EPA-600/2-76-096


April 1976
Environmental Protection Technology Series
                         ;CTION m

                STACK  SAMPLING  SYSTEMS  FOR

                      EMISSIONS IN  FLUE GASES
                                sS
                                 J0> sr.
                               I
                               55
                               O

                               v
                 #
                  •&
                  ^
                  in
                  O
                  T
Sis,
'-*>.      ^
                                  ^i PRO^^
                                Office of Rosaarch and Development

-------
               RESEARCH REPORTING SERIES

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

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

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

This report has been reviewed by  the U.S.  Environmental
Protection Agency, and approved for publication.  Approval
does not signify that the contents necessarily reflect the
views and policy of the Agency, nor does mention of trade
names or commercial products constitute endorsement or
recommendation for use.
This docBnfent is available to the public through the National Technical Informa-
tion Service. Springfield. Virginia 22161.

-------
COLLECTION EFFICIENCIES OF STACK SAMPLING SYSTEMS FOR
          VANADIUM EMISSIONS IN FLUE GASES
                        by
          H.  L.  Goldstein  and  C.  W.  Siegmund
        Exxon Research  and Engineering  Company
              Products  Research Division
                    P.  0.  Box  51
              Linden, New  Jersey   07036
              Contract  No.  68-02-1748
                   Project Officer
                   James L.  Cheney
    Stationary Source Emission Measurement Branch
      Environmental Science  Research Laboratory
    Research Triangle Park,  North Carolina  27711
        U.  S.  ENVIRONMENTAL PROTECTION AGENCY
          OFFICE OF RESEARCH AND DEVELOPMENT
      ENVIRONMENTAL SCIENCE RESEARCH LABORATORY
    RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711

-------
                                - ii -
                              DISCLAIMER
           This report has been reviewed by the Environmental Science
Research Laboratory, U. S. Environmental Protection Agency, and approved
for publication.  Approval does not signify that the contents necessarily
reflect the views and policies of the U. S. Environmental Protection
Agency, nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.

-------
                               - Ill -


                              CONTENTS



LIST OF FIGURES	

LIST OF TABLES	   vii

ACKNOWLEDGEMENTS	   ix

1.  INTRODUCTION	    1

2.  SUMMARY	    4

3.  CONCLUSIONS	    9

4.  RECOMMENDATIONS	   10

5.  BACKGROUND	   11

    5.1  Vanadium in the Environment	   11

         5.1.1  Ambient Air Measurements	   11
         5.1.2  Toxicity of Vanadium Compounds	   13
         5.1.3  Natural Occurrence of Vanadium in Fuels	   14

    5.2  Nature of Vanadium in Emissions	   16

         5.2.1  Mechanism of Particulate Formation	   16

6.  EXPERIMENTAL PROGRAM	   21

    6.1  Research Plan	   21
    6.2  Combustion Test Facility	   23

         6.2.1  50 hp Cleaver Brooks Package Boiler	   23

    6.3  Program Design	   27
    6.4  Boiler Tube Deposits	   32
    6.5  Stack Sampling Systems	   33

         6.5.1  Method 5 EPA	   33
         6.5.2  ER&E Collection System	   35

    6.6  Particulate Sampling	   37
    6.7  Particulate Isolation Method - General Procedure	   46
    6.8  Determination of Vanadium in Particulate	   46

-------
                                 -  iv  -
7.  RESULTS AND DISCUSSIONS ......................................    52

    7.1  Basis for Comparing Vanadium Recovery ...................    52
    7.2  Vanadium Material Balance Around the
         Cleaver Brooks Boiler ...................................    52

         7.2.1  Total Vanadium Recovery Using
                All-Venezuelan Resids ............................    52
         7.2.2  Erratic Vanadium Recovery
                Using Light Arab Residuum ........................    56
         7.2.3  Statistical Analysis to Compare
                Vanadium Recovery ................................    59
         7.2.4  Vanadium-Bearing Particulate Inventory ...........    61
         7.2.5  Size Distribution of Vanadium-Bearing
                Particulates .....................................    67

    7 . 3  Sampling System Collection Efficiency ...................    70
    7 . 4  Vanadium Input - Fuel Analysis ..........................    71
    7.5  Vanadium Recovery in the Boiler .........................    75

         7.5.1  Preliminary Tests - Evaluation of
                Boiler Pass Tube Inserts .........................    75
         7.5.2  Factorial Program Results - Vanadium
                Recovery at High Firing Rate .....................    79
         7.5.3  Vanadium Recovery Made at Low Firing Rate ........    81
         7.5.4  Further Consideration of
                Total Vanadium Recovery ..........................    83

    7.6  Recovery of Vanadium in the Stack - Summation ...........    86
    7.7  The Problem of Sulfate Formation
         in Water Impingers ......................................    90

         7.7.1  Sulfuric Acid Formation ..........................    90
         7.7.2  The Nature of Artificial Particulate .............    91

    7.8  Determination of Oxidation State
         of Vanadium Particulate in Flue Gas .....................    94

8 .  REFERENCES [[[    96
APPENDIX 1   STATE OF THE ART REVIEW
             VANADIUM ANALYSIS AND PARTICULATE
             SAMPLING METHODOLGY	    98

APPENDIX 2   LABORATORY ISOLATION OF PARTICULATE
             COLLECTED FROM STACK SAMPLING TRAINS	   116

APPENDIX 3   THE DETERMINATION OF VANADIUM IN PARTICULATES	   122

APPENDIX 4   OPERATIONAL DATA FROM FACTORIAL PROGRAM	   133


-------
No.
                                   - v -

                            LIST OF FIGURES


                                                                    Page
5.1    >10 MICRON PARTICULATE EMISSIONS
           (CENOSPHERES)	   18

5.2    INTERMEDIATE  SIZE RANGE  (1-10  MICRONS)
       PARTICULATE		.....  19

5.3    THE EFFECTS OF TEMPERATURE AND OXYGEN AVAILABILITY
       ON  EQUILIBRIUM COMPOSITION FOR THE REACTION
       2V 0   j—7+    4VOo  + 02        	   20
        2. j    ~"

5.4    VAPOUR PRESSURES  OF FUEL  OIL ASH CONSTITUENTS	   20

6-l    THE FOUR PASS CONSTRUCTION OF A TYPICAL C-B BOILER	   24

6.2    50 HP CLEAVER BROOKS BOILER	   25

6.3'    ARRANGEMENT OF FLUE GAS MONITORING SYSTEMS	   26

6.4    LONGER RESIDENCE  TIME FAVORS
       BURNOUT OF CENOSPHERES	   29

6.5    FOUR PASSES OF CLEAVER BROOKS BOILER	   34

6.6    EPA METHOD 5 SAMPLING TRAIN	   35

6. 7    ER&E  STACK SAMPLING TRAIN	   36

6.8    OVERALL VIEW OF BOILER, STACK  SAMPLING SYSTEMS
       AND CONTROL MODULES	   38

6-9    ARRANGEMENT FOR SIMULTANEOUS OPERATION OF EPA AND ERE  STACK
       SAMPLING TRAINS.  (POSITION 1) EPA TRAIN LOCATED
       ON LEFT HAND SIDE	   39

6.10   CLOSE-UP VIEW OF  PORT HOLE CONFIGURATION FOR
       SIMULTANEOUS SAMPLING	   40

6.11   OVEN ASSEMBLY OF  ER&E SYSTEM SHOWING ANDERSEN CASCADE
       IMPACTOR, IMPINGER  SYSTEM AND MAKEUP AIR LINE.  GLASS
       TUBE UPSTREAM OF  IMP ACTOR IS CONNECTED TO PROBE	  41

6.12   VOLATILES KNOCKOUT  SYSTEM CONNECTED TO ER&E TRAIN
       FEATURES A DRY ICE ACETONE COOLED CONDENSER	   42

7.1    TOTAL VANADIUM RECOVERED  IN BOILER AND STACK -
       SHORT RESIDENCE CASE	   54

7.2    TOTAL VANADIUM RECOVERED  IN BOILER AND STACK -
       LONG RESIDENCE CASE	   55

-------
           -  vi -





LIST OF FIGURES (CONT'd.)
No.
7.3
7.4
7.5
TOTAL VANADIUM RECOVERED - LIGHT ARAB FUEL 	
VANADIUM RECOVERY IN BOILER -
SHORT RESIDENCE CASE 	
VANADIUM RECOVERY IN BOILER -
LONG RESIDENCE CASE 	
Page
	 58
	 80
	 82

-------
                                  - vii -


                             LIST OF TABLES


No.                                                                  Page

5.1   CONCENTRATIONS OF V IN AMBIENT AIR	   12

5.2   CHARACTERISTICS OF MIDDLE EAST AND
      VENEZUELAN CRUDES AND FUEL OILS	   14

5.3   CONCENTRATIONS OF VANADIUM IN DOMESTIC COALS	   15

6.1   RESIDUAL FUEL OILS USED IN V COLLECTION
      EFFICIENCY STUDY	   27

6.2   HEAVY FUEL OIL INSPECTIONS	   28

6.3   EXPERIMENTAL PROGRAM DESIGN	   31

6.4   HEAT EXCHANGE SURFACE IN THE CLEAVER BROOKS	   32

6.5   INFLUENCE OF COMBUSTION CONDITIONS ON
      BOILER AND STACK TEMPERATURE	   33

6.6   STACK VELOCITY PROFILE FROM BOILER	   43

6.7   LOW LEVEL SOURCE PROFILE	   44

6.8   MASS CONCENTRATION PROFILE IN STACK
      IS UNIFORM	   45

6-9   EFFECT OF ASHING METHOD ON VANADIUM
      ANALYSIS BY ATOMIC ABSORPTION	   48

6.10  COMPARISON OF INSTRUMENTAL METHODS
      OF VANADIUM ANALYSIS	   49

6.11  ELEMENTAL ANALYSIS BY EMISSION SPECTROSCOPY	   51

7.1   AVERAGE  PARTICIPATE AND VANADIUM  INVENTORY
      FOR LIGHT ARAB FUEL	   60

7.2   ANALYSIS OF VARIANCE - TOTAL VANADIUM
      RECOVERY IN BOILER AND STACK	   62

7.3   TOTAL VANADIUM RECOVERY AS AFFECTED BY STACK
      SAMPLING SYSTEM AND COMBUSTION CHAMBER RESIDENCE TIME	   63

7.4   AVERAGE PARTICULATE AND VANADIUM INVENTORY
      HIGH ASH ALL VENEZUELAN FUEL OIL (F!>	   64

-------
                               - viii -

                        LIST OF TABLES (Cont'd.)


No.                                                                 Page

7.5   PARTICULATE COMPOSITION INFLUENCED BY SIZE	   65

7. 6   DISTRIBUTION OF V205 IN PARTICULATE ASH	   66

7.7   AVERAGE PARTICULATE AND VANADIUM INVENTORY
      INTERMEDIATE ASH ALL VENEZUELAN, FUEL OIL (F2)	   68

7-8   RELATIVE SIZE DISTRIBUTION OF STACK PARTICULATE	   69

7.9   SAMPLING SYSTEM COLLECTION EFFICIENCY	   72

7.10  DRUM TO DRUM VARIATION OF VANADIUM ANALYSIS	   73

7.11  VARIATION OF SAMPLES WITHIN A DRUM OF
      ALL-VEN FUEL OIL	   74

7.12  VARIATION OF PARTICULATE DISTRIBUTION
      IN BOILER (PRELIMINARY TESTS)	   76

7.13  VARIATION OF VANADIUM DISTRIBUTION IN BOILER -
      PRELIMINARY TESTS	   78

7.14  PARTIAL ELEMENTAL COMPOSITION OF BOILER SOLIDS	   78

7.15  PARTICULATE AND VANADIUM INVENTORY IN BOILER                   84
      (SHORT COMBUSTION CHAMBER RESIDENT TIME -
       2 LB./MIN. F. R. )	

7.16  PARTICULATE AND VANADIUM INVENTORY IN BOILER
      (LONG COMBUSTION CHAMBER RESIDENCE TIME -
      0.5 LBS./MIN. F. R. )	   85

7.17  VANADIUM RECOVERY IN THE BOILER	   87

7.18  RECOVERY OF VANADIUM IN STACK SAMPLING SYSTEMS	   88

7.19  COMPARISON OF PARTICULATE SULFATE/
      SULFURIC ACID INVENTORY	  92

7.20  ARTIFICIAL PARTICULATE FORMATION IN
      WATER IMPINGERS OF METHOD 5	  93

7.21  COMPOSITION OF SOLIDS FROM METHOD 5
      WATER IMPINGERS	   95

-------
                           ACKNOWLEDGEMENTS

The authors wish to express their appreciation to Messrs.  J. Singleton
and J. DeHoff for performing the laboratory work described in this
report.

We also wish to acknowledge the efforts of the personnel of the
Analytical Division who were responsible for developing the analytical
methodology and carrying out the vanadium analysis,particularly
Drs. R. A. Hofstader, P. R. Gaines and M. C. Grandolfo.
                                - ix -

-------
                          1.   INTRODUCTION
          Studies of vanadium concentration in ambient air have
shown a strong link between residual fuel oil combustion and high
vanadium levels.   There is also growing evidence which suggests that
the major part of the vanadium emissions from stationary combustion
sources are concentrated in the fine sized particles (1»2).  As there
are no direct reading instruments for measuring vanadium in stack
gases, emission factors for this element must be based on conventional
particulate sampling methods, i.e.  collecting dust or fumes on filters,
electrostatic precipitators and/or impingers.  These methods have
in common several potential drawbacks:

     (1)  the possibility that fine particles may not be collected
          very effectively and therefore pass through the sampling
          device.

     (2)  the possibility that volatile vanadium compounds will be
          lost from the sample by continuously drawing flue gas over
          the particulate.

          The study performed for this contract was to compare experimentally
the effectiveness of the U. S. Environmental Protection Agency's (EPA)
standard Method 5 particulate sampling train and a stack sampling
system developed by Exxon Research and Engineering Company (ER&E)
for collecting vanadium emissions in flue gas from heavy fuel oil
combustion.  The ER&E system overcomes some of the major shortcomings
traditionally associated with stack samplers more commonly used.  It
permits the determination of size distribution of the particulate down
to about Q.4 ym and it collects finer particulate in several impingers
containing a liquid medium which can operate at stack temperatures.  -This
avoids any possibility that S02 will oxidize to sulfate in the impingers
which would create false particulates.

          To assess collection efficiency of the two sampling trains
in a real environment under realistic but closely controllable
conditions, an actual boiler was used.  The unit consisted of a completely
instrumented 50 hp, four pass package boiler manufactured by Cleaver
Brooks.  For particulate sampling, a platform mounted above the boiler
allowed access to two 3" ports which had been drilled at right angles
to each other in the 12" diameter boiler stack.  Using this arrangement,
both sampling trains were operated simultaneously during the tests.

          The package boiler burns heavy fuel oil under combustion
conditions typical of those in a field installation.  The gases moving
through the system are cooled in a realistic manner so particulate burn-out
and condensation from the vapor phase is realistic.  Thus, this boiler is
the best equipment of reasonable size for generating particulate
approximating those which will be encountered in the field.  A large boiler
would be too cumbersome in terms of fuel requirements and operating
difficulties to carry out such a study under closely controlled conditions.

-------
                                  - 2 -
          In order to establish the effectiveness of the collection
systems, i.e. their ability to inventory all the vanadium particulate
entering the stack, a vanadium material balance was carried out
around the boiler.  In essence by knowing the vanadium  content of
the fuel which is burned and determining the amount of vanadium deposited
in the boiler, it is possible to establish quantitatively how much of
the vanadium particulate is carried into the stack with the flue gas.
Since the Cleaver Brooks is a four pass boiler, that is, the combustion
products pass the length of the boiler four times before entering the
stack, vanadium deposits in each of the passes must be carefully inventoried.
Ideally, the firetube, which constitutes pass 1, and the 32 heat
exchanger tubes making up the 3 other passes would be completely cleaned
out and the amount of vanadium collected would be measured after each
run.  However, cleaning all the tubes after each run would be extremely
time consuming and tedious.  To avoid this and at the same time provide
a high degree of accuracy, two methods for measuring the amount of
vanadium deposition in the boiler were originally proposed.  The first
involved isokinetically sampling the combustion gases at the end of
the firetube.  The second involved collecting fallout in stainless steel
liners placed in selected heat exchanger tubes in the passes.  This
latter was selected by the EPA as the most practicable technique.

          The amount of vanadium entering the boiler during a given run
was determined by knowing the vanadium content of the fuel and the fuel
firing rate.  This was compared with the vanadium output from a run
consisting of:  the amount measured in the flue gas leaving the boiler
using the sampling systems mentioned previously plus, the amount of
vanadium collected in the boiler tubes and passages.  In order to
measure the latter without going through a long and tedious boiler
clean-out after each run, tube liners were used.  These were stainless
steel inserts which were machined to fit snugly inside certain of the
32 heat exchanger tubes in the four passes.  At the end of each run,
these liners were removed and the particulates in them were weighed
and analyzed for vanadium.  The vanadium collected in each insert
was considered to be typical of the amount collected in the other tubes
in the same pass.

          The difference between the vanadium entering the boiler with
the fuel, that accounted for in boiler deposits and that collected
in the sampling systems gave a measure of the inefficiency of the sampling
system and/or inaccuracies in the other measurements involved.

-------
                                 - 3 -
          In order to determine how accurate the sampling systems
would be over a range of vanadium particulate levels in the flue gas
and particle size distributions, runs were made on several fuel oils
over a range of operating conditions.  The fuels were chosen on the
basis that they were typical of those in use now and in the future in
the U. S. and that their vanadium contents would span the range which
would normally be encountered in U. S. fuel oils.

          The primary variable used to adjust combustion conditions
was residence time in the combustion zone.  Previous (company sponsored)
studies by Exxon Research and Engineering Company'-^ had shown that
increasing residence time decreased total particulate level by
burning out carbonaceous particles and shifted size distribution
towards smaller particles as the large ones burned out.  It was
felt that if size distribution had a significant influence on the
effectiveness of the sampling systems, it would be primarily because
of their inability to collect very small particles quantitatively.

          The basic information being sought in this study was (1) what
percent of the fuel vanadium was accounted for by the collection systems
after correcting for boiler deposits, (2) whether there was a significant
difference between the effectiveness of the two systems, and (3) how
much the effectiveness changed as vanadium content and size distribution
was varied.

          Ultimately, four variables were involved in this investiga-
tion viz., sampling system, fuel oil, combustion chamber residence time
and sample location in the stack.  The latter was selected to rule out
the possibility that particulate maldistribution in the flue gas would
produce a bias in the results.  The two sample systems were tested si-
multaneously in each run using sample ports located 90° apart.  A sys-
tematic maldistribution might have caused one or the other port to give
higher vanadium recovery.  A factorial design of the type 2x2x2x3
was used to study the responses of the variables and also provide a frame-
work for analysis and interpretations of the results.

          While the factorial program constituted our major effort, some
limited experiments were conducted to determine the oxidation state of
vanadium particles collected on the filter of the Method 5 sampling train.
The methodology consisted of exposing several vanadium oxides to hot flue
gas from combustion of a conventional high sulfur fuel oil and then check-
ing for possible changes in the oxidation state.

          Prior to the start of the factorial program, a literature survey
to assess methods of sampling and analyses of vanadium emissions from
combustion sources was made.  Sources which were searched included the
American Petroleum Institute, Literature and Patents files from 1964 through
1975, NTIS (Government Reports) 1964 through current, and Chemical Abstracts
1970-1975.  A bibliography covering this state-of-the-art review is pre-
sented in the first Appendix.

-------
                                - 4 -
                             2.  SUMMARY
          Exxon Research and Engineering Company (ER&E) has conducted
an experimental study to measure the efficiency of a standard EPA
Method 5 stack sampling train to collect vanadium emissions in flue
gas and compare it with an ER&E developed stack sampling system.
The ER&E system employs an Andersen Cascade Impactor and several
high velocity impingers filled with silicone oil.  The impactor gives
size distribution down to approximately 0.4 microns, finer material
is collected in the impingers which now can operate at stack
temperature and, therefore do not produce false sulfate type particulate.
The stack solids are not soluble in silicone oil so separation of this
material is easily accomplished.  In these tests the entire sampling
system was maintained at 400°F.  The full EPA Method 5 sampling train
consisting of a probe, a cyclone, a fiber glass filter (MSA-1106BH)
and two water impingers were also employed.  Sampling was accomplished
using the published EPA methodology (4) with the exception that the probe,
cyclone and filter were heated uniformly to 400°F rather than the
prescribed 225-250°F.

          The ability of these sampling systems to collect vanadium
emissions was measured in a series of runs on particulates generated
by combustion of typical residual fuel oils in a 50 hp four pass
package boiler modified to simulate boilers of a variety of sizes.
This latter was accomplished by changing combustion chamber residence
time.  Experimentally, this was carried out by varying the fuel oil
firing rate and/or adding an additional section of refractory to
the existing combustion chamber.  With these procedures residence time
could be changed almost 7 fold from roughly 50 msec to 350 msec.  In
comparison the combustion zone of a large utility boiler may provide a
residence time of more than 1 second.

          Variables in this study were selected on the basis that they
might affect the amount and size distribution of vanadium in the
flue gas.  Two levels of combustion chamber residence time and three
residual fuel oil compositions, which would normally be encountered
in the U. S., were chosen.  Included were two all-Venezuelan resids
with vanadium contents of 359 ppm and 149 ppm respectively and a light
Arab resid with a vanadium content of 39 ppm.

          An important part of the experimental procedure was the
simultaneous operation of both particulate sampling trains during each
run.  This served to eliminate sampling bias attributable to fluctuations
in boiler operation.  To accomplish simultaneous sample withdrawal, two
sampling ports located 90 degrees apart in the stack were used.  As
a further precaution, to safeguard against possible maldistribution of
particulate in the stack, the sampling trains were systematically
rotated between sampling port locations.  Thus half of the test program
was carried out with a sampling train at one port and the other half
with the sampling train at the second port.   The selection of  sampling
port location was accomplished using a restricted randomization.

-------
                                  -  5-
           Ultimately, four variables were involved in this study:
sampling system, fuel oil, combustion chamber residence time and
sample port location in the stack.  To evaluate these factors and provide
a framework for statistical analysis, a factorial experimental design
of the type 2x2x2x3 was employed.  A total of 36 experiments were
made consisting of 12 base case runs each with two levels of replications.

           The key to evaluating sampling system collection efficiency
was a vanadium balance in the boiler.   B7 knowing the vanadium
content of the fuel which is burned and by determining the amount of
vanadium that deposits in the boiler, it is possible to establish
how much vanadium enters the stack.  The amount not accounted for in
the sampling train represents a combination of collection inefficiency
and/or experimental error in measurement and other inaccuracies.
Vanadium analysis of the boiler and stack solids was accomplished using
Atomic Absorption Spectroscopy.

           The amount of vanadium deposited within the boiler was
determined by using removable stainless steel liners in one of the
heat exchanger tubes in each pass of the boiler.  At the end of each
run, the tubes were removed and the deposits in each were collected
and analyzed for vanadium.  Total deposits were estimated by assuming
that the vanadium deposited in a liner was equal to the vanadium deposited
in each of the other tubes in the same pass.  Fire-tube deposits were
measured by cleaning the fire-tube after each run.  Any vanadium
not accounted for in the boiler was assumed to have entered the stack
with the flue gas.

          The flue gas in the stack was sampled  isokinetically using
either Method 5 or ER&E systems.  Particulate samples  from  the probe,
cyclone, filter and impingers of  the EPA system were measured separately
and analyzed for vanadium.  Particulate samples  from the ER&E system
were combined into size fractions: >10 microns,  1  to 10 microns and
<1 micron.  Each fraction was analyzed for vanadium.

           For each run the material balance obtained with the EPA
Method 5 system was compared with that obtained with the ER&E system.
In addition, the material balances between runs were compared to
determine the influences of fuel type and vanadium content, combustion
residence time and sampling position within the stack.   These comparisons
were made using analysis of variance.

           It was determined that vanadium recovery was dependent on
only two variables, combustion chamber residence time and sampling
system; fuel type, i.e. amount of vanadium in the fuel, and sampling
position in the stack were not statistically significant.

-------
                                 - 6 -
           Results obtained with the light Arab fuel oil were very
erratic and indicated a high degree of experimental error.  This
was attributed to the very low V content in the fuel and consequently
in the particulate.  Vanadium recovery with either sampling train
averaged in excess of 100% and ranged between 70% and 174% of the fuel
V input.  Tests made with this particular fuel could not be used
to measure collection system efficiency.  In contrast, results
obtained using the two all-Ven (Venezuelan) fuels were very consistent
and therefore served as the basis for the evaluation.

           At short combustion chamber residence time (high fuel
firing rate), particulate distribution was predominantly in the
coarse size range, total vanadium recovery (boiler plus stack) was
highest using the EPA Method 5 system and averaged 88.6 wt. % of the
fuel V input.  Based on-the number and potential inaccuracies of the
measurements used to make the overall vanadium material balance (i.e.
vanadium deposition in the boiler headers is not accounted for) this
value represents nearly quantitative recovery.  In contrast, total
V recovery using the ER&E system was lower by 7.7% and averaged
81.8 wt. %.  This difference was statistically significant (90 percent
confidence limit) but still within acceptable limits.

           At the short residence time, recovery of vanadium in
the fire-tube and boiler passes averaged 29.3 wt. % of the fuel V input
or about 34% of the total recovered.

           In  tests made at  long  residence  time   (low fuel  firing  rate),
which by comparison  is  about  1/3 less  than  found in a large  utility
boiler, the  particulate distribution was  shifted to the  submicron size
range.  This  resulted  in  a  significant  decrease  in  total vanadium
recovery  (boiler  plus  stack sampling systems) averaging  about  21%.   This
gross change  in vanadium  recovery,  however,  is not  fully ascribed to
collection inefficiency.  Part,  if  not  all  of  the problem,  may  be
caused  by  the  failure  to  inventory  completely  the vanadium which
deposits in  the boiler.   At the  firing  rate  used in these tests,  the
volumetric flow rate of the combustion  product  gas  stream is very low
and may produce some maldistribution as it  goes  through  the  boiler
passes.  Since our estimate of  the  quantity  of vanadium  sampled in
the stack  is  predicated on  the  amount  recovered  in the boiler,  it is
evident that  an error  here  would distort  the recovery value  in  the
stack.  Vanadium  recovery in  the boiler for  this case amounted  to 22.7
wt. % of the  fuel V  input,  which represents  a decrease of 23%  compared
to  the  previous  tests  made  at short residence time.  On  this basis
total vanadium recovery using Method 5  averaged  73.9 wt.  % versus
62.3 wt. % obtained  with  the  ER&E  system.


           The difference in  recovery between the two sampling  systems
is  real and  significant at  long  residence  times.  This indicates  that
V collection  efficiency in  the  ER&E sampling system is lower and  is
primarily  due  to  the inability  of  the  silicone oil  impingers to retain

-------
 very  fine  vanadium  particles.   In  contrast,  the  fiber  glass  filter
 used  in  the  Method  5  system  has a  lower  particle size  cut-off  limit
 i.e.,  it retains  more of  the very  fine vanadium  particles  than do
 the silicone oil  impingers.

           The inefficiency of  the  high velocity  impingers  to collect
 the very fine submicron particles  was general  and not  confined to
 the particular collection medium.   In Method 5 several water filled
 impingers  were used as back-up for the filter.  Essentially  none
 of  the vanadium was found in these impingers even at long  residence
 time  where  the data  suggests that fine vanadium particles were passing
 through  the  filter.

           Total particulate  emissions which were also measured with
 each  sampling system  were in reasonable  agreement.  This was taken
 as  a  further indication that only fine highly  concentrated vanadium-
 bearing  particulate were  being lost through the  silicone oil impingers.
 At  short residence  time,  the two stack  sampling  trains had total
 particulate  inventories less than 1% apart. At  long residence time
 stack loading was significantly reduced  and total particulate
 inventories  differed  by about 13%.  The  higher inventory was obtained
 with  Method  5 and was consistent with  the higher vanadium assay.


           A variable and highly erratic amount of solids isolated
 from the chilled water  impingers of the Method 5  train were not
 included in  the total particulate  inventory.   These solids were composed
 predominantly  of ammonium sulfate  and ammonium bisulfate and were
 therefore  regarded  as artificial particulate,  i.e. an artifact of the
 chilled water  impinger  system only.  The origin of ammonia in  the flue
 gas is not known.

           Vanadium collection efficiency of the sampling systems as
 defined in this study is the ratio of the amount of V collected in the
 respective stack sampling train to the amount of V in the fuel minus
 the amount deposited  in the boiler/   Vs    \  .   This presupposes
                                   ( V]7-FB  J
                                   \^       S
 that all losses of vanadium occur  from the stack sampling system.  Since
 there is a   finite possibility that some of these losses are due
 to inaccuracies in  the boiler measurement, the  calculated collection
 efficiency may be taken as representing only a  minimum value.  In
 addition, because of  the lower base used  in these calculations, collection
 efficiency is always lower than the corresponding vanadium recovery.

           At short combustion chamber  residence  time, typical  of a
 small  boiler, the EPA Method  5 stack sampling train had a minimum
vanadium collection efficiency of 83.9%.   The ER&E train operated

-------
                                 -  8 -
at the same condition had a collection efficiency of 74.3%.  At
long combustion chamber residence time, typical of a large boiler,
where V emissions are mostly in the submicron size range, collection
efficiency of the EPA Method 5 system apparently decreased.  Collection
efficiency amounted to 66.2% compared to 51.1% obtained with the
ER&E system.

           Several experiments were also conducted to determine
the oxidation state of vanadium emissions collected on the Method 5
fiber glass filter.  Using selected vanadium oxides and exposing them
to flue gas at elevated temperature, it was found that some
vanadium (V) was partially reduced to vanadium (IV).  Although this
is by no means definitive, the results do suggest the possibility
that both V ** and V" emissions may be present in the flue gas.

-------
                                 - 9 -
                           3.  CONCLUSIONS
          On the basis of this completed factorial study to evaluate
and compare collection efficiency of Method 5 and ER&E type stack
sampling systems, the following conclusions can be made:

     1.  Both EPA Method 5 and ER&E sampling systems show good
         vanadium recovery at short residence time.

     2.  Recovery in both sampling systems decreases at long residence
         time but the decrease was significantly greater in the ER&E
         system.

     3.  The EPA Method 5 sampling system generally produced a better
         vanadium material balance because it collected more of the
         fine vanadium particles.

     4.  The Method 5 system collects no fine vanadium particles in
         the water impingers.

     5.  The fiber glass filter used in Method 5 has a lower fine particle
         cut-off size limit than the high velocity Greenberg-Smith type
         impingers containing the silicone oil or water.

     6.  The low  vanadium content of the light Arab fuel oil made it difficult
         to obtain a reliable vanadium material balance around the
         boiler.

     7.  Total particulate inventories measured by ER&E and Method 5
         (front half) sampling trains at 400°F are similar and within
         experimental error.

     8.  The back half of the EPA system (water impingers) collects a
         highly variable quantity of artificial particulate matter.

     9.  Oxidation states of vanadium emissions in the flue gas may
         include V1"^ as well as

-------
                                 - 10 -
               4.  RECOMMENDATION FOR FURTHER TESTING
          The vanadium material balance measured with the Method 5
train at short residence time probably represents nearly quantitative
recovery.  The main inaccuracy appears to be associated with the
boiler inventory.  For example, the amount of particulate fallout
occurring at the headers at either end of the boiler is not known.
This represents about 10% of the surface area of the boiler and
therefore could contribute to the error noted in these measurements.

          At long combustion chamber residence time, there is some
uncertainty as to how accurate vanadium collection is.  The probability
that at the low firing rate the combustion gas stream is maldistributed
in the boiler passes must rank high.  To obtain a better vanadium
material balance, i.e., to determine where the losses are, additional
efforts may be directed either to improving the collection in the
stack or in the boiler.  Since the stack sampling train results
suggest that very fine vanadium particles are being lost, the chances
of collecting them with existing equipment is slight.  Therefore
primary emphasis in any further testing should be initially placed
on improving the vanadium material balance in the boiler.

          Vanadium collection in the boiler may be accomplished in
several additional tests by two techniques.  The first consists of
making a complete collection and inventory of all material which
deposits in the boiler after a run.  While this will be tedious and
quite time consuming, it does provide the highest reliability and
accuracy.  The second employs isokinetic sampling of the fire-tube.
While this requires less time to accomplish it, it may also be less
accurate particularly if the combustion gases in the fire-tube are
maldistributed.  Both of these techniques, however, can provide a
better and potentially more accurate measure of vanadium particle
fallout in the boiler than the single tube insert which was used.

-------
                                   - 11 -


                             5.   BACKGROUND

5.1  Vanadium in the Environment

     5.1.1  Ambient Air Measurements

          A number of ambient air particulate sampling studies have determined
the concentration of specific elements in the collected particulate samples.
These surveys have shown, among other things, that the suspended particulate
from certain urban areas have significantly higher concentrations of vanadium
(V) than corresponding samples  from rural areas,  or for that matter from
certain other urban areas. ( 5,6,7,8)  The highest  concentrations, either as
fraction of total particulate or  yg/m^ in ambient air, have been found in
cities along the East Coast of  the U.S.  Measurements in New York City during
the 1950's showed annual averages in the range of 2.5 yg/m^ . (9)  Data gathered
in 1967 showed that a number of sampling locations in the New York-New Jersey-
Connecticut area averaged about 0.5 yg/m  on an annual basis• (8)  Other East
Coast sampling locations averaged 0.1-0.2 vig/m3.This is shown in Table 5.1.

          By wav of contrast urban areas in the south and midwest averaged
.01 to .05 Mg/m  of V in ambient air.This level was only a little higher
than the V concentrations in rural areas of the same portions of the U.S.

          Information of this type on ambient concentration of V particulates
has led a number of investigators to the conclusion that the primary source
of vanadium in ambient air is fuel oil combustion.  Most of the fuel oil
used in the United States is burned in the major  metropolitan areas along
the East Coast, where the highest ambient V concentrations were found.
During the period in which much of the data was obtained the fuel oil used
was almost exclusively of Venezuelan origin which meant that it had rela-
tively high V content — about 200-400 ppm by weight — compared with fuel
oils from U.S. or Middle East fuels which usually contain well under 100
ppm V.  About 1970 fuel sulfur  regulations went into effect along the East
Coast, forcing a drastic change in fuel oil composition.  A comparison of
National Air Sampling Network data on ambient V concentration in 1967 and
1970 for Philadelphia and Washington, D.C. indicates a reduction in level
which is consistent with the lower V content of the low S fuel oil, about
                                    196/8)           19?0 (10)

          Philadelphia              .264              .14
          Washington, D.C.          .165              .09

50-100 ppm, rather than 200-400 ppm.  Many investigators have noted the
seasonal variation of ambient V in the areas of high concentration.(9,11}  jt ^
highest during the winter months when large quantities of fuel oil are burned
for space heating.

          The relatively low ambient concentrations of V in midwestern urban
areas, indicate that coal combustion is probably  not a significant source
of V particulates.  These areas are as highly industrialized as those on
the East Coast, but they use coal as the major fuel.  Although many coals
contain a certain amount of vanadium, its concentration is low compared with
total ash content.  Evidently most of the V is retained with the other ash

-------
                                  - 12 -
                                TABLE 5.1
                  CONCENTRATIONS OF V IN AMBIENT AIR
    YEAR
     1967
                   1970
New York, N. Y.
Bayonne, N. J.
Philadelphia, Pa.
New Haven, Conn.
Washington, D. C.

Chicago, 111.
E. St. Louis, 111.
Hammond, Ind.
Montgomery, Ala.
St. Louis, Mo.
Pittsburgh, Pa.
Akron, Ohio

Phoenix, Ariz.
Denver, Cola.
Dubuque, Iowa
Monroe State Forest, Ind.
Rio Arriba County, N.M.
Annual
Avg.
905
445
264
49
165
Ug/m3
Max.
1.4
.99
.43
.74
.23
                                                      4th Quarter   Annual Avg
.06
.006
.034
.0033
.0116
.016
.0063

.0037
.0034
.0062
.0013
.0014
.10
.007
.054
.0033
.025
.034
.013

.0042
.0034
.0062
.0017
.0014
.12

.09

.07



.03
.14

.09

.06

-------
                                   - 13 -
constituents in the electrostatic precipitators or other particulate collec-
tion devices on coal fired combustors.

          In areas where low ambient concentrations of V have been found,
soil is considered to be the most likely source.   On the basis of V/A1
ratio Marten et al^-^) concluded that soil dust was responsible for a
major portion of the V found in large particles in the San Francisco Bay
area.  They also concluded that fuel oil combustion sources were responsible
for V in smaller particles, 
-------
                                   - 14 -
          Animal exposures to concentrations of 8,000 to 18,000 yg/m  of
 ^0  for 2 hours daily for a period of 9 to 12 months resulted in acute and
 nronic poisoning.' '
          These studies and episodes were all related to airborne concentra-
tions which were order of magnitudes higher than those encountered in ambient
air.  There is no clear cut evidence that exposure to airborne particulate
vanadium compounds at the concentrations normally encountered in ambient air
has any acute toxic effects.  However, exposures to concentrations of 20 to
900 Pg/rn^ have been related to reduced cholesterol concentrations.
                                              (13)
          In a statistical study, Hickey et al     found a moderate correla-
tion between ambient vanadium concentration and "diseases of the heart" in
25 communities in the U.S.  They also found high correlations with bronchitis,
pneumonia and lung cancer.  However, there was no way of establishing a cause
and effect relationship in these studies.

          The National Academy of Science has recently completed a study of the
health effects of ambient concentrations of V particulates.  They concluded
that there is no health hazard at the concentrations normally encountered in
ambient air.  There may be a health hazard associated with high localized
concentrations which might be encountered in industries which form or use
vanadium compounds, i.e. manufacture of high alloy steel or the suppliers
of the vanadium for this use.

     5.1.3  Natural Occurrence of Vanadium in Fuels

          Vanadium is one of the most common inorganic ash constituents of
crude petroleum.  Its concentration varies widely according to crude source.
North African crudes from Libya and Nigeria are generally very low in V
content about 1 to 5 ppm.  Middle East crudes tend to be intermediate —
25 to 50 ppm and Venezuelan crudes tend to be high — 100 to 500 ppm V.
Table 5.2 shows the concentration of V in typical Middle East and Venezuelan
crudes and fuel oils which have been used in U.S. fuel oils.

                                 TABLE 5.2
                   Characteristics of Middle East and
                   Venezuelan Crudes and Fuel Oils(14)
                        Lt. Arabian
                               Fuel
                        Crude  Oil
  Kuwait
       Fuel
Crude  Oil
Tia Juana Med    Libyan
        Fuel
 Crude  Oil
       Fuel
Crude  Oil
Gravity, °API
Sulfur, Wt.%
Pour Point, °F
Viscosity, SSU @ 100°F
Viscosity, SSF @ 122°F
Vanadium, ppm
Nickel, ppm
34.7
1.7
-55
44
-
13
4
15.
3.
+55
-
175
37
11
5
0





31.4
2.5
-20
56
-
31
7
15.5
5.0
+55
-
175
61
14
26.
1.
-40
116
-
156
20
5
5





16.
2.
+20
-
170
350
31
3
0





39.
0.
+40
40.
-
3
-
2
2

5



22.2
0.4
+105
-
65
10
-

-------
                                  - 15 -
          The vanadium in these crudes is in the form of organo-metallic
compounds which are thought to include many porphorins or porphorin like
compounds.  A small part of the vanadium is in the form of volatile
compounds in the vacuum gas oil boiling range, but by far the most is
non-volatile and is associated with the asphaltene fraction.  A number of
studies have been carried out at Exxon Research and elsewhere in an
to characterize this fraction and the organo-metallics it contains. (15,16,17)
However, because of its extreme complexity very little separation has been
achieved.

