, 289-92, New York.

11.   Seylers, H., Z. Physiol. Chem., 348, 1688 (1967).

12.   Danielsson,  H., "Present Status of Research  on Catabolism  and Ex-
     cretion of Cholesterol."  Adv.  in Lioid Res.. 1,335  (1963).

13.   Cook, R. P., et.al., Biochem.  J..  62,  225 (1956).

14.   Aylward, F,  and Wills, P, A,,  Brit. J. Nutr.. 16, 339  (1962),

15.   Rosenfeld, R. S., Arch. Biochem. Biophys,,  108,  384  (1964),

16.   Snog-Kjaer, A., Prange,  I.  and Dam, H.,  J.  Gen.  Microbiol.,  14,
     256  (1956).

17.   Coleman, D. L.  and  Bauman,  C.  A.,  Arch.  Biochem.  Biophys.. 72,
     219  (1957).

18.   Rosenfeld,  R.  S.,  et al., J. Biol. Chem., 211,  301  (1954).

19.   Rosenfeld,  R.  S.,  et al.,  J. Biol.Chem..  222,  321 (1956).

20.   Rosenfeld,  R.  S.  and Gallagner, T. F., Steroids, 4,  515 (1964).
                                 109

-------
                            REFERENCES


21.  Rosenfeld,  R.  S,,  Arch Biochem.  Biophys.. 112, 621 (1965).

22.  Eneroth,  P.,  et al.,  J.  Lipid Res..  5, 245 (1964).

23.  Grundy,  S.  M.,  et  al, J.  Lipid Res.. 6, 397 (1965).

24.  Miettinen,  T. A.,  et  al., J.  Lipid Res.. 6, 411 (1965).

25.  Spritz,  N., et  al., J.  Clin.  Inves., 44, 1482 (1965),

26.  "Hormones and Atherosclerosis," Academic Press, 157  (1959).

27.  Rosenfeld,  R.  S. and  Hellman, L., J. Clin. Invest..  38, 1334  (1959)

28.  Fieser,  L.  F. and  Fieser, M., "Steroids," Reinhold Publishing Com-
     pany.

29.  Handbook of Chemistry and Physics. 40th Edition, C-259.

30.  Saad,  H.  Y. and Higuchi, W. I., J. of Pharmaceutical Sciences, 54,
     1205-1206 (1965).

31.  Brown, H. H., et al., Anal. Chem.. 26,397 (1954).

32.  Zak, B.,  et al., Amer.  J. Med. Tech.. 23, 283 (1957).

33.  Gerson,  T., Biochem.  J., 77,  446 (1960).

34.  Jorgensen,  K. H. and  Dam, H., Acta Chemica Scandinavica. 11,  1201-
     1208  (1957).

35.  Vandenheuvel, W. J. A.,  Seeley, C. C. and Horning, E, C., J.  Am.
     Chem.  Soc.. 82, 3481  (I960).

36.  Wells, W. W.  and Makita, M.,  Anal. Biochem., 4, 204  (1962).

37.  Luukkainen, T., et al.,  Biochem. Biophys. Acta.. 52, 599  (1961).

38.  Pomeroy, R. and Wakeman, C. M., Indust. Engng. Chem. Analyt.  Ed.,
     13, 795-801 (1941).

39.  Knechtges,  0.  J.,  et  al., Sewage Wks. Jr., 6, 1082-1093  (1934).

40.  Loehr, R. C.  and Higgins, G.  C., Int. J. Air Wat.  Poll.. 9, 55-57
     (1965).

41.  Mills, P. A., et al., JAOAC. 46,186  (1963).
                                110

-------
                         REFERENCES


42.  Burchfield, H. P., "Drug and Pesticide Analysis by Gas Chromato-
     graphy in Toxicology, Metabolism and Residue Analysis Studies,"
     Presented at Course on Biomedical and Clinical Applications of
     Gas Chromatography, Washington University, Saint Louis, Missouri
     (1969).

