vvEPA


Combustion
Research on the Fate
of Fuel-nitrogen
Under Conditions
of Pulverized Coal
Combustion

Interagency
Energy/Environment
R&D Program Report

-------
                 RESEARCH REPORTING SERIES


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

    1. Environmental Health Effects Research

    2. Environmental Protection Technology

    3. Ecological Research

    4. Environmental Monitoring

    5. Socioeconomic Environmental Studies

    6. Scientific and Technical Assessment Reports (STAR)

    7.  Interagency Energy-Environment Research and Development

    8.  "Special" Reports

    9.  Miscellaneous Reports

 This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
 RESEARCH AND DEVELOPMENT series. Reports in this series result from the
 effort funded under the 17-agency Federal Energy/Environment Research  and
 Development Program.  These studies relate to EPA's mission to protect the public
 health and welfare from adverse effects of pollutants associated with energy sys-
 tems. The goal of the  Program is to assure the rapid development of domestic
 energy supplies in an environmentally-compatible manner by providing the nec-
 essary  environmental data and control technology. Investigations include analy-
 ses of the transport of energy-related pollutants and their health and ecological
 effects;  assessments of, and development of, control  technologies for energy
 systems; and integrated assessments of a wide-range of energy-related environ-
 mental issues.
                         EPA REVIEW NOTICE
 This report has been reviewed by the participating Federal Agencies, and approved
 for publication. Approval does not signify that the contents necessarily reflect
 the views and policies of the Government, nor does mention of trade names or
 commercial products constitute endorsement or recommendation for use.

 This document is available to the public through the National Technical Informa-
 tion Service. Springfield. Virginia 22161.

-------
                                    EPA-600/7-78-165

                                          August 1978
Combustion  Research on the Fate
of Fuel-nitrogen  under Conditions
  of  Pulverized Coal Combustion
                         by

              J.M. Levy, J.H. Pohl, A.F. Sarofim, and Y.H. Song

                Massachusetts Institute of Technology
                 Department of Chemical Engineering
                  Cambridge, Massachusetts 02139
                     Grant No. R803242
                       Task No. 2
                  Program Element No. EHE624A
                 EPA Project Officer: John H. Wasser

               Industrial Environmental Research Laboratory
                Office of Energy, Minerals, and Industry
                 Research Triangle Park, NC 27711
                       Prepared for

               U.S. ENVIRONMENTAL PROTECTION AGENCY
                 Office of Research and Development
                    Washington, DC 20460

-------
                            ABSTRACT



     An experimental investigation of coal pyrolysis and oxidation



and char oxidation was performed with the objective of determining



the effects of temperature and fuel/oxygen equivalence ratio on



the conversion of coal-nitrogen to NO .  Experiments were carried
                                     a


out in a laboratory laminar flow furnace under conditions repre-



sentative of pulverized coal combustors, i.e., heating rates of



10  - 10 K/sec., temperatures of 1000 - 2100K, and residence times



of 2-1000 milliseconds.



     Data on the devolatilization of coal were obtained at tempera-



tures up to 2100K, both as a function of time and under conditions



of asymptotic yield.  The devolatilization experiments showed that



no nitrogen loss occurred until 10 - 15% of the coal had been



devolatilized, consistent with the hypothesis that coal-nitrogen



is contained mostly in heterocyclic rings.  Nitrogen could be



completely removed from the char by prolonged heating at temperatures



above^1750K, implying that nitrogen does not form condensed stable



structures at pulverized flame temperatures as does carbon.  After



initial fracture, loss of nitrogen and total weight loss are linearly



correlated with a nitrogen to carbon slope of 1.25 to 1.5.  The



pseudo-first order rate constant for initial nitrogen loss was



found to be:
     k = 9.3xl03 exp | -22,700/RT I
I -22,
     Data on the conversion of coal-nitrogen to NO , under oxidative
                                                  J^



conditions, were obtained at temperatures up to 1750K, for






                             ill

-------
residence times up to one second.  The conversion of coal-nitrogen



to NO  decreased monotonically with increasing fuel/oxygen
     Ji


equivalence ratio, and decreased slightly with increasing tempera-



ture.  Oxidation experiments were also carried out on char in



order to separate the individual contributions to NO  emissions
                                                    J^


of volatile and char-bound nitrogen.  The char used was prepared



by the pyrolysis of coal at temperatures and residence times



corresponding to those of the oxidation experiments.  The



conversion to NO  of char-nitrogen was lower than the corresponding
                Jv


value for coal-nitrogen, while following trends similar to that



of coal-nitrogen.  It was found that volatile nitrogen compounds



accounted for the major fraction of NO  produced from coal-nitrogen,



especially at high temperatures and low fuel/oxygen equivalence



ratios.  The results suggest that low NO  emissions from pulverized
                                        X


coal combustors are favored by a two stage design,  the  first



stage operated fuel-rich at high temperature,  the second stage



operated fuel-lean at low temperature.
                               iv

-------
                            CONTENTS
Figures	viii
Tables 	  x

     I.  Introduction	   1
    II. .  Background	   5
             II-l  Fuel NO and Thermal NO	   6
             II-2  Char NO and Volatile NO	   8
             II-3  Reduction of NO by Char	11
   III.  Equipment and Background	13
            III-O     Flow Furnace	14
            III-l.O   Pyrolysis Experiment 	  15
            III-l.l.O Laminar Flow Furnace 	  16
            III-l.l.l Coal Feeder	19
            III-1.1.2 Particle Collector 	  21
            III-1.1.3 Gas Preheater	23
            IIX-1.1.4 Suction System 	  24
            III-1.1.5 Experimental Procedure 	  25
            III-1.2.0 Free Fall Experiment with
                      Bronze Collector 	  27
            III-1.2.1 Experimental Procedure 	  30
            III-1.2.2 Free Fall Experiment with
                      Alumina Collector	31
            III-1.2.3 Experimental Procedure 	  32
            III-1.3.0 Crucible Experiment	33
            III-2.0   Oxidation Experiments	34
            III-2.1.0 Free Fall Experiment	35
            III-2.1.1 Experimental Procedure 	  36
            III-2.2.0 Laminar Flow Experiment	37
            III-2.2.1 Particle Collector 	  39
            III-2.2.2 Gas Preheater	41
            III-2.2.3 Vacuum System	42
            III-2.2.4 Preparation of Char	43
            III-2.2.5 Experimental Procedure 	  44
                               v

-------
       III-3     Procedures  for Measuring
                 Nitric Oxide  Reduction by Char	46
       III-4.0    Gas  Phase Analyses	47
       III-4.1    Chemiluminescent Gas  Analysis
                 of NO	48
                      Ji
       III-4.2    Colorimetric  Analysis
                 of NH3	49
       III-4.3    Standard Specific Ion Electrode
                 Analysis of HCN	49
       III-4.4    Gas  Chromatbgraphic Analysis
                 of N2/ CO,  C02, and CH4	50
       III-4.5    Non-Dispersive Infrared Analysis
                 of CO and O>2	51
       III-5     ASTM Proximate Equipment	52
IV. Coal Characterization	54
        IV-1.0    Montana Lignite 	  56
        IV-1.1    Visual Appearance of  Montana
                 Lignite	57
        IV-1.2    Particle Size Distribution	61
        IV-1.3    Surface Area	64
        IV-1.4    Proximate  Analysis	66
        IV-1.5    Heat of Combustion	68
        IV-1.6    Elemental Analysis	69
        IV-1.7   Ash Analysis	71
        IV-1.8   Functional Group Analysis 	  73
        IV-2.0   Pittsburgh Seam #8 Bituminous
                 Coal	75
        IV-2.1   Visual Appearance of Pittsburgh
                 Bituminous Coal	76
        IV-2.2   Particle Size  Distribution	80
        IV-2. 3   Surface Area	83
        IV-2.4   Proximate Analysis	84
        IV-2.5   Heat of Combustion	85
        IV-2.6   Elemental Analysis	86
        IV-2.7   Ash  Analysis	88
        IV-2.8   Functional Group Analysis 	  89
        IV-3.0   Physical Properties of Coal  	  92
        IV-4.0   Characterization of Other
                 Coals and Chars	                94
                        vi                     	

-------
 V.  Coal Results	96
        V-1.0   The Coal Flame	96
        V-2.0   Devolatilization Results	100
        V-2.1   Devolatilization Results from
                Crucible Experiments	100
        V-2.1.1 Pittsburgh Seam #8 hvA-Bituminous
                Coal	100
        V-2.1.2 Montana Lignite-A	102
        V-2.2   Devolatilization of Dispersed
                Coal Particles	105
        V-2.2.1 Pittsburgh Seam #8 Bituminous Coal	105
        V-2.2.2 Montana Lignite	HO
        V-2.2.3 Nitrogen Pyrolysis from Lignite
                and Bituminous Coal	113
        V-2.3   Analysis of Volatile Products
                of Coal Pyrolysis	117
        V-3.0   Oxidation Results 	 123
        V-3.1   Oxidation of Lignite and Bituminous
                Coal-Fixed Residence Time 	 123
        V-3.2   Oxidation of Lignite Char-Fixed
                Residence Time	129
        V-3.3   Oxidation of Sub-Bituminous Coal Fixed
                Residence Time- Effect of Temperature . . . 134
        V-3.4   Time-Resolved Oxidation of Char	136
VI. Discussion of Experimental Results	140
       VI-1.0   Devolatilization Results	140
       VI-1.1   Asymtotic Crucible Results	142
       VI-1.2   Time-Resolved Flow Furnace Results	145
       VI-1.3   Primary Nitrogeneous Pyrolysis
                Products	153
       VI-2.0   Oxidative Results 	 155
       VI-2.1   Coal Oxidation	158
       VI-2.2   Char Oxidation	160
       VI-3     The Contributions of Volatile and
                Char Nitrogen to Nitric Oxide	164
       VI-4     Conclusions and Implications for
                Control of NO  Emissions	169
                             J\
References      	172
                         vii

-------
                           FIGURES
Number                                                    Pa9

III-l  Laminar Flow Furnace System 	
III-2  Coal Feeder	20
III-3  Water Cooled Collector  	  22
III-4  Free Fall Experiment with Bronze Collector  .  .  .  .  28
III-5  Free Fall Experiment with Alumina Collector ....  38
       and Crucible Experiments
 IV-1  Scanning Electron Micrograph of Raw Montana ....  58
       Lignite-A
 IV-2  Photomicrograph of Polished Raw Lignite-A   ....  59
 IV-3  Rosin-Rammler Accumulative Size Distribution.  ...  62
       of Lignite Particles
 IV-4  Lignite Particle Size Frequency Distribution  ...  63
 IV-5  Scanning Electron Micrograph of Raw
       Pittsburgh Seam #8 hvA Bituminous Coal	''
 IV-6  Photomicrograph of Polished Raw Pittsburgh  ....  73
       Seam 18 hvA Bituminous Coal
 IV-7  Rosin-Rammler Accumulative Size Distribution  ...  81
       of Bituminous Coal Particles
 IV-8  Bituminous Particle Size Frequency Distribution .  .  82
  V-l  Schematic of a Single Coal Particle Burning ....  91
       in a Pulverized Coal Flame
  V-2  Crucible Devolatilization of Pittsburgh Seam  . .   101
       18 hv A-Bituminous Coal
  V-3  Crucible Devolatilization of Montana lignite-A. .   103
  V-4  Element Loss with Devolatilization of a
       Pittsburgh Seam 18 hvA-bituminous Coal	
  V-5  Element Loss with Devolatilization of a	
       Montana Lignite-A
  V-6  Retention of Nitrogen in Devolatilized Lignite. .   114
       and Bituminous Coal Chars
  V-7  Percentage of Original Coal- Carbon 	
  V-8  Percentage of Original Coal- Nitrogen 	
  V-9  Conversion of Coal Nitrogen as a Function of
       Fuel Equivalence Ratio in a 1500K Furnace  * * *  124
  V-10 Fate of Fuel Nitrogen During Oxidation: Conversion
       to Nitric Oxide, Retention by Unburned Char, and
       Combustion Efficiency Defined by Solid Weioht Loss
       Montana Lignite at 1750K  ....
                             rtii           	

-------
 V-ll Fate of Fuel Nitrogen During Oxidation:  Conversion.  .  f  m  128
      to Nitric Oxide, Retention by Unburned Char, and
      Combustion Efficiency Defined by Solid Weight Loss.
      Montana Lignite at 1250K.

 V-12 Conversion of Nitrogen vs. Fuel Equivalence Ratio ....  130
      in a 1500K Furnace for a Lignite Char Previously
      Devolatilized at 1500K for 1 second.

 V-13 Fate of Char Nitrogen During Oxidation:  Conversion.  .  .  .  132
      to Nitric Oxide, Retention by Unburned Char, and
      Combustion Efficiency.  Char from Montana Lignite
      Pyrolyzed and Oxidized at 1750K.

 V-14 Fate of Char Nitrogen During Oxidation:  Conversion.  ...  133
      to Nitric Oxide, Retention by Unburned Char, and
      Combustion Efficiency.  Char from Montana Lignite
      Pyrolyzed and Oxidized at 1250K.

 V-15 Effect of Temperature on Conversion of Coal	135
      Nitrogen to Nitric Oxide.  Montana Sub-bituminous
      Coal.

 V-16 Weight Loss of Char Prepared from a Montana Lignite  .  .  .  137
      as a Function of Distance to the Collector During
      Oxidation.

 V-17 Nitrogen/Carbon Ratio as Percentage of that in the.  .  .  .  138
      Original Char as a Function of Distance to the
      Collector During Oxidation.

VI-1  Coal Nitrogen Retention in the Char under Conditions.  .  .  143
      of Asymptotic Weight Loss.

VI-2  Comparison of Rate of Nitrogen Removal from Coal	147
      with Pyrolysis of Model Nitrogen Compounds and
      Fundamental Rate Constants.

VI-3  Correlation between Nitrogen Loss and Total Weight.  .  .  .  150
      Loss during Pyrolysis of Coal.

VI-4  Contribution of Volatiles to NOX Emissions: Percent  .  .  .  166
      of Total NOX Contributed by Volatiles, and Percent of
      Conversion of Volatiles to NO, .
                              ix

-------
                          TABLES
Number                                                     Page
IV-1   Coal Characterizations	67
IV-2   Characterization of Montana Lignite from the. ...  70
       Savage Mine in Richland County, Ground and
       Classified to 38-45y.
IV-3   Ash Characterization	72
IV-4   Functional Group Analysis of a Lignite	74
IV-5   Characterization of Pittsburgh Seam 18               87
       Bituminous Coal From the Ireland Mine Ground*
       and Classified to 38-45y.
IV-6   Functional Group Analysis of a High Volatile....  90
       Bituminous Coal.
IV-7   Characterization of Coals and Chars	95
 V-l   Carbon Distribution During Combustion of  	 120
       Lignite at T=1750 K and 4=3*4.
 V-2   Nitrogen Distribution during Combustion of	121
       Lignite at T=1750 K and 4

-------
                           SECTION I



                         INTRODUCTION





     Coal constitutes one of the major United States fossil fuel



reserves.  Exploitation of these reserves is, however, complicated



by a number of problems, including their high nitrogen content



which can contribute significantly to the emissions of nitrogen



oxides (NO ) .
          X


     United States coals have nitrogen contents that average 1.4%



nitrogen by weight.  Typically, 20 to 60 percent of the nitrogen



in a fuel is converted to nitrogen oxides in practical combustion



devices.  Higher conversions are obtained when the nitrogen content



of the fuel is low or when the combustor is operated lean.  At



current levels of coal consumption of over 500 x 10  tons/year,



a conversion efficiency of fuel-nitrogen to NO  of 20% (assuming
                                              X


an average of 1.4% fuel-nitrogen in coal) would yield emissions



exceeding 4.6 x 10  tons/year of NO  (evaluated as NO0).   This is
                                   X                 ^


significant when compared with current estimates (Department of



Transportation, 1976) of 10.3 x 10  tons/year emitted from station-



ary sources and 11.6 x 10  tons/year emitted from mobile sources.



     The anticipated increase in utilization of fuels derived from



coal provides the motivation for developing combustion processes



that will yield low NO  levels when fired with fuels having a high
                      J


nitrogen content.  This report presents the results of a study



conducted at M.I.T. aimed at elucidating the mechanism of

-------
fuel-nitrogen conversion to NO  in pulverized coal flames with the
                              J^


specific goal of developing a strategy for minimization of NOx



emissions through careful control and optimization of combustion



conditions.



     Most coal burned for power generation is done so in large unit



pulverized coal flames (hundreds of megawatts of electrical capacity).



In such combustors, coal, ground to a particle size of less than



200 microns, is injected through a nozzle into a region of preheated



air where the particles are rapidly heated by conduction and radiation



in about 1 millisecond to temperatures of 2000K or higher.  During



the next tens of milliseconds, volatile matter is evolved which



undergoes homogeneous gas phase oxidation in a mode intermediate to



premixed and diffusion flames.  Burning of the volatile matter is



usually complete in about 10-100 milliseconds.  Simultaneously,



the devolatilized or partially devolatilized char undergoes



heterogeneous oxidation.  Char burnout is usually complete in about



300 milliseconds.



     Coal combustion can thus be conceptualized as rapid heating



causing devolatilization, followed by two distinctly different



(and, to first approximation, independent) elementary oxidation



mechanisms.  Clearly, emission of any combustion product, in partic-



ular, NOx, is affected by the relative rates and efficiencies of all



three aforementioned physical processes, each of which, in turn, is



differently affected by alteration of any of the parameters acces-



sible to the combustion engineer, e.g. heating rate, temperature,



fuel/air equivalence ratio,  degree of mixedness, residence times, etc.

-------
It is certainly not clear .-__ priori, what combinations of conditions



will lead to minimization of product emissions, especially trace



pollutants whose fundamental formation paths are not, in any case,



well characterized.  It is obvious, therefore, that systematic



development of an optimum combustion modification strategy designed



to minimize NO  emissions without otherwise adversely affecting
              Ji


combustor operation awaits an understanding of the coupled physical



and chemical mechanisms by which fuel-nitrogen is converted to NO .
                                                                 vt


     Ideally, a full mechanistic exposition of the problem would



entail a microscopic or molecular description of the pathway by



which fuel-nitrogen is converted to NO .  This would, in principle,
                                      Ji


start with the thermal or catalytic rupture of specific bonds in



the carbonaceous coal matrix leading to well-characterized pyrolysis



products, followed by parallel homogeneous and heterogeneous oxida-



tive routes described in the former case by numerous coupled,



elementary reactions and in the latter case by consideration of



surface and pore transport effects as well.  Unfortunately, develop-



ment of such a fundamental mechanism of fuel-nitrogen conversion



is yet rather distant, and, given, among other problems, the current



dearth of necessary rate data, only marginally tractable at present.



     These considerations, however, do not preclude the possibility



of characterizing the various steps of fuel-nitrogen conversion in



a somewhat more phenomenological vein, while retaining a quantitative



sense of "mechanism" adequate to the needs of the fuels engineer.



That is, the routes by which fuel-nitrogen is converted to either



NO  or N, (the desired end product) may still be mechanistically
  X     A

-------
traced, and, at each juncture, the influence of systematic variation



of combustion parameters may be quantitatively examined.  This more



empirical mechanism of coal-nitrogen conversion may be described as



follows:
                               ni  ^NO
  Coal-N
                    Volatile!
                                                XM  ^ NO.
              *   'X.          v
                    Char	i-Burned Char
                                   Unburned Char
     This study  focusses on the conversion of fuel-nitrogen in



pulverized  coal  flames  for two representative United States coals ,



a Pittsburgh Seam bituminous and a Montana lignite.  Each essential



step in the empirical mechanism has been quantitatively investigated



with respect to  variation of combustion parameters and /or reaction



rate.   The  results  are  applied qualitatively to the development of



a staged combustion process designed  to minimize NO  emissions.
                                                   Jm


Quantitative results, essential to the design of such a combustor,



are presented.

-------
                          SECTION  II



                          BACKGROUND





     The formation of nitrogen oxides during coal combustion is a



complicated process involving contributions both from the fixation



of atmospheric nitrogen  (thermal NO ) as well as from the oxidation
                                   Ji


of nitrogen bound chemically in the fuel  (fuel NO ).  It is,
                                                 J\


primarily, to the latter of these two processes that this study



is directed.  For an excellent review of the literature through



mid-1976, the reader is further referred to Pershing (1976).



     It is recognized that combustion of pulverized coal particles



occurs in three main stages: evolution of the volatile components



or devolatilization, followed by homogeneous volatile combustion,



and, finally, heterogeneous combustion of the residual char.



