EPA-460/3-77-023
December 1977
          HYDROGEN CYANIDE

                 EMISSIONS FROM

                    A THREE-WAY
         CATALYST PROTOTYPE
      U.S. ENVIRONMENTAL PROTECTION AGENCY
          Office of Air and Waste Management
          Mobile Source Air Pollution Control
         Emission Control Technology Division
            Ann Arbor, Michigan 48105

-------
                                        EPA-460/3-77-023
HYDROGEN CYANIDE EMISSIONS

          FROM A THREE-WAY

       CATALYST  PROTOTYPE
                         by

               Eugene L. Holt and Mary H. Keirns

                Exxon Research and Engineering
                  Product Research Division
                      P.O. Box 51
                  Linden, New Jersey 07036


                  Contract No. 68-03-2485

               EPA Project Officer: Robert Garbe


                      Prepared for

            ENVIRONMENTAL PROTECTION AGENCY
               Office of Air and Waste Management
            Office of Mobile Source Air Pollution Control
               Emission Control Technology Division
                  Ann Arbor, Michigan 48105

                     December 1977

-------
This report is issued by the Environmental Protection Agency to report technical data of
interest to a limited number of readers. Copies are available free of charge to Federal
employees, current  contractors and grantees, and nonprofit organizations—in limited
quantities—from the Library Services Office (MD-35), Research Triangle Park, North
Carolina 27711; or, for a fee, from the National Technical Information Service, 5285
Port Royal Road, Springfield, Virginia 22161.
This report  was furnished to the Environmental Protection Agency by Exxon Research
and Engineering, Product Research Division, P.O. Box 51, Linden, New Jersey, 07036, in
fulfillment of Contract No.  68-03-2485. The contents of this report are reproduced
herein as received from Exxon Research and Engineering. The opinions, findings, and
conclusions expressed are those of the author and not necessarily those of the Environ-
mental Protection Agency. Mention of company or product names is not to be considered
as an endorsement by the Environmental Protection Agency.
                        Publication No. EPA-460/3-77-023

-------
                           TABLE OF CONTENTS


                                                            Page No.


  I.  Summary	      1

 II.  Introduction	      4

III.  Experimental Conditions 	      6

      III.l  Vehicle Description	      6

      III. 2  Test Fuel	      6

      III.3  Test Sequence	      6

             III.3.1  Driving Modes 	      6
             III.3.2  Test Design 	      6

      III.4  Three-Way Catalysts	      6

      III.5  Vehicle Malfunctions	      9
             III.5.1  02 Sensor Disconnected	      9
                      III.5.1.1  Air-Fuel Ratio Enrichment.      9
                      III.5.1.2  Air-Fuel Ratio
                                 Properly Set	     10
                      III.5.1.3  Induced Misfire	     10

      III.6  Chemical Analyses	     10

             III.6.1  HCN Analysis	     10
             III.6.2  NH3 Analysis	     12
             III.6.3  Regulated Emissions 	     12
             III.6.4  Sulfate Emissions 	     12
             III.6.5  Fuel Economy	     12
             III.6.6  Bydrogen Sulfide Emissions.  	     12

 IV.  Experimental Results	      13

      IV.1  Catalyst Comparisons Under
            Normal Operating  Conditions  	      13

      IV.2  Catalyst Comparisons Under
            Malfunction Operating Conditions   	      13
            IV. 2.1  HCN	      13
                    IV.2.1.1  Federal Te,st Procedure.  ...      13
                    IV.2.1.2  Congested  Freeway Driving
                              Schedule  (CFDS)  	      15
                    IV.2.1.3  64 kph Cruise	      16
                    IV.2.1.4  80 kph Cruise	      16
                    IV.2.1.5  Idle	      16

-------
                                                            Page No.
            IV.2.2  NH3	     17
                    IV.2.2.1  Federal Test Procedure. ...     17
                    IV.2.2.2  CFDS	     18
                    IV.2.2.3  64 kph Cruise	     19
                    IV.2.2.4  80 kph Cruise	     19
                    IV.2.2.5  .Idle	     20
            IV,2.3  Carbon Monoxide 	     20
                    IV.2.3.1  Federal Test Procedure. ...     20
                    IV.2.3.2  CFDS	     21
                    IV.2.3.3  64 kph Cruise	     21
                    IV.2.3.4  80 kph Cruise	     22
                    IV.2.3.5  Idle	     22
            IV.2.4  Hydrocarbons	     23
                    IV.2.4.1  Federal Test Procedure. ...     23
                    IV.2.4.2  CFDS	     23
                    IV.2.4.3  64 kph Cruise	     23
                    IV.2.4.4  80 kph Cruise	     24
                    IV 2.4.5  Idle	     25
            IV.2.5  Nitrogen  Oxides 	     25
            IV.2.6  Sulfate Emissions 	     28
            IV.2.7  Hydrogen  Sulfide Emissions	     30

  V.  References	     34

Appendix A - Determination of Cyanide in Sodium Hydroxide
             Impinger  Solutions Using an Ion-Selective
             Electrode	     35

Appendix B - Considerations in the Measurement of Cyanide
             Ion Using Ion-Selective Electrodes ......     38

Appendix C - Instruction Manual - Cyanide  Ion Activity
             Electrode Model  94-06.  ..."	     45

Appendix D - Determination of Ammonia in Sulfuric Acid
             Impinger  Solutions Using a Gas-Sensing
             Electrode	     54

Appendix E - Fuel Economy	     56

Appendix F - Procedure for the Determination of Hydrogen
             Sulfide in Automotive Exhaust Using the
             Methylene Blue Method	".....     58
                                   ii

-------
                            LIST OF TABLES


Table No.     	Title	     Page No.

 III-l        Test Fuel Blend	    7

 III-2        Summary of Sixteen Test Program of
              HCN Emissions	    8

 III-3        Steady State Air-Fuel Ratios - 02 Sensor
              Off-Idle Enrichment	    9

 III-4        Steady State Air-Fuel Ratios - 02 Sensor
              Off-Normal Idle Adjustment	   10

  IV-1        FTP Emissions and Fuel Consumption for
              Experimental Catalysts - Normal Vehicle
              Operation	   14

  IV-2        Emission:  HCN mg/km - Cycle:   FTP	   15

  IV-3        Emission:  HCN mg/km - Cycle:   CFDS	   15

  IV-4        Emission:  HCN mg/km - Cycle:   64 kph	   16

  IV-5        Emission:  HCN mg/km - Cycle:   80 kph	   17

  IV-6        Emission:  HCN mg/Min	   17

  IV-7        Emission:  NH3 mg/km - Cycle:   FTP	   18

  IV-8        Emission:  NH3 mg/km - Cycle:   CFDS.	   18

  IV-9        Emission:  NH3 mg/km - Cycle:   64 kph	   19

  IV-10       Emission:  NH3 mg/km - Cycle:   80 kph	   19

  IV-11       Emission:  NH3 mg/Min - Cycle:  Idle	   20

  IV-12       Emission:  CO g/km - Cycle:  FTP	   20

  IV-13       Emission:  CO g/km - Cycle:  CFDS.	   21

  IV-14       Emission:  CO g/km - Cycle:  64 kph	   21

  IV-15       Emission:  CO g/km - Cycle:  80 kph	   22

  IV-16       Emission:  CO g/Min - Cycle:  Idle	   22

  IV-         Emission:  HC g/km - Cycle:  FTP	   23
                                  iii

-------
Table No.     	Title	     Page No.

  IV-18       Emission:  HC g/km - Cycle:  CFDS	    24

  IV-19       Emission:  HC g/km - Cycle:  64 kph	    24

  IV-20       Emission:  HC g/km - Cycle:  80 kph	    25

  IV-21       Emission:  HC g/Min - Cycle:  Idle	    25

  IV-22       Emission:  NOX g/km - Cycle:  FTP	    26

  IV-23       Emission:  NOX g/km - Cycle:  CFDS	    27

  IV-24       Emission:  NOX g/km - Cycle:  64 kph	   27

  IV-25       Emission:  NOX g/km - Cycle:  80 kph	   27

  IV-26       Emission:  NOX g/Min - Cycle:  Idle	   28

  IV-27       Emission:  804™ mg/km -  Cycle:  FTP	   28

  IV-28       Emission:  SQf mg/km -  Cycle:  CFDS	   29

  IV-29       Emission:  SO^ mg/km -  Cycle:  64 kph ....   29

  IV-30       Emission:  SOf mg/km -  Cycle:  80 kph ....   29

  IV-31       Emission:  SOf mg/Min - Cycle:   Idle	   30

  IV-32       Emission:  804* g/km - Cycle:  Average ....   30

  IV-33       Emission:  H2S mg/km - Cycle:  FTP	   31

  IV-34       Emission:  H2S mg/km - Cycle:  CFDS	   31

  IV-35       Emission:  H2S mg/km - Cycle:  64 kph	   31

  IV-36       Emission:  H2S mg/km - Cycle:  80 kph	   32

  IV-37       Emission:  H2S mg/Min -  Cycle:   Idle	   32

    B-l       Measurement  of Cyanide Using an  Orion Cyanide
              Ion-Selective Electrode  and an Orion Single-
              Junction Reference Electrode 	   41

    B-2       Measurement  of Cyanide Using a Metrohm Cyanide
              Ion-Selective Electrode  and an Orion Double-
              Junction Reference Electrode 	   42
                                    iv

-------
Table No.     	Title	     Page No.




  E-l         Emission:  Fuel Lbs. - Cycle:  FTP	    56




  E-2         Emission:  Fuel Lbs. - Cycle: CFDS.	    56




  E-3         Emission:  Fuel Lbs. - Cycle:  Idle	    56




  E-4         Emission:  Fuel Lbs. - Cycle:  64 kph ....    57




  E-5         Emission:  Fuel Lbs. - Cycle:  80 kph ....    57

-------
                            LIST OF FIGURES
Figure No.     	Title	     Page No.

  B-l          Orion Cyanide Electrode/Single Junction
               Reference Electrode	    43

  B-2          Metrohm Cyanide Electrode/Double Junction
               Reference Electrode	    44
                                    vi

-------
                              I.  Summary

          The research described in this report was conducted under
EPA Contract No. 68-03-2485.  It consisted of testing four catalysts,
nominally identical except for Pt/Rh ratios, for their emission levels
of regulated gaseous pollutants as well as the unregulated pollutants
HCN, NH3, S04** and H2S during simulated three-way catalyst vehicle
malfunction modes.  Fuel economy measurements were also made.  The
catalysts were each tested under Federal Test Procedure (FTP) and
Congested Freeway Driving Schedule (CFDS) cycles, 64 and 80 kph
cruises, and at idle.  The simulated malfunction modes were:

          1)  02 sensor disconnected - idle adjustment correct

          2)  02 sensor disconnected - idle adjustment enriched

          3)  13% misfire

          The catalysts were prepared by Engelhard Industries under
conditions similar to those used in manufacturing the three-way cata-
lyst used in Volvo TWC automobiles.   A prototype of this vehicle was
used as the test car.

          Of the fifteen possible test conditions (four catalysts and
no catalyst x three simulated malfunction modes), it was decided to
run only twelve conditions and replicate four of these,  providing a
total of sixteen run sequences,  each sequence consisting, as mentioned,
of the FTP, CFDS, 64 and 80 kph cruise, and idle.  The replicates
allowed calculation of the standard deviation of measurement for the
various pollutants looked at.

HCN
                                                                        *
          Normal operation of the catalysts, under FTP conditions, showed
no HCN tailpipe emissions, at least to the detectability limits for this
cycle of 1.4 rag/km.  However, under malfunction conditions, HCN emissions
were generally observed.  During malfunction, for the cyclic FTP and
CFDS runs, the highest levels were found for the no-catalyst case,
indicating that HCN was produced upstream of the catalyst bed.  It is
not known, however, whether this occurred in the combustion chamber or
in the exhaust system.  In any case, when catalyst beds were present,
tailpipe emissions of HCN were invariably lower, indicating that, at
least under cyclic driving modes, all the catalysts tested were actually
removing HCN from the exhaust gas.

          A different situation was observed under steady state mal-
function test conditions.  The 100% Rh catalyst appeared to produce
HCN in the exhaust gas when the 02 sensor was disconnected, with the
idle adjustment maintained correct.  Under these conditions, the other
catalysts removed HCN, as in the cyclic driving modes.  When the idle

-------
                                 - 2 -
adjustment was enriched, the 100% Rh catalyst produced still more HCN,
and even the 6% Rh catalyst now made HCN.  Under this condition, only
the 100% Pt catalyst continued to remove HCN.

          At misfire, all catalysts removed HCN from the exhaust,
although there seems to be a correlation between catalyst Rh content
and tailpipe emission level.

NH
          Ammonia emission levels were much higher under malfunction
conditions than under normal operating conditions.  Although all four
catalysts showed this trend, the pure Rh gave the highest levels, which
appeared to increase with degree of enrichment.  The highest single
value, 724 rag/km, was observed under misfire conditions, however.  At
idle, NH3 emission levels were quite low under all catalyst-malfunction
combinations.

CO

          Generally, for the two sensor-disconnected malfunction modes
the pure Rh catalyst provided the most control, with the pure Pt worst.
For example, under FTP conditions, the 100% Rh catalyst removed about
50% of the CO, giving levels of 10-24 g/km.  The 100% Pt catalyst also
removed about 50% of the CO when the idle was correctly set, but none
when the idle was enriched.  In no case, however, did the conversion
efficiency approach the levels achieved during normal operation, where
emissions ranged around 2 gm/km.  Under misfire conditions, all four
catalysts were generally equivalent, giving about 40-70% conversion
efficiency.

HC

          Under misfire conditions the pure Pt catalyst was invariably
most efficient for HC removal, and the pure Rh the least.  However,
under the 02 sensor disconnect modes the situation was more complex.
The pure Pt was the worst catalyst under FTP and cruise conditions, but,
for the CFDS was the best catalyst when the idle adjustment was properly
set.  Under idle enrichment, the Rh-containing catalysts were again best,
however.  HC emission levels were generally much higher for malfunction
modes than during normal operation, although the best of the malfunction
runs equalled the normal runs.  At idle little, if any, HC conversion
was achieved, except under the misfire mode, where perhaps 75% control
was seen.