43.  Clark, S. J. and Wotiz, H. E., Steroids. 2, 535, (1963).

44.  Wotiz, H. H., et al.t Steroids. 10, 127 (1967).

45.  Exley, D., Biochem. J., 107, 285 (1968).

46.  Turfitt, G. E., J. Bact., 54, 557 (1947).

47.  Arima, K., et al., Agr. Biol. Chem.. 33, 1636-39 (1969).

48.  "Personal Communication,"  Mr. Jerry Brown, Game and Freshwater
     Fish Commission, State of Florida, Gainesville, Florida.

49.  Geldrich, E. E., JWPCF. 34, 295 (1962).

50.  Hull, T. G., "Diseases Transmitted from Animals to Man," Charles
     C. Thomas Publisher, 4th Edition, 640-643  (1955).

51.  Control of Communicable Diseases in Man, llth Edition, Abram S.
     Benenson Editor.  Amer. Pub. Health Assoc. (1970).

52.  Tabak, Henry H. and Bunch, R. L., "Coprostanol, A Positive Marker
     of Domestic and Run-Off Pollution," Missouri River Basin Sterol
     Assay Project Report (1970).

53.  Environmental Protection Agency,'Methods for Chemical Analysis of
     Water and Waste," Training Manual  (1971).

54.  Daniel, Cuthbert and Wood, F.  S., Fitting  Equations  to  Data, Wiley-
     Interscience  (1971).

55.  Hanel, K. K. and Dam, H., Acta Chemical Scandinavia,  9,  677  (1955).

56.  Well, W. W., and Mores,  P.  A., Nature. Lomnd.,  189,  483 (1961).

57.  Gerson, T.,  Biochem. J..  77,  446 (1960).

58.  Smith, L. L.,  et  al.,  Lipids, 3,  301 (1968).

59.  Kirchmer, C.'  J.,  Ph.D.  Dissertation, University of Florida (1961).
                                Ill

-------
                         REFERENCES


 60.  Remieux, R. U. and von-Rudloff, E., Can. J. Chem.. 33, 1701 (1955).

 61.  Von-Rudloff, E., ibid, 34, 1413 (1956).

 62.  Wan, M. E., Zerota, S., J. Org. Chem., 24, 741  (1959).

 63.  Fieser, L. F., JAGS, 75, 4386  (1953).

 64.  Henry, R. A., Schmit, J. A. and Dieckman, J. F., J. Chromatogr.
     Sci., 9, 513 (1971).

 65.  Fitzpatrick, F. A., Siggia, S. and Dingman, J., Anal. Chem.t 44,
     2211 (1972).

 66.  Strain, H. H., JACS. 57, 758  (1935).

 67.  Pippen, E. L., Eyring, E. J.  and Nonaka, M., Anal. Chem., 29, 1305
     (1957).

 68.  The Chromagrography of Steroids," by  I. E.  Bush, Pergamon Press
     (1961).

69.  Reich, H., Sanfilippo, S. J.  and Crane, K.  E., J. Biol. Chem., 198,
     713 (1952).

70.  Janovsky, J. V., et al., Ber.. 19, 2155 (1886) :24, 917 (1891).

71.  Feigl, F., "Spot Test in Organic Analysis," p. 308 and 438, 6th
     Edition, Elsvier Publishing Co.  (1960).

 72.  Strickland, J. D. H. and Parson, T. R., "A  Practical Handbook of
     Sea Water Analysis," Rish. Res. Bd. Can. Bull., 167, 311 pp (1966).

 73.  Hauser, T. R. and Shy, C. M., Env. Sci. & Technol.. 6, 890 (1972).

 74.  Danielsson, H. and Tchen, T.  T., "Steroid Metabolism," p. 132, in
     Metabolic Pathways, Vol. II,  Academic Press, New York and London.
     (1968).

 75.  Danielsson, H., Adv. in Lipid Res., 1,335  (1963).

 76.  Sternberg, J. C., et al., and Parkins, G.,et al., "Gas Chromatogra-
     phy," p. 321 and 269, Ed. N.  Brenner, et al.,  Academic Press  (1962).