          Basically, however, the vanadium containing compounds are known
to be non-volatile, which means that essentially all of the vanadium in
the crude remains in the residual fraction which is normally the basic
component of heavy fuel oil.  Furthermore, it is primarily associated with
the asphaltenes in that residuum.  These materials are the most resistant
to cracking and volatilization during the combustion process.  They form
the carbonaceous residue of the spray droplets during the fuel oil com-
bustion process after the volatiles have burned off.  These residues,
called cenospheres, contain essentially all of the vanadium from the fuel.
As the cenospheres burn out in the combustion zone, the vanadium is oxi-
dized, primarily to V 0   which forms the ash residue.(18)

          Vanadium concentration in U.S. coal is normally in the range of
15-50 ppm.  It is usually in the form of organo-metallic compounds similar
to those found in crude oil.  The relatively low total V content in con-
junction with the high ash content (olO%) of these coals makes vanadium
a minor constituent of the total ash.  As indicated in Table 5.3, V amounts
to .01-.05% of the ash from coal combustion.  As such most of it is removed
from the flue gas in the electrostatic precipitators or other ash collection
devices.  Emission of V particulate from coal firing is only a minor source
of atmospheric V as indicated by the low ambient concentrations in indus-
trialized sections of the Midwest where coal is the primary fuel.

                      TABLE 5.3 (FROM REFERENCE 8)

              Concentrations of Vanadium in Domestic Coals

                                       Vanadium in Ash     Vanadium in Coal
Coal Source                            	%	     	ppm	

Northern Great Plains                    .001 - .058              16
Eastern Interior Region                       —                  35
Appalachian Region                            —                  21
Texas, Colorado, N.Dakota, S.Dakota     0.01 - 0.1
West Virginia                           0.018 - 0.039
Pennsylvania (Anthracite)               0.01 - 0.02
Buck Mountain Bed       .                    0.11                 176
Diamond Bed                                 0.09                  92

-------
                                    - .16 -
5.2  Nature of Vanadium in Emissions

     5.2.1  Mechanism of Particulate Formation

          Fuel oil is sprayed into the combustion zone of a boiler in the
form of 100 - 300 Vtai droplets which are mixed with combustion air as they
enter the flame.  Single droplet studies have given a good picture of what
happens to the droplets in the combustion zone.  As the droplet heats up,
the volatile materials boil off, especially those in the outer layer and
burn in the vapor phase.  This leaves a shell of non-volatile solid material
around the surface of the droplet.  As the droplet temperature increases,
volatile material from the interior of the droplet vaporizes and ruptures
the outer shell, leaving holes in it and forming a solid layer further in.
This process is repeated until all of the volatile material has vaporized
and as much of the non-volatile material as will crack into volatile hydro-
carbons has done so.  This leaves a solid, carbonaceous residue roughly the
size of the original droplet but containing a high percentage of void volume.
This residue is called a cenosphere.  Most of the material in the cenosphere
is the asphaltene fraction of the fuel oil.  We have found that the asphaltene
content of a fuel oil is a very accurate predictor of the amount of cenosphere
which will be formed during combustion, and of the total weight of stack solids
which will be emitted at a given set of combustion conditions.  Figure 5.. 1
shows electron microscope pictures of cenospheres which are >10ym and Figure
5.2 shows partially burned out material in the 1-lOym.
          Since essentially all of the vanadium in the fuel is contained in
the asphaltenes, it will also remain in the cenospheres.  When the ceno-
spheres form, most of the vanadium is still in the form of organo-metallic,
porphoryn-like compounds.  As the cenospheres burn out, the vanadium is
converted into the oxide form.  No studies are available to tell what
oxide is present in the partially burned out cenospheres, but after complete
burn out, V-0,. is formed at combustion conditions.  Almost all of the ceno-
spheres do burn out before they leave the combustion zone.  The asphaltene
content of a typical high S (2.2%s) Venezuelan fuel oil is 10-12%, whereas
total particulate emissions from a well run boiler are less than 0.1-0.2
wt.% on fuel.
          When the carbonaceous portion of the cenosphere burns out, the V
in the ash residue tends to form V_0,-.  Figure 5.3, which is from reference
19, shows the equilibrium between V20, and Vo^V  Increasing 0- content favors
the formation of V 0  but increasing temperature favors V«0  formation.  At
normal combustion conditions significant V 0_ formation is savored.  Under
combustion zone conditions, V 0  has a significant vapor pressure, as shown
in Figure 5.4, also from reference  19.  It is high enough so that all of
the V present even in the highest ash fuel oil can be vaporized.  As the
combustion gases move through the heat exchanger sections of the boiler,
they are cooled and the vapor pressure of V 0  drops.  Particulate V 0
forms.  The bulk of it is in the very small size range because of the rapid
cooling rate.  When the gases leave the boiler at 5-600°F or less, the
vapor pressure of V 0,.  is  low enough so  that only a negligible amount is
in  the vapor phase,   vanadyl chloride VC1, which boils at 300°F has a vapor
pressure,  and would be  in  the vapor phase in the flue gas.  However, no
one has reported any  formation of VC1, in the combustion of fuel oil.

-------
                                   - 17 -
          Thus vanadium particulates in the flue gases leaving a boiler
are distributed over the entire range of sizes.  The cenospheres which
are 10-50ym in diameter contain vanadium in the carbonaceous matrix;
partially burned out cenospheres in the l-10ym range contain a higher
percentage of V; and submicron particles of V?0,. are formed by condensation
from the vapor phase.

-------
                                   FIGURE 5.1

                        >10 MICRON PARTICULATE EMISSIONS
                       	(CENOSPHERES)	
10 Microns
                                                                                                       00

-------
                       - 19 -
                    FIGURE 5.2




INTERMEDIATE SIZE RANGE (1-10 MICRONS) PARTICULATE




                      1.0 Microns

-------
               50
             o
             630
             en
            0*20
            O
            <
             >-,
             en
             o
               -10
^•/". .    ^ — ~Z_	i —*-<-__Excess a\r
/  — ~^£- ~- ~~   ^sf IV.Excess a\r

 "' *?         ^"^ '
    ./__ __ [—,x-5V. Excess a\r
                                                                                                                                                             woo
                 OV.V204
80
20
60
AO
40
60
20
60
  OV.V205
100V. V204
Fig.  5.3     .  The cffi-cls uf toinpcr.tturc and oxygen availability on equilibrium
                                 fur ilii; reaction 2V2O. -,- -VVOj + Oj
                                                                                       700     800    900  1.000  1.100    1.300  1.500
                                                                                                      Temperature. °C

                                                                                        J . 4>   Vapour pressures of fuel oil ash constituents
                                                                                                                                                                            N)
                                                                                                                                                                            O

-------
                                  -  21 -


                        6.  EXPERIMENTAL PROGRAM

 6.1  Research Plan

          This section discusses combustion facilities, the design of the
experimental program, test methodology, analytical procedures used for
evaluating vanadium collection efficiency of the EPA Method 5 and ER&E
stack sampling trains.  The general approach taken in this investigation
consisted of three phases:

     1.  State-of-the-Art-Review:  An initial literature review was
made to assess stack sampling and analytical methodology for collecting
and analyzing vanadium-containing particulate (both gaseous and solids)
from fossil fuel combustion sources.  The survey indicated that Atomic
Absorption Spectroscopy followed by Arc Emission Spectroscopy are the two
most precise and widely used techniques to analyze for vanadium-containing
particulate (both gaseous and solids) from fossil fuel combustion sources.
Collection of particulates in flue gas has been carried out with an ever-
increasing number of devices.  However, there is nothing in the literature
to indicate that the standard fiber glass filter employed in the Method 5
sampling train is less efficient than other collectors.  For particulate
sizing, the Cascade Impactor, in spite of its many "stated problems",
still remains the method employed by most investigators.

     2.  Develop Test Program:  A factorial experimental program was
developed and approved by the EPA.  This program, conducted in our combus-
tion facility,was designed to permit evaluation of vanadium particulate
collection efficiency for a standard EPA Method 5 stack sampling train and
an ER&E developed sampling system.  It employs a factorial design to mea-
sure the interactions between several variables which were chosen because
they might affect the amount of vanadium in the flue gas and the size
distribution of the vanadium emissions.  These variables were combustion
chamber residence time and fuel composition (ash and asphaltene content).
Two levels of residence time and three typical heavy fuel oil compositions
were selected for evaluation.  In addition,  two  sampling  locations,
 90° apart  in  the stack, were  included as variables  in  the  study
 to eliminate  any bias attributable  to particulate maldistribution  in
 the flue gases.  To  evaluate  these  factors, 36 experiments were
 planned -  12  base runs, each  with two levels of  replications.  An  important
aspect of the test procedure was the simultaneous operation of both sampling
trains in each test.  This feature not only served to double the data base,
but it also eliminated sample bias attributable to run to run fluctuations
in boiler performance.  This influenced our decision to use restricted
randomization for selecting the sequence of experiments.

          The key to evaluating sampling system collection efficiency was
a vanadium balance around the boiler.  By knowing the vanadium content
of the fuel which is burned and determining the amount of vanadium that
deposits in the boiler, it is possible to establish how much vanadium
enters the stack.  The amount not accounted for in the sampling train

-------
                                  -  22 -
represents the combination of collection  inefficiency and/or experimental
error in measurement and other inaccuracies.

          To accomplish a vanadium material balance in the boiler, two
methods were originally proposed.  The method actually used in the study
consisted of inserting a removable, thin wall, stainless steel liner
into one heat exchanger tube in each boiler pass.  At the end of a test,
the inserts were removed, their contents were inventoried and assayed for
vanadium.  Assuming that the combustion gases were not maldistributed,
the amount of vanadium found in one insert times the number of heat ex-
changer tubes in that particular pass equalled the amount of vanadium
deposited.  This method has two potential causes of inaccuracy.  The
first is the supposition that combustion gases flow equally through all
the heat exchanger tubes.  The other shortcoming results from not being
able to inventory the two water backed face plates in which the heat
exchanger tubes are set.  These plates account for about 10% of the
total surface area of the passes.

          In order to overcome these problems, a second method was pro-
posed involving isokinetic sampling of the combustion gases at the end
of the firetube.  The advantage of sampling there is that the combustion
gases have not as yet gone through any heat exchanger tubes or return
sections, which tend to collect particulates.  Thus a material balance
would obviously be facilitated.

          The disadvantages of sampling in the firetube are that the
gases are still very hot and that there may be greater maldistribution
of particulates than in other sections.  At the point where the sample
would be taken, gas temperature is 1100-1600°F depending upon combustion
conditions.  The gas sample would have to be quenched rapidly from fire-
tube temperature to 400°F in the collection system.  Since some of the
ash components might still be vaporized at firetube temperature, this
quenching would cause them to precipitate in a different form than they
might have in a heat exchanger.  Thus size distribution might be affected.

          Particulate maldistribution may come about from inhomogeneities
in the flame and the swirl imparted to the combustion gases by the burner.
This problem simply means that greater care would have to be taken in tra-
versing to get a good sample of the overall particulate loading.

          In view of budget limitations it was decied only one method
should be used.  Since the tube inserts appeared the most practicable
means of measuring vanadium deposits in the boiler, they were chosen.
During the initial phase of the program several preliminary experiments
were conducted to debug this technique. Samples from these runs were also
used to establish the analytical methodology for determining vanadium
content of the particulate.

     3.  Conduct Test Program:  Paired tests were conducted in the package
boiler during each week of operation.  Experimental pairings were based on
alternating combustion chamber residence time.  This technique was employed
as a means of minimizing the effects of possible long term fluctuations in
boiler performance.  Of the 36 tests constituting the factorial program
only one (Run 61) had to be terminated and the results eliminated because

-------
                                 -  23 -

of problems traced to boiler operation.  The entire test program required
5 months to complete.  Additionally several experiments were made to
determine changes in oxidation state of vanadium compounds retained on a
filter when subjected to the stack gas environment.

6.2  Combustion Test Facility

     6.2.1  50 hp Cleaver Brooks Package Boiler

          All combustion experiments were carried out in a 50 h.p package
boiler having a nominal firing rate of 2 Ibs/min. (15.0 gph) of residual
fuel oil.  This boiler is a horizontal firetube type; by means of appro-
priate baffles and heat exchanger tubes, the combustion gases are forced
to pass the length of the boiler four times before being emitted into the
stack.  A cut-a-way view of a typical C-B boiler showing the four pass
construction is illustrated in Figure 6.1.  In these tests, firing rates
of 2 Ibs/min. and 0.5 Ibs/min. were used.  This difference represents a
four-fold change in combustion chamber residence time; the reciprocal of
firing rate Cp^) is roughly proportional to residence time.  To amplify
even more the effect of residence time, in those runs using the low firing
rate an added second refractory chamber was butted to the existing com-
bustion chamber.  This increased the overall combustion volume by about
66>% and residence time by an additional 1.6 times.  When not required, the
added refractory was removed from the boiler.  Firing rate and fuel con-
sumption were accurately metered in each test by periodically monitoring
the loss in weight of the oil supply drum.  Atomization of the residual
fuel oil was accomplished by maintaining a constant volume of air at low
pressure (10 psi) on the burner gun.  In all tests the fuel oil viscosity
at the atomizing nozzle was maintained at 30 centistokes by heating the
fuel to a predetermined temperature with an electrical heater.

          Secondary air for combustion was furnished by a centrifugal
blower mounted in the boiler head.   The air was forced through a diffuser
plate to thoroughly mix with the atomized oil before combustion.  The
amount of the secondary air was controlled by means of a damper which was
regulated to keep oxygen concentration in the flue gas, normally at 2%
(VLO% excess air).

          For monitoring boiler performance, the concentration of CO,
C02 and Q£ in the flue gas was continuously measured using Beckman instru-
ments.  Both carbon monoxide and carbon dioxide were measured using non-
dispersive infrared analysis while oxygen was determined polarographically.

          Figure 6.2 is a front view photograph of the Cleaver Brooks boiler
illustrating the oil supply drum monitoring arrangement and air control
damper/burner housing.  Figure 6.3 shows the instrument bank (foreground)
for monitoring composition of flue gas.  Controls for the stack sampling
systems are positioned alongside the boiler.

          To insure operational stability, the package boiler was allowed
to warm up for a minimum of one hour before the start of a run which would
normally last from 1-2 hours.  During the test, minor adjustment of the
secondary air to maintain flue gas oxygen concentration at 2% was
normally the only control required to hold conditions constant.  The oil
firing rate was set immediately upon start up of the boiler.  Temperature
of the combustion gases at passes 1, 2 and 3 as well as in the stack were
recorded continuously.  Other ancillary measurements included  pressures  in

-------
                                 THE FOUR PASS CONSTRUCTION OF A TYPICAL C-B BOILER
                                                                                             Pass 3
                                                                                                                 Pass 4
                                                                                                                 Pass 2
                                                                                                                            to

                                                                                                                            I
 FIGURE  6.1

This cutaway view shows the gases are constrained to flow through the four passes in the order indicated.  The
combustion air enters the burner through an Adjustable Damper.  The blower forces air through the diffusor and
into the combustion chamber, this constitutes Pass (No. 1).   Baffling allows gases to pass to the front of the
boiler only through Pass (No. 2).  A plate constrains gases to travel to the rear of the boiler through Pass (No.
From the rear head,gases are forced through Pass (No. 4) to the stack.
3).

-------
                                 -25-
FIGURE 6.2   - 50 HP CLEAVER BROOKS BOILER

FRONT VIEW OF BOILER SHOWING OIL SUPPLY  MONITORING  SYSTEM, AIR  DAMPER
CONTROL AND BURNER HOUSING.

-------
                                 - 26 -
FIGURE 6.3 - ARRANGEMENT OF FLUE GAS MONITORING SYSTEMS

FOREGROUND:  INSTRUMENTS INCLUDE CO, C02, 02 AND NOX MONITORS.   CONTROL
EQUIPMENT  FOR PARTICULATE SAMPLING TRAINS IS POSITIONED NEAR BOILER.

-------
                                 -27-
the windbox and end of first pass, and back pressure on the oil line and
atomizing air line.

 6.3  Program Design

          The object of our experimental program was to measure and compare
the effectiveness of the standard EPA type stack sampling system and the
Exxon Research developed system for collecting vanadium emissions from com-
bustion of appropriate residual fuel oil.  A standard EPA type train was
used which consisted of an 18 inch glass lined probe, a cyclone, a glass
fiber filter (125 mm, Type MSA-1106BH) and two chilled water impingers.
With the exception of the latter  all components were heated to 400°F.
The ER&E system comprises an 18 inch glass lined probe,  an 8-stage
Andersen Cascade Impactor and two silicone oil-containing impingers, all
heated to a temperature of 400°F.  As back-up to this system, a dry ice-
acetone  finger trap at -100°F was used in the line after the last impinger
to condense any volatile vanadium compounds.

          The reliability of the two systems, i.e., the ability to collect
all of the vanadium-bearing materials entering the stack sampling system
was determined on the basis of a vanadium balance made around the boiler.
Inventorying the particulate fallout in the four boiler passes was accom-
plished with removable, thin  wall, stainless steel liners inserted in
one of the heat exchanger tubes in each pass.  At the end of a test the
inserts were removed, their contents were collected and weighed by the
same means used for isolating sample probe solids.  Analysis for vanadium
in the particulate was made using atomic absorption spectroscopy.

          It was expected that particle size distribution of the vanadium
particulate would have a major bearing both on the amount of material that
deposits in the boiler tubes and on the collection efficiency of the stack
sampling systems.  Since the concentration and chemistry of vanadium in
the fuel during combustion could influence this distribution, residual fuel
oils encompassing a range of ash contents and asphaltenes (hexane insolu-
ble material) were employed.  Selected resids included a high and an  inter-
mediate  sulfur  fuel  oil  produced  from Venezuelan  (Tia  Juna  Medium)  crude
and a high sulfur  Light  Arab  fuel  oil.   All fuels are typical of those
currently being used in the United States.  Key inspections are presented
in Tablt  6.1.  Complete analyses are shown in Table  6.2.


                               TABLE  6.1

         Residual Fuel Oils Used in V Collection Efficiency Study

                                                           Asphaltene Content
Fuel Designations         Sulfur  Vanadium  Nickel  Ash   (Hexane Insolubles)
& Type	    Wt. %   wppm      wppm    wt.%   	Wt.  %

F , Hi. S (Venezuelan)
F2, Med. S
F-, Hi. S. (Light  Arab)
                          2.8      39       10      0.01
2.2
1.0
359
149
63
26
0.09
0.05
12
5

-------
                                                   TABLE 6.2
      Inspection
% Ash
B S & W
BTU///
% Carbon
% Hydrogen
% Sulfur
Con Carbon
Flash Point, °F
API Gravity, 60°F
% Hexane Insolubles
% Nitrogen
Pour Point
Sed. by Hot Filtration
Vis. SSU @ 100
Total Oxygen %
Metals, ppm
    V (Range)
    Na
    Ni
    Fe
Heavy
Low Ash
Light Arab
Test Fuel F0
0.01
0.06
18,708
85.22
11.04
2.84
7.18
220
16.4
3.74
0.21
+40
0.02
1444
0.20
37-42
Trace
10
10
Fuel Oil Inspections
High Ash
Venezuelan
Test Fuel Fn
0.09
0.16
18,540
84.73
11.23
2.20
12.70
178
14.7
11.79
0.48
+20
0.20
4694
0.42
335-383
24
63
9
Intermediate Ash
Venezuelan
Test Fuel F0
0.04
0.80
18,976
86.63
12.24
0.96
5.63
216
20.7
4.98
0.28
-10
0.20
500
0.40

138-158
10
26
6
to
00

-------
                                 - 29 _
          Of the two Venezuelar  fuel  oils it may be noted that ash and
asphaltene contents are roughly proportional to their sulfur content.
This results from the desulfurization process used.  Since ash and
asphaltene in the Venezuelan residua cause intolerable catalyst deac-
tivation in the desulfurization process, the residuum is vacuum dis-
tilled before it is hydrotreated.  Only the overhead cuts from the
vacuum tower, which are free of ash and asphaltenes, are desulfurized.
A portion of the high sulfur vacuum bottoms is blended back into the
desulfurized vacuum gas oil to make the final low S fuel oil.  The
amount which can be added is limited by the fuel oil S target.  A 1%
sulfur fuel oil contains only about half as much vacuum resid as the
high sulfur (2%) fuel oil.  Since all of the ash and asphaltenes are
contained in the vacuum resid the concentrations of these components
are proportional to the sulfur content.  Thus a 1% S resid would have
about half that of a 2% S fuel oil.

          Residence time in the combustion zone is a major variable
affecting particle size distribution and distribution of vanadium among
the particle size range.  Residence time is also one of the primary
differences between different sized boilers.  Previous studies at Exxon
Research have measured the change in particulate distribution as a
function of residence time.  This effect is illustrated in Figure  6.4.
As residence time in the combustion zone increases, more of the large
carbonaceous particles are consumed, shifting size distribution toward
submicron particles.  These are generally the most difficult to collect
in a sampling system.  A major change in size distribution resulted
from cutting firing rate in half with an extra combustion chamber.  Total
particulates decreased by about half to 0.11 wt.% and almost all the
cenosphere and intermediate size particles were consumed.  Inorganics
and refractory carbon material comprised the remaining particulate.

                               FIGURE 6.4
                    LONGER RESIDENCE TIME FAVORS
                       BURNOUT OF CENOSPHERES
  200
              BASE
              1 LB/MIN
              1 CHAMBER
0.5 LB/MIN
2 CHAMBERS
0.5 LB/MIN
3 CHAMBERS

-------
                                  -  30  -
          In the experimental program, two levels of residence time were
studied.  For the base level  (H) tests were run at a firing rate of
2 Ibs/min.  (15 gal/hr.) using tne normal combustion chamber configura-
tion.  At this combustion condition the calculated volumetric flow rate
was on the order of 450 cu. ft./min., which gave a residence time of
roughly  50 msec in the combustion chamber.  In this case large particu-
late dominated the size distribution.  For the second level ('T-f ) , firing
rate was decreased to 0.5 Ibs/min. (3.5 gal/hr.) and an additional
refractory chamber was added.  These changes produced about', a 7-fold
increase in residence time  to 350 msec.    More  importantly, at this
combustion condition carbonaceous material burned off and the particle
size distribution was shifted toward the submicron range.

          As both residence time and fuel composition are main variables
which may influence vanadium  collection efficiency a factorially designed
experimental program was employed.  The program plan is outlined in
Table 6.3.  There were 12 different experiments each replicated three
times for a total of 36 runs.  As previously indicated both EPA and ER&E
sampling trains were operated simultaneously in each experiment.  This
was possible by using two ports located at right angles to each other
in the boiler stack.  To prevent the possibility of sample bias, since
it was not feasible to make a standard traverse in each port for every
run, the sampling train positions were rotated as indicated in
the table.  On the basis of finding no statistical difference between
results from the two positions, the experimental design allows the
pooling of these data.    In  essence, this would double the number of
replications and degrees of freedom in defining the experimental error.

          Selection of the order in which the experiments were run
was based on a restricted randomization.  With truly randomized tests,
there is a finite probability that a block of experiments involving
a particular residence time (jTo or'TiO will be grouped toward either
end of the program.  Any long term change in boiler performance would
influence residence time and  seriously prejudice the results.  Since
we were interested in comparing collection efficiency of the sampling
systems for vanadium particulate emitted in different size ranges,
a restricted test approach was used.  Accordingly, paired tests, a
test at one residence time followed by the same test but at the
other residence time were used.  Since only two tests a week could be
run, this grouping was most efficient.
          To analyze for vanadium in the collected particulate, Atomic
Absorption spectroscopy was used.  This method combines both high precision
with rapid sample throughput.  The latter means that many samples can
be run in a short time period.  This was an important feature since over
the life of this study approximately 475 samples were analyzed.

-------
                                                     TABLE 6.3
                                          Experimental Program Design


Residence Time
TO = 50 msec.
Standard - Normal
Combustion Chamber,
15 Gnls/hr. firing rate
Residence Time
T j = 350 msec.
*vr fold increase with
double chamber and
3.75 Gals/hr. firing rate
Stack Sampling Probe Location
(1) EPA 	 p->
EPA Vs. ER&E 1
Residual Fuel Type
Hi S Ven.
Run No.
55
57
68

54
56
67
F2
Med S Ven .
Run No.
48
69
75

49
70
76
F3
Light Arab
Run No .
47
61*
78

46
60
77
Stack Sampling Probe Location
(2) 1 +4 	 EPA
ER&E Vs. EP/I 1
Residual Fuel Type
Hi S Ven..
Run No .
52
71
79

53
72
80
(finish)
F2
Med S Ven .
Run No.
45
63
74

44 (start
62
73
F3
Light Arab
Run No.
50
58
65

51
59
66
                                                                                                               Cycle 1
                                                                                                               Cycle 2

                                                                                                               Cycle 3


                                                                                                               Cycle 1
                                                                                                               Cycle 2

                                                                                                               Cycle 3
•  Experimental order based on restricted randomization of pairs.
•  Total experiments = 36
•  Time frame = 5-  months
                              Notes on Experiments
                              *Run No. 61 was terminated after boiler upset.  No data obtained in this run.

-------
                               -  32 -
6.4  Boiler Tube Inserts

          Each of the passes in the Cleaver Brooks boiler consisted of
cooled heat exchanger tubes, 121 inches long by 2-1/4 inches internal
diameter.  There are fourteen of these tubes in pass 2, ten in pass 3
and eight in pass 4 providing a total of 192 ft.  of heat exchanger
surface area.  Pass 1 is significantly different from the others.
It consists of a refractory lined combustion chamber 30 inches long x
12 inches diameter expanding into a water backed firetube,  19 inches  in
diameter.' As summarized in Table 6.4, the firetube (pass 1) accounts for
roughly 15% of the total heat exchanger surface area.   Pass 2 has
the greatest surface area, accounting for 33%.   Pass 3 accounts
for 24%, while pass 4 makes up 19% of the area.  There are water
backed face plates at either end of the boiler which could not be
inventoried for vanadium particulate.  These account for 9% of the
surface area.

                             TABLE 6.4

             Heat Exchange Surface in the Cleaver Brooks
Pass

 1
 2
 3
 4
Ends
No. of Tubes

 (firetube)
    14
    10
     8
Total Surface Area, Ft.'

       37.6
       84.0
       60.0
       48.0
       22.0
% of Total

   15
   33
   24
   19
    9
                                       251.6
                                                       100
          Particulates are thought to deposit within the heat exchanger
section by two mechanisms.  Large particles drop out by impaction on
the walls of the tube or in the headers where gases reverse direction
between passes.  It was not possible to collect and inventory particulates
from the headers after each run so this is a potential source of error.

          The other collection mechanism is thermal or molecular
diffusion of fine particles to the walls of the tubes.  The temperature
gradient between flue gas and water backed metal surface served as a
driving force for diffusion.  Since metal surface was essentially
constant,this varied from pass to pass and between runs.  Table  6.5
shows temperature levels measured  in   each pass at high and low firing
rates.  By the end of the second pass, temperature has dropped to
within a few hundred degrees of stack temperature, especially at low
firing rate.  Thus driving force for thermal diffusion is highest
in the first two passes.

-------
                                  -  33  -
                              TABLE 6.5

  Influence of Combustion Conditions on Boiler and Stack Temperatures


                                  Fuel Firing Rate, Ibs./min.
Temperature °F                      0.5               2.0

End Pass 1                         1060               1625
End Pass 2                          380                705
End Pass 3                          283                500
End Pass 4                          225                360
Stack                               220                355
          Removable thin wall stainless steel liners were used to
measure particulate and vanadium fallout in the passes.  These were
machined from 304 seamless 2-1/4 inch OD x .065 wall OD pipe.   These
liners were inserted during a run into one tube in each pass and then
removed and their contents inventoried at the end of each run.  Since
it was not possible to sample the firetube by this method, a manual
cleaning method was used, (sweeping out the firetube).  In Figure 6.5,
photographs A and B show the back end of the Cleaver Brooks boiler
with removable liners inserted in the various passes.  The pictures
show two tubes per pass, but in most runs only one per pass was used.
The 14 second pass tubes are around the firetube in the bottom half of
the boiler.  The chalk line drawn across the boiler facing separates
the 10 tubes in the third pass from the 8 tubes in the fourth pass.
Combustion gas enters the third pass at the front end of the boiler.
Since this area is not easily accessible, a locking collar was welded
into three of the tubes to prevent channeling of the gas stream between
liner and tube.  Channeling was prevented at the back end of the boiler
by a collar welded on the liners.  The collar fits tight against the
tube and face plate providing a seal.  Photographs C and D show the
firetube .refractory lined combustion chamber and a front view where
the burner housing has been  exposed.

6.5  Stack Sampling Systems

     6.5.1  EPA Method  5

          The EPA Method 5 stack sampling system used in this study was
a commercially available Joy Emission Parameter Analyzer.  This unit
consisted of an 18 inch glass lined probe, a cyclone, a 125 mm glass
fiber filter (filter type MSA-1106BH) and two water impingers.  The
configuration is shown schematically in Figure 6.6.  With the exception
of the impingers, the entire sampling train was maintained at 400°F
in every run.  This insured  that stack gases entering the sampling system
were well above the acid dew point.  Therefore sulfuric acid could not
condense on the collected particulate or the walls of the equipment.

-------
                                         - 34 -
                 FIGURE  6.5  -   FOUR  PASSES OF CLEAVER BROOKS BOILER
                                                     it
A. THIN  WALL STAINLESS STEEL REMOVABLE LINERS
   FITTED  INTO SELECTED HEAT EXCHANGER TUBES
   IN PASSES 2, 3 AND 4.  (BACK END VIEW)  -
   COLLAR  ON END OF LINER PREVENTS CHANNEL-
   ING OF  THE COMBUSTION  PRODUCT GAS STREAM.
                                           B.  BACK TO FRONT VIEW OF BOILER ILLUSTRAT-
                                               ING CENTRAL  FIRETUBE AND HEAT EXCHANGER
                                               TUBES.  REMOVABLE LINERS UAVE BEEN IN-
                                               SERTED  INTO  HEAT EXCHANGER TUBES. CHALK
                                               MARK OUT LINE INDICATES  TUBES  IN PASS
                                               3 AND 4.
>v
                              -*.
             J    C*   J    '•
C.  CENTRAL FIRETUBE LOOKING TOWARD BURNER
    GUN.  AREA OF REFRACTORY LINING IS THE
    COMBUSTION CHAMBER.
                                         D.  FRONT VIEW OF BOILER WITH BURNER  HOUSING
                                            AND ACCESS DOOR  REMOVED.  SHELF SEPARATES
                                            PASS 4 FROM PASS 3.

-------
                                      -  35 -



                                FIGURE 6.6

                         EPA Method 5 Sampling Train
      PROBE
REVERSE-TYPE
 PITOT TUBE
                     HEATED AREA
                                    FILTER HOLDER
                                                        THERMOMETER
                       VELOCITY
                      PRESSURE
                       GAUGE
                                                                    CHECK VALVE
                                                                    VACUUM LINE
                    ORIFICE
                    GAUGE
          COARSE VALVE
                       DRY TEST METER
AIR-TIGHT
  PUMP
       6.5.2  ER&E Collection System

             Because of the problems with  existing  stack sampling trains
  and to meet the Company's program needs,  ER&E  developed a system several
  years ago with more analytical capability than    "<.od  5.   This system
  permitted segregation of particulate  into several size fractions;
  it eliminated the formation of "artificial type particulate" and it
  allowed collection of relatively large samples which could be quantitatively
  recovered for analysis.  Rather than  spend considerable time and effort
  in hardware development, the  best available components were selected
  and built into a sampling system.  Thus  the stack sampling train illustrated
  in Figure  6.7 evolved.

             The heart of  the system is an Andersen eight stage Cascade
  Impactor and several Greenberg-Smith, high velocity impingers filled
  with silicone oil.  These are all mounted in an oven maintained at 400°F.
  The high temperature impingers make it possible to obtain a true inventory
  of "real particulate" with no sulfuric acid formation or hydrocarbon
  condensation.  Both can  and do occur  in  cold water impingers to a
  different degree than they do in stack plumes. Since particulate loading
  is reported on a mass basis,  a more accurate inventory is possible with
  the ER&E system

-------
          FIGURE 6.7
    ER&E STACK SAMPLING TRAIN
         DRY AIR METERING SYSTEM
                                   OPTIONAL  BACK-UP SYSTEM
                                    Ji      Jl
                                s* ~ ~~^y   f ~ ~~~SJ      '
                                O     tf       !
  (MICRO   L—I
MANOMETER)
    CERAMIC LINED OVEN AT 400°F
                                     ICE WATER
                                       TRAP
ORIFICE
BY-PASS
 VALVE       MAIN
            VALVE
                      DRY ICE
                    CONDENSER
                                                  VAC.
                                                 GAUGE
                          (      JL- LEAKPROOF
                          V    J  VACUUM PUMP
                           V—"\

-------
                                   - 37 -
          In most of the experiments, a sampling rate of about 0.90
SCFM was used with this system.  With the  isokinetic sampling
requirement, however, when that amount of flue gas could not be withdrawn,
dry filtered makeup air was provided to the unit upstream of the cascade
impactor.  With this provision, it was possible to still sample
isokinetically while maintaining a fixed flow rate through the impactor
and impingers.  The makeup air was heated to 400°F and supplied at
atmospheric pressure.  To alleviate the problem of wall loss and particle
bounce in the cascade impactor, it was found necessary to have leak-proof
operation and during sampling maintain a constant volumetric flow rate
through the system.

           Particle distribution was divided into three (mass) size
ranges:  >10 urn, 1-10 Mm and 1 ym.  The amount in each of these ranges
was determined mathematically by plotting on logarithmic probability
paper the aerodynamic particle diameter versus cumulative weight percent
of the collected particulate.  Aerodynamic diameter was determined
based on a computer, solution of  the Ranz-Wong equation^  ' for the
individual stages of the Andersen Impactor.  The 50% cut-off points
(050) calculated for the stages were generally in agreement with the
manufacturer's values.  At the conditions employed, stage 0 had a
cut-off of approximately 10 urn while stage 5 had a cut-off of roughly
1 ym.  On this basis, all material isolated from stages 6 and 7 plus
the impingers made up the submicron particulate fraction.

          As  further illustrated  in Figure 6.7, the ER&E system uses a
dry ice-acetone cold trap at -100°F  to insure knock-out of any
volatile compounds.  The two impingers outside the oven section, one
empty and the other containing chilled water, remove silicone oil or
other materials entrained by the  flue gas.  The probe was designed
to obtain even heating over its entire length.  To minimize particle
deposition  the probe length was kept to a minimum, about 18 inches.

           Figures  6.8- 6.12 are photographs of the EPA and ER&E sampling
trains and major components making up the ER&E system.  Figure 6.9 and  6.10
illustrate the arrangement used for sampling simultaneously with the
two trains at the ports located 90° apart in the stack.  This arrangement
was designated in  the factorial program as location 1.  Location 2 was
obtained by rotating the position of the'sampling trains.

6.6  Particulate Sampling

           Sampling of flue gas from combustion of heavy fuel oil in
the Cleaver Brooks Boiler was carried out in a 12 inch circular stack.
A single sample point located midway  on the radius was employed for
sampling particulate.  This procedure was experimentally justified
by our previous findings of uniform gas flow and particulate dispersion
across the stack.  Velocity profiles across the stack were  determined
with a standard 6-point pitot tube traverse (in the north-south
direction).   The traverse points were located at the center of each of

-------
                                      - 38 -
FIGURE 6.8   -  OVERALL VIEW OF BOILER,  STACK SAMPLING  SYSTEMS
                AND CONTROL MODULES.

-------
                                       - 39 -
FIGURE 6.9 - ARRANGEMENT FOR SIMULTANEOUS OPERATION OF EPA AND ERE STACK
             SAMPLING TRAINS. (POSITION 1)   EPA TRAIN  LOCATED  ON  LEFT  HAND  SIDE.

-------
                                    -40-

FIGURE 6.10 - CLOSE-UP VIEW OF PORT HOLE CONFIGURATION FOR SIMULTANEOUS SAMPLING.

-------
                                        -41-

•
     FIGURE 6.11 - OVEN ASSEMBLY OF ER&E SYSTEM SHOWING ANDERSEN CASCADE IMPACTOR,
                   IMPINGER SYSTEM AND MAKEUP AIR LINE.GLASS TUBE UPSTREAM OF IMPACTOR
                   IS CONNECTED TO PROBE.

-------
                                       - 42 -
FIGURE 6.12 - VOLATILES KNOCKOUT SYSTEM CONNECTED TO ER&E TRAIN FEATURES A
              DRY ICE ACETONE COOLED CONDENSER.

-------
                                      -43-
    six equal area zones.  Representative results are illustrated in
    Table 6.6.  Normally test conditions (firing rate and excess combustion
    air) are such that gas velocities on the order of a few feet/second
    are obtained.  Hence calculated, flue gas velocity based on combustion
    conditions, is used for setting the isokinetic sampling rate.  The
    agreement between calculated and measured values is quite good.
    The data in Table 6.6 also show the uniform flow behavior of the flue
    gas in the stack.

                                   TABLE 6.6

                       Stack Velocity Profile from Boiler
                      (Fuel, firing rate, 1.1 Ibs./min. 10% excess air)
Traverse
Point
Number

  1
  2
  3
  4
  5
  6

Midpoint
(N/S)
Distance
into Flue
Inches

   0.5
   1.8
   3.5
   8.
  10,
  11.5
   6.0
 Pitot Tube
 AP
Inches H?0

  0.011
  0.011
  0.010
  0.010
  0.011
  0.011

  0.010
Stack Temp.
   °R
   680
   700
   740
   752
   750
   748

   743
Stack Velocity, ft./min.
Measured    Calculated
  401
  407
  399
  402
  421
  420

  400
368
379
401
407
406
405

402
               A second port drilled into the stack at the same height
     but 90° removed (east-west direction) permits simultaneous withdrawal
     of particulate.  Characteristic velocity profiles were developed for
     both sampling planes and are tabulated in Table 6.7.  The data indicate
     a balanced flow across the stack.  At the midpoint location,measured
     velocities were within 10% of each other and of the calculated velocity.
     Pitot tube readings (AP) were measured with an inclined Dwyer micro-
     manometer.

-------
                                 - 44 -
                              TABLE 6.7
                       Low Level Source Profile
                  (Fuel firing rate ^0.65 Ibs/min)
Traverse
Point
Number
N/S
1
2
3
4
5
6
Distance
Into Flue
Inches
0.5
1.8
3.5
8.5
10.2
11.5
Pitot Tube
AP
Inches, H20
0.004
0.005
0.004
0.005
0.005
0.008
Stack Temp.
°R
706
712
712
736
736
742
Midpoint    6.0
  E/W
  1
  2
  3
  4
  5
  6
 0.5
 1.8
 3.5
 8.5
10.2
11.5
Midpoint    6.0
             0.005
0.005
0.005
0.005
0.006
0.005
0.008

0.005
                                                  Stack Velocity, ft/min.
                                                   Measured   Calculated
               738
730
728
732
734
738
744

734
               249
               279
               250
               284
               284
               360

               298
283
282
283
310
283
361

283
           265
           267
           267
           276
           276
           278

           277
274
273
275
275
277
279

275
          In order to insure that the particulate mass flow rate was
uniform from point to point within the cross section of the stack a
similar type 6-point traverse was also made.  The results are summarized
in Table 6.8 and compared with the particulate catch obtained by simply
sampling only at the midpoint in the stack.  Both tests were conducted
with marine bunker fuel.  On average the mass concentration of particulates
determined by the two procedures differed by only a negligible amount,
about 1%.  Comparing size distribution, however, only the mass concentration
of submicron particulates was in agreement.  The intermediate and large
size fractions obtained by traversing the stack were almost the reverse
of the distribution from single point sampling.  Normally an 18 inch
long sample probe is used to withdraw the flue gas.  To accomplish the
traverse, however, a 5 foot probe was employed.  With the latter length

-------
                                 -45-
there was considerable deposition of solids on the wall of probe.
Although both large and intermediate size particles were probably
present, the probe sample (by convention) was considered to be
large particulate.  This would, of course, bias the results in the
direction observed in the table.  Any decrease in the amount of
solids attributed to the >10 ym fraction would produce an equivalent
weight increase in the 1-10 ym fraction and, overall, would place the
total particulate distribution in better agreement with the single
point sample.

