United States
     Environmental Protection
     Agency
                                       Environmental Monitoring and
                                       Laboratory
                                       Cincinnati OH 45268
     Research and Development
                                       EPA-600/S4-82-022  May 1982
v

     Project Summary

         etermination of Benzidines in
     Ylh(te^trial  and  Municipal
            stewaters
x  ,
     Ralph M. Riggin and C. C. Howard
                  www.elsevier
       A
      benzi
      (DCB
      Thisn
      perfor
      with (
      selecti
      compel
      0.1 ppl
      types c
      surface1
      industri
      several
      many
      decrease
      to abou W-ppirroTBenzidine and DCB.
      Apparent levels of 10-20 ppb of
      benzidine were present in the dye
      plant effluents. For one dye plant,
      effluent benzidine was determined to
      be 9 ppb and 12 ppb using  two
      different sets of chromatographic
      conditions, thus supporting the belief
      that  benzidine is present at the level
      stated.
       Precision and accuracy of the
      method were estimated from the
      results  for five wastewater samples
      spiked at levels between 1 and 50 ppb.
      For this group of  samples the re-
      coveries were 69 ± 15% for benzidine
      and 76 ± 9% for DCB.
       Storage of several wastewater
      samples for (two or seven days) at 4° C
      and pH 2 resulted in degradation of the
      compounds in several cases, probably
      due to oxidation or irreversible adsorp-
      tion to  paniculate matter. Therefore,
  -.-Yl
COM I
       to obtain accurate values for
       » and DCB, it is believed to be
       r to assay the sample as soon
       llection as possible. For
        Containing chlorine, a reduc-
        such as sodium thiosulfate
        added, since chlorine was
        rapidly degrade benzidine

        \iect Summary was devel-
             Environmental Mon-
           Support Laboratory,
          OH, to announce key
         'he research project that is
 'uity documented in a separate report
of the same title (see Project Report
ordering information at back).

Introduction
  Under provisions of the Clean Water
Act, the Environmental Protection
Agency is  required to promulgate
guidelines establishing test procedures
for the analysis of pollutants. The Clean
Water Act Amendments of 1977
emphasize the control of toxic pollutants
and declare the 65 "priority" pollutants
and  classes of pollutants to be toxic
under Section 307(a). This report is one
of a series that investigates the ana-
lytical  behavior .of selected  priority
pollutants and suggests a suitable test
procedure of their measurement.
  It presents the results obtained under
Contract No. 68-03-2624, in which a
method was developed for the determi-
nation  of benzidine, dichlorobenzidine,
and  diphenylhydrazine in aqueous

-------
effluents. This work was conducted in
two phases, wherein Phase I consisted
of evaluation of various analytical
methodologies, and Phase II consisted
of validating the most promising method
using several aqueous effluent samples.
Data from the Phase I study, which were
presented  in  a report  dated  May  19,
1978, are also presented herein.
  The objective of this study was to
develop an optimized analytical method
for the  determination of benzidine,
dichlorobenzidine (DCB)  and  1,2-di-
phenylhydrazine (DPH) and to validate
the method  on a variety of aqueous
effluents.
  The successful completion of this
program  involved  the fulfillment of
certain  directives  set forth  in  the
contract by EPA. An extensive literature
review was first conducted to evaluate
the previous work in the area.  Subse-
quent work was directed toward deter-
mination and then full  evaluation of an
appropriate  measurement technique,
which best satisfied the requirements
for sensitivity and selectivity, as well as
the considerations of sample cost, that
is, equipment, time, and training, which
would be needed for the method.  The
stability of the benzidine compounds in
water  miscible solvents and their
instability in  chlorinated and unchlorin-
ated buffered water at different pHs and
storage temperatures were studied over
the prescribed time periods. Extraction
efficiency of two organic solvents was
also studied for the standard compounds.
The remainder of the program involved
the study of the sample preparation and
clean up steps which would be neces-
sary to eliminate sample interferences.
The complete method was then applied
to several representative wastewater
samples and an assessment was made
of the precision and  accuracy of the
complete procedure.