NOX

          Comparing FTP results for the 02~sensor disconnect modes with
the emissions during normal operation, it is seen that pure Rh is still
the most effective catalyst for NOX removal.  However, its activity does

-------
                                 - 3 -
not appear as great, especially when the idle is correctly adjusted.
The Ft catalyst was about the same in activity under normal and 02-
sensor disconnect modes.  The pure Rh catalyst was also the best under
CFDS and cruise conditions, with the pure Pt worst.  At misfire and
idle conditions, NOX tailpipe emissions were extremely low with all
catalysts.
          Sulfate emissions were low under all conditions tested, in
no case exceeding 3.9 mg/km.  For the two sensor disconnect conditions,
the pure Pt catalyst seemed to produce some S0^~, compared to raw engine
exhaust valves, while the pure Rh catalyst removed S04= from the raw
exhaust.  Under misfire conditions the situation was reversed between
the two catalysts.

H2S

          Under normal operating conditions a maximum value of 0.3
mg/km was observed for H2S.  Under malfunction conditions increased
values were observed, although in no case did the level exceed about
8 mg/km, this during misfire.  Under 02 sensor disconnect conditions,
values no higher than about 3 mg/km were seen.  No correlation with
catalyst composition was apparent.

-------
                                   -  4  -
                         II.  Introduction

          In 1975 a report was published by Voorhoeve, et al., showing
that HCN could be formed over a platinum automotive exhaust oxidation
catalyst in the laboratory when the catalyst was exposed to synthetic
mixtures of NO, CO and H2 in He at temperatures from 400 to 800°C.(D
It was also found that the presence of small amounts of H20 or S02 se-
verely suppressed HCN formation.  Because of the presence of NO, CO and
H£ in automotive exhaust and the similarity between the laboratory
catalytic and temperature conditions with those observed in vehicles,
along with the high toxicity level of HCN, several studies have been
carried out to confirm the initial laboratory results and to deter-
mine if HCN can be formed in real automotive exhaust gases.  In some
unpublished work done at Exxon Research and Engineering Company a
laboratory mixture of CO, NO, C3Hg and H2, forced over a PTX-IIB
automotive catalyst at 700°C, yielded a similar level of HCN to that
reported by Voorhoeve.  However, when gas mixtures simulating lean
and rich exhaust gases were passed over the same catalyst, the former
yielded no detectible HCN and the latter only around 1 ppm, compared
to levels of up to 700 ppm observed with the less realistic gas mix-
tures.  To extend this work to actual exhaust gases, a 1975 car,
equipped with a 350 C.I.D. engine and a Pt-Pd pelleted catalyst, was
run under FTP and 64 kph cruise conditions with the catalyst fresh
and aged and under normally lean as well as rich carburetion.  In
no case did the HCN emissions exceed 3 mg/km, or the equivalent
of about 1.5 ppm in the exhaust gas.

          Shortly after this work was completed, indicating that HCN
formation should not be a problem with conventional oxidation catalyst
vehicles, even under abnormally rich operating conditions, the U.S.
Environmental Protection Agency obtained some three-way catalyst
vehicles.  These were manufactured by Volvo and had previously been
used for certification testing.  Bradow and Stump of the EPA found that
under normal stoichiometric  air-fuel ratio, and lean malfunction oper-
ating conditions, no  detectable HCN was formed.(2)  However, when the
system was adjusted to operate rich, easily detectable amounts of HCN
were formed over the  catalyst.  At the request of the EPA, Exxon repeated
these types of experiments, using similar, and, in one case an iden-
tical vehicle, to those used by the EPA.  Our results basically
confirmed those obtained by the EPA.

          This raised the question of why, when exposed to rich exhaust
gas, the three-way catalyst  system produced HCN.  In contrast, earlier
Exxon work with a conventional oxidation catalyst vehicle, also exposed
to a rich exhaust gas, did not find such an effect.  A possible answer
was catalyst composition, and part of this mystery was cleared up by
additional EPA studies.  Bradow and Stump tested a number of catalysts
on an engine dynamometer stand, with the carburetor adjusted to yield
a rich exhaust gas.  It was  found that rhodium-containing catalysts
tended to give significantly higher levels of HCN than did Pt or Pt-Pd
catalysts.  Since the Volvo  catalysts contained RH, and the catalyst

-------
                                 - 5 -
used by Exxon did not, these results appear to explain the discrepancy.
They support also some earlier laboratory work at Imperial Chemical
Industries, Ltd., which reported Rh to be a much more effective catalyst
for the reduction of NOx than is Pt, and further showed the reduction to
be less selective to N2 than in the case of Pt, at least at temperatures
below 600°C(3)

          Because of the implication of Rh and of rich operation in
HCN formation, Exxon Research and Engineering undertook this program
to study more closely the effects of catalyst Rh content and vehicle
malfunction mode on the rate of HCN production from a three-way system
vehicle.  Also, because of the report of Bradow and Stump that high
speed driving increased HCN formation,  the effects of several different
cyclic and steady-speed driving modes were investigated.

-------
                                  -  6 -




                   III.  Experimental Conditions

III.l  Vehicle Description

          The car used in this program was a prototype of the 1977
three-way system vehicle introduced by Volvo in California.  It was
loaned to us by Engelhard Industries Division, where it had been used
for fleet testing.  The car had accumulated approximately 27,000 km
prior to this program.  With the exception of routine maintenance
and of catalyst changes, to be discussed in Section III.4, it was
used as received.

III.2  Test Fuel

          The test fuel contained 305 ppm S, <0.0025 g/1 of Pb and
approximately 0.25 mg/1 of P-  The fuel was blended from unleaded
components and the desired S content obtained by the addition of
equal S-bearing amounts of thiophene and di-t-butyl disulfide.  A
complete Inspection report of  the fuel is given  in Table  III-l.

III.3  Test  Sequence

     III.3.1 Driving Modes

          During  each of  the catalyst-system condition combinations,
the vehicle  was driven over the  Federal Test Procedure, the  CFDS, and
one-half hour periods of  64 and  80 kph cruises and of  idle.   During each
driving_mode measurements of emissions of HC, CO, NOX, HCN,  NH3,  H2S,
and S04~ were made.   In addition, fuel economy was also determined.

     III.3.2 Test Design

          Instead of  running each of the possible fifteen  combinations
of  catalyst  and system (four catalysts and no catalyst times  three
system conditions) it was felt to be more useful to omit several
and replace  them  with some replicates of the remaining combinations.
Accordingly, twelve of the possible  combinations were  actually run
and four were replicated, making a total of  sixteen test sequences.
The final arrangement is  shown in Table III-2.

III.4  Three-Way  Catalysts

          The four catalysts in  this program were prepared by Engelhard
Industries.  Although the details of catalyst preparation  and compo-
sition are proprietary to Engelhard, they were requested  to  make  the
catalysts in a manner similar  to that used  in producing the  catalyst
used  commercially by  Volvo in  their  1977 three-way system vehicles
being sold in California.  All four  catalysts were similar in such
properties as substrate,  geometry, and total noble metal  content.

-------
                             - 7 -
                      TABLE III-l
                    TEST FUEL BLEND
Research Octane, ASTM

Motor Octane, ASTM

Gravity, API @ 60°F

Breakdown Time, Minutes

Sulfur, ppm

Phosphorous, mg/100 ml

Lead, g/gal

FIA, Vol. %

     Aromatics
     Saturates
     Olefins

RVP, kPa

Evap. @ °C

     38

     57

    101

    173

Initial Boiling Point, °C

Final Boiling Point, °C
 93.5

 84.0

 61.7

960+

305

  0.025

 <0.01
 24.5
 63.1
 12.4

 78.6 (11.4 psi)
  5%

 20%

 50%

 90%

 28

210

-------
                                           TABLE III-2
                        SUMMARY OF SIXTEEN TEST PROGRAM OF HCN EMISSIONS
Catalyst/

Operating
   Mode
100% Rh
Pt-17% Rh
Pt-5% Rh
100% Pt
None
02 sensor
disconnected

On sensor
disconnected
A/F enriched

Misfire
  XX*
  X
                      X
                      XX
                                    X
                                    XX
                     XX
                                        X
                                                                                       00
                                                                                       I
*  Double X indicates replicate.

   Each catalyst was broken in under normal operating conditions for one day using the
   modified AMA cycle.  Following this, FTP emissions for each catalyst were measured,
   again under normal operating conditions.

-------
                                   - 9 -
They differed only in noble metal composition.  One  catalyst  contained
Pt as the only noble metal component, another contained  only  Rh.   The
other two had Pt/Rh ratios of 83/17 and 94/6.  An empty  catalyst
canister was used for the "no catalyst" runs.

III.5  Vehicle Malfunctions

          Since the primary purpose of this program was  to determine
the HCN and NH3 emissions of three-way system vehicles under  abnormal
operating conditions, three failure modes, considered typical of
those which might occur in consumer use, were induced in the  test  car.
Two of these involved 02 sensor failure, with and without enrichment
of the idle adjustment.  The third was a random misfire  simulating
partial ignition failure.

     III.5.1  Op Sensor Disconnected

          III.5.1.1  Air-Fuel Ratio Enrichment

          In this operating mode the 02 sensor was disconnected, to
simulate its total failure.  In addition, the idle adjustment of the
fuel injection system was set richer than normal.  This provided
substantial deviations from stoichiometry, as demonstrated by steady
state measurements, shown in Table III-3.
                            Table III-3

                   Steady State Air-Fuel Ratios

                   QO Sensor Off-Idle Enrichment


     Speed, km/h          Average A/F Ratio*            Range

        idle                     13.0                 12.7-13.2

         64                      13.8                 13.7-13.9

         80                      13.8                 13.2-14.1
     *  Average of 4 runs.
It can be seen that in all cases the air-fuel ratios were substantially
lower than the stoichiometric value of about 14.6.  Although air-fuel
ratios were not measured during cyclic operation, it is likely that
their values remained below stoichiometric during these driving
modes as well.

-------
                                 -  10 -
          III.5.1.2  Air-Fuel Ratio Properly Set

          In this malfunction mode the idle adjustment was set accord-
ing to specifications.  Thus the only vehicle malfunction was the
sensor disconnection.  It can be seen from Table III-4 that under
these conditions the system more closely approached stoichiometric
conditions than when the idle was misadjusted.  However, air-fuel
values were still rich.
                            Table III-4

                   Steady State Air-Fuel Ratios

               QO Sensor Off-Normal Idle Adjustment


     Speed, km/h          Average A/F Ratio*            Range

        idle                     14.1                 14.0-14.2

         64                      14.4                 14.3-14.5

         80                      14.4                 14.3-14.5


     *  Average of 4  runs.


           III.5.1.3   Induced Misfire

           The third malfunction mode studied  in  this  program was
partial misfire.  For this mode, the idle  adjustment  and  Q£  sensor
were operated normally, but a  random misfire  generator, developed
at Exxon Research and Engineering Company, was used to randomly
prevent plug firings, thereby  simulating a poorly operating  ignition
system.  Approximately a 13% misfire rate  was used.

III.6  Chemical Analyses

     III.6.1  HCN Analysis

           The HCN content of the exhaust gas  was determined  by drawing
a sidestream of the CVS diluted gas through a bubbler train.   The flow
rate was approximately 2 1/min.  Generally three bubblers are  used
in series.  The first contains dilute ^804,  and is used  for NH3
sampling,  to be described in Section III.6.2. The last two  contain
80 ml. each of a solution made up of 50 ml. of IN NaOH in deionized
water and  30 ml. of deionized  H20.  The second of these bubblers,
third in the train, is used as a precaution in the unlikely  event
that HCN is not completely tra'pped in the  first  NaOH  bubbler.  Other

-------
                                - 11 -
precautions included the use of glass sampling lines wherever possible
Short lengths of stainless steel are acceptable, but plastics, including
Teflon, were avoided.  Since water condensation may cause sampling
losses, heated lines should be used if the gas stream is water saturated.
Finally, I^S may interfere with specific ion electrode detection of CN~.
Therefore, if H2S is suspected, the sample solution should be treated
with Cd(OH)2 to precipitate sulfide prior to electrode analysis.

          After gas sampling is completed the NaOH bubblers are worked
up for analysis by emptying each one into a 100 ml. volumetric flask
and adding deionized rinse water from the bubbler to make up 100 ml.
total volume.  A 20 ml. aliquot is pipetted from each solution in sep-
arate plastic containers.  To each is added 1 ml. of ION NaOH solution.
Potential readings are then taken with a cyanide specific ion electrode
according to the manufacturers instructions.   If sulfide is suspected
in the solution, add about 30 ml. of Cd(OH)2 or Cd2CC»3 solution, stir and
filter before taking readings.

          The cyanide solutions should be stirred at a constant rate
during the potential reading, and values should be taken only after
the potential stabilizes.  The cyanide concentration of each scrubber
solution is determined by comparing the potential reading with a cali-
bration curve, made up as described in Appendix A.  This appendix also
contains further details of the cyanide analysis procedure.  Appendix B
presents the results of a brief study by George Milliman of the Analytical
and Information Division of ERE on the effects of varying amounts of C02
passage through the bubbler solutions, and our method of counteracting
any such effects, the responses of single and double junction reference
electrodes, and a comparison of two brands of reference electrodes.  Finally,
Appendix C is an operating manual for one such specific CN~ ion electrode,
offering a theoretical discussion of factors  affecting electrode operation.

          It should be noted that the first bubbler, containing dilute
H2S04» *s also capable of trapping small quantities of HCN along with
the intended NH3.  Therefore the solution from this impinger must also
be analyzed for HCN and the results added to  those from the two NaOH
bubblers.  The solution is worked up as described in Section III.6.2
and a 20 ml. aliquot taken.  It is made strongly basic with 1 ml. of
concentrated NaOH and analyzed in the same manner as the solutions
from bubblers 2 and 3.

          The limits of detectability for HCN using the above procedures
was based on the lower limit of 0.04 Ug CN~/ml  capable of detection
in the original samples.  With exhaust gas flow rates of 2 1/min and
80 ml of solution in each scrubber, a lower limit of 1.4 mg/km was obtained
over the FTP.  The 64 kph and 80 kph cruises  gave 0.8 and 0.6 mg/km
respectively for gas sampling periods of one-half hour.