                              TABLE 6.8
            Mass Concentration Profile in Stack is Uniform
Particulate Spectrum                 I                       II
	mg/SCM	          6-Point Traverse      Single Point Sample
                                (5 ft. probe)          (18 in. probe)

      >10 microns                  76.07                   46.18
      1-10 microns                 50.72                   92.37
      <1 micron                   126.78                  118.03
      - Total Mass                253.57                  256.58

      %A                            —                      1.2

      - Wt. % on Fuel               0.32                    0.32
          Since both the total mass concentration of solids, and
gas velocity, in the stack were steady  (no variation with time) and
uniform (no variation from point to point within the cross section)
only one measurement is required for accurate results  '   '.  While this
measurement may be taken from anywhere within the cross section, a
midpoint sample station has been utilized.  It should be noted that
even with an unsteady but uniform source condition, measurements are
needed at only one location.  In the later case, however, the measurements
must be taken over an entire cycle, or over as long a time period
as possible, for noncyclic operation.  Although the Cleaver Brooks
operation is indeed noncyclic  a sampling time of up to three hours
was used in the program.

-------
 6. 7  Particulate Isolation Method - General Procedure

          Recovery of particulate from the stainless steel liners
inserted into the heat exchanger tubes of the boiler was accomplished
using the same procedure employed for collecting the sampling probe
solids.  A water/organic solvent wash was normally all that was required
to remove the solids quantitatively.  In several experiments, an
acid rinse of the tubes was employed as a back-up wash.  Subsequent
analysis for vanadium in the acid showed only trace quantities.
Particulate from the firetube and the three other passes were kept
segregated and analyzed for vanadium content separately.  Prior to
analysis each sample was dried under vacuum for a minimum of A hours,
at a temperature between 250 and 300°F.  Analysis of all particulate
samples was carried out by Atomic Absorption Spectroscopy.  The
detailed procedure is presented in Appendix 3.

          Isolation of the solids in both sampling trains were
accomplished using the following general scheme:

          Particulate matter from the various parts of the sampling
train were isolated separately in order to determine net particulate
weight.  This was accomplished with water and organic solvent washings.
Solvents were removed by evaporation and all particulate material was
dried under vacuum at 300°F prior to weighing.  The isolated solids
were combined into three fractions for chemical analysis  (normally
only vanadium) and characterization.  In the ER&E train these consisted
of  probe material and stage 0 of the impactor; stages 1-5 of the
impactor; stages 6,7 and the impingers.  For the EPA Method 5 train
they consisted of probe and cyclone; filter, and impingers.

          The detailed procedures used for isolation and characterization
of the particulate are presented in Appendix 2.

6.8  Determination of Vanadium in Particulate

          Atomic Absorption was the primary instrumental method used
to determine vanadium content of the particulate in this study.  It
has high precision and it is fast.  However it does require that the
vanadium be in solution.  To accomplish this, two procedures were
initially evaluated.

          In the first method, a series of particulate samples held
in porcelain crucibles were ashed at 1450°F overnight.  (This is
a modification of the standard method of test for ash from petroleum
product; ASTMD-482).  While there was some concern that at this
temperature V^s would volatilize, there has been no evidence to support
this.  On the contrary Foster et al'22)f in a study of phase equilibrium.

-------
                                - 47 -
relationships in the system Na20-SC>3-V205 found that a crucible
containing V2C*5 showed absolutely no weight loss in six hours
at ^1600°F.  An identical test showed that a mixture of 20% Na2SC>4 -
80% V2C>5 lost no weight during 27 hours of exposure at 1600°F.  It
was further demonstrated by Milan ^  ) that in vacuo ^^65 begins
to dissociate into V2C>4 and 02 at temperatures only slightly above
its melting point of 1274°F.  For example, at 1382°F the oxygen
partial pressure in the gas phase corresponded to a melt containing
0.56% V20s while at 1832°F the melt contained 2.75% V204-  It was
also noted that the dissociation was reversible provided that the
composition of the melt or condensed phase remained constant.

          While V205 may not volatilize during the high temperature
ashing, it does melt and fuse on the interior surfaces of the crucible.
The melt also exhibits a tendency to creep up the side walls increasing
the possibility for direct material loss.  Dissolving the fused
vanadium is extremely difficult and time consuming and there is no
real assurance that all of the metal will go into solution.  To avoid
these complications, a second treatment labeled "liquid fire" was
also evaluated.  In brief, this method employs a digestion of the
particulate sample using a mixture of concentrated acids, HN03, HC104
and I^SO^.  In Table 6.9 vanadium analyses run by Atomic Absorption
are compared for the two different ashing procedures.   Samples of
boiler deposits were obtained from several preliminary experiments
(runs 33 and 34) made to check the reproducibility of the heat
exchanger tube inserts.  The deposits from the four passes in run 34
were ashed using both methods.  In run 33 only the samples from passes
3 and 4 were tested in this manner, the other two were run using
only the liquid fire method.

          The results indicate a significantly lower vanadium content
from the high temperature ashing procedure.  Of the 6 samples run using
both methods, only one showed agreement  (9.3% vs. 9.2%), two other
samples were about 10% apart.  Where similar samples were subjected
to, high temperature ashing, the results also showed considerable
variation in vanadium content.  For the liquid fire treatment, samples
from the two test series generally were in good agreement.  As noted
previously, high temperature ashing caused the V205 to fuse on the
crucible wall making complete removal extremely difficult.  Any loss
of material in this manner would, of course, lower the vanadium analysis.

-------
                                  - 48 -
                              TABLE 6.9

  Effect of Ashing Method on Vanadium Analysis by Atomic Absorption


                               	Wt. % Vanadium in Sample	
 Sample Identification         Ashing at High Temp.*        Liquid Fire

Run 34

   Firetube Deposit                   4.7                      7.8
   Pass 2 Deposit                     9.3                      9.2
   Pass 3 Deposit                     2.3                      6.2
   Pass 4 Deposit                     3.7                      4.7

Run 33

   Firetube Deposit                    -                       7.5
   Pass 2 Deposit                      -                       8.2
   Pass 3 Deposit                     3.9                      4.3
   Pass 4 Deposit                     4.3                      4.8


*Ashed at 1450°F overnight.
           Arc Emission  Spectroscopy was also evaluated as an
 alternate means of instrumental analysis.  The major advantage of
 this method is the ability to obtain multi-element analysis on the
 same sample.  The primary disadvantage is that a much larger
 sample is required than with Atomic Absorption.  In this study,  the
 firetube deposit from a third experiment, run 40, was used.   Since
 there appeared to be pieces of refractory from the combustion chamber
 contaminating the sample a simple sieving separation was employed.
 To provide a homogenous sample, only material passing through a
 -200 mesh standard sieve  £v75vtn) was used.  This amounted to
 roughly 95% of the total.  Ten separate samples were analyzed using
 the spectrometric procedure.  Ten other samples of the sieved material
 were analyzed using liquid fire/Atomic Absorption.  A comparison of
 the results are presented in Table  6.10.  The two methods showed good
 agreement and both had  satisfactory precision.

-------
                           - 49 -
                         TABLE 6.10
  Comparison of Instrumental Methods of Vanadium Analysis
     Emission Spectroscopy

             6.8
             6.8
             6.9
             6.7
             6.4
             6.8
             6.8
             6.5
             6.6
             6.9

          x  6.7
Atomic Absorption

      6.6
      6.7
      6.6
      6.8
      6.8
      6.5
      7.1
      6.1
      7.0
      6.9

      6.7
s =
                                       (x-x)2 = 0.73
s = 0.17
   = 0.27
  = + 2.5%
   = + 4.0%

-------
                                - 50 -
          In Table 6.11 the complete multi-elemental analysis
obtained using Emission Spectroscopy has been summarized.  The
first determination was made on the fire chamber sample before
sieving.  The ash level is significantly higher, which may be
attributed to refractory material rich in silica and probably
some rust particles.  After sieving, the ash determinations were
quite uniform.  The ratio of Ni to V in the sample is on the same
order as in the residual fuel oil; Na content, however, is slightly
higher.

          To complete the analytical procedure, a method was
developed that permitted quantitative extraction of vanadium from
the particulate collected on a standard EPA-type, fiber glass
filter mat.  The fiber glass filter was supplied by Western
Precipitation as a part of their stack sampler  (Emission Parameter
Analyzer) and was stated to be equivalent to MSA-1106BH.  In
practice, the glass filter (125 mm) after being cut into strips
was extracted with a heated solution of 5% nitric acid for about
one hour.  The solution was filtered to remove carbonaceous residue
and then analyzed by Atomic Absorption.    The vanadium content is
obtained by relating the measured absorbance of the solution to the
absorbance of standard solutions.  To insure that the filter itself
did not contain appreciable vanadium several blanks were run.  The
results were all negative.  In this respect, the filter may be
similar to the spectro grade type A glass fiber filter distributed
by Gelman Instrument Company.  Based on a product description
bulletin, for the typical 8" x 10" filter sheet V content is <10 Mg.

          Carbonaceous particulate  remaining from the extraction
process was similarly analyzed for vanadium.  The solids were
first digested using liquid fire and then analyzed by Atomic Absorption.
Vanadium content proved to be negligible, amounting to <0.2% of the
total vanadium associated with the filter particulate.

-------
                                    - 51 -
                                 TABLE  6.11
Elemental Analysis by Emission Spectroscopy
(Run 35, Fire Chamber)
Determination Ash*%
1 35.2
After sieving
thru 200 mesh
2 31.4
3 " " 30.9
4 " " 31.2
5 " " 30.7
6 " " 31.0
7 " " 30.9
8 " 31.1
9 " " 31.3
10 " " 30.9
11 " " 31.0
Fe %
8.5
8.3
8.2
8.3
7.8
7.7
8.2
8.3
8.2
8.1
8.6
Si %
0.4
<0.2
<0.2
<0.2
<0.2
<0.2
<0.2
<0.2
<0.2
<0.2
<0.2
Ni %
0.9
1.1
1.1
1.0
1.0
1.0
1.1
1.1
1.0
1.0
1.1
Na %
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
V %
6.9
6.8
6.8
6.9
6.7
6.4
6.8
6.8
6.5
6.6
6.9
*Ashed at 1450°F

-------
                                   - 52 -
                      7.  RESULTS AND DISCUSSIONS

7.1  Basis for Comparing Vanadium Recovery

           This section discusses the results of the factorial test
program to measure vanadium particulate collection efficiency of the
EPA Method 5 sampling train and compare it with the ER&E developed
system.  The operating data for the 32 tests constituting this program
are presented in Appendix A Tables 1-10.  Complete particulate and
vanadium inventories are presented in Appendix 5 Tables 1-12.  This
section also discusses the several experiments run to determine changes
in oxidation state of selected vanadium compounds exposed for several
hours to flue gas at elevated temperatures.

           The strategy for establishing vanadium collection efficiency
was based on making a vanadium material balance around the boiler.
The elements of this balance involved establishing the amount of vanadium
entering the boiler from the fuel and then comparing it with the amount
of vanadium collected in the solids deposited in the boiler passes
plus the amount collected in the respective stack sampling train.  The
amount of vanadium entering the boiler was easily determined knowing
the fuel firing rate and vanadium analysis of the oil.  The amount
deposited in the boiler was based on collection of all deposits from
the tube inserts as discussed in section 6.4.  The amount sampled in
the stack, however, represents only a small portion of the total
vanadium being carried in the flue gas leaving the boiler and is
normally expressed in terms of concentration, i.e. mg/SCM, flue gas.
To relate this fraction to the total volume of flue gas produced during
combustion in these experiments, a stoichiometric calculation was
made based on the elemental analysis of the fuel and the excess
combustion air.  Since the three test fuels which were used had fairly
similar C/H ratios, the amount of combustion gas produced per pound
of fuel fell within a narrow range; 195-200 SCFDB/lb. fuel at 110% of
stoichiometric air.  Using this relationship, it is possible to convert
particulate concentration expressed as mg/SCM into weight percent on
fuel burned, and hence into part per million (ppm), on a fuel basis.
This latter nomenclature is used throughout this report to express
vanadium content in particulates.  The concentration of particulate
in flue gas has also been corrected to a standard basis of 3% 02 at 1
atmosphere  pressure and 77°F.  In addition all gas volumes are reported
on a dry basis (i.e. SCMDB).

7.2  Vanadium Material Balance Around
     the Cleaver Brooks Boiler	

     7.2.1  Total Vanadium Recovery Using All-Venezuelan Resids

           This section compares the overall vanadium balance around
the Cleaver Brooks boiler which then provides the basis for comparing
the collection efficiency of the stack sampling systems.  The results

-------
                                  - 53 -
are illustrated graphically in Figures  7.1 -  7.3.  Vanadium recovery,
which is defined as the amount of vanadium recovered in the boiler
plus the amount of vanadium recovered in the respective stack sampling
systems as a percent of the vanadium input from the fuel, is represented
in bar chart form as a function of fuel type and combustion conditions.
At high firing rate (short combustion chamber residence time,/|o), total
vanadium recovery from the all-Ven (Venezuelan) fuels (F^ and F2) ranged
between 58 wt. % and 104 wt. %.  The combined average, however, based
on the 24 runs made with both sampling systems in this block of the
study amounted to 85.3 wt. % + 3.7 (95 percent  confidence limit).
This recovery is excellent and probably represents the best that may
be obtained in our system.  For example, consider that the vanadium
inventory in the boiler made no allowance for possible V fallout at
the two headers.  Additionally vanadium recovery was based on the analysis
of vanadium in a minimum of ten separate particulate samples and
that each of these measurements had an associated precision of + 4.5%.
Comparing the two stack sampling system invididually, it is readily
apparent that Method 5 recovered on average more vanadium, 88.6 wt. %
than the ER&E system, 81.8 wt. %.
           At this combustion condition, the amount of vanadium collected
in the boiler was fairly consistent averaging 29.3 wt. %.  This
represents about one-third of the total vanadium recovered.  The stack
sampling systems had combined recovery of 56.0 wt. %.

           At the low firing rate (residence time was increased by a
factor of  almost  7), vanadium  inventoried  in  the  boiler  passes plus  the
stack sampling systems was significantly lower.  Total vanadium recovered
amounted to 68.1 wt. % + 3.7 (95 percent  confidence limit).   This
represents a decline of 20% relative to the low firing rate runs
(long residence time).  This may be observed in Figure 7.2.  Vanadium
recovered  in the boiler averaged 22.7 wt. % versus 29.3 wt. %  in the
previous series of runs.  The reason for the reduction in the  amount
of vanadium fallout in the boiler passes is not understood.  It may be
speculated that  the  low volumetric  flow rate  of  the  combustion gases  at
reduced  firing results in a maldistribution of the combustion
products as they go through the passes.  Since only one  insert
wa= used per pass to collect the vanadium-bearing particulate, any mal-
distribution would easily upset the inventory.  These points will be amplified
at greater length in section 7.5, where vanadium recovery in the boiler
is considered in detail.  The main point to be noted here, however, is
that a decrease in V recovery in the boiler affects the material balance
and has a  disproportionate effect on sample system collection  efficiency
because of the difference in bases used for the comparisons.

-------
                           FIGURE  7.1
TOTAL VANADIUM RECOVERED IN BOILER  AND STACK-SHORT RESIDENCE CASE



_l
LU
Z>
LL
•z.
o
^ 100
i—
£
o
!/) 80
Q
"Z.
<

o:
LU
5 ^O
O
CO

LU
CC
^ 40
^^
o
o
LU
CE
^
13
< 20

•^.
<
**



FUEL Fj ALL-VEN. V = 359 FUEL F? ALL-VEN. V = 149
V ppm 230 201 249 180 254 182 EPA 80 96 102 81 106 59
in
Stack 205 190 211 185 232 150 ERE 80 78 77 78 70 70
EPA ERE EPA ERE
X" 89.1 WT.% 83.4
~ s

—

-



—


-

—





—



_



X
^^^m





^
l-l


C3
Q_
CO
n




11.4
—




_^
*-•


z
_J
Q_
CO
LU
rv





CO
LU
CO
CO
^
o.
LU
_J
O
CO

































































































I^B^

















10.3
^^^



















































—

































































r— 1


















—





















































^•M





































































^T
1— (
|—

z:
Ou
CO
d.
LU








•zr
»— «
|—

tD
Q_
CO
LU
or
LU




CO
LU
CO
CO
Q_
or
LU
_J
1— 1
0
CO



























^^^m




































































^•^^H

















88.2 WT. % 80.3
14.8
—



















































^•^H































































^^^


















7.0

~—






























































^^^
























-



—


-

—



_

—



-













i
Ul

1




















55 57 68 52 71 79 48 69 75 45 63 74
Sampling Location 12 12

-------
                                          FIGURE 7.2
              TOTAL VANADIUM  RECOVERED IN BOILER AND STACK-LONG RESIDENCE CASE


	 1
UJ
ID
O
=5 100
-
o
00 80
0
UJ
	 1
§ 60
Q
UJ
Cd
UJ
o 40
o
1 1 1
on

:>
^
a
i 20
>





/•>
FUEL FT ALL-VEN. V = 359 FUEL F2 ALL-VEN V = 149
V ppm 189 181 163 188 180* 173 EPA 79 71 84 69 80 88
in
Stack 158 170 116 139 175 126 ERE 43 63 62 49 57 69
EPA ERE EPA ERE
I = 71.2 WT. % 62.4 X" = 76.6 WT. % 62.3
sx = 4.8 9.7 sv = 6.6 6.7
A A

_^

-
—




—







—

™





^^^m


l—t
t—
(T

	 1
Q.


CO
eC
Q-
UJ




-

1— 1
CtT
1—
o
z
a.


CO
UJ
of




CC. CO
Ul UJ
_l CO
•— ' CO
o <:
CO CL.





































—











































—























































fm^m
















^—^m


















































—













































^^^

















__







































-^


«


Z
1— 1
Q_

*t


Ul




S
a:

^
_j
a.

CO

Ul
Ul


a: co
Ul Ul
_l CO
•— " CO
o ^c
CQ Q-






















^^^B















	









































—















	











































	























































	
















	
























































_























"
-




—









™





URun: 5A 56 67 53 72 80 49 70 76 44 62 73
V « v / - j \ ( 1
Sampling Location

-------
                                  - 56 -
           The stack sampling system accounted for most of the decrease
(^60%) in vanadium recovery.  The combined average for both EPA and ER&E
trains amounted to 45.4  wt. % representing a decline of 19% relative
to recovery   at  the high firing rate Clo) runs.  The ER&E train showed
the lowest recovery of vanadium averaging 39.5 wt. %, exclusive of
the boiler inventory.  In contrast, the EPA system averaged 51.2 wt. %.
The difference between the two systems, roughly 25%, represents a
measure of collection inefficiency - the inability to collect fine
particles of vanadium.  It is evident that the  silicone  oil-impingers
of the ER&E system do not collect all the fine vanadium-bearing
particles which pass the last stage of the Andersen Cascade Impactor.
The latter has a size cut-off (d5g) of roughly 0.4 microns.  These
fine particles also  appear  to pass  through  the  dry  ice-acetone  cooled  condenser.

          The inefficiency of the Greenberg-Smith type impingers
to collect fines is not attributable to use of only our liquid medium
but appears to be general since the same problem was encountered
with the water impingers used in Method 5.  Fine V particles which
are capable of passing through the fiber glass filter  (MSA-1106BH)
also pass cleanly through the impingers.  This was particularly
apparent at long residence time where at most only a few ppm of
vanadium were ever isolated from the impingers even though up to 20%
more vanadium fines  (by weight) appeared to have passed through the
filter.   Relative to the combination of Cascade Impactor/impingers,
the filter has a lower particle size cut-off thus it collects more of
the vanadium-bearing submicron particulate.

      7.2.2  Erratic Vanadium Recovery
            Using Light Arab Residuum

           The runs made with the light Arab fuel oil  at both high
and low firing rates were characterized by extreme  fluctuations in
vanadium recovery and generally poor reproducibility.  As  illustrated
in Figure  7.3 with the exception of a few run^ most of the  inventories
were over 100%.  At  the high firing rate  CT^) recovery  in  the boiler
ranged from 36 wt. %  to 67 wt. % and over the five  runs  averaged 51.3
wt. %.  This was significantly higher than the  inventory obtained
with either of the all-Yen  fuel oils.  Recovery  of  vanadium in  the
stack, based  on the  combined sampling systems, averaged  71.5 wt. %.
Taken  together, the  total of boiler and stack inventories  equaled
122.8  wt. % + 18.8  (95 percent confidence limit).   The confidence limits
are directly  proportional to the spread of the data and  therefore
serve  as a measure of experimental  error.  The  lack of run to run
reproducibility is indicated by the large limits  placed  around  the
average + 18.8%.

-------
                                 - 57 -
           At the reduced firing rate ( I i) recovery in the boiler
was somewhat more uniform but still substantially lower than that at
the higher firing rate.  Directionally, these results confirm the
difference noted in boiler inventory with the two all-Ven fuels
fired at the respective combustion conditions.  The V inventory
from the various passes in the boiler averaged 32.3 wt. % and the stack
sampling systems averaged 73.2 wt. %.  Total recovery therefore
equaled 105.5 + 28.4%  (95 percent confidence limit).  The large
experimental uncertainty is associated in this case with the stack
sampling systems, particularly the inventory made in Run 51 where
recovery amounted to 174% (ER&E system) versus 87.1% (Method 5).

           Since results obtained in the tests with light Arab fuel
were so erratic, it served no useful purpose to carry them through
a statistical analysis.  Therefore, this test series was not included
in the overall analysis of variance (ANOVA).

           The very low vanadium content in the particulate is
probably the main cause for the large fluctuations observed in these
runs.  Since V inventory in these tests was based on the sum of the
individual V analyses of a large number of different samples,
handling and/or analytical errors were probably compounded.  For
example, tests conducted at high firing rates (short residence time),
showed the largest absolute variation in the amount of V collected.
Solids deposited in the boiler varied by as much as 12 ppm  (on fuel)
or roughly 30% of total V input from the fuel.  In the stack, the
range was about 15 ppm.  At this combustion condition, a large amount
of carbonaceous particulate was emitted; the boiler and stack solids
totaled 361 mg/SCM, equivalent to 0.46 wt. % on fuel.  Included in
this total was the fuel ash which accounted for about 0.01 wt. % or
roughly 2% of the particulate weight.  Since the carbonaceous particles
each contain some ash, the concentration of vanadium in the particulate
was low (i.e. the V content in the solids averaged about 1.0  wt. %).

           At long residence time, 79% of the carbonaceous solids
were consumed.  This resulted in a particulate inventory of 77 mg/SCM;
equivalent to 0.10 wt. % on fuel.  The concentration of ash in the
particulate was increased to 10%.  The average vanadium content in the
solids amounted to 7.8 wt. % compared to 1.0 wt. % at the short
residence time.  While the vanadium content was higher, the weight of
solids collected was quite low.  This latter proved to be a limiting
factor, affecting sensitivity of  the analtytical method.  These data
are summarized in Table 7.1.

           Comparing the stack particulate inventory made with  the
entire ER&E system but only the front half of the Method 5, there
was excllent agreement.  The largest difference between the two sampling
trains occurred at the high firing rate  (short residence) and amounted
to 16 mg/SCM or about  7% on total weight.  At the low  firing rate

-------
                                         FIGURE 7.3

                        TOTAL VANADIUM RECOVERED-LIGHT ARAB FUEL
u_
z
o
o
   100-
    80-
o
CQ
Qi
UJ
>
O
o
UJ
O
60-
40-
    20 -
     0
—
Run:
ocatior
^^^m
»— 1
1— 1
Q.
s:
D-
UJ
<~0
n
• — i
•K
i
z
1 — 1
a.
to
UJ
a:
UJ
to
UJ
to
to
Q.
UJ
_l
t— «
0
CO

IT)
CO
CO
I— 1
*
CVJ
1 — 1
*




—



10
f— 1
I-H
*

F
uel V input = 39 ppm
SHORT RESIDENCE//
^^^m


47 78 50 58 65
i 1
\ /
V
2
LONG RESIDENCE/


EPA SAMPLING TRAIN |
| ERE SAMPLING TRAIN |
UJ
to
2
C£
UJ
»— i
O
CQ

^H^M



J
1
—
»—)
-K



•^^
^^^H

r
r
i

0
if
sx
~N^/
1
sr
sx
•M^HI
EPA ERE
111.3 134.4
26.6 23.7
94.0 116.9
11.0 34.9"


60 77 51 59 66
1
\ /
2
:
—

                                                                                             Ul
                                                                                             00

-------
                                 - 59 -
(long residence) where particulate emissions were lower, a relative
difference of about 4% (1 mg/SCM) was obtained.

           The back half or impinger part of the Method 5 train was
also found to contain solids.  However, it was determined that this
material, consisting mostly of sulfate, was an artifact of the chilled
water impingers.  In runs made with the two all-Ven fuels, the same
type solids ,in varying amounts were also isolated from the water
impingers.    Since these solids were not real particulate emitted
during combustion of fuel oil, they were not counted in the total
inventory.  A more detailed discussion of the nature of these
artificial particulates is presented in section 7.7.

     7.2.3  Statistical Analysis to Compare
            Vanadium Recovery	

           An analysis of variance (ANOVA) was made to compare
the influence of stack sampling system  (S) fuel oil (F), combustion
chamber residence time (T), and port location in stack  (L), on
total vanadium recovery.  The ANOVA which is presented in Table 7.2
indicates that fuel type and sample port location are not statistically
significant and therefore do not affect vanadium recovery.  The main
effects, as indicated by the F ratio in excess of the critical table
value P(F >5.57) = 0.05 at fy-^, $32 (degrees of freedom) are residence
time and sampling system.  Since there were no significant interactions,
vanadium recovery may simply be represented as a,function of these
two variables.  The effect of residence time ( T) is to reduce vanadium
recovery by 20% and this appears to be independent of the tested
levels of the other variables (none of the interactions T-F, T-S,
T'F-S, T-L, T-S-L, T-F-S-L are large).  The effect of sampling
system(S) is to reduce vanadium recovery by 11%  (81.3 wt. % to 72.1 wt. %)
and this also appears to be independent of the tested levels of the
other variables (none of the interactions TtS, F«S, T-F-S, S-L,
T-S-L, F-L-S, T-F-S-L  are large).  Residence time, with one degree of
freedom, accounted for 42.6% of the total experimental variation
(sum of squares/total error).  Residence time  ('T ) and sampling system (S)
combined,with two degrees of freedom, accounted  for more than half
of the variation, 54.8%.  Pure experimental error (residual effect)
estimated with 32 degrees of freedom amounted to  35.1% of the total
variation.  The remainder was associated with the interaction effects
which were very minor, amounting to 10.1% of the  total calculated
experimental variation.

           Pooling the two nonsignificant variables, fuel and port
location, a two way analysis of variance was made to obtain a better
estimate of experimental precision.  The new variance with 44 degrees
of freedom was 83.31 whereas the previous variance had  32 degrees of

-------
                                                        TABLE 7.1
Particulate Distribution

•  Boiler Solids

•  Stack Particulate

   -  ER&E Train
        l-10ym
        
-------
                                  -  61 -
freedom and was equal to 90.61.   The twelve additional degrees of
freedom permitted by folding over the factorial design gave about
an 8% improvement in estimating the experimental precision.

           The relative variance (percent of the total experimental
variance of 83.31) associated with vanadium recovery in the boiler
passes amounted to 23%, while the variance associated with vanadium
recovery in the stack  (both sampling systems) accounted for 59%.  The
remainder, 18%, was the unexplained variance and represented the
sum of all unaccountable experimental errors, i.e. errors attributed
to sampling, handling and V analysis.  It may be noted that the stack
measurement part of the experiment had the lowest precision.  However,
this represents the totality of all  treatments, a large part of this
was probably associated with the decrease in ER&E system collection
efficiency particularly at the low firing rate  (/Tl).

           The highest vanadium recovery was obtained using the
EPA Method 5 train at short combustion residence time.  Total recovery
as summarized in Table 7.3 averaged 88.6 wt. % + 5.2  (95 percent
confidence limit).  This value which was based on an overall vanadium
material balance conducted around the boiler probably represents close
to quantitative recovery of V considering the number of measurements
that were involved and the potential inaccuracies.  In comparison
vanadium recovery based on the ER&E stack sampling system was lower
by 7.7% and averaged 81.8 wt. % + 5.2 (95 percent confidence limit).
The difference in averaged recovery between the two sampling systems
was not found to be significant at the 95 percent level using
"Student's  t-test".  However, when tested at the 90 percent level the
difference was significant.  Overall, the two sampling trains gave
acceptable vanadium recovery.

            At  long  combustion chamber residence time, vanadium  recovery
measured  by either  stack  sampling  system was  significantly lower,  however,
the decrease obtained  in  the ER&E  system was  much greater.   Based  on
Method  5,  total V recovery went  from 88.6 wt.  % to  73.9 wt.  % + 5.2
(95 percent confidence limit) representing  a relative decrease  of  16.6%
Based on  the ER&E system,  recovery was decreased  by 24%,  to a level of
52.3 wt.  %.  Comparing the  two sampling  systems the change in vanadium
recovery  amounted to 16%.

      7.2.4 Vanadium-Bearing Particulate Inventory

            Although the amount of  vanadium  collected  by the
sampling systems  was different,  the total  amount  of particulate
collected was  similar  and well within experimental  error.    For the
high ash  (V =  359 ppm)  all-Ven fuel  oil  (Fj). presented  in  Table 7.4
at short  residence  time,  total particulate  emissions  inventoried  in the
ER&E stack sampling system amounted  to 305  mg/SCM equivalent  to 0.39  wt.  %
on fuel oil.   The Method  5  train collected  301  mg/SCM equivalent  to
0.38 wt.  %  on  fuel  oil.   The vanadium content  in  the  total particulate

-------
                                                                         TABLE 7.2
LEVELS OF  FACTORS
                                                             ANALYSIS OF VARIANCE - TOTAL VANADIUM
                                                                 RECOVERY  IN BOILER AND STACK
F f'p- 2 ri><:
, £PA 5££ t *^?"
U ,«.«.«• ^ ^
D / 1, 7 .T^
^^ f.J"«e. (
•J-5~J

S.S1



S"-$7
- g-j
'
£'57
5.60334
6.45336 * / sy 1 -r -.
1.84081 CJPVC.' ^^''•L "r£-&'OP*%£ /!*/'
0 18749 i *- y
90'61 - s' AM ^Jtl^ly

-------
                                             - 63 -
                                           TABLE 7.3

                  TOTAL VANADIUM RECOVERY AS AFFECTED BY STACK SAMPLING SYSTEM
                             AND COMBUSTION CHAMBER RESIDENCE TIME
                  EPA Method 5
                                         ER&E Train
                                                                   t Test

/
/
         N
             88.6 wt.  %
(Short    \J (83.4 - 93.8 wt.
Residence/1
V
Time
        /
P (.95)
81.8 wt. %

(76.6 - 87.0 wt. H

P (-95)
t calcd. =1.54

0 = 44

At 95%,Critical Table

Value =1.96

Difference not significant

At 90% difference is
 significant
             73.9 wt.  %

             (68.7 - 79.1 wt.

             P (-95)
                                     62.3 wt. %

                                     (57.1 - 67.5 wt. %)

                                     P (-95)
                                                               t calcd = 4.41

                                                               0 = 44

                                                               At 95%,Critical Table

                                                               Value = 1.96

                                                               Difference is significant
             NOTE:
             Experiment  variance; =  s  = 83.31
             Standard  error  =  s  = f-3»~. = 2.63
                               X  I - \S

             95  percent  confidence  interval  = t  (.025,44)= 1.96 x 2.63 + X


             Probability -5.2  <_ X <_ +  5.2  =  0.95

-------
                                                   TABLE 7.4
Particulate Distribution

•  Boiler Solids

•  Stack Particulate

    - ERE Train

      >10um
      1-1OMm

                 Total

    - EPA Train

      Probe and Cyclone
      Filter

      Water Impingers
                 Total
AVERAGE PARTICULATE AND VANADIUM INVENTORY
.HIGH ASH ALL VENEZUELAN FUEL OIL (Fi)
Short Residence Time
Particulate
mg/SCM
222
104
145
56
305
184
117
(301)
8
309
Vanadium
Analysis
wt. %
3.68
1.81
5.36
10.09
2.43
10.67
Vanadium
Content
ppm (on fuel)
104
24
99
72
195
57
159
(216)
Trace
216
Long Residence Time
Particulate
mg/SCM
61
3
12
48
63
7
65
(72)
5
77
Vanadium
Analysis
wt. %
9.91
5.23
15.04
19.95
3.36
21.26
Vanadium
Content
ppm (on fuel)
77
2
23
122
147
3
176
(179)
Trace
179
  NOTES;

  (1)   All  data corrected to 3% 0 .

  (2)   Results from individual runs  summarized in Tables 1-4  ,  Appendix 5.

-------
                                  - 65 -
collected by the ER&E train averaged 5.02 wt. % compared to 5.63
wt. % in Method 5.  On an absolute basis, this difference amounted
to 21 ppm V and was mainly associated with the filter catch.  Typically the
smallest sized particulate have the highest vanadium contents.  In this
case, however, the particulate on the filter had a higher V concentration than
particulate isolated in the submicron size range, 10.67 wt. % versus
10.09 wt. % V.  While this may appear paradoxical, the situation could
arise if the filter collected finer submicron particles than the
impingers, i.e., the filter had a much lower size range cut-off.  If
it is assumed that the missing 21 ppm of vanadium (as V205>  passed
through the ER&E system, then the difference in particulate inventory
necessary to account for it would amount to about 3 mg/SCM.  Even
if the V content in the missing particulate was 10.09 wt. %, the
difference would amount to only about 17 mg/SCM or about 7% on  the
total inventory.  This magnitude is well within the expected experimental
variation and would not be detected.

           Some insight into the nature  of particulate emissions was
obtained in previous company sponsored studies.  Table 7.5  shows  the
C, N and ash composition of the particulate  in the different size
ranges.

                               TABLE 7.5

               PARTICULATE  COMPOSITION  INFLUENCED BY  SIZE
                         (High  Sulfur Fuel  Oil)

Particulate          	Analysis, wt. %
Fraction

>10 micron
1-10 micron
<1 micron
            The largest particles,  mostly cenospheres,  have a high
 carbon content, 79.4%, and a relatively low ash content, about 10%.  The
 intermediate size range has a slightly lower carbon content and a somewhat
 higher ash content, about 18%.  The submicron particles have a relatively
 low carbon content, only 30.1% and a much higher ash content, about 57%.
Carbon
79.4
70.7
30.1
Nitrogen
1.8
1.2
0.3
Inorganic Ash
10.1
18.1
57.1

-------
                                 -  66 -
           Analysis of the ash residue from the three size ranges,
Table 7.6  indicated that a large part of each was vanadium.  Expressed
relative to fuel oil the vanadium oxide in the particulate ranged
from as little 24 ppm in the very large particles to 137 ppm in the
smallest particles.
                               TABLE 7.6

                DISTRIBUTION OF V20s IN PARTICULATE ASH
V2°5

Na00
                 Wt. % in Ash Fraction
                          1-10       <]
41.6

 1.7
63.7

 1.8
58.2

 3.4
Wt. % on Fuel in Size Range
            1-10        <1
.0024

.0002
           0.0110

           .0003
.0137

.0008
           At long residence time, the total difference in particulate
inventory between the two sampling systems amounted to about 14%, which
on an absolute basis was only 9 mg/SCM.  The ER&E system collected
63 mg/SCM of particulate,(0.08 wt. % on fuel)versus 72 mg/SCM (0.09
wt. % on fuel)for Method 5.  The net effect of increased residence time
is that most of the carbonaceous residues burn-out, so that the total
weight of particulate is substantially reduced.  Additionally, the
size distribution of these particles is shifted toward the submicron
range.  Ash material, mainly vanadium oxide, which would normally have
been distributed among the carbonaceous particulate is probably
vaporized into the combustion gas stream, then recondenses in the cooler
part of the boiler as submicron particles.  As may be noted in Table 7.4,
about 83% of the vanadium particles collected by the ER&E sampling system
were in the submicron size range compared to 37% at short residence time.
This shift in size distribution is also borne out by the Method 5 train.
Vanadium collected on the glass filter (<5 micron) amounted to 74% at
short residence time compared to 98% at long residence time.  The
concentration of vanadium in the fineparticulate was also substantially
higher than in the previous runs at/YQ.   Submicron particulate
collected in the silicone oil impingers had a vanadium content of
20 wt. % and those collected on the fiber glass filter had a V content
of 21.3 wt. % versus roughly 10.5  wt. %.   Since  the  concentration of

-------
                                  -67-
V in the particulate is high, the loss of only a small amount of these
fine particles may have a critical effect on the overall V recovery.
For example, the difference between sampling system particulate inventory
amounted to 9 mg/SCM or on a relative basis about 14%.  If the
vanadium content in this amount of particulate was roughly 27 wt. %,
which is not much higher than the concentration actually assayed, it
would account for the 34 ppm difference obtained in vanadium inventory.
On the other hand in the Method 5 sampling system, if the fines which
passed through the filter were V205, an increase in total particulate
inventory of about 10% would be sufficient to bring vanadium recovery
up to 90%.

           In tests made with the intermediate ash, all-Ven fuel oil
(F2> summarized in Table 7.7, total particulate collected by Method 5
(front half only) and ER&E stack sampling trains also showed good
agreement.  Directionally the results were very similar to those
obtained with fuel F^.  At short residence time particulate inventories
were identical, 127 mg/SCM versus 128 mg/SCM, equivalent to 0.16 wt. %
on fuel oil.  However, the average vanadium content was higher  in the
solid collected by Method 5, 4.64 wt. % versus 5.34 wt. %.  This
difference on an absolute basis amounted to 12 ppm V or roughly 15%
of the total inventoried in  the stack.   Again, it may be concluded
that the loss of vanadium from the ER&E train was associated with
the passage of a relatively  small amount of fine, vanadium-bearing
particulate out the impingers.  This situation also occurred in the
runs made at long residence  time.  However, in this case, it was
further aggravated by the shift in size distribution to nearly  all
submicron particles and the  higher vanadium content in these fines.
The EPA system collected about 13% more particulate, which on an
absolute basis amounted to 5 mg/SCM.  The vanadium content in the
particulate was also higher  averaging 14.93 wt. % versus 12.43  wt.  %.

     7.2.5  Size Distribution of
            Vanadium-Bearing Particulates

           The size distribution of  the vanadium-bearing particulate
which deposited in the boiler was also influenced by combustion
conditions.   At short residence time, the vanadium content in  these
deposits averaged 3.9 wt. %.  Based  on a similar concentration  in the
stack, 3.3 wt. %, the particulate were predominantly  in the 1-10 micron
size range.  At long residence time, the boiler deposits averaged
7.6 wt. %,  indicating a mix  of submicron and  1-10 micron particles.