Although the stages may overlap to some degree (Howard and Essenhigh



1967a, 1967b), it is clear that char combustion is generally of



much longer duration then volatile combustion.  Based on this model,



the conversion of coal-nitrogen to NO  can,therefore, also be
                                     J^


regarded as a three-stage process: coal-nitrogen devolatilization,



volatile-nitrogen combustion, and char-nitrogen combustion.  It is



evident that the extent of devolatilization prior to ignition and



the distribution of nitrogen between char and volatiles will



strongly influence NO  emissions.  Finally, since the destruction
                     J^


of NO  is as important as the formation of NO  in determining the
     X                                       X


net NO  emission from a coal combustor, a review on the reduction
      Ji


of NO  by char is also included.

-------
II-l  FUEL NO AND THERMAL NO
     An early investigation of the production of NO during fluidized
bed combustion of coal was reported by Jonke, e_t al^ (1970) .  In
their experiment, after steady-state combustion had been reached
with a synthetic air mixture of nitrogen and oxygen, argon was
substituted for the nitrogen.  It was found that the NO level of
the flue gas was not materially affected by the substitution of
argon for the nitrogen.  This indicated that the major source of
nitrogen for the production of NO was the nitrogen bound chemically
in the coal.  This conclusion, while establishing the significance
of fuel-NO, did not, however, rule out the possibility of significant
contributions from thermal NO under the higher temperature conditions
of pulverized coal flames.   (Thermal NO, known to be formed via the
highly temperature sensitive Zeldovich mechanism (Zeldovich 1946) ,
is not expected to contribute significantly under the low tempera-
ture conditions of a fluidized bed combustor.)
     Pershing (1976) and Pershing and Wendt  (1976)  isolated fuel
NO by using, instead of air, a synthetic oxidant mixture of the
same heat capacity containing 21% O2, 18% CO2,-and 61% Ar.  Com-
parison between total NO with preheated air as the oxidant and
fuel NO with synthetic mixture as the oxidant was accomplished
under nearly identical combustion conditions.  Prom the studies
of four different coals and one coal char, they found that,
typically, fuel NO contributed over 80% of the total NO emissions.
In addition, they also concluded that fuel NO formation is rela-
tively insensitive to flame temperature over a wide range of
practical interest.

-------
     The reaction mechanism for the formation of thermal NO from



the fixation of atmospheric nitrogen was first proposed by



Zeldovitch (1946).  This mechanism has since been confirmed



and modified by a number of investigators.  The Zeldovitch



mechanism is highly temperature dependent so that in typical



laboratory scale pulverized coal flames, little thermal NO is



expected due to the comparatively low flame temperatures of small-



scale combustors with cold walls.  This explains why little thermal



NO has been observed and why fuel NO contributes most of the total



NO emission in small-scale coal-fired systems.

-------
                                                                  8
II-2  CHAR NO AND VOLATILE NO
     Until now, very little quantitative information has been
available on the relative contributions to fuel NOx emissions of
the nitrogen evolved from the coal as volatiles (and oxidized
homogeneously )   and the nitrogen retained in the char   (and
oxidized heterogeneously).  In addition, no information is avail-
able in the literature on the pyrolysis of nitrogen from coal at
pulverized coal flame conditions, although some data are available
on the distribution of the nitrogen in coal among pyrolysis pro-
ducts at coking conditions.  The coking data, however, are not
applicable to pulverized coal flames because of differences in
heating rates, final temperature, residence times, and particle
interactions.
     Based on these fragmentary data, Sternling and Wendt  (1972)
drew up a scheme for the distribution of coal nitrogen between
volatiles and char during pyrolysis and oxidation.  They indicated
that a significant portion of the nitrogen in the coal would not
be volatilized but would remain in the char, so the oxidation of
nitrogen in the char might contribute significantly to NO emissions
     Coal char combustion is a very complicated process.  Since
the char structure is porous, both internal and external combustion
occur.   (Anson, Moles, and Street, 1971).  In addition, pore
diffusion may sometimes play an important role in the combustion
process  (Mandel 1977).
     The physical and chemical characteristics that influence the
conversion of fuel nitrogen to NOx during coal char combustion
were further examined theoretically by Wendt and Schulze  (1976) .

-------
They showed that diffusion-reaction interactions are important in



determining the selectivity of the char particle toward nitric



oxide production.  Their model predicts that char combustion is a



potentially high source of NO.



     The work by Blair et "al.(1976) has also shown that a large



fraction of the nitrogen may remain in the char after devolatiliza-



tion.  Studies of the behavior of fuel-nitrogen under simulated



combustion conditions by Pohl (1976) , Pohl and Sarofim (1976), and



Song, Be4r, and Sarofim (1977) also support this picture for tem-



peratures below 1750K, but the trend of the data suggests that



much of the fuel nitrogen may be devolatilized at higher tempera-



tures .



     In a study utilizing a fluidized bed combustor, Pereira and



Beer (1975) and Pereira (1975) have assessed the influence of



temperature on the relative contributions of char-nitrogen and



volatile nitrogen to the total NO emissions.  They showed that the



char contribution to the total NO emissions predominated at low



temperatures  (800C and less), whereas at higher temperatures the



contribution of the volatile fraction became more important.



     It is useful to indicate the similarity between the combustion



of coal-volatiles and ordinary gaseous flames.  The volatile com-



bustio.n process may be generally regarded as proceeding in two



stages: first, the volatiles must mix with oxygen, and secondly,



chemical reaction must occur.  It is apparent that the combustion



of coal volatiles is analogous to ordinary gaseous flames.  It is



easiest to investigate kinetic mechanisms in systems that have

-------
                                                                 10
uniform mixing, consequently much laboratory work has been done
on the conversion of simple nitrogen compounds in premixed, gaseous
hydrocarbon flames.
     The earliest relevant flame study, performed by Shaw and Thomas
(1965), is only historically significant and the data must be
reinterpreted to include in the fuel equivalence ratio the amount
of oxygen necessary to burn the nitrogen additive.  Subsequent
studies by Fenimore (1972), De Soete (1973), and Sarofim, Williams,
Modellf and Slater (1975) have, however, resulted in the following
generalizations:
     1.  The conversion of fuelrnitrogen to NO  is close to
                                              J\
         quantitative under fuel lean conditions, but decreases
         dramatically for fuel rich conditions,
     2.  An increase in fuel nitrogen concentration results
         in an increase in total NOX emissions  but a
         decrease in the conversion efficiency to NOV, and
                                                    Ji
     3.  Conversion efficiency increases slightly with
         increasing flame temperature.
     These conclusions, derived from studies on idealized labora-
tory systems, are also supported by the results obtained on
practical industrial systems  (Martin and Berkau, 1972; Turner et a^,
1972; and Bartok ejb al., 1972) .

-------
                                                                 11
II-3  REDUCTION OF NO BY CHAR



     In an experimental effort to investigate the NOV emissions
                                                    J^


from a fluidized bed coal combustor, Gibbs, Pereira, and Beer



(1976),  observed a reaction between NO and char.  Although char



was known to react readily with NO in the absence of oxygen



(Edwards, 1972), an NOr-char reaction had not, previously, been



predicted under oxidizing conditions.  An investigation was then



conducted to study the reaction between NO and char in a fluidized



bed operating over a temperature range of approximately 600C to



900C.  They reported that the reaction had a relatively low



activation energy ( ^16 Kcal/mole) and a reaction order of one



with respect to NO.  In addition, they also found that the nitric



oxide reduction was strongly dependent on the char particle size,



with increasing conversion as particle size decreased.



     These conclusions are partially supported by the recent work



done in Japan  (Furusawa, Kunii, Oguma, and Yamada, 1977).  The



Japanese researchers have conducted a kinetic study of NOX reduction



in a fixed bed reactor with a dilute mixture of carbonaceous



materials (i.e., char or activated carbon) over a wide range of



temperatures (750-1250K).  They also found that the reduction of



NO  was first order with respect to NOV concentration.  Interestingly
  x                                   *


though,   they observed three temperature ranges with different



activation energies: 17.3 Kcal/mole for 773K to 993K;



57.1 Kcal/mole for 993K to 1056K; and 39.7 Kcal/mole for 1056K



to 1183K.



     Apparently, the char/NOx reaction is analogous to the char/O2

-------
                                                                 12
reaction, and in fact, NO may compete with O2 for the char



gasification reaction.  It is important to consider the NO/char



reaction in determining net NO  emissions from a coal combustor.
                              J\


A better understanding on the NO/char reaction will, of course,



provide information pertinent to the reduction of NOX emission



by the modification of the combustion processes.

-------
                                                                  13
                         SECTION III



                    EQUIPMENT AND PROCEDURE





     The equipment to be described has been developed for use in



several coal research projects in M.I.T.'s Fuels Research Labora-



tory.  The equipment was designed to cover experimental conditions



of interest for pulverized coal flames and can simulate the



extremely rapid heating expected in the pulverized coal flame;



for example, data has been taken on coal devolatilization, in a



dispersed phase, with time resolution of about 0.5 milliseconds,



heating rate of about 2 x 10 K/sec, and a maximum temperature



of 2100K.  Experimental conditions of gasifiers are of secondary



interest and have not been matched nearly so well.

-------
                                                                  14
III-Q    FLOW FURNACE
     Flow reactors have been used widely by various researchers.
Kimber and Gray (1967) applied such a system to coal studies at
temperatures up to 2800K.  One of the main advantages of flow
furnaces is the similarity to actual systems.  By controlling both
gas and wall temperatures, a wide range of simulation is possible,
and the results may be readily applicable to real systems.  A
typical flow furnace system includes a temperature controlled
furnace, a high temperature gas supply device, a coal feeder, and
a sampling probe.  The relative complexity of the system and material
problems at high temperatures are major detriments of such an
apparatus.
     For the present study, a laminar flow furnace was chosen for
its versatility.  Similar systems were developed by BCURA researchers
 (Field, 1970 and Badzioch and Hawksley, 1972) , but some important
improvements were made in system designs and characterizations.
These improvements include a partially fluidized vibrating feeder,
a water-quenched collector with a bronze filter and particle velocity
measurements by a laser doppler anemometer.   In the BCURA studies,
char particles were quenched by thermal conduction from water cooled
collector walls and then  separated by cyclone separators.  Good
recovery of particles was not achieved because of the adhesion of
char particles  to the collector wall.  In the present study this
problem is  solved by collecting char particles with a bronze filter
at  the  mouth  of the collector.  Water or argon  is  injected at the
mouth of  the  collector  in order to quench the coal particles rapidly.

-------
                                                                  15
III-r.Q  PYROLYSIS EXPERIMENT



     The objective of the pyrolysis experiments is to elucidate the



mechanism and rate of nitrogen devolatilization from coal.  Regula-



tion of the evolution of nitrogenous volatiles is of major



importance in the control of nitric oxide emission.

-------
                                                                  16
III-l.l.O  LAMINAR FLOW EXPERIMENTS
     A schematic of the laminar flow furnace system is shown in
Figure III-l.  The system is built around a modified ASTRO model
1000A graphite furnace (ASTRO Industries, Inc., Santa Barbara,
California).  A 3.5 in. I.D., 12 in. long graphite resistance
heating element is used to control furnace temperatures.  A 4 in.
I.D. graphite radiation shield and packed graphite powder provide
thermal insulation.  The outer shell is about 11 inches in diameter
and water-cooled for easy access.  The furnace is equipped with a
programmed automatic temperature control which allows temperature
changes to be made at desired rates.  This feature is especially
important when ceramic muffle tubes are used, since they are
susceptible to thermal shock.  The temperature is controlled through
a tungsten-rhenium thermocouple which is located just outside the
heating element in the center of the furnace.  The maximum operating
temperature of the furnace is about 3000K with a graphite muffle
tube.  Three pairs of observation ports are located along the axis
of the furnace at intervals of 1.5 inches.  The ports have quartz
windows which supply an observation area of 0.5 in. in diameter.
Optical instruments such as a radiation pyrometer or a laser doppler
anemometer  are used through these windows.  Coal particles are
injected  into the center of the 2 in. I.D. graphite muffle tube
through a 0.047 in. I.D.,  3/16 in. O.D. water-cooled stainless steel
injector.   A 5/8  in. O.D.  graphite  tube is used as a heat insulator
to minimize the heat loss  to the water-cooled injector.  Preheated
gas  is introduced horizontal to the axis of the furnace and

-------
                   FIGURE  III-l.
                                                             17
                      Cool end Carrier Gas
      Radiation
      Shield
 Graphite
 Honeycomb
 Flow Stroightener
  Plasma Gun
 0
c
  Argon
Quartz
Observation
Ports
           Water Out 
                                             Graphite
                                             Heating
                                             Element
                                             Graphite
                                             Muffle
                                             Tube
                                                  Bronze
                                                  Filter

                                                  Water
                                                  Cooled
                                                  Collector
                                              Exhaust
                                              Cooling
                                              Coil
                                               =: Water
           Cooling

           Quenching H?O
           Vacuum    	
                   LAMINAR FLOW FURNACE  SYSTEM

-------
                                                                  18
straightened through a 1 inch thick graphite honeycomb.  The
honeycomb holes are 0.067 inches in diameter and the distance
between holes is 0.083 inches.  The lower surface of the honeycomb
is 0.5 inches above the center of the first observation port.
Six 2-3/8 inch disks are placed 1 inch apart above the honeycomb
to prevent radiant heat loss.  The coal particles are heated as
the carrier gas mixes with the preheated main gas flow.  The char
is collected in a water-cooled bronze filter after being quenched
by water jets.  The rest of the gas is cooled by an exchanger and
exhausted.  Since the muffle tube is constructed from graphite,
atmospheric air must be excluded at all times.  Two ball valves
provide a convenient method for quickly closing the passages
through which the collector is removed or for shutting off access
to the plasma gun.  The graphite muffle tube and the honeycomb
are  replaced by an alumina muffle tube and an alumina honeycomb
for experiments in oxidizing atmospheres.

-------
                                                                  19





III-l.l.l  COAL FEEDER



     The coal feeder (Figure III-2) employs a mechanical vibrator



and partial fluidization of coal particles.  The combined system



results in relatively uniform feed rates.  The main body is made



of polished plexiglas which allows observation of the coal level



and feed rate.  The I.D. of the plexiglas section is 0.5 inches



and the length 3 inches; about 5 gin. of coal can be charged.  The



feed rate is controlled by the needle valve opening, the vibrator



strength, and the carrier gas flow rate.  The carrier gas is



introduced through the hollow needle valve and injected radially



through four equally spaced, 0.010 inch holes at the tip of the



needle valve.  The gas fluidizes :the coal particles locally, thus



preventing packing and clogging in the narrow region.  An external



electric vibrator shakes the entire feeder assembly.  The exit of



the feeder is connected to a 0.065 in. O.D. and 0.047 in. I.D.



stainless steel tubing which vertically enters the furnace.  The



vertical arrangement of the feeder-injector line allows the use of



a fine tube with a very small amount of carrier gas flow (as small



as 5 cc/min).  Less carrier gas allows more rapid heating of the



injected particles.  Feed rates between 0.01 gm/min and 1.0 gm/min



can be obtained with reasonable uniformity.  In normal operations



about 0.15 gm/min  is used.

-------
            FIGURE III-2.
  Vocuum
  Argon
  Flush
Feed  Rote
Control
Valve
                      r
Carrier Gas
                                                20
                     II
Carrier Gas
Jet
1 1
L I

I I


         Coal Feed
           Port
        Coal


        Plexiglas
         Body
             Coal and  Carrier das
                 COAL FEEDER.

-------
                                                                  21





III-1.1.2  PARTICLE COLLECTOR



     The collector (Figure III-3) consists of three concentric



copper tubes.  The external cooling water passes through two of



the passages and the third is used to withdraw the quenched gases.



Eight 1/8 in. O.D. copper tubes pass through the second water



passage to carry the quenching water to be directly injected.  Heat



transfer analysis shows that cooling by conduction to the water-



cooled wall is insufficient to achieve char quenching because of the



relatively large hot gas flow rate (up to lA/sec) and the neces-



sarily large collector diameter.  This problem is solved by directly



injecting quenching water at the mouth of the collector through



twelve holes drilled in the flared upper section of the sintered



bronze filter.   (The filters are manufactured specially by Thermet



Inc., Gloucester, Mass.).  The porous bronze filter allows some of



the quenching water to pass through the whole area of the upper



section of the filter, preventing coal particles from adhering and



being exposed to intense radiation from the furnace wall.  The



bronze filter is 0.740 in. O.D. at the open end, 0.4 in. O.D. and



4 inches long at the straight section, and the wall thickness is



0.060 inches.  The average size of bronze is about 15 microns



and coal particles larger than 5 micron are retained in the filter.

-------
FIGURE III-3
                           22
water jets-
. . *
water __0  	
out lu* 	

1j-
-*~
1 |

(-LjQ in i -._i -iTz~.

water in  -t."
-
i *~r>i-il rirtH
SL ' " f c

1 A>
^^ A r~
1 f
: j
1
i
i !

IT  
_ *
	
*

vac

* 	 Tt->^-
*
L 	 ..
i
-- 
mm
X, hot gos
A
VS
bronze
filter
^-*fl
jj
WATER COOLED COLLECTOR

-------
                                                                  23





III-1.1.3  GAS PREHEATER



     An AVCO plasma gun model PG 040 is attached to the furnace to



preheat the main gas stream.  The plasma gun has a water-cooled



copper anode.  An AIRCO model 1500-C7 welding rectifier supplies



D.C. power to the gun.  The maximum power input to the plasma gun



is 40 Kw.  The plasma heats the gas with a 30 percent thermal



efficiency-  The temperature of the preheated gas is controlled by



changing the power input and the gas flow rate.  Argon, helium,



nitrogen or  hydrogen can be used as working gases.  A graphite



muffle tube restricts the working gas, therefore only inert gases



can be used.

-------
                                                                 24





III-1.1.4  SUCTION SYSTEM



     A mechanical vacuum pump is used to withdraw the injected



quenching water and the quenched argon gas.   Water is separated



from the gas by a gravitational separation vessel.  The quenched



gas is dried through a packed bed of calcium sulfate and metered



by a Fischer and Porter flowmeter.

-------
                                                                  25




III-1.1.5  EXPERIMENTAL PROCEDURE



     The furnace is heated to a desired temperature using the



automatic temperature controller.  Three tungsten-rhenium thermo-



couples, introduced through :-the observation ports, are used to



monitor the temperature inside the muffle tube.  After steady-



state is reached, the main gas is introduced and the plasma gun



is started.  The temperature of the main gas is controlled to the



same temperature as the furnace wall temperature.  This is done



by matching the thermocouple readings before and after the intro-



duction of the main gas.  The main gas flow rate is controlled by



a Fischer and Porter flowmeter to give desired main gas velocities.



The pressure of the furnace is maintained at about 2 inches of



water.



     A predried bronze  filter is weighed and placed in the collector



Suction rates vary between 1/3 and 1/2 of the main gas flow rates



depending on the furnace temperature .   The flow rate of the



quenching water jets is maintained at 0.7 fc/min for most runs.



After a desired suction rate and a water flow rate are obtained,



the collector is raised into the furnace and positioned at a pre-



scribed distance from the coal particle injector.  About 0.05 to



0.5 gm of coal is weighed and charged into the feeder.  Then, a



carrier gas flow rate and a needle valve opening are set and the



vibrator is turned on.  The coal level in the feeder is continuously



monitored.  As char particles accumulate in the bronze filter, a



cake builds up.  In order to maintain the same suction rate, the



valve opening of the suction line has to be increased to compensate



for the increased pressure drop across the filter.  The maximum

-------
                                                                  26
pressure drop allowed is about 20 in. Hg.  Upon completion of the
coal particle injection, the collector is removed from the furnace
and replaced with a new bronze filter.  The collector is raised
again to a different location and the next run is started.  Normally
about ten runs are repeated at the same experimental conditions with
different collector locations.  After all runs are finished, the
filters are dried at 110C for three hours in a nitrogen atmosphere.
The filters are then cooled in a desiccator and weighed as soon
as the room themperature is reached.  Char samples may be analyzed
for ash content or sent to Galbraith Laboratories, Inc., Knoxville,
Tennessee  for the ultimate analysis.  Weight loss of coal is
determined by the difference of the weights between the coal fed
and the collected char.