-------
                                - 12 -
     III. 6. 2  NH3 Analysis

          The solution used in the impingers for NH3 analysis is prepared
by adding 1.1 ml of concentrated H2S04 to a 1 litre volumetric flask and
filling to the mark with deionized H20.  The impinger is filled with
80 ml of this solution and after sampling, at 2 1/min, an additional 20
ml is used for rinsing and added to the first 80 ml in a volumetric
flask to make a total of 100 ml of sample solution.  Analysis with a
gas sensitive NH3 electrode is carried out on 20 ml aliquots of this
sample according to the electrode manufacturer's instructions.  Further
details of this procedure are found in Appendix D.  The limits of
detectability of NH3 were 6.5 mg/km over the FTP and 2 and 1.5 mg/km
for the one-half hour cruises at 64 and 80 kph respectively.

     III. 6. 3  Regulated Emissions

          Determination of HC, CO, and NOX emissions were carried out
in the standard manner, using CVS bag samples and instrumental analyses.
These included flame ionization detection for HC, non-dispersive IR
for CO and chemilumine s c en ce for NOx.  Details of the CVS unit can be
found in a recent report prepared for the EPA by Exxon Research and
Engineering Company ^ .

     III. 6. 4  Sulf ate Emis s ion s

          Sulfate emissions were measured simultaneously with the H2S,
HCN, NH3 and regulated emissions described in Sections III. 6. 1-3,
using the Exxon exhaust particulate sampling system.  Details of this
system and of the analytical procedure used to measure sulfate collected
on the sampling system filters may be found in Reference 4.  The limits
of detectability were 0.4 mg/km for the FTP and 0.3 mg/km for the other
driving modes.

     III. 6. 5  Fuel Economy

          Fuel economy was determined by the carbon balance method,
with the results checked by direct weighing of fuel before and after
each run.  A high degree of consistency was found both internally, in
comparing successive runs where only the catalyst was changed, and
between the carbon balance and weight methods.  The fuel weight con-
sumptions for each run and driving mode are given in Appendix E.

     III. 6. 6  Hydrogen Sulfide Emissions
              was determined by passing a known amount of exhaust gas,
generally at 1 1/min, but on occasion at 2 1/min, through scrubbers
containing a buffered zinc acetate solution.  After sampling, solutions
of formaldehyde, N,N-dimethyl-p-phenylene diamine sulfate in H2S04, and
ferric ammonium sulfate were added.  Absorbance was measured at 667 nm
and in 5 mm cells vs water.  Total yg of H2S were then determined by
reference to a calibration curve.  A more detailed discussion is found
in Appendix F.  Under FTP conditions, the detectability limit was 0.07
mg/km.  For the CFDS it was 0.06 or  0.11 mg/km, depending on the exhaust
gas flow rate.  At  cruise for  one-half hour it was about 0.04 mg/km.

-------
                                   - 13 -
                     IV.  Experimental Results

IV.1  Catalyst Comparisons Under
      Normal Operating Conditions

          Prior to starting the test matrix comparing  the various
catalysts under three malfunction modes, the four  catalyst  formulations
were each run on the Federal Test Procedure with the test vehicle
set up for normal, stoichiometric operation.  This served to establish
a base case against which to compare the effects of the malfunctions
and to determine the levels of the unregulated emissions HCN and NHg
which could be expected from properly operating three-way systems.
It also provides a measure of catalyst composition effects  under
normal operating conditions.

          Table IV-1 gives the regulated and unregulated emissions,
as well as fuel consumption, for these runs.  No trend with catalyst
composition is evident for HC, CO, HCN or S0^= emissions, all of
which are very low.  Ammonia emissions are also low, especially
compared to values observed in malfunction tests, as will be indicated
in later sections.  The pure Rh and Pt catalysts gave the highest
values, but in view of the overall low levels it is not clear if this
is significant.  The most interesting effect is for NOX emissions,
which correlate inversely with Rh content.  This bears out  the generally
held position that Rh is an important contributor to three-way catalyst
activity, especially for NOX reduction.

          Comparison of the fuel consumption figures, which are of
course independent of catalyst composition, shows that the  range of
values is. only about 3%, verifying the reliability of our carbon
balance determinations.

IV.2  Catalyst Comparisons Under
      Malfunction Operating Conditions

     IV.2.1  HCN

          IV.2.1.1  Federal Test Procedure

          Table IV-2 shows the HCN results obtained for the 16 FTP
runs made at the conditions of varying operating mode and catalyst
composition.  As discussed in Section III, only 12 of the possible
15 combinations were tested, with four replicates, for a total of 16
runs.  The replicates from the FTP, CFDS, 64 kph, and 80 kph tests
were combined to provide an estimate of the overall HCN test standard
deviation.   This value was 2.9.

-------
TABLE IV-1
FTP

EMISSIONS AND

Relative Noble Metal Amounts

100% Rh
17% Rh/83% Pt
6% Rh/94% Pt
100% Pt
HC
0.12
0.15
0.13
0.13
FUEL CONSUMPTION FOR EXPERIMENTAL CATALYSTS
Normal Vehicle Operation
Emissions
g/km
CO NOX HCN
1.7 0.21 <1.4
2.1 0.30 <1.4
2.0 0.41 <1.4
1.9 0.63 <1.4


NHq
11.7
<6.5
<6.5
15.5
'
mg/km
s°4" H2S
0.4 <0.1
0.6 <0.1
0.6 0.3
0.2 0.2
Fuel Consumption,
Ibs/test

3.80
3.68
3.74
3.69

-------
                                  - 15 -
                            Table IV-2
             Emission:  HCN mg/km
                                           Cycle:  FTP
         Catalyst/   100% Rh     17% Rh     6% Rh       0% Rh        No
Malfunction            0% Pt     83% Pt    94% Pt     100% Pt     Catalyst
°2 Sensor
Idle Correct  V        2.4

02 Sensor Off          s 9
Idle Enrichened J       O*

13% Misfire            2.0
            2.5
            1.4
                     1.7
                     3.0
                     3.6
                                                        1.7

                                                        2 ,


                                                       <1.4
                                                                    9.5

                                                                    5.5
                                                                    8.5

                                                                    8.0
          The most striking feature shown in this table are the
higher emissions of HCN for the runs made with no catalyst.  It
appears that all of the catalysts tested, at all three operating
modes, actually removed HCN from the exhaust gas.  As between the
four catalysts themselves, tailpipe emissions appeared similar,
except that the all rhodium catalyst showed a higher value for the
idle enrichened case.  It should be noted also that in nearly every
case the FTP emissions of HCN were higher than those measured during
normal operation, discussed in Section IV. 1.  It is not known whether
HCN is formed upstream of the catalyst in the combustion chamber or
exhaust system.

           IV. 2. 1.2  Congested Freeway Driving Schedule (CFDS)

          As in the case of the FTP, the catalysts showed a tendency
to remove HCN from the engine out exhaust, as evidenced by the higher
values for the no-catalyst case shown in Table IV-3.   As between the
catalysts, again the all rhodium system was highest,  but still did
not appear to be a net producer of HCN during this test cycle.
                            Table IV-3
             Emission:  HCN mg/km
                                           Cycle:  CFDS
         Catalyst/   100% Rh     17% Rh     6% Rh       0% Rh
Malfunction            0% Pt     83% Pt    94% Pt     100% Pt
                                                                     No
                                                                  Catalyst
02 Sensor Off
Idle Correct

02 Sensor Off
Idle Enrichened

13% Misfire
4.2
4.6

7.4

5.2
                                  3.3
                                  2.3
                                            lt8

                                            1.1
                                            2.9
                                                        1.1


                                                        2.0

                                                        2.1
8.1

7.7
2.6

7.5

-------
                                 - 16 -
          IV.2.1.3  64 KPH Cruise

          The results from this cruise mode are more complex than those
for the FTP and CFDS cycles discussed in the previous two sections.
With the sensor disconnected but the A/F ratio idle adjustment
properly set the pure Rh catalyst produced HCN beyond that already
present in the raw exhaust, while the low Rh (6%) and pure Pt catalysts
removed it.  As the exhaust became richer, with the idle A/F adjust-
ment enriched, the pure Rh catalyst produced still more HCN and even
the 6% Rh catalyst made some HCN.  In this case only the pure Pt
catalyst appears to remove HCN,.  .Under misfire conditions all three
catalysts removed most of the HCN from the exhaust.  These results
are shown in Table IV-4.
                             Table IV-4
              Emission:   HCN mg/km
          Cycle:  64 kph
          Catalyst/   100% Rh
 Malfunction            0% Pt
17% Rh     6% Rh       0% Rh        No

83% Pt    94% Pt     100% Pt     Catalyst
02 Sensor Off
Idle Correct
             tf

Oo Sensor Off  A
Idle EnrichenedJ

13% Misfire
                        7.2
                       16.2


                       39.8

                        1.4
 0.6
 0.8
           2.3

           9.2
           8.3
 1.2


 2.3

<0.8
5.0

3.1
4.1

4.9
           IV.2.1.4  80 KPH Cruise

           This cruise condition resembled the 64 kph cruise in that
 the pure Rh catalyst produced HCN with the 02 sensor disconnected,
 especially under the enriched mode.  In the latter mode the 6% Rh
 catalyst also made some HCN.  All other combinations of catalyst-
 driving mode removed HCN from the exhaust.  These results are shown
 in Table IV-5.
           IV.2.1.5  Idle

           Under idle conditions no significant quantities  of  HCN are
 produced,  either in the raw exhaust or over'any of the catalysts at
 any malfunction mode, as shown in Table IV-6.

-------
                                 - 17 -
                             Table  IV-5
              Emission:  HCN mg/km
                      Cycle:  80 kph
         Catalyst/   100% Rh     17% Rh     6% Rh
Malfunction            0% Pt     83% Pt    94% Pt
                                                      100% Pt
                                                No
                                             Catalyst
02 Sensor Off
Idle Correct

02 Sensor Off
Idle Enrichened

13% Misfire
                       5.8
                      17.2

                      27.3

                       3.8
              4.6
              1.1
                       3.8

                      11.7
                      11.8
                                                        1.4


                                                        5.0

                                                        1.3
                          8.4

                          6.2
                          7.1

                         10.0
                            Table IV-6
             Emission:  HCN mg/min
                      Cycle:   Idle
         Catalyst/
Malfunction
02 Sensor Off
Idle Correct
02 Sensor Off   \
Idle EnrichenedJ

13% Misfire
       IV.2.2  NH-
100% Rh
  0% Pt
                                 17% Rh
                                 83% Pt
  0.6

 <0.9

 <0.8
                                <0.8
                                <0.8
 6% Rh
94% Pt


<0.6

 0.8
                                                        0% Rh
                                                      100% Pt


                                                        0.6


                                                        0.9

                                                      <0.8
   No

Catalyst


  0.8


  0.8
 <1.1
     +
  0.6
            IV.2.2.1  Federal Test Procedure

            As shown in Table IV-7, NH3 levels are generally much higher
  than for the normal operating case described in Section IV.1.  This
  appears to apply even to the raw exhaust, although the values do overlap
  within one standard deviation, and it is not known if the catalysts
  under stbichiometric operation removed NH3 from the raw exhaust, or
  if the engine itself under these malfunction modes produced more NH3
  than the same engine at stoichiometric conditions.  Combining the
  replicated runs from the FTP, CFDS,.and 64 and 80 kph cruise modes
  gave a standard .deviation of 32.

            The four catalysts under the malfunction modes all increased
  the NHs level in the exhaust, again however, a few of the values
  overlapped by one standard deviation.  Maximum production was shown by
  the all Rh catalyst, under enriched and misfire conditions.  No catalyst
  composition effect was evident for the correct idle-sensor disconnected
  case.

-------
                                 -  18 -
                            Table IV-7
             Emission:  NHg mg/km
                     Cycle:   FTP
         Catalyst/   100% Rh     17% Rh      6% Rh       0% Rh        No

Malfunction            0% Ft     83% Ft     94% Ft     100% Ft     Catalyst
02 Sensor Off
Idle Correct

02 Sensor Off
Idle EnrichenedJ

13% Misfire
                       103
                       100


                       198


                       214
             113
              68
                        105


                         68
                        134
                         95


                        142


                         63
                           50


                            7
                           20

                           24
          IV.2.2.2  Congested Freeway Driving  Schedule

          This cycle showed much higher levels of catalyst NEU
production than did the FTP, even though the raw exhaust level was
lower.  These results are given in Table IV-8.  Again the pure Rh
catalyst gave the highest levels, during the enriched and misfire
operating modes.  Also following the FTP pattern was the lack of a
catalyst composition effect for the idle correct-sensor disconnected
case.
                            Table IV-8
             Emission:  NH~ mg/km
                     Cycle:  CFDS
          Catalyst/

Malfunction	


02 Sensor Off  I
Idle  Correct  J

02 Sensor Off  1
Idle  EnrichenedJ

13% Misfire
100% Rh

  0% Ft


  165
  158


  279


  724
17% Rh

83% Ft
 6% Rh

94% Ft


 158


 157
 239
 195
 165
                                                        0% Rh

                                                      100% Ft


                                                        188


                                                        189


                                                        103
   No

Catalyst


   11


    7
   17

-------
                                 -  19 -
     IV.  IV.2.2.3  64 kph Cruise

          For this cruise test mode the raw exhaust NH  levels were
lower than observed for the FTP and CFDS cycles, as shown in Table IV-9.
Once again the pure Rh catalyst showed the highest levels, similar
to those for the FTP, at the enriched and misfire operating conditions.
Also, as before, no catalyst composition effect was seen for the
idle correct-sensor disconnected case.  One difference from the FTP
and SET results was a correlation between Rh content and Nl^ production
for the enriched case, with pure Rh highest and pure Pt lowest.
                              Table IV-9

             Emission:   NHs mg/km          Cycle:  64 kph


         Catalyst/   100% Rh     17% Rh     6% Rh       0% Rh        No

Malfunction	      0% Pt     83% Pt    94% Pt     100% Pt     Catalyst
0, Sensor Of f 1
Idle Correct  J
           •1 OQ

           193                  17°         15A
02 Sensor Off   I      o-in                  154          s<;          <2
Idle EnrichenedJ                           154                       2

13% Misfire            195         99                    64           5
                                   86
          IV. 2. 2. 4  80 kph Cruise

          This test mode appeared very similar to the 64 kph case.
Unfortunately one data point is missing due to a mechanical problem,
but if we assume this value to follow the pattern, then the same
conclusions stated for the 64 kph case hold here, although the overall
production level is higher.  These results are shown in Table IV- 10.