           The size distribution of  particulate in the stack as
previously noted was significantly altered by combustion conditions.
The results which have been  determined with  the ER&E  sampling train are
summarized  in Table 7.8.  At short residence  time, irrespective of
fuels, approximately 29% of  the  total particulate weight was emitted
as cenospheres  (>10 microns) while 76% of  the particulate was emitted
in the size range >1 micron.  Submicron  particulate amounted to 24% of
the emissions.  At the  increased residence  time,  combustion of  the

-------
                                                 TABLE 7.7
Particulate Distribution

•  Boiler Solids

•  Stack Particulate

    - ERE Train

      >10ym
      1-lOym

                 Total

    - EPA Train

      Probe and Cyclone
      Filter

      Water  Impingers-
                 Total
AVERAGE PARTICULATE AND VANADIUM INVENTORY
INTERMEDIATE ASH ALL VENEZUELAN, FUEL OIL (F2)
Short

Particulate
Mg/SCM
89
30
57
40
127

51
73
(128)
13
141
Residence
Vanadium
Analysis
wt. %
3.88
1.57
3.30
8.83


2.00
7.96



Time
Vanadium
Content
ppm (on fuel)
44
6
24
45
75

13
74
(87)
Trace
87
Long

Particulate
Mg/SCM
37
4
7
25
36

5
36
(41)
10
51
Residence
Vanadium
Analysis
wt. %
7.64
1.96
5.61
16.01


1.57
16.79



Time
Vanadium
Content
ppm (on fuel)
36
1
5 i
51 ON
57
i
1
77
(78)
Trace
78
NOTES;

(1)  All data corrected to 3% 0-.

(2)  Results from individual runs summarized in Tables 5-8, Appendix 5.

-------
                                       -  69 -
                                   TABLE 7.8
               RELATIVE SIZE DISTRIBUTION OF STACK PARTICULATE

                 NOTE:   Based on Data From ER&E Sampling System
Size Distribution

    >10Mm
    l-10ym
    10ym
    1-lOym
                        100%
             Fuel
                                             ~ Intermediate Ash, All-Ven
24
45
31
8
32
60
11
19
70
2
9
89
                100%
                100%
                100%
                                     Fuel FT - Low Ash, Light Arab
    >10ym
    1-lOum
 28
 49
 23
100%
 13
 28
 59
100%
 11
 18
 71
100%
  3
  3
 94
100%

-------
                                  -  70  -
carbonaceous residues was facilitated and resulted in a pronounced
shift in particle size distribution to the submicron range.  At
this condition, roughly 72% of the total particulate were emitted
as submicrons compared to 24% in the previous run.  The amount of
cenospheres  was decreased by almost 70% and now accounted for
only 9% of the total weight.  As cenosphere and cenosphere derived-
residue  burn out the volatile ash material present in these particles,
particularly vanadium, is liberated into the gas stream to condense
in the cooler part of the boiler as fine particles.  Only 2% of
the vanadium remained associated with these large particles compared
to 11% at the short residence time.  Approximately 12% of the vanadium
was in >1 micron particles while 88% was in the submicron particles.
From the preceding it may be concluded that at the long residence
time, sampling system collection efficiency is influenced by the
presence of a large amount of submicron particles.

7.3  Sampling System Collection Efficiency

           Vanadium collection efficiency as used in this report
is the ratio of the amount of vanadium collected in the stack
sampling train divided by the amount of vanadium in the fuel mjLnus
the amount of vanadium that deposited in the boiler
Thus, it is apparent that for a particular level of vanadium recovery,
the calculated collection efficiency will be lower because of the
different basis used.  Collection efficiency will be equal to 100%
only if all of the vanadium introduced into the boiler from the
fuel is recovered.   Any vanadium which is not accounted for in the
deposits from the passes is assumed to enter the stack.  If the
amount of vanadium inventoried by the stack sampling is insufficient
to close the balance then the difference is attributed to the
inefficiency of the collection system for collecting V fines.   In this
study, the highest vanadium recovery averaged 88.6 wt. % using the
Method 5 train at short residence time.  However, this probably
represents the best material balance obtainable with the resources
employed; primarily the boiler tube inserts.  Consequently, there is
little direct data to indicate where the loss of vanadium, in this
particular case 11.4% of the input, has occurred.  Assigning it
entirely to the sampling system as has been done in Table 7.9, the
resulting vanadium collection efficiency of 83.9% represents the
minimum value for Method 5.  If the entire loss of vanadium occurred
in the boiler passes then collection efficiency of Method 5 would
naturally amount to 100% and the other values could be scaled up
accordingly.  The change in collection efficiency from 83.9% with
Method 5 to 74.3% with the ER&E train is a real effect.  This
difference is attributed to the inability of the hot silicone oil
impingers to collect all of the very fine vanadium particles.

-------
                                 - 71 -
           At long combustion chamber residence time, the collection
efficiency of Method 5 decreased from 83.9% to 66.2%, roughly
a change of 21%.  However of this total, more than half, 14%, was
related to a lower vanadium inventory in the boiler passes.  If the
same amount of vanadium had been deposited in the passes at both
residence times,sampling system collection efficiency would be
equal to 72.4%.  The collection efficiency of the ER&E system relative
to Method 5 decreased by an additional 23% to a level of 51.1%.  At
short residence time, the relative loss in efficiency was only 11%.
The steeper decline in ER&E stack train performance at long residence
time is associated with the emissions of significantly smaller
vanadium particles which pass unchecked through the hot oil impingers.
In contrast, the fiber glass filter used in Method 5 retains more of
these fine particles.  Vanadium fines which do pass through the filter
are not collected in the water impingers again indicating the inefficiency
of this type of system to collect submicrons.

           To sum up, the fiber glass filter used in Method 5 train
has a significantly lower particle size cut-off than the high velocity
type Greenberg Smith impingers.  With this type of filter, Method 5
collection efficiency for vanadium particles in combustion-produced
flue gas will amount to, at the very least, 66-84%.  Large boilers
which normally have bigger combustion chambers and therefore longer
combustion chamber residence time will tend  toward  the  lower
vanadium collection efficiency.

7.4  Vanadium Input - Fuel Analysis

           The first element of the material balance involved establishing
the vanadium content of the three fuel oils.  Although each fuel
was drawn from a bulk lot, they were segregated and stored in 55 gallon
drums.  Each test was run with a new drum of fuel oil.  Thus over the
course of the program 36 drums of oil were sampled and analyzed for
vanadium content prior to use.  As illustrated in Table  7.10 the all-Ven,
high ash fuel designated F-^ had a vanadium content averaging 359 ppm
+ 8.4 (95 percent confidence limit) which expressed as V^Oij accounted
for about 70% of the fuel ash.  Nickel podium and iron were the other
components in the ash.  The intermediate ash all-Ven fuel oil, F2, had
a vanadium content averaging 149 ppm + 4.2  (95 percent confidence limit).
Expressed as V205, it accounted for about 66% of the ash weight.  The
light Arab low ash fuel designated F3 had an average vanadium content
of 39 ppm + 2.0 (95 percent confidence limit) accounting for about 70%
of the ash, on  the basis of V20^.  Relative analytical precision was
about the same, roughly 4% for the vanadium analysis in the three fuels.

-------
                                  - 72 -
                                     TABLE  7.9
                        SAMPLING  SYSTEM  COLLECTION  EFFICIENCY
                                EPA Method  5
                        ER&E
  Short Residence
   Time,
83.9%
74.3%
 Long Residence
   Time
66.2%
51.1%
           As the precision of these analyses represented drum to
drum variation only, fuel samples were drawn periodically from one
drum and analyzed.  This procedure served to test the hypothesis
that the drum samples came from the same population and therefore
the average value was representative of the vanadium content in the
test fuel lot.  The mean, as shown in Table 7.11, for the 11 vanadium
determinations was 351 ppm with a standard deviation of 23 ppm or
on a relative basis, 6.5%.  Thus from a statistical standpoint the
value of 359 ppm representing the average of all drums adequately
describes the vanadium content of test fuel F]_.  By inference, the
same holds true for fuels F2 and F3.  The vanadium content of the
fuel oil multiplied by the fuel consumption gives the vanadium input.

-------
                  - 73 -
                TABLE 7.10
Drum to Drum Variation of Vanadium Analysis
    Mean  =   x

              s
  Std. dev.  =  x

Std. Error  = 
-------
                                           - 74 -
                                       TABLE 7.11




                    Variation of Samples Within A Drum of All-Ven Fuel Oil
Date:     8/21   9/23   10/24   11/27   12/31   1/30   1/2/20   2/28   3/28  4/30  5/30




V,ppm:    330    300     330     373     350    358     365     353    360   380   360









          *11 = 350'8 PPm





          sx  = 22.84






          *%  = 6.9








          Relative std. dev.  =6.5%

-------
                                - 75 -
7.5  Vanadium Recovery in the Boiler

     7.5.1  Preliminary Tests - Evaluation of Boiler
            Pass Tube Insers	

           Several preliminary runs were made to determine how well
the heat exchanger tube inserts worked and if replication of particulate
inventory was possible.  Two removable tube inserts were randomly
placed in each of the three heat exchanger passes.  Tests were then
made using firing rates of 1 Ib./min. and 2 Ibs./min. at 10% excess
combustion air.  Results are presented in Table 7.12 and indicate that
the inserts provide a reliable measure of particulate fallout in the
boiler.  To facilitate comparisons, the weight of particulate was
divided by the calculated combustion gas volume CVL95 SCFDB/lb. fuel)
then multiplied by the total number of tubes in the pass to give
mg/SCMDB  (corrected to 3% 02).  Runs 33 and 34 may be directly
compared since a similar firing rate, 1 Ib./min. was used.

           The difference between tube inventories in the same pass
ranged from less than 1% to as much as 13%.  However, for all tests
the average was 5%.  Run to run variation of fallout averaged about 9%.
This somewhat higher difference was most likely associated with some
facet of boiler operation (i.e. slightly different firing rates, stability,
etc.) rather than handling errors.  The major variation in particulate
fallout between runs occurred in pass 2.  This pass had the most
surface area and therefore was prone to fluctuations or changes in
boiler operation.  For example, comparing Run 33 with Run 35, the fuel
firing rate was increased from 1 to 2 Ibs./min.  The largest change
in particulate was in pass 2 where the amount of fallout tripled in
weight.
            In Run 35,  the effect on particulate fallout on changing
 combustion chamber residence time may be noted.  As  a result of
 increasing the fuel firing rate and therefore decreasing residence
 time,  the amount of solids deposited in the passes was doubled.  Since
 burnout of a carbonaceous particle in the combustion zone of a boiler
 is believed to take place by diffusion of oxygen to  the particles'
 surface - a fairly slow process, shortening residence time increases
 the probability that the particle will escape unburned.

-------
                                             TABLE 7.12
                   VARIATION OF PARTICULATE DISTRIBUTION  IN  BOILERfpRELIMINARY TESTJ
Pass
Pass 1 (firetube)
Pass 2
Pass 3
Pass 4
             Total
         Run 33
Total Particulate/Pass
	mg/SCM*	
  A                 B
 29.2
 18.0
 11.7
 99.0
          40.1
 31.4
 18.1
 12.9
102.5
                     Run 34
            Total Particulate/Pass
           	mg/SCM*	
             A                 B
 35.2
 21.0
 12.0
110.5
                                       Run 35
                               Total Particulate/Pass
                              	mg/SCM*	
                                A                  B
42 .3
         36.9       106.3
Sample Contaminated  36.9
         12.1       __1±1
                    226.7
                                                   82.2
105.0
 32.3
  1.4
220.9
Fuel Consumption; Lb.
 - Inserts in Place
 - Total Run
Firing Rate, Lbs./Min.
         194.50
         320.50
           1
                    198.25
                    305.75
                     1
                                       373.50
                                       556.25
                                         2
*Corrected to 3% 0,

-------
                                 - 77 -
           Residence time also had considerable influence on the
concentration and distribution of vanadium in the boiler solids.
For a particular combustion condition, as shown in Table 7.13, there
was little variation between runs in vanadium content of particulate
from the same pass.  On average the difference amounted to less than
5%.  This is on the same order as the analytical precision for the
Atomic Absorption method.  Since somewhat more variation between
runs was found in the absolute weight of particulate, the amount of
vanadium collected also varied.  In these preliminary trials the
difference in the total amount of V inventory between runs was on
the order of 20%.  Considering the relatively small amount  of solids
collected, this difference is about as good as can be expected.

           Increased combustion chamber residence time serves to
promote burnout of cenospheres.  Given sufficient residence time,
all of the carbonaceous matter in the particle would be consumed
and only ash (vanadium particles) would be emitted into the combustion
product gas stream.  With short residence time, i.e., high firing
rate, most of the ash is retained in the cenosphere.  Thus the
deposits in the boiler, depending upon combustion conditions should
have distinct vanadium contents.  In runs 35 and 36, the solids
isolated in the boiler averaged 3.54 and 3.31 wt. % V, respectively.
Almost all the vanadium fallout,99%,was in the first three passes.
In Runs 33 and 34, fallout in the first three passes accounted for 90%
of the total deposits but 78% was in the first two passes alone.
The vanadium content in these solids averaged 7.07 and 7.64 wt. %
respectively or about double the previous runs.  These results are
summarized in Table 7.13.

           Some idea of the nature of the deposits in the various boiler
passes were obtained from their elemental analysis.  In Table 7.14
C, H, N and S analysis of deposits from Run 33 are presented.

           The solids from pass 1 and 2 which had the highest vanadium
content roughly 8% (ash) also had the lowest carbon content, averaging
about 53%.  In passes 3 and 4 when vanadium content averaged about 4.5
wt. % (lower ash) carbon content was considerably higher and amounted
to 74%.  While a sulfur analysis is presented, most of the sulfur
is believed to be associated with the ash as sulfate.  Sulfuric acid
would not be a factor since the high temperature in the passes
would preclude its formation.  Sulfur content appears highest in
the solids containing the least amount of carbon, therefore, the
most ash.

-------
                                   - 78  _
                               TABLE 7.13

    VARIATION OF VANADIUM DISTRIBUTION IN BOILER - PRELIMINARY TESTS
    	(ALL-VEN HIGH ASH RESID FUEL)
   FUEL FIRING RATE 1 Ib./MIN

                             Run 33                     Run 34
                    Particulate Inventory      Particulate Inventory
Boiler Deposits     mg/SCM   Wt.%V   ppm V     mg/SCM   Wt.% V   ppm V

Pass 1 (firetube)     40      7.5     38         42      7.8       42
Pass 2                38      8.2     32         36      9.2       42
Pass 3                18      4.3     11         21      6.2       17
Pass 4                12      4.8      9         12      4.7      	7_
             Total   100              90        111               108
t  FUEL FIRING RATE 2 LBS./MIN.

Boiler Deposits
                           Run 35                       Run 36
Pass 1 (firetube)     82      3.0     31         75       2.7      26
Pass 2               106      4.0     54         93       3.8      45
Pass 3                35      3.3     15         27       3.1      11
Pass 4               	1      6.6    	1        	2       5.9       1
             Total   224             101        197                83
                               TABLE  7.14

            PARTIAL ELEMENTAL COMPOSITION OF BOILER SOLIDS
                (FUEL FIRING RATE 1 LB./MIN. - RUN 33)
                      Elemental Composition
Boiler Deposits         C     H     N     S

Pass 1 (firetube)      47.1  0.8   0.9   10.2
Pass 2                 58.8  0.8   1.0    8.0
Pass 3                 76.8  0.7   1.2    5.9
Pass 4                 72.2  0.8   1.0    7.0

-------
                                - 79 -
   7.5.2    Factorial Program Results - Vanadium Recovery
            at High Firing Rate	

           Once it was established that the tube inserts provide
a reasonably consistent measure of vanadium particulate fallout in
the boiler, the factorial experimental program was conducted.  In
each run, total vanadium was determined in the boiler solids and
in each stack sampling system.  In Figures 7.4 - 7.5 bar
charts graphically illustrate vanadium recovery in the four passes
(V recovered in boiler solids as a percent of V input from fuel)
as a function of fuel type and combustion chamber residence time.
Figure 7.4 summarizes the tests conducted at short residence time
(r\ o)> i-e. 2 Ibs./min. fuel firing rate.  It shows that there was
no difference in percent recovery of vanadium between the high and
intermediate ash all-Yen fuels.  The absolute amount in the deposits,
however, differed considerably because of the difference in the
amount of V in the fuel oils.  Over the twelve tests, recovery in
the boiler tubes averaged 29.3 wt. %.  Run to run fluctuations were
minimal, amounting to less than 30 ppm.  In contrast, vanadium
recovery in the five runs made with the low ash, light Arab fuel oil
averaged 51.3 wt. % of the V in the fuel and showed considerable
variation between runs.  This increase in variability was further
indicated by the large standard deviation.  In this case standard
statistical techniques, i.e., F-Ratio and Bartlett's chi square tests
were used to compare variability of the three groups of runs.  It was
established that the variances associated with the means from the
all-Ven fuel oils were from  the same population distribution but were
different from the variance obtained in the light Arab runs.

           A primary cause for 'the lack of repeatability in tests
conducted with light Arab fuel oil was the low concentration of
vanadium in the boiler solids.  The V content as shown in Table 5.15
ranged from as little as a few tenths of one percent up to 2.5 wt. %
but averaged about 1.0  wt. %.  On the other hand, the V content of
solids from the high ash all-Ven ranged from 2 wt. % to 10 wt. %,
averaging 3.7 wt. %.  Those  from the intermediate ash all-Ven fuel
varied from 2.0 wt. % to 5.5 wt. %, averaging 3.9 wt. %.

           There was more particulate deposition in the boiler passes
with light Arab fuel oil even though it was lower in asphaltenes
(hexane insolubles) than the all-Ven fuel ^2-  Tne all-Ven fuel F^,
had the highest asphaltene content and produced the most solid emissions.
The propensity of heavy oil  to form carbonaceous solids is directly

-------
CO
UJ
CO
CO
cc.
o
CO
o:
LU
>
o
o
LU
Cf.
                                          FIGURE 7.4

                      VANADIUM  RECOVERY IN BOILER - SHORT RESIDENCE CASE
y   60
    50
    40
     30
20
     10
-
FUEL Fi ALL-VEN.
HIGH ASH, V = 359

Epm V 108 91 107 105 120 94
FUEL F2 ALL-VEN.
MED. ASH, V = 149
43 57 51 32 40 43

F
L
20
Mean = X" = 29.0 WT. % I = 29.7 WT. %
Std. Dev. = s = 2.9 s = 5.9
X X

—

—
MMH
P4
P3
P2
^^mm
PI



	
m^mm


^^mt
••^^
^^^m



_^_
««
^M^M



^•i^

^^^M



^—
^^^


P4
P3
P2
•^M
PI






^^•i




	

^•^H



^^M
^••M



—
	



•^•H
^•^^
^i^^





^^^m
P4
P3


P2
PI





rer
)W /
7F

^^^M




F3 LT. ARAB
SH, V = 39
14






•^IMM
•^^^

	



23
^^^^
^^^^





17
-
X" = 51.3'
s = 12.1
X




^^^M
•^^H
	

	
:
—

—
                                                                                              00
                                                                                              o
        Run: 55  57  68  52  71  79
                                  48  69  75 45  63  74
47  78  50  58  65

-------
                                  - 81 -
related to its asphaltene content.  However, this relationship
may be modified by the type and molecular weight range of these
complex multi-ring structures.  The light Arab fuel oil is believed
to contain highly  aromatic molecules which may form very refractory
carbonaceous solids in the combustion chamber.  The higher output
of carbonaceous solids containing low ash (V) dilutes even further
the concentration of vanadium in the particles.

     7.5.3   Vanadium Recovery Made at Low Firing Rate

           At long residence time.burnout cf carbonaceous material
was promoted, as illustrated in Figure  7.5.  Therefore, particle
size was shifted toward the submicron and the relative ash content in the
particles was increased.  The increase of V concentration in the
solids produced from the light Arab fuel improved run to run
repeatability.  There was no statistical difference in precision
between the three groups of runs.  Considering overall V recovery
for the five runs on light Arab fuel, the average was 32.3 wt. % of the
fuel V.  This was considerably higher than the recoveries made with
fuel F-^, which averaged 21.4 wt. % or fuel F£ which averaged 24.0
wt. %.  Combining the latter two, the overall average for all-Ven
fuels was 22.7 wt. %, or about 40% less than for the light Arab.

           The concentration of vanadium in the boiler deposits
from the long residence time runs were considerably higher than
obtained at short residence time.  The average V content of deposits
in the boiler burning the high ash all-Ven fuel (F^) was about
9.9 wt. %.   Roughly 80% of the total vanadium fallout was obtained
in the first two passes while close to 95% of the total was in the
first three passes.  This latter value also coincided with total
particulate inventory in the first three passes.  At short residence
time, the distribution was quite similar , 93% of the V in the boiler
was recovered in the first three passes and about 95% of the particulate
was obtained there.

           Combustion of the intermediate ash all-Ven fuel oil
resulted in deposits having an average V content of 7.6 wt. %.  In
this case about 90% of the V was accounted for in the first three
passes along with 87 wt. % of the particulate fallout.  At short
residence time the distributions were similar, 90 wt. % and 89 wt. %
respectively.

           The light Arab fuel oil produced boiler solids which had
an average V content of 2.0 wt. %.  The V and total particulate distributions
with this fuel were somewhat more erratic and showed less correspondence.
In the first three passes approximately 87% of the V and 83% of the
particulate were collected while at short residence time 93% of the V
and 89% of the particulate were collected.  These data are summarized  in

-------
                                  FIGURE 7.5
               VANADIUM RECOVERY IN BOILER - LONG RESIDENCE CASE


60
LJ
o
^ 50
i—
CO
LJ
CO
< 40
n

Qi
LJ
1
_J
0
CD
z 30
a
1 i 1
LLJ
on
LJ
>
8 20
LJ
Q-
— i
Q
1 "
n
FUEL Fi ALL-VEN.
HIGH ASH, V = 359
_ppm V 78 79 82 65 100


—
-

Mean = I = 21.4 WT. %
~Std.Dev. s = 4.1
X

-


^^ ^^

-

_


_

—
-
—


_P4
P3
^••M


P2
"PI













^^•^


^IBM













•^^H









•^^

"^


MMM














^i^MI













57



















M^M
^^^

^^H
FUEL F2 ALL-VEN. FUEL F3 LT. ARAB
MED. ASH, V = 149 LOW ASH, V = 39
35 27 40 44 32 37 13 14 13 13 10 -


^™
-

X" = 24.0 WT. % X" = 32.3 WT. %
s = 4.0 sx = 3'9
X
^_^^














P4
^—
P3


MM
P2
^•^
PI








^^

—
—

^^f^






—
^^™

•^




^^B











^~~







^^^^
^^•1












^^^





^••M







^"^

^"™


— I


•^^H









P4

^^^
P3





P2

^^^m
PI















^•^^



HI^B^







"

























•^M
























^^•^



^^











^^^


^ ••

__


-


-

—











•

00
1












Run:  54  56  67 53  72  80
49  70 76  44 62  73
60  77  51  59  66

-------
                               - 83 -
Table 7.16.  While the concentration of V in the particulate obviously
increases the amount of particulate actually collected is so very low
that handling and work up constitute the major source of errors.  This
is further indicated in Run 65 where duplicate samples were analyzed
for V.  The largest difference between analyses was about 22%, however,
the V content of the samples were quite low amounting to 0.61 wt. %
and 0.75 wt. %.  On the total amount of V recovered, i.e. 17 ppm the
difference in analysis was only 1 ppm or about a 6% relative error.
      7.5.4  Further Consideration of
            Total Vanadium Kecovery

           Regarding overall vanadium recovery at the long residence
time, the fact that it was lower than at short residence time is
somewhat surprising.  Considering the temperature differential existing
between the flue gas and heat exchanger tube surfaces and the
overall length of these tubes (equivalent to a probe 40 feet long)
it was thought that thermal diffusion would cause deposition of a
large amount of fine particles.   As previously noted in section 7.2.5
at long residence time upwards of 80% of the vanadium-bearing
particulate were in the submicron size range.  Since these particles
had the highest V content, total recovery from the passes should have
been equal to or exceeded recovery obtained at short residence time.
Temperature differentials are highest between the flue gas and
tube surfaces in passes 1 and 2.  Therefore thermal forces would be
maximized.  Consistent with this mechanism at long residence
time, particulate with the highest V content were deposited in the
first two passes.  In passes 3 and 4 the V contents of the solids
were lower and more in line with the concentration found in the
particulate deposited at short residence time.  There was no apparent
V concentration gradient in the particulate deposited in the passes
at the latter combustion condition.

           To determine if there was any interaction between fuels
and vanadium recovery, a computer analysis of variance (ANOVA)
procedure was employed.  However, only the runs using the all-Ven
fuels FI and F2 were compared.  Results from the light Arab fuel
experiments were so different that these tests could not be considered
as having come from the same population as the others.  In Table 7.17
the analysis of variance is reproduced.  Based on an F ratio test,
the only significant variable affecting vanadium recovery in the
boiler was residence time; fuel type had no effect.

-------
                                                                      TABLE  7.15

                                                     PARTICULATE AND VANADIUM INVENTORY  IN  BOILER
                                              (LONG COMBUSTION CHAMBER RESIDENCE TIME  -  0.5 LBS./MIN.  F.R.)
f FUEL FI, ALL-VEN, HIGH ASH (V



Total
= 359 PPM
Run 54
Vanadium
Particulate Analysis
Particulate Distribution
Boiler Solids
Pass 1 (firetube)
Pass 2
Pass 3
Pass 4
Total
mg/SCM

12
29
11
7
59
• FUEL F?, ALL-VEN, MEDIUM ASH,


Pass 1 (firetube)
Pass 2
Pass 3
Pass 4
Total
• FUEL FT, LIGHT ARAB,
Boiler Solids
Pass 1 (firetube)
Pass 2
Pass 3
Pass 4
Total

13
9
18
10
50
LOW ASH

12
19
12
6
49
W Wt.%

10.70
12.26
7.80
7.00

(V = 149 PPM)
Run 49
5.33
9.41
4.60
3.00

(V = 39 PPM)
Run 60
2.70
1.91
1.07
2.62



Vanadium
Content
ppm
(on fuel)

16
45
11
6
78


9
11
11
4
35


4
5
2
2
13



Run 56
Total
Part. V
mg/
SCM

13
24
8
6
51


14
10
4
3
31


3
24
12
7
46
Anal.
Wt.%

10.12
14.34
10.93
8.93


Run 70
2.53
11.97
8.73
7.90


Run 77
5.60
2.17
1.88
3.29



V
Con-
tent
ppm

17
43
12
7
79


4
16
4
3
27


2
7
3
2
14



Run 67
Total
Part. V
mg/
SCM

19
27
14
4
64


7
12
7
3
29


10
22
8
8
48
Anal.
Wt.%

11.92
12.00
4.91
6.96


Run 76
11.67
13.08
6.78
6.32


Run 51
0.73
3.48
0.98
1.05



V
Con-
tent
Ppm

28
42
8
4
82


11
20
6
3
40


1
10
1
1
13



Run 53
Total
Part. V
mg/
SCM

16
31
10
8
65


19
7
18
6
50


5
18
13
11
47
Anal.
Wt.%

7.82
9.15
6.51
5.62


Run 44
4.90
9.71
8.40
6.87


Run 59
4.06
2.54
1.34
1.29



V
Con-
tent
ppm

15
36
8
6
65


12
8
19
5
44


3
6
2
2
13



Run 72
Total
Part. V
mg/
SCM

23
26
11
5
65


8
13
6
3
30


13
20
15
11
59
Anal.
Wt.%

10.29
15.58
9.62
9.60


Run 62
6.88
11.51
5.49
4.64


Run 66
1.82
1.82
0.78
0.85



V
Con-
tent
ppm

30
51
13
6
100


7
19
4
2
32


3
5
1
1
10



Run 80
Total
Part. V
mg/
SCM

20
25
12
8
65


8
11
7
4
30







Anal.
Wt.%

4.72
9.90
5.53
4.70


Run 73
8.65
11.71
8.24
9.40










V
Con-
tent
ppm

12
32
8
5
57


9
17
7
4
37







NOTES:

   All weights corrected to
                               V

-------
                                                                      TABLE 7.16

                                                     PARTICULATE AND  VANADIUM INVENTORY IN BOILER
                                              (SHORT  COMBUSTION  CHAMBER RESIDENCE TIME - 2 LB./MIN. F.R.)
• FUEL Fi, ALL-VEN, HIGH ASH (V = 359 PPM)


Total
Run 55
Vanadium
PartiQulate Analysis
Particulate Distribution m
Boiler Solids
Pass 1 (firetube)
Pass 2
Pass 3
Pass 4
Total
• FUEL F2, ALL-VEN, MEDIUM

Pass 1 (firetube)
Pass 2
Pass 3
Pass 4
Total
• FUEL F3, LIGHT ARAB, LOW
Boiler Solids
Pass 1 (firetube)
Pass 2
Pass 3
Pass 4
Total
B/SCM1-1^

35
81
38
8
162
ASH (V =

23
31
15
9
78
ASH (V =

28
83
37
11
159
Wt.%

5.31
5.19
4.82
7.10

149 PPM)
Run 48
3.85
5.30
3.59
3.19

39 PPM)
Run 47
0.76
1.38
0.29
0.68


Vanadium
Content
ppm
(on fuel)

24
54
23
7
108


11
21
7
4
43


3
15
1
1
20

Run 57
Total
Part. V
mg/
SCM

61
148
47
11
267


23
35
21
10
89


36
71
39
16
162
Anal.
Wt.%

3.34
2.12
3.17
4.44


Run 69
5.45
5.60
4.66
2.86


Run 78
1.67
1.22
1.02
1.10


V
Con-
tent
ppm

26
40
19
6
91


16
25
12
4
59


8
11
5
2
26

Run 68
Total
Part. V
mg/
SCM

60
69
27
7
163


28
45
24
10
107


36
77
35
13
161
Anal.
Wt.%

4.29
5.15
6.32
7.38


Run 75
3.22
4.48
3.50
3.29


'Run 50
1.02
0.58
0.53
0.69


V
Con-
tent
ppm

33
45
22
7
107


11
26
10
4
51


6
6
2
1
14

Run 52
Total
Part. V
mg/
SCM

79
117
52
19
267


19
31
17
11
78


22
71
39
17
149
Anal.
wt.%

2.40
3.58
2.99
3.33


Run 45
2.07
4.21
2.86
2.67


Run 58
2.55
1.07
0.75
1.22


V
Con-
tent
ppm

24
53
20
8
105


5
17
6
4
32


Run 71
Total
Part. V
mg/
SCM

45
102
35
8
190


20
44
23
9
96

Anal
Wt.%

4.39
4.93
4.67
10.06


Run 63
3.68
3.47
2.94
3.64



V
Con-
tent
ppm

25
64
21
10
120


9
19
8
4
40


Run 79
Total
Part. V
mg/
SCM

78
145
46
13
282


21
35
20
11
87

Anal.
Wt.%

1.86
2.11
4.50
5.42


Run 74
4.43
3.90
3.43
2.91



V
Con-
tent
ppm

19
39
27
9
94


12
18
9
4
43

Run 65(2)
7
10
4
2
23
22
60
35
20
137
1.59-1
1.07-1
0.61-0
0.35-0

.69 5
.00 8
.75 3
.42 1
17















                                                                                                                                                                00
                                                                                                                                                                01
NOTES:
(1)


(2)
All weights corrected to 3% 0~-


Separate sample submitted under blind designation.

-------
                                - 86 -
           Vanadium recovery for the 24 runs averaged 26.0 wt. %
+ 2.04 (95 percent confidence limit) of the fuel V input.  The
particular fuel employed did not affect vanadium recovery.  Combining
both residence times using fuel FI, recovery averaged 25.2 wt. %
versus 26.8 wt. % obtained with fuel F2.  On this basis, the six
runs made with each fuel at the particular residence time were
pooled to provide the respective total averages.  Thus at short
residence time, vanadium recovery averaged 29.3 wt. % + 2.04
(95 percent confidence limit) versus 22.7 wt. % + 2.04  (95 percent
confidence limit) at long residence time.  This difference of
roughly 23% is   statistically significant effect at the 95
percent confidence level.

           The variance associated with the boiler inventory
amounted to 18.97 out of the total experimental variance of 83.31
determined for the entire program.  Therefore, 23% of the total
variance could be explained by the variation in V recovery
determined in the passes.

7.6  Recovery of Vanadium in the Stack - Summation

           In each test, both stack sampling systems were operated
simultaneously to prevent sample bias.  For the same reason, they
were alternated between the two sample ports 90 degrees apart.
The four variables, i.e. sampling system, residence time, fuel and
sample port location, were tested for their effect on vanadium
recovery using the analysis of variance procedure  (ANOVA).  Since
runs made with the light Arab fuel, as previously noted, gave V
recoveries in many cases well over 100 percent, these tests were
eliminated from the ANOVA.

           The results of the analysis of variance indicated there
were only two primary effects:  residence time  (T) which produced
a change in V recovery of 10.5 units or roughly 20% and sampling
systems which produced a change in recovery of about 9.3 units or
18%.

           Since sample port  (the flue gases were not maldistributed
in the stack) and fuel type did not have an effect on vanadium
recovery, tests with these variables were pooled.  A two way analysis
of variance using only residence time and sampling system variables
was run to obtain a better estimate of experimental precision.   This
served as the basis for Table  7.18 which compares average vanadium
recovery in the two sampling  systems as a function of combustion
chamber residence time.  Under conditions of short residence time,

-------
                                                             TABLE  7.17

                                                  VANADIUM RECOVERY  IN THE BOILER

                              ANALYSIS OF VARIANCE
S3JRCE
COLUMNS

R3dS X COLUMNS

ERROR

              TOTAL

3RAND viEAM

ERROR STD DEV
              ss

        0.2660004E 03

        0.1650041E 02

        0.5510421E 01

        0.37936H3E 03

        0.6673795E 03

0.26020B3E  02   r 3(t,.D

            01   -
  ROW MEAMS
  0.2269166E  02
  0.293<»999E  02
  COL MEANS
  0.2519166E  02
  0.268^.999E  02
                      ._ ^
                     ^j. ' I,
                   p
                   /, 7
DF

 1

 I

 1

20

23
                                              */t>
0.2660004E  03   O.U02333E 02

0.16500^1E  02   0.8698887E 00

0.5510421E  01   0.-2905051E 00

0.1896841E  02  •=.
                                                                                                  T, \ ^
           t


-------
                                -  88 -
both sampling trains gave acceptable vanadium recovery.   The Method 5
train, however, had a higher collection efficiency inventorying
about 11-1/2% more vanadium than the ER&E system.  When residence
time was increased, both sampling systems suffered a loss in collection
efficiency but the ER&E system declined the most.  While the EPA
train went from 59.3 wt. % down to 51.2 wt. %, a decrease of nearly
15%, the ER&E system decreased by 25%, going from 52.5 wt. % to 39.5
wt. %.  At this lower level, vanadium recovery in the Method 5 train
was almost 23% greater than in the ER&E system.  It can be concluded
from the preceding that the glass filter used in the Method 5 system
is more efficient than the silicone oil-containing Greenberg-Smith
impingers in collecting fine vanadium-bearing particles.  It is these
submicron particles which apparently pass through the last stage of
the Andersen Cascade Impactor, which has a cut-off of 0.4 ym and then
out the impingers and dry-ice acetone cooled condenser.  The inability
of the impingers to trap the very fine vanadium particulate was also
evident in the EPA system.  In most of the tests no more than a trace
of vanadium <0.5 ppm was isolated from the chilled water impingers.
In only one of the 36 runs did the amount of vanadium exceed 1 ppm
and that resulted from combustion of light Arab fuel oil where about
3 ppm was found.

           Regarding the Method 5 train, since there was an apparent
decrease in V recovery from 59.3 wt. % to 51.2 wt. %, it was expected
that downstream of the fiber glas filter a small amount of vanadium
would be detected.  As noted, however, there was no difference in
the V content of the water impingers at either residence time.  The
silica gel trap backing the impingers, which is employed to dry the
flue gas prior to metering, was also checked for the presence of
vanadium.  In Runs 72, 76 and 80 the silica gel, approximately 200
grams, was extracted with water and analyzed for V by Atomic Absorption.
In all three cases, the vanadium content of the silica gel traps
amounted  to no more than a few ppm out of the more than 20 ppm that
was missing.  The uncertainty in analysis was largely due to the
dilute state of the sample and subsequent high noise to signal ratio.
While this amount of vanadium does not materially alter the total material
balances in the three runs, it does suggest that some fine particles
may be passing through the sampling system unimpeded.  A check of the
ER&E system was similarly attempted by submitting the contents of the
oiler..attached to the vacuum pump for V analysis.  The results were
inconclusive since here too only a few ppm V were found.

           To sum up, if vanadium is passing through the sampling system
as the results  imply,   the particles are  in an extremely fine size
range probably <0.1ym,  and are uncollectable  using  standard  techniques,
i.e.  filter, impingers  and impaction devices.

-------
                                     -  89 -
                                 TABLE  7.18
                 RECOVERY  OF  VANADIUM IN STACK SAMPLING SYSTEMS
                 (Basis:   No  statistical  difference between fuel
                  and  F2 or stack sampling location)
(ft
               EPA  Method  5 Train
             59.3 wt.  %
             vanadium recovery
                (55.3  -  63.3)
                 51.2 wt.  %
                (47.2 -  55.2)
                                           ER&E Train
                                              52.5 wt.
                                          (48.5  - 56.5)
                                           39.5 wt.  %
                                          (35.5 - 43.5)
      * Test
     = 2.02
   \= 22
  /%025,44)':rom Table
   zl.96
     = 4.99
   ^=22
Xl0.025,44)From Table
   = 1.96
NOTE:
Experimental variance = 49.56
Std. error = .M!k=12  = 2.03
95 percent confidence interval
Probability  -4.0  <_ X  <_  +  4.0
                                         44^ =
                                       =  0.95
                                                  x  2"03   X

-------
                                   -  90  -
7.7  The Problem of Sulfate Formation
     in Water Impingers	

     7.7.1  Sulfuric Acid Formation

           The sulfur trioxide (803) content of flue gas from
combustion of a typical high sulfur fuel oil burned in the Cleaver
Brooks is about 20 ppm.    There is sufficient water vapor present
to convert this entire amount to sulfuric acid.  If this occurred
the acid, on condensing, would produce about 80 mg/SCM.  Provided
the stack temperature and gas stream are maintained above the acid
dew point (280°F) condensation will not occur.  Sulfur dioxide can
also be converted to acid.  While this reaction was  thought  to be
slow unless a catalyst was employed, recent data published by Battelle^   '
indicates that a significant part of the S02 in the flue gas will
dissolve in the Method 5 water impingers and may oxidize over a
period of several hours.  The rate of oxidation is proportional to
the amount of dissolved oxygen in the water.  With a high sulfur fuel
oil, each water impinger can dissolve as much as 20 mg S02/100 ml
H20.  Therefore, the oxidation of only a small amount of this will
produce a measurable quantity of acid particulate.  Since the potential
to form significant amounts of 1^504 is so high, it is not surprising
that  the water impingers used in the Method 5 system have been a
source of varying quantities of particulate.

           To overcome this problem in our Company sponsored studies,
we switched from using water in the impingers to a liquid which
could be heated to temperatures >350°F and in which S02 is insoluble.
A number of experiments were run to evaluate the concept of a hot
liquid impinger system as a means of overcoming the sulfate problem.
Several cross comparisons were made with water impingers.

           The initial evaluation was made with a sampling train
consisting  of a heated probe, Andersen Cascade Impactor and high
velocity chilled water impingers.  Under normal base load combustion
conditions in the Cleaver Brook  using a high sulfur fuel oil,
approximately 25-30% of the particulate passed through the system and
was collected in the cold water impingers.  In this case about 45 mg/SCM.
Approximately 90% of this material was water soluble.  Analyses indicated
almost an equal split between S04= tied up with metals such as V, Ni,
Na, Fe and as free sulfuric acid.  The acid was determined by potentio-
metric titration of the H+, the sulfate by gravimetric precipitation
with BaCl2-  On a mass basis sulfuric acid accounted for roughly 15 mg/
SCM of the total impinger particulate.