Analytical Methods
Development
  Three analytical  techniques were
evaluated for potential use in  the
determination of benzidines: (1) direct
GC/AFD, (2) GC/ECD following derivati-
zation with a fluoroacy group! and (3)
HPLC/EC.


Direct GC/AFD
  Gas chromatographic  properties of
the three compounds were investigated
using a Hewlett Packard 5730  gas
chromatograph equipped with dual
FID/AFD. Columns were 6ft x 2 mm I.D.
glass, packed with a 3% loading of the
liquid stationary phase on Gas Chrom Q
100/120. The  stationary phases eval-
uated were OV-1, OV-17, SP2250DB,
and OV-225, which represent a wide
range of polarities. GC/FID was used to
evaluate all four stationary phases and
GC/AFD  was evaluated  for the
SP2250DB stationary phase.
  The four stationary phases investi-
gated  represent  a  wide range of
polarities. Both OV-17 and SP2250DB
were found to give satisfactory results.
OV-1 was acceptable but some degree
of tailing was  noted,  especially at low
concentrations. OV-225  gave an un-
acceptable degree of column bleed.
  The use of AFD f or the detection of the
benzidines  was investigated using
SP2250DB.  By operating isothermally
at 230°C, a sensitivity of about 1  ng on
column can  be achieved  for benzidine
and DCB; but  under these conditions,
the DPH peak appears in the solvent
front.
  It was noted for each column that two
peaks arose from the injection of DPH,
and furthermore, that the larger of the
peaks elutes much earlier than would
be expected for a  reasonably polar
compound of this molecular weight. In
order to confirm  the identity  of the
eluted components, GC/MS was used,
employing  a 30 meter  SE-30  glass
capillary column.
  A  single  sharp peak was  observed
eluting  at  210°  from a temperature
programmed run. The mass  spectrum
revealed it  to  be a component with a
molecular ion at  182 amu,  whereas
DPH has a molecular ion at  184. The
component resulting from DPH injection
was suspected to be azobenzene (M.W.
182) which was  confirmed by the
injection of pure  azobenzene.  It  is
apparent, therefore, that DPH instan-
taneously decomposes to azobenzene in
the GC injection port.


GC/ECD
  The  use of various derivatization
reagents to form highly electron captur-
ing derivatives was investigated. The
following derivitization reagents were
evaluated: (1) trifluoroacetic anhydride
(TFA); (2) pentafluoropropionyl anhydride
(PFPA); (3) heptafluorobutyric anhydride
(HFBA); (4) triftuoroacetyl  imidazole
(TFI); and (5) heptafluorobutryl imiazole
(HFBI).
   Numerous problems were encountered
using this approach, the most serious of
which was the production of multiple
peaks and the incomplete reaction of
DPH with  all  of the reagents studied.
Benzidine  reacted well with  the HFBI
but not any of the anhydrides or the TFI.
DCB derivatized the best of the three
amines and failed to form a derivative
only with  TFA. Similar results were
obtained using both the anhydrides and
imidazoles. The DCB-HFBA derivative
was formed with good yield, whereas
there was a relatively poor yield of about
10% of benzidine-HFBA derivative. No
underivatized amine was present in
either case,  so some decomposition
must have occurred in the benzidine
derivatization  reaction. Both benzidine
and DCB were successfully derivatized
using HFBI.  As  stated earlier, HFBI
derivatization of  DPH was  not  suc-
cessful, since no derivative formation
was observed.
  On the  basis of these  results,  this
approach did  not seem useful for the
analysis of benzidines and was  not
further investigated, since an alternate
technique, HPLC/EC, was working
quite well. However,  it appears that
HFBI would   be the best derivatizing
reagent to use for benzidine and DCB, if
one wished to assay these compounds
by GC/ECD.