                              Table IV-10

             Emission:  NH3 mg/km          Cycle:  80 kph


         Catalyst/   100% Rh     17% Rh      6% Rh       0% Rh        No
Malfunction            0% Pt     83% Pt     94% Pt     100% Pt     Catalyst
f 1
 J
02 Sensor Off 1      '   431                   33g         2go
Idle Correct           309
02 Sensor Off  ^      ,,-                   245         ....           5
Idle EnrichenedJ                            298         ^           2
13% Misfire            ---                               111

-------
                                - 20 -
          IV.2.2.5  Idle

          In all cases the NH3 values were low, about the levels of the
stoichiometric case in Section IV.1, with the catalysts showing no
increase over the raw exhaust.  As in the HCN case, idle production
of this unregulated emission is too low to allow further comparisons.
These results are found in Table IV-11.

                              Table IV-11

              Emission:  NH3 mg/Min          Cycle:  Idle


         Catalyst/    100% Rh     17% Rh      6% Rh       0% Rh        No

Malfunction	       0% Ft     83% Ft     84% Ft     100% Ft     Catalyst

°2 Sensor Off *\         7.2
Idle Correct  J         6.6                  10'1         8'8         9<9

02 Sensor Off  1                             6.3                      2.7
Idle Enrichened J                             5.6                    <2.0

13% Misfire             6.3        ?"?                    5.8         4.9
     IV.2.3  Carbon Monoxide

          IV.2.3.1  Federal Test Frocedure

          For the two cases with disconnected sensors similar patterns
are found, as shown in Table IV-12.  The pure Rh catalyst provides the
greatest degree of CO conversion, the Pt/Rh catalysts are next and the
pure Ft catalyst the lowest.  In the enriched case the pure Ft catalyst
shows no conversion activity at all, since CO emissions are equal to the
no catalyst case.  For the misfire mode, all three catalysts provide
approximately 60% conversion.  The  standard deviation calculated  from
the replicated runs for the FTP, CFDS,  and cruise modes  is  2.3.
                              Table IV-12

                Emission:  CO g/km          Cycle:  FTP


         Catalyst/    100% Rh     17% Rh      6% Rh       0% Rh        No

Malfunction             0% Ft     83% Ft     94% Ft     100% Ft     Catalyst

°2 Sensor Off A        10.3                      fi
Idle Correct  J         9.6                              L '*         a"1

02 Sensor Off  *\      ,                      32.2                    37.0
Idle Enrichened J      "'y                   24.2       J8*1         39.0

13% Misfire             9.6         '                     9.3         22.7

-------
                                 - 21 -
          IV.2.3.2  Congested Freeway Driving Schedule

          No clear pattern is discernible for this cycle.  The only con-
clusions apparent are that the Pt catalyst in the sensor-off, enriched
mode provides very little CO conversion and that all three catalysts
provide about 50% conversion during misfire.  These results are given
in Table IV-13.

                              Table IV-13
         Catalyst /

Malfunction

02 Sensor Off
Idle Correct

02 Sensor Off
Idle Enrichened

13% Misfire
.on: CO g/km
100% Rh 17% Rh
0% Pt 83% Pt
6.7
7.2
13.1
Cycle :
6% Rh
94% Pt
5.8
13.1
10.3
CFDS
0% Rh
100% Pt
5.1
17.5

No
Catalyst
12.8
23.4
19.1
                       10.,

11.2
22.2
          IV.2.3.3  64 kph Cruise

          With the sensor off but idle A/F control properly adjusted all
three catalysts were equivalent, as shown in Table IV-14.  However, when
the A/F control was enriched, the pure Pt catalyst gave no CO control,
while the two Rh containing catalysts were partially effective.  Under
misfire conditions the three catalysts were equivalent to each other,  *
at about the 70% level.

                              Table IV-14
         Catalyst/

Malfunction

02 Sensor Off
Idle Correct
             +

02 Sensor Off  1
Idle Enrichenedj

13% Misfire
i: CO g/km Cycle: 64 kph
100% Rh 17% Rh
0% Pt 83% Pt
5.8
4.2 .
11.3
/ o 5.2
4'9 5.8
6% Rh
94% Pt
5.3
12.0
10.3

0% Rh
100% Pt
5.4
20.0
5.2
No
Catalyst
8.8
18.6
16.2
15.7

-------
                                - 22 -
          IV.2.3.4  80 kph Cruise

          As shown in Table IV-15, a similar pattern to the 64 kph
cruise case is observed for the two 02 sensor-off modes.  However, in
the misfire mode the pure Pt catalyst is now more effective for CO
control than are the two Rh-containing catalysts.

                              Table IV-15
.: CO g/km
100% Rh
0% Pt
9.9
7.2
14.5
13.2
Cycle: 80
17% Rh 6% Rh
83% Pt 94% Pt
8.2
15.8
12.1
15.2
15.3
kph
0% Rh
100% Pt
6.7
22.9
9.5

No
Catalyst
18.1
21.4
15.9
24.5
         Catalyst/
Malfunction

02 Sensor Off
Idle Correct

02 Sensor Off   ^
Idle Enrichenedj

13% Misfire

          IV.2.3.5  Idle

          Under  idle conditions, as shown in Table IV-16, little if any
CO control was achieved for the two sensor-off cases, except for the
pure Rh catalyst with the properly set idle adjustment.  Here CO con-
version, on average, was about 40%.  Under misfire conditions considerably
more conversion  was achieved, with the pure Rh being most active and pure
Pt the least.

                              Table IV-16
         Catalyst/
Malfunction

02 Sensor Off
Idle Correct
             ^
02 Sensor Off   1
Idle Enrichenedj

13% Misfire
>n: CO g/Min
100% Rh 17% Rh
0% Pt 83% Pt
1.7
2.5
9.6
0 3 °'4
°'J 0.4
Cycle : Idle
6% Rh 0% Rh
94% Pt 100% Pt
3.3 2.7
9.3 9 Q
8.8 y'°
0.5

No
Catalyst
3.4
9.5
10.5
3.7

-------
                                -  23 -
     IV.2.4  Hydrocarbons

          IV.2.4.1  Federal Test Procedure

          As shown in Table IV-17, HC emissions for the two 02 sensor
disconnect modes were least for the pure Rh catalyst and greatest for
the pure Pt catalyst.  The enriched mode gave higher values than did
the case with the proper idle adjustment.  During misfire, all three
catalysts gave substantial reductions in HC emissions, but the pure
Pt was most effective in this case, and pure Rh least effective.  Tlje
standard deviation for these HC measurements, excluding those at idle,
is 0.06.

                              Table IV-17

                Emission:  HC g/km          Cycle:  FTP


         Catalyst/    100% Rh     17% Rh      6% Rh       0% Rh        No

Malfunction             0% Pt     83% Pt     94% Pt     100% Pt     Catalyst

02 Sensor Off*\         0.25                  0>32        Q>39         0<83
Idle Correct J         0.25

02 Sensor Off  *\        n  „                  0.79        n R,         1.14
Idle Enrichenedj        "'**                  0.64        U'O:>         1.11

13% Misfire             3.48       2*87                   2'55       12<48

          IV.2.4.2  Congested Freeway Driving Schedule

          During this cycle, Table IV-18, the two 02 sensor disconnect
modes showed opposite effects.  With the A/F idle adjustment correctly
set, the pure Pt catalyst was more effective for HC control than either
of the Rh catalysts.  However, when the idle A/F setting was enriched,
the two Rh catalysts were better.  In addition,  as expected, the overall
control efficiency decreased with enrichment.  Under the misfire conditions,
the pure Pt catalyst was most effective and the pure Rh the least,  similar
to the FTP cycle misfire.

          IV.2.4.3  64 kph Cruise

          As shown in Table IV-19, the two sensor disconnect modes had
better HC control over the pure Rh catalyst than the other two catalysts.
The pure Pt was worst, and in the enriched case actually gave no control.
Under the misfire condition the situation was reversed, with the pure Pt
catalyst better than the two Rh-containing catalysts.

-------
                                  24 -
                              Table IV-18
Emission: HC g/km
Catalyst/
Malfunction
02 Sensor Off!
Idle Correct J
02 Sensor Off ^
Idle Enrichened-f
13% Misfire
Emission
Catalyst /
Malfunction
02 Sensor Of f 1
Idle Correct J
02 Sensor Off \
Idle EnrichenedJ
13% Misfire
100% Rh 17% Rh
.0% Pt 83% Pt
0.18
0.17
0.35
i ?fi 2'84
3-26 2.69
Table IV-19
: HC g/km
100% Rh 17% Rh
0% Pt 83% Pt
0.17
0.12
0.36
2 31 *'47
2'31 2.45
Cycle:
6% Rh
94% Pt
0.16
0.38
0.28
Cycle: 64
6% Rh
94% Pt
0.19
0.43
0.44

CFDS
0% Rh
100% Pt
0.13
0.46
2.55
kph
0% Rh
100% Pt
0.20
0.52
2.08

No
Catalyst
0.50
0.75
0.70
11.32

No
Catalyst
0.41
0.55
0.52
6.55
          IV.2.4.4  80 kph Cruise

          Under this cruise, as shown in Table IV-20, the pattern was
mostly similar to the 64 kph cruise, except that under the first mal-
function mode, with sensor disconnect but idle adjustment proper, all
three catalysts were similar in HC control effectiveness.  However, as
before, pure Pt was poorest when the idle was enriched, giving very
little control, but was best under the misfire condition.

-------
                                - 25 -
                              Table IV-20
Emission: HC g/km Cycle: 80 kph
Catalyst/ 100% Rh 17% Rh 6% Rh 0% Rh
Malfunction 0% Pt 83% Pt 94% Pt 100% Pt
No
Catalyst
02 Sensor Off j 0.32
Idle Correct J 0.22
02 Sensor Off ^
Idle EnrichenedJ 0.39 jj'^ 0.63 jj'Jj
13% Misfire 2.91 ^.03 2<45 ?>62
IV. 2. 4. 5 Idle
Under idle conditions, little or no control was achieved in the
two sensor disconnect modes, with the possible exception of pure Rh for
the enriched case. For the misfire case, as before, the pure Pt catalyst
gave best HC control. These results are shown in Table IV-21.
Table IV-21
Emission: HC g/Min Cycle: Idle
Catalyst/ 100% Rh 17% Rh 6% Rh 0% Rh
Malfunction 0% Pt 83% Pt 94% Pt 100% Pt
°2 Sensor Off \ 0.20
Idle Correct J 0.18 U<1/ °'2U
02 Sensor Off *1 n , 0.31 n -
Idle EnrichmentJ * 0.27 '^
n LL
13% Misfire 0.62 JJ' ** 0.46
IV. 2. 5 Nitrogen Oxides
No
Catalyst
0.19
0.34
0.32
2.11
          For both of the sensor disconnect modes the pure Rh catalyst
gave the highest level of NOX control,  while the Pt-containing catalysts
were roughly equivalent to each other.  This held true for both cyclic
driving modes and 'the two cruises.   It  ig surprising that lower levels
of NOx emissions were not observed, since the electronic control unit

-------
                                - 26 -
tended to drive the air-fuel ratio rich when the 02 sensor was dis-
connected, especially with the idle adjustment enriched.  Two possible
explanations for this effect can be proposed.  First, under cyclic
driving conditions there may be sufficient lean excursions to prevent
efficient NOX reduction at all times.  Secondly, there may not be
sufficient exothermic heat from CO oxidation under these rich conditions
to keep the catalysts at an efficient operating temperature.  As was
shown in Tables IV-12 to IV-15, only partial CO removal was achieved
under malfunction driving modes.

          Obviously, the pure Rh catalyst was much more efficient, but
whether the poor activity of Pt was due to poorer low temperature
activity or less ability to compensate for lean excursions of the air-
fuel ratio cannot be determined from this data.  The only clear
observation is that pure Rh was a better catalyst for NOX reduction
than Pt-containing catalysts when the 02 sensor was disconnected.

          At the misfire condition, however, catalyst activity reversed,
with 'the two Pt-containing catalysts showing better activity, although
all three were quite good.  This result held true for the CFDS and the
two cruises as well.
          At idle, all three catalysts controlled most of the
emissions under all operating conditions, and no judgements can be made
between them.  These results, as well as those for all the test cycle
NOX results, are given in Tables IV-22 to IV-26.  The standard deviation
for these NOX measurements, calculated from the replicate FTP, CFDS, and
cruise runs, was 0.12.
                              Table IV-22

               Emission:  NOX g/km         Cycle:  FTP


          Catalyst/    100%  Rh      17% Rh       6% Rh       0% Rh        NO

Malfunction _      0%  Pt      83% Pt     94% Pt      100% Pt     Catalyst
\
J
 02  Sensor  Of f           0.38
 Idle Correct             0.53
02  Sensor  Off                                  0.49        Q 6g        1.22
Idle EnrichenedJ       °'26                  0.71        °'b8        1.30

13% Misfire              0.16       °Q%                   0>Q5        1<5?