-------
                                                                 27
III-1.2.0  FREE FALL EXPERIMENT WITH BRONZE COLLECTOR



     In order to obtain longer reaction times, coal particles are



allowed to fall under the influence of gravity through a preheated



furnace.  The schematic of the system (Figure III-4) is similar



to the flow furnace shown in Figure III-l.  The main component of



the system is an ASTRO model 100A graphite furnace.  The same



temperature control system, which is used for the laminar flow



furnace, controls the furnace wall temperature.  A  24 inch long,



3 in.O.D., 2 1/2 in.I.D. alumina muffle tube shields the graphite



heating element from the reaction zone.  A 15 inch  long, 2 1/4 in.



O.D., 2 in. I.D. alumina tube supports an alumina honeycomb flow



straightener.  The honeycomb, made by E.I. du Pont  de Nemours, has



a nominal pore diameter of 1/16 inch.  The maximum  temperatures of



the alumina tubes and honeycomb are 2100K and 1900K respectively.



A 9 inch long, 3 1/2 in.I.D. cooling section is attached to the



bottom of the furnace.  The coal feeder and the water-cooled



injector are the same as the ones used in the laminar flow system,



except the diameters of the tubes.  The water-cooled injector is



0.250 in. O.D. and 0.101 in.I.D.  Small amounts of main gas (about



0.05 A/sec) are introduced from the top of the furnace.  The gas is



heated by conduction from the muffle tube wall and honeycomb.  The



honeycomb induces uniform velocity and temperature profiles in the



radial direction.  The temperatures along the axis of the furnace



are measured by a chromel-alumel thermocouple up to 1500K and by



a platinum-rhodium or an iridium-rhodium thermocouple above 1500K.



Since the main gas Reynolds number is small, the flow is practically

-------
                FIGURE III-4
                                                       28
                                Cool Particles
                                and Carrier Gas
       Main Gas
Honeycomb
Flow  Straightener
      Water Cooled
      Particle  Injector
                                             Graphite
                                             Heating
                                             Element

                                             Alumina
                                             Muffle
                                             Tube

                                             Alumina
                                             Liner Tube
                                             To Support
                                             Honeycomb

                                             Water
                                             Cooled
                                              Shell
         Cooling
         Section
         Water In 
              Vacuum
Water Out
 Particle  Collecting
 Porous Bronze Disc
 Optical  Grade  Quartz
 Plate
           FREE PALL EXPERIMENT WITH BRONZE COLLECTOR

-------
                                                                 29






fully developed for the entire region.  Therefore the temperature



of the main gas is considered to be close to the wall temperature



of the muffle tube.  Coal particles are introduced into the center



line of the main flow through the water-cooled injector.  Carrier



gas flow rates between 2 cc/min. and 25 cc/min.  are  used.  Char



particles are collected on a 3 inch diameter, 1/8 inch thick



sintered bronze disk at the bottom of the cooling section.  The



bronze discs are manufactured by Thermet, Inc., Gloucester, Mass.



They have an average pore diameter of about 5 microns.  The gas



exits through a 1/4 inch pipe fitting under the center of the bronze



disc.

-------
                                                                  30





III-1.2.1  EXPERIMENTAL PROCEDURE



     The furnace is heated to a desired temperature at a heating



rate between 250C/hr. and 350C/hr., using the temperature



programmer.  After a steady state is reached, a preweighed bronze



filter is placed on the bottom flange.  Then the furnace is alter-



natively evacuated and filled with argon several times to eliminate



atmospheric air.  About 1 gram of coal is weighed and charged to



the feeder.  The suction rate is controlled so that the furnace



pressure is maintained at atmospheric pressure for fixed main gas



and carrier gas flows.  The vibrator is started and coal is fed in



about 10 minutes, giving a feed rate of 0.1 gm/min.  After all the



particles are fed, the bronze disc is removed and quickly weighed



to avoid adsorption of moisture.  The collected char is transferred



to a sample bottle for analyses, such as ash content, elemental



composition, and particle size distribution.

-------
                                                                  31





III-1.2.2    FREE FALL EXPERIMENT WITH ALUMINA COLLECTOR



     The object of this experiment is to obtain the asymptotic



weight loss values at high heating rates and very long reaction



times.  The furnace is the same as used in the free fall experiment



with the bronze filters.  A 40 mm. O.D., 165 mm. long cylindrical



alumina crucible  (Coors CN-170) is supported by a 1 7/8 in. O.D.,



3/8 in. I.D. alumina tube.  A thermocouple is placed between the



support disc and  the bottom of the collector to monitor the temper-



ature during the  experiment.  The crucible temperature is recorded



continuously on a chart recorder.  The collector is raised or lowered



by a motor-driven linear actuator at a constant speed.  This is to



protect the ceramic material from thermal shock caused by a sudden



temperature change.  At 1830K it takes about 20 minutes to position



or remove the alumina collector.

-------
                                                                 32





III-1.2.3  EXPERIMENTAL PROCEDURE



     After a desired furnace temperature is attained, a preweighed



alumina collector is raised to the hot zone of the furnace by the



linear actuator.  The furnace is continuously purged by argon to



prevent contamination by air leaks.  The temperature at the bottom



of the collector is recorded and when a steady state is reached,



about 1 gram of coal is fed in approximately 10 minutes.  Main gas



flow rates between 0 and 0.05 a/sec and carrier gas flow rates



between 2 cc/min. and 24 cc/min are used.  After all the particles



are fed, the collector is kept for 10 minutes at the same position



in order to assure sufficient reaction time for the particles fed



last.  The collector is then lowered into the cooling section and



transferred to a desiccator.  About half of the char sample and the



material left in the crucible are analyzed for ash content.

-------
                                                                  33
III-1.3.0  CRUCIBLE EXPERIMENT



     The object of the crucible experiments is to determine weight



loss under slow heating conditions and extremely long residence



times.  Instead of feeding and collecting coal particles, as in



the case of free fall and laminar flow experiments, about 1 gram



of coal is weighed and placed in a 15 mfc alumina crucible (Coors



CH-16), then the crucible is raised into the hot zone of the furnace



using  the linear actuator discussed above.  The temperature between



the crucible and the support disk is recorded by a thermocouple



during the whole cycle of heating and cooling.  Heating rates



between 0.5K/sec and 5.0K/sec can be obtained.  It should be



noted  that this experiment is close to the ASTM proximate analysis



for the determination of volatile matter, where one gram of coal



is placed in a 15 mJl platinum crucible, lowered into a furnace



heated to 950C, and held for seven minutes.




III-1.3.1  EXPERIMENTAL PROCEDURE



     The procedure is the same as that of the free fall experiment



with alumina collector, except no coal if fed.

-------
                                                                  34
III-2.0  OXIDATION EXPERIMENTS
     Two types of oxidation experiments serving two different
purposes have been performed, the former a coal or char oxidation
experiment at a fixed residence time of one second (free fall mode),
the latter a time-resolved char oxidation experiment (laminar flow
mode).  The objective of the fixed residence time experiments is
to investigate the conversion of fuel nitrogen to nitrogen oxides
under different burning conditions.  Comparison of coal and char
oxidation results  (See Section VI-3 )  enables the separate con-
tributions to NO  emissions of volatiles and char to be identified.
                a
Variables of greatest interest which are investigated are flame
temperature and fuel equivalence ratio.  The distribution of
nitrogenous products in the gas phase and condensed phase are of
particular interest for the evaluation of the contribution of fuel
nitrogen to noxious bound nitrogen products.  The objective of the
time-resolved experiments is to determine the kinetics of char
oxidation with a view towards establishing residence time require-
ments  for completing char burnout and to further determine the fate
of fuel nitrogen undergoing heterogeneous oxidation.

-------
                                                                  35



III-2.1.0  FREE FALL EXPERIMENT



     Coal and coal char oxidation experiments at a fixed residence



time of one second are run in the alumina muffle tube furnace in



free fall mode as described in Section III-1.2.0.  The coal-char



used in these studies is prepared by pyrolysis of the coal at the



temperature and residence time corresponding to those of the



oxidation experiments.  Either air or a helium-oxygen mixture,



simulating air, is used as the oxidizing atmosphere in all



oxidation experiments.

-------
                                                                  36
II1-2.1.1  EXPERIMENTAL PROCEDURE
     The experimental procedure is the same as that described in
Section III-1.2.1  except that an oxidizing gas is now introduced
into the furnace as the main gas.  The main gas flow is typically
on the order of 3000 cc/min.  A small amount of additional oxidizing
gas, about 30 cc/min., is used to carry the coal into the furnace.
These conditions result in a small flame stabilized on the injector
nozzle.  The flame length depends somewhat on operating conditions,
but is typically about two inches long.  Large amounts of soot and
unburned coal are observed leaving the flame under fuel rich flame
conditions, but only ash is recovered under lean conditions.  The
furnace wall temperature is controlled in the manner described
previously.  The wall temperature is measured both with an optical
pyrometer and with thermocouples.  Nitric oxide concentrations are
measured with a chemiluminescence NO meter from a slip stream of
the total gas flow out of the furnace  (See Section III-4).
     Quenched char and condensable material are collected at the
bottom of the furnace on a porous sintered bronze disk.  The
recovered material is weighed to determine the percent of the
material burned and some samples are sent to Galbraith Laboratories,
Inc., Knoxville, Tennessee for nitrogen analysis by the Kjeldahl
procedure.

-------
                                                                  37
III-2.2.0  LAMINAR FLOW EXPERIMENT



     The alumina muffle tube furnace described in Section III-1.2.0



can also be used in a laminar flow mode by removing the water-



cooled bottom section and adding water-cooled flanges for the



accommodation of a  water-cooled  probe.  A schematic of the experi-



mental apparatus is shown in Figure III-5.  This alumina muffle



tube laminar flow  furnace is similar to the graphite muffle tube



laminar flow furnace described  in Section III-l.l.O, except for



the differences  in particle collector, gas preheater, and suction



system.

-------
                 FIGURE  III-5.
                                                         38
                         Cool & CorrUzr  Gas
Main Gas
                                Honeycomb
                                Flow Stroightener
      Graphite
      Heating
      Clement
                                Alumina
                                Rod
                                With
                                Thermocouple
                                        .-Cool Collector
                                               Crucible
Water
Cooling
Coil
     Vacuum
        or
      Purge
         FREE FALL  EXPERIMENT WITH  ALUMINA COLLECTOR
                 AND CRUCIBLE EXPERIMENT

-------
                                                                  39

III-2.2.1  PARTICLE COLLECTOR
     The collector probe is essentially the same as the one
described in Section III-1.1.2.  However, as the furnace studies
have progressed-, the water quench initially used was dropped in
favor of a gas quench for the following reasons:

     1.  With water quenching, the bronze collecting filter
         quickly becomes clogged, probably due to the
         tendency of water to wash small ash particles
         off the char surface into the filter pores.
         With gas quenching, there has been no sign
         of filter clogging.

     2.  Use of a gas quench simplifies the experimental
         apparatus.  With only gases, but no water,
         evolved from the furnace, the large reservoir
         required to separate gases from water is
         eliminated.

     3.  Since for these fuel-nitrogen studies gas phase
         analyses are essential, elimination of the water
         quench assures higher accuracy by reducing possible
         gas losses due to solution in the excess water.
         Dilution of the effluent gases does, however,
         become a potential problem at very low concen-
         trations .

     Nitrogen, argon, and helium have been tested for efficiency
of quenching.  Argon has been found to cool the hot main gases
most efficiently.  Typical gas quench flow rates required are
about 25 A/min  with a  main gas flow rate through the collector
of 3 t/min. (corrected to room temperature),  which is half of the
total main gas flow rate supplied to the furnace of 6 Z/min.

-------
                                                                    40
The reason the gas quench is feasible in the present system is that
the quantity of hot main gas removed through the collector is much
smaller than that in the system descibed in Section III-l.l.O
(3 A/min. versus 30 t/mint) .

-------
                                                                  41
III-2.2.2  GAS PREHEATER



     The main gas enters the furnace through the top plate at room



temperature.  Heat transfer calculations show that at any furnace



temperature and the maximum gas flow rate of 6 A/min. (measured at



room temperature), the gas reaches the furnace temperature before



leaving the alumina honeycomb.  These predictions are confirmed by



measuring the gas temperature with a Pt-Rh thermocouple as it passes



through the honeycomb.  The system has nevertheless been equipped



with a packed bed preheater for use with higher gas flow rates.



The preheater is placed upstream of the furnace main gas inlet



and consists of a Lindberg model 55081-A tubular heater (120V, 8.5 A)



capable of reaching a temperature of 1000C.  The packed bed, made



of about 5 mm. alumina pellets, is 12 inches in depth and is con-



tained in a 1 inch O.D. stainless steel tube.

-------
                                                                 42





III-2.2.3  VACUUM SYSTEM



     Suction for the drawing of gases through the collector probe



is accomplished by the use of two Cenco Megavac pumps in parallel.



These pumps (115 V, 6 A, 1/3 hp) have individual capacities of



57 fc/min. maximum at 1 atmosphere.  The vacuum flow rate is simply



controlled by a regulating valve.  All vacuum lines are 1/2 inch



in diameter or larger to help minimize pressure drops and maximize



pumping speeds.

-------
                                                                  43





III-2.2.4  PREPARATION OF CHAR



     In selecting a char for the oxidation and kinetics studies,



two criteria must be satisfied.  First, the char should contain



a minimum of volatiles (largely carbon and hydrogen) at the



temperatures and residence times pertinent to the oxidation



experiments.  A volatile-free char would permit isolating the



contributions of purely heterogeneous processes in the study of



fuel nitrogen conversion to nitric oxide.  Second, the char should



retain as large a percentage as possible of the nitrogen originally



present in the coal so that the fate of the char-nitrogen during



oxidation can be accurately studied with a minimum sampling error.



Taking into account both criteria, the char used for this study



was produced by pyrolyzing lignite coal particles at 1750K for a



residence time of one second.  The char produced was then size-



graded to select particles in the 38 to 44 micron size range for



further oxidation experiments.  For the production of char, a minor



modification is made to the furnace in free-fall mode (Section III-



2.1.0) in order to permit the collection of the char in a holding



flask.

-------
                                                                  44





III-2.2.5  EXPERIMENTAL PROCEDURE



     The furnace is heated at 250 to 350C per hour to the desired



temperature by using the automatic temperature controller and pro-



grammer.  The final temperature is checked with the platinum-



rhodium thermocouple extending just below the honeycomb.  Typical



temperature profiles along the muffle tube axis are determined by



moving a platinum-rhodium thermocouple along the axis from the



bottom of the furnace.  Once the furnace has reached the desired



temperature, a preweighed bronze filter is placed in the collector



probe.  About 1 gram of char is weighed and placed in the feeder



with the needle valve closed.  The quenching gas flow rate and



vacuum are set to just counter each other at about 25 A/min. so



that the furnace pressure does not change rapidly as the probe is



put into the muffle tube.  The collector is then raised into the



furnace and secured at a particular height.  An O-ring, through



which the probe is pushed, seals the muffle tube from the external



atmosphere.  The main gas is next adjusted to provide the desired



oxidizing gas partial pressure using helium as a diluent.  The total



main gas flow rate is always 6 -/min. STP with a maximum of 3 fc/min.



being drawn into the collector for purposes of proper quenching.



The remainder of the gas passes out the bottom of the furnace, its



flow rate being controlled by a rotameter in conjunction with the



vacuum control valve.  As the main gas is being set, the furnace



pressure is continuously monitored and maintained at a slight



positive pressure  (less than 1 psig), also by adjustment of the



vacuum.  Once the muffle tube has been purged for 2 min., the



carrier gas is set at about 30 cc/min.  (only 1/2 of 1 percent

-------
                                                                  45





of total main gas flow)  and the needle valve and vibrator are



adjusted so that the char feeds fairly consistently in 6-7 min.



or at a rate of about 0.15 g/min.  When the feeding is finished,



the main gas is turned off  and the collector probe is removed.



The quenching gas and vacuum are then turned off and the filter



is very quickly removed and weighed to avoid adsorption of moisture.



Finally, the samples are stored in sealed glass vials to await



further analyses.

-------
                                                                   46
III-3  PROCEDURES FOR MEASURING NITRIC OXIDE REDUCTION BY CHAR



     The NO-char experiments are carried out in the alumina muffle



tube furnace operated in the free-fall mode as described in



Section III-2.1.0.  Mixtures of nitric oxides in helium (or argon),



supplied by Matheson Gas Company, are used as main gases.  Three



different concentrations of nitric oxides, i.e. 530 ppm and 6,400 ppm



(all balanced in Helium) and 1100 ppm (balanced in Argon), have



been used.  The char used in this study was prepared by the pro-



cedures described in Section III-2.2.4.  The main gas flow is



typically on the order of 3 A/rain.  A Chemiluminescent NO-NO
                                                            X


Gas Analyzer and a chart recorder are used to monitor the NO



concentration in the furnace exit stream  (See Section III-4).  Two



to three minutes after the NO containing stream is first introduced



into the furnace, the NO concentration in the exit stream stabilizes



at the concentration of the inlet stream.  After steady conditions



have been reached, char feeding is commenced and the decrease in



NO concentration in the exit stream is observed and recorded con-



tinuously.  After completion of the char feeding, the NO concentra-



tion in the exit stream returns gradually to its original level.



When the NO concentration in the exit stream has returned to its



original level and stabilized, the main NO gas is turned off and



a secondary helium gas is introduced into the furnace to purge the



furnace until no more NO can be detected.  The secondary helium gas



is then turned off and the residue char is removed from the sintered



bronze disc filter for analysis.

-------
                                                                  47




III-4.0 GAS PHASE ANALYSES



    Analysis of the volatile pyrolysis and oxidation products



formed in the furnace complements the weight loss data obtained



through analysis of the char.  Emphasis was placed here on analysis



for nitrogen-bearing species, especially NO, HCN, and NH3/ though



some efforts were made to detect N2 by gas chromatography.  Carbon



containing species of interest included CO, CO2f and CH4, analysis



for higher hydrocarbons being outside the scope of this work.  In



all cases, volatile products passed through the furnace hot zone,



obscuring any distinction between primary and secondary products.



Methods employed for analysis are detailed below.

-------
                                                                  48
III-4.1  CHEMILUMINESCENT GAS ANALYSIS OF NO
   ^^^ "            ^^^^^^^^^^^~      ^-^^-     X

     A Thermo Electron Model 10B Cheitiiluminescent NO-NO  Gas
                                                       
-------
                                                                  49
III-4.2  COLORIMETRIC ANALYSIS OF NH3




     The basic gas sampling unit consists of two bubblers, a flow



regulator, and a wet test meter.  Off-gases from the furnace



are passed through the bubblers containing 0.1 N H2SO4 solution



to absorb NH3.  A sample aliquot from the bubbler is then tested



for ammonia content by the direct Nesslerization method in



accordance with ASTM procedures.



III-4.3  STANDARD SPECIFIC ION ELECTRODE ANALYSIS OF HCN



     HCN gas from the furnace exit stream is collected in a gas



sampling unit.  The gas sampling unit is essentially identical



to the unit used for NH- sampling except that the absorbing agent



in the bubblers is now changed to 0.1 N NaOH solution.  After



absorption has been completed, an aliquot of 0.01 N Pb(NO3)2 is



added to the sample solution for the removal of the potential



interference by Sion derived from the coal surface.  The



cyanide content in the sample solution is then measured by an



Orion cyanide ion electrode model 94-06.  The electrode is



calibrated frequently against a series of standard NaCN solutions



which are prepared daily.

-------
                                                                  50
III-4.4  GAS CHROMATOGRAPH1C ANALYSIS OF N2, CO, CQ2, and CH4
     Calibrated Instruments^ five-layer material gas sampling bags
are used to collect gases from the furnace exit streams.  The
sampling bag is then connected to a gas chromatographic system
and gases are introduced into a sample loop for GC analyses.  The
GC system is a Fisher/Hamilton Model 29 Gas Partitioner.  This
instrument is equipped with two columns.  Column Number One is
packed with 30% Di-2-ethylhexylsebacate on 60-80 .mesh Chromosorb P
and Column Number Two is packed with 40M50 mesh Molecular Sieve 5A.
Both columns are operated at room temperature.  Alternatively, the
packing materials in both columns can be simultaneously replaced
by Porapak Q with Column Number One operating at room temperature
and Column Number Two operating at dry ice temperature (-78C).
The chromatograph is calibrated against a standard gas before and
after each sample run to ensure accuracy.  The standard gas,
supplied by Matheson Gas Company, consists of known concentrations
of CO, CO2, CH4, N2, and H2 in Helium.  Total CO, CO2, CH., and N2
emissions from the furnace are calculated by normalizing the GC
signals to those of the standard gases.