-------
                                  - 91 -
           With the sampling system consisting of a probe, Andersen
Cascade Impactor and silicone oil in place of water impingers,
particulate passing into the latter accounted for 25 wt. % of the total
inventory, nearly 30 mg/SCM.  Most noteworthy was the absence of sulfuric
acid.  Total sulfate tied up with metals in this case amounted to 16%
of the impinger solids whereas in the former case, with the water
impingers the amount was nearly 25% of the catch.  In another modification
of this experiment, a glass fiber filter was placed after the Andersen
Cascade Impactor and the particulate catch in the silicone oil impingers
system was measured.  A small amount of particulate did pass through
the filter, ^4 wt. % of the total inventory, about 5 mg/SCM.   Analysis
indicated the particulate contained a small amount of sulfate but
no free sulfuric acid.  The weight of filter catch plus impinger
material equaled 35 mg/SCM compared to the previous case of 30 mg/SCM
where particulate not retained by the Andersen Impactor passed directly
into the impinger system.  While the weight of these particulate catches
were nearly the same, the sulfate content was quite dissimilar.  With
the filter present in the system, sulfate comprised almost half  (50%)
of the solids weight compared to less than a quarter (20%) without the
filter.  This difference may be an indication that there was interaction
between particulate retained on the filter and flue gas.

           A final experiment used the Andersen Impactor, fiber glass
filter and water impingers.   Particulate catch from the Andersen
and filter were the same as in the previous case.  However, the  impinger
catch was much higher amounting to 20 mg/SCM compared to 5 mg/SCM with
silicone oil.  Based on a potentiometric titration of the H"1" about
12 mg of this total was free sulfuric acid.  Other type sulfate accounted
for 2 mg.  These data are summarized in Table  7.191

      7.7.2  The Nature of Artificial Particulate

           In the factorial study, a highly variable and unpredictable
quantity of artificial particulate was isolated in each run from the
chilled water impingers of  the Method 5 train.  These solids which
consisted predominantly of  ammonia and sulfate were recovered  in
amounts ranging from 2 to 44 mg/SCM.  Attempts to correlate the  formation
of this material with any of the obvious variables including fuel oil
sulfur content were not successful.  As shown in Table  7.20 when either
the light Arab  (2.8 wt. % S) or intermediate ash all-Ven  (1 %  S) fuels
were burned, a similar quantity of solids, approximately  13 mg/SCM,
was isolated from the impingers.  When the high ash 2.2%  S  fuel oil
was burned, a considerably  lower inventory was obtained averaging 8 ppm.
Combustion chamber  residence time had only a slight effect  on  the
amount of particulate.

-------
    - H2S°4
                                                       TABLE 7.19
                               COMPARISON OF PARTICULATE SULFATE/SULFURIC ACID INVENTORY
Filter, mg/SCM

    - Sulfate
                                            Cascade Impactor
                               	   	Cascade Impactor Plus Filter	
Water Impingers   Silicone Oil Impingers   Water Impingers   Silicone Oil Impingers
                                                 16
                                            16
    - Total Particulate
                                                                                  30
                                                                         30
Impingers, mg/SCM

    - Sulfate
      10

      15
  5

None
 2

12
0.5

None
    - Total Particulate
      45
 30
20

-------
                                     -  93 -
                                   TABLE 7.20

                         ARTIFICIAL PARTICULATE FORMATION
                          IN WATER IMPINGERS OF METHOD 5
                                            Combustion Conditions
                                 Short Residence Time       Long Residence Time
*uel Type                        Impinger Solids. mg/SCM    Impinger Solids. mg/SCM

2.2% Sulfur, High Ash
 (All-Ven(Fl)                             8                           5

1.0% Sulfur, Intermed.
 Ash  All-Ven (F£)                       13                          10
2.8% Sulfur, Low Ash, Lt
  Arab (F3)
                The composition of the solids isolated from the water
     impingers was determined for a number of runs, and is shown in
     Table 7.21.  These solids, which were totally water soluble, contained
     very little ash (at 1450°F), generally less than 5 wt. %.  Vanadium,
     which would be one of the major metallic constituents of the ash was
     present but in an exceedingly low concentration, ranging from 0.5
     wt. % to 1.5 wt. % (expressed as X^O^).  Carbon content of the solids
     was generally under 10.00 wt. %.  This taken together with the low
     ash content tends to rule out the possibility that the material
     represents "real" particulate.

                The major constituents of the solids are ammonia and
     sulfate.  The latter is present in concentrations of up to 70%.  In
     some of the preliminary runs, these artificial particulate were
     analyzed by  X-RAY spectroscopy and found to consist of a mixture
     of ammonium sulfate and ammonium bisulfate exclusively.   In the runs
     where both ammonia and sulfate analysis have been reported,the ratios
     also suggest a mixture of the two compounds.  In several cases, the
     solids were dissolved in water and a standard acid-base titration was
     made to determine the hydrogen ion concentration.  This has been converted

-------
to the equivalent amount of sulfuric acid and is reported as such
in the table.  The amounts generally range from 0.2 to .5 wt. %.  Since
the solids were dried under vacuum at 300°F for up to 8 hours, it
is possible that a significant amount of free acid was removed
prior to the analysis.  Therefore, too much emphasis should not be
placed on these titrations.

           Ammonium sulfate is stable at temperatures up to 400°F.
Since this material was not found in the hot silicone oil impingers
but was only recovered from the water impingers of Method 5, it can
be regarded as an artifact of that system.  This artificial particulate
would not be emitted from a boiler during combustion.  Therefore,
its inclusion in a particulate inventory is not corrected and will
result in an erroneously high stack loading.

           The presence of NH  in the flue gas in these runs is
puzzling and has not been resolved.

7.8  Determination of Oxidation State
     of Vanadium Particulate in Flue Gas

           A short series of experiments were run to determine the
change in oxidation state of vanadium exposed to flue gas.  Milligram
samples of three vanadium oxides:  V203  (III), ^2°^  (V) were placed
separately on a fiber glass filter, heated to 375°F and each exposed
for 2 hours to flue gases produced by firing the high sulfur all-Ven
heavy fuel oil with 10% excess air.  Only the V205 gave any indication
of reaction, changing color from mustard yellow to green.  After
exposure each of the samples was transferred under a nitrogen blanket
into a flask and dissolved in sulfuric acid.  A colorimetric determination
run on the sulfuric acid solutions of the oxides indicated that
approximately one third of the pentavalent vanadium  (V) was reduced
to tetravalent V (IV).  While these experiments are by no means
definitive, they do"~indicate that some of the vanadium in the particulate
may be emitted in V 4 state.  In a previous Company sponsored study,
color changes were noted for particulate which had been collected
on a hot fiber glass filter used to back up the Andersen Cascade
Impactor.  In that work the filter was normally taken apart in the
laboratory and the particulate were, therefore, exposed to air.  The
fines which were usually green in color at 375°F became black or grey
on cooling to ambient temperature.  Subsequent reheating would not
produce the reverse color change.
                                                              +4
           Since the normal vanadyl  type  compounds  formed by V
are not volatile at stack  temperatures,  the presence of V™ would not
explain the  loss of vanadium  from  the sampling systems.

-------
                                                    TABLE  7.21
Run No.

  52
  53
  54
  55
  56
  57
  67
  68
  71
  48
  49
  62
  63
  69
  46
  47
  50
  59
  60
            Fuel
           High Ash
            it   ii
            it   n
            tt   n
       All-Ven
        it   n
        it   n
            it
            it
            n
            n
 n
 n
n
n
            Interned. Ash-All-Ven
n
it
n
n
it
ii
it
                tt
                it
            Low Ash-Lt .  Arab
  n
  n
.on of Solids from
Solids
Isolated, mg/SCM
14
9
6
8
4
8
4
10
6
4
12
4
5
11
20
6
37
11
23
Method 5 Water

V C
0.31
-
0.66
None 5 .
0.56
0.26
0.25
0.14
0.10
None 12 .
0.40 10.
Trace
0.20
0.83
None 2 .
1.3 4.
None 1.
Trace
0.38
Impingers

H N
_ —
-
- -
01 2.29 8.81
-
-
- -
- -
0.47*
29 3.65 8.65
98 3.43 8.29
0.22*
0.26*
_ _
83 5.28 16.24
51 3.60 8.35
21 4.18 11.86
0.3*
0.2'-


NH^
_
-
-
-
-
-
-
1.73
Trace
—
9.08
Trace
-
9.3
18.82
2.87
13.96
-
Trace
                                                                                                             69.76
                                                                                                             23.0

                                                                                                             51.31
                                                                                                             26.10
                                                                                                             40.02
                                                                                                             50.25
                                                                                                             32.01

                                                                                                             46.17
                                                                                                             60.62
*Reported as H SO,

-------
                                 - 96 -


                              8.   REFERENCES
 1.  Natusch, D.F., Wallace, J.R. and Evans, C.A.,  "Toxic Trace
     Elements: Preferential Concentration in Respirable Particles",
     Science 183:202 (1974).

 2.  Lee, R.E.,Jr.and Von Lehmden, D.J., "Trace Metal Pollution in the
     Environment", J. Air Poll. Control Assoc., 23:853 (1973).

 3.  Goldstein, H.L. and Siegmund, C.W., "Factors Affecting Particulate
     Emissions from Combustion of Residual Fuel Oil", First Chemical
     Congress of the North American Continent, Symposium Combustion
     Generated Air Pollution, Paper 40, Mexico City, 1975.

 4.  "EPA Standards of Performance for New  Sources", Federal
     Register, Vol. 36, No. 247, Part II, Dec. 23, 1971, pp. 24876-
     24895.

 5.  Brief,  R., Sand Morrow, N.Y., "Elemental Composition of Suspended
     Particulate in Metropolitan New York",  Environmental Science and
     Technology, Vol. 5 p.  786 (1971).

 6.  Sugimae, A. and Hasegawa, T., "Vanadium Concentrations in  Atmosphere",
     Environmental Science and Technology, Vol. 7,  No.  5, May 1973.

 7.  Kneip,  T.J. et al "Airborne Particulates in New York City", J.  Air
     Pollution Control Association, 2:144 (1970).

 8.  Athanassiades, Y.C., "Air Pollution Aspects of Vanadium and Compounds",
     Report  PB188 093, Contract No. PH-22-68-25, September 1969.

 9.  Schroeder, H.A., Air Quality Monograph //70-13, "Vanadium", American
     Petroleum Institute, October 1970.

10.  Lee, R.E., Jr., Soranson, S.S.,  Enrione, R.E., and Morgan, G.B.
     "National Air Surveillance Cascade Impactor Network II. Size
     Distribution of Trace Metal Components", Environmental and
     Technology 6:1025 (1972).


11.  U. S. Department of Health, Education,  Air Quality Data from
     the National Air Surveillance Networks and Contributing State
     and Local Networks. 1966 Edition.  Publication APTD 68-9.
     Durham, N.C.

12.  Marters, C.S. et al, "Sources of Vanadium in Puerto Rican  and
     San Francisco Aerosols",  Env. Science and Technology  7:817 (1973).

13.  Hickey, J.R. et al "Relationship Between Air Pollution and Certain
     Chronic Disease Rates  - Multivariant Statistical Studies",
     Arch. Environmental Health 15 (6) 728 (1967).

14.  Siegmund, C.W., "Low Sulfur Fuels Are Different",  Hydrocarbon
     Processing, February 1970.

-------
                                  - 97 -
REFERENCES (Cont'd.)


15.  Biefe, H. et al Volatility of Metal Porphyrin Complexes in Petroleum,
     J. Chem. and Eng. Data 5 (4) October 1960.

16.  Wolsky, A.A. and Chapman, G.W., "A Study of an Intrinsic Contaminant
     in Petroleum - The Nature of Vanadium in Asphaltenes", API Division
     of Refinery Mtg. Detroit, May 1960.

17.  Beach, L.K. and Shewmaker, J.E., "The Nature of Vanadium in Petroleum",
     Ind. Eng. Chem. 49 (7) 1157  (1957).

18.  Sulzer, P.T. et al, "The Influence of Some Chemical and Physical
     Factors on the Formation of Deposits from Residual Fuel Oil",
     Trans. A.S.M.E. 17:995, 1955.

19.  MacFarlane, J.J., "Combustion Conditions and Errosion of Metal
     Surfaces by Fuel Ash Deposition" - The Mechanism of Corrosion by
     Fuel Impurities, Butterworth, London 1963.

29.  Ranz, W.E. and Wong, J.B.,"Impaction of Dust and Smoke Particles",
     I & E Chem., Vol. 44, No. 6, p.1371-1380, June 1952.

21.  Achinger, W.C. and Shigehara, R.T., "A Guide for Selected Sampling
     Methods for Different Source Conditions", J. Air Pollution
     Control Assoc. 18, No. 9, September 1968, p. 605-609.

22.  Foster, W.R., Leipold, M.H., and Shevlin, F.S., "A Simple Phase
     Equilibrium Approach to the Problem of Oil-Ash Corrosion",
     Corrosion, 12, 23-31 (1956).

23.  Milan, E.F., "The Dissociation Pressure of Vanadium Pentoxide",
     J. Nuclear and Inorg. Chem. 33, 498-508, (1929).

24.  Hillenbrand,  L.  J.  et  al,  "Final Report  on  Chemical  Composition
     of Particulate Air  Pollutants from  Fossil-Fuel  Combustion  Sources",
     Battelle Columbus Laboratories, March  1,  1973.

25.  Crow, E. L., Davis, F. A., and Maxfield, M. W., "Statistics Manual",
     p. 103-104, New York, Dover Publications, Inc. 1960.

-------
                       - 98 -
                      APPENDIX 1






                 STATE OF THE ART REVIEW




VANADIUM ANALYSIS AND PARTICIPATE SAMPLING METHODOLOGY

-------
                                - 99 -
  BIBLIOGRAPHY

 1.   Yoshida,  T..,  Kousaka,  Y.  and  Okuyama,  K.,  A New Technique  of
     Particle  Size Analysis of Aerosols  and Fine Powders  Using  An
     Ultra-Microscope,  Ind. Eng. Chem. Fundam.  Vol.  14, No.  1,(1975).
 2.   Duce,  R.A.,  Hoffman,  G.L. (Univ.  R.I.),  Zoller, W.H. (Univ. Md.)
     Atmospheric  Trace  Metals  at Remote  Northern and Southern
     Hemisphere Sites:   Pollution  or Natural?,  Science  187,  #4171:59-61,
     (1/10/75).
 3.   Nader,  J.S.,  Current  Technology for Continuous  Monitoring  of
     Particulate  Emissions, U.S. Environmental  Protection Agency,
     Journal of the Air Pollution  Control Association,  Vol.  25, No.  8,
     p.  814-821 (August 1975).
 4.   Benson, A.L., Levins,  P.L., Massucco,  A.A., Sparrow, D.B.,
     Valentine, J.R.,  Development  of A High Purity Filter for  High
     Temperature  Particulate Sampling and Analysis,  J.  of the  Air
     Pollution Control  Assoc., Vol.  25,  No. 3,  p. 274-277, (March  1975).
 4.   Begnoche,  B.C., Risby, T.H.  (Pa.  State Univ.),  Determination  of
     Metals in Atmospheric Particulates  Using Low-Volume  Sampling  and
     Flameless  Atomic  Absorption  Spectrometry, Anal. Chem., Vol.  47,
     #7:1041-45 (June 1975).
 6.   Jackson,  M.L. (Univ.  Idaho),  Patterson,R.G. (Gonzago Univ.),
     Fine Particle Collection in  (Single and) Multihole,  Sonic-And
     Supersonic-Flow Impingers, 74th AIChE Natl. Meeting  (New Orleans
     3/11-15/73)  AIChE Symp. Ser.  71,  //147:47-51 (1975).
 7.   Soffian,  G,  Westlin,  P.R., Particulate Emissions from Apartment
     House Boilers and  Incinerators, U.S. Environmental Protection Agency,
     J.  of the Air Pollution Control Assoc., Vol. 25, No. 3 p.  269-273.
 8.   Smith, W.B., Gushing, K.M.,  Lacey,  G.E., McCain, J.D.,  Particulate
     Sizing Techniques  for Control Device Evaluation, Southern Research
     Institute, Birmingham,Ala.,  Environmental  Protection Technology
     Series, EPA-650/2-74-102-a (August  1975).
 9.   uibbs, H.P., Marier,  P.,   (Dept. Environ.,0nt.)  Some  Environmental
     Effects,  Sources,  and Control Methods for  Fine Particulates,  4th
     AIChE-Can. Soc. Chem. Eng. Jt.  Chem. Eng.  Conf. (Vancouver 9/9-12/73)
     AIChE Symp.  Ser. 71,  #147:60-69 (1975).
10.   McCain, J.D., Gooch,  J.P., Smith, W.B. (Southern Res. Inst.), The
     Results of Field Measurements of Industrial Particulate Sources and
     Electrostatic Precipitator Performance, Fine Particle Electrostatic
     Precipitation Symp. (Pensacola Beach,  Fla. 9/30-10/2/74)  J. Air
     Pollut. Control Assoc. 25, #2:117-21  (Feb. 1975).
11.   Tullar, I.V., (Jacobs Eng. Co., Cherry Hill,N.J.), Suffet, I.H.
     (Drexel Univ., Phila, Pa.) The Fate of Vanadium in an Urban Air
     Shed:  The Lower Delaware River Valley, J. of the  Air Pollut. Control
     Assoc. Vol.  25, No. 3, p. 282-286,  (March  1975).
12.   McCain, J.D., Gooch,  J.P., Smith, W.B. (Southern Res. Inst.)  The
     Results of Field Measurements of Industrial Particulate Sources and
     Electrostatic Precipitator Performance, Fine Particle Electrostatic
     Precipitation Symp. (Pensacola Beach,  Fla. 9/30-10/2/74)  J, Air Pollut,
     Control Assoc. 25, #2:117-21  (Feb.  1975).

-------
                               - 100 -
13.  Mancini, C. (Univ. Roma), Vanadium Determination in Fuel Oil by
     Neutron Activation Analysis with a Californium-252 Source, Riv.
     Combust. 29, #3:91-93, (March 1975).
14.  Bakke, E. (U.S. Filter Corp., MikroPul Div.) Wet Electrostatic
     Precipitators for Control of Submicron Particles, Fine Particle
     Electrostatic Precipitation Symp. (Pensacola Beach, Fla. 9/30-10/2/74)
     J. Air Pollut. Control Assoc., 25, #2:163-67 (February 1975).
15.  Sugimae, A. (Environ.,Pollut. Control Cent., Osaka Prefecture)
     (A Method For) The Emission Spectrographic Determination of Trace
     Elements in Airborne Particulate Matter, Anal.  Chem. 46, #8:1123-25,
     (July 1974).
15.  Lee, R. E., Hein, J. (Environ. Prot. Agency), A Method for the
     Determination of Carbon,  Hydrogen, and Nitrogen in Size-Fractionated
     Atmospheric Particulate Matter, Anal. Chem. 46, #7:931-33 (June 1974).
16.  Chang, H.C. (Bechtel Corp.), A (New) Parallel Multicyclone Size-
     Selective Particulate Sampling Train, Am. Inc.  Hyg. Assoc., J. 35
     #9:538-45 (September 1974).
17.  Cooper, D.W., Spielman, L.A. (Harvard Univ.), A New Particle Size
     Classifier:  Variable-Slit Impactor With Photo-Counting,
     Atmos. Environ, 8, #2:221-32, (March 1974).
18.  Griever, P.M. (Res. Applicance Co.) A Review of Atmospheric Sampling
     Methods and Devices, Instrum. Soc. Am. Power Div. Meet. (Boston,
     May 1974); (Abstr.) Instru., Technol. 21 #8:7,10 (August 1974).
19.  Monkman, J.L., Halman, R. et al (Environ. Health Center), Accurate
     Determination of Environmental Pollutants (Including Lead and
     Vanadium:  A Discussion), Abstr. #ANAL-45, 167th ACS Natl. Meet.
     (Los Angeles, 3/31-4/5/74) Abstr. Paper.
20.  Dingle, A.N., Joshi, B.M. (Univ. Mich.) Ammonium Sulfate Crystallization
     in Andersen Cascade Impactor Samples, Atmos. Environ. 8, #11:1119-30
     (November 1974).
21.  Ranweiler, L.E., Moyers,  J.L., Atomic Absorption Procedure for
     Analysis of Metals in Atmospheric Particulate Matter, Environ.
     Science & Technology, Vol. 8, No. 2, p. 152-156,(February 1974).
22.  Cunningham, P.T., Johnson, S.A. (Argonne Natl.  Lab.) Characterization
     and Analysis of Airborne Particulate Material By Infrared Spectroscopy,
     167th ACS Natl. Meet.  (Los Angeles, 3/31-4/5/74) ACS Div. Environ.
     Chem. Prepr. 14, #1:42-43 (1974).
23.  Giauque, R.D., Goda, L.Y., Brown, N.E.  (Univ. Calif. Berkeley)
     Environ. Sci. Technol. 8, #5:436-41,  (May 1974).
24.  Dennis, R., Collection Efficiency As A Function of Particle
     Size, Shape and Density:   Theory and Experience, J. of Air Pollut.
     Control Assoc., Vol. 24,  No. 12, p. 1156-1163,(December 1974).
25.  Shannon, L.J. (Midwest Res. Inst.), Control Technology For Fine
     Particulate Emissions, U.S. Dep. Commer. Natl.  Tech. Inf. Serv.
     PB #236,646, p. 224 (May 1974).
26.  Seeley, J.L., Skogerboe,  R.K. (Colo. State Univ.)  (In) A Combined
     Sampling-Analysis Method for the Determination of Trace Elements
     in Atmospheric Particulates, Anal. Chem. 46, #3:415-21  (March 1974).
27.  Gladney, E.S., Zoller, W.H., Jones, A.G., Gordon, G.E., Composition
     and Size Distributions of Atmospheric Particulate Matter in the
     Boston Area, Environmental Science & Technology, p. 551, (June 1974).

-------
                                 - 101 -


28.  von Lehmden, D.H., Jungers, R.H., Lee, R.E. (Natl. Environ. Res.
     Center, Res. Triangle Park) Determination of Trace Elements in Coal,
     Fly Ash, Fuel Oil, Gasoline...A Preliminary Comparison of Selected
     Analytical Techniques, 165th ACS Natl. Meet. (Dallas April 1973)
     Anal. Chem. 46, #2:239-45 (February 1974).
29.  Flinn, J.E., Reimers, R.S. (Batelle Mem. Inst., Columbus Lab.),
     Development of Predictions of Future Pollution Problems, U.S.,
     Environ. Prot. Agency, Socioecon. Environ. Stud. Ser. Rep. #600/5-74-005
     p.213 (March 1974).
30.  Myers, R.L., (Univ. Pitts. Diss.) Electrical Detection of Airborne
     Particulates Using Surface-Ionization Techniques, (Abstr), Diss.
     Abstr. Int. B 34, //9:4560B, p. 137 (March 1974).
31.  Anderson, D. (Environ. Prot. Agency) Emission Factors for Trace
     Substances, U.S. Dept. Commer., Natl. Tech. Inf. Serv. PB//230.894
     p. 79, (December 1973).
32.  Heindryckx, R., Dams, R. (Inst. Nucl. Sci., Rijksuniv. Gent.,
     Ghent, Belg.), Evaluation of Three Procedures for Neutron Activation
     Analysis of Elements in Atmospheric Aerosols Using Short-Lived
     Isotopes, Radiochem. Radioanal. Lett. 16(4), p.209-25, (Eng.) (1.974).
33.  Dworetsky, S.H. (Bell Lab.) The Feasibility of An Automatic
     Isokinetic Stack (Emission) Sampler, Environ. Sci. Technol. 8, #5:464,
     (May 1974).
34.  Nixon, D.E., Fassel, V.A., Kniseley, R.N. (At. Energy Comm.;
     Iowa State Univ.) Inductively Coupled Plasma-Optical Emission
     Analytical Spectroscopy..(Sample Introduction Via) Tantalum Filament
     Vaporization of Microliter Samples, Anal. Chem. 46,  //2:210-13
     (February 1974).
35.  DeBonno, G., (U.S. Army Mobility Equip. Res. & Dev.  Cent.) Investigating
     Waste Oil Disposal By Combustion, U.S. Dep. Commer., Natl. Tech.
     Inf. Serv. AD //772, 991 p. 13 (January 1974).
36.  Patterson, R.K., Wagman, J. (Environ. Prot. Agency), Mass and
     Composition of An Urban Aerosol As A Function of Size for Several
     Visibility Levels, 167th ACS Natl. Meet. (Los Angeles, 3/31-4/5/74)
     ACS, Div. Environ. Chem. Prepr. 14, //l:102-7 (1974).
37.  McCain, J.D., Gushing, K.M.,Smith, W.B., Methods for Determining
     Particulate Mass and Properties Laboratory and Field Measurements,
     J. of Air Pollution Control Assoc., Vol. 24, No. 12, (December 1974).
38.  Wolcott, O.K. (La. State Univ. & Agric. & Mech. Coll., Diss.)
     Methods for the Direct Determination of Heavy Metal  Pollutants in
     the Environment, (1973), p.168; (Abstr.) Diss. Abstr. Int. B 34,
     //9:4257B (March 1974).
39.  Taylor, C.E., Taylor, W.J. (Southeast Environ. Res.  Lab.) Multielement
     Analysis of Environmental Samples by Spark-Source Mass Spectrometry,
     U.S. Environ. Prot. Agency. .Environ. Prot. Tech. Serv. Rep. //660/2-74-001,
     p. 23 (January 1974).
40.  New Lab Scrutinizes Air Pollution, Can. Chem. Process, 58 //3:30,
     p.32-33 (March 1974).
41.  Robertson, D.C., Carpenter, R., Neutron Activation Techniques for
     the Measurement of Trace Metals in Environmental Samples,
     National Academy of Sciences, National Res. Council, Washington, D.C.
     *USAG (4400000), p. 84, NAS-NS-3114,  (January 1974).

-------
                                 - 102 -
42.  Gordon, G.E., Zoller, W.H. (Univ.  Md.) Normalization and Interpretation
     of Atmospheric Trace-Element Concentration Patterns, 1st Natl.
     Sci. Found. Trace Contaminants Annual Conf. (Oak Ridge Natl. Lab.
     8/8-10/73) U.S. Dept. Commer. Natl. Tech. Inf. Serv. CONF #730802:314-25
     (March 1974).
43.  Ruch, R.R., Gluskoter, H.J., Shimp, N.F. (111. State Geol. Surv.)
     Occurrence and Distribution of Potentially Volatile Trace Elements
     in Coal, U.S. Environ. Prot. Agency, Environ.  Prot. Technol. Ser.
     Rep. #650/2-74-054 p. 105 (July 1974).
44.  Detrie, J.P., Particulate Emission from Combustion in Furnaces,
     Rev. Ass. Fr. Techns. Petrole, (224), p.99-107 (March/April 1974).
45.  Particulate Emissions from the Combustion of Heavy Fuel Oil in
     Large Furnaces, Cardinal (Electr.  Fr.) Rev. Assoc. Fr. Tech. Pet.
     #224:104-7 (March/April 1974).
46.  Detrie, J.P., Particulate Emissions Resulting from  (Industrial)
     Combustion, Rev. Assoc. Fr. Tech.  Pet. //224:99-103  (March/April 1974).
47.  Coughlin, R.W. (Lehigh Univ.) Physiochemical Aspects of Sampling
     Particulates, AIChE Natl. Meet.(1972), AIChE Symp. Ser. 70
     #137:273-76  (1974)
48.  Scale, L.M.  (GCA Corp.), Proceedings  (of a) Symposium on the Use of
     Fabric Filters for the Control of  Submicron Particulates (April 8-10,
     1974, Boston, Mass.), U.S. Environ. Prot. Agency, Environ. Prot.
     Technol. Ser. Rep. #650/2-74-043 p.316,  (May 1974).
49.  Coughlin, R.W., Siegel, R.D., Charanjit Rai,  (Editors)
     Recent Advances in Air Pollution,  AIChE Symp. Series 137, Vol. 70,
     (1974).
50.  Wagman, J. (Environ. Prot. Agency), Recent Developments in
     Techniques for Monitoring Airborne Particulate Emissions from
     Sources, AIChE Natl. Meet. (1972), AIChE Symp. Ser. 70 #137:277-84
     (1974).
51.  Winefordner, J.D. (Univ. Fla.), Vickers, T.J. (Fla. State Univ.)
     Review of Fundamental Developments in Analysis...Flame Spectrometry,
     Anal. Chem. 46, #5:192R-227R  (April 1974).
52.  Modowski, E.R. (Babcock & Wilcox)   Stack Particulate Sampling,
     Mechanical Engineering,  (October 1974).
53.  Hettsche, H.O., Ten Years of Trace Metal Analysis and Its
     Evaluation, "Heavy Metals as Air Pollutants" Symp.  (Duesseldorf
     2/22-23/73) Stauf-Reinhalt. Luft 34, #1:17-18 (January 1974).
54.  Hamil, H.F., Bamann, D.E., Thomas, R.E.  (Southwest Res. Inst.)
     The Collaborative Study of EPA Methods 5,6, and  7 in Fossil Fuel-
     Fired Steam Generators...Final Report, U.S. Dept. Commer. Natl.
     Tech. Inf. Serv. PB #237,695 p.39, (May 1974).
55.  Hanson, H.A.  (Univ. Minn., Diss.), The Deliquescent Properties of
     Aerosols, p. 241, (Abstr.) Diss. Abstr.  Int. B 35 #2:847B
     (August 1974).
56.  Natusch, D.F.S., Wallace, J.R. (School of Illinois, Urbana),
     Evans, C.A., Jr.  (Materials Res. Lab., Univ. of  111.)  Toxic Trace
     Elements: Preferential  Concentration  in Respirable  Particles,
     SCIENCE, Vol. 183 (January 18, 1974).
57.  Gordon, W. A.  (N.A.S.A., Lewis Res. Cent.) Quantitation and
     Detection of Vanadium in Biologic and Pollution Materials, N.A.S.A.
     Tech. Memo #X-71497, p.33,  (January 1974).

-------
                               - 103 -
58.  Novakov, T. (Univ. Calif..Berkeley),  Sulfates in Pollution
     Particulates,  67th Air Pollut. Control Assoc. Annu. Meet.
     (Denver 6/9-13/74) Paper #74-200, p.  29.
59.  Little, P., Martin, M.H. (Univ. Bristol), Tables, Map, Contour
     Maps, and 21 References, Environ. Pollut. 6, #1:1-19 (January 1974).
60.  Lochmueller, C.H., Galbraith, J.W., Walter (Duke Univ.), Trace
     Metal Analysis in Water By Proton-Induced X-Ray Emission Analysis
     (Pixea) of Ion-Exchange Membranes, Anal. Chem. 46, #3:440-42
     (March 1974).
61.  Natusch, D.F.S., Wallace,  J.R., Urban Aerosol Toxicity:  The
     Influence of Particle Size, SCIENCE,  Vol. 186, #4165, p. 695-699
     (November 22,  1974).
62.  O'Neal, A.J. (Long Island Lighting Co.) Vanadium Recovery from
     the Combustion of Venezuelan Fuel Oil, Combustion 46, #5:18-21
     (November 1974).
63.  Preis, H. (Eidg. Materialpruefungs.- & Versuchsans Bauwesen &
     Gewerbe), Method for Quick Colorimetric Determination (As Little
     As 1 PPM) Of Vanadium In Petroleum Products Without Their Ashing,
     Erdoel Kohle,  Erdgas, Petrochem, Brennst.-Chem. 26 #9:503-5
     (Sept. 1973).
64.  Goldberg, A.J. (Environmental Protection Agency, Washington, D.C.
     Office of Research and Monitoring), A Survey of Emissions and Con-
     trols for Hazardous and Other Pollutants, PB-223 568/7, p. 185
     (February 1973).
65.  Cooper, D.W. and Spielman, L.A.  (Div. of Engineering & Applied Physics,
     Harvard Univ.,Camb.,Mass.), A New Particle Size Classifier: Variable-
     Slit Impactor with Photo-Counting, Atmospheric Environment, Vol. 8
     pp. 221-232 (1973).
66.  Gladney, E.S., Zoller, W.H. and Gordon, G.E.  (Univ. Md.), Size
     Fractionation of Trace Elements  (Including Vanadium from Oil
     Combustion) On Urban Aerosoles, Abstr. No. NUCL-51, 18-30827.
67.  Cropper, W.P.  (Standard Oil Co. - Indiana), The Concentration of
     Solid Particles Suspended in a Gas Stream is Determined,
     U.S. 3,459,947, i. 8/5/69, f.7/17/68, CL250/218 (C.I.P. of abandoned
     Ser. 468,750,  f. 7/1/65).
68.  Shinko Chemical Co. Ltd.,  Treating Waste Waters Containing Vanadium -
     By Redn. of Vanadium and Copper with Iron, JA-4850559-Q.
69.  Environmental Protection Agency  (Research Triangle Park, N.C. - Office
     of Air and Water Programs), Air Quality Data for Metals 1968 and 1969
     from the National Air Surveillance Networks, PB-224 823/5, p. 182
     (June 1973).
70.  Everett, G.L., West, T.S., and Williams, R.W.  (Dept. Chem., Imp. Coll.
     Sci. Technol., London, Eng.), Atomic Absorption and Fluorescence
     Spectrometry With a Carbon Filament Atom Reservoir. XIV. Determination
     of Vanadium in Fuel Oils, Anal.  Chim. Acta. 66  (2), 301-3 (Eng.)(1973).
71.  Michels, Savine (Med. Fak., Rheinisch-Westfael. Tech. Hochsch.,
     Aaxhen, Ger.), Biological Effects of Vanadium with Respect to
     Emission Limits, Staub-Reinhalt, Luft,  33(7),  276-9 (Ger.)(1973).
72.  Hillenbrand, L.J., Engdahl, R.B., and Barret,  R.E.  (Battelle Mem.
     Inst.), Chemical Composition of  Particulate Air Pollutants from
     Fossil-Fuel Combustion Sources,  U.S. Dept. Commer., Natl.Tech.Inf.
     Serv. PB #219,009, 174 p., 21-30869  (3/1/73).

-------
                                -  104  -
73.   Gallorini,  M.,  Geneva,  N.,  Meloni, S., Maxia, V. (1st Chim. Gen.
     Inorg.,  Univ.  Pavia,  Pavia, Italy), Determination of Lead, Manganese,
     Aluminum, and Vanadium in Atmospheric Dust Using Flame-Emission
     Atomic Absorption, Inquinamento 15(9), 13-18 (Ital.) (1973).
74.   Vogg, H., Hartel, R.  (Lab.  Isoteopentech., Kernforschungszentrum,
     Karlsruhe,  Ger.), Determination of Trace Elements in Airborne Dust
     Samples by Neutron Activation Analysis, Fresenium Z. Anal Chem. (1973).
75.   Bonner,  N.A.,  Bazan,  F.,  Camp, D.C. (California Univ.,  Livermore,
     Lawrence Livermore Lab),  Elemental Analysis of Air Filter Samples
     Using X-Ray Fluorescence, UCRL-51388, p. 50 (June 1, 1973).
76.   Anderson, D. (Environmental Protection Agency, Research Triangle
     Park, N.C., Office of Air Quality Planning and Standards), Emission
     Factors for Trace Substances, EPA-450/2-73-001, PB-230 894/8, p. 80
     (Dec. 1973).
77.   Imai, S., Ito,  K., Hamaguchi, A., Kusaka,  Y., Warashina, M. (Public
     Health Res. Inst., Kobe,  Japan), Emission Spectrographic Determination
     of Trace Elements in Airborne Particulates by Using Membrane Filter,
     Bunseki Kagaku 22(5), 551-8 (Japan)(1973).
78.   Tullar,  I.V.(Day & Zimmerman, Inc.) and Suffet, I.H. (Drexel Univ.),
     The Fate of Vanadium in An Urban Air Shed, The Lower Delaware River
     Valley,66th Air Pollut. Control Assoc. Annu. Meet. (Chicago), Pap. //
     73-117:23p. (6/24-28/73).
79.   Calvert, S., Goldshmid, J.  et al  (A.P.T. Inc.), Feasibility of Flux
     Force/Condensation Scrubbing for Fine Particle Collection, Dep. Commer.,
     Natl. Tech. Inf. Serv.  PB //227.307, 149 p. (Oct. 1973).
80.   May, W.R.,  Zetlmeisl, M.J.  and Bsharah, L. (Petrolite Corp.), High-
     Temperature Corrosion in Gas Turbines and Steam Boilers by Fuel
     Impurities —2.  The Sodium Sulfate/Magnesium Sulfate/Vanadium Pentoxide
     System,  Ind. Eng. Chem.,  Prod. Res. Dev. 12, #2:140-45  (June 1973).
81.   Scanlon, J. (Clin. Pediatr. - Phila), Human Fetal Hazards from
     Environmental Pollution with Certain Non-Essential Trace Elements,
     11,135-41 (Ref. 41) TOXBIB/72/6555327 (March 1973).
82.   Robbins, W.K.,  Micro-Metal Analysis Using Heated Vaporization Atomic
     Absorption (HVAA)(Abstract), pres. at International Symposium on
     Microchemical Techniques, 1973, Penn State Univ., Univ. Park, Pa.
     (August 19-24,  1973).
83.   Rancitelli, L.A., Cooper, J.A., Perkins, R.W.  (Battelle Pacific
     Northwest Labs., Richland,  Wash.)(1158200), Multielement Characterization
     of Atmospheric Aerosols by Instrumental Neutron Activation Analysis
     and X-Ray Fluorescence Analysis, BNWL-SA-4671, p. 23 (1973).
84.   Brandone, A.,  Maxia,  V. and Orvini, E.  (Cons. Na. Ric.; Univ. Pavia),
     Nuetron Activation Analysis of Al, As, Br, Mn, Sb, Si and V in
     Airborne Particulate Matter, llth Soc. Chim. Ital. Natl. Congr.
     (Perugi - 10/19-21/72)  Chim. Ind. (Milan) 55,  //ll: 873-76  (Nov. 1973).
85.   Shelby,  W.D.,  Persian! (GTE Lab Inc.), Nondestructive Instrumental
     Monitoring of Fuel Oil for Vanadium, Sodium and Sulfur, Environ. Sci.
     Technol V7 N.2 125-27,  Corrosion  and Deterioration Measurement Methods
     (Feb. 1973).
86.   Woolrich, P.F., Occurrence of Trace Metals in  Environment: An Overview,
     Am. Ind. Hyg.  Assoc.  J. 34, 217-26, TOXBIB/72/4626402 (May 1973).