-------
  - 27 -
Table IV-23
Emission: NOX g/km
Catalyst/
Malfunction
02 Sensor Off \
Idle Correct J
02 Sensor Off "^
Idle Enrichened J
13% Misfire
Emission
Catalyst/
Malfunction
02 Sensor Of f I
Idle Correct J
02 Sensor Off ^
Idle Enrichened J
13% Misfire

100% Rh 17% Rh
0% Pt 83% Pt
0.30
0.43
0.26
°'04 S'.S
Table IV-24
: N0xg/ta
100% Rh 17% Rh
0% Pt 83% Rh
0.01
0.03
0.04
°-°2 S:°oi
Table IV-25
Emission: NOx g/km
Catalyst/
Malfunction
02 Sensor Of f ^
Idle Correct J
02 Sensor Off T
Idle Enrichened J
13% Misfire
100% Rh 17% Rh
0% Pt 83% Pt
0.00
0.19
0.39
n n/ °-02
°-04 o.oo
Cycle : CFDS
6% Rh 0% Rh
94% Pt 100% Pt
0.70 0.88
1%
0.02
Cycle: 64 kph
6% Rh 0% Rh
94% Pt 100% Pt
0.35
0.31 n ,9
0.23 J
0.01
Cycle: 80 kph
6% Rh 0% Rh
94% Pt 100% Pt
0.41 0.56
0.71 0 6Q
0.60 *
0.01

No
Catalyst
1.51
1.18
1.61
1.36

No
Catalyst
0.72
0.53
0.50
0.88

No
Catalyst
1.24
1.02
1.15
1.28

-------
                                - 28 -
                              Table IV-26
              Emission:  NQx g/Min
                                            Cycle:  Idle
         Catalyst/

Malfunction
Idle Correct

02Sensor Off    1
Idle Enrichened l

13% Misfire
100% Rh

  0% Pt

 0.004
 0.004


 0.007


 0.001
                                  17% Rh

                                  83% Pt
                                              6% Rh

                                             94% Ft
                                              0.000
                                              0.002
                                                          0% Rh

                                                        100% Pt


                                                         0.003


                                                         0.001


                                                         0.000
   No

Catalyst


  0.024


  0.023
  0.022

  0.031
     IV. 2. 6  Sulf ate Emissions

          The S04~ emissions from these runs are presented in Tables IV-27
through IV-31.  In addition, since the emissions were so low, Table IV-32
shows them averaged together over the four driving modes, excluding idle,
for each catalyst and operating condition.  In no case was an emission
level exceeding 4 rag/km observed.  It can be seen from Table IV-32
that for the two sensor disconnect cases the pure Pt catalyst had the
highest emissions, exceeding those from the raw engine exhaust.  The pure
Rh catalyst, however, in showing the lowest emissions, was lower even than
the no catalyst, or raw engine exhaust case.  Under misfire conditions
the situation was reversed, the pure Rh catalyst having highest emissions,
exceeding the raw exhaust case, and the pure Ft the lowest.  In all
these cases, the Pt-Rh catalysts fell between the two pure catalyst
emissions .
                              Table IV-27
              Emission:  S04= mg/km
                                             Cycle:  FTP
         Catalyst/    100% Rh

Malfunction	       0% Ft

02 Sensor Off  \         1.0
Idle Correct  J         0.4

02 Sensor Off   1       Q ^
Idle EnrichenedJ

13% Misfire             3.0
            17% Rh

            83% Ft
                                               6% Rh

                                             94% Ft
                                               2.7
                                               0.6
                                                          0% Rh

                                                        100% Ft
                                                          2 Q
                                                          2'°

                                                          0.6
                                                                       No

                                                                    Catalyst
                                                                      0.9
                                                                      0.8

                                                                      1.3

-------
  - 29 -
Table IV-28
Emission: 804 mg/km
Catalyst/
Malfunction
02 Sensor Off I
Idle Correct J
02 Sensor Off \
Idle Enrichened J
13% Misfire

100% Rh 17% Rh
0% Pt 83% Pt
0.4
0.9
0.4

!:!
Table IV-29
Emission: S04= mg/km
Catalyst/
Malfunction
°2 Sensor Off 1
Idle Correct J
02 Sensor Off ^
Idle Enrichened (
13% Misfire

Emission:
Catalyst/
Malfunction
Oo Sensor Off I
Idle Correct J
02 Sensor Off \
Idle Enrichened J
13% Misfire
100% Rh 17% Rh
0% Pt 83% Pt
0.0
0.2
0.0
0 3 °'2
°'J 0.1
Table IV-30
804' mg/km
100% Rh 17% Rh
0% Pt 83% Pt
0.1
0.2
0.2
n fi °'2
°'6 1.0
Cycle :
6% Rh
94% Pt
0.8
1.4
1.0


Cycle :
6% Rh
94% Pt
0.0
0.4
0.2


Cycle :
6% Rh
94% Pt
0.6
1.4
0.7

CFDS
0% Rh
100% Pt
2.0
3.0

0.6

64 kph
0% Rh
100% Pt
1.0
1.6
0.1

80 kph
0% Rh
100% Pt
1.3
3.9
0.1

No
Catalyst
1.0
0.8
0.7
0.6


No
Catalyst
0.4
0.2
O.,0
0.2


No
Catalyst
0.3
0.3
0.1
0.1

-------
                                 - 30 -
                              Table IV-31
             Emission:
         Catalyst/

Malfunction
02 Sensor Of f
Idle Correct
              1
              J
                1
02 Sensor Off
Idle Enrichened J

13% Misfire
             Emission:
          Catalyst/

Malfunction
 Q£  Sensor Off
 Idle  Correct
               1
               /
02 Sensor Off   \
Idle Enrichenedj

13% Misfire
S04 mg/Min
100% Rh 17% Rh
0% Pt 83% Pt
0.0
0.1
0.0
°0.l
Table IV-32
S04= g/km
100% Rh 17% Rh
0% Pt 83% Pt
0.42
0.41
0.31
Cycle :
6% Rh
94% Pt
0.1
0.0
0.0

Idle
0% Rh
100% Pt
0.0
0.1
0.0

No
Catalyst
0.0
0.1
0.0
0.1
Cycle : Average
6% Rh
94% Pt
0.56
1.51
0.65
0% Rh
100% Ft
1.42
2.60
No
Catalyst
0.83
0.55
0.40
                           20
                          .ZU
                                    °'80
                                                          0.35
0.58
      IV.2.7  Hydrogen Sulfide Emissions

           While carrying out these runs,  ERE also made measurements  of
 H2S emissions.   The results are given in  Tables  IV-33 to  IV-37  and the
 analytical procedure in Appendix F.  The  standard deviation for the  true
 replicates, those cases where identical sampling  flow rates  were used,
 is 1.27 if all nine true replicate cases  are used.   However,  during  FTP
 testing,  during misfire conditions, a pair of replicates  gave 7.02 and
 1.46 mg/km of H2S.   If this pair is excluded from the calculation, the
 standard deviation is reduced to 0.59.

-------
                                - 31 -
                              Table IV-33
Emission: H2S mg/km Cycle: FTP
Catalyst/ 100% Rh 17% Rh 6% Rh 0% Rh
Malfunction 0% Pt 83% Pt 94% Pt 100% Pt
^
°2 Sensor Of f I 0.94 0>?6 1<69
Idle Correct J 0.40
°2 Sensor Off \ 0>go 1.25 Q<74
Idle Enrichenedj 0.95
13% Misfire 0.79 £°* 3.78
Table IV-34
Emission: H2S mg/km Cycle: CFDS
Catalyst/ 100% Rh 17% Rh 6% Rh 0% Rh
Malfunction 0% Pt 83% Pt 94% Pt 100% Pt
02 Sensor Off] 1.69 0<75 ^Q
Idle Correct J 0.72
02 Sensor Off *\ - « 2.13 Q ,R
Idle Enrichenedj '^ 1.02 u'°°
13% Misfire 8.23 ^02 5*U
Table IV-35
Emission: H2S mg/km Cycle: 64 kph
Catalyst/ 100% Rh 17% Rh 6% Rh 0% Rh
Malfunction 0% Pt 83% Pt 94% Pt 100% Pt
02 Sensor Off] '1.91
Idle Correct J 0.91 u'4b u'^8
02 Sensor Off "\ ~ -, 0.41 n
Idle Enrichenedj ' 0.20 U
-------
                                 - 32  -
                              Table IV-36
             Emission:   H2& rag/km
                    Cycle:   80 kph
         Catalyst/

Malfunction

02 Sensor Off
Idle Correct

02 Sensor Off   }
Idle Enrichened j

13% Misfire
                      100% Rh

                        0% Ft

                        3.16
                        2.10


                        3.03

                        0.37
           17% Rh

           83% Pt
           6% Rh

          94% Pt


           0.44


           0.64
           0.64
            2.72
            1.50
            0% Rh
          100% Pt


            0.87


            0.25

            3.33
             No
           Catalyst


           <0.04

             0.07
             0.56**

             0.11
                              Table IV-37
              Emission:  H2S mg/min
                      Cycle:   Idle
         Catalyst/

Malfunction	

02 Sensor Off!
Idle Correct J

02 Sensor Off   \.
Idle EnrichenedJ

13% Misfire
100% Rh
  0% Pt

  0.33
 <0.05

 <0.05

  0.19
  0.04
17% Rh
83% Pt
 6% Rh
94% Pt

 0.52

<0.04
<0.09
  0.12
  0% Rh
100% Pt

  1.29


  0.06

  0.29
                                                                        No

                                                                     Catalyst


                                                                      <0.04


                                                                      <0.10
                                                                      <0.04**

                                                                       0.08
          As was  shown in Table IV-1, the four experimental catalysts
under normal operating conditions gave a maximum of 0.3 mg/km of H2S for
the FTP.  Table IV-33 shows  that under malfunction conditions increases
in H2S  during  the FTP are observed,  although the absolute emission levels
are still low.  There appears  to be  no pattern with catalyst composition
changes.  As the  malfunction mode changes, it appears that misfire can
sometimes lead to the highest  values, especially if the 7.02 mg/km
figure  is real.   Another suspect number is the 15.29 mg/km observed
with no catalyst  when the sensor was disconnected and the idle enriched.
In no other non-catalyst run;  over cyclic or cruise modes, was a value
higher  than about 0.6 mg/km  seen.

-------
                                - 33 -
          The CFDS cycle also has its highest H2S emissions under mis-
fire conditions, as much as 8.23 mg/km with the pure Rh catalyst.  For
the other conditions there is again no discernible pattern with catalyst
composition.

          At cruise, H2S emissions over the catalysts are also generally
increased compared to the engine out H2S levels.  The pure Rh catalyst
gives the highest values when the 02 sensor is disconnected, but the Pt
catalyst is highest at misfire conditions.  Again, however, the absolute
emission levels are low, a maximum of 3.88 mg/km with the pure Pt
catalyst at misfire and 64 kph.

          At idle, the catalyst H2S emissions are sometimes higher than
the engine-out emissions, but with the exception of the pure Pt catalyst
when the idle adjustment was peoperly set, not appreciably higher.

-------
                                - 34 -
                            V.  References
1.  Voorhoeve, R. J. H., et al., "Hydrogen Cyanide Production During
    Reduction of Nitric Oxide Over Platinum Catalysts," Science, 190,
    149, October 10, 1975.

2.  Bradow, R. L., and Stump, F. D., "Unregulated Emissions from Three-
    Way Catalyst Cars," SAE Paper 770369, March, 1977.

3.  Ashmead, D. R., et al., "Nitrogen Oxide Removal Catalysts for
    Purification of Automobile Exhaust Gases," SAE Paper 740249,
    February, 1974.

4.  Griffith, M. G., et al., "Assessment of Automotive Sulfate Emission
    Control Technology," EPA Report 460/3-77-008, June, 1977.

-------
                                  - 35 -

                               APPENDIX A

    DETERMINATION OF CYANIDE IN SODIUM HYDROXIDE  IMPINGER SOLUTIONS
    	USING AN ION-SELECTIVE ELECTRODE	


Introduction and Scope

          This method is applicable for the determination of cyanide  in
0.5M sodium hydroxide impinger solutions.  A cyanide ion-selective
electrode is used for the measurement.

Summary of Method

          The impinger solution is treated with additional sodium hydroxide
in order to (1) level differences in ionic strength between the scrubber
solution and the calibration standards and (2) increase the pH of the
scrubber solution to insure that all the hydrogen cyanide  is converted
to cyanide ion.  Cyanide is then measured using an ion-selective electrode.
Sulfide interferes with the method.  If it is present, it is removed by
treatment of the solution with cadmium carbonate.

Apparatus

          (1)  Orion 94-06 Cyanide Ion-Selective Electrode, or equivalent.

          (2)  Orion 90-01 Single Junction Reference Electrode, or
equivalent.

          (3)  Orion 801A Digital Ion analyzer, or equivalent.

Reagents

          (1)  10M sodium hydroxide ionic strength and pH adjustment
solution.  Dilute 400 g of 50% sodium hydroxide solution to 500 ml with
deionized water.  Store in a plastic bottle.

          (2)  0.5M sodium hydroxide solution.

Cyanide Solutions

     NOTE .1;  Cyanide is extremely toxic and should be
              handled using appropriate safety precautions.

          Solution 1;  1000 yg CN~/ml.  Prepare by weighing 2.503 g
potassium cyanide and making up 1 liter with 0.5M sodium hydroxide.

          Solution 2;  100 yg CN~/ml.   Dilute 20 ml of Solution (1)  to
200 ml with 0.5M sodium hydroxide.

          Solution 3;  10 yg CN~/m.  Dilute 20 ml of Solution (2) to
200 ml with 0.5M sodium hydroxide.

-------
                                  - 36 -
          Solution 4;  5 yg CN~/ml.  Dilute 10 ml of Solution (2) to
200 ml with 0.5M sodium hydroxide.

          Solution 5;  1 yg CN~/ml.  Dilute 20 ml of Solution (3) to
200 ml with 0.5M sodium hydroxide.

          Solution 6;  0.5  yg CN~/ml.  Dilute 10 ml of Solution  (3) to
200 ml with 0.5M sodium hydroxide.

          Solution 7;  0.1 yg CN~/ml'.  Dilute 20 ml of Solution. (5) to
200 ml with 0.5M sodium hydroxide.

          Solution 8;  0.05 yg CN~/ml.  Dilute 10 ml of Solution  (5) to
200 ml with 0.5M sodium hydroxide.

          Solution 9;  0.01 yg CN~/ml.  Dilute 20 ml of Solution  (7) to
200 ml with 0.5M sodium hydroxide.

     NOTE 2;   Store all the cyanide solutions in
               plastic bottles.

     NOTE 3;   Discard Solutions 3-9 after one week.

Calibration

          Refer to the manufacturer's instruction manuals for detailed
operating instructions.

          (1)  Connect the electrodes to the meter.  Pipet 20 ml  of the
5, 1, 0.5, 0.1, 0.05, and 0.01 yg CN~"/ml cyanide solutions into plastic
containers, add 1.0 ml of 10M sodium hydroxide solution, and obtain
potential readings.  The cyanide solutions should be stirred at a constant
fate, and the  readings should be taken after the potential stabilizes,
a.5 jain,.  The. response  of the cyanide electrode drifts and, therefore.,
the calibration should be checked occasionally.  If the electrode drifts
excessively, polish it with an Orion 94-82-01 polishing strip and
recalibrate.

          (2)  Obtain a calibration curve by plotting the logarithm of
the cyanide concentrations vs. the potential readings.  This curve is
linear from 0.04 to 5 yg CN~/ml with a slope of approximately 50  mV per
ten-fold change in cyanide concentration.  Below 0.04 yg CN /ml the curve
deviates from  linearity.