-------
                                                                 51





III-4.5  NON-DISPERSIVE INFRARED ANALYSIS OF CO AND CO2




     Co and CO2 from the furnace are also measured by Beckman Non-



Dispersive Infrared Analyzers.  Model 315A is used for CO- analysis



and Model 315B for CO.  Gases from the furnace exit stream are



introduced directly into thses instruments for a continuous



monitoring of CO and CO,.  The instruments are calibrated fre-



quently against standard CO and CO2 mixtures  respectively.

-------
                                                                  52
III-5  ASTM PROXIMATE EQUIPEMENT



     A drying oven, a cylindrical furnace, and a brick-rinsulated



furnace are used to determine moisture, volatile matter, and ash



contents.  ASTM standards  (D3173-73, D3174-73, and D3175-73, 1972)



are followed wherever possible.  Some modifications are made,



however, to accommodate smaller sample sizes and different



experimental equipment.  A Hoskin cylindrical furnace (Model 303,



5.0 amps) is the main part of the system to determine volatile



matter.  The furnace is capable of a maximum temperature of 1100C



at full load.  The heated cylindrical tube is 1.38 inches in



diameter and 13 inches long.  The equipment has a cooling section



15 inches long and 1.5 inches I.D. in order to provide adequate



cooling for the devolatilized samples in an inert atmosphere.  This



prevents oxidation of the hot samples.  The whole assembly is



sealed with a high temperature alumina cement.  The furnace is



continuously purged with nitrogen when determining proximate



volatile matter.  A chromel^alumel thermocouple is centered in the



heating  zone 5.0 inches from the bottom.  A porcelain crucible



 (5 m capacity) containing the samples to be analyzed is suspended



in the hot  zone of the furnace about 0.2 inches above the tip of  the



thermocouple.  A reasonably steady hot zone  ( _ 20C) is maintained



4.3  inches  to  8.6 inches from  the bottom.



     Volatile  matter determined by this system has been compared



with the results of two independent laboratories  (Galbraith Labora-



tories,  Inc.,  Knoxville, Tennessee and M.I.T. Central Analytical



Facility) which followed the ASTM specifications.  Agreements within

-------
0.5 percent of the weights of original coals were obtained.  The
agreement justifies the present system.
     Ash content is determined in a magnesium brick furnace.  The
inside dimensions of the furnace are a base 8.5 inches by 4 inches
and a height of 9 inches.  This furnace is large enough to accomo-
date the cylindrical alumina collector used in the free-fall
experiment.  Four silicon carbide heating elements (Norton XL 3/8 in.
O.D. by 13 inches long) supply necessary heat to the furnace.  Glazed
porcelain crucibles are used to hold the samples.  It was confirmed
that the use of different crucible sizes did not affect ash content.
ASTM standard procedures were used with the exception of the size
of the crucibles and samples.

-------
                                                                 54
                         SECTION IV



                   COAL CHARACTERIZATION





     Coals differ widely depending on geological age, geographical



location and geological conditions.  Major differences may be found



in coal composition even between different positions in a particular



coal seam.  Classifications developed to differentiate between



different coals have been based on rank, physical appearance,



elemental analysis, and functional group analysis.



     Rank is a gross measure of properties such as heat of com-



bustion, carbon content, agglomerating and aggluting tendencies



of coals with definitions differing between countries.



     Because of ease of utilization, rank, sometimes supplemented



by elemental analysis, has been the classification adopted by



industry for providing a measure of those properties that determine



the end use of coal.  Research workers have been seeking a more



fundamental measure of coal behavior and limited success has been



achieved using petrographic analysis.  Functional form analysis



provides a more fundamental measure of coal structure but its use



is qualified by the difficulty of the measurements, which usually



require solvation  of  the coal with possible modification of the



coal structure.



     It is a long-range objective of fundamental coal studies to



correlate the behavior of coals with a classification scheme,

-------
                                                                 55
preferably simple.  For the current study, a lignite and a



bituminous coal, two widely different coals, have been selected.



The characterization of these is given below.

-------
                                                                 56
IV-1.0  MONTANA LIGNITE


     The Montana lignite was chosen as representative of a low


rank, low sulfur, non-caking western coal which could be strip


mined.  These coals are likely to find future use as raw material


for gasifiers and possibly as fuel for local power generation.


The reserves of lignites in the United States represent approxi-

             4
mately 6 x 10  quadrillion BTUs or 28.8% of the ultimate United


States Coal reserves.

-------
                                                                 57
IV-1.1  VISUAL APPEARANCE OF MONTANA LIGNITE
     Scanning electron micrographs of the Montana lignite used in
these studies are shown in Figure IV-1.  The lignite has been
ground and classified to nominally 38-43 microns.  The low
magnification micrograph shows many small particles have escaped
mechanical size classification and a few large, irregularly shaped
particles are also present in the supposed narrowly-sized fraction.
The micrographs at higher magnification show the particle surface
is relatively smooth, has few visible macropores, and has ash
particles on the order of several microns physically held on the
surface.
     The photomicrographs of polished sections under oil immersion
are shown in Figure IV-2.  The micrographs show that the coal
particles are not uniform in internal macropore area or in re-
flectance.  Four different levels of reflectance have been iden-
tified in the polished section slides.  Most particles appeared
to be composed of only one petrographic element although occasion-
ally one particle will contain as many as three different elements.
     The top micrograph in Figure IV-2 shows three particles of
different reflectances.  The whitish particle on the right side of
the micrograph is tentatively identified as micronite; the highly
reflective fusanite particles are very rare in lignites and are
not evident in the samples; the dull, barely visible particle just
to the left of the micronite particle has been identified as exinite;
and the particle below and to the left of the exinite particle has
been identified as vitrinite.  The top micrograph shows some macro-
pore structure in the vitrinite and exinite, but most of the

-------
       RAW MONT>'-\~ UGNl   ,^8-45
             0           400 Jrm
                  100 X
                 H  K.    W
0            40um
     1000X

                                   30OOX

FIGURE VI-1.  Scanning electron microgra:    raw Montana

            lignite-A.

-------
      POLISHED  MONTANA LIGNITE
             under  ceder oil  450X
FIGURE VI-2.  Photomicrograph of polished raw

-------
                                                                60





macropore structure is in the medium reflective micronite.  The



macropores appear either in micronite in the form of circular voids



as large as ten microns in diameter or as dentritic-like structures



with stem diameters on the order of several microns.  Fewer pores



are visible in vitrinite and exinite; these pores are either



irregular shaped voids on the order of 5-10 microns in diameter,



or relatively straight fissures, about one micron in diameter.

-------
                                                                 61
IV-1.2 PARTICLE SIZE DISTRIBUTION



     The Montana lignite was received as 1/8 inch particles.  These



particles were ground in a ball mill and separated through roto-



tapped Tyler screens.  Most of this study was conducted with the



38-43 micron size range, although some data was taken using the



74-88 micron size range.  These two size ranges were chosen as



being representative of coal fed to a pulverized coal flame.



     The electron micrograph in Figure IV-1 showed that the closely



sized fraction had a number of undersized particles and a few



oversized particles, shown more quantitatively by the particle size



distribution in Figures IV-3 and IV-4.  The cumulative mass dis-



tribution was fitted with the Rossin-Rammler (1939) distribution



function, as for example in Figure IV-3.  Scanning electron micro-



graphs of lignite particles were counted on a Zeiss particle counter



and Rosin-Rammler parameters were developed using a least square



fit of the particle frequency.  The mean particle diameter derived



from the Rosin-Rammler distribution was 52.3 microns.  The Rosin-



Rammler derived curve is compared in Figure IV-4 with the original



data.  The original data gave a mean particle size distribution of



47.6 microns and a standard deviation of 15.2 microns.  The fit of



the data is considered satisfactory, in spite of the differences in



calculated and measured frequency at large particle diameters,



because the size interval is small and relatively few oversized



particles are present.

-------
                                 RAW   LIGNITE
                          ROSIN -RAMMLER    DISTRIBUTION
     ~   10
          10
16'
     u
     z
     LJ
     O
     LJ
     LJ
          -2
     <    10
     2
     o
     u
          -3
          10
 I   I  I  I I I II

x  52.3
s   3.42
b -  1.33x10"'
798 Particles
                                   I   I   I I  I I .1
           I  I I
                  I	I  MUM
                                           I   I  I  I I I I
I   \  I  I 1 III
             10               101               10z              X)3

                             DIAMETER  (MICRON)
                                                             .
FIGURE  IV-3. Rosin-Rammler accumulative size distribution of lignite particles.
                                                                    o\
                                                                    N>

-------
     X  30
     V)
     I
     cc
     H
     CO
     cc
      u

      O
      UJ
      tt
      Lu
         25
20
          15
 10
            0
                           RAW   LIGNITE

                   FREQUENCY   DISTRIBUTION
            20
                                I
x - 47.6

(T- 15.2

s .  3.42

b - 1.33X106

798 Pdrticlcs
                                          I
                                                 \
                                         i rJn
 40       60        80

 DIAMETER ( MICRON)
100
FIGURE  IV-4.  Lignite particle size frequency distribution.
                                                                         u>

-------
                                                                 64

TV-1.3 SURFACE AREA

     Numerous measurements of surface areas of raw coal and

processed coals have been reported.  Spencer and Bond (1966),

however, have questioned the use of adsorption isotherms to derive

absolute surface areas of microporous solids.  Van Krevelen (1961)

showed that the areas determined by conventional adsorption methods

at liquid nitrogen temperatures are two orders of magnitude lower

than values derived from adsorption at room  temperature or values

derived from methanol heat of wetting data.  The diffusion process

appears to have an activation energy of about 4 kcal/gmole.  The

activated diffusion process accounts for the low values of surface

area obtained at low temperature, because the adsorbate cannot

penetrate completely into the microstructure of coal in a reasonable

period of time at low temperatures.

     Marsh  (1965) has reviewed the various methods of determining

surface area and concluded that adsorption of CO- near room tempera-

ture gives the most reasonable values for surface areas of coal and

cokes.  Typical surface areas for coals with less than 85% carbon
                                2
 (dmmf) are in the range of 200 m /g.  Adsorption of inert gases at
                                     2
77K gives values on the order of 2 m /g.  Adsorption of krypton at

77K for the lignite used in this study gave a surface area of
   2                                                             o
2 m /g when analyzed using the Dubinin-Polanyi equation and 1.2 m /g

when analyzed using the BET equation.  These values are not repre-

sentative of the total surface area but are considered to be

representative of that fraction of the surface area contained in

pores with diameters greater than about 12 A.   (See Van Krevelen,

Marsh, Can, et al, 1972, and Tingey and Morrey, 1973) .

-------
                                                                 65
     Surface areas from CO2 adsorption at room temperature have


not been measured for the coals used in this study.  The surface


area is expected to be reasonably close to the surface area of

     2
183 m /g measured by Nsakala et al (1975) for a North Dakota


lignite with a daf carbon content of 71%.  THis value is slightly


lower but in reasonable agreement with the surface areas found


by Can et al (1972) of 225-308 m /g for North American lignites


containing between 63 and 72% carbon.  Swann et a_l (1974) report


a value for the CO- adsorption area of Australian Yallourn brown

             2
coal of 290 m /g comparable to that of American lignites.


     The integrated particle size distribution curve yields a

                                  2
superficial surface area of 0.11 m /g.  This value is 20 times


smaller than the krypton adsorption area and 2000 times smaller


than the normal CO, areas for lignites.

-------
                                                                  66





IV-1.4  PROXIMATE ANALYSIS



     Proximate analysis- moisture, volatile matter, and ash content-



were performed at three independent laboratories with similar re-



sults.  The proximate analyses are reported in Table IV-1 along



with other information for the two coals used in this study.  The



volatile matter was approximately 36.2%, the equilibrium moisture



content was close to 13.6%, and the ash was 7.8% on an as received



basis.  The lignite was treated as a sparking coal.

-------
                                           TABLE IV-1
                                     COAL CHARACTERIZATIONS
                                  SSSSF            ULTIMATE ANALYSIS          SOLpoR pORMS
                     BTU/lb       A.R. Wt.%               A.R. Wt.%
   COAL      MINE    MOIST
101 "'"
                                                                          Bv    p..,..
                              H.O    VM   ASH     C     H     N     S    D?ff . . itL   SO,
PITTSBURGH
 SEAM 18    IRELAND  14,377   2.2   4O.7  1O.8  67.81  5.O3  1.20  4.83   8.55  2. SO  0.69  1.62
  hvA-b      .

LIGNITE-A   SAVAGE    8'809  13'6   36'2   7'8  54'5  4'96  *88  '84  1?*14  0'U  *18  *55


*  HIGHER HEATING VALUES WERE CALCULATED ON THE BASIS OF EQUILIBRIUM MOISTURE WITH MINERAL MATTER,
   AS DETERMINED BY PADIA (1976) FROM LOW TEMPERATURE ASHING, REMOVED.  THE MINE MOUTH MOISTURE
   CONTENT OF THE BITUMINOUS COAL WILL BE CLOSE TO ITS EQUILIBRIUM MOISTURE CONTENT SO THE HEATING
   VALUE WILL REMAIN UNCHANGED.  LIGNITE FROM THE SAVAGE MINE HAS BEEN FOUND BY PAULSON ET AL. (1972}
   TO CONTAIN 28.2% MOISTURE; THE HEATING VALUE CALCULATED ON THIS BASIS WOULD BE 7346 BTU/lb.

                                                                                              a\

-------
                                                                  68





IV-1.5  HEAT OF COMBUSTION



     Rank classification of coal, below 69% daf carbon content,



is based on the heating value and the agglomerating and aggluting



properties of the coal.  The heating value is the high heating



value of the coal (water as a liquid) on moist mineral free basis.



The mineral matter is usually calculated based on the ash and sulfur



content of the coal by empirical equations, most frequently the Parr



formula, but is best obtained by use of low-temperature ashing.



     The heat of combustion was measured on a dry sample of the



Montana lignite and corrected to a moist mineral free basis using



the mineral matter determined by low temperature ashing and either



the equilibrium moisture content or the moisture content found by



Paulson et al (1972), for samples from the same mine.  The heating



values are in reasonable agreement with heating values previously



calculated using the Dulong formula.  The measured heating value



is reported in Table IV-1 as 8809 BTU/lb with equilibrium moisture



content and 7346 BTU/lb with mine moisture; the latter value clas-



sifies the coal used in this study as lignite-A.

-------
                                                                 69
IV-1.6  ELEMENTAL ANALYSIS

     Ultimate analysis of the coals used in this study were

performed at the Galbraith Laboratories, Inc., Knoxville,
              
Tennessee and are reported in Table IV-1 on an as received basis

and in Table IV^-2 on several common bases.  Sulfur form analysis

is also reported in Table IV-1.

     The Montana lignite contained 73.3% carbon on a dry mineral

matter free basis, 0.83% nitrogen and 0.98% sulfur on a dry basis.

Approximately 65% of the sulfur is organic sulfur.  Oxygen was

determined by difference and is not considered very accurate.  Oxygen

is determined by substracting ash and all other elements from 100%.

This procedure can lead to rather substantial errors in the oxygen

concentration because mineral matter changes weight when sulfides

and carbonates react during ashing.  The procedure gives 22% too

high an oxygen concentration for the Montana lignite.

-------
                                                                     70
                                 TABLE  IV-2


LIGNITE
CHARACTERIZATION OF MONTANA LIGNITE


COMPONENT
HtO
C
B
N
Pyritic S
SO*
ORGANIC S
TOTAL S
ASH
0(DIFF.)
FROM

ULTIMATE
AIR
13.60
54.90
4.96
0.72 .
0.11
O.18
0.55
O.84
7.84
17.14
THE SAVAGE MINE IN RICHLAND COUNTY,
GROUND AND CLASSIFIED TO 38-45y
ULTIMATE ULTIMATE ULTIMATE ULTIMATE
DRY daf dmf* C-H- BAS?.S
 *
63.54 69.88 73. 3O 71. 3O
5.74 6.31 6.62 6.44
0.83 0.92 1.23
0.13 - - . -.
0.21
O.64 . O.73 . 0.73
0.98 - - -
9.07 . -
19.84 22.16 18.12 22.26
*  MINERAL MATTER FOUND FROM LOW TEMPERATURE ASHING BY PADIA  (1976)- TO



   BE 11.50% (WET BASIS)

-------
                                                                 71
IV-1.7  ASH ANALYSIS



     The mineral elements of ash (Table IV-3) were determined by



atomic absorption of the metallic elements, reported as if the



elements are in their natural oxidation state.



     The principal component of mineral matter in coals is common



clay  such as SiO- and A12O3.  The ash from a Montana lignite



contained 27% silica and 16% alumina.  The ash also had high



concentrations of calcium (28% as CaO), magnesium (9% as MgO), and



a surprisingly large amount of sulfur (13.5% as SO-).  Iron and



phosphors  are  the  only other significant elements; Fe2O^ accounts



for 3.4% of the ash and P25 acounts for 1%.  Titanium, potassium,



and sodium are the remaining major elements in the lignite.ash.

-------
                                                         72
                      TABLE IV-3
                ASH CHARACTERIZATION
Wt.% on
ignited
basis
Ash
P*0,
SlO*
FcPi
AlaOi
TiOx
CaO
Hgo
80s
KtO
NatO
Montana
 Lignite-A
7.84
1.02
26.80
3.41
16.41
0.50
28.44
9.02
13.45
O.35
0.27
PlttSDurgn
High Voli
Bituminoi
10.55
0.30
. 37.52
29.34
19.15
O.57
4.65
O.73
4\10
- 1.29
O.38
Analysis by Galbraith Laboratories, Inc., Knoxville,
Tennessee and Padia (1976).

-------
                                                                  73




IV-1.8  FUNCTIONAL GROUP ANALYSIS



     No functional group analysis has, as yet, been performed on



the coals used in this study, nor has a functional group analysis



of the ligriit* used in this study been found in the literature.



Tingey and Morrey have compiled average functional group analyses



for various rank coals from literature values.  Values for the



functional group composition for a lignite of 70.6% carbon will be



used for comparison purposes until direct measurements become



available.



     Tingey and Morrey's functional group distribution is shown in



Table IV-4.  A lignite can be expected to have an average ring size



of about 2-4, 27% of the carbon will be aliphatic carbon and 65%



of the hydrogen will be attached to aliphatic carbons.  Most of the



carbon (51.7%) in such small ring systems is peripheral carbon.  The



aliphatic hydrogen is mostly attached to aliphatic carbon that is



attached either to o or 3 position to an aromatic ring; hydroxyl



hydrogen accounts for only 4.6% of the hydrogen.



     Oxygen concentration and distribution are not included in the



functional group summary presented by Tingey and Morrey.  Oxygen



concentrations can be estimated by difference, after allowance for



a concentration of nitrogen and sulfur of about 2%.  The oxygen



concentration obtained by difference is 23%, which is a high but



reasonable value for an American lignite.  Hydroxyl oxygen may then



account for a third of the oxygen present with most of the remaining



oxygen present as carbonyl oxygen.  These results are consistent



with generalized analysis reported by Dryden (1963).

-------
                                                             74
                         TABLE IV-4
                       i  ^^^^^^^^^^^^M^

          FUNCTIONAL GROUP ANALYSIS OF A LIGNITE

               From Tingey and Morrey (1973)
                     and Dryden (1963)
                        Hydrogen
                      Distribution
                      % of  total H
   Carbon
Distribution
% of total C
Total daf
Average Ring Sizes
Aromatic
  Honoring
  Condensed Ring
  Peripheral C
  Interior C
Aliphatic
  Methylene Bridges
  a CHX
  0 
-------
                                                                 75



IV-2.0  PITTSBURGH SEAM 18 BITUMINOUS COAL


     The Pittsburgh bituminous coal was chosen as a coal


representative of a deep mined, high sulfur, caking eastern


coal.  This coal is highly swelling and caking and is repre-


sentative of coals that might be used in coking operations or


in direct fired utility boilers.  United States reserves of


bituminous coal represent about 46% of total coal reserves or

                      4
approximately 9.6 x 10  quadrillion BTUs.

-------
                                                                  76
IV-2.1  VISUAL APPEARANCE OF PITTSBURGH BITUMINOUS COAL



     The bituminous coal was ground and classified in the same



manner as the lignite.  Scanning electron micrographs of the



classified bituminous coal are shown in Figure IV-5.  The electron



micrograph at low magnification shows, as for lignite, irregular



shaped particles and poor classification.  There are more fines



present in the bituminous coal than in the lignite, and this is



thought to be because the bituminous coal has much lower equilib-



rium moisture content and exhibits more pronounced static attraction



The bituminous coal, when viewed under higher magnification, appears



to have a hard, smooth surface with some visible fracture patterns.