-------
                                - 105 -
 87.   Bird,  A.  Jr.,  McCain,  J.D.,  Harris,  D.B.,  Particulate Sizing Techniques
      For Control Device Evaluation,  (March 15,  1973).
 88.   Morgen,  E.A.,  Rossinskaya,  E.S.,  Vlasov,  N.A.,  Zhes'ko,  V.P.
      (Irkutsk, Cos. Univ.,  Irkutsk,  USSR), Photometric Determination of
      Vanadium (V)  Using the Reagent  Magneson IREA,  Zavod.  Lab.  39(4),
      409-10 (Russ.)(1973).
 89.   Mitrache, F.,  Physical and  Chemical Aspects of the Toxicology of
      Vanadium, Farmacia (Bucharest), Vol.  21,  Issue 6, P325-30 (1973).
 90.   Kneip, T.J.,  Kleinman, M.T.  and Eisenbud,  M.  (N.Y. Univ.  Med. Cent.),
      Relative Contribution  of Emission Sources to the Total Airborne
      Particulates  in New York City,  3rd Int. Union  Air Pollut.  Prev. Assoc.-
      VDI-Komm. Reinhaltung  Luft  Int. Clean Air Congr.  (Dusseldorf 10/8-12/73)
      Proc:  C96-C97  (1973).
 91.   Kerby, R.C. (Can.  Dep. Energy,  Mines Resources) and Wilson, J.R. (Queen's
      Univ., Kingston, Ont.), Solid-Liquid Phase Equilibria for the Ternary
      Systems V.O -Na_0-Fe00», V00,.-Na00-Cr 0,.,  and  V-O.-Na-O-MgO, Can.J.Chem.
      51 //7:1032-40 (5/1/73).              2 5       252
 92.   Zoller,  W.H.,  Gordon,  G.E.,  Gladney,  E.S., Jones, A.G. (Dept. Chem.-
      Univ.  Maryland, College Park, Md.),  Sources and Distribution of
      Vanadium in the Atmosphere,  Advan. Chem.  Serv., 123,  31-47 (Eng)(1973).
 93.   Martens,  C.S., Wesolowski,  J.J.,  Kaifer,  R.,  John, W., Harriss, R.C.
      (Lawrence Livermore Lab., Livermore,  Calif.),  Sources of Vanadium in
      Puerto Rican and San Francisco  Bay Area Aerosols, Environ. Sci.
      Technol., 7(9), 817-20 (Eng.)(1973).
 94.   Greifer,  B. and Taylor, J.K. (Natl.  Bur.  Stand.), Survey of Various
      Approaches to the Chemical Analysis of Environmentally Important
      Materials, U.S. Dep. Commer., Natl.  Tech.  Inf.  Serv.  COM //74-10460,
      226 p. (July 1973).
 95.   Spittler, T.M., Thompson, R.J., Scott, D.R. (Environ. Prot. Agency,
      Res. Triangle Park, N.C.),  System for Collection and Measurement of
      Elemental and Total Mercury in Ambient Air Over a Concentration Range,
      Am. Chem. Soc. Div.-Water,  Air Waste Chem. Gen. Pap.  v 13 n 1,
      p 88-89,  Meeting held  April 8-13, 1973.
 96.   Loon,  J.C.V.,  Toronto's Precipitation Analyzed for Heavy Metal
      Content,  Water Pollut. Control, Ont.  Ill #2:38-41 (1973);  (Abstr.)
      Water Pollut.  Abstr. (U.K.)  46  //4:No. 681 (April 1973).
 97.   Lee, R.E. and Von Lehmden,  D.J. (Environ. Prot. Agency), Trace Metal
      Pollution in the Environment, J.  Air Pollut.  Control Assoc. 23,
      #10:853-57 (Oct. 1973).
 98.   Of Vanadium Bonding in Petroleum, (Univ.  South. Calif.) 166th ACS
      Natl.  Meeting 8/26-31/73, ACS Div. Pet. Chem.  Prepr.  18, #4:648-51
      (1973).
 99.   Sugimae,  A.,  Kasegawa, T. (Environ.  Pollut. Control Cent., Osaka,
      Japan), Vanadium Concentrations in Atmosphere, Environ.  Sci. Technol.
      7(5),  444-8 (Eng.)(1973).
100.   Rolla, A.  (Stn. Sper.  Combust.),  Frigeri, P.  (Cent. Inf. Stud.
      Esperienze, Segrate) et al,  X-Ray Fluorescence Analysis of Metals in
      the Atmosphere of Milan, 23rd Chem. Days Conf. (Milan 4/19-20/72)
      Chim.  Ind. (Milan) 55, #8:623-28  (August 1973).

-------
                                 - 106 -
101.  Gilfrich, J.V., Burkhalter, P.G. and Birks, L.S. (Nav. Res. Lab.),
      X-Ray Spectrometry For Particulate Air Pollution...A Quantitative
      Comparison of Techniques, Anal. Chem. 45 #12:2002-9 (Oct. 1973).
102.  Miller, M.S., Friedlander, S.K. and Hidy, G.M.  (Calif. Inst. Tech.),
      A Chemical Element Balance For The Pasadena Aerosol, 161st ACS Natl.
      Meet. (Los Angeles 3/28-4/2/71) J. Colloid Interface Sci. 39
      #1:165-76 (April 1972).
103.  Whitby, K.T., Husar, R.S. and Liu, B.Y.H. (Univ. Minn.), The Aerosol
      Size Distribution of Los Angeles Smog, 161st ACS Natl. Meet. (Los
      Angeles 3/28-4/2/71) J. Colloid Interface Sci.  39 #1:177-204
      (April 1972).
104.  Nuttonson, M.Y. (Am. Inst. of Crop Ecology, Silver Spring, Md.)
      (023 700), AICE Survey of USSR Air Pollution Literature, Volume XVII.
      A Compilation of Technical Reports (4th) on the Biological Effects
      and the Public Health Aspects of Atmospheric Pollutants, AICE-AIR-72-17,
      p. 153 (1972).
105.  Air Pollution Instrumentation:  A Trend Toward Physical Methods,
      Science 177 #4050:685-87 (8/25/72).
106.  Hirling,  J., Pavlik, 0., Czeller, G. (Izot. Intez., Magy. Tud. Akad.,
      Budapest, Hung.), Air Purification by the Separation of Radioactive
      Molecular Impurities and Dispersed Solid Particles, Izotoptechnikai
      Kut., 95-113  (Hung.). Magy.Tud.Akad.Izotop.Intez.: Budapest, Hungary
      (1972).
107.  Oddo, N.  (Perkin-Elmer Ital. S.p.A.), Application of Atomic Absorption
      Spectrophotometry to the Analysis of Polluting Elements, 23rd Chem.
      Days Conf. (Milan 4/19-20/72) Chim. Ind. (Milan) 54 //ll: 1009-11
      (Nov. 1972).
108.  Martens,  C.S., Atmospheric Chemistry of Coastal Aerosols, Cloud Droplets
      and Rain, Florida State Univ. Diss (1972), 172 p. (Abstract) Diss.
      Abstr. Int. B V33 N.5 2237B-2238B (November 1972).
109.  Eisenbud, M. & Ehrlich, L.R. (N.Y. Univ. Med. Cent.), Carbon Monoxide
      Concentration Trends in Urban Atmospheres, Science 176 #4031:193-94
      (4/14/72).
110.  Sulkowski, W., Izycki, J., Adamiak-Ziemba, J. and Stroszejn, G.,
      Changes in the Upper Respiratory Trace in Workers Exposed to Vanadium,
      Otolaryngol Pol. 26, 165-72 (1972).
111.  Rolla, A. (Stn. Sper. Combust.), Determination By X-Ray Fluorescence of
      Elements  in Dust Collected on Filters from the Milan Atmosphere,
      Riv. Combust. 26 #4:156-63 (April 1972).
112.  Rolla, A. (Std. Sper. Combus., San Donate Milanese, Italy), Determination
      of Elements in the Particles Collected in Filters from Air in Milan by
      X-Ray Fluorescence, Riv. Combust. 26(4), 156-63 (Ital.)(1972).
113.  Gallorini, M., Geneva, N. et al (Univ. Pavia), Determination of Lead,
      Manganese, Aluminum, and Vanadium in Atmospheric Particulates by Atomic
      Absorption and Flame Emission (Spectroscopy), llth Soc. Chim. Ital.
      Natl. Meet. (Perugia, 10/19-21); (Ital. Abstr.) Riv. Combust. 26
      #11:392-98 (November 1972).
114.  Watanbe,  H., Berman, S., Russel, D.S. (Div. of Chem., National Research
      Council of Canada, Ottawa, Ontario, Canada), Determination of Trace
      Metals in Water Using X-Ray Fluorescence Spectrometry, Talanta 1972,
      Vol. 19,  pp. 1363 to 1375, Pergamon Press, Printed in No. Ireland (1972).

-------
                                -  107 -
115.   Duke, C.R.,  Cho, B.Y. (Ind. Nucleonics Corp., Columbus, Ohio),
      Development of a Nucleonic Particulate Emission Gage, U.S. Nat. Tech.
      Inform.  Serv. PB Rep. 1972, No. 209954, 64 pp (Eng).   Avail. NTIS from
      Gov. Rep. Announce. (U.S.)(14) 181 (1972).
116.   Burkitt, P.L., Nickless, G. (Sch. Chem.,  Univ. Bristol, Bristol, Engl.),
      Distribution of Heavy Metals in the Vicinity of an Industrial Complex,
      Nature (London), 238(5363), 327-8 (1972).
117.   Kometani, T.Y. (Bell Lab.), Bove, J.L. (Cooper Union Adv. Sci. & Art),
      Nathanson, B. (N.Y. Dept. Air Resources)  et al, Dry Ashing of Airborne
      Particulate Matter (Including Lead) on Paper and Glass Fiber Filters
      for Trace Metal Analysis by Atomic Absorption Spectrometry, Environ.
      Sci. Technol. 6 #7:617-20 (July 1972).
118.   McHugh,  J.A. and Stevens, J.F. (Gen. Elec.Co.), Elemental Analysis of
      Single Micrometer-Size Airborne Particulates by Ion Microprobe Mass
      Spectrometry, Anal. Chem. 44 #13:2187-92 (November 1972).
119.   Hasagawa, T. and Sugimae, A., Emission Spectrographic Determination
      of Trace Metals in Airborne Particulates, Bunseki Kagaku 20:840-45
      (1971):   (Abstr.) Air Pollut. Abstr. (U.K.); No. A18527  (July 1972).
120.   Tumanovskii, A.G., Zel'kind, M.E. (Vses.  Teplotekh, Inst. Moscow -
      USSR), Formation of Corrosive Sulfur and Vanadium Compounds from
      These Elements Contained in Fuel Oils During Burning in Combustion
      Chambers of Gas-Turbine Apparatus, Neftepererab, Neftekim, (Moscow)
      (3), 9-13 (Russ.)(1972).
121.   Moyers,  J.L., Zoller, W.H., Duce, R.A. , Hoffman, G.L.  (Dept. Chem.,
      Univ. Hawaii, Honolulu, Hawaii), Gaseous Bromine and Particulate Lead,
      Vanadium, and Bromine in a Polluted Atmosphere, Environ.Sci.Technol.
      6(1), 68-71 (1972).
122.   Louria,  D.B., Joselow, M.M., Browder, A.A., Human Toxicity of Certain
      Trace Elements, Ann. Intern. Med. Vol. 76, p. 307-319  (Ref. 154)
      (Iss. February 1972).
123.   Fish, R.A.  (Sci. Branch, Greater London Counc., London, England),
      Metals in the Atmosphere, Proc. Annu. Conf. Nat. Soc. Clean Air,
      39(Pt. 1), 11/1-11/19 Eng.) (1972).
124.   Gordon,  G.E. (Univ. Md.), Hopke, P.K. (State Univ. Coll. Fredonia,
      N.Y.) et al, Neutron Activation Study of Trace Elements on Boston-Area
      Atmospheric Particles (Including V from Residual Oil Combustion, Na and
      Cl from Marine Aerosols, and Pb and Br (Plus Some Cl) from Automobile
      Exhaust), 163rd ACS Natl. Meeting, Abstr. //NUCL-4, Boston, Mass.
      (4/9-14/72).
125.   Gusev, M.I., Elfimova, E.V., Ryzhkovskii, V.L., Markina, N.A.  (USSR),
      Nonferrous Metallurgy As a Source of Atmospheric Pollution, Aktual. Vop.
      Gig. Usloviyakh Krain, Sev. 1972, 76-80 (Russ).  From Ref. Zk. Met. 1972,
      Abstr. No.  12G28.
126.   Fenimore, C.P. (Gen. Elec. Corp. Res. Div.), Oxidation of Sulfur Dioxide
      in Stainless Steel or Inconel Sampling Probes, Combust. Flame  18
      //3:469-70 (June 1972).
127.   Nifong,  G.D., Boettner, E.A. and Winchester, J.W.  (Univ. Mich.),
      Particle Size Distributions of Trace Elements in Pollution Aerosols,
      Am. Ind. Hyg. Assoc., J. JJ3 //9:569-75 (September 1972).
128.   Webb, R., Webb, M., Rapid Emission Spectrographic Method for the Analysis
      of Air Filters, Atomic Energy Research Establishment, Harwell  (England),
      AERE-R-6966, p. 7  (January 1972).

-------
                                 - 108 -


 129.   Heisler,  S.L.,  Friedlander,  S.K.,  Husar,  R.B.,  The Relationship of Smog
       Aerosol Size and Chemical Element Distributions to Source Characteristics,
       W.M.  Keck Laboratory of Environmental  Health Engineering; Calif.  Inst.
       of Tech., Pasadena,  Calif.  (July 1972).
 130.   Morrow, N.L.,  Brief, R.S.,  Bertrand,  R.R.(Exxon Res. & Eng.)Sampling and
       Analyzing Air Pollution  Sources,  Chemical Engineering, p. 85-98
       (January 24, 1972).
 131.   Frigieri, P.,  Trucco, R.  et al. (Centre Inf. Stud. & Experienze,  Milan),
       Spectroscopic Analysis of Elements Present in Airborne Dust,  16th
       Spectrochem Seminar  (Sirmione 6/9-12/71)  Chim.  Ind. (Milan) 54 +1:12-17
       (January 1972).
 132.   Atkins, D.H.F., Fisher,  E.M.R., Salmo, L., Studies of the Analysis of
       Airborne Particulate Material by Radioactivation and Sodium Iodide
       Spectrometry,  Atomic Energy Research Establishment, Harwell (England),
       AERE-R-6724, p. 25  (October 1972).
 133.   Hwang, J.Y. (Instrum. Lab., Inc.), Trace Metals in Atmospheric
       Particulates and Atomic Absorption (AA) Spectroscopy), Anal.  Chem. 44,
       #14:20A-24A, 26A-27A (December 1972).
 134.   U.K.  Monitors Trace Elements in Air, Chem. Eng. News 50  #12:40 (3/20/72).
 135.   Lininger, R.L., Vertical and Temporal Variations of Trace Elements in
       Surface Air, Univ. Mich. Diss. (1972) 222 p.;  (Abstr.) Diss.  Abstr. Int.
       B, 33, //5:2265B (November 1972).
 136.  Thatcher, L.L., Johnson,  J.O.,  Welsh,  C.N.,  A Comprehensive
       Program in Neutron Activation Analysis in Water Quality,
       (edited by) International At.;  p.323-328  (1971).
 137.  Mayer, W.A. (AEG-Telefunken), Activation  Analysis  of Vanadium
       in Oil, Erdoel Kohle, Erdgas, Petrochem.  24  //6:416-17
       (June 1971).
 138.   Dittrich, T.R., Cothern,  Richard C.,(Univ. of Dayton), J. of  the
       Air Pollution  Control Assoc., Vol.  21, No. 11,  p.  716 (November 1971).
       Harrison, P.R.  et  al (Univ.  Mich.),  Brar,  S.S., Nelson, D.M.
       (Argonne  Nat.  Lab.), Areawide Trace  Metal  Concentrations  Measured
       By Multielement Neutron  Activation Analysis, J. Air Pollut. Control
       Assoc. 21 //9:563-70  (September  1971).
 139.   Guicherit,  R.  (Inst. Gezondheidstech.  TNO, Neth.)  Automatic
       Measurement of  Air Pollutants,  Chem.  Weekbl. 67(9), p.18-23 (1971).
       Colucci,  J.M.,  Begeman,  C.R. (Ge.  Mot. Res.  Labs.) Carcinogenic
       Air Pollutants  in  Relation  to Automotive  Traffic In New York,
       Environ.  Sci.  Technol. 5,  //2:145-50  (February 1971).
 140.   Dickson,  F.E.,  Kunesh, C.J.  et  al.  (Gulf  Res. & Devel. Co.)
       The Characterizations of  Vanadium IV  Environments  in Petroleum
       by Electron Spin Resonance  (ESR),  161st ACS  Nat. Meet.
       (Los  Angeles 3/28-4/2/71),  ACS  Div.  Petrol.  Chem.  Prepr. , 16
       //1:A37-A46  (February 1971).
 141.   Abramov,  V.N.,  Ofitserov,  S.I., Khmyrov,  V.I.,  (USSR),
       Combustion  of Mazut  with  Vanadium-Containing Ashes, Probl.
       Teploenerg,  Prikl. Teplofiz  1971,  No.  7,  9-19 (Russ.)
       From  Ref.  Zh.,  Teploenerg,  1972,  Abstr. No.  2T60.
142.   Zhigalovskaya,  T.N., Egorov, V.V., Malakhov, S.G.,  Shilina, A.I.,
       Krasnopevtsev,  Yu. V.  (USSR), Content  of Heavy  Metals in  the  Air
       In Some Regions of the USSR, Meteol.  Aspekty Zagryazneniya  Atmos.
       1971  320-9  (Russ.) From Ref. Zk.,  Khim 1972  Abstr.  No.  41423.
143.   Persiani,  C. (Bayside Res.  Cent.,  GTE  Lab, Inc., Bayside,N.Y.)
       Detection of Selected Air Pollutants  Using a Laborator Neutron
       Generator,  Amer. Ind.  Hyg.  Assoc., J.  1971,  32(9),  573-8  (Eng..;

-------
                                  - 109 -
144.    Grennfelt,  P.  (Swedish Inst.  for Silicate Res.  Gothenburg),
        Akerstron,  A.K.,  Brosset,  C.,  Determination of  Filter-Collected
        Airborne Matter by X-Ray Fluorescence, Atmos.  Environ. Vol. 5,
        No.  1, p.1-6,  (January 1971).
X45     Shinkarenko, L.S., Bavika, L.I., Determination  of Sulfuric Acid
        in Air By the Vanadate Method, All U Nion Sci Res Inst Syn Fats.
        Neftepererab Teftekhim (Mosc)  N.9 40-41 (1971)  (In Russian Measurement
        Methods.
146.    Pillay, K.K. Sivasankara;  Thomas, Charles Carlisle, Jr., (West
        New York Nucl. Res.Cent.,  Inc., Buffalo,N .Y.) Determination of
        the Trace Element Levels in Atmospheric Pollutants by Neutron
        Activation Analysis, J. Radioanal. Chem. 7(1),  p.107-18 (1971)  (Eng.)
147.    Omang, S.H. (Central Inst. Ind. Res. Oslo) Determination of Vanadium
        and Nickel in Mineral Oil By Flameless Graphite Tube Atomization,
        Anal. Chem. Acta 56 #3:470-73  (October 1971).
148     Pilz, W., Komischke, S., Prior, G., The Determination of Vanadium
        in Aqueous Solutions, Air and  Biological Material-1.  The  Determination
        of Vanadium in Aqueous Solutions, Int. J. Environ. Anal. Chem.  1(1),
        1971, p. 47-61,  (Reed 1972).

 149.   Bodri, Zh.  (USSR), Effect  of  Fuel Composition on  Pollution of
        the  Atmos.  Harmful Products  of Combustion,  Zagryazneniya Vozdukha
        Vykhlopn. Gazami  Tr.ansp.  1971, 21-8 (Russ.)  From  Ref.  No.  201426.
 150.   Morrow,  N.L.,  Brief, R.S.  (Exxon Res.  & Eng. Co.) Elemental
        Composition of Suspended Particulate Matter in  Metropolitan
        New York, Environ. Sci.  Technol. 5,  //9:786-89 (September 1971).
 151.   Robinson, E.,  Robbins,  R.C.,  Emissions, Concentrations,  and
        Fate of Particulate Atmospheric Pollutants,  prepared  for API,
        Washington,  SRI Project No.  SCC-8507,  API Pub.  No.  4076,
        (March 31,  1971).
 152.   Hasagawa, T.,  Sugimae,  A.  (Environ.  Pollut.  Control Cent.,  Osaka,
        Japan),  Emission  Spectrographic Determination of  Trace Metals
        in Airborne Particulates,  Bunseki Kagaku 1971,  20(7),  840-5 (Japan).
 153.   Hasagawa, T.,  Sugimae,  A.  (Japan) Heavy Metals  in Atmospheric
        Particulates Size Distributions, Kuki  Seijo 1971, 9(1),  1-9.
 154.   Sem,  G.J.,  Borgos, J.A., Olin, J.F., (Thermo Systems,  Inc.  Minn.)
        Monitoring  Particulate  Emissions, Chem. Engineering Progress
        Vol.  67,  No.  10,  p.83-89,  (October 1971).
 155.   Davis, W.E.  &  Associates,  National Inventory of Sources  and
        Emissions:  Vanadium...1968, U.S. Dep.  Comm.  Natl. Tech.  Inf.  Serv.  PB
        #221,  655 p.  60 (June  1971).
 156.   Pillay,  K.K.  Sivasankara;  Thomas, Charles  Carlisle,  Jr., Hyche, C.M.
        (Western New York'Nucl.  Res. Cent.,  Inc.,  Buffalo,  N.Y.),
        Neutron Activation Analysis of Inorganic Constituents  of Airborne
        Particulates,  Nucl.  Technol.  1971 10(2),  224-31 (Eng.)
 157..   Colucci,  J.M.,  Begeman,  C.R.  (Gen.  Mot. Res. Lab.,  Warren,  Mich.)
        Polynuclear  Aromatic Hydrocarbon and Other  Pollutants  in Los
        Angeles  Air,  Proc.  Int.  Clean  Air Congr.,  2nd 1970  (Pub. 1971)
        28-35  (Eng.)
 158.   Bando, S.,  Yamane,  Y.,  Murakami, Y.  (Japan  At.  Energy  Res.  Inst.
        Tokyo, Japan)  Radioactivation  Analysis for  the  Determination  of
        Trace  Elements  in Air  Pollutants,  Nucl. Tech. Environ. Pollut.
        Proc.  Symp.  1970  (publ.  1971)  169-81 (Eng.)  IAEA: Vienna,Austria.
 159.   DeBerry,  D.W.  (Tracer,  Inc.),  Sladek,  K.H.  (Univ.  Texas)
        Rates  of  Reaction of S02 With  Metal  Oxides,  Can.  J.  Chem.  Eng.  49
        #6:781-85 (December 1971).

-------
                                -  110 -


160.   "State  of  the  Art:  1971  - Instrumentation  for  Measurement  of
       Particulate  Emissions  from Combustion  Sources" ,  Vol.  I  and  II,
       Report  of  Contract  CPA 70-23 by Thermo-Systems,  Inc.  for EPA.
       Documents  PB 202-665 and PB  202-666, Natl.  Tech.  Inf.  Serv.
       Springfield, Va.
161..   Zoller, W.H.,  Gordon,  G.E. et al  (Univ.  Md.)  The Sources and
       Distribution of Vanadium in  the Atmosphere, 162nd ACS Nat. Meet.
       (Washington  9/12-17/71)  ACS  Div.  Water,  Air Waste Chem.  Prep.  11
       //2:159-65  (1971).
162.   Gordon, G.E., Zoller,  W.H.,  Gladney,  E.S., Jones, A.G. (Dep.  Chem.
       Univ.  Md,  College Park,  Md.)  Trace Elements in the Urban, Nucl.
       Methods Environ, Res.  Proc.  Amer. Nucl.  Soc.  Top. Meet.  1971 30-7
       (Eng.).
163.   Laamanen,  A., Lofgren, A.,  Partanen,  T.J.   (Inst. Occup.  Health,
       Helsinki,  Finland), Trace Metals  in Particulates in the  Air of
       Helsinki and Turku, (1964-69)  Work,  Environ, Health 1971, 8(2),
       63-9 (Eng.).
164..   Pilz, W.,  Komischke, S., Prior, G., (Physiol-Chem Anal.  Lab.,
       Farbenfabriken Bayer A.-G.,  Leverkusen-Bayerwerk, Ger.)  Vanadium
       Determination  in Aqueous Solutions, Air, and Biological  Material  I.
       Vanadium Determination in Aqueous  Solutions.    Int. J. Environ.
       Anal. Chem.  1971, 1(1),  47-61, (Ger.)
165.   Vanadium May Be the Next Metal to  Join the  Environmental Blacklist,
       Chem. Eng. (New York)  78 //13:46 (6/14/71).
166.   Gilfrich,  J.V.,  Burkhalter,  P.G.,  Birks, L.S.  (U.S. Naval  Res. Lab.)
       X-Ray Fluorescence  Analyses  (of Vanadium, Molybdenum,  Lead, Etc.)
       in Particulate Air  Pollutants,  16th Int.  Colloq. Spectrosc.  (Heidelberg
       10/4-9/71; (Ger. Abstr.) Chem. Ing. Tech.  44 #11:734-38  (June  1972).
167.   Byers,  R.L.,  Davis, J.W., Matson,  W.L., McKinstry,  H.A.  (  Pa.  State
       Univ.)   A  New  Technique  for  Characterizing  Atmospheric Aerosols
       for  Size and Shape, CAES Publication No. 163070  (Center  for
       Air  Environment Studies, University Park,  Pa.) APCA 70-65.
168.   Bando,  Shoji (Japan),  Activation  Analysis  of  Airborne Particulate
       Matter  in  the  Chiba Area,  Genshiryoku Kogyo 1970,  16(9),  68-72.
169.   Kneip,  T.J.,  Eisenbud, M., Strehlow, C.D.,  Freudenthal,  P.C.   (Med.
       Center,  New  York Univ.)  Airborne  Particulates  in New  York City,
       J. Air  Pollut.  Control Assoc., 1970 20(3).
170.   Airborne Solids' Makeup  Subjects  of New  Study  Automobile Manufacturers
       Association,   California Air Resources Board;  Los Angeles  Air  Pollution
       Control DI;  Western Oil  & Gas Assoc., Chem. Eng.  V77,  N.17 62
       (8/10/70), Motor Fuels;  Air  Pollution.
171.   Prinz,  B.  (Landesantalt  fuer Immissions  -  und  Bodennutzungsschutz  des
       Landes  NW, Essen, W. Germany)   Application of  Statistical Methods
       for  Taking Random Samples to Investigate Large-Space  Immissions,
       Staub - Reinhaltung Luft, Vol. 30, No. 5,  p.204-10  (May  1970).
172.   Rollier, Mario A.,  Brandone,  Alberto, Geneva,  Nicla (1st Chim. Gen.,
       Univ. Pavia,  Pavia, Italy),   Atmospheric  Pollution Control
       Neutron-Activation  Analysis  of Elements  in  Solid  Atmospheric Dust.
       Chim. Ind. (Milan)  1970, 52(10),  977-94.
173.   Miles,  J.W.,  Fetzer, L.E., Pearce, G.W., (Tech.  Devel. Lab.,
       U.S. Dept. of  Health,  Education &  Welfare,  Ga.)   Collection and
       Determination  of Trace Quantities  of Pesticides  in  Air,
       Environmental  Science  &  Technology, Vol. 4, No.  5,  p.420-425,
       (May 1970).

-------
                                  - Ill -
 174.   Shah,  K.R.,  Filby,  R.H.,  Haller,  W.A.  (Washington State Univ.,
       Nuclear Radiation Center, Pullman,  Wash.)  Determination of Trace
       Elements in  Petroleum by  Neutron Activation Analysis,  J. Radioanal.
       Chem.  6,p.413-422 (1970).
 175.   Noll,  K.E.  (Univ. of Washington,  Seattle),  Pilate, M.J.
       Inertial Impaction of Particles Upon Rectangular Bodies,  J. Colloid
       Interface Science,  Vol.  33, No. 2,  p.  197-207, (June 1970).
 176.   Zoller, W.H., Gordon, G.E. (Mass. Inst.  Techn.)  Instrumental
       Neutron Activation Analysis of Atmospheric  Pollutants Utilizing
       Ge(Li) X-ray Detectors,   Anal. Chem. 42, #2:257-65 (February 1970).
 177.   Brandon, J.H. (Betz, Trevose,  Pa.)   Measurement of Particulate
       Emissions,  Ind.  Coal Conf. Purdue Univ.  Lafayette, Ind., p. 16
       (October 7-8, 1970).
178.   Dams,  R., Robbins,  J.A.,  Rahn, K.A., Winchester,  J.W. (Great
       Lakes  Res.  Div., Univ. of Mich.,)  Nondestructive Neutron Activation
       Analysis of  Air  Pollution Particulates,   Anal. Chera.  1970, 42(8),
       861-7.
 179.   Boldrino, F., Principles  of Atomic  Absorption Spectrometry,
       (Sta.  Sper.  Combust.)  Riv. Combust. 24//6:272-83 (June 1970).
 180.   Jackson, M.R. (IIT Res.  Inst., Chicago,  111.) Lieberman, A.,
       Townsend, L.B.,  Romanek,  W.,    Prototype Fly Ash Monitor for
       Municipal Incinerator Stacks,   ASME, Proc.  of Natl. Incinerator
       Conf., Cinn. Ohio,  p .1.82-8 (May 17-20, 1970).
 181.   Saltzman, B.E.,   Significance  of Sampling Time,  J. Air Pollut.
       Control Assoc.,  Vol. 20,  No.  10,  p.660-5, (October 1970).
 182.   Ziesse, N.G., Penney, G.W.,   Some  Effects  of Particle Size on
       Measurement  of Fine Airborne Particulates,   Ashrae Trans.,
       Vol.  76, Pt. 1,  Paper 2127, p.26-37 (1970).
 183.   Brown, R.,  Vossen,  P.G.T. (AEI Sci. App. Ltd., Manchester)
       Spark  Source Mass Spectrometric Survey Analysis of Air Pollution
       Particulates,  16th ASTM  Mass  Spectrometry  Conf., (Pittsburgh May 1968)
       Anal.  Chem.  42,  #14:1820-22,  (December 1970).
 184.   Brar,  Sarmukh S., Nelson, D.M., Kanabrocki, E.L., Moore, C. E.,
       Burnham, C.D., Hattori,  D.M.  (Div.  of  Biol. & Radiol. Physl, Argonne
       Nat. Lab.,  111.)  Thermal Neutron Activation Analysis of Particulate
       Matter in Surface Air of  the Chicago Metropolitan Area One Minute
       Irradiations, Environ. Sc. Technol., 4(1),  p.14-17 (1970).
 185.   Schroeder,  H.A.  (Dartmouth Med. Sch.)   Vanadium,   API Air Qual.
       Monogr. #70-13 p. 32 (October  1970).
 186.   Mamuro, T.,  Fujita, A., Matsuda,  Y. (Radia. Center Osaka Pref.  Sakai, Japan)
       Activation Analysis of Airborne Dust,   Annual Rep. Radia. Center
       Osaka  Prefect, 10,  9-20 (1969).
 187.   Spurny, K.R., Lodge, J.P. Jr., et al  (Nat.  Center Atmospheric Res.)
       Aerosol Filtration By Means of Nuclepore Filter...Structural
       and Filtration Properties,  Environ. Sci. Technol. 3, #5:453-64
       (May 1969).
 188.   Kneip, T.J., Eisenbud, M. et al  (New York Univ. Med. Center)
       Airborne Particulates in New York City,   62nd Air Pollut. Control
       Assoc., Annual Meeting (New York 6/22-26/69) Paper #69-166:37p.
 189.   Athanassiadis, Y.C., Air  Pollution Aspects  of Vanadium and Its
       Compounds,  Litton Systems, Inc.  Bethesda,  Md. Environmental
       Systems Div., PB-188 093, p. 105 (September 1969).

-------
                                - 112 -
 190.   Thompson, R.J., Morgan, G.B., Purdue, L.J. (Div. Air Qual. Emiss. Data,
       Natl. Air Pollut. Control Admin., Cincinnati, Ohio),   Analysis
       of Selected Elements in Atmospheric Particulate Matter By
       Atomic Absorption,  Anal. Instrum. 7, 9-17 (1969).
 191.   Das, H.A., Application of Activation Analysis to the Problem
       of Air Pollution,  Chem. Weekbl. 65(30) 7-8 (1969).
 192.   Mammarella, L., Biondi, A., Atmospheric Pollution in Rome (Multistage
       Monitoring of Particle Size and Concentration of Aerosols By A
       System of Several Apparatuses),(1st. Ig. "G.  Sanarelli", Univ. Roma,
       Rome, Italy) Nuovi Ann. Ig. Microbiol. 20(3)  208-37 (1969).
193.   Flesch, J.P., (Natl. Center for Air Pollut. Control)  Calibration
       Studies of a New Sub-Micron Aerosol Size Classifier,  153rd
       ACS Nat. Mtg. (Miami Beach 4/9-14/67) J. Colloid Interface Sci.,
       29 #3:502-9  (March 1969).
194.   Brandon, J.H.,   Can A Fuel Treatment Program  Control Stack
       Emissions,  Combustion Vol. 41, No. 4, p20-24, Air Pollution;
       Fuel Oils (October 1969).
195.   Lees, B., (Brit. Petrol. Co. Ltd.)  Chemistry of Deposits in Oil -
       Fired Boilers,   J. Inst. Fuel 42 #344:356-57  (September 1969).
196.   Liu, B.T.G., Marple, V.A., Yazdani, H.,  Comparative Size
       Measurements of Monodisperse Liquid Aerosols  by Electrical
       and Optical Methods,  Univ. Environ. Sci. Technol. 3,  #4:381-86,
       (April 1969).
197.   Control Techniques for Particulate Air Pollutants, U.S. Nat.
       Air Pollut. Control Admin. Publ.  //AP-51, p.551 (January 1969).
198.   Bando, S., Tsuyoshi (Japan At. Energy Res. Inst. Oarai, Japan)
       Determination of Vanadium in Deposit and Airborne By Neutron
       Activation Analysis, Kagaku 18(12), 1447-82 (1969).
199.   Householder, M.K. (Youngstown State Univ., Ohio) Goldschmidt, V.W.,
       Impaction of Spherical Particles on Cylindrical Collectors,
       J. Colloid Interface Sci. Vol. 31, No. 4 p. 464-78 (December 1969).
200.   Battigelll (Univ. N.C. Sch. Med.)  Particulates:  Air Quality
       Criteria Based  on Health Effects, API, Air Qual. Monogr. //69-2
       p. 15 (February 1969).
201.   Byers, R.L. (Pa. State Univ.), Particulate Collection Equipment
       and Characteristics — 2.  West Scrubbers and Electrostatic
       Precipitators,   Lehigh Valley Air Conser. Council Joint
       "Eng. Aspects Air Pollut. Contr." Conf. (Bethlehem 5/22/69)
       Proc. p. 29.
202.   Perman, J., Perman, E.,  Spectrographic Determination of Metal
       Traces in the Atmosphere, Bol, Geol. Minero 1969, 80(5), 476-84.
203.   Himi, Y. Mori,  0., Studies on Sampling Points for Measurement
       of Dust Concentration in Horizontal Straight Flue, J.  Chem.
       Soc. Japan, Vol. 72, No. 11, p.2347-51  (November 1969).
204.   Halstead, W.D., (Chem. Div., Cent. Elec. Res. Lab., Leatherhead, Eng.)
       Thermodynamics  of Fuel Oil Ash Constituents in Combustion Systems,
       J. Inst. Fuel,  42 (346), p.419-24 (1969).
205.   Yen, T.F., Boutcher, L.J. et al. (Carnegie-Mellon Univ.) Vanadium
       Complexes and Porphyrins in Asphaltenes, J. Inst. Petrol. 55,
       #542:87-99 (March 1969).

-------
                                   - 113 -
 206.    Hoffman,  G.L.,  Duce,  R.A.  (Univ. Hawaii),  Zoller, W.H.  (Mass.  Inst.
        Technol.)  Vanadium,  Copper,  and Aluminum  in  the Lower  Atmosphere
        Between California  and  Hawaii,  Environ. Sci.  Technol.  3,  //11:1207-10,
        (November 1969).
 207.    Air  Quality  Data  from the  National  Air  Surveillance  Networks  and
        Contributing Seate  and  Local  Networks 1966 Edition
        National  Air Pollution  Control  Administration, Durham,  N.C.,
        APTD-68-9, p.167  (1968).
208.    Rama, S., Paggi, A. (Univ. Siena)  An Apparatus for Collecting
        Smoke and Other Particulate Samples on Microscope Slides in
        (Subsequent) Particle-Size Determinations,  Riv. Combust.  23,
        //6:325-28  (June 1968).
209.    Keane, J.R., Fisher,  E.M.R. (Health Phys. Med. Div. at  Energy Res.
        Establ., Harwell, Eng.)  Analysis of Trace Elements in  Airborne
        Particulates By Neutron Activation and X-Ray Spectrometry,
        Atmos. Environ. 2(6), p.603-14  (1968).
210.    Andreatch, A.J., Innes, W.B., Characterization of Petroleum Waste
        Products By  Selective Combustion Thermal Effects, 155th ACS
        Natl. Mtg. (San Francisco  3/31-4/5/68) ACS Div. Petrol  Chem. Inc.
        Preprints Vol. 13, No.  2, Motor Fuels; Air Pollution
211.    Wiele, H., Kasten, G. , Combustion and Determination of Compounds
        Containing Inorganic  Sulfur,  Inorganic Fluorine, and Organic
        Phosphorus,  Arsenic,  Vanadium and Lead by  the Wickbold  Method
        Under Application of  a Metal  Burner, Fresenius'Z. Anal. Chem.
        235(4), p.335-41  (1968).
212.    Osamu, T., (Div. Hyg.,  Inst.  Sci. Labor, Tokyo, Japan),
        Determination of Toxic  Substances in Air Inorganic Compounds,
        Rep. Inst. Sci. Labour, No. 69, p.10-22 (1968).
213.    Volodarskii, I. Kb..,  Grekhov, I.T., Panin, V.I., Shmuk, E.I.,
        Shpirt, M. Ya.  (USSR), Distribution of Some Rare and Nonferrous
        Metals During Combustion of Coals, Khim. Tverd. Topi (2), p.103-6
        (1968).
214.    Lundgren, D.A.  (Univ. Calif., Statewide Air Pollut. Res. Center)
        The Effects  of Various Fluoride Sources Citrus Growth and Fruit
        Production,  Univ. Calif.  Statewide Air Pollut. Res. Center
        Univ. Calif  Riverside Extension "Air Pollut. Calif.:  Past,
        Present & Future Conf.  "Air Pollut: Causes & Effects Symp.
        (Riverside,  Calif. 9/26/68) p.15-23.
215.    Shiraiwa, T., Matsuno, F., Identification  of Vanadium-Attack
        Products,  Sumitomo Kinzoku 20  //2 (1968) (Abstr) Ccrros. Abstr.
        7, //5:328, (September 1968).
216.    Jerman, L.,  Jettmar,  V. , Polarographic Determination of Vanadium
        in Air of Work Areas  (Hyg-Stat. Bezirksnat Ausschusses, Prague, Czech),
        Z. Gesamte Hyg. Ihre  Grenzgeb 1968 14(1),  12-14.
217.    Bergmann, J.G., Ehrhardt,  C.H.  et al. (Am. Oil Co.)  Trace Analysis
        of Metals By Ion Exchange  Concentration and X-Ray Fluorescence  (XRF)
        153rd ACS Natl. Mtg.  (Miami Beach 4/9-14/67) Abstr., Papers:
        Abstr. No. B14.
218.    Bergmann, J.G., Ehrhardt,  C.H.  et al (Am.  Oil Co.)  Determination
        of Sub-PPM Nickel and Vanadium  in Petroleum by Ion Exchange
        Concentration and X-Ray Fluorescence, 153rd ACS Mtg. (Miami Beach
        April 1967), Anal. Chem. 39 //11:1258-61 (September 1967).
219.    Laamanen, A.  (Tyoterveyslaitos, Helsinki), Air Pollution in Helsinki,
        Kern. Teollisuus, 23(7), p.619-25 (1966).
220.    Guest, R.J., Ingles,  J.C.  (Extraction Met. Div. Mines Branch, Dept.
        Mines Tech.  Surv., Ottawa) Analytical Procedures for a  Vanadium
        Recovery  Process, Can.  Dept.  Mines Tech. Surv., Mines Branch Tech.
        Bull. TB  79, p.62  (1966).