Procedure

          (1)  Pipet 20 ml of the scrubber solutions into plastic containers,
add 1.0 ml of  10M sodium hydroxide solution and obtain potential  readings
as indicated above.

          If sulfide is present, add approximately 30 mg of cadmium carbonate,
stir for 0.5 hr., and filter before obtaining the potential readings.

-------
                                   - 37  -
Calculation

          (1)  Obtain the cyanide concentration of the scrubber solutions
from the calibration curve.  Values for cyanide concentrations below
0.04 ug CN /ml are only approximate,  and conclusions drawn from solutions
that are below 0.04 yg CN~/ml should  be tempered.

-------
                              - 38 -


                            APPENDIX B

               CONSIDERATIONS IN THE MEASUREMENT OF
            CYANIDE ION USING ION-SELECTIVE ELECTRODES
 Introduction


          Cyanide ion is measured using an ion-selective electrode by
recording the potential developed between this electrode and a reference
electrode.  The cyanide electrode is composed of a pressed disk of silver
bromide and senses cyanide ion by the attack on the electrode to form
Ag(CN)2~".  This attack, which results in the dissolution of the electrode,
limits its lifetime at high cyanide concentrations, and therefore, the
electrode is not generally used in solutions containing more than 5-10
pg CN~/ml.

          Since the electrode response results from attack on the silver
bromide disk; AgBr, Kgp = 3 x 10-13; to form Ag(CN) ~, K^NSTAB - 4 x 10"19,
 it follows that any material which chemically attacks the electrode gives
 rise to a response, and thus interferes with cyanide measurement.  Sulfide,
Ag«S, Kgp = 10"-*, is a very serious interferent.  Iodide, bromide and
 chloride.interfere to a much less extent; Agl Kgp = 8 x 10""^, AgBr K   »
 3 x 10-13, AgCl Kgp - 2 x 10-10.  The electrode cannot be used in strongly
 reducing solutions since AgBr would be reduced to elemental silver.

          The electrode responds  to free cyanide ion.  Therefore, the pH
 of the solution must be sufficiently high, pH ^O, so that cyanide ion
 greatly predominates in solution  over hydrogen cyanide, K  = 4 x 10-10.
 Furthermore, the presence of materials which complex cyanide ion,_such as
 iron  (II) and iron  (III) ions; K^.,.,... Fe(CN),4~ - 10-37, Fe(cmJ- =
 10-44; mU8t be absent.          TNSTAB       6                  6


           Since  the  response of the  electrode  is  Nernstlan,  the  potential
 due to cyanide  ion is  proportional to the logarithm of its  activity.
 Since activity  is dependent on ionic strength,  the electrode must  be
 calibrated in solutions which have ionic strengths which are similar to
 those in which  cyanide is to be measured.

 Scrubbing of Carbon Dioxide
 Containing Gases	

           When a sodium hydroxide scrubber is used to remove hydrogen
 cyanide from carbon dioxide containing gases,  (1) the pH of the solution
 vill drop as hydroxide is converted to carbonate and then to bicarbonate
 eventually reducing the effectiveness of the solution for hydrogen cyanide
 removal , and (2) the ionic strength of the solution will change as the
 hydroxide, carbonate, and bicarbonate change in the scrubbing solution.

-------
                                  - 39 -
          This report is directed to the discussion of a procedure
(Attachment I) for the measurement of cyanide in hydroxide scrubbing
solutions.  The procedure incorporates steps to insure (1) the pH of solu-
tions are sufficiently high so that cyanide ion predominates over hydrogen
cyanide, (2) that ionic strength differences between calibration and meas-
urement solutions are leveled, and (3) removal of sulfide from measurement
solutions.  The response of two ion-selective electrode/reference electrode
combinations are also discussed.

Experimental

          Cyanide potentials were obtained using an Orion 801A digital
voltmeter and an Orion cyanide electrode-Orion single junction reference
electrode combination or a Metrohm cyanide electrode (manufactured by
Orion to Metrohm specs)-Orion double junction reference electrode combina-
tion.  The electrodes were allowed to stabilize in cyanide containing
solutions before data were recorded.  The data that are given in Tables B-l
and B-2 are the average of two to three measurements on the solutions.

Results and Discussion

          In order to evaluate the procedure which is employed for
cyanide measurements in sodium hydroxide scrubbers used to remove cyanide
from carbon dioxide containing gases, see Appendix A, cyanide measure-
ments were made in synthetic model solutions.  The procedure was developed
on the basis that a 0.5M sodium hydroxide solution is used for scrubbing
the gas.  Before the measurement of cyanide, the solution is prepared for
measurement by adding additional sodium hydroxide to make the solution
approximately 1M in sodium ion.  This assures that the pH is sufficiently
high BO that hydrogen cyanide is dissociated into cyanide ion and that the
ionic strength of the scrubbing solution is comparable to the calibration
solutions, which are prepared in approximately 1M sodium hydroxide.

          The model solutions which were used in this study contained
0.005 to 1.0 pg cyanide per ml prepared by dissolving sodium cyanide in
(1) 1M sodium hydroxide,  (2) 0.75M sodium hydroxide and 0.125M sodium
carbonate, and  (3) 0.5M sodium hydroxide and 0.25M sodium carbonate.
These solutions represent, respectively, the calibration solution, a
scrubber solution with 50 percent conversion of hydroxide to carbonate
and prepared for cyanide measurement, and a scrubber solution with 100
percent conversion of hydroxide to carbonate and prepared for cyanide
measurement.  The 100 percent conversion case represents the maximum
amount of carbon dioxide  that can be contained in a scrubber without the
possibility of rendering  it ineffective for hydrogen cyanide scrubbing.

          The potential measurements obtained on these solutions using
an Orion cyanide electrode and a single junction reference electrode are
given in Table B-l and those obtained using a Metrohm cyanide electrode
and a double junction reference electrode are given in Table B-2.

-------
                                   - 40 -
          The conclusions that can be drawn from the data in the Tables
are, that above 0.04 yg CN~/ml and within the experimental error of  the
measurementt

          (1)  The response of the two  electrode combinations are equiv-
               alent with, respect to the potential  change observed for a
               10-fold increase  in cyanide concentration.

          (2)  The response of the two  electrode combinations are inde-
               pendent of the matrices  which were studied.  This indi-
               cates that the procedure that is used for cyanide measure-
               ments in  scrubber solutions levels the  pH and ionic
               strength  effects  on cyanide measurements due to  dissocia-
               tion or activity  considerations.

           (3)  Below  0.04 yg  CN""/ml,  differences in matrices affect   the
               cyanide  response.

           (4)  The  useful limit for  both measurement Systems are
               equivalent,  approximately 0.04  yg CN~"/ml.

           Since  the potential measurements were found  to be independent
 of matrix,  above 0.04 yg CN~/ml, they were  averaged for each concentration
 for the electrode systems.   These averages  are presented in Tables B-l and B-2,
 and plotted in Figures B-l  and B-2.  As can be seen from the graphs,  the cali-
 bration is linear above 0.04 yg CN""/ml.  The  linear regression  lines ob-
 tained from the data between 0.04 and 1.0 yg  CN~/ml are indicated on the
 graphs.

 Summary and Coneluaions

           Using the procedure which is given in Appendix A for  cyanide
 measurements in scrubbers used to remove hydrogen cyanide  from carbon
 dioxide containing gas streams, up to the maximum carbon dioxide content
 consistent with efficient hydrogen cyanide removal,

           •  The two electrode systems are equivalent.

           •  Matrix effects  are negligible*

           9  The cyanide content can be measured to 0.04 yg CN /ml
              with confidence.  Below this concentration, large errors
              are to be expected in cyanide measurements due to matrix
              effects and deviation of  the cyanide  response from
              Nernstian behavior.

-------
  - 41 -
TABLE B-l
MEASUREMENT OF CYANIDE USING AN ORION CYANIDE
ION-SELECTIVE ELECTRODE AND AN ORION SINGLE-JUNCTION
REFERENCE ELECTRODE
-mV
CN~. US/ml l.OM NaOH
0 14.1
0.005 16.1
0.01 22.1
0.02 29.2
0.04 41.6
0.06 48.8
0.10 58.8
0.50 94.8
1.00 110.5

0.75M NaOH
0*125M Na«CO,
3.8
10.6
17.8
23.0
40.1
48.4
58.6
95.4
112.2

0.5M NaOH
0.25M Na*C04
-2.5
8.1
17.4
30.4
40.8
48.9
64.6
95.8
113.2

Avg.
	
	
	
	
40.8
48.7
60.7
95.3
112.0
50.9 mV
decade

-------
  - 42 -
TABLE B-2
MEASUREMENT OF CYANIDE USING A METROHM CYANIDE
ION-SELECTIVE ELECTRODE AND AN ORION DOUBLE-JUNCTION
REFERENCE ELECTRODE
-mV
CN~, llg/ml l.OM NaOH
0 20.3
0.005 23.1
0.01 28.8
0.02 36.3
0.04 48.7
0.06 56.4
0.10 66.4
0.50 101.0
1.00 116.5

0.75M NaOH
0.125M Na^CO,
10.2
17.6
24.0
30.8
46.8
55.6
65.0
101.7
117.1

0.5M NaOR
0.25M Na0C00
1.4
13.2
22.3
35.0
45.1
53.4
70.0
101.4
120.4

Avg.
	
	
	
	
46.9
55.1
67.1
101.4
118.0
50.7 tnV
decade

-------

-------
20
10
  0.001
                                                             CH~/ml
                                                                                                                          1.0

-------
               -  45  -
             APPENDIX C
instruction  manual
cyanide ion activity electrode
model 94-06

introduction
     .required equipment
     before using the electrode
     measurements with a Specific Ion Meter
chemistry of cyanide
     activity and concentration
     effect of pH
     metal ion complexation

cyanide measurements

     activity standards
     general procedure
     activity measurements
     concentration measurements

electrode characteristics
     potential behavior
     electrode life
     precision
     temperature effects
     response time
     interferences

appendices
     I estimating total ionic strength
     II values of 2.3 RT/F

specifications

Orion Research Incorporated
11 Blackstone Street
Cambridge, Massachusetts 02139
617-864-5400
   ©1972 Orion Research Incorporated/Printed in U.S.A.
    Form IM94-06/2701

-------
                                                   -  46  -
introduction
The  IONALYZER® Cyanide Ion Activity Electrode.
Model  94-06, is one of a family of electrochemical
sensors developed by Orion Research. Designed to be
used like a conventional  pH electrode, it has a solid
state sensing element and an unbreakable epoxy body
which  is highly resistant to chemical attack.
At the present time, the 94 series also includes elec-
trodes for fluoride, sulfide, silver, bromide, iodide, and
chloride ion. All Series 94 electrodes, which have solid
state sensing elements, may be used in a wide variety
of aqueous and non-aqueous solutions.
Orion also offers Series 92 electrodes with liquid ion
exchanger membranes. Nitrate, calcium, chloride, di-
valent cation (water hardness), perchlorate, and cupric
 ion  electrodes are  currently available in the series.
These electrodes are designed to be used in aqueous
 solutions only.

 required equipment
 In addition to the electrode, the following equipment
 is needed to perform cyanide determinations:
     IONALYZER Specific Ion Meter, or a pH meter
                     with an expanded millivolt scale
     Single Junction Reference Electrode, Model 90-01,
                                       or equivalent
     Electrode Holder, Model 92-00-01

  before using the electrode
  A  protective rubber cap is placed over the bottom of
  the electrode to prevent damage to the sensing element
  during  shipment.  Remove the cap before using the
  electrode. It is advisable to replace the cap when the
  electrode is not in use.
measurements with a Specific Ion Meter

The upper (red) scale of the Specific Ion Meter is used
for direct reading  of ion  concentration or activity.
Readings can be scaled by any power of 10: thus, for
example, a center scale reading can be assigned a value
of 1.0 ppm or 0.10 M.  If  the center scale reading is
1 ppm, then the extreme left hand reading is 0.1 ppm,
and the extreme right hand reading is 10 ppm.

To read cyanide, set
the Function Switch
to the F" position.
Place the electrodes
in a standardizing so-
lution with a value
that is an even power
of 10. Turn the cali-
bration control  for a
center scale reading.

NOTE:   Moving the
temperature  control
does not affect a cen-
ter scale reading.

To obtain the correct
slope setting, place
the  electrodes  in  a
second standard with
a  value  within one
decade  of the first
standard. Now adjust
the temperature con-
trol until a  correct
reading is obtained.
Move the slope con-
trol until the sample
temperature  is  read
on the temperature
(0-100°C) scale.  Now
the % of theoretical
slope can be read on
the slope (80-100%)
scale.
                                                                                              TEMPERATURE
                                                                 The  meter  is now
                                                                 calibrated and ready
                                                                 for use.
                                                                                                100
                                                                                      s.

-------
                                                  - 47 -
                                                            chemistry  of cyanide
1.00H
0.95-
0.90-
0.85-
0.80-
0.75-
0.70-
                               figure 1
           cyanide ion activity coefficient
                  vs. total ionic strength
 Activity
Coefficient
                Total Ionic Strength (M)
           10
              -3
              10
                       -2
10
                                -1
10°
 1.0-
 0.8-
 0.6-
                               f igure 2
                        ionized fraction
                             of cyanide
                         vs. solution pH
 0.4-
 0.2-1
                             PH
                              activity and concentration

                              Cyanide in solution may be present as the free ion
                              (CN"), or may  be complexed to hydrogen or metal
                              ions. The "chemical effectiveness" of the free portion
                              of the cyanide depends on the overall composition of
                              the solution and the solution temperature, as well as
                              on  the actual concentration of free, dissociated cya-
                              nide ion.

                              The cyanide electrode responds to the activity (or
                              "effectiveness") of the dissociated cyanide ion.  The
                              activity of cyanide ion, rather than the concentration,
                              determines the rate and extent of chemical reactions,
                              and is, in general, a more significant parameter than
                              cyanide ion concentration.
                              Total cyanide concentration (Ct) consists of  both
                              free (Cf) and bound (Ct,) ions:
                                                            1)
                                                             c, -
                  9
                   10
    T
    11
   I
   12
                                                     The free cyanide concentration (Cf) is related to the
                                                     cyanide ion activity (A) by a factor called the activity
                                                     coefficient (7):
                                                            2)
The activity coefficient is variable and depends on the
total ionic strength of the solution.  (See Appendix I.)
At  low pH's, very little cyanide is in the ionic form.
and the effect of total  ionic strength is relatively un-
important.