The lignite had very few macropores but the bituminous coal appears



to be almost entirely free of such large external openings.  A few



micron-size ash particles were observed clinging to the bituminous



coal particle surface.



     Polished surfaces of bituminous coal, shown in Figure IV-6,



viewed under cedar oil exhibit similar petrographic elements to



those observed earlier for lignite.  The predominance of a low



reflecting element, most likely vitrinite, and two elements that



reflect more light are seen in the top micrograph.  The larger and



darker of the two highly reflecting elements appears similar to the



element identified as micronite in the lignite; the smaller highly



reflective element appears to be a rare fusanite element.  The



macropore structure of the micronite elements appears to be less



pronounced than was found for the lignite, but the lower reflective



elements in bituminous coal appear to have greater macropore

-------
..W Pi'TSBUPGH SEAM 8  BITUMINOUS COAL
                    38-45um

                '  isrfp
                y^WrN?
                 ^      /*
                 0          400um
                      100X
     0          40UFT
         1OCK


3C
  FIGURE VI-5. Scanning electron micrograph of raw Fittsbur

            seam #8 hvA bituminous coal.

-------
                                                   78
POLISHED PITTSBURGH SEAM  #8 BITUMINOUS  COAL
              under ceder oil  450 X
                        50 urn

 FIGURE VI-6.  Photomicrograph of polished raw Pittsburgh
             seam #8 hvA bituminous coal.

-------
                                                                 79
structure than the lignite.  As in the lignite, most particles are



petrographically pure/ although several petrographic forms are



occasionally observed; for example, three different reflectivities



are evident in one particle in the bottom micrograph.

-------
                                                                 80





IV-2.2  PARTICLE SIZE DISTRIBUTION



     Mass particle size distribution, calculated in the same manner



as for the  lignite gave  a  poorer fit of the data for bituminous



coal than for lignite, even though more than three times as many



particles were counted.  The fit of cumulative particle mass by



the Rosin-Rammler distribution function is shown in Figure IV-7.



The mass average particle size determined from the cumulative



distribution is 64.9 microns, while average particle diameter



determined from the raw data is 57.5 microns and the standard



deviation is 22.2 microns.  The average particle size determined



from both methods is larger than the maximum screen size of



43 microns.  A 43 micron opening could allow a particle with a



maximum dimension in at least one direction of about 60 microns



to pass.  Narrow particles with two dimensions less than the



maximum size opening could pass through the screen with a very



large size in one dimension.  Electron micrographs were checked



and a few odd shaped particles appeared in the pictures.  A few



large particles can skew a mass average distribution to much higher



mass average particle sizes.  Another possibility that cannot be



completely eliminated, but which could account for the large



particles present in the classified sample is rents or enlarged



holes in the screens.  The original data and the derived Rosin-



Rammler distribution curve are shown in Figure IV-8.  The agreement



is considered satisfactory with the reservations previously mentioned



in the section on lignite distribution.

-------
                         RAW BITUMINOUS  COAL
                   ROSIN -RAMMLER   DISTRIBUTION
    5   10' |=
    o
    Id

    8
    UJ
    >

    <

    2
    D
    U
       r   i  i n ni
        i

       X 64.9
       s   2.54

       b -  2.49 xlO5

   ^   2574  Particles
        101
102
         103
T  T~l I I II

                              I I bllll
 I  1  I  I MM
           10
                   10 1             10 2

                  DIAMETER ( MICRON)
           10
                                                                             00
FIGURE  IV-7. Roein-Rammler accumulative size distribution of bituminous coal

            particles.

-------
                             RAW  BITUMfNOUS

                       FREQUENCY   DISTRIBUTION
     ce
     <
     ce
     I-
     5
     o:
     U
     z
     U
     3
     O
     LJ
     cc
     u.
         30
         25
20
          15
10
             I
x57.5
(T22.2

s n 2.54

b- 2.49X105

2574 Porticlcs
                                I
                                 I
                      20
                      40        60

                    DIAMETER (MICRON)
                                    80
100
FIGURE  IV-8. Bituminous particle size frequency distribution.
                                                                             00
                                                                             to

-------
                                                                 83



IV-2.3  SURFACE AREA


     Surface areas, determined using krypton adsorption, are as in


the case of lignite significantly below the true surface area,


representing only pores with restrictions greater than about 12 A.


Calculations of the surface area from krypton adsorption data at

                                                  2
77K using the Dubinin-Polanyi equation gave 2.1 m /g; the BET

                    2
equation gave 1.25 m /g. Can et al (1972) measured the C0_ adsorp-


tion area for a Pittsburgh seam bituminous coal with 82.4% carbon

                                        2
(dmmf) and found a surface area of 141 m /g.  This value is con-


sistent with the value of about 150 m /g obtained from neon


adsorption at 298K for a coal with 83% C  (Dryden, 1963) .

-------
                                                                  84





IV-2.4  PROXIMATE ANALYSIS



     Proximate analyses for the bituminous coal of this study is



reported on an as received basis in Table IV-1.  The Pittsburgh



seam 18 bituminous coal has a low equilibrium moisture content



of about 2.2%.  This value is taken to be representative of mine



samples.

-------
                                                                 85




IV-2.5  HEAT OF COMBUSTION



     The heat of combustion of the Pittsburgh Seam #8 bituminous



coal, determined in the same way as that of lignite, is 14,377



BTU/lb on a moist mineral matter free basis.  This value is con-



sistent with other values reported for Pittsburgh seam coals and



classifies this coal as a high volatile A bituminous coal.

-------
                                                                  86





IV-2.6  ELEMENTAL ANALYSIS



     Elemental analysis of the bituminous coal and sulfur form



analysis are presented in Table IV-1 on an as received basis, and



in Table IV-5 on other commonly used bases.  The bituminous coal



has 1.04% nitrogen and a high sulfur content of 4.55% on an as



received basis.  The major portion of the sulfur is pyritic (2.29%)



but the organic sulfur is much higher (1.58%)  than that of ordinary



plant materials.  The carbon content is 82.27% on a dry mineral



matter free basis.

-------
                                                                   87
                                 TABLE IV - 5
BITUMINOUS COAL
CHARACTERIZATION OF PITTSBURGH SEAM f8
BITUMINOUS COAL FROM THE IRELAND MINE
GROUND AND CLASSIFIED TO 38-45y
COMPONENT
HjO
C
H
N
Pyritic S
SO*
Organic S
Total S
ASH
O(DIFF.)
ULTIMATE
A.R.
2.20
67.90
4.91
1.04
2.29
O.68
1.58
4.55
10.52
8.88
ULTIMATE
DRY
-
69.46
5.01
1.06
2.35
0.65
1.52
4.53
10.76
9.19
ULTIMATE ULTIMATE
daf dmf*
-T.
77.80 82.27
5.62 5.95
1.19 1.26
-
-
1.81 1.91
-
-
13.58 8.61
                                                                ULTIMATE
                                                               C-H-O  BASIS
                                                                   83.31

                                                                    6.01
                                                                  10.87
*  Mineral matter found from low temperature ashing by Padia  (1976) to be
   15.27% (wet basis).

-------
                                                                 88





IV-2.7  ASH ANALYSIS



     The ash composition of the two coals are compared in Table IV-3



The bituminous coal has a higher ash content than lignite, about



10.5 weight percent versus about 7.8 percent.  A very large portion



of the bituminous coal's ash is composed of iron compounds, mostly



present as pyrites (Padia, 1976) .  The high pyrites content of the



mineral matter accounts for the major portion of the sulfur content



of the bituminous coal.  Iron, calculated as ferric oxide, accounts



for 30% by weight of the ash compared with 38% for silica and 19%



for alumina.  Calcium, residual sulfur trioxide and potassium are



the only other constitutents of consequence in bituminous ash.

-------
                                                                 89
IV-2.8  FUNCTIONAL GROUP ANALYSIS



     Functional group analysis has not as yet been performed on



the bituminous coal sample of this study.  A partial functional



group analysis has been reported in the literature on a vitrain



concentrated sample of a Pittsburgh seam bituminous coal.  Heredy



and Fugassi (1966) 'investigated a benzene extract for hydrogen



distribution by proton NMR for a vitrain concentrate from the



same mine  (Ireland) as the bituminous coal used in this study.



Their results showed that 16.3% of the hydrogen was substituted



on condensed aromatic rings, 14.7% of the hydrogen was attached



to either monocyclic aromatic rings or phenolic oxygen; 69% of the



hydrogen was attached to aliphatic carbons, 29.6% to o carbons and



39.4% to 3 carbons.  These figures agree only fairly with the



figures reported in Table IV-6 and have not been considered in



constructing Table IV-6 since the benzene extract accounted for



only about 10% of the total coal.



     Retcofsky and Friedel (1970) have also investigated functional



group distribution of a vitrain rich sample of the coal from  a



Pittsburgh seam bituminous coal.  Their findings are consistent



with the results of Tingey and Morrey reported in Table IV-6.



Retcofsky and Friedel studied high resolution proton NMR of pyridine



and carbon disulfide coal extracts.  The pyridine extract contained



21.4% of the coal and had an elemental analysis close to that of the



whole coal; the carbon disulfide extract contained only 2.5% of the



coal and had significantly higher carbon and hydrogen contents and



lower oxygen content than the original coal.  Aromatic carbon

-------
                                   90
TABLE IV-6
FUNCTIONAL GROUP ANALYSIS OF
A HIGH VOLATILE BITUMINOUS COAL
From

*
Tingey and Morrey,
and Dry den, (1963)

(1973)


Hydrogen Carbon Oxygen
Distribution Distribution Distribution
% of total H % of total C % of total 0
Total daf
Average Ring Size
Aromatic
Honoring
Condensed Ring
Peripheral C
Interior C
Aliphatic
Methylene Bridges

-------
                                                                 91
accounted for 73% of the carbon in the pyridine extract, 35% of



the hydrogen was aromatic, the o aliphatic hydrogen accounted for



26% and B  hydrogen accounted for 36% of the total hydrogen.



Hydroxyl hydrogen accounted for only 3% of the total hydrogen.



The values derived from the carbon disulfide extract were similar



although a higher percentage of aliphatic compounds seem to be



present in the carbon disulfide extract.



     The average ring size of a high volatile bituminous coal



with 83% carbon content may be 5-8; the distribution of the



carbon may be about 70% aromatic, 30% aliphatic, with only 6% of



the carbon in methylene bridges; the distribution of the hydrogen



may be 20% aromatic, 75% aliphatic and about 5% in the phenolic



OH; the oxygen may be split equally between hydroxyl and carbonyl



oxygen with negligible amounts in carboxyl and ether groups.

-------
                                                                  92
IV-3.0  PHYSICAL PROPERTIES OF COAL



     Estimates of some of the physical properties of coal have



been used in this study.  These values are not very accurate and



have been used only for crude calculations.  The values of interest



are specific gravity, thermal conductivity, and specific heat.  The



interpretation of these values for a substance such as coal that



is very heterogenous and changes both chemically and physically



when mildly heated is questionable.  Even when no changes occur



in the coal the values are uncertain because the values strongly



depend on coal rank, mineral matter content, and moisture content.



     Average values of the physical properties needed have been



taken from McCabe and Boley (1945), Badzioch et al^  (1964) and



Kirov (1965).  Specific gravity will increase roughly 0.01 unit



for a percent increase in ash content.  Thermal conductivity of


                                        4          -4
raw coals below 400C varies from 5 x 10   to 8 x 10   cal/g-em-sec-K



Above about 300-400C, where carbonization reactions begin, the



thermal conductivity rises rapidly until at 900C (the maximum



temperature in Badzioch e_t al's study) it has a value of about


       -4
60 x 10   cal/g-cm-sec-K.  This high value in all probability repre-



sents conductive heat transfer augmented by radiative transfer



through pores and conveotive transfer by volatile products flowing



through cracks and pores.  The latter two effects are difficult to



separate from the former experimentally.  The thermal conductivity



of the particle is needed for purposes of calculating the rate of



temperature rise in the particle.  Since the resistance to heat



transfer by conduction within the particle is negligible relative

-------
                                                                  93
to the external resistance, uncertainty in the value of the thermal



conductivity is of little significance.



     Measurements of specific heat will also be effected by the



heterogeneous nature of coal and by the physical and chemical



changes coal goes through as it is heated.  The specific heat of



raw coal increases with increasing volatile matter and increasing



moisture content and decreases with increasing ash content.  The



specific heat of raw coals with less than 10% volatile matter



(d.a.f.) will be between 0.2 and 0.25 cal/g-K, while coals with



volatile matter  (d.a.f.) between 10 and 50% will have specific



heats between 0.25 and 0.35 cal/g-K. Coals with volatile matter in



the range of 40% (d.a.f.) will have specific heats around 0.3 cal/g--K.



     Char residuals (based on weight of raw coal) could have a



maximum specific heat of 0.42 cal/g-K at 320C for a char from a



coal with 50% d.a.f. volatile matter.  The specific heat of chars



produced from coals with volatile matter (d.a.f.) greater than 30%



have a specific heat maximum in the region of rapid devolatilization,



The specific heat then decreases towards an asymptotic value at



temperatures above 1100C.  The position of the specific heat maxi-



mum is displaced towards higher temperatures as the volatile matter



of the coal decreases.  Chars produced from coals with d.a.f. vola-



tile matter in the range of 30-40% would be expected to have a



specific heat value of about 0.33 cal/g-K based on the weight of



the raw coal.

-------
                                                                  94
IV-4.0  CHARACTERIZATION OF OTHER COALS AND CHARS



     Most experiments were run using the carefully characterized



lignite and bituminous coals described in Sections IV-1.0 and



IV-2.0.  Several additional experiments were performed using a



Montana sub-bituminous coal.  Char oxidation experiments utilizing



chars prepared from the Montana Lignite were also performed.



Proximate and ultimate analyses of these additional samples are



shown in Table IV-7.

-------
                         TABLE  IV-7.CHARACTERIZATION OF COALS AND CHARS
^^M^MMiHflVMM^HHMBMnMI^^ManBI^MMHMH
Type*

Montana Lignite-A
Montana
Sub-bituminous**

Proximate Ultimate
Analysis Analysis
A.R. WtZ A.R. WtZ
Specification VM H,0 Aali C U M S 0
z (by Diff .)

Savage 36.20 13.60 7.80 54.40 4.96 0.88 0.84 17.42
Powder 35.16 21.23 9.34 53.26 3.35 0.87 0.78 11.17
River
Region
1250 K
  Lignite Char

1750 K
  Lignite Char

1750 K
  Lignite Char***
MM*     4.20  .12.08   66.94  2.09   1.02   0.56   13.11
MM      2.02   18.03   76.42  0.69  "0.55   1.06    1.23
MM      2.10   19.00   76.11;  0.36   0.58   1.11    0.74
"MM - Mot Measured

**
  Supplied by Aero therm Division of Acurex Corporation
***
  "Used in time resolved measurements.
                                                                                                         vo
                                                                                                         in

-------
                                                                 96
                           SECTION V



                          COAL RESULTS





     This section reports the results obtained on evolution and



oxidation of nitrogen contained in coal using the equipment and



procedures described in Section III.  The results described



include nitrogen loss from coal held in crucibles, nitrogen loss



from coal rapidly heated in the dispersed phase, and conversion



of coal and char nitrogen to nitric oxide when burned in a small



coal flame.



V-1.0  THE COAL FLAME- MOTIVATION FOR THE EXPERIMENTS



     Coal supplies slightly over half of the fossil fuel burned



to generate electric power.  Its use can be expected to increase



as the other fossil fuels become less abundant and more costly.



Most of the coal burned for power generation is burned in pul-



verized coal flames in large units (hundreds of megawatts



electrical capacity).  Combustion of coal yields higher emissions



of sulfur dioxide, particulate matter, and nitrogen oxides than



other fuels.  The data reported in this section were taken in an



effort to understand some of the complex aspects of the oxidation



and reduction of the nitrogen contained in the coal with the



ultimate goal of investigating strategies  for  control of the



fuel nitrogen contribution to total nitrogen oxide emissions.



     A schematic of a pulverized coal flame is shown in Figure V-l



Coal is ground to less than 200 micron particles  (most particles

-------
                          COAL  DSVOLATILIZATION  AND COMBUSTION
           HEATING DEVOLATILIZATION     VOLATILE BURNING HCTEROGENOUS COMBUSTION
   SECONDARY AIR
                                     CO,.H,0
    PRIMARY AIR

     COAL
  COAL
PARTICLE
     SECONDARY AIM




     TIME SCALE  2MSEC


     EXPERIMENTS    INERT
                   DCVOLATILIZATION
                             DIFFUSION
                           S FLAME
                     100 MSEC
                  COAL OXIDATION
 300 MSEC.
CHAR OXIDATION
FIGURE   V-l.   Schematic of a single coal particle burning in a pulverized

                 coal flame.

-------
                                                                 98






are less than 100 microns), and fed through the primary injection



nozzle.  A fraction of the total air (usually 15-20% of the



stoichiometric air requirements) is used to carry the coal into



the flame region.  A large 500 mw utility boiler typically feeds



about 200 tons/hr coal and 3300 tons/hr air.  If four burners



are used, each burner will be fed 50 tons/hr coal and about



50 tons/hr air through the primary nozzle.  The remaining air is



preheated and fed through swirl vanes surrounding the primary



nozzle.



     The coal particles exit the nozzle in a slightly fuel rich



atmosphere.  The reducing nature of this atmosphere is increased



further by recirculation of partially combusted gases from the



flame region.  The coal particles are heated, possibly as high



as 2200K, by the hot recirculating gases in about one millisecond.



The coal particle rapidly loses volatile matter over the next tens



of milliseconds.  This volatile matter, depending on the velocity



and uniformity of ejection, can exclude oxygen from the particle



surface and burn as a diffusion flame surrounding the particle.



The volatile matter may not be able to form an effective oxygen



shield in all cases, and the particle will simultaneously undergo



gas phase volatile matter burning and heterogeneous solid phase



combustion.  Burning of the volatile matter is usually complete



in about 100 milliseconds.



     Volatile burning will leave a char residue, even if the solid '



has been burning simultaneously with the volatile matter.  The



char will be heterogeneously oxidized in about 300 milliseconds.

-------
                                                                 99
     Experiments that might be used to isolate the different



regions of coal combustion are listed at the bottom of Figure V-l.



Volatile matter removal can be investigated by devolatilizing coal



under conditions of reduced oxygen concentration which supress



ignition of the particle.  Devolatilization of coal particles



under inert atmospheres will be reported in this section.  Coal



oxidation under the condition when the coal particle is surrounded



by a diffusion flame of the volatile matter can be investigated



by burning the coal particle in a dilute phase so that particles



have little interaction with each other.  Only overall oxidation



experiments where the coal particles burn as small diffusion



flames are reported in this section.  Char oxidation kinetics can



be investigated by manufacturing the char in an atmosphere where



the coal particles will not ignite and then reinjecting the coal



particles into an oxidizing atmosphere where the particles burn



heterogeneously as individual particles.  Some results from char



oxidation are reported in this section.  The gas phase reactions



of the products produced from coal devolatilization and partial



combustion, descriptions of which are needed to complete the



picture of the oxidation of the nitrogen content in coal, are best



studied in simplified systems such as flat flame burners, and do



not constitute part of the present study.

-------
                                                                  100




V-2.0 DEVOLATILIZATION RESULTS



     Coal pyrolysis studies include asymptotic  elemental and



total weight loss from crucible held samples as well as time-



resolved weight loss experiments conducted in the flow reactors



for residence times up to one second.  Limited analyses of gas-



phase pyrolysis products are also presented in an attempt to close



the carbon and nitrogen material balances.



V-2.1 DEVOLATILIZATION RESULTS FROM CRUCIBLE EXPERIMENTS



     The purpose of the crucible experiments is to investigate



the effect of temperature on the behavior of coal devolatilization



in inert atmospheres at long residence times.  The nature of the



experiments precluded the attainment of the rapid heating con-



ditions typically found in pulverized coal flames.  Heating rates



were typically on the order of several degrees Kelvin per second.



The data reported here are for times at which the coal weights



had reached their asymptotic values.