-------
                                  - 114 -
221.   Rich, T.A., (General Electric, Res. & Development Center, Schenectady,
       New York),  Apparatus and Method for Measuring the Size of Aerosols,
       J. de Recherches Atmospheriques, p.79-86, (1966).
222.   Kuz'micheva, M.N., Determination of Vanadium in the Air of Working
       Areas and in Biological Media, Vop. Gig. Tr. Prof. Patol, 369-79
       (1966).
223.   Kuzmicheva, M.N. (F.F. Erisman Inst. Hyg., Moscow) Determination
       of Vanadium in the Atmosphere, Gigiena & Sanit  .31(2), 53-4 (1966)
224.   Wehrberger, F., Problems in Oil-Fired Boilers of Industrial Power
       Plants, Mitt.  Ver. Grosskesselbesitzer #98:349-57 (October 1965);
       (Abstr.) Combustion 37 #12:47 (June 1966).
225.   Suboch, W.P. (Brit. Am. Oil Co. Ltd.) Vanadium Analyzer Useful Tool,
       Oil Gas J. 64 #9:116-17,119 (2/28/66).
226.   Kuz'micheva, M.N. (F.F. Eoisman Sci. Res. Inst. Hyg.  Moscow)
       Method for Detecting Vanadium in the Air, Gigiena & Sanit. 30(5),
       71-2(1965).
227.   Bird, R.J., Small, N.J.H. (Shell Res. Ltd.)  The Combustion of
       Heavy Fuel Oil:  Some Observations on Carbonaceous Droplet Residues
       and on the Ash from the Gas Stream, J. Inst. Petrol., 51 #494:71-77
        (February 1965).
228.   Bol'shakov, G.F., An Investigation of Ash Composition in Hydrocarbon
       Fuels, Khim. i Teknol. Toplivi Massel 9 #10:48-52 (October 1964).
229.   Bender, R.J.,  How Low for Low Excess Air, Power 108 #3:63-65 (March 1964);
       (Abstr.) Corrosion Abstr. 3, #5:367 (September 1964).
230.   Guinn, V.P., Wagner, C.D. (Shell Develop. Co., Calif.) Instrumental
       Neutron Activation Analysis, Anal. Chem. 32, 317-23 (1960).
231.   Hansen, J., Hodgkins, C.R. , Wet Ash Spectrochemical  Method for
       Determination of Trace Metals in Petroleum Fractions,  Exxon Research
       and Engineering Company, Linden, Analytical Chemistry, Vol. 30,
       No. 3, pages 368-372, March 1958.
232.   Hirt, R.C., Doughman, W.R. and Gisclard, J.B. (American Cyanaraid,
       Stamford, Conn.) Application of X-Ray Emission Spectrography to
       Airborne Dusts in Industrial Hygiene Studies, Anal. Chem. 28, 1649-51
       (1956).
233.   Allen, N.P., Kubaschewski, 0., Goldbeck, von, 0., (Nat. Physical Lab,
       Teddington, Middlesex, Eng.) The Free Energy Diagram of the Vanadium
       Oxygen System , J. of the Electrochemical Soc., p. 417-423
       (November 1951).
234.   Milan, E.F., The Dissociation Pressure of Vanadium Pentoxide, Vol. 33,
       p. 498-508 (1929).
235.   Ammonium Sulphate and Vanadium Recovery from Flue Dust Obtained
       During Heavy or Crude Oil Combustion, Azuma Kako Co., Ltd. JA-7141211-R.
       Analysis of Waters Containing Vanadium - By Complex,   JA-4850559-Q
       Collecting Metal Oxide Particles from Furnace Waste Gas - For
       Addition to Heavy Oil to Reduce Pollution on Combustion, Showa Denko KK
       JA-7J29283-R.
236.   Zegel, W.C. (Doylestown, Pa., assignor to Scott Research Lab., Pa.)
       Method and Equipment for Catalytic Analysis of Gases, Filed June 3, 1969
       Ser. No. 830,012 - U.S. 3,687,631.
237.   Pollard, A.J.  (Rosecroft Park, Md. assignor to the United States of America
       as represented by the Secretary of the Navy), Method  for Removing
       Vanadium Deposits from the Fire Side of Heat Transfer Surfaces.,
       #3,369,934

-------
                                - 115 -
238.  Whigham, W,  (Ville de'Anlon,  Quebec,  Canada),  Production  of
      Vanadium Values  from  Crude  Oil,  #3,416,882.
239..  Richardson,  H.L.,  (Pittsburgh, Pa., assignor  to  Chemical  Construction
      Corporation, New York)  Removal  of Solids  from Flue  Gas,  #3,568,403.

-------
              -  116  -
             APPENDIX 2

LABORATORY ISOLATION OF PARTICIPATES
COLLECTED FROM STACK SAMPLING TRAINS

-------
                              APPENDIX 2
                 Laboratory Isolation of Particulates
                 Collected from Stack Sampling Trains
MATERIALS
     4    800 ml beakers
     2    1,000 ml beaker
     1    2,000 ml beaker
     8    30 ml porcelain crucibles previously marked with run no.
          and appropriate designations .
          crucibles placed in furnace @ 1400°F for several hours.

     1    Metier 5 decimal place balance Model H20T
     1    Blackstone ultra-sonic cleaner
     1    Millipore stainless steel 47 mm pressure filter holder
REAGENTS

     Hexane
     Isopropyl Alcohol
     Dist. H20
     Chloroform

FIRE CHAMBER

          The solids accumulated in the fire chamber are swept into
a teflon container then sieved through a 100 mesh screen to remove
extraneous particles such as pieces of fire brick.

          The two fractions, <100 and >100 are transferred to tared
bottles and put into the vacuum oven to dry at 275°F for four hours »  The
weight is recorded and the sampler submitted for chemical analysis.

BOILER TUBE INSERTS

          After completion of the warm up period, the boiler was shut down
and the rear door opened.  A single stainless steel tube was inserted
into a randomly selected boiler tube in each of the three passes of the
Cleaver Brooks Boiler.  The rear door was closed and the boiler restarted.
As soon as run conditions were achieved, stack sampling began.  Weights
of fuel at restart and at start of stack sampling and at completion
of run were recorded.

          The boiler tube inserts were removed and labelled as to their
location in the passes.

          The tubes were cleaned in the following manner; the tube insert
was mounted vertically to platform with clamps and a labelled 1000 ml
beaker placed under it.  The tube was washed alternately with dist. H20
and isopropyl alcohol. A plunger consisting of a rubber disc conforming
to the inner diameter of the tube and mounted on a long stainless steel
rod was passed up and down the length of the tube several times in order
to remove any material that might adhere to the sides of the tube.

-------
                               - 118 -
Following this the tube was again rinsed with dist. H20 and isopropyl
alcohol.  The beaker was removed and put on steam bath and evaporated to a
manageable volume.  This was transferred to a tared and labelled
crucible.  Then evaporated to dryness.  The solids were placed in a
vacuum oven at 275°F to dry.   After weighing,samples from each of
the three batches of solids were submitted for vanadium analysis.
In several tests, the final wash of the tubes was accomplished using
about 500 ml of 0.5 NH Cl. The solution was submitted directly for
V analysis.

PROBE

           Cap off both ends of probe and  rinse off all material  that
was collected on  outside  of the probe during  run.  Remove  cap  from
nozzle  tip and wash inside  of nozzle and probe with dist.  1^0.  When
probe is  full place caped end into ultra-sonic bath for approx. 30  sec.
Pour into 250 ml  beaker.  Repeat using  isopropyl  alcohol.  Remove cap
and rinse probe from  capped and directly into beaker  with  dis.  H20  and
isopropyl alcohol.

           Glass  connector, probe to Andersen:  Repeat same  procedure
used in washing probe.  Rinse into 250  ml  beaker  containing  probe
wash.   Place on steam bath  to evaporate.
 CYCLONE

           Solids  are recovered from cyclone using probe  isolation
 procedure.   Solids  are dried under vacuum at 300°F.

 ANDERSEN  CASCADE  IMPACTOR

           8 separate stages  with a corresponding  stainless  steel
 plate  are washed  in same  manner.   (The  head of  the Andersen is  washed
 along  with the "0"  stage  and plate).  Carefully remove stage, wash rim
 and face  with dist. H20 and  isopropyl  alcohol  into  beaker.  Reverse
 stage  and repeat  procedure.   If any material appears to  remain, use
 rubber policeman  to remove.   Remove stage and using  forceps place
 corresponding plate into  the washings,  cover with water  glass.  Place
 beaker with washings and  plate into ultra-sonic bath for approx. 30
 sec. or until all solids  are off plate.   Remove plate from  washings.
 Rinse  both sides  into beaker with dist.  water and isopropyl alcohol.
 Place  beaker on steam bath  to evaporate.

           After the last  plate has been  removed the  base of the
 Andersen  neck, and  connector leading to  the impingers will  remain.
 They are  rinsed with dist. H20 and isopropyl alcohol into a 1,000
 ml  beaker marked  "impinger water wash".   Cover  beaker and keep  it
 aside  to  be used  again later on in work-up.

-------
                               - 119 -
GLASS CONNECTORS

          Remove stop-cock grease.  Fill connector with hexane.
Cap both ends and place in ultra-sonic bath for 1 minute.  Pour into
2,000 ml beaker with impinger washings.  Repeat again, then rinse
connectors until all traces of silicone oil are removed.  Allow to
dry.

FILTRATION OF SILICONE OIL

          Stir oil and hexane washings in 2,000 ml beaker to be
filtered.  Weigh  1 millipore teflon .20 mu filter paper and place in
pressure filter and lock.  Add approx. 300 ml of impinger wash from
beaker to sample container.  Lock.  Use approx. 60 psi pressure
to filter.  All of sample should filter in 3 hours.  At end of
filtration rinse sample container with 3 x 100 ml increments of
hexane and pressure these through filter.  When filtration is complete
remove filter paper and dry in vacuum oven at 300°F for four hours.
Remove filter paper, place in desicator to cool.  Weigh when filter
paper is at room temperature.

WATER WASH OF IMPINGERS
          (This is done while impingers are filtering) check dry
impingers for any traces of silicone oil.  If there are traces of
silicone oil, wash with hexane again adding washings to impinger sample,
when there is no further trace of oil present, the dry impingers are
washed with distilled water and isopropyl alcohol in the same manner
as with the hexane wash.  Should any trace of material remain after the
water isopropyl alcohol wash an additional wash using chloroform should
be used.  All washings should be added to previously obtained "impinger
water wash" beaker and placed on steam bath to evaporate.

WATER IMPINGERS

          After removal from sampling train removal all traces of
stop-cock grease from impingers.  Place each impinger in ultra-sonic
bath for approximately five minutes, empty contents into a 2,000 ml
beaker labelled "impinger H20 wash".  Each impinger is washed by the
same method; thoroughly  rinse all inner surfaces of impinger, neck and
arm into impinger with 100-150 ml of water.  Place in ultra-sonic
bath for two cr three minutes.  Empty into 2,000 ml "impinger wash"
beaker.  Repeat same procedure using isopropyl alcohol.  Pour wash
into beaker.  If any material remains, an additional wash using
chloroform is used.   Evaporate contents of "impinger wash" to dryness
on steam bath.  Transfer material from beaker to tared 30 ml crucible

-------
                                  - 120 -
using water, isopropyl alcohol as a rinse.   Evaporate crucible to
dryness.  Place in vacuum oven at 300°F for four hours.    Cool in
dessicator to room temperature.  Weigh and  record weight on particulate
emissions data sheet.  Submit dry sample in screw neck vials for
typical analysis.

PROBE, PLATES, IMPINGER WATER WASH SAMPLES

           Weigh 30 ml crucibles and record tare weights.  Transfer
samples on steam bath to appropriate crucibles rinsing with water and
isopropyl alcohol.  Evaporate to dryness.  Place in vacuum oven at
300°F.  Remove crucibles, place in dessicator, weigh when crucibles
are at room temperature.  Record weights of all fractions on particulate
emissions data sheet.
 SILICONS  OIL  IMPINGERS

           Open  all  impingers  and wipe  off  stop-cock  grease  from  ground
 glass  joints.   Place  on steam bath.  This  will help  make  silicone  oil
 less viscous  and  easier to handle  (this  can  be done  prior to work-up
 of Andersen so  silicone oil will be  reasonably fluid for  work-up).
 Pour  contents of  all  impingers into  2,000  ml beaker, rinsing with
 hexane.   Pour approx.  250 ml  hexane  into each impinger.   Place impinger
 into  ultra-sonic  bath for approx.  2  mins.  while impinger  is in ultra-
 sonic  bath, use squeeze bottle and wash  inside of impinger  neck and
 arm into  the  impinger.   Pour  hexane  into 2,000 ml beaker.  Wash
 impinger  bottle and neck thoroughly  with hexane until no  trace of
 silicone  oil  is present.  Repeat with  each impinger.  Allow impingers
 to dry.


 COMBINATION OF  FRACTIONS

            After  all  weights  are  recorded, samples are combined into
 three fractions:   probe and "0";  plates  1  thru 5; plates  6,7 and impinger,
 and submitted for vanadium analysis  by Atomic Absorption  Spectroscopy.
 The filters,  one  from the filtration of  the  silicone oil  in the ER&E
 train, and the  fiber  glass filter  from the Method 5  train,  after drying
 under  vacuum  at 300°F for four hours  are  also submitted  for vanadium
 analysis.

 ASH PROCEDURE

            Using  three  weighing boats  that have been previously heated
 in a  furnace  at 1450°F  overnight and cooled  to room  temperature in a
 dessicator.  Obtain tare weights on  boats.  Place approximately 10 mg.
 sample of material  from each  of the  three  fractions  (dried  at  500°F
 overnight) into the boats. Weigh  and record.  Place boats in furnace
 overnight at  1000°F.   Following morning  remove boats, place in dessicator
 to cool  to room temperature.   Weigh  and record sample loss.   Calculate
 ash.

-------
                                       COMPARISON  OF  TIME  REQUIREMENTS
                                       FOR  WORK  UP OF SAMPLING  SYSTEMS


Probe
Cyclone

Filter
Impinger
(+ Connectors)
Leak Check
Totals


EPA SYSTEM
Prep. Work Up*
5 10
5 10

15 15
10 45

5
40 mins. 80 mins.


ER&E

Probe + 0
Andersen 1-5

Andersen 6,7
Impingers (filtering)
(+ Connectors)
Traps
Leak Check
Totals

SYSTEM
Prep. Work Up*
5 15
45
5
15
15 210

15 20
5 - i
45 mins. 305 mins. M
i
Total Work Up + Prep.  = 2 hours
Total Work Up + Prep.   5 Hrs. 50 mins.
*Exclusive of drying isolated solids.

-------
                    -  122  -
                 APPENDIX 3
THE DETERMINATION OF VANADIUM IN PARTICULATES

-------
                                 - 123 -


                              APPENDIX  3

              The Determination  of Vanadium in Particulates


                       ABSORPTION SPECTROSCOPY
          Atomic absorption spectroscopy  is a  single element  technique
requiring at least 3 cm^ of sample  solution per  determination.  Due  to
the small sample sizes encountered  in this project  only  the vanadium
determination was evaluated.

Sample Preparation

          The solid particulate samples were dried  before weighing.   The
dried sample was then digested with a mixture  of perchloric,  nitric, and
sulfuric acids.  This acid digestion was  found to be quicker  than dry
ashing and is not as subject to error through  loss  when  working with
small samples.

          The teflon and silica filter particulate  samples  were extracted
with 5% HN03 acid.  It was found that at  least 99%  of  the vanadium is
removed by this  technique.  Aqueous samples were analyzed directly.


Accuracy

          The accuracy of an atomic absorption analysis  is  dependent
upon the accuracy of the calibration standards,  as  well  as  physical
and chemical  interferences.  Reagent grade  VOS04 , ^'205 and  Fisher
1000 p.p.m. standards were used.  Calibration  standards  prepared  from
each of these sources agreed quite  well.

          Several chemical species  have  been  reported  to chemically
interfere with the vanadium determination.  Sachaleu,  Robinson, and
West^1' have reported using the addition  of A1C13  to  overcome  these
interferences.  This technique was  therefore used for  all vanadium
determinations associated with this project.

          The accuracy of vanadium  by optical  emission has  been well
established and was therefore used  to check  the  atomic absorption
results.  Good agreement was obtained as  can be  seen from an  inspection
of Table I.

-------
                                -  124  -
                                TABLE 1

         COMPARISON OF VANADIUM ANALYSIS  BY  ATOMIC ABSORPTION
                   AND OPTICAL EMISSION SPECTROSCOPY
                            	wt% V	
     Sample                    A. A.           0.  E.  S,
       1                        3.7             3.8
       2                        3.3             3.8
       3                        4.1             3.6
       4                        2.3             2.2
       5                        3.6             3.7
       6                        2.6             2.4
       7                        5.0             4.9
       8                        8.4             8.4


Sensitivity

          The sensitivity of the vanadium determination is  such that  a
minimum of 50 yg of vanadium is required per analysis.   This means  that
a 2 mg particulate sample containing less than 2.5 wt70  V could  not  be
determined by this technique.
  'S.  L. Sachaleu. J. W. Robinson, and P.  W.  West,  Anal.  Chim.  Acta.,
   37 (1967) 12.

-------
                                 -  125  -
Precision

          A day to day precision study was conducted using a large sample
which had been sieved, dried, and thoroughly mixed.  Aliquots of this
sample were put through the entire procedure and the resulting sample
solutions were saved.  Each day a new sample solution was prepared and
analyzed with previous preparations.  The columns of data in Table II
are the repetitive results for each sample preparation and were used to
calculate the day-to-day precision of the measurement step.  The top
diagonal of data in Table II are the first results obtained for each
new sample preparation and were used to calculate the day-to-day precision
for the entire procedure.  An average relative standard deviation (la)
of 2.8% was calculated for the measurement step as compared to 4.3% for
the entire procedure.  These results show that the sample preparation
can be carried out as precisely as the measurement.

                      PROCEDURE-ATOMIC ABSORPTION

Apparatus

          A P&E Model 403 atomic absorption spectrometer, equipped with
a vanadium hollow cathode lamp and a strip chart recorder were used in
this study.

Instrumental Parameters
          The following parameters apply to the P&E Model 403 unit.

     Wavelength  -  319.2 nm; UV setting
     Lamp Current  -  14 mA
     Slit 4 (Spectral Band Pass - 0.2 nm)
     Fuel - Acetylene
     Oxidant - Nitrous oxide
     Flame Stoichiometry -  Reducing; red cone
     1 - 1 1/2 cm in height

                            REAGENTS (A.A.)
                                     3
Standard Vanadium Solution 1000 Mg/cm

          Available from Fisher Scientific Company.

Digestion Acid Mixture

          To 150 cm  of cone. HN03 add 50 cm3 of 72% HCIO^ and 180 cm3 of
cone ^SO,/;.  After stirring and cooling, the solution is made up to 400 c
with cone. HN03.

-------
TABLE II



,^Date Prepared
Date Analyze'3""*--^^
1/21
1/22
1/24
1/30
1/31
2/11
2/13
2/21
2/24
2/25



1/21

6.58
6.36

6.22
6.46
6.67
6.76
6.31
6.62
6.64



1/22


6.28

6.28
6.28
6.64
6.82
6.37
6.64
6.73
VANADIUM


1/24



6.48

6.71
6.78
6.78
6.56
6.56
6.59
PRECISION


1/30




6.56
6.36
6.92
6.92
6.50
6.74
6.83
STUDY


1/31 2/11 2/13 2/21 2/24 2/25





6.39
6.86 6.26
6.76 6.62 7.07
6.66 6.15 7.16 6.86
6.80 6.50 7.25 6.95 6.84
6.77 6.50 7.09 6.07 7.02 6.94

-------
                                 - 127 -

                            3
Aluminum Solution 3000 ug/cm
                                                      3
     -    Weigh 6.00 g of aluminum metal into a 400 cm  beaker.  Add
40 cm  of aqua regia and a few milligrams of a mercury salt.  Transfer
to a 2-litre volumetric flask after dissolution is complete and dilute
to volume with deionized water.

Nitric Acid Solution 5% (v/v)

          Add 50 cm-^ of concentrated nitric acid to a 1-litre Volumetric
flask and dilute to volume with  deionized water.

Calibration

          (1)  Add 0, 1, 5, 10,  and 15 cm  of the 1000 ug/cm  stock vana-
dium solution to each of 5-100 cm^ volumetric flasks.
                                               3
          (2)  Add to each of these flasks 1 cm  of the acid digestion
mixture and 10 cm^ of the aluminum solution.  Then dilute up to volume
with deionized water and mix well.  This makes 0, 10, 50, 100, and 150 ug/cm
vanadium standards, respectively.

       2  (3)  Aspirate each of  the standards into the flame using the
0 Mg/cm  solution as the blank.

     T    (4)  Prepare a calibration curve plotted as absorbance versus
Mg/cm  vanadium.

Sample Analysis

          The samples associated with this project consisted of three
types:  free particulate, particulate on filter paper (teflon or silica),
and aqueous extract solutions.

     A.  Free Particulates

          (1)  Dry the sample for at least 1 hour at 110°C and allow to
cool in a desiccator containing  anhydrous magnesium perchlorate or
calcium chloride as the desiccant.

          (2)  Weigh between 2 and 100 milligrams of sample, depending
upon availability, into a 30 ml Kjeldahl digestion flask.

          (3)  Place the flask onto a Kjeldahl digestion rack equipped
with a glass hood and bubbler system.

          (4)  Add 2 cm  of the acid digestion mixture and heat slowly
until the brown fumes disappear and the solution is clear.

          (5)  Bring the mixture to fumes of perchloric acid and continue
to heat for 15 minutes.

-------
                                 - 128 -
          (6)  Cool, add 5 cm  of water, and bring to a boil.

          (7)  Transfer the digestate quantitatively to a 10 cm  volu-
metric flask after cooling.

          (8)  Add 1 cm  of the aluminum solution and dilute to volume
with deionized water.

          (9)  Measure the absorbance of the sample solution and determine
the corresponding ug/cm3 of vanadium from the calibration curve.  If the
absorbance is higher than the highest calibration standard, dilute to
within the working range making certain that the final solution contains
300 yg/cm3 of Al+3.
         (10)  Calculate the wt. % vanadium using the following equation:

                             > (cm  of final solution) (c
                             sample weight (milligrams)
    „.     ,.     (ug/cm  vanadium) (cm  of final solution) (dilution factor) x 10
wt.  7, vanadium = -JJ=a	
     B.  Filter Samples

          (1)  Cut the entire filter into small pieces being careful not
to lose sample and place into a 400 cm3 beaker.
                        3
          (2)  Add 50 cm  of 5% nitric acid to the beaker.

          (3)  Place a watch glass over the beaker and boil for 30 minutes.

      ^   (A)  Cool and filter through //I Whatman filter paper into a
100 cm  volumetric flask.

          (5)  Add 10 cm  of the aluminum solution and dilute to volume
with deionized water.

          (6)  The remainder of the sample preparation is described in
steps 9 and 10 of Part A. (Free Particulates).

     C.  Aqueous Solutions
                                       3                           3
          (1)  Accurately transfer 9 cm  of the sample into a 10 cm
volumetric flask.

          (2)  Dilute to volume with the aluminum solution and mix well.

          (3)  The remainder of the sample preparation is described in
steps 9 and 10 of Part A. (Free Particulates).

-------
                                   - 129  -
                         EMISSION SPECTROSCOPY
          Emission spectroscopy is a multi-element analysis  technique yielding
a relative precision of at least 10%.  Our existing capability at the
onset of this project allowed the simultaneous determination of Ni,
Fe, V, Si and Na in relatively large samples.   Due to the  small
sample sizes to be encountered, work was done  to extend the  lower
working limit without affecting the accuracy or precision  of the
method.  Only solid particulate samples were studied although this
technique could be extended to filter and aqueous samples.

Sample Preparation

          The particulate samples were dry ashed in porcelain
crucibles.  The formation of a friable ash was accomplished  through
the addition of glycerin to the sample before  ashing.

Calibration Standards

          Standards consisted of the oxides of Ni, Fe,  V,  and Si,
and  NaCl.  Each element was introduced into the standard  to cover
the low, high, and intermediate points of the  range of  interest.  The
2.750 gm total for each standard consisted of  the metal oxides,
sodium chloride, and lithium tetraborate.  A straight line curve was
obtained for each element indicating that the  amount of matrix does
not produce variation in the calibration curves over the range studied.
Five standards were fired to calibrate the instrument and  prepare  the
standard curves.  Scale deflection units (Y-axis) were  plotted versus
mgs of each element (X-axis).

          The emission spectroscopy procedure  is calibrated  to cover  the
range of 0.5 to 5 mg V, 0.05 to 0.5 mg Ni, 1.00 to 25 mg Si, 1.00 to  25
mg Na, and 1.00 to 25 mg Fe.

                   PROCEDURE - EMISSION SPECTROSCOPY

Apparatus

          Direct reading ARL Quantometer.  Excitation source is a
voltage spark of the ARL multisource unit, Model #4700.

          ARL power driven hydraulic press with a 12.7  mm  die (1/2").

-------
                                 - 130 -


Instrumental Parameters

          5 sec prespark

          50 sec exposure controlled by integration of a constant
amount of energy from the internal standard.

          Set in 10-100 scale deflection units.

          Wavelengths of analytical emission lines are:
                                           o
                                           A
               Li               4972.0 (reference line)
               V                3184.0

               Ni               3414.8

               Si               2881.6
               Na               5896.0

               Fe               3020.6

Reagents

          (1)  Certified Spectranalyzed Glycerin.

          (2)  1,128407 - 4-9's pure - Spex Industries.

          (3)  SP-1 grade graphite - Spex Industries.

          (4)  Graphite rod with a flat tip - 6.35 x 50.8 mm - high purity
Spex Industries.

Procedure

          (1)  Weigh a minimum of 10 mg into a porcelain crucible and
add 1 cm  of glycerin.

          (2)  Ash the sample at 540°C for one hour.

          (3)  Weigh the ash.

          (4)  Transfer the ash to a graphite crucible.

          (5)  Cover the ash with 1126407 so that the total weight of ash
plus 1,126407 equals 2.750 g.   (Example:   If ash weight is 2.0 mg add
2.748 g  Li2B407) •

          (6)  Fuse at 980°C for 10 minutes.

-------
                                  - 131-
          (7)  Grind the glass-like bead obtained from the fusion in
a tungsten carbide mill.

          (8)  Sieve the ground material through a 200 mesh cloth.

          (9)  Blend one part of sample with two parts of SP-1
pelletizing graphite.

         (10)  Press a portion  ( ^  1  g) into a 12.7 mm briquette with
the press set at 40,000 psi.

         (11)  Place the briquette in a brass holder.  The counter
electrode is a 6.35 x 50.8 ram graphite rod.

         (12)  Set in and fire the standards and then fire the samples
on the Quantometer.

         (13)  Read and record scale deflection units for each element.
Determine the mgs of the element from the appropriate curve.

-------
                                    SUMMARY  OF

                       ATOMIC ABSORPTION SPECTROSCOPY
             Lam:


•M
-:ama ccrirair.ing
sample
j ^
H I
Monc:'.-..'orr>ator
-_^— -^ " '
'^^^-^ 1

Detecto: &
recorder
s
               Wavelength. X
Hollow cathode lamp emits
line spectrum of element to
be determined
Resultant spccVfwim after
ibcorption by sample
                                                                          A
                                                                                                              U)
                                                                                                              NJ
Photodctcctor sees only the
resonance line, diminished
by sample absorption

-------
                - 133 -
               APPENDIX 4
OPERATIONAL DATA FROM FACTORIAL PROGRAM

-------
APPENDIX 4

EXPERIMENTAL
Experiment No.
Fuel Type
Residence Time
Stack Position
Vanadium Analysis in Fuel, ppm
Operational Data
Flue Gas Oxygen, %
Nozzle Temperature, °F
Stack Temperature, °F
Fuel Consumption, Lbs.
• Firetube
• Heat Exchanger Inserts
• Particulate Sampling
Bacharach Smoke Tape No.
Boiler Deposits, grams
Firetube
Pass 2 (1 tube of 14)
Pass 3 (1 tube of 10)
Pass 4 (1 tube of 8)
Corrected Distribution, mg/SCM (Combustion Gas)
Firetube
Pass 2 (all tubes)
Pass 3 (all tubes)
Pass 4 (all tubes)
Total
Vanadium Analysis, Wt. %
Firetube
Pass 2
Pass 3
Pass 4
Vanadium Distribution In Boiler Solids, ppm (on fuel)
Firetube
Pass 2
Pass 3
Pass; 4
Total
Vanadium Recovery
Average V Content in Solids, wt. %
TABLE 1
DATA SUMMARY FOR
44
F2
TT
2
145

2.0
147
213

194
111
104
2.5

21.76
0.32
1.19
0.51

19
7
18
6
50

4.90
9.71
8.40
6.87

12
8
19
5
44

6.9

BOILER
45
£2
To
2
150

2.0
150
318

450
264
237
5.5

51.61
3.49
2.66
2.20

19
31
17
11
78

2.07
4.21
2.86
2.67

5
17
6
4
32

3.2

OPERATION
46
,13
Ti
i
40

2.0
186
220

245
172
162
1.5

19.30
1.56
0.77
0.80

14
22
8
7
51

2.91
8.81
3.81
3.64

5
25
4
3
37

5.7
 47

Ti
  1
 37


 2.0
 187
 328

 579
 452
 362
   6


94.75
15.36
 9.67
 3.42


  28
  83
  37
  11
 159


 0.76
 1.38
 0.29
 0.68


   3
  15
   1

  20


 1.0
                                               48
                                               ¥.
                                                i
                                               138
                                              2.0
                                              150
                                              318

                                              403
                                              262
                                              237
                                              5.5


                                             56.15
                                              3.42
                                              2.40
                                              1.72


                                               23
                                               31
                                               15
                                               _9
                                               78
                                                85
                                                30
                                                59
                                              3.19


                                               11
                                               21
                                                7
                                               _4
                                               43


                                              4.3
 49

 T?
  1
 140
2.0
140
215

 99
 99
 74
2.5


7.76
0.38
1.08
0.71


 13
  9
 18
 IP.
 50

5.33
9.41
4.60
3.00
  9
 11
 11
  4
 35
5.3

-------
Experiment No.
Fuel Type
Vanadium Analysis, ppm
Residence Time
Stack Position

ERE Sampling Train, Operating Data

•  Isokinetic Sampling Rate, %
•  Sample Volume SCFDB
•  Gross Particulate Weights, mg
        Probe & 0 Stage
        Andersen Stages 1-5
        Andersen Stages 6,7
        Impingers
                               Total

•  Vanadium Analysis, Wt.  %
        Probe & 0 Stage
        Andersen Stages 1-5
        Andersen Stages 6,7
        Impingers
•  Particulate Distribution, Wt. %
        >10u
EPA Sampling Train, Operating Data

•  Isokinetic Sampling Rate ,  %
•  Sample Volume, SCFDB
•  Gross Particulate Weights, mg
        Probe
        Cyclone
        Filter
        Impingers
                               Total

•  Vanadium Analysis, Wt. %
        Probe & Cyclone
        Filter
        Impingers

  *Based on Cumulative Distribution.
EXPERIMENTAL
44
F2
145
T'
2
93
43.23
11.79
11.99
5.13
24.37
53.28
1.53
2.60
15.99

23
25
52
98
56.60
17.56
2.50
59.56
50.78
130.40
2.15
14.61
None Detected
APPENDIX 4
TABLE 2
DATA SUMMARY FOR
45
F2
X¥o
2
99
78.91
55.51
141.67
18.30
75.31
290.79
1.26
2.76
10.70

21
47
32
98
78.91
14.05
54.28
221.82
103.89
394.04
1.30
6.31
None Detected
STACK SAMPLING
A6
F3
40,
Ti
1
95
.65.49
4.82
15.39
10.95
13.33
44.49
1.56
1.36
2.28

12
35
53
1.06
83.70
7.11
3.70
52.17
50.05
113.03
0.74
7.86
1.30
SYSTEMS
4?
F3
iV
1
93
112.99
218.38
385.48
22.09
126.85
752.80
Trace
0.11
1.16
6.84
31
49
20
92
121.26
119.88
209.39
459.41
20.45
809.13
0.14
1.24
None Detected

48
F2
To?
i
93
83.52
49.05
136.44
11.60
62.40
259.49
1.63
3.44
10.80
13.95
24
47
29
95
81.34
23.31
38.59
190.13
9.71
261.74
1.37
7.73
None Detected
  96
39.83

 8.90
11.03
 4.89
21.85
46.67
 0.21
   92
   56
15.26

  21
  24
  55
 1.05
43.26
 5.10
 1.63
50.32
15.25_
72.30
 Trace
15.70
 0.40

-------
                                                                APPENDIX It
Experiment No.
Fuel Type
Vanadium Analysis, ppra
Residence Time
Stack Position
Operational Data
 Flue Gas Oxygen, %
 Nozzle Temperature,°F
 Stack Temperature, °F
 Fu2l Consumption, Lbs.
   o Fire tube
   o Heat Exchanger Inserts
   o Particulate Sampling
 Bacharach Smoke Tape No.

Boiler Deposits, grams
   Firetube
   Pass 2
   Pass 3
   Pass 4
             (1 tube of 14)
             (1 tube of 10)
             (1 tube of 8)
                   Total
                                 (Combustion Gas)
 Corrected  Distribution,  mg/SCM	
    Firetube
    Pass  2    (all tubes)
    Pass  3    (all tubes)
    Pass  4    (all tubes)
                   Total

 Vanadium Analysis,  wt.  %
    Firetube
    Pass  2
    Pass  3
    Pass  4

 Vanadium Distribution in Boiler Solids,  ppm (on fuel)
    Firetube
    Pass  2
    Pass  3
    Pass  4
                    Total

 Vandadium  Recovery
Average V Content in Solids, wt. %
TABLE
DATA SUMMARY
50
F3
39
T°
2
2
185
325
469
264
241
6
102.54
8.41
5.33
2.54
118.82
36
77
35
13
161
1.02
0.58
0.53
0.69
5
6
2
1
14
3

FOR BOILER OPEM
51
F3
40
Tl
2
2
178
216
233
154
141
2
13.65
1.40
0.73
0.93
16.71
10
22
8
8
48
0.73
3.48
0.93
1.05
1
10
1
1
13
52
Fl
350
To
2
2
215
316
240
129
90
6
108.83
6.21
3.86
1.74
120.64
79
117
52
19
267
2.40
3.58
2.99
3.33
24
53
20
8
105
                                                                                            53
                                                             0.7
                                                                       2.1
                                                                                  3.1
                                                                                             368
                                                                                             Ti
                                                                                               2
  2
205
214

122
 70
 48
  2

10.87
 0.88
 0.40
 0.41
12.56

 15
 31
 10
  8
 64
                                                                                             7.82
                                                                                             9.15
                                                                                             6.51
                                                                                             5.62

                                                                                             15
                                                                                             36
                                                                                              8
                                                                                              6
                                                                                             65
                                                                                             8.0
54
Fl
350
Ti
1
2
201
218
173
80
65
2
11.70
0.94
0.50
0.41
13.55
12
28
11
7
58
10.70
12.26
7.80
7.00
16
45
11
6
78
55
Fl
358
(To
1
2
205
346
556
301
114
5
111.94
10.15
6.50
1.70
130.29
35
81
38
8
162
5.31
5.19
4.82
7.10
24
54
23
7
108
56
Fl
^
1
2
198
217
203
90
67
2
14.90
0.87
0.43
0.41
16.61
13
24
8
6
51
10.12
14.34
10.93
8.93
17
43
12
7
79
57
.FI
363
1
2
200
341
427
214
118
5
148.93
13.04
5.84
1.75
169.56
61
148
47
11
267
3.34
2.12
3.17
4.44
26
40
19
6
91
                                                                                                       10.6
                                                                                                                  5.2
                                                                                                                            12.2
                                                                                                                                          2.7

-------
                                                              APPENDIX  4
                   ppm
Experiment No.
Fuel Type
Vanadium Analysis,
Residence Time
Stack Position
ERE Sampling Train, Operating Data
  • Isokinetic Sampling Rate, %
  • Sample Volume, SCFDB
  • Gross Particulate Weight, mg
          Probe + 0 Stage
          Andersen Stages 1-5
          Andersen Stages 6, 7
          Impingers
                  Total
  • Vanadium Analysis, Wt. %
          Probe + 0 Stage
          Andersen Stages 1-5
          Andersen Stages 6,7
          Impingers
  • Particulate Distribution*
                                                                                52
                                                                                           53
                                                                                                                 55
                                                                                                                            56
                                                                                                                                        57
                                 Wt.  %
          1-10 X.
EPA Sampling Train, Operating Data
  • Isokinetic Sampling Rate, %
  • Sample Volume, SCFDB
  • Gross Particulate Weight, mg
          Probe
          Cyclone
          Filter
          Impingers
                  Total
  • Vanadium Analysis, Wt. %
          Probe + Cyclone
          Filter
          Impingers
92
75
114.5
282.9
1 .0
173.1
98
61
2.8
7.6
19.2
23.5
94
26
71.6
142.7
26.0
43.1
96
55
6.1
17.6
16.2
61.4
95
74
4.9
18.4
22.4
80.0
98
41
93.8
113.5
9.5
40.2
93
78
11.4
21.0
24.9
88.6
92
41
112.5
207.4
13.3
43.0
                                                                              283.4
                                                                                         101.3
                                                                                                    125.7
                                                                                                                257.0
                                                                                                                           145.9
                                                                              352.6
                                                                                  02
                                                                                  99
                                                                                0.31
135.7

  5.40
 25.06
  None
171.1

  8.33
 22.86
  0.66
381.4

  3.78
 15.63
  None
186.1

  5.21
 21.95
  0.56
                                                                                                                                       376.2
0.44
0.46
1.18
3.28
23
57
20
93
86
30.3
58.4
229.0
93.2
2.50
2.73
2.50
17.76
6
18
76
105
75
3.3
1.7
42.4
16.6
1.10
4.18
3.82
8.25
27
51
22
96
34
57.9
66.7
213.6
14.4
4.17
13.11
18.38
19.09
6
22
72
96
62
3.9
7.4
107.0
17.4
7.00
18.86
25.74
21.77
4
17
79
95
79
4.2
3.3
150.5
13.1
2.43
8.90
16.80
13.12
35
46
19
105
44
45.1
112.9
112.6
10.8
3.88
22.97
18.52
23.71
7
17
76
94
84
9.5
7.3
158.5
10.8
1.66
4.62
14.29
9.31
32
54
14
105
48
43.2
260.3
128.7
9.3



i
OJ
-J
1






441.5

  2.29
 11.96
  0.26
      *Based  on  Cumulative  Distribution.

-------
                                                                    APPENDIX 4
Experiment No.
Fuel Type
Vanadium Analysis, ppm
Residence Time
Stack Position

Operational Data
  Flue Gas Oxygen, %
  Nozzle Temperature, °F
  Stack Temperature, °F
  Fuel Consumption, Lbs.
    • Firetube
    e Heat Exchanger Inserts
    • Particulate Sampling
  Bacharach Smoke Tape No.