 77.  Dobriner, K. and Lieberman, S., "Steroid Hormons," pp. 46-88, Ed.
     E. S. Gordon, The University  of Wisconsin Press  (1950).
                                112

-------
                             APPENDIX
                           A.  Standards

1.  Coprostanol was obtained from Applied Science Laboratories, Inc.
(State College, Pa.),  100.0 mg of coprostanol was dissolved in
100.0 ml of hexane and 2.00 ml pipetted into test tubes and dried.
The working standard was prepared by dissolving the 2.00 mg of copro-
stanol in the tubes in 10.00 ml of hexane, 0.250 ml of which was taken
in another test tube with a micro pipet and dried.  The tubes contain-
ed 50.00 ug of coprostanol.

2.  Cholesterol was obtained from K & K Laboratories, Inc. (Plainview,
N. Y.).  100.0 mg was dissolved in 100.0 ml of absolute methanol (or
ethanol) and 2.00 ml aliquots of the solution were pipetted into test
tubes and dried.  The working standards were prepared by dissolving
2.00 mg of cholesterol in 10.00 ml of hexane, 0.50 ml (or 0.250 ml)
was measured into test tubes by micro pipets.  Each tube contains 100.0
Ug (or 50.0) of cholesterol.

3.  Coprostanone, A^-cholesten-3-one and cholestanol were obtained
from K & K Laboratories, Inc.  The hydrocarbons (n-docosane and n-octa-
cosane) were obtained from Applied Science Laboratories, Inc.

                           B .  Reagent

1.  Sarett reagent:  20 grams of chromic acid anhydride  (0.2 mol) was
dissolved in the mixed solvent of 31,6 grams of pyridine and 150 ml of
glacial acetic acid by adding the chromic acid slowly and by keeping
the temperature below 20° C.  The solution was diluted with glacial
acetic acid to 250 ml.  The  reagent is photosensitive.   It was stored
in a brown bottle.  The working solution was prepared by diluting the
stock solution with an equal volume of glacial acetic acid and filt-
ering with a glass fiber filter using vacuum.  One milliliter of the
working solution contained 40 mg of chromic trioxide  (
2.  Saponification  reagent  for  coprostanol  esters:  0.5N NaOH in abso-
lute methanol.   Two grams of  sodium hydroxide was dissolved in 100 ml
of absolute methanol (analytical  grade)  and stored in a glass bottle
with a screw  cap.   It was kept  out  of  contact with atmospheric carbon
dioxide .

3.  2,4-DNPH  reagent for the  standard  and hydrazonation of coprostanone:
0.1019 gram of  2,4-DNPH crystals  was placed in  a 50 ml beaker, dissolved
in two milliliters  of concentrated  sulfuric acid and 3 ml of distilled
water added.  The mixture was diluted  exactly to 200 ml with carbonyl-
free methanol in a  volumetric flask.   One milliliter of the solution
contained  1.000 yg  of 2,4-DNPH  (as  coprostanol  equivalent).
                                113

-------
                             APPENDIX
4.  2,4-DNPH-sulfate  reagent for preparation of carbonyl-free water
and methanol:  To 0.4 gram of 2,4-DNPH in a 50 ml beaker was added  2
ml of concentrated sulfuric acid.. The 2,4-DNPH was dissolved by  adding
3 ml of distilled water.  The mixture was then diluted with 10 ml of
absolute methanol.

5.  Periodate solution for diazonium formation:  The  stock solution
was prepared by dissolving 2.0 grams of  sodium periodate in 100 ml
of carbonyl-free water.  2,00 ml of the  stock solution was diluted
with 48 ml of the carbonyl-free water, for the working solution.

6.  Coupling reagents:   (a)  a-naphthylamine solution—2.00 grams of
a-naphthylaminehydrochloride was dissolved in one liter of distilled
water to which 1 ml of concentrated hydrochloric acid was added.  It
was stored in a dark bottle.  The working solution was prepared by
diluting the stock solution by 20 fold with carbonyl-free water.  1.00
ml of the working solution was contained 100 yg of a-naphthylaminehy-
drochloride, which was equivalent to about 220 yg of  coprostanol; (b)
N-(l-naphthyl)-ethylenediamine solution—0.50 gram of the dihydrochlo-
ride was dissolved in 500 ml of distilled water.  The solution was
stored in a dark bottle.