V-2.1.1  PITTSBURGH SEAM #8 hvA-BITUMINOUS COAL



     Figure V-2 shows the retention of various elements in the



coal and ultimate weight loss for crucible held coal as the



temperature increases from 600 to 2200K.  The chars were analyzed



after times considered to represent asymptotic weight loss and



are represented on a dry basis.  The weight loss for temperatures



above about 1200K is asymptotic at about 40%, below 1200K the



weight loss drops.  The weight loss observed after 12 hours at



600K is only about 10%.  Carbon loss levels off at 1200K, when



the weight loss is about 30%.  Hydrogen is initially removed more



slowly than oxygen but after both oxygen and hydrogen contents of

-------
   100 -
z
o
I-
z
UJ

UJ
o:
                            I	1	1
                    HVA  Bituminous Cool
                    ASTM  Prox. Char
          600    800
1000  120O  1400  1600  1800  2000

    MAXIMUM   TEMPERATURE ( *K )
 FIGURE V-2.  Crucible devolatilization of Pittsburgh Seam #8hv A-bituminous

             ooal.

-------
                                                                  102
the char have dropped to about 40% of their original values, the



rate of fractional loss of hydrogen and oxygen become



approximately equal.  Oxygen, hydrogen, and nitrogen are com-



pletely removed from the char at high temperatures,  under the



conditions of asymptotic weight loss.  Nitrogen is much more



stable in the char than either hydrogen or oxygen.



V-2.1.2 MONTANA LIGNITE-A



     Element loss from the lignite samples held in crucibles is



shown in Figure V-3.  The retention trends of the various elements



are very similar to the trends observed for similar experiments



using the bituminous coal.  The proximate analysis volatile matter,



dark symbols, agrees well with the ultimate weight loss obtained



by slowly heating coal samples.  Carbon loss and total weight loss



increase only slightly above the proximate analysis temperature of



1223K.  The asymptotic carbon loss was about 20% although the



carbon loss appeared to increase slightly as the devolatilization



temperature increased above 1700K.  The char showed no additional



weight loss above about 1600K.  The asymptotic weight loss was



44% which is only slightly greater than the proximate volatile



matter of 41%.  Little weight loss of carbon occurs at low temper-



atures .



     Oxygen, hydrogen, and nitrogen are completely removed at



high temperatures.  Hydrogen and oxygen loss as the percent of the



element originally present are very nearly the same.  Initially



there is approximately five times as much hydrogen present as



oxygen on a mole basis.  Hydrogen and oxygen contents are reduced

-------
                      i	1	r

                      LIGNITE

                      ASTM  PROX  CHAR
   100
-7  80
 *
O   60
LJ

UJ
o:
    20
         T
          D
          1
 D
a
    I
1
         600   800
         1000   1200   1400  1600   1800   2000

             MAXIMUM  TEMPERATURE (*K)
  FIGURE V-3.   Oruoible devolatilization of Montana lignite-A.
                                                                                o
                                                                                LJ

-------
                                                                 104
to low values at temperatures in the range of 1600K and are both

reduced to zero above 1800K.

     Nitrogen is, again, the most refractory of the elements that
                                                         
are completely removed from coal.  The lignite coal used in this

study has approximately 28% of the original nitrogen removed

under proximate analysis conditions.  This value agrees well with

the amount of nitrogen removed under the crucible tests of this

study.

-------
                                                                  105





V-2.2  DEVOIATILIZATION OF DISPERSED COAL PARTICLES



     Experiments were conducted on devolatilization of coal



falling freely through an inert gas surrounded by hot walls.   The



purpose of this get of experiments was to obtain devolatilization



data for coal that is rapidly heated (^10 K/sec) in a dispersed



phase for relatively long residence times, on the order of one



second.  A slight modification of this experiment was used to



gather data on rapidly heated coal devolatilization at longer



residence times by dropping the coal into a crucible where



further devolatilization of the char  occurs on a sample held at



the maximum temperature of the experiment.



V-2.2.1  PITTSBURGH SEAM 8 BITUMINOUS COAL



     Data gathered on element loss from the bituminous coal under



a variety of experimental conditions, including free fall condi-



tions, is shown in Figure V-4.  The free fall data has a residence



time of approximately one second and the free fall into a crucible



data has residence times varying from 3 x 10^ to 6 x 105 seconds.



The laminar flow experiments gave devolatilization residence time



from several milliseconds to a maximum of about 300 milliseconds.



The long residence time data is replotted from the crucible exper-



iments discussed in Section V-2.1.1.



     The retention of the various elements are plotted in



Figure V-4 versus time in milliseconds with furnace temperature



as a parameter.  A log time scale is used so that the data can be



represented on one plot.  The retention of each element (weight



percent of the original element retained after devolatilization)

-------
                                                                 106
   100
 O 50
   TOO
 X
  50
*  0
 too
 u
 *
 50

 .0
100
 *
 or
   50
  0
10O
 g 50

 0
                PERCENT OF ELEMENT  RETAINED  IN  CHAR
                PITTSBURG  .SEAM  & 8  BITUMINOUS  COAL
                                        JL
                                                  JL
                                          r
                                           MQK. Tamp,
                                         Y610*K     4
                                                  JLL
                                                    ir
                                         JL
                                                  JL
                                                    ir
                                         JL
                                                  JL
                 1O          1O2
                  Lominer Flow        Free Foil    .   Crucible
                  RESIDENCE  TIME   MILLISECONDS
PIGURE V-4.    Element loss with devolatilization of a
               Pittsburgh Seam #8 hvA~bituminous coal.

-------
                                                                 107





is plotted with a common time axis to facilitate comparisons of



the stability of the different elements in the coal matrix.  The



solid lines extended past one second represent retention of the



element for coal that has been dropped through the furnace, caught



in a crucible, and held at furnace temperature for the time



indicated.  The dotted lines connecting two points represent the



change in retention with additional devolatilization when the



char produced at the shorter residence time is reinjected into



the furnace and held at the same temperature for additional time.



     An asymptotic carbon retention of 75% was approached at



300 milliseconds for 1260K, of 65% at 100 milliseconds for



1510K, and of 60% at 50 milliseconds for 1740K.  Although the



total weight loss appeared to attain an asymptotic value of 52%



dry at 1940K and 32 milliseconds and 57% dry at 2100K and



21 milliseconds, the carbon loss does not appear to be complete.



This implies that further slow carbon loss might be obtained at



longer residence times at these high temperatures.  The carbon



retention was about 50% at 32 milliseconds and 1940K and about



40% at 21 milliseconds and 2100K.  Kobayashi (1976)  presents a



more complete discussion of the same data.



     Hydrogen is relatively easy to remove from coal during



pyrolysis.  Below 1260K, hydrogen loss is relatively rapid until



about 50% of the hydrogen has been removed; the remaining hydrogen



is removed much more slowly.  However, hydrogen can be almost



completely removed at 1260K for residence times on the order of



ten minutes.  At high temperatures the complete removal of hydrogen

-------
                                                                  108
is rapid, 21 milliseconds at 2100K.  Large quantities of hydrogen



are still retained in the char after pyrolysis at 740K for 22



hours.  Slow evolution of hydrogen probably continues for even



longer periods at 740K, unless further evolution of hydrogen is



constrained by thermodynamics.



     Nitrogen is the most difficult element to ultimately remove



from the coal matrix after carbon.  At low temperatures, around



700K. nitrogen loss is very small; at 1260K nitrogen loss is



also small but can be seen to continue at a slow rate by the two



square points joined by dotted lines.  Nitrogen removal is com-



plete at temperatures above 1500K at long, but reasonable



residence times, as indicated by the data around five minutes



residence times.  Nitrogen removal by pyrolysis and oxidation is



the major concern of this study and will be discussed in more



detail from an expanded plot of nitrogen removal.



     Sulfur is not as difficult to remove from coal as carbon and



nitrogen, but is more difficult to remove than hydrogen and oxygen.



The data indicates that at temperatures below about 1900K sulfur



removal is slow once 50% of the sulfur has been removed.  The



amount of sulfur easily removed from coal is apparently tempera-



ture dependent.  The asymptotic retention of sulfur in the char



at 1260K is about 70% but at 1510 and 1740K is about 40%.  Sulfur



was almost completely removed at 2100K and 21 milliseconds.



     Oxygen appears to be slightly easier to remove than hydrogen



at higher temperatures but somewhat more difficult to remove at



lower temperatures.  All the oxygen can be removed at 1260K and

-------
                                                                 109
ten minutes residence time or at times less than 7 milliseconds
at 1940K.  Slow evolution of oxygen at 1260K is evident from
the decrease in oxygen concentration between the two points
connected by a dotted line.

-------
                                                                 110
V-2.2.2  MONTANA LIGNITE



     Data on the element retention of devolatilized lignite



(shown in Figure V-5) used in this study versus residence time



at different furnace temperatures is similar to the results for



the bituminous coal just presented.  The release of hydrogen



appears to be more rapid under certain conditions for lignite



than bituminous coal; all the other elements showed a similar



behavior for bituminous coal and lignite.  The curves for element



retention are, in general, similar in shape and magnitude to the



retention curves developed for bituminous coal.



     Carbon loss appears to become asymptotic for all temperatures



at about 100 milliseconds.  The asymptotic carbon retentions are:



75-80% at 1260K, 65% at 1510K, 55-60% at 1740K, -^50% at 1940,



and 'x40% at 2100K.  Little additional carbon appears to be re-



moved under conditions of low temperature and long residence times.



     Hydrogen loss is complete at either moderate temperatures and



long residence times or high temperatures and short residence times.



For example, hydrogen was completely removed from char devolatilized



at 1750K for five minutes or at 2100K for 30 milliseconds.



Hydrogen is more easily removed from lignite than from bituminous



coal at temperatures below 1510K; above this temperature the rate



of hydrogen removal from the two coals appears to be approximately



equal.



     Nitrogen, as for bituminous coal, is the element most resistant



to ultimate devolatilization after carbon.  The rates of nitrogen



removal from bituminous coal are slightly faster than from lignite



at all temperatures.  The difference is, however, not great.  The

-------
                                                               111
  10O
 *
 8 50
 u
 u
 O
 8
 O
  O

,100


 50


  O

 100
 u
 O 50
 2
    O

   100


   50
   100
g  50
i.
                  PERCENT OF  ELEMENT  RETAINED  IN  CHAR
                            MONTANA  LIGNITE
                                               OQ
                                        J I
                                        i i
                                            Mpx. Temp.
                                            io'K
                                          4720-740*K  O200*K
                                          OlOOO'K
                                          OI220-I26O*K
                                           O740-1760*K
                                        J I  CS
                                              0-8
                                         I I
                                        1 I
                                             fi
                                                 Cructbl>
                                                  10'
                    RESIDENCE  TIME  MILLISECONDS
                                                           10'
FIGURE  V-5.    Element  loss with devolatilization of a
                Montana  lignite-A.

-------
                                                                 112





nitrogen removal from both coals  is discussed in more detail



later.



     The percent removal from the two coals of sulfur show similar



behavior, with the exception of the greater sulfur removal at



1260K from the bituminous coal.  As for bituminous coal, sulfur



removal is relatively slow.  Indications are that sulfur can be



completely removed at high temperatures but can not be effectively



removed at low temperatures.  Asymptotic retention of sulfur at



temperatures below 1510K is approximately 50%.



     Percentage oxygen removal from lignite is approximately equal



to the removal for bituminous coal at temperatures below 1510K,



but oxygen removal from bituminous coal is considerably faster at



temperatures above 1940K.  Oxygen can be completely removed within



reasonable times from lignite char at all temperatures above 1260K.

-------
                                                                 113




V-2.2.3  NITROGEN PYROLYSIS FROM LIGNITE AND BITUMINOUS COAL



     The nitrogen retention data from Figures V-4 and 5 for the



fast flow and free fall experiments is replotted in Figure V-6 with



expanded scales.  The retention data at 1260K for the bituminous



coal shows some scatter.



     There are two features of the retention curves for lignite



that also should be noted.  One datum point has higher retention



than a smooth curve through the rest of the data would indicate.



This point was obtained in an attempt to increase the residence



time of the fast flow furnace past the residence times that the



furnace yielded when operated in the normal mode.  The plasma gun



could not be used to heat the large quantity of gas used to carry



the particle at the low flow rate required; it is, therefore, likely



that this point had a significantly lower temperature than the rest



of the data at 1510K.  Two data points taken at a temperature of



1740K and short residence times show approximately the same small



nitrogen loss.  The nitrogen content of the original coal has been



checked many times but, the nitrogen content determined by the



Dumas method gave consistantly lower results than the nitrogen



content determined by the Kjeldahl method.   This can only be caused



by incomplete reduction of nitrogen oxides generated during the



burning of coal in the Dumas method.  The Dumas method gave nitrogen



results about 20% lower than the Kjeldahl method for lignite, di-



gestion of the nitrogen compounds from butuminous coal is known to



be more difficult (Jodidi (1910)  and Shacklock and Drakeley (1927))



so the nitrogen content of the bituminous coal was measured using

-------
                 X
                 0
                 o
                 lit
                 z
                 u
                 a:
                 z
                 IU
                 o
                 o
                 oc
                            RETENTION OF NITROGEN  IN PYROLIZED
                                           CHARS
100
 90
 60
 70
 60
 50
 40
 30
 20
 10
  0
                              I
                                 LI6NI
                                     \
_L
                            BITUMINOU
I
I
                              10      100    1000
                                             1     10     100
                                 EXPOSURE  TIME (MILLISECONDS)
                                             1000
FIGURE  V -6.   Retention of nitrogen in devolatilized lignite and bituminous
              coal chars.

-------
                                                                 115





a 24 hour Kjeldahl digestion.  Use of the nitrogen contents



determined by the Dumas method gave inconsistent results for



nitrogen in the early stages of devolatilization, therefore the



higher Kjeldahl results for nitrogen content of the original coal



have been used.  Even when the Kjeldahl results are accepted there



is a residual uncertainty in the nitrogen measurements.  This



uncertainty could indicate that the two points showing nitrogen



retention of approximately 97% actually represent conditions where



no nitrogen evolution occurs; this explanation agrees with the



hypothesis on nitrogen devolatilization to be presented in the



discussion section.



     Figure V-6 shows more clearly that nitrogen is a very refrac-



tory element in coal.   At the most extreme conditions of the



experiments presented here, no more than 70% of the nitrogen has



been removed after one second.  Additionally nearly all the nitrogen



is retained in the char for an induction period that depends on



temperature.  Bituminous coal has an induction period of about



100 milliseconds at 1260K, 60 at 1510K, ^20 at 1740K, 12 at



1940K, and near zero at 2100K.  The induction times for lignite



are: 100 milliseconds at 1260Kf 30 at 1510K, 20 at 1740K,



10 at 1940K, and 2 at 2100K.  The induction times for lignite



and bituminous coal appear to be consistent considering that some



extrapolations were necessary.



     The nitrogen retention for lignite at 1510K levels off at



67% and about 100 milliseconds.  The shape of the curve appears



different when compared with the rest of the data.  However, the



asymptotic value at one second is very well established so the

-------
                                                                  116





shape of the curve must be nearly correct.

-------
                                                                 117





V-2.3  ANALYSIS OF VOLATILE PRODUCTS OF COAL PYROLYSIS



     A set of experiments was conducted in which the gas-phase



products of coal pyrolysis were partially analyzed in an attempt



to determine the fate of the carbon and nitrogen evolved from the



coal.  The lignite coal in a 38-44 micron size cut was fed into the



free-fall reactor under an inert atmosphere of helium, and with a



residence time in the hot zone of about one second.  Off-gases from



the furnaces were collected in a gas sampling bag for gas chromatog-



raphic analysis, or bubbled through scrubbers for electrochemical



analysis.  The  reader is referred to  Section III-4  for a  review



of the analytical procedures.  Results pertinent to the carbon and



nitrogen material balances are shown in Figures V-7 and V-8 respect-



fully  and summarized in Tables V-l and V-2.



     The carbon material balance cannot, of course, be closed by



the limited analyses performed, insofar as hydrocarbons other than



methane are not accounted for and tars, known to form under these



conditions are, likewise, not detected.  However, in view of



Suuberg's recent results (1977), tar and heavier hydrocarbon con-



tributions can be roughly estimated, as shown in Table V-l, yielding



a generally satisfactory closure of the carbon material balance.



HCN contributes negligibly.



     Closure of the nitrogen balance is somewhat less certain,



however, because of large instrumental uncertainties in the analysis



for N2 as well as the possiblity of evolution of nitrogen bearing



species other than those indicated (N2, HCN, NH3).  No NO was de-



tected  (detection limit: -2.5 ppm) under the conditions of the

-------
                                              118
o
CD
o:
<
U

i
O
U
o
o:
O

u.
O
U!
O
U
U
cr
LJ
OL
     o
     UJ
     I-
     cc
     U
U
   12

   10

    8
o
UJ


UJ
Q:62
     O
     U
o
UJ
    1




    4

    3
O   1
u

cc

580

5 70
o
y 60
    UJ
    o:
                    125O        15OO

                  TEMPERATURE ; K

                      FIGURE V-7.

-------
                                               119
UJ
o
o
o:
O
U
O

o:
O

U.
O
u
O
Id
U
ct
UJ
a.
      

8
g
o
cc
UJ
>

O
U
o
UJ
25h

20

 15

 1O
  8

  6

  4

  2




  8

  6
a I  .
LJ z  4
>
z    p
O    *
o


<  80


I 70

S  60

<  50

u
(
                      1250       15OO

                    TEMPERATURE  K
                       FIGURE V-8.

-------
                                                                    120
                   '  TABLE V-l

      CARBON DISTRIBUTION DURING COMBUSTION OF
            LIGNITE AT T=1750 K and =3~4
Combustion                 Weight % of the Original
 Products                     Carbon in Lignite
   CO                              72 - 75
   C02                             11 ~ 16
   CH4                              1-2
   Other CH                    Not Measured
   TAR                         Not Measured
   Char                     -  '      4-5
   Total                            88-98

-------
                                                                 121
                     TABLE V-2







     NITROGEN DISTRIBUTION DURING COMBUSTION OF



           LIGNITE AT T=1750 K AND =3~4
Combustion                Weight % of the Coal-Nitrogen

 Products                            in Lignite
   N2                                71  - 75



   HCN                                6-8



   NH3                                Trace



   N0y                                1~3
     ^\


   Char                               3~5
   Total                             81  ~ 91

-------
                                                                 122
experiment.  Furthermore, reported HCN levels may not strictly
represent primary yields of HCN  as additional experiments
provide  indication of thermal decomposition of HCN at the highest
temperature conditions in the furnace.  The results in this section
show the consequence of both primary pyrolysis of the coal and
secondary pyrolysis of the volatile products.  They are reported
to provide a measure of confidence in the data on the char through
an ability to close a material balance.

-------
                                                                  123
V-3.0  OXIDATION RESULTS
     In order to investigate the regime of volatile burning and
heterogeneous combustion depicted in Figure V-l, the burning and
conversion of coal nitrogen to nitric oxide for both raw and
devolatilized coal were investigated.  The premise of conducting
oxidation experiments on both char and coal was that if volatile
burning and char burning could be treated as nearly independent,
then the contribution of the volatiles to total nitric oxide could
be obtained by subtracting the contribution of char oxidation from
the total nitric oxide emission obtained form oxidation of the raw
coal.
V-3.1  OXIDATION OF LIGNITE AND BITUMINOUS COAL-
       FIXED RESIDENCE TIME
     Lignite and bituminous coal were burned in 21% oxygen in helium
mixtures at varying percentages of the stoichiometric oxygen re-
quirement at three different temperatures according to the procedure
outlined in Section III-2.  The unburned char was recovered and
residual combustibles and nitrogen were determined.  Nitrogen oxide
production was determined by measuring the concentration of the
gas effluent from the furnace.
     The results obtained from oxidizing raw coals at various fuel
equivalence ratios in a 1500K furnace are shown in Figure V-9.
Little difference was apparent in nitrogen converted to nitric oxide
and nitrogen retained in the char for the two coals and the limited
number of particle sizes studied.  There is some indication that the
burnout of bituminous coal may be more difficult than the burnout
of lignite.