Boiler Deposits, grams
    Firetube
    Pass 2  (1 tube of 14)
    Pass 3  (1 tube of 10)
    Pass 4  (1 tube of 8)
                  Total

Corrected Distribution, mg/SCM
    Firetube
    Pass 2  (all tubes)
    Pass 3  (all tubes)
    Pass 4  (all tubes)
                  Total

Vanadium Analysis, wt. %
    Firetube
    Pass 2
    Pass 3
    Pass 4

Vanadium Distribution in Boiler Solids, ppm  (on fuel)
    Firetube
    Pass 2
    Pass 3
    Pass 4
                  Total
Vanadium Recovery
Average V Content in Solids, wt. %
(Combustion Gas)
TABLE 5
EXPERIMENTAL DATA SUMMARY
58
F3
51

2
2
188
357
518
264
184
6
66.05
7.71
5.96
3.16
82.88
22
71
39
17
149
2.55
1.07
0.75
1.22
7
10
4
2
23
FOR BOILERS
59
F3
42
Of
2
2
186
234
267
109
55
2
8.00
0.82
0.84
0.88
10.54
5
18
13
11
47
4.06
2.54
1.34
1.29
3
6
2
2
13

60
F3
38

1
2
178
218
159
75
62
2
10.52
0.57
0.50
0.33
11.92
12
19
12
6
49
2.70
1.91
1.07
2.62
4
5
2
2
13
                                                                          62
                                                                         F2
                                                                         158
                                     1.2
                                                  2.1
                                                               2.1
   2
 145
 222

 213
 108
48/82
   2


 10.47
  0.59
  0.37
  0.26
 11.59


   8
  13
   6
   3
  30


  6.88
 11.51
  5.49
  4.64


   7
  19
   4
  _2
  32

  8.3
              63
             F2
                                                                                       2
                                                                                     145
                                                                                     341

                                                                                     687
                                                                                     325
                                                                                     178
                                                                                       6

                                                                                     81.94
                                                                                      6.09
                                                                                      4.39
                                                                                      2.20
                                                                                     94.62
                                                                                      20
                                                                                      44
                                                                                      23
                                                                                       9
                                                                                      96

                                                                                      3.68
                                                                                      3.47
                                                                                      2.94
                                                                                      3.64

                                                                                       9
                                                                                      19
                                                                                       8
                                                                                      _2
                                                                                      40

                                                                                      3.3
  65
 F3
  40
   2
 186
 346

 667
 324
120/160
   6


 83.94
  8.01
  6.54
  4.57
103.06
  22
  60
  35
  20
 137

1.59/1.69
1.07/1.00
0.61/0.75
0.35/0.42

   5
   8
   3
  _2
  17

  1.0

-------
                                                                APPENDIX 4

                                                                 TABLE 6

                                               EXPERIMENTAL DATA SUMMARY FOR STACK SAMPLING SYSTEMS
Experiment No.
Fuel Type
Vanadium Analysis, ppm
Residence Time
Stack Position
ERE Sampling Train, Operating Data
  • Isokinetic Sampling Rate, %
  • Sample Volume, SCFDB
  • Gross Particulate Weight, mg
        Probe + 0 Stage
        Andersen Stages 1-5
        Andersen Stages 6,7
        Impingers
                     Total
  • Vanadium Analysis, We. %
        Probe + 0 Stage
        Andersen Stages 1-5
        Andersen Stages 6,7
        Impingers
  • Particulate Distribution*  , Wt.  %
          >10y
         1-lOu
         
-------
                                                                  APPENDIX 4
                                                                  TABLE 7
                                                      EXPERIMENTAL DATA SUMMARY FOR BOILERS
Experiment No.
Fuel Type
Vanadium Analysis, ppm
Residence Time
Stack Position
Operational Data
  Flue Gas Oxygen, %
  Nozzle Temperature, °F
  Stack Temperature, °F
  Fuel Consumption, Lbs.
    • Firetube
    • Heat Exchanger Inserts
    • Particulate Sampling
  Bacharach Smoke Tape No.
Boiler Deposits, grams
                                                               66
67
                                                                                       68
                                                                                                   69
                                                                                                               70
                                                                                                                           71
                                                                                                                                       72
    Firetube
    Pass 2   (1
                tube  of  14)
    Pass 3
    Pass 4
             (1 tube  of  10)
             (1 tube  of  8)
                    Total
                                 (Combustion Gas)
Corrected Distribution, mg/SCM
    Firetube
    Pass 2  (all tubes)
    Pass 3  (all tubes)
    Pass 4  (all tubes)
                     Total
Vanadium Analysis, wt. %
    Firetube
    Pass 2
    Pass 3
    Pass 4
Vanadium Distribution in Boiler Solids, ppm  (on fuel)
    Firetube
    Pass 2
    Pass 3
    Pass 4
                      Total
Vanadium Recovery
Average V Content in Solids, wt. %
5
/Ti
2
2
183
219
202
91
69
2
14.79
0.81
0.81
0.74
17.15
13
20
15
11
59
1.82
1.82
0.78
0.85
3
5
1
1
10
343
/IT
1
2
199
217
216
97
53
2
22.85
1.09
0.75
0.29
24.98
19
27
14
4
64
11.92
12.00
4.91
6.96
28
42
8
4
82
363
at
i
2
205
354
296
296
152
6
100.94
8.36
4.54
1.58
115.42
60
69
27
7
163
4.29
5.15
6.32
7.38
33
45
22
7
107
F2
/ft
1
2
150
446
627
296
153
6
84.00
4.49
3.69
2.14
94.32
23
36
21
10
90
5.45
7.34
4.66
2.86
16
33
12
4
65
!§£>
(h
1
2
151
218
222
66
56
2
18.36
0.29
0.14
0.14
18.93
14
10
4
3
31
2.53
11.97
8.73
7.90
4
16
4
3
27
Fl
383,

2
2
197
413
820
314
179
6
212.04
13.12
6.30
1.71
233.17
45
102
35
8
190
4.39
4.93
4.67
10.06
25
64
21
10
120
Fl
335;
nr
2
2
195
220
197
102
91
2
26.13
1.08
0.62
0.38
28.21
23
26
11
5
65
10.29
15.58
9.62
9.60
30
51
13
6
100
                                                               1.3
                                                                            10.1
                                                                                        5.2
                                                                                                    5.8
                                                                                                                6.9
                                                                                                                            5.0
                                                                                                                                      12.2
                                                                                                                                                   o
                                                                                                                                                   i

-------
Experiment No.
Fuel Type
Vanadium Analysis, ppm
Residence Time
Stack Position
ERE Sampling Train, Operating Data
                              mg
• Isokinetic Sampling Rate, %
• Sample Volume, SCFDB
• Gross Particulate Weight,
     Probe + 0 Stage
     Andersen Stages 1-5
     Andersen Stages 6,7
     Impingers
                     Total
• Vanadium Analysis, Wt . %
     Probe + 0 Stage
     Andersen Stages 1-5
     Andersen Stages 6,7
     Impingers
• Particulate Distribution
                                 Wt .  %
         1-lOp
EPA Sampling Train, Operating Data
   • Isokinetic Sampling Rate, %
   • Sample Voluem SCFDB
   • Gross Particulate Weight, mg
           Probe
           Cyclone
           Filter
           Impingers
                      Total
   • Vanadium Analysis, Wt . %
           Probe + Cyclone
           Filter
           Impingers
APPENDIX 4
TABLE 8
EXPERIMENTAL DATA SUMMARY








0
2











1

0
66
F3
38

-------
APPENDIX 4


TABLE 9


EXPERIMENTAL DATA SUMMARY FOR BOILERS
Experiment No.
Fuel Type
Vanadium Analysis, ppm
Residence Time
Stack Position
Operational Data
Flue Gas Oxygen, %
Nozzle Temperature, °F
Stack Temperature, °F
Fuel Consumption, Lbs.
• Firetube
• Heat Exchanger Inserts
• Particulate Sampling
Bacharach Smoke Tape No.
Boiler Deposits, grams
Firetube
Pass 2 (1 tube of 14)
Pass 3 (1 tube of 10)
Pass 4 (1 tube of 8)
Total
Corrected Distribution, mg/SCM (Combustion Gas)
Firetube
Pass 2 (all tubes)
Pass 3 (all tubes)
Pass 4 (all tubes)
Total
Vanadium Analysis, wt. %
Firetube
Pass 2
Pass 3
Pass 4
Vanadium Distribution in Boiler Solids, ppm (on fuel)
Firetube
Pass 2
Pass 3
Pass 4
Vanadium Recovery Total
Average V Content in Solids, wt. %
73
F2
150
T*
2

2
150
238

199
88
66
2

9.06
0.42
0.35
0.25
10.08

8
11
7
4
30

8.65
11.71
8.24
9.40

9
17
7
4
37
9.8
74
F2

«*
2

2
150
455

662
358
252
6

83.55
5.36
4.32
2.99
96.22

21
35
20
11
87

4.43
3.90
3.43
2.91

12
18
9
4
43
3.8
75
F2
150
'/*
1

2
150
410

378
207
117
6

63.22
4.02
2.91
1.50
71.65

28
45
24
10
107

3.22
4.48
3.50
3.29

11
26
10
4
51
3.7
76
F2
145
rfi~
1

2
150
225

209
91
78
2

9.08
0.47
0.35
0.23
10.13

7
12
7
3
29

11.67
13.08
6.78
6.32

11
20
6
3
40
10.7
77
F3
38
T
2
185
219
174
98
89
2
2.70
0.98
0.66
0.46
4.80
3
24
12
7
46
5.60
2.17
1.88
3.29
2
7
3
2
14
2.4
78
F3
38
.-TO
2
185
391
482
196
119
6
100.85
5.72
4.37
2.30
113.24
36
71
39
16
162
1.67
1.22
1.02
1.10
8
11
5
2
26
1.2
79
&
sfr-
2
2
200
384
463
191
115
6
207.56
11.40
5.00
1.78
225.74
78
145
46
13
282
1.86
2.11
4.50
5.42
19
39
27
9
94
2.6
80
s,
  2
202
219

198
 84
 63
  2

 22.85
  0.86
  0.57
  0.47
 24.75

 20
 25
 12
  8
 65

  4.72
  9.90
  5.53
  4.70

 12
 32
  8
  5
 57
 6.9

-------
                                                                APPENDIX  4
Experiment No.
Fuel Type
Vanadium Analysis, ppm
Residence Time
Stack Position
ER&E Sampling Train, Operating Data
 • Isokinetic Sampling Rate, %
 • Sample Volume, SCFDB
 • Gross Particulate Weight (uncorrected) ,Mg
     Probe + 0 Stage
     Andersen Stages 1-5
     Andersen Stages 6,7
     Impingers
                    Total
 o Vanadium Analysis, wt. %
     Probe + 0 Stage
     Andersen Stages 1-5
     Andersen Stages 6,7
     Impingers
 • Particulate Distribution*  wt. %
EPA Sampling Train, Operating Data
  • Isokinetic Sampling Rate, %
  • Sample Volume, SCFDB
  • Gross Particulate Weight (uncorrected) , rag
     Probe
     Cyclone
     Filter
     Impingers
                    Total
Vanadium Analysis, wt. %
   Probe + Cyclone
   Filter
   Impingers
TABLE
EXPERIMENTAL
73
F2
150,
f\\
2
95
79
5.0
9.3
9.7
58.7
82.7
3.27
12.70
17.38
16.29
6
14
80
96
90
7.0
3.1
82.5
4.3
96.9
3.47
21.71
1.18
10

DATA SUMMARY
74
F2
150
/TO
2
92
83
57.6
142.9
25.3
87.4
313.2
1.99
3.05
8.29
6.66
18
46
36
97
86
29.0
49.9
156.0
20.7
255.6
1.86
6.60
0.09
75
F2
150
MO
1
97
42
45.6
82.5
10.9
39.9
178.9
1.89
3.59
5.56
8.44
29
42
29
98
42
22.6
64.8
91.5
5.0
183.9
2.11
9.18
0.83
76
F2
145
1
96
101
4.0
13.8
7.7
67.0
92.5
2.41
8.23
10.50
18.70
5
14
81
98
104
3.3
13.9
100.0
12.8
130.0
2.96
19.60
2.50
77
Fi
38^
1
94
112
8.7
25.1
6.1
44.8
84.7
1.76
2.76
6.46
13.96
12
27
61
102
122
5.2
15.5
95.4
5.5
121.6
1.88
7.10
0.49
78
F3
 99
 42

 72.9
126.6
 12.8
 66.9
279.2

  0.63
  0.80
  1.56
  2.21

  26
  45
  29

 101
  43

 43.6
 96.8
130.1
  5.1
275.6

  0.59
 15.22
  1.41
79
Fl
365
2
95
42
165.1
218.8
18.4
52.7
455.0
1.27
3.20
7.55
9.20
40
44
16
100
44
66.2
271.6
156.1
7.3
501.2
1.35
9.18
0.28
80
FI
373 .
''2
93
78
8.2
30.4
17.8
87.4
143.8
3.16
8.99
14.63
20.11
6
18
76
100
83
8.2
29.9
161.6
6.6
206.3
1.79
20.57
0.64
  *Based  on  Cumulative  Distribution.

-------
                  - 144 -
               APPENDIX 5

PARTICULATE INVENTORY AND VANADIUM BALANCE
(All data corrected to 3% 0- in flue gas)

-------
                                                                      TABLE 1
                                                 Particulate Inventory and Vanadium Balance From
                                                     Combustion of High Ash Venezuelan Resid
Particulate Distribution

    Boiler Solids

    Pass 1 (firetube)
    Pass 2
    Pass 3
    Pass 4
    Stack Particulate
    •  ER&E Train
                          Total
       1-lOp
       EPA Train

       Probe & Cyclone
       Filter

       Water Impingers
                          Total
                          Total


Total
Particulates
mg/SCM
35
81
38
8
162
75
97
39
211
123
86
(209)
8
217
(Short
Run 55
Vanadium
Analysis
Wt. %
5.31
5.19
4.82
7.10

2.43
8.90
14.53

3.78
15.63

-

Combustion Chamber Residence

Vanadium
Content
ppm(on Fuel)
24
54
23
7
108
23
110
72
205
59
171

-
230

Total
Particulates
mg/SCM
61
148
47
11
267
99
167
43
309
215
91
(306)
7
313
Time)
Run 57
Vanadium
Analysis
Wt. %
3.34
2.12
3.17
4.44

1.66
4.62
12.87

2.29
1.96

0.26



Vanadium
Content
ppm(on Fuel)
26
40
19
6
91
21
98
71
190
63
138

Trace
201
                                                                                                                                Run 52
Total        Vanadium
Particulates Analysis
mg/SCM	 Wt. %
     97
    183
     79
    359
    128
    214
        (
     14
    356
(342)
        1.10
        4.18
        7.37
1.02
5.99

0.31
                Vanadium
                Content
                ppm(on Fuel)
79
117
52
19
267
2.40
3.58
2,99
3.33

24
53
20
8
105
            14
            97
            74
           185
  17
 163

Trace
 180

-------
                                                                            TABLE 2

                                                      Particulate Inventory and Vanadium Balance From
                                                      	Combustion of High Ash Venezuelan ResId
                                                           (Short Combustion Chamber Residence Time)
                                                    Run 68
                                                                                                   Run 71
                                                                                                                                        Run 79
Particulate Distribution

  Boiler Solids

  Pass 1 (firetube)
  Pass 2
  Pass 3
  Pass 4


  Stack Particulate

  • ER&E Train
Total
     l-10y
                         Total
          Total
          Particulate
          ma/SCM	
             60
             69
             27
            	7_
            163
             55
            115
             60
            230
                                                   Vanadium
                                                   Analysis
                                                   Wt.  %
          it.29
          5.15
          6.32
          7.38
          2.59
          7.83
         10.24
                     Vanadium
                     Content
                     ppm(on fuel)
                 33
                 45
                 22
                	7
                107
                 18
                115
                 78
                211
                               Total
                               Particulate
                               mg/SCM	
                  45
                 102
                  35
                 	8
                 190
                 152
                 152
                  58
                 362
                               Vanadium
                               Analysis
                               Wt.  %
                4.39
                4.93
                4.67
               10.06
                2.16
                5.54
               11.08
                           Vanadium     Total        Vanadium  Vanadium
                           Content      Particulate  Analysis  Content
                           ppm(on fuel) mg/SCM	  Wt.%	 ppm  (on fuel)
                        25
                        64
                        21
                        10
                       120
                       42
                      108
                       82
                      232
                           144
                           158
                            58
                           360
                                        1.86
                                        2.11
                                        4.50
                                        5,42
                          1.27
                          3.20
                          8.52
                                      19
                                      39
                                      27
                                      _9
                                      94
                  23
                  64
                  63
                 150
  • EPA Train

     Probe & Cyclone
     Filter

     Water Impingers
                         Total
            102
            101

             10
            213
(203)
 4.50
14.85

 0.14
 58
191

Trace
249
280
 92

	6_
378
(372)
 2.87
13.00

<0.10
102
152

Trace
254
                                                                                    256
                                                                                    118
                                                                                                                              380
(374)
1.35
9.18

0.38
 44
138

Trace
182

-------
                                                                      TABLE 3
                                                  Particulate Inventory and Vanadium Balance From
                                                  	Combustion of High Ash Venezuelan Resld
                                                      (Long Combustion Chamber Residence Time)
                                                  Run 54
                                                                                         Run 56
                                                                                                  Run 53
Particulate Distribution

    Boiler Solids

    Pass 1 (firetube)
    Pass 2
    Pass 3
    Pass 4


    Stack Particulate

    •  ER&E Train
Total
       Total        Vanadium
       Particulates Analysis
       mg/SCM	 Wt. %
                  Vanadium
                  Content
                  ppm(on Fuel)
12
29
11
7
59
10.70
12.26
7.80
7.00

16
45
11
6
78
                        Total
                        Particulates
                        mg/SCM
                                    13
                                    24
                                     8
                                    _6
                                    51
                                                                                        Vanadium
                                                                                        Analysis
                                                                                        Wt. %
                                                             10.12
                                                             14.34
                                                             10.93
                                                              8.93
                                     Vanadium
                                     Content
                                     ppm(on Fuel)
                                                    17
                                                    43
                                                    12
                                                    J_
                                                    79
                                  Total        Vanadium
                                  Particulates Analysis
                                  mg/SCM	 Wt.  %
                                              Vanadium
                                              Content
                                              ppm(on Fuel)
16
31
10
8
65
7.82
9.15
6.51
5.62

15
36
8
6
65
        >10p
        l-10u
    •  EPA Train

        Probe & Cyclone
        Filter

        Water Impingers
                             Total
           2
          10
          45.
          57
                             Total
           3
          64.
            (
          I2.
          89
(67)
          7.00
         18.86
         23.35
 8.33
22.86

 0.66
              2
             23
            133
            158
   3
 186

Trace
 189
                 4
                11
                47
                62
(70)
                                    _4
                                    74
            3.88
           22.97
           22.83
 5.21
21.95

 0.56
             2
            31
           137
           170
   4
 177

Trace
 181
                 4
                14
                44^
                62
 6
58
  (
_9
73
(64)
           4.17
          13.11
          20.16
                                                                                      5.40
                                                                                     25.06
                       2
                      23
                     114
                     139
                                                              4
                                                            184
                                                                                                                                           188

-------
                                                       Partlculate Inventory and Vanadium Balance From
                                                            Combustion of High Ash Venezuelan Resid
Particulate Distribution
  Boiler Solids
  Pass 1 (firetube)
  Pass 2
  Pass 3
  Pass 4
  Stack Particulate
  • ER&E Train
                       Total
    1-10M
  • EPA Train

     Probe & Cyclone
     Filter

     Water Impingers
                       Total
                        Total
(Long Combustion Chamber Residence Time)

Total
Particulate
mg/SCM
19
27
14
4
64

3
11
54
68
7
70
(77)
4
81
Run 67
Vanadium
Analysis
Wt. %
11.92
12.00
4.91
6.96


1.51
12.29
14.38

5.00
17.74
0.25

Vanadium
Content
ppm(on fuel)
28
42
8
4
82

1
17
98
116
4
159
Trace
163

Total
Particulate
mg/SCM
23
26
11
5
65

3
15
46
64
5
67*
72 (37)
4
76
Run 72
Vanadium
Analysis
Wt. %
10.29
15.58
9.62
9.60


6.51
18.46
23.33

5.71
24.65
0.63

Vanad ium
Content
ppm(on fuel)
30
51
13
6
100

3
35
137
175
3
100
Trace
103

Total
Partlculate
mg/SCM
20
25
12
8
65

4
11
47
62
15
^ (80)
3
83
Run 80
Vanadium
Analysis
Wt. %
4.72
9.90
5.53
4,70


3.16
8.99
18.53

1.79
20.57
0.64

Vanad ium
Content
ppm (on fuel)
12
32
8
5
57
t-
*•
i
2
13
111
126
3
170
Trace
173
         "Average of  Runs  53  and  80.

-------
                                                        TABLE 5
                                      Particulate  Inventory and Vanadium Balance From
                                       Combustion  of  Intermediate Ash Venezuelan Resid
Particulate Distribution .

  Boiler Solids

  Pass 1 (firetube)
  Pass 2
  Pass 3
  Pass 4
               Total

  Stack Particulate

  • ER&E Train

    >10y
    l-10y
               Total
    EPA Train
     Probe & Cyclone
     Filter

     Water Impingers
(Short Residence Time Case)
Run
Total
Particulate
mg/SCM
19
31
17
11
78
26
58
40
124
29
95
(124)
44
168
45, Fuel F2
Vanadium
Analysis
Wt. %
2.07
4.21
2.86
2.67

1.26
2.76
10.70

1.30
6.31

Trace

Vanadium
Content
ppm(on fuel)
5
17
6
4
32
4
20
54
78
5
76

-
81
Run 48, Fuel
Total
Part.
mg/SCM
23
31
15
9
78
26
51
31
108
24
77
(101)
4
105
V
Anal.
Wt. %
3.85
5.30
3.59
3.19

1.63
3.44
13.42

1.37
7.73
F2
V Cont.
ppm
(On fuel)
11
21
7
4
43
5
22
53
80
4
76

Total
Part.
mg/SCM
20
44
23
9
96
34
51
46
131
88
62
Run 63
V
Anal.
Wt. %
3.68
3.47
2.94
3.64

1.33
3.44
7.18

2.70
9.66

V Cont.
ppm
(on fuel)
9
19
8
4
40
6
22
42
70
30
76
(150)
Trace

-
80
5
155
0.20

Trace
106
                                                                                                                           vo

                                                                                                                           I

-------
                                                             TABLE 6
Particulate Distribution

  Boiler Solids

  Pass 1 (firetube)
  Pass 2
  Pass 3
  Pass 4
                Total
                                      Particulate Inventory and Vanadium Balance From
                                 Combustion of Intermediate Ash Containing Venezuelan Resid
(Short Residence Time Case)
Run 69 Run 74
Total
Particulate
mg/SCM
Vanadium
Analysis
Wt. %
Vanadium
Content
ppm(on fuel)
Total
Part.
mg/SCM
V
Anal.
Wt. %
V Cont.
ppm
(on fuel)
Run 75
Total
Part.
mg/SCM
V
Anal.
Wt. %
V Cont.
ppm
(on fuel)
 23
 36
 21
 12
 90
5.45
7.34
4.66
2.86
16
33
12
_4
65
21
35
20
11
87
4.43
3.90
3.43
2.91

12
18
9
4
43
28
45
24
10
107
3.22
4.48
3.50
3.29

11
26
10
4
51
  Stack Particulate
  • ER&E Train

     >10y
     l-10y
                Total
 29
 65
 39
133
2.05
3.90
7.6
8
32
38
78
23
58
45
126
1.99
3.05
7.15

6
23
41
70
41
59
41
141
1.89
3.59
7.65

10
27
40
77
                                                                   Ul
                                                                   o
  t EPA Train

     Probe & Cyclone
     Filter
     Water Impingers
                Total
 69
 71
 11
151
                                    (140)
1.80
8.72
0.83
16
79
              _
              96
31      1.86
62      6.60
   (93)
           	8
           101
        0.09
   7
  52
Trace
  59
69      2.11
72_      9.18
  (141)

        0.83
18
84
                                               102

-------
                                                       TABLE 7

                                      Particulate Inventory and Vanadium Balance From
                                      Combustion of Intermediate Ash Venezueland Resid
                                                  (Long Residence Time Case)
                                  Run 44, Fuel F,.
                                         Run 49, Fuel F
                                                                     Run 62

Particulate

Distribution
Total
Particulate
mg/SCM
Vanadium
Analysis
Wt. %
Vanadium
Content
ppm(on fuel)
Total
Part.
mg/SCM
V
Anal.
Wt. %
V Cont.
ppm
(on
fuel)
Total
Part.
mg/SCM
V
Anal.
Wt. %
V Cont.
ppm
(on fuel)
Boiler Solids

Pass 1 (firetube)
Pass 2
Pass 3
Pass 4
            Total

Stack Particulate
19
 7
18
_6
50
        4.90
        9.71
        8.40
        6.87
12
 8
19
_5
44
13
 9
18
10
50
  Na
  Na
  Na
 3.00
                                                Est.(44)
8
13
6
3
30
6.88
11.51
5.49
4.64

7
19
4
2
32
• ER&E Train

   >10y
   l-10y
            Total
 9
10
II
41
        1.53
        2.60
       15.99
 2
 3
44.
49
8
10
22
40
0.21
2.92
14.10

Trace
4
39
43
3
5
26
34
1.34
2.13
16.61

1
1
55
57
• EPA Train

   Probe & Cyclone
   Filter

   Water Impingers
            Total
12
_36

_30
78
(48)
        2.15
       14.61
 3
66
                      69
 5
li

12.
56
15.70
                (44)
79
                              79
 5     Flask Broke  -
_35_      17.90      80
   (40)
 4      <0.1     Trace
44                 80

-------
                                                          TABLE 8
                                        Particulate Inventory  and  Vanadium Balance From
                                         Combustion of  Intermediate  Ash Venezuelan Resid
Particulate Distribution

  Boiler Solids

  Pass 1 (firetube)
  Pass 2
  Pass 3
  Pass 4
              Total

  Stack Particulate

  t ER&E Train
              Total
(Long Residence
Run 70
Total
Particulate
mg/SCM
Vanadium
Analysis
Wt. %
Vanadium
Content
ppm(on fuel)
Time Case)
Run 73
Total
Part.
mg/SCM
V
Anal.
Wt. %
V Cont.
ppm
(on fuel)
Total
Part.
mg/SCM
Run 76
V
Anal.
Wt. %

V Cont.
ppm
(on fuel)
     1-10
  • EPA Train

     Probe & Cyclone
     Filter

     Water Impingers
              Total
14
10
4
3
31
2.53
11.97
8.73
7.90

4
16
4
3
27
 3
 4
2£
36
 2
_42

J3
52
(44)
        5.00
       10.00
       14.9
 2.50
13.11

 0.25
               2
               6
              55
              63
   1
  70

Trace
  71
8
11
7
4
30
8.65
11.71
8.24
9.40

9
17
7
4
37
7
12
7
3
29
11.67
13.08
6.78
6.32

11
20
6
3
40
2
5
28
35
3.27
12.70
16.82

1
8
60
69
2
4
25
31
2.41
8.23
17.58

1
5
56
62
i
M
Ln
1
 4      3.47       2
31     21.71      86
   (35)
_2      1.18      
-------
                                                          TABLE 9

                                         Particulate  Inventory and  Vanadium Balance
                                        From Combustion  of  Low Ash  Light  Arab Resid
                                                 (Short Residence Time Case)
Particulate Distribution
                                           Run 47
Total
Particulate
mg/SCM
                                         Vanadium
                                         Analysis
                                         Wt. %
          Vanadium
          Content
          ppm (on fuel)
                                                  Run 50
                            V Cont.
                             pprt
                             (on fuel)
                            Total
                            Part.
                            mg/SCM
                                                                Run 58
                            V
                            Anal.
                            Wt. %
                           Cont.
                           .ppm
                          (on fuel)
  Boiler Solids

  Pass 1 (firetube)
  Pass 2
  Pass 3
  Pass 4
               Total
Stack Particulate
•  ER&E Train
                Total
    l-10y
   EPA Train
    Probe & Cyclone
    Filter

    Water Impingers
                 Total
  28
  83
  37
  11
 159
  69
 109
  45
 223
  90
 128
 224
        (218)
0.76
1.38
0.29
0.68
Trace
0.11
6.84
0.14
1.24
 3
15
 1
 1
20
 2
34_
36
 2
20
              22
36
77
35
13
161
1.02
0.58
0.53
0.69

5
6
2
1
14
22
71
39
17
149
2.55
1.07
0.75
1.22

7
10
4
2
23
50
123
43
216
0.44
0.46
2.90

3
7
16
26
59
101
38
198
0.59
0.83
5.23

4
11
25
40
 35
 90

 37
162
0.36
0.92
                               (125)
 2
11
                               13
140
 79
                    225
0.69
1.52

0.65
                                                                                                                           Ul
                                                                                                                           OJ
 12
 15

Trace
 27

-------
                                                        TABLE 1,0
                                         Particulate Inventory and Vanadium Balance
                                        From Combustion of  Low Ash Light Arab Resid
Particulate Distribution

  Boiler Solids

  Pass 1 (firetube)
  Pass 2
  Pass 3
  Pass 4
               Total
 Stack Particulate

 •  ER&E Train

     >10y
     l-10y
               Total
 •  EPA Train
     Probe & Cyclone
     Filter

     Water Impingers
               Total
(Short Residence Time Case)

Total
Particulate
mg/SCM


22
60
35
20
137
48
98
54
200
115
93
(208)
13
221
Run 65
Vanadium
Analysis
Wt. %
(1)
Duplicate
1.59 - 1.69
1.07 - 1.00
0.61 - 0.75
0.35 - 0.42

0.67 - 0.61
0.77 - 0.71
2.3

0.67 - 0.57
1.53

0.25 - 0.47


Vanadium
Content
ppm (on fuel)

Average
5
8
3
1
17
4
9
16
29
9
18

Trace
27

Total
Particulate
mg/SCM


36
71
39
16
162
58
100
64
222
109
101
(210)
4
214
Run 78
Vanadium
Analysis
Wt. %


1.67
1.22
1.02
1.10

0.63
0.80
2.08

0.59
15.22

1.41


Vanad ium
Content
ppm (on fuel)


8
11
5
2
26
i
i
4
10
17
31
8
20

<1
28
 (1)
   Separate samples submitted under blind designation.

-------
                                                        TABLE 11
                                         Particulate Inventory and Vanadium Balance
                                         From Combustion  of  Low  Ash  Light Arab Resid
Particulate Distribution

  Boiler Solids

  Pass 1 (firetube)
  Pass 2
  Pass 3
  Pass 4
               Total
Stack Particulate

•  ER&E Train

    >10y
    1-10M
               Total
   EPA Train
    Probe & Cyclone
    Filter

    Water Impingers
               Total
(Long Residence

Total
Particulate
mg/SCM
14
22
8
7
51
3
8
12
23
4
21
(25)
20
45
Run 46
Vanadium
Analysis
Wt. %
2.94
8.81
3.81
3.64

1.56
1.36
2.28

Trace
7.86

1.30


Vanadium
Content
ppm(on fuel)
5
25
4
3
37
<1
1
4
5
_
21

3
34
Time Case)

Total
Part.
mg/SCM
5
18
13
11
47
6
4
Run 59
V
Anal.
Wt. %
4.06
2.54
1.34
1.29

2.21
3.04
23 10.67
33
6
22
(28)
11
39

3.54
9.14
Run 60
V Cont.
ppm
(on fuel)
3
6
2
2
13
2
1
31
34
3
25
Total
Part.
mg/SCM
12
19
12
6
49
2
4
19
25
4
29
V
Anal.
Wt. %
2.70
1.91
1.07
2.62

1.95
3.40
9.35

2.35
6.00
V Cont.
ppm
(on fuel)
4
5
2
2
13
< 1
2
23
25
1
22
(33)
0.76

1
29
23
56
0.10

Trace
23
Ul
Ul

-------
                                                       TABLE 12
                                       Partlculate  Inventory  and  Vanadium Balance
                                      From Combustion of  Low  Ash  Light Arab Resid
Particulate Distribution

  Boiler Solids

  Pass 1 (firetube)
  Pass. 2
  Pass 3
  Pass 4
              Total
 Stack Particulate

 •  ER&E Train
      l-10y
              Total
 •  EPA Train

     Probe & Cyclone
     Filter

     Water Impingers
              Total


Total
Particulate
mg/SCM
13
20
15
11
59
3
4
23
30
3
26
(29)
3
32
(Long
Run 66
Vanadium
Analysis
Wt. %
1.82
1.82
0.78
0.85

0.43
2.81
7.6

1.82
6.27

0.13

Residence Time Case)
-
Vanadium
Content
ppm (on fuel)
3
5
1
1
10
<1
1
.22
23
1
21

Trace
22

Total
Particulate
mg/SCM
3
24
12
7
46
3
7
16
26
6
26
(32)
2
34
Run 77
Vanadium
Analysis
Wt. %
5.60
2.17
1.88
3.29

1.76
2.76
12.26

1.88
7.10

0.49


Vanadium
Content
ppm (on fuel)
2
7
3
2
14
1
2
25
28
1
24

Trace
25
Ln

I

-------
                                - 157 -







                               APPENDIX 6




                           CONVERSION FACTORS




                        ENGLISH TO METRIC UNITS
To Convert From                     To                 Multiply By



    OF                              »c                 (»F _32) 5/9




   "H20                            mm Hg                 1.8682




    ft3                            m3                    0.02832




    Ibs/min                        Kg/min                0.45359




    Inches                          mm                  25.4




    Btu/lbs                       Kcal/Kg                0.555




    Feet                          Meters                 0.30480

-------
                                              - 158  -
                                   TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing)
 l. REPORT NO.
   EPA 600/2-76-096
                                                          3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
  COLLECTION EFFICIENCIES  OF STACK SAMPLING SYSTEMS FOR
  VANADIUM EMISSIONS  IN FLUE GASES
             5. REPORT DATE
                April,  1976
                                                          6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
                                                          8. PERFORMING ORGANIZATION REPORT NO.
  H. Lawrence Goldstein  and  C.  W.  Siegmund
9. PERFORMING ORGANIZATION NAME AND ADDRESS
  Exxon Research and  Engineering Company
  Products  Research Division
  P. 0. Box 51
  Linden, N. J.  07036
             10. PROGRAM ELEMENT NO.

                1AA010
             11. CONTRACT/GRANT NO.
                68-02-1748
 12. SPONSORING AGENCY NAME AND ADDRESS
   Environmental  Sciences Research Laboratory
   Office of Research  and Development
   U. S. Environmental  Protection Agency
   Research Triangle Park,  North Carolina 27711
             13. TYPE OF REPORT AND PERIOD COVERED
               Final    7/74 - 6/75
             14. SPONSORING AGENCY CODE
                EPA-ORD
 15. SUPPLEMENTARY NOTES
 16. ABSTRACT
 SUMMARY - An experimental  program has been conducted to measure  and  compare the effi-
 ciency of two stack sampling  systems to collect vanadium-bearing particulate emissions
 in flue gas.  One sampling system was EPA's Method 5, the other  was  developed by
 Exxon Research and Engineering Company for specialized in-house  studies.

    To evaluate collection  efficiency, an extensive factorial  study was  carried out in
 a 50 hp four-pass firetube boiler burning typical residual  fuel  oils.  In  each test the
 sampling systems were operated simultaneously in the stack  to collect  the vanadium-
 bearing particulate emissions.   Three residual fuel oils were tested:  two Venezuelan
 (359 and 149 ppm V) and one Arabian (39 ppm V).  A vanadium balance  was established
 for each experiment by inventorying both the particulate emissions and  the particulate
 remaining in the boiler.   Test variables, in addition to the  sampling  systems and fuel
 oils, also included two combustion residence times and two  sampling  probe locations.

    The results of the study show that vanadium collection efficiency depends on two
 variables: combustion residence time and sampling system.   For both  systems efficiency
 decreases as combustion residence time increases, which results  in a particulate size
 distribution shift to the  submicron range.  Where particulate emissions are in the
 coarse size range, collection efficiency in both sampling systems is almost quantita-
 tive. The oxidation states of vanadium in fuel oil emissions are briefly  discussed.	
17.
                               KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                             b.IDENTIFIERS/OPEN ENDED TERMS
                                                                        c. COSATI Field/Group
  Air Pollution
  Particles
 *Vanadium
 *Residual Oil
  Flue gases
 *Collecting methods
 *Efficiency
  Chemical analysis
                               13B
                               07B
                               21D
                               21B
                               14B
                               07D
18. DISTRIBUTION STATEMENT

     Released to Public
19. SECURITY CLASS (This Report)
21. NO. OF PAGES

   167
                                             20. SECURITY CLASS (This page)
                                                Unclassified
                                                                        22. PRICE
EPA Form 2220-1 (9-73)

-------
                                                          INSTRUCTIONS

    1.   REPORT NUMBER
        Insert the EPA report number as it appears on the cover of the publication.

    2.   LEAVE BLANK

    3.   RECIPIENTS ACCESSION NUMBER
        Reserved for use by each report recipient.

    4.   TITLE AND SUBTITLE
        Title should indicate clearly and briefly the subject coverage of the report, and be displayed prominently. Set subtitle, if used, in smaller
        type or otherwise subordinate it to main title. When a report is prepared in  more than one volume, repeat the primary title, add volume
        number and include subtitle for the specific title.

    5.   REPORT DATE
        Each report shall carry a date indicating at least month  and year.  Indicate the basis on which it was selected /e.g., date of issue, date of
        approval, date of preparation, etc.).

    6.   PERFORMING ORGANIZATION CODE
        Leave blank.

    7.   AUTHOR(S)
        Give name(s) in conventional order (John R. Doe, J. Robert Doe,  etc.). List author's affiliation if it differs from the performing organi-
        zation.

    8.   PERFORMING ORGANIZATION REPORT NUMBER
        Insert if performing organization wishes to assign this number.

    9.   PERFORMING ORGANIZATION NAME AND ADDRESS
        Give name, street, city, state, and ZIP code. List no more than two levels of an organizational hirearchy.

    10.  PROGRAM ELEMENT NUMBER
        Use the program element number under which the report was prepared. Subordinate numbers may be included in parentheses.

    11.  CONTRACT/GRANT NUMBER
        Insert contract or grant number under which report was prepared.

    12.  SPONSORING AGENCY NAME AND ADDRESS
        Include ZIP code.

    13.  TYPE OF REPORT AND PERIOD COVERED
        Indicate interim final, etc., and if applicable, dates covered.

    14.  SPONSORING AGENCY CODE
        Leave blank.

    15.  SUPPLEMENTARY NOTES
        Enter information not included elsewhere but useful, such as:  Prepared in cooperation with, Translation of, Presented at conference of,
        To be published in. Supersedes, Supplements, etc.

    16.  ABSTRACT
        Include a brief (200 words or less) factual summary of the most significant information contained in the report. If the report contains a
        significant  bibliography or literature survey, mention it here.

    17.  KEY WORDS AND DOCUMENT ANALYSIS
        (a) DESCRIPTORS - Select from the Thesaurus of Engineering and Scientific Terms the proper authorized terms that identify the major
        concept of  the research and are sufficiently specific and precise to be used as index entries for cataloging.

        (b) IDENTIFIERS AND OPEN-ENDED TERMS - Use identifiers for project names, code names, equipment designators, etc. Use open-
        ended terms written in descriptor form for those subjects for which no descriptor exists.

        (c) COSATI FIELD GROUP - Field and group assignments are to  be taken from  the 1965 COSATI Subject Category List. Since the ma-
        jority of documents are multidisciplinary in nature, the Primary Field/Group assignment(s) will be specific discipline, area of human
        endeavor, or type of physical object. The application(s) will be cross-referenced  with secondary Field/Group assignments that will follow
        the primary posting(s).

    18.  DISTRIBUTION STATEMENT
        Denote releasability to the public or limitation for reasons other than security for example "Release Unlimited." Cite any availability to
        the public,  with address and price.

    19. &20.  SECURITY CLASSIFICATION
        DO NOT submit classified reports to the National Technical Information service.

    21.  NUMBER OF PAGES
        Insert the total number of pages, including this one and unnumbered pages, but exclude distribution list, if any.

    22.  PRICE
        Insert the price set by the National Technical Information Service or the Government Printing Office, if known.
EPA Form 2220-1  (9-73) (Reverse)

-------