HPLC with Electrochemical
Detection
  This approach proved useful almost
immediately, so following a short period
of column evaluation, a set of parameters
was selected which  gave excellent
resolution, sensitivity, and reproduc-
ibility. Several reversed phase and ion
exchange  columns were  preliminarily
evaluated, including: Zorbax  ODS  and
CN,  M Bondapak C-18,  Lichrosorb  RP-
18, RP-8,  and RP-2, Spherisorb ODS,
and Whatman Partisil SCX. All the ODS,
CN, and  RP-8 packings gave poor
efficiency  for the  amines—less  than
2000 plates in all cases—although they
gave efficiencies of greater than 10,000
plates for  anthracene,  which is com-
monly  used  for  determining column
efficiency. The Whatman  SCX column
was very poor in that it gave badly tailing
peaks.  The RP-2 column was found to
give excellent efficiency for the amines—
greater than 6000 plates. Better resolu-
tion of the DCB and DPH was achieved
using acetonitrile  instead of  methanol
under the  conditions evaluated.
  Based on  the  results mentioned
above,  a set of HPLC  conditions  was
selected which gave optimum resolution
and sensitivity for the compounds of

-------
interest.  These conditions are listed
below:

  Column—RP-2 5 micron 25 cm x4.6
           mm I.D.
  Mobile phase—50% acetonitrile - 50%
                sodium acetate buffer
                0.1 M pH4.7
  Detector—Electrochemical  (glassy
           carbon) 3  mm  diameter,
           Bioanalytical  Systems
           model LC2
  Detector potential-—0.8V
  Injection volume—25 microliters

  Using the electrochemical detector,
the minimum detectable quantities
were 0.05 ng for benzidine, 0.1 ng for
DCB, and 0.3  ng for DPH at a signal to
noise  ratio  of  five  to one. Linear
response for the components is obtained
from 0.1  - 400 nanograms injected for
DCB and benzidine (a linear dynamic
range of 103).  Above 400 nanograms, a
rapid fall  from linearity is observed for
DCB and benzidine. Precision  of peak
heights for  eight replicate injections of
benzidine was ±2.8% at the 50 nano-
gram level.
  The  electrode potential (0.8V) was
chosen based on our study of response
versus electrode potential for  each of
the compounds. This study shows that
full (diffusion  controlled)  response is
achieved  for  each compound  at 0.8V
versus Ag/Ag Cl and above.

Solvent Stability  Studies
  The stability of the three amines in
water  miscible solvents was  studied
over a 90-day period. For  DCB and
benzidine,  acetonitrile and  methanol
were  used. Both components  were
completely stable over  the  90-day
period. DPH was found to decompose
completely  in three days or  less in all
solvents investigated, including benzene,
methylene chloride,  methanol,  tri-
ethylamine,  acetonitrile, and acetic
acid. It is apparent that DPH standards
must be prepared fresh daily and thus
no further studies were conducted.

Extraction  Studies
  Extraction studies were conducted to
evaluate  the  extraction efficiencies of
methylene  chloride  and chloroform
preserved  with 2%  ethanol  for  the
benzidines in  water at pHs 2,7, and 10.
  Fifty microliters of standard solution
in acetone  was added to 500 ml of
appropriately  buffered water to yield a
concentration of 10 ppb for each of the
compounds in a 1000 ml separatory
funnel.
  The benzidines were extracted with
50 and .then 30 mL of the appropriate
solvent. The extract was washed with
20  mL  water,  and the solvent  was
exchanged to methanol by concentrating
to 5 mL on a rotating evaporator at 35° C.
Finally, the extract was concentrated to
2 mL  on a vortex  evaporator  and
prepared for HPLC by dilution to 4 mL
with 0.1 M sodium acetate buffer.
  The  extraction  methodology was
found to be a very delicate area since
numerous unanticipated problems
developed which led to low recoveries,
especially for benzidine and DPH. The
stability of DPH was found to be a major
problem which  was not completely
solved.
  The  following  are  some  of  the
problems encountered:
  1. Benzidine is substantially ad-
     sorbed on  Na2SC>4 if a  normal
     solvent drying step is employed.
     The use of KaCOs corrected this
     problem. Elimination of the drying
     step was found to be useful when
     using HPLC analysis.
  2. Benzidine  is  heat labile  so  that
     Kuderna-Danish concentration
     techniques  gave low  recoveries.
     The  use  of rotary evaporation
     eliminated this problem.
  3. Decomposition of benzidine  oc-
     curred  when  concentrating
     methylene chloride or chloroform.
     The addition of 15% MeOH prior to
     concentration stabilizes the
     benzidine.
  Generally, chloroform  at pH  7  was
found to give a more efficient extraction,
expecially for benzidine. As expected,
no benzidine  was  recovered at pH 2,
whereas  the  less basic  DCB was
extracted at all  pH values. DPH  de-
graded  readily  in  both aqueous  and
organic  media  and  although both
chloroform and methylene chloride
gave about 70% recoveries at pH 10, no
set of extraction parameters was found
which  gave  greater than about 70%
recovery. At pH 2, DPH was found to
degrade to benzidine to some extent.
Based on these data,  it was concluded
that chloroform is a satisfactory solvent
for the extraction of benzidines.