Note that free cyanide (Cf) in this context refers only
to the cyanide which is present as the CN~ ion, and
should not be confused with the electroplating term
meaning all cyanide not bound to metal complexes.
effect of pH

Hydrogen ion complexes cyanide ion to form the weak
acid HCN.  The greater the hydrogen ion activity, the
greater the amount of cyanide ion which is complexed.
Figure 2 shows the fraction of free cyanide ion as a
function of pH. At pH 8. less than 10% of the cyanide
is free; 50% is free at pH 9.2; and at pH 12 virtually
all the cyanide is in the form CN .

-------
                                                   -  48  -
metal ion complexation

Many transition and  heavy metal ions (Fe** Fe***,
Co+++. Ni++, Cu+, Ag* Cd~. Hg~ etc.) form very stable
cyanide complexes.  The extent of complexation de-
pends on the relative levels of cyanide and the com-
plexing metal, as well as on solution pH. In acid solu-
tions, there is less metal complexation of cyanide, but
greater hydrogen ion complexation.


cyanide measurements

 activity standards

 Standardizing solutions of fixed pH and total ionic
 strength are  prepared by diluting a 0.1 M cyanide so-
 lution with 0.1 M NaOH.

 1.  Prepare a stock solution of approximately 0.1 M
    NaOH. Dissolve 4.0 grams of solid NaOH in 1 liter
    of distilled water. Store in a plastic bottle. Pre-
    pare fresh solutions weekly.

 2.  Prepare a stock solution of 0.1 M CN~.  Dissolve
    4.90 grams of reagent grade NaCN in 1 liter of dis-
    tilled water.  (Alternately,  6.51  grams of reagent
    grade KCN may be dissolved in 1 liter  of distilled
    water.) Store in  a plastic bottle. Prepare fresh so-
    lutions weekly.
 3. Prepare a 10~3  M CN" solution. Pipet 1 ml. of the
    0.1 M CN" solution into a 100 ml. volumetric flask
    and dilute to the mark with the 0.1 M NaOH solu-
    tion.  This solution must  be prepared each day.

 4. Prepare  a  10~4 M CN" solution by diluting the
    10~3 M  CN" solution 10:1 with the 0.1  M NaOH
    solution. If desired, a 10~5 M CN" solution may be
    made by diluting the 10~3 M CN" solution 100:1
    with the stock NaOH solution. These standardizing
    solutions must also be prepared daily.

  NOTE:  For activity measurements, cyanide  standard-
          izing solutions must be assigned their correct
          activity values.  When using standardizing so-
          lutions with a 0.1 M NaOH background, mul-
          tiply the cyanide ion concentration by 0.77
          to obtain the cyanide ion activity value.
general procedure

1. Insert the reference electrode connector into the
   reference electrode input jack on the Specific Ion
   Meter or pH meter.  Make sure the rubber band or
   plug used to cover the electrolyte filling hole has
   beer) removed.

2. Insert the cyanide electrode  connector into the
   sensing electrode input jack.  Make sure the cap has
   been removed from the electrode membrane.

3. Set the Function Switch of the Specific Ion Meter
   to the  F" position. If a  pH  meter is used,  set the
   Function Switch to the expanded millivolt position.

4. Place both electrodes in appropriate standardizing
   solutions and record the readings.  Blot the  elec-
   trodes  between  solutions to prevent sample carry-
   over.  Stirring of both samples and standards at a
   fixed rate is necessary during  measurement

   Always blot solutions from the sensing element
   with a soft tissue. Never wipe the sensing element
   except to remove adherent deposits. Scratches on
   the  membrane surface cause slow  response to
   changes in ion  level.

5. If a Specific Ion Meter is used, calibrate as recom-
   mended  in  the  instrument  instruction  manual.
   Brief directions are also given on page 3.

   If a pH meter with an expanded millivolt  scale is
   used, plot the activity or concentration values of
   the standardizing  solutions on the logarithmic axis
   of a piece of semi-logarithmic graph paper.  Plot
   the potentials developed in the standardizing  solu-
   tions on the linear axis.

6. Place both  electrodes in the unknown solution.
   Read the value of  the unknown  on the direct-
   reading specific ion scale of the Specific Ion Meter.
   If electrode potentials are read out on a pH meter,
   convert the millivolt reading to sample activity or
   concentration using the calibration curve.

   Samples and standardizing  solution* should  be at
   the same temperature.

 7. After  use,  rinse the electrode with distilled water
   and blot dry. Store the electrode upright with the
   protective cap in place.

-------
                                                  - 49 -
activity measurements

The  electrode responds directly to cyanide ion acti-
vity, without regard to sample ionic composition, total
ionic strength, pH, or complexing species.  To deter-
mine the activity in an unknown solution, standardize
the electrodes in solutions of known cyanide ion acti-
vity  and measure the activity of the unknown sample
following the directions in the section genera! pro-
cedure.

All  standardizing  solutions and  samples should be
stirred at about the same rate for good reproducibility.

Sample activities can be measured over the range 10~3
to 10~6 M,  and between 10~2 and 10~3 M on an in-
termittent basis only.  The electrode should not be
used in samples with cyanide activites above 10~2 M,
as excessively rapid corrosion of the sensing element
will  occur.

The  cyanide electrode develops a  potential propor-
tional to the logarithm of the activity of the free cya-
nide ion in the sample solution. At 25°C, the electrode
exhibits about a  59 mv change in potential for each
tenfold change in  cyanide ion activity.  Potentials are
increasingly positive in more dilute solutions and in-
creasingly negative in more concentrated  solutions.
Figures (broken line) shows typical potential response
of the cyanide electrode to changes in cyanide ion ac-
tivity.  Actual electrode potentials can vary as much
as ±20 mv from the values shown, depending on the
stirring rate, the reference electrode used, sample tem-
perature, etc.
The Single Junction  Reference  Electrode, Orion
Model 90-01, is recommended.
figure 3
typical electrode response to cyanide ion activity
(broken line) and concentration (solid line)
(all solutions 0.1 M in NaOH)
                                                I
                                               •o
                                                I
                                               •o
                                          JO
                                           o
                                          2
                                           CO
                                           I
                                          1
                                          V)
                                                                                                               I
                                                                                                              •o
8
                                                                        o
                                                                        to
                                                                        CM
                       O
                       in
o
o
T
o
to
I

-------
                                                  -  50  -
                                                              electrode  characteristics
concentration measurements

If  measurements of total cyanide concentration are
required, estimate the total ionic strength and cyanide
level of the sample and use Figure 4 to obtain the re-
commended sample  dilution factor.  Prepare pure
NaCN standards which bracket the expected value of
the unknown and dilute these standards with NaOH
in  exactly the same manner as the unknown sample.
Diluted samples and  standards have the  same total
ionic strength  and a  sufficiently high pH to destroy
the HCN complex. The cyanide concentration o"f the
unknown can now be determined following the direc-
tions in the section general procedure.

AH samples and standardizing solutions must be stirred
at a fixed rate to insure good reproducibility.

Samples with  metal complexes present, if measured
by this technique, give approximate values for the to-
tal free cyanide (as defined by electroplaters; i.e., all
cyanide not bound to metal complexes).  Values are
approximate since sample dilution can cause some dis-
sociation of the metal-cyanide complex.

Samples with  high cyanide levels are diluted to the
recommended operating  range (10~3 M  to 10~s  M
cyanide) as required with 0.1 M NaOH.  Standards are
prepared with 0.1 M NaOH as outlined  in the section
standardizing solutions.

Standards with high total ionic strengths must be meas-
ured in a concentrated background solution so that
variations in the original sample ionic strength do not
affect  the measurement.  If 1.0M NaOH is used to
prepare the samples,  then standards must also be pre-
pared with 1.0 M NaOH.
titrations
The cyanide electrode can be used as an end point for
the titration of cyanide solution with silver nitrate. If
titrations  are to be  done routinely, the silver elec-
trode, Orion  Model 94-16, is preferred,  as it can be
used in more concentrated solutions and will  give
larger end-point breaks. Using either electrode, there
will be  two end-point breaks, corresponding to the
formation of Ag(CN)2" and AgCN respectively.
potential behavior

The sensing element of the cyanide electrode is a solid
mixture of inorganic silver compounds.  This mem-
brane is an ionic conductor which allows silver ions to
pass between  the sample solution and the internal re-
ference solution, which is kept  at a fixed silver ion
level.  The distribution of silver ions between the two
solutions develops a potential which depends on the
silver  ion activity in the sample solution:
3)    E = Ea  +
2.3 5J logAA
                                9
where: E = the measured total potential of the system
      Ea = the portion of the total potential due to
          choice of reference electrodes and internal
          solutions
2.3 RT/F = Nernst factor (59.16 mv at 25°C): R and F
          are constants, and T is the temperature in
          degrees Kelvin
    AAQ* - the silver ion activity in the sample solution
figure 4
recommended sample dilution vs. sample total ionic
10-,


5
a, '•u"
Q.
E
•CO
V)
010-'-
£
O)
0)
is
o 10-2-
c
5
"55
° 10'3-






O
CO
5
5
JC

1
o
1
• •
T—
J3
Q


+^
'i X
Iz
** S
3 o
2 <~
5


^
*;
o O
O CO
7 z
^> "*
5 5
o
strength





Dilute 1:1 with
1.0 M NaOH








Dilute 1:1 with
0.1 M NaOH
i i i
               0'1    10~2    10~2    10'4

                Cyanide Level in Sample (M)
                            10
                                                -s

-------
                                                    - 51 -
The operation of the electrode depends on the reaction
of cyanide ion with sparingly soluble silver salts from
the membrane surface:

4)    AgX  + 2CIST -» Ag(CN}2- + X"

The ions (X") released by this reaction fix the activity
of silver ion at the membrane surface through the sol-
ubility product relationship:
5)
where Ksp is the solubility product of the silver salt
and  v depends on the stirring rate.  Substitution in
Equation  (3) yields:
6)    E =
                                              or:
                               2K
7)    E = [E. + 2.3^


           2.3^.logACN-

When the temperature does not vary and the stirring
rate is kept reasonably constant, the terms in brackets
are constant and can be defined by a new constant Eb:
8)    E = Eb -

electrode life
                  2.3^JlogACN-
Operation of the electrode depends on the slow solu-
tion of  one of the electrode membrane components
(see the section potential behavior).  Ultimately the
membrane  will erode to a  point where satisfactory
operation  is no longer possible.  Lifetimes decrease at
high sample cyanide levels and at high rates of stirring.
At  moderate stirring rates,  the  electrode should last
for several  hundred hours of continuous operation at
a 10~3 M cyanide ion level.  At 10~s M cyanide levels
the electrode  should operate for over  four months.
The electrode  can be used in NaOH solutions as con-
centrated as 5 M, and in mixed water and water-soluble
organic  solvents.
precision

The electrode will detect low levels of cyanide Ion;
however, since it is a logarithmic device, it cannot be
used to detect very small changes in cyanide ion acti-
vity at high cyanide levels. Factors such as electrode
temperature,  drift, noise, and stirring rate limit the
precision.  With frequent calibration, measurements
should be reproducible  to ±15%  of the cyanide ion
activity of the sample solution.

temperature effects

Since electrode potentials are affected by changes in
temperature,  samples  and   standardizing  solutions
should be close to the same temperature.  The abso-
lute potential  of the  reference electrode changes
slowly with  temperature  because of changes in the
solubility equilibria on  which the electrode depends.
The slope  of the  cyanide electrode also varies with
temperature,  as indicated by the  factor 2.3 RT/F in
the Nernst equation. Values  of the Nernst factor are
given in Appendix II.

response time

With stirring, typical response varies from several se-
conds  in 10~3 M cyanide ion  solutions to several min-
utes around the limit of detection.  Response is more
rapid when going from dilute to concentrated solu-
tions than in  the other direction.

Deposits which accumulate  on the sensing element
degrade response time and should be removed.

interferences

The electrode will malfunction if ions which form very
insoluble salts of silver are present at sufficiently high
levels to form a layer of silver salt on the membrane sur-
face. In addition, the electrode must not be placed in
strongly reducing  solutions, such  as photographic de-
veloper, which form a layer of silver metal on the elec-
trode membrane.  The electrode can be restored to nor-
mal operation  by wiping off the membrane surface.

-------
                                                    -  52  -
                                                              appendices
Table 1 gives the maximum allowable concentration
of the more common  interfering ions, expressed as
the ratio of the interfering ion concentration to the
sample cyanide concentration.  If  the ratio is less than
that listed  in the table, neither the accuracy of the
measurement nor the surface of  the electrode mem-
brane will  be affected.  The more concentrated the
cyanide solution to be measured, the higher the allow-
able concentration of interfering ions.

Hydrogen ion and many metal ions complex cyanide.
Although the electrode is not affected by these ions,
their presence lowers the level  of free cyanide ion in
the sample.

Sufficiently high concentrations of species which form
extremely  stable complexes with silver (NH3, S2O3=)
also interfere with measurements and result in the
reading of higher values of cyanide ion activity than
actually exist in the sample.  Intermittent exposure
to high levels of these species will not, however, dam-
age the electrode.

table 1

maximum allowable ratio of interfering ion to cyanide
ion (concn. of interfering ion/concn. CN")
Interference
cr
1"
Br"
sr
Maximum Ratio
106
0.1
5x103
must be absent
 To illustrate the use of Table 1: What is the maximum
 allowable Br" concentration at which a 3.3 x 10~s M
 NaCN solution may be measured?

 The maximum allowable ratio of Br" to CN" is given
 in Table 1 as:

         Max. Ratio = 5x103

 Therefore:

         Max. Br" = 5 x 103 (3.3 x 10"s) M Br

                  = 1.6x10"'  M Br"
appendix I

estimating total ionic strength of mixed electrolytes

The total ionic strength of a solution is given by the
equation:
Ionic Strength  = 1/
                                     ci
that is, the concentration of each and every ion in the
solution (Cj) is multiplied by the corresponding square
of the charge on the ion (Z--,), and the sum of these
products is divided by 2.