-------
                         CONVERSION OP COAL NITROGEN
        g 100
        K  80
           60
      O
      hi

      II
      go
       o
       z
       O
       V)
       U

       8
       #
 60
 40
 20
  O

100
 90
 80
 70
 60
 50
 40
 30
 20
 10
  0
Wall Tamp.  1500 *K
Floma  Tamp 1600 *K
 A Llgnlta 75-90 pM
 A Llgnita 38-45 /j M
 a Bituminous 75-90/jM
  Bituminous 38 -
                                                J.
                               234
                            FUEL   EQUIVALENCE  RATIO
FIGURE  V-.9,  .Conversion of coal nitrogen as  a function of fuel equivalence
             ratio in a 1500K furnace.
                                                                          N>

-------
                                                                  125

     As expected, the char burnout is near 100% at lean conditions
but residual char remains as the fuel equivalence ratio approaches
stoichiometric and burnout continues to decrease until an asymp-
totic value of about 70% at a fuel equivalence ratio of 1.5 is
obtained.  Nitrogen retained in the char follows a curve compli-
mentary to the burnout curve.  When char burnout is complete there
is obviously no nitrogen retained in the char; however, when char
oxidation is incomplete nitrogen is retained in the residual char.
Nitrogen retained in the char reaches an asymptotic value of 45% at
fuel equivalence ratios above 1.75.
     The conversion of fuel nitrogen to nitrogen oxides approaches
60% at very lean fuel equivalence ratios but decreases as fuel
equivalence ratio increases.  At a fuel equivalent ratio of 0.7
the conversion of fuel nitrogen to nitrogen oxides is about 30%, at
1.2 about 15% and above 2 the conversion is asymptotic at about 5%.
     Similar conversion of lignite in a 1750K furnace is shown in
Figure V-10.  Several differences are apparent between the oxidation
of coals at 1500K and at 1750K:

     1)  Burnout at the higher temperature is complete even at
         a fuel equivalence ratio of 1.5, and even at very rich
         conditions the burnout is still 95%.  This may be
       "  indicative of an oxidative pyrolysis mechanism and

     2)  The conversion efficiency of fuel nitrogen to nitrogen
         oxides is marginally lower: about 5 absolute percent.

The conversion of fuel nitrogen to nitrogen oxides was about 60% at
very lean fuel equivalence ratios, 22% at 0.7, 10% at 1.2, and
reaches an asymptotic conversion efficiency of about 1% for fuel
equivalence ratios above 2.

-------
 ft
 o

80
60

*****  "T -XA*~ *  i "" . u 2A1 i2r-Ta &
  
- -
till
   ~    60
.  I S 40
0  2  20
   X
   o
o  >
"  c
 o  O
5*  O
70
60
50
40
30
20
10
                                          Furnace  Temperature  1750 K
                                          Lignite  Coal   38~44jj
                                                                  A f..A
                           1               2               3
                                              
                            Fuel / Oxygen  Equivalence Ratio

                  Figure V-.10. Fate of Fuel Nitrogen During Oxidation:  Conversion
                             to Nitric Oxide (Bottom), Retention by Unburned Char
                             (Middle), and Combustion Efficiency Defined by Solid
                             Weight Loss (Top).  Montana Lignite at 1750 K.

-------
                                                                 127
     Similar conversion of lignite in a 1250K furnace is shown



in Figure V-ll.  Burnout, of course,  is less efficient at the



lower temperature.  Conversion efficiency of fuel nitrogen to



nitrogen oxides is, however, quite similar to that at 1500K.

-------
1
    
^ *
> o
T> "O
 
O #
o
80
60

60
40
20
   X
   o
  "O
o  v
#.
70
60
50
40
30
20
10
                                 Furnace  Temperature 1250 K
                                 Lignite Coal  38~44
                           1               2                3
                            Fuel / Oxygen Equivalence  Ratio

                    Figure V-ll.Fate of Fuel Nitrogen During Oxidation:  Conversion
                                to Nitric Oxide (Bottom),  Retention by Unburned Char
                                (Middle), and Combustion Efficiency Defined by Solid
                                Weight Loss (Top).  Montana Lignite at 1250 1C.
                                                                            to
                                                                            00

-------
                                                                  129





V-3.2  OXIDATION OF LIGNITE CHAR- FIXED RESIDENCE TIME



     Lignite coal was devolatilized at 1500K for one second and



then oxidized in a 1500K furnace supplied with varying amounts



of oxidizer.  Char produced in this manner has about 47% (d.a.f.)



weight loss and 0.67% nitrogen.  The results are shown in



Figure V-12 and show similar trends, but with some differences,



to oxidation of the raw coal.  Asymptotic burnout in one second



was not apparent for oxidation of char as it was for oxidation



of coal at the highest fuel equivalence ratio of 2.6.  This may



result because the char has already been devolatilized at 1500K



and has very little volatile matter left to lose .  For  the  case



of raw coal ,  however, about 50 percent of the initial weight may



be lost by pyrolysis in addition to any oxidation that occurs.



     The nitrogen retained in the char is higher than that found



for raw coals but  there is more char remaining at a given con-



dition.  The fractional retention of the nitrogen in the char is



expected to increase  as  the fuel equivalence ratio is increased,



with a consequent increase in the amount of residual char.



     The conversion of char nitrogen to nitrogen oxides follows a



trend similar to the one observed for oxidation of raw coals.  One



striking difference between the conversion of coal nitrogen from



chars and from coals is that the conversion for chars is signifi-



cantly lower.  Under fuel-rich conditions, the conversion of the



nitrogen in the char reaches an asymptotic value of about 4% but



about 50% of the nitrogen is retained in the solid residual, whereas



at the same conditions, the conversion of the nitrogen in the raw



coal is about 15% and the nitrogen retained in the char from the

-------
                                                           .130
    ^
  z T:
   o*
   z
   o
   111
   fr-
   ee
   bl
        1OO
        5O
       100
5O




40




30




20




 10




 0
                        I
I
                       1.O          2.0          3.O


                (FUEL  EQUIVALENCE RATIO  BASED ON  CHAR)
FIGURE V-12_  Conversion of nitrogen vs fuel equivalence ratio


              in a 1500K furnace for a lignite char previously


              devolatilized at 1500K for 1 second.

-------
                                                                 131
coal is about 36%.



     The residual unburned material from char oxidation is greater



than from coal oxidation at similar conditions and the nitrogen



retention of the residual from the char is correspondingly higher.



However, even under conditions where no residual is left for either



char or coal oxidation/  the conversion efficiency of  the



char is significantly lower than that for the raw coal.



     Similar results are obtained for oxidation of chars at two



other temperatures.  A lignite coal was devolatilized for one



second at 1750K and then oxidized in a 1750K furnace.  The results



are shown in Figure V-13.  Similarly, a lignite coal was devola-



tilized for one second at 1250K and then oxidized in a 1250K



furnace.  These results are shown in Figure V-14.

-------
o    80
    60

.   60
     o
     CO
        c o
     o 2g20
        *
        k
        O
        z
        a
        -o
        O
        U
     70
     60
     50
     40
     30
     20
     10
                              1
FUrnoce  Temperature  1750 K     -
1750 K  Lignite  Char
                               Fuel /Oxygen   Equivalence Ratio
                                         F1gureV-l3.
Fate of Char Nitrogen During Oxidation:  Conversion to Nitric Oxide (Bottom):  Retention by Unbumed Char
 (Middle); and Combustion Efficiency (Top).  Char from Montana Lignite Pyrolyzed and Oxidized at 1750 K.
                                                                                      Ul
                                                                                      to

-------
I
   5   80
      60
      
   c  
   '^   60
   c  u 40
   o
   <
   0)
   ce
       o -g20
         
o
              70
              60
              50
              40
              30
              20
              10
                                       Furnace Temperature 1250 K
                                       1250 K  Lignite  Char
                                 1              2              3              4
                                    Fuel  /  Oxygen  Equivalence Ratio
                                         Figure V~14.
Fate of Char Nitrogen During Oxidation:  Conversion to Nitric Oxide (Bottom); Retention by IJnbumed Char
                (Middle); and Combustion Efficiency (Top).  Char from Montana Lignite
                               Pyrolyzed and Oxidized at 1250 K.
                                                                                      U)
                                                                                      CO

-------
                                                                  134




V-3.3  OXIDATION OF SUB-BITUMINOUS COAL



       FIXED RESIDENCE TIME- EFFECT OF TEMPERATURE



     In an attempt to elucidate the effect of furnace temperature



on the conversion of fuel-nitrogen to NO , an additional series
                                        X


of experiments were performed in the flow reactor.  A Montana



sub-bituminous coal in a 38-44 micron size cut was fed into the



reactor for a 1 second residence time at three temperatures and



a series of fuel/oxygen equivalence ratios varying from about



0.8 to 1.2.  The results are shown in Figure V-15.  Consistent



with results of Section V-3.1, it can be seen that conversion



efficiency increases with decreasing fuel/oxygen equivalence ratios

                                                   i

at a fixed temperature.  Similarly, at a fixed fuel/oxygen



equivalence ratio, increasing the temperature from 1250K to



1500K renders little change in conversion efficiency, but a further



temperature increase to 1750K results in a decrease in conversion



efficiency.  A more detailed discussion of the effect of temperature



on NO  production is presented in Section VI.

-------
z
o
c
o
u
O1
o
o
    30
     20
S   10
 o
 9
           T^
                         I      T
                      O 1250K*
                      D 1500K
                      A 1750K
                   I
I
J
          .0.4.    0.8     1.2     1.6
         Rjct/ Oxygen Equlvolancc
                 Ratio
                      Co)
                      (b)
                      (c)
                         I
                                      I
                                   0.8
                                    1.0
                                    1.2
                                           1
t
I
                1200 1400.   1600  1800
                      Tmpcrotupc*K
                                  Figure V-15.
                           Effect  of Temperature on Conversion of
                           Coal Nitrogen to Nitric Oxide. Montana
                           Sub-bituminous Coal.
                                                                            UJ
                                                                            .tn

-------
                                                                  136





V-3.4  TIME-RESOLVED OXIDATION OF CHAR



     As can be seen from Figures V-10 and V-ll, with increasing



fuel/oxygen equivalence ratio, the percent conversion of coal



nitrogen to NO  decreases.  Under the same fuel-rich conditions,
              Ji


however, an increasing amount of the coal nitrogen may escape



the first stage of a staged combustor with the unburned char.



The kinetics of char oxidation have, therefore, been investigated



in order to establish residence time requirements for completion



of the char combustion and to determine the fate of the char-



nitrogen during oxidation.



     A lignite char was prepared by feeding the lignite coal into



a 1750K furnace for a residence time of one second.  The char,



so prepared, was then fed into the furnace at temperatures of



1250K, 1500K, and 1750K.  Total weight loss and nitrogen and



carbon elemental weight losses were then measured for two different



oxygen partial pressures, and for residence times varying from



about 200 milliseconds to about 600 milliseconds by positioning



the collector probe at varying positions relative to the feeder



according to the procedure outlined in Section III-2.  The data



are shown in Figures V-16 and V-17.  As can readily be seen, total



weight losses and initial weight loss rates increase with tempera-



ture and oxygen concentration.  Nitrogen to carbon ratios, similarly,



remain essentially unchanged or decrease only slightly with time



at the lowest temperatures, but decrease significantly with



increasing temperature.  The decrease in N/C ratio is more pro-



nounced at lower oxygen concentrations.  These data are consistent



with the contention that nitrogen is removed from the char both by

-------
o  -90
1   80
O*
~   70
   60
    50
*
01  40
I   30
S   20
5.   10
                              7T
               1250K
           O  1500K
           A A1750K
Op2n Symbols Po20.2 otm
Shodd Symbols P0 aQ.4 otm
               2
 III!
              1      2     34      5678
           Distance  Between. Feeder  And Collector; inch
                              Figure V-16.
                      Weight Loss of Char Prepared from a Montana
                      lignite as a Function of Distance to the
                      Collector During Oxidation.
                                                                           Ul

-------
#    140
J~ 120
a 100
z  o  SO
*'60
01
   40
   20
T
                   T
T
                                  T
T
              Open Symbols
              Shaded  Symbols  PO20.4otm
             I
                     I
I
                   I
             I
      1
                         I      I
                          O0 1250 K
                          D  1500K
                          A   17 50 K
I
I
            1-2     3     4      5     6-7.8
           Distance  Between  Feeder And  Collector: Inch
                                 Figure V-17.
                        Nitrogen/Carbon Ratio as Percentage of
                        that in the Original Char as a Function
                        of Distance to the Collector During
                        Oxidation.
                                                                           u>
                                                                           00

-------
                                                                139
continued devolatilization and heterogeneous oxidation,  whereas



carbon, thought to form more stable structures in the char,  is



further removed only by oxidative attack.

-------
                                                                   140
                         SECTION VI
            DISCUSSION OF EXPERIMENTAL RESULTS

VI-1.0  DEVOLATILIZATION RESULTS
     At the start of this period, no data existed in the open
literature on devolatilization of nitrogen from coal at conditions
applicable to pulverized coal flames.  A relatively large body of
data is available on nitrogen behavior in coal subjected to coking
conditions, i.e. low temperatures, 950C; slow heating rates,
C/min.; and in packed beds (Konle, 1928, Fielder and Davis, 1934,
Davis and Parry, 1939, Kirner, 1947, Hill, 1945, Lowry, 1942, and
Bronshtein, 1970).  With the growth in interest in control of
conversion of coal nitrogen to nitric oxide in pulverized coal
flames, the data available on nitrogen behavior under coking
conditions has been extrapolated to the conditions of pulverized
coal flames and used to construct speculative models.  Such
extrapolations of coking data have led to misinterpretation of
coal nitrogen behavior in pulverized coal flames.
     Insofar as devolatilization can be viewed mechanistically as
the first step in coal combustion  (See, again, Section V-1.0),
development of a strategy for control of NOV emissions based upon
                                           Jv
combustion modification requires an accurate knowledge of the time/
temperature history of devolatilization under conditions appro-
priate to pulverized coal combustion.  Such experiments are also
necessary for future development of a more detailed mechanistic

-------
                                                                  141





model of coal pyrolysis which would hopefully relate evolution of



volatile species to thermal rupture of specific bonds or functional



groups in the complex matrix which comprises coal.



     All of the pyrolysis data in this study were obtained from



coal devolatilized in an inert atmosphere.  In a real pulverized



coal flame, of course, coal devolatilization occurs in a rich but



not oxygen-free atmosphere.  The pyrolysis data of this study were



taken in an oxygen-free environment because the inert atmosphere



allowed the effects of thermal decomposition to be clearly separated



from the effects of oxidation.  It is realized that thermal and



oxidative pyrolysis, as well as oxidation, may not proceed by



entirely independent mechanisms,but the experimental division out-



lined here represents the clearest, first-order approach to the



problem.  Future extensions of this work might involve a more



detailed exploration of the coupling of the mechanisms of devola-



tilization and oxidation.

-------
                                                                  142





VI-1.1  ASYMPTOTIC CRUCIBLE RESULTS



     Crucible results (Figures V-2 and V-3) in this study extended



the maximum temperature of available data from about 1500K to



2200K.  The nitrogen retention in the char and weight loss (dry



basis) are shown versus maximum temperature in Figure VI-1.  The



nitrogen retention and weight loss are reported at the times at



which weight loss approached an asymptotic value.  Under conditions



of asymptotic weight loss the chars may still undergo slow nitrogen



loss.



     In Figure VI-1, the low temperature portion of the nitrogen



retention curve for lignite shows little nitrogen loss at tempera-



tures below 750K; for bituminous coal, 5% of the nitrogen is lost



at the same temperature, even though the dry weight losses are about



equal.  In general, the nitrogen loss from the bituminous coal under



coking conditions seems to be slightly greater than from the lignite.



This slight difference might be explained by some small amount of



the coal nitrogen being bonded in exinite resinous bodies that can



distill without significant thermal degradation.



     The retention of nitrogen in lignite and bituminous coal char



is proportional to temperature  between 80 and 20% retention in



agreement with other data.  The first and last portion of the



nitrogen removal appears to be slower.  The slow removal of nitrogen



from coals implies that nitrogen is contained in coal in a relatively



stable form, probably predominantly in heterocyclic nitrogen com-



pounds .

-------
                                          A Lignite Nitrogen Retension
                                          A Bituminous  Coof Nitrogen Retension
                                          O Lignite Weight  Loss
                                          O Bituminous Coal  Weight Loss
                                                   ASTM WL  Bituminous Cool
                                                                  _
                                                     ASTM WL Lignite	'
                                         ASTM N Lignite
                                         ASTM N Bituminous Cool
                                 I   I   I   I
            500
100O
      1500

TEMP   K
2000
             Asymptotic  Nitrogen Retension  ond  Weight  Loss Crucible
                                    Experiments
FIGURE Vi-1.   Coal nitrogen retention in the  char under conditions of asymptotic

                 weight loss.
                                                                                               U)

-------
                                                                  144






     Some of the data at 1223K were obtained using the ASTM



Proximate Analysis test and some were obtained by heating to the



ASTM test temperature in the furnace.  The nitrogen retention under



ASTM Proximate Analysis conditions is 72% for the lignite and 63%



for the bituminous coal.  There appears to be little difference



between nitrogen retention for either coal between actual and



simulated ASTM Proximate Analysis tests.  There was no obvious



correlation between ASTM nitrogen retention and the nitrogen



retention of chars in pulverized flames.



     The nitrogen loss from both coals is complete at high tempera-



tures.  Very little nitrogen (less than 10%)  is retained in the char



at temperatures above 1750K.  Nitrogen is completely removed for



both coals at a temperature of 2200K, even though the weight loss



is only 44% dry, and char has retained about 70-75% of the original



carbon.  The carbon retained in the crucible experiments is much



higher than the carbon retained under simulated pulverized coal



flame conditions.  The conclusion is that under slow heating in



packed beds, carbon can form complexes that are stable to at least



2200K but that nitrogen compounds which may form higher molecular



weight complexes are not stable at pulverized coal flame conditions.



     The total weight loss at 2200K is only marginally greater than



the ASTM volatile matter, 44% compared to 42% dry.  The ASTM volatile



levels indicated by the dashed lines are well established by repeated



tests from different laboratories.  The weight loss shown at 1223K,



the ASTM temperature, was slightly lower for both lignite and



bituminous coal but the difference is not thought to be significant.

-------
                                                                  145





VI-1.2  TIME-RESOLVED FLOW FURNACE RESULTS



     The pyrolysis results of greatest interest for purposes of



developing emission control strategies  are those obtained under



conditions most nearly approximating those found in pulverized coal



combustors, namely, high heating rates, 104-105C/sec, and high



temperatures, 1500-2200K.  It is expected that high heating rates



will result in weight losses differing somewhat from those found



in the crucible experiments.  As will be seen in a later section,



these pyrolysis data are necessary for elucidating the mechanism



of NO formation during coal burnout, in particular, for distin-



guishing  NO formed through homogeneous gas phase oxidation from



that formed via heterogeneous burnout of char.  This matter is



returned to in detail in Section VI-3.  Since the amount of nitrogen



devolatilized from coal under different conditions may have a large



effect on control strategies, it is desirable, as indicated above,



to develop a predictive model for nitrogen pyrolysis from coal that



will be applicable to conditions in a pulverized coal flame.



     Empirical models can be developed that adequately represent



pyrolysis data.  Little physical significance can, however, be



ascribed to these models  nor can they reliably be extended beyond



the rartge of experimental conditions employed or coals studied.



This, of course, severly limits the usefulness of these models.



However, where fundamental models are difficult to develop, empirical



relations do, often, serve useful engineering functions.



     Coal pyrolysis has, in the past, generally been considered too



complicated to model except empirically.  New information developed,

-------
                                                                  146
however, on the chemical and physical structure of coal does allow



some physically realistic models of complex coal processes to be



proposed.  While the pyrolysis results presented here do lead to



some conclusions concerning the nature of devolatilization and



condensation reactions likely to predominate in coal pyrolysis,



proposal of a full mechanistic pyrolysis model is considered beyond



the scope of this work and will not, therefore, be attempted.  We



do, however, in the tradition of coal kineticists, present an



empirical interpretation of coal nitrogen pyrolysis.



     The usual first order treatment of nitrogen pyrolysis from



coal, except that the rate is restricted to the first 10% of



nitrogen loss, is compared with fundamental first-order rate con-



stants in Figure VI-2.  The rate constants were taken from a com-



pilation by Benson and O'Neal (1970) and represent reliable



experimental values which have been heavily weighted by absolute



rate theory considerations.  Also shown on the plot are some



pyrolysis data taken on nitrogen compounds that might be considered



to represent the form of fossil fuel nitrogen.  The nitrogen com-



pound pyrolysis data was taken from Kurd and Simon (1962) and



Axworthy (1975) .



     The pseudo-first order rate constant derived from the initial



slopes of the data presented in Figure VI-2 yields a value of





                 kN = 9.3 x 103 exp(-22,700/RT) sec'1





for the Montana lignite-A and Pittsburgh Seam #8 hy-A bituminous



coals studied here.  It is apparent that the rate constant derived



in this study does not resemble elementary first order reaction

-------
                     10"
                     10'
                      10"
                  a  '
                  
-------
                                                                  148
rate constants.  Rate constants derived from  pyrolysis of model



nitrogen compounds are reasonably in agreement with the rate con-



stants derived for nitrogen removal from the coals of this study



but are in poor agreement with fundamental first order rate



constants.