Storage Stability Studies
  The stability  of  the benzidines in
water  under a variety of  storage
conditions was investigated. Two
temperatures, 4°C and room tempera-
ture, three pH levels, 2, 7, and 10, and
two chlorine levels, 0 and 2 ppm, were
evaluated by preparing duplicate  500
mL samples spiked with 10 ppb of DCB
and benzidine or DPH. DPH  was done
separately since it can degrade  to
benzidine  under certain storage con-
ditions. The solutions  were stored  in
amber glass bottles for 7 days prior to
extraction.
  None of the amines were detectable
in the solutions to which chlorine was
added. A pH of 2 was found to give the
best  results for benzidine and  DCB.
However,  at pH 2  DPH degrades  to
benzidine, thus creating an  undesired
artifact. This problem can be overcome
by employing  pH 4.7 acetate buffer
where both benzidine and DCB are well
preserved, and  DPH degrades to two
unidentified  components,  not  to
benzidine. At pH 4.7, two peaks were
obtained which did  not interfere with
DCB or benzidine, neither of which was
DPH (based on chromatographic reten-
tion times).
  Based  on these results,  it was
concluded that a pH between 2 and 4 is
best for preservation of the benzidines.
Some artifacts due to DPH decomposi-
tion to benzidine are likely at pH 2 and
below if DPH is present in the sample.
Addition of a reducing agent (such  as
NazSOa) to destroy chlorine is required.
  DPH was found to be very unstable in
wastewater, specifically secondary
sewage. DPH,  spiked at the  100 ppb
level, disappeared with  a half life  of
about 15 minutes in the presence  of
oxygen and about 60 minutes with
oxygen removed. This  result indicates
that  DPH  analysis in  wastewater  is
virtually meaningless,  since the DPH
level determined cannot be  directly
related to the DPH in the sample at the
time of collection.

Wastewater Studies

  Based  on the above studies, the
following procedure was applied to the
analysis of actual wastewater samples.
  • Adjust pH of 500 mL to 1 Laliquotof
   wastewater  to pH  7 and serially
   extract with 100,  50 and 50 mL
   volumes of chloroform.
  • Extract  the  benzidines from the
    solvent with three 25 mLaliquotsof
    1M H2S04.
  • Neutralize the acid solution to pH 6-
   7 and serially extract with 30, 20
   and 20 mL volumes of chloroform.
  • Wash the  solvent  with  20 mL
   distilled water and exchange the
   solvent  to  methanol while con-
   centrating  to 5  mL  on a  rotating
   evaporator.

-------
  • Concentrate the methanol extract
    to 1.0 mL with a gentle stream of
    nitrogen.
  • Add 4.0 ml pH 4.7 acetate buffer
    and analyze by HPLC with electro-
    chemical detector.
The  following  water samples were
selected for analysis:
  1.  Surface water (Olentangy River)
  2.  Secondary sewage effluent (Col-
     umbus, Ohio)
  3.  Primary sewage (Columbus, Ohio)
  4.  Final effluent from plant produc-
     ing various organic chemicals
     (such as nitrobenzene, nitrophenols,
     o-dichlorobenzene,  and chloro-
     anilines).
  5.  Final effluent from plant produc-
     ing benzidine based dyes.
  6.  Oxidation process stream from a
     plant producing benzidine based
     dyes.
  Each  wastewater was analyzed,
unspiked, in triplicate. Six 500 mL
aliquots of each  sample  were  spiked
with benzidine  and DCB at a level at
least five times background.  Three
aliquots were analyzed immediately and
three were analyzed after storage, at pH
2 and 4°C, for a period of time—either
two or seven days. The pH 2  level was
attained by addition of 1M H2SC>4. One
gram per liter of sodium thiosulfate was
added to each sample prior to spiking.
The analytical results for all samples are
tabulated in Table 1.
  Sample 2  was a secondary sewage
effluent  spiked at the 4 ppb level.