                    C.   Solvents and Solutions

1.  Carbonyl-free water:  One milliliter of concentrated sulfuric acid
and 2 ml of 2,4-DNPH sulfate reagent were added to one liter of distill-
ed water, allowed to stand overnight, and then distilled.

2.  Carbonyl-free methanol:  One milliliter of concentrated sulfuric
acid and 2 ml of 2,4-DNPH-sulfate reagent were mixed  with one  liter of
absolute methanol  (analytical grade), refluxed for one hour and  then
distilled.

3.  Hexane, ethanol and  acetonitrile:  Analytical  grade.

4.  Pyridine base  solution:  Pyridine-(sp.  grade)-Glacial acetic  acid-
distilled water  0-5:5:80 by volume).
                                114            *US. GOVERNMENT PRINTING OFFICE: 1974 546-319/426 1-3

-------
  SELECTED WATER
  RESOURCES ABSTRACTS
  INPUT TRANSACTION FORM
                    t. RcpfrtNo.
                                        W
  4. Title ANALYSIS OP COPROSTANOL, AN INDICATOR OF FECAL
    -     CONTAMINATION,
                                         5. Rt ;,ortDtrte

                                         6.
                                         8. ••
 .7. Author(s) singley,J. E., Kirchmer, "C.  J.;  and Mlura, R.
  9. Organization
    Florida University, Gainesville, Department of Environ-
    mental Engineering Sciences.


  12, Spons oriog O rganizs' '"o
                                                                            o.
                                                                  10.  Project &:.,  •

                                                                      16020 EVG
                                        11,
                                        13.  Type of Report
                                            Period Covered
  IS. Supplementary Notes-   •         >                  .'.,
      Environmental Protection-Agency Report No.  EPA 660/2-74-021, March  1974.
  16. Abstract
       Gas chromatogfaphlc (GC) analysis"of coprostanol was improved by showing that
  sample esterificattonbefore injection is not necessary.  The GC method  can detect
  20 ng/1, which was estimated to be equivalent to approximately two conforms per
  100 ml.   Comparison of coprostanol analyses to total and fecal coliform  analyses
  confirmed the predicted advantages of  a chemical method over a biological method.
  Samples were preserved.with concentrated sulfuric acid (1 ml/1 sample).

       The GC method was used in extensive field surveys, degradation studies, and
  treatment plant efficiency studies.  A reasonably good correlation between  coprostanol
  concentration and treatment plant efficiency (BOD, COD, and TOG measurements) was
  found, but further measurements are needed.

       A colbrimetric method was developed to determine coprostanol at a concentration
  of l>ig/l.   Analysis time was estimated to be two hours, but several samples were
  analyzed simultaneously.  Colorlmetric analyses of. coprostanol in field  samples
  gave  higher: results than GC analyses;  high colbrimetric results were probably due
  to the presence of other fecal steroids.
  i?a. Descriptors * Analytical techniques,  * Gas chromatography, * Colorimetry,  * Coliforms,
     Sewage bacteria, Water pollution,  Chemical analysis, Domestic wastes
    b. identifiers  * Coprostanol, Sewage  treatment efficiency, 2,4-dinitrophenylhydrazine,
      Sample preservation !                                  I
  17c. COWRR Field & Group
                        02A.
  IS.  Availability
19. Sf-writy C'ass.
   (Report)

20. Security Class.
   (Page)
21. JV.G. of
   Pbges

22. Price
Send To:
                                                       WATER RESOURCES SCIENTIFIC INFORMATION CENTER
                                                       U.S. DEPARTMENT OF THE INTERIOR
                                                       WASHINGTON. O. C- 2O24O
  Abstractor  A. L. Alford
            I  institution Southeast Environmental Research Lab.
WRSIC 1O2 (REV JUNE

-------