For example, a solution of the following composition:

        0.05 moles per liter NaCN
        0.20 moles per liter K2S04
        0.15 moles per liter NaOH

has a total ionic strength calculated as follows:
 ion
Na*
K*
S04=
OH"
CN"


1
1
4
1
1

Ionic
x
x
x
x
x

Strength
0.05 + 0.15
0.20 + 0.20
0.20
0.15
0.05
total =
= L|0 = o.80 M
0.20
0.40
0.80
0.15
0.05
1.60

 appendix II
 values of 2.3 RT/F in millivolts
t(°C)
0
10
20
25
2.3 RT/F
54.20
56.18
58.16
59.16
t(°C)
30
40
50
100
2.3 RT/F
60.15
62.13
64.11
74.04

-------
                 - 53 -
specifications
       activity range   10~2 M to 10~6 M CN
      recommended
     operating range
   temperature range
           pH range
     electrode body

  electrode resistance
       electrode life
       interferences
               size
            storage
10'3 M to 10'* M CN"
0°to100°C
pHOtopHU

acid, base, and solvent resis-
tant epoxy
less tnan 30 megohms

about 200 hours in a 10~3 M
cyanide solution: about 1000
hours in a 10~5 M cyanide
solution

sulf ide must be absent:
iodide level must be one-tenth
the cyanide level

electrode is 6" long, with a
30" cable and a U.S.Standard
electrode connector: longer
cables available on special
order

electrode must be stored dry:
soaking before use is not
required
minimum sample size   10-20 ml. in a 50 ml. beaker

-------
                                 - 54 -
                               APPENDIX D

       DETERMINATION OF AMMONIA IN SULFURIC ACID IMPINGER SOLUTIONS
                     USING A GAS-SENSING ELECTRODE
Introduction and Scope

          This method is applicable for the determination of ammonia in
0.2% sulfuric acid impinger solutions.  A gas sensitive ammonia electrode
is used for the measurement.

Summary of Method

          The impinger solution is treated with sodium hydroxide to
convert ammonium ion to ammonia.  Ammonia is then measured using a gas
sensitive ammonia electrode.

Apparatus

          (1)  Orion 95-10 Gas Sensitive Ammonia Electrode, or equivalent.

          (2)  Orion 90-01 Single Junction Reference Electrode, or
equivalent.

          (3)  Orion 801A digital Ion Analyzer, or equivalent.

Reagents

          (1)  10M sodium hydroxide solution.  Dilute 400 g of 50% sodium
hydroxide solution to 500 ml with deionized water.  Store in a plastic
bottle.

Ammonia Solutions

          Solution 1:   1000 ug NH3/ml.  Prepare by weighing 3.140 g
ammonium chloride and making up to 1  liter with deionized water.

          Solution 2;   100 ug NH3/ml.  Dilute 20 ml of  Solution  (1) to
200 ml with deionized water.

          Solution 3;   50 ug NH3/ml.  Dilute 10 ml of Solution  (1) to
200 ml with deionized water.

          Solution 4;   10 ug NH3/ml.  Dilute 20 ml of Solution  (2) to
200 ml with deionized water.

          Solution 5;   5 ug NH3/ml.   Dilute 20 ml of Solution  (3) to
200 ml with deionized water.

          Solution 6;   1 ug NH3/ml.   Dilute 20 ml of Solution  (4) to
200 ml with deionized water.

-------
                                     - 55 -
          Solution 7;  0.5 yg NH3/ml.  Dilute 20 ml of Solution  (5)  to
200 ml with deionized water.

          Solution 8;  0.1 yg NH3/ml.  Dilate 20 ml of Solution  (6)  to
200 ml with deionized water.

Calibration

          Refer to the manufacturer's instruction manuals for detailed
operating instructions.

          (1)  Connect the electrodes to the meter.  Pipet 20 ml of  the
100, 50, 10, 5, 1, 0.5, and 0.1 yg NH3/ml solutions into plastic containers.
Add 0.3 ml of 10M sodium hydroxide solution just before obtaining potential
readings.  The solutions should be stirred at a slow,  constant rate, and
the readings should be taken after the potential stabilizes.

          (2)  Obtain a calibration curve by plotting the logarithm of
the ammonia concentration vs.  the potential readings.

Procedure

          (1)  Pipet 25 ml of the scrubber solutions into plastic containers,
add 0.3 ml of 10M sodium hydroxide solution, and obtain potential readings
as indicated- above.

Calculation

          (1)  Obtain the ammonia concentration of the scrubber solutions
from the calibration curve.

-------
    -  56  -
APPENDIX E

Fuel Economy
Table E-l
Emission: Fuel Lbs.
Catalyst/
Malfunction
02 Sensor Off I
Idle Correct J
02 Sensor Off ^
Idle Enrichenedj
13% Misfire

100% Rh 17% Rh
0% Pt 83% Pt
3.59
3.63
3.81
- oq 3.80
3'89 3.81
Table E-2
Emission: Fuel Lbs.
Catalyst/
Malfunction
©2 Sensor Offl
Idle Correct /
02 Sensor Off ^
Idle Enrichened j
13% Misfire

100% Rh 17% Rh
0% Pt 83% Pt
3.23
3.52
3.57
o •>/ 3.51
3'74 3.75
Table E-3
Emission: Fuel Lbs.
Catalyst/
Malfunction
02 Sensor Off I
Idle Correct J
02 Sensor Off 1
Idle Enrichenedj
13% Misfire
100% Rh 17% Rh
0% Pt 83% Pt
1.49
1.44
1.58
1.59 !-44
Cycle : FTP
6% Rh 0% Rh
94% Pt 100% Pt
3.58 3.55
3.89 fi
3.81 3'76
3.94
Cycle : SET
6% Rh 0% Rh
94% Pt 100% Pt
3.22 3.36
3.49 . ,n
3.59
3.85
Cycle: Idle
6% Rh 0% Rh
94% Pt 100% Pt
1.57 1.50
1.57
1.53 1<46
1.42

No
Catalyst
3.59
3.74
3.96
3.98

No
Catalyst
3.31
3.38
3.67
3.82

No
Catalyst
1.36
1.45
1.47
1.51
        1.42

-------
 - 57 -
Table E-4
Emission
Catalyst/
Malfunction
02 Sensor Off \
Idle Correct J
02 Sensor Off \
Idle Enrichened j
13% Misfire
Emission
Catalyst/
Malfunction
02 Sensor Off I
Idle Correct J
02 Sensor Off I
Idle Enrichened J
13% Misfire
: Fuel Lbs.
100% Rh 17% Rh
0% Pt 83% Pt
3.86
3.74
3.95
ft ft 1
4'35 4.' 19
Table E-5
: Fuel Lbs.
100% Rh 17% Rh
0% Pt 83% Pt
5.46
5.27
5.56
6 29 6'22
6'29 6.28
Cycle: 64 kph
6% Rh 0% Rh
94% Pt 100% Pt
3.79 3.90
3.99 , 1S
3.78
4.31
Cycle: 80 kph
6% Rh 0% Rh
94% Pt 100% Pt
5.14 5.33
5.62
5.35 ^'^
6.20

No
Catalyst
3.77
4.02
3.91
4.23

No
Catalyst
5.53
5.47
5.33
6.22

-------
                                - 58 -

                              APPENDIX F

        PROCEDURE FOR THE DETERMINATION OF HYDROGEN SULFIDE IN
           AUTOMOTIVE EXHAUST USING THE METHYLENE BLUE METHOD
Reagents

  1.  Acetate Buffer - dissolve 54.9 g of zinc acetate
      and 8.2 g of sodium acetate  (NaC2H302)  in deionized water.  Dilute
      to one litre with deionized  water.

  2.  Amine Solution - dissolve 0.93 g of N,N-dimethyl-p-phenylene
      diamine sulfate in 75  cc of  deionized water.   Cautiously, add
      187 cc of cone. H2S0.4,  cool,  and dilute to  1  litre with  de-
      ionized water.

  3.  Ferric Solution - dissolve 120.6 g of ferric  ammonium sulfate in
      750 cc of deionized water.   Add 27 cc of cone. H2S04,  cool, and
      dilute to 1 litre with deionized water.

  4.  Formaldehyde Solution  - dilute 10 cc  of formaldehyde  (37%) to
      100 cc with deionized  water.


 Standardization and  Calibratipn

   1.  Weigh approximately  8  g of Na2S"9 H20 in a  beaker. Rinse the
       crystals with  deionized water before  dissolution  in a nitrogen
       purged  one  litre volumetric  flask.  Dilute  to the mark with de-
       ionized water.

   2.   Dilute  10  cc of this solution to  one  litre  in another nitrogen
       purged  volumetric with deionized water. This solution contains
       approximately 10 ppm sulfide.

   3.   Using a 5 cc micro buret,  add 0.1,  0.25, 0.5, 1,  2,  3, and 5 cc of
       the sulfide soln to 100 cc volumetric flasks, which  contain 10 cc
       of the acetate buffer.  Rinse down the  added sulfide with approximately
       40 cc of deionized water.   Swirl to insure  mixing.  Prepare a
       blank of the acetate buffer.

   4.   Immediately add 50 cc of  the sulfide  soln to each of  3 Erlenmeyer
       flasks containing 10 cc of the acetate buffer and mix.  Prepare
       two blanks containing the buffer and 50 cc  deionized water.

   5.   Pipet 5 cc of a 0.01N 12  soln to one of the Erlenmeyer flasks.
       Add 10 cc of cone. HC1 and titrate to a starch end point with
       standardized 0.01 N sodium thiosulfate using a 5 cc micro buret.
       Repeat this procedure with the remaining flasks.

-------
                                - 59  -
  6.  Calculate the concentration of the sulfide solution and the total
      micrograms of sulfide added to each calibration flask.

  7.  To the sulfide blank calibration flask, add 1 cc of the formaldehyde
      soln and mix.  Add 10 cc of the amine solution down the side of the
      flask.  Taking care not to introduce air bubbles, swirl the flask
      gently to mix the reagent.  Add 2 cc of the ferric solution and
      swirl an additional 30 seconds.  Dilute to the mark with water and
      place in the dark.  Repeat this procedure with the other flasks.

  8.  After 15 minutes, measure the absorbance on a spectrophotometer
      at 667 nm in 5 cm cells against water.

  9.  Plot the absorbance of each flask against yg sulfide contained on
      linear graph paper.
Note:  The calibration curve may not necessarily follow Beer's Law (i.e.,
       not a straight line).  In that case, plot the best curve between
       the points.
Procedure for Samples

  1.  Samples are collected by passing a measured amount of gas at a
      determined rate through specially designed scrubbers containing
      10 cc of the acetate buffer and 40 cc deionized water.
                                                                     *
  2.  To one of the samples add 1 cc of formaldehyde solution and mix
      well.  Add 10 cc of the amine solution through the inlet tube of
      the scrubber.  Mix the reagent by drawing the samples up and down
      through the inlet tube.  This can be done using a rubber suction
      bulb.  Add 2 cc of the ferric solution through the inlet tube and
      repeat the mixing procedure described for 30 seconds.  Quantitatively,
      transfer the scrubber contents to a 100 cc volumetric flask and
      dilute to the mark.  Set the flask in the dark.  Repeat with the
      remaining samples.

  3.  After 15 minutes, measure the absorbance at 667 nm in 5 cm cells
      against water.

  4.  Read from the calibration curve the total yg of sulfide present
      in the sample.

-------
a. TI 11.1 AND sunin LE
   Hydrogen Cyanide Emissions From a
   Three-Way Catalyst Prototype
                                   TECHNICAL REI'OHT DATA
                                rciiil IniJj'iictiuns on lilt ;nvnv hcjiirc t'<>tiii>lrtiiiK)
  il I'HU 1 N( V
             6. PERFORMING ORGANIZATION CODE
                                                           i. ill cii'it N I •;; ACCI SSION-NO.
             5. RCPORT DATE
              11/77
7. AUTHOR(S)
   Eugene L. Holt, Mary  H.  Keirns
             8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORG \NIZATION NAME AND ADDRESS
   Products Research  Division
   Exxon Research  and Engineering
   P.O. Box 51
   Linden, New Jersey   07036
                                                           10. PROGRAM ELEMENT NO.
             11. CONTRACT/GRANT NO.
                EPA No. 68-03-2485
12. SPONSORING AGENCY NAME AND ADDRESS
   EPA, MSAPC, ECTD
   2565 Plymouth  Rd.
   Ann Arbor, MI   48105
             13. TYPE OF REPORT AND PERIOD COVERED
                Final report
             14. SPONSORING AGENCY CODE
 15. SUPPLEMENTARY NOTES
 16. ABSTRACT
             Four  catalysts, nominally  identical except for Pt/Rh ratio, were tested
 for emissions  of  regulated pollutants  and for HCN, NH3, S04~ and H2S.  Fuel economy
 was also measured.   The tests were conducted over the FTP, CFDS, and  64 and 80 kph
 cruise modes,  using three simulated malfunction modes:  02 sensor disconnected-idle
 adjustment  correct; Q£ sensor disconnected-idle adjustment enriched;  13%  misfire.
 The catalysts  were  prepared by Engelhard Industries under conditions  similar to those
 used in making commercial three-way catalysts.  A prototype Volvo TWC vehicle was the
 test car.   The four catalysts contained respectively:  Rh only, 17% Rh-Pt;  5% Rh-Pt;
 Pt only.   Tests were also run with no  catalyst under the malfunction  modes  and with
 each catalyst  under normal engine conditions.  Of the fifteen possible test combi-
 nations  (four  catalysts and no catalyst x 3 malfunction modes), twelve were chosen to
 be run, with four replicates, giving sixteen runs total.

              No HCN was detected  at the tailpipe under normal operating conditions over
 the FTP.   However,  under the malfunction modes, HCN emissions were  generally found.
 During the FTP and CFDS cycles all of  the catalysts appeared to remove .some HCN from
 the exhaust gas under all malfunction  modes, but during cruise some catalysts under
 certain malfunctions caused a further  increase in tailpipe HCN emissions.  The maximum
 level  observed was about 40 mg/km.	
17.
                                KEY WORDS AND DOCUMENT ANALYSIS
a.
                  DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS  C.  COSATI Held/Group
   Catalytic Converters
   Emissions
   Hydrocyanic Acid
13. DISTRIBUTION STATEMENT

   ReleaseN Unlimited
19. SECURITY CLASS (ThisReport)
                                                                         21. NO. OF PAGES
                                              20. SECURITY CLASS (This page)
                                                                         22. PRICE
EPA Form 2220-1 (9-73)

-------