Table 1.    Data for Wastewater Analyses
Results  obtained were comparable to
those for  the surface  water sample.
Storage  of this effluent, spiked at the 4
ppb, for 48  hours at  4°C, and pH 2,
resulted in  essentially no loss in
recovery. The shorter  storage period
was used, since  it was felt that the
stability  of benzidine in dilute aqueous
solution  was so  unpredictable that
samples should be assayed as soon as
possible. Forty-eight hours appeared to
be the shortest time period which could
reasonably be expected for the comple-
tion of sample collection and transport
to the analytical laboratory.
  Sample 3  was  a primary sewage
sample  and  was  spiked at  the 1  ppb
level. This spike level was  chosen in
order to  validate the procedure at levels
approaching the detection limit, although
as  a  general practice  higher  spiking
levels were used so that accuracy could
be  more easily determined  and poor
recoveries  would still result  in  a
measurable peak. As shown in Table 1,
recoveries were quite good and precision
was about ±10-20%.  However,  after
storage  for seven  days, low recovery
and poor reproducibility were observed
for DCB, but not  for benzidine. This
result appears to indicate that DCB may
be  irreversibly bound to  particulate
matter,  since this  sample had a  large
amount  of suspended matter.
  Sample 4 was a final effluent sample
from a plant producing  various organic
chemicals such as nitrophenols, o-
dichlorobenzene, chloroanilines, and
    nitrobenzenes.  Samples  1-3 were
    selected for this  study  because they
    represent matrices which receive
    inputs from a diversity of industrial
    processes, and thus benzidine determi-
    nation might be applied to such samples
    in the future. Sample 4, on the other
    hand, was selected because it represents
    a matrix suspected of containing various
    aromatic amines (anilines) and phenols
    which would be likely interferences for
    the HPLC/EC method (since they are
    readily oxidized). Sample 4 was spiked
    at the  1  ppb  level. Recovery and
    particularly precision for benzidine
    were poor both  before  and  after a
    storage time of 48 hours. Recovery and
    reproducibility for DCB were better than
    for benzidine, both  before and after
    storage. The reason for this lack of
    reproducibility  for  this sample  is not
    clear, although  it  may  be  due to
    chemical oxidation, since benzidine is
    much more easily oxidized than DCB.
      Sample  5  was obtained from a plant
    producing benzidine  based dyes. This
    sample was about one year old upon
    receipt and had been  stored at pH 2 and
    4°C  during  that  period of time.  The
    sample was diluted 10:1 with distilled
    water prior to analysis. Recoveries were
    very good before storage-but decreased
    after storage.  The drop in benzidine
    level was  particularly disturbing, since
    there was a detectable level of benzidine
    in the unspiked sample, which had been
    stored for  over  a  year.  It seemed
    possible therefore that  the benzidine
Sample Sample
No. Description
Distilled water

1 Surface water

2 Secondary sewage**

3 Primary sewage

4 Final effluent**
for organic chemicals

Benzidine
DCB
Benzidine
DCB
Benzidine
DCB
Benzidine
DCB
Benzidine
DCB
Background Spike
(ppb) (ppb)
<.1 4
<1 4
<./ 4
<.1 4
<./ 4
<./ 4
.1 1.0
.1 1.0
<. 1 1.0
<.1 1.0
Recovery
0 days
70 ± 7*
50 ± 5
63 ± 6
48 ± 5
73 ± 8
63 ± 7
64 ± 5
55 ± 10
47 ±46
89 ± 8
7 days
	
—
—
—
60 ± 12
58 ± 6
71 ±21
19 ± 10
52 ± 17
60 ±21
            plant
     5      Final effluent**             Benzidine
            from plant producing        DCB
            benzidine based dyes
     6      Process effluent**           Benzidine
            from oxidation process       DCB
            in plant producing
 	benzidine based dyes	
  *Mean ± standard deviation.
 **Stored for 48 hours instead of seven days.
                    12 ±3
50
50

10
10
79 ±
78 ±

82 ±
98 ±
14
 7

 3
 5
47 ±
23 ±

75 ±
95 ±
8
7

7
3

-------
background was due to matrix  inter-
ferences, rather than benzidine. The
extract from Sample 5 was injected at
various electrode potentials to see if a
lower electrode potential would  result
in a lower  background for benzidine.
The apparent  level of benzidine in the
unspiked sample  decreased  in  going
from 0.8 to 0.6 volts. This indicates that
an  interfering component,  with an
oxidation  potential greater  than
benzidine, is present. However, it was
still not certain whether the benzidine
level at 0.6 volts was accurate. In order
to check this, a second set  of  HPLC
conditions was used. The value obtained
for benzidine using the original chroma-
tographic conditions (12 ppb) correlated
closely with the value obtained  using
the alternate  colum  (9 ppb),  thus
indicating that the peak probably is
benzidine.
  The  final sample,  sample 6,  run
through the analytical scheme, was an
oxidation process effluent from  the
benzidine based dye plant from which
sample 5 was taken. This sample was
also diluted 10:1  with distilled  water
before  analysis. Due to a  high  back-
ground at  0.8 volts, benzidine was
determined at 0.6 volts. Recoveries for
benzidine and DCB spiked at the 10 ppb
level were  excellent both before  and
after storage for 48 hours at 4°C and pH
2.
  An  estimate of  the precision and
accuracy of the method can be obtained
using  the  results for  all five waste
samples. For these samples, the average
recovery was 69 ±  15% for benzidine
and 76 ± 9%  for DCB.
 and DCB are stable at pH 2 and pH
 4.7.  However, at elevated  pH
 benzidine is more readily degraded,
 due to oxidative reactions.
 Analysis of several wastewater
 samples showed that the analytical
 methods developed herein work
 well for the analysis of benzidine
 and DCB down to the 1 ppb level for
 most samples. Dye plant samples
 are  a special  case since those
 utilized in this program  all showed
 apparently detectable  10-20 ppb
 levels  of benzidine. Confirmed
 measurements on two  HPLC
 columns indicate that benzidine is
 present at the levels stated, although
 more  definitive measurements
 should be made.
Ralph M. Riggin and C. C. Howard are with Battelle Columbus Laboratories,
  Columbus. OH 43201.
James E. Longbottom is the EPA Project Officer (see below).
The complete report, entitled "Determination of Benzidines in Industrial and
  Municipal Waste waters," (Order No. PB82-196 32O; Cost: $10.50, subject to
  change) will be available only from:
        National Technical Information Service
        5285 Port Royal Road
        Springfield, VA 22161
        Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
        Environmental Monitoring and Support Laboratory
        U.S. Environmental Protection Agency
        Cincinnati, OH 45268
Summary and
Recommendations
  The  conclusions drawn from this
study are as follows:
  • DPH is too unstable in both organic
    solvents and aqueous solutions to
    permit meaningful analytical data
    to  be obtained.
  • Benzidine and DCB are completely
    stable for at least 90 days in both
    methanol  and acetonitrile  when
    stored in the dark in sealed ampules.
  • Benzidine and DCB can be efficiently
    extracted from water with chloro-
    form at pH 7 and  pH 10. At pH 2,
    DCB  is  readily  extracted but
    benzidine is not.
  • All three compounds are destroyed
    when stored in aqueous solutions
    containing 2 ppm of chlorine. In the
    absence of chloroform, benzidine
                                                                                        *USGPO: 1982 — 559-092/3406

-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Postage and
Fees Paid
Environmental
Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
                 PS   0000329
                 U S  ENVIk  PROTECTION  AGENCY
                 REGION  5 LIBRARY
                 230  S DFARHORN  STREET
                 CHICAGO IL 60604

-------