     Kobayashi (1976) has derived rate constants from the same



data in a parallel study and found an overall weight loss constant



for both coals of about:





                 = 6.6 x 104 exp(-25,000/RT)  sec"1





The rate constants derived for loss of hydrogen, carbon, and oxygen



were also very similar to the constants for nitrogen loss and over-



all weight loss.  Kobayashi has tabulated previously derived coal



pyrolysis rate constants and found the constants, within a range



of scatter, to be consistent with each other and to be consistent



with rate constants derived similarly in this study.  The rate



constants do not agree with the theoretical values that would be



predicted if the kinetic models used to interpret the data were



physically realistic.  The conclusion to be drawn is that first



order kinetic models can explain individual sets of data, provided



enough parameters are used, but cannot predict consistent rate



constants for coal pyrolysis.



     It is probably obvious that coal pyrolysis involves an entire



spectrum of reactions of different orders, including parallel,



sequential, and competitive paths.  As pyrolysis is traditionally



interpreted as occurring in a single reaction which is first order

-------
                                                                  149





in material remaining to be reacted, it is not surprising that rate



parameters so derived do not resemble rate constants for elementary



reactions.  It is extremely difficult, therefore, to derive physi-



cally meaningful pyrolysis rate parameters for a substance as



complicated as coal.




     Perhaps a more useful analysis of the coal pyrolysis data,



from the viewpoint of development of a mechanistic model, is the



correlation of elemental nitrogen loss with total volatile loss



under conditions applicable to pulverized coal flames.  This is



demonstrated in Figure VI-3.



     The rate of nitrogen loss during pyrolysis is not independent



of the loss of other elements during pyrolysis because C, H, 0, N,



and part of the S are bonded together in intricate chemical struc-



tures.  Kobayashi (1976) has demonstrated the similarity between



individual element removal rates and the rate of  total  weight loss



Gray, et a_l (1975)have also produced empirical correlations between



the extents of various element loss during hydrogasification of coal.



     Figure VI-3 shows nitrogen retention versus d.a.f. weight loss



for the two coals of this study subjected to pyrolysis in flow



experiments (open points), free fall into crucible (half-open points),



and crucible results (solid points).  The correlation between



nitrogen loss and total weight loss is good for both coals when only



the laminar flow data is considered.  The data for pyrolysis of the



coals in crucibles show a much higher rate of nitrogen loss than




carbon loss.

-------
                           NITROGEN  LOSS V3 WEIGHT  LOSS
                            LIGNITE
BITUMINOUS
                 x
                 u
                 o
                 LJ
                 U
                 a:
                 LJ
                 o
                 o
                 QL
                       80  60   40   20   0 80   60  40  20   0

                                 /. WEIGHT L6SS ( d.O.f.)
FIGURE  vi-3.   Correlation between nitrogen loss and total weight loss during

               pyrolysis of coal.
                                                                                       en
                                                                                       o

-------
                                                                  151

     In the long residence time crucible experiments reported in
Section V-2.1 it was demonstrated that the total weight loss
reached an asymptotic value only slightly higher than would be
predicted by the ASTM Proximate Matter Test, but that elements
other than carbon were completely removed.  This can be explained
because carbon is the only element that can form a highly con-
densed substance that is stable at pulverized coal flame tempera-
tures. Nitrogen is  not thought to have sufficient bonding electrons to
participate in high order condensation reactions with carbon.
Nitrogen must, therefore, be present only on the periphery of the
char structure where it can be removed under the conditions of a
pulverized coal flame.
     Gray, e_t al.  (1975) observed similar behavior in his experiments
on hydrogasification of coal.  Under conditions similar to those
used in this work Gray found that nitrogen loss and carbon loss were
related linearly with a slope of 1.8.  However, when coal was
pyrolysized in his  "hot rod" reactor with heating rates and
environmental conditions more nearly like those encountered in
the crucible experiments of this experiment, nitrogen loss was
no longer linearly  related to carbon loss.  Nitrogen loss was more
rapid than carbon loss and nitrogen could be completely removed
with removal of as  little as 30% of the carbon.
     Nitrogen loss  does not commence with carbon loss but is
delayed until approximately 10-20% of the carbon is lost.  This
is consistent with  the belief that some of the carbon is contained
in alphatic structures that are easy to remove compared to destruc-
tion of aromatic stuctures, while most of the nitrogen is probably

-------
                                                                  152




held in heterocyclic ring stuctures.



     Nitrogen is removed from coal more rapidly than total weight



loss occurs, after the initial delay in nitrogen removal, as in



evidence from the slope of the curves; the ratio of nitrogen loss



versus weight loss is 1.25 for the lignite and 1.5 for the



bituminous coal.



     Nitrogen loss from different coals at pulverized coal flame



conditions in an inert atmosphere can then be crudely estimated,



probably within 10 to 20 percent, from total weight loss data by



use of the shape and nature of Figure VI-3.  No nitrogen loss should



be assumed until approximately 15% of the total weight  has been



lost, then nitrogen loss can be estimated as approximately 1.25-



1.5 times the total weight loss until approximately 60% d.a.f.



weight loss has occurred.  Extrapolation of this relationship



beyond 60% weight loss is uncertain and would predict that, for



pyrolysis of coals under inert pulverized coal flame conditions,



complete nitrogen loss will occur at 80-90% total weight loss.

-------
                                                                   153
VI-1.3  PRIMARY NITROGENEOUS PYROLYSIS PRODUCTS




     It is possible at this juncture to draw some educated



conclusions as to the primary fate of fuel-nitrogen released as



volatiles from coal.




     Pyrolysis data of this study indicates an induction period



after heating prior to release of fuel nitrogen from coal.  This



is though to reflect the common speculation that most of the fuel



nitrogen is contained in reasonably stable, heterocyclic structures,



rather than in labile side chains.  Consistent with this belief



is the recent observation of Blair, e_t a_l  (1976) that under pyrolysis



conditions most of the coal nitrogen is,  in fact, evolved as tars



and heavy, hetrocyclic molecules.  These  large molecules are then



thought to undergo secondary pyrolysis to yield HCN as the major



nitrogen-bearing species.  Studies by Axworthy (1976)  on the



pyrolysis of model compounds indicate rapid rates for this pyrolysis



step which, when extraplated to temperatures pertinent to pulverized



coal combustion, suggest that formation of HCN from these heavier



molecules is very rapid and certainly not rate controlling.



     The results of this study seem to confirm this mechanism?



closure of the nitrogen material balance at high temperatures



(See Section V-2.3) resulting in HCN and N2 showing up as the major



stable species.  The N2 so observed, however, is almost certainly



a product of the secondary pyrolysis of HCN, and not a primary



pyrolysis product.  This is easily understood from the complementary



observation that under fuel-lean, low-temperature conditions (See



Figure VI-4) , conversion of volatile fuel-nitrogen to NOx is

-------
                                                                 154
essentially complete.  Insofar as, under these conditions,  N~ is



not thought to be a reasonable precursor to NO, N2 cnnot be a



primary pyrolysis product of coal.

-------
                                                                  155



VI-2.0  OXIDATIVE RESULTS




     The primary objective of this work is the elucidation of the




mechanism of NOx formation during pulverized coal combustion with




a specific view towards development of a strategy for the minimi-




zation of NO  emissions.  As will be seen, NO  can originate both
                                             J\.


from gas phase oxidation of nitrogen bearing volatiles emitted from



the initial pyrolysis of the rapidly heated coal as well as from




heterogeneous oxidation of the devolatilized char.  Insofar as the




chemical mechanisms of these two processes are rather different,




variation of combustion conditions by the standard techniques




available to the combustion engineer, e.g.  change of temperature,




fuel/air equivalence ratio, degree of mixedness, staging of com-




bustion, etc., may effect the two processes quite differently; and




it is certainly not immediately obvious what combinations of




combustion conditions will lead to a minimum in NO  emissions.  In
                                                  Ji



an attempt to gather at least an empirical understanding of the




factors affecting the split of fuel-nitrogen between volatiles and




char and the oxidation of each, a series of independent coal and




char burnout experiments were undertaken.  These results, combined




with the pyrolysis data of the previous section, will be seen to




throw valuable light on the problem.



     All experiments were run in the laminar flow reactor described




above in Section III.  As such, aerodynamic conditions were not



varied.  In a somewhat parallel study, however, Pershing and Wendt




(1976) investigated the effect of mixing in a 6 Ib/hr unit and




found that efficiency of fuel nitrogen conversion to nitric oxide



was affected by burner aerodynamics but not affected by flame

-------
                                                                 156





temperature, provided the flame temperature was not too high.   The



results of Pershing and Wendt on fuel nitrogen conversion on a



similar coal agree with the conversions of this study.  For instance,



at 15% excess air a Pittsburgh coal with 1.20% nitrogen showed



about 28% conversion; approximately the same value was obtained



for a similar coal in this study -



     Pershing and Wendt injected coal through a divergent nozzle



and an axial injector and found, when everything else was equal,



about a 10% absolute difference in conversion efficiency between



the two nozzles for a Western Kentucky hvC bituminous coal and



Montana sub-bC coal.  The injector system used in this study is



close to the axial injector system used by Pershing and Wendt.



The burner design had no effect on char oxidation.  The injector



used in this study can not match the mixing conditions encountered



in a pulverized coal flame and so represents a different flame.



Data, however, can be compared on a relative basis for this mixing



condition and valid conclusions can be drawn about the effects of



variables other than mixing for systems in general.



     Pershing and Wendt used adiabatic flame temperatures as a



comparison of temperature effects on fuel nitrogen behavior since



they had no direct measure of the flame temperature.  Comparison



of the effects of flame temperatures on thermal and fuel nitrogen



emission showed that thermal nitric oxide emissions increased



greatly at higher temperatures but fuel nitrogen conversion did



not increase until the adiabatic flame temperature reached about



2400K.  The stoichiometric adiabatic flame temperatures in helium



for the coals of this study, disregarding dissociation, were about

-------
                                                                  157




2100K for the lignite and about 3000K for the bituminous coal;




corrections for dissociation will lower the lignite temperature




about 200K and the bituminous coal 300K.

-------
                                                                  158
VI-2.1  COAL OXIDATION



     In this study, coal particles, nominally in the size range of



38-45 microns,or,occasionally75-90, were injected with an equal



mass of gas through a 1-2 mm nozzle into a hot wall furnace.



Oxygen content of the helium/oxygen carrier stream corresponded to



15-25% of the stoichiometric requirement which corresponds well



with the values of the primary air utilized in industrial practice.



The secondary air  (again stimulated by a helium/oxygen mixture)



was injected through a honeycomb flow straightener surrounding the



coal-feed nozzle.   In view of the difference  between the mode



of operation in this study and practical systems, the gross agree-



ment in the data is remarkable.  The conversion of fuel nitrogen to



nitric oxide in our system for a fuel equivalence ratio of 0.7 is



about 30 percent which is towards the upper end of the range of



estimates for field units.



     Comparison of the results in Figures V-9 and V-10 suggest tnat



higher temperatures appear to have lower fuel nitrogen conversions.



At a normal operating fuel equivalence ratio of 0.7, the percentage



of fuel nitrogen conversion to nitric oxide decreased by several



percentage points absolute when the furnace temperature was increased



from 1500 to 1750K.  At the higher temperature, as would be expected,



the burnout was markedly improved and the amount of nitrogen retained



in the char drastically reduced.



     Because the data of Figures V-9 and V-10 were gathered over a



long time span, and because of the current interest in the effect



of temperature on NO  emissions from pulverized coal flames, the

-------
                                                                 159
effects of temperature were rechecked in an independent series of



experiments using a Montana sub-bituminous coal.  These results,



shown in Figure V-15, confirm the tendency towards decreasing



conversion efficiencies with increasing temperature.  The effect



of temperature on controlling NO  emissions is returned to in
                                X


Section VI-3.

-------
                                                                  160
VI-2.2  CHAR OXIDATION



     Essential to elucidating the mechanism of fuel nitrogen



conversion to NO   is an understanding of the split of fuel
                Ai


nitrogen between volatiles which are homogeneously oxidized to



NO  and char which is heterogeneously oxidized.  Such information



can be gleaned from a comparison of coal oxidation results, which,



inherently, include contributions from both processes and char



oxidation results, which, necessarily, involve only the latter



process.



     The conversion efficiency of fuel nitrogen to nitric oxide



for a char was lower by about 10% absolute than the corresponding



value for a coal at a fuel equivalence ratio of 0.7.  The effect



of fuel/oxygen ratio on conversion of the fuel nitrogen to nitric



oxide in the coal was similar to that for the char  (See Figures



V-9 through V-14).  The fractional burnout of the char was lower



than the corresponding values for coal at fuel rich equivalence



ratios.  In addition, more nitrogen was retained by the unburned



char than the unburned coal.



     Pershing and Wendt  (1976) report values for the conversion



efficiency of the nitrogen in a char of about 13 percent at a fuel



equivalence ratio of 0.7 for both a divergent nozzle and an axial



injector.  This conversion efficiency agrees well with the value



found in this study using a char devolatilized at 1500K for one



second.  It seems reasonable that mixing should have much  less



effect on the slow heterogeneous oxidation reactions than  on the



oxidation of volatiles.

-------
                                                                   161





      Figures V-9 through V-ll demonstrate that with increasing fuel



 equivalence ratios, the percent conversion of coal nitrogen to



 nitric oxide decreases.  Under the same fuel rich conditions,



 however,  an increasing amount of the coal nitrogen may  escape  the



 first stage of the combustor with the unburned char.  The  study



 of the kinetics of oxidation of char was therefore undertaken  with



 the objective of establishing the residence time requirements  for



 completing the combustion of the char and determining the  fate of



 the char-nitrogen during oxidation.   The data in Figures V-16  and



 V-17 are  plotted as a function of distance.   For determining kinetic



 parameters,  residence time measurements are  necessary.  Such mea-



 surements were made by Kobayashi (1976)  who  used essentially the



 same system and showed that the average particle velocity  may  be



 approximated by 1.4 times the average main gas  velocity.   For  the



 interpretation of results of the present study,  distances  were con-



 verted to residence times,  using Kobayashi's  approximation.  The



 results with times reported in milliseconds  are  summarized below:
^*\^ Distance, inch
Temperature, K ^^-^^^
1250
1500
1750

3
256
213
183

4
340
283
243

5
427
356
305

6
512
426
365

7
599
500
428
     It has been previously shown (Figures V-4 and V-5)  that



during coal pyrolysis the carbon loss appears to become  asymptotic

-------
                                                                 162

at times exceeding 100 milliseconds at 1750K, the temperature used
to produce the char in this study, but that nitrogen release con-
tinues for longer times.  The rationalization of the results was
that while carbon is present in relatively stable compounds in the
char, no comparable stabilized nitrogen structures are formed;
consequently, char-nitrogen continues to be released until it is
completely eliminated from char.  Therefore, when a char particle
is oxidized, the nitrogen loss will be due to both devolatilization
and oxidation, but the carbon loss will be exclusively due to
oxidation.
     Assuming the fuel nitrogen is uniformly distributed throughout
the char, the oxidation rate of the nitrogen should therefore equal
the product of the oxidation rate of the carbon and the mole ratio
of the fuel nitrogen to carbon in the char (Wendt and Schulze, 1976),
Since the pyrolysis loss and oxidation loss of char-nitrogen are
additive, the consumption rate of char-nitrogen can be written as:
  Total Consumption-Rate l
  of Char Nitrogen      J
  OR
  Since
        dN _ /3N
        dT   \3t / Pyrolysis
          dC  = /dC
          dt    Idt } Oxidation
    Consumption-Rate of
    Char-Nitrogen Due
    to Pyrolysis

    Consumption-Rate of
    Char-Nitrogen Due
    to Oxidation
3t }  Oxidation

-------
                                                                163



Therefore,
     -   1            + (dc\             (\
   dt    ^ty Pyrolysis    \dt/ Oxidation   \C) Following Pyrolysis



This model indicates that for low oxygen pressure and high tempera-


ture char oxidation, the nitrogen to carbon ratio in the char will


decrease as a function of reation time.  As the oxygen pressure


increases and temperature decreases, the oxidation process tends


to dominate, resulting in a nitrogen/carbon ratio which is approx-


imately constant during the course of the combustion process.  The


experimental findings support this conclusion.  The data can be


used to derive kinetic parameters for the oxidation and pyrolysis


of the char nitrogen.  The experiments were carried out with large


excess of oxygen so as to maintain a constant oxygen partial


pressure and thus facilitate interpretation of the results.  The


basic kinetic parameters derived from the data can be used to


analyze char burnout under the varying oxygen concentrations that


are encountered in practical combustors.

-------
                                                                  164
VI-3.  THE CONTRIBUTIONS OF VOLATILE AND CHAR NITROGEN

       TO NITRIC OXIDE




     Very little information exists on the contribution of nitric




oxide produced from the oxidation of char and from the oxidation




of volatiles.  Pereria et al_ (1975) have shown the volatile




contribution in a fluidized bed to be negligible at temperatures




below 1000K but to increase until the nitric oxide produced from




volatile matter accounted for 2/3 of the total emission at 1200K.




These figures agree qualitatively with the results of this study,




although the systems, temperatures, and mixing patterns must be



considerably different.




     The data from Figures V-9 through V-14 can be used to obtain




preliminary estimates of the fractions of the NO  formed in pulver-
                                                X.



ized coal flames that are contributed by the volatiles and char.




As a first order approximation it will be assumed that the conver-




sion efficiencies to nitric oxide of the nitrogen in the char and




volatiles are independent.  The fate of the fuel nitrogen during




combustion of coal can be represented by the following simplified



scheme:
          1-a
                     VOLATILES
                     CHAR
                              i_n   N0 + OTHER
                                 11     ~
                                 BURNED CHAR
                                             1-n
                          1-Y
UNBURNED CHAR
                      + OTHER

-------
                                                                  165



where



     a  =  fraction of the coalTN that is released as  volatiles



  1  - a  =  fraction of the coal-N that is retained in  the  char



    n,  =  fraction of the volatile-N that is  converted  to NO
                                                           x


     Y =  fraction of the char-N that is consumed




  Yr|2 =  fraction of the char-N that is converted to NO
                                                       J\



The overall  fractional conversion,  n*,  of coal-nitrogen  to nitric


oxide is  then:
              n   =  an,  + (1 - a)  y
                                  n
                                  n2
         Overall   Volatile    Char Contribution

      Conversion   Combustion

      to NO
The value  of  (1  - a)  as  a function of  temperature  is obtained  from

                                               *
the pyrolysis  data of Figure  V-6 .   Values  of  n  and Yn2 as  functions



of temperature and fuel  equivalence ratio  are obtained from the coal



oxidation  data of Figures V-9 through  V-ll and from the char oxida-



tion data  of Figures  V-12 through  V-14.  From this information and



the above  scheme for  fuel-nitrogen partitioning, it is possible to



derive values  for both the efficiency  of volatile  nitrogen  conver-



sion to NO , HI ,  and  the function  of total NO contributed  by
           x    J.                               x
volatiles  (ann/n) -  These values  are plotted as a function of fuel/



oxygen equivalence ratio  for  the two temperature extremes of these



experiments,  1250K and 1750K, in Figure VI-4 .  The efficiency of



conversion to NO  of nitrogenous species evolved as volatiles (r^)
                li


and the fraction of overall conversion to NOx which originates from

-------
                                                            166
 o
o    90
}|    8O
T.    7O
      6O
  ^  50
O 2
= O  40
S~  30
>    20
     10
o
o
8
o o
I"
      9O
      80
      70
      60
      5O
      40
      30
      2O
      10
                              ^^-^^
              Fuel / Oxygen  Equivalence  Ratio
                           Figure VI-4.
                Contribution of Volatiles to NO  Emission:
                Total N0x Contributed by Volatiles (Top), and
                Conversion of Volatiles to NO  (Bottom).

-------
                                                                   167
(               \
            	I ) are both seen to decrease monotonically with
            H* /
 increasing fuel/oxygen equivalence ratio.  The former result is
 entirely consistent with previously reported data  on  the conversion
 of  volatile fuel-nitrogen to NO  in both premixed  (Fenimore  1972,
                                J\
 Sarofim,et al 1975)  and laminar diffusion flames  (Sarofim, e_t al_  1975)
 and,  in  fact, can be shown to fall between the two extremes.   The
 latter result,  however,  would not have been obvious,  a priori, and
 represents the fact that though c;    .litrogen conversion efficiency
 to  NOx also decreases with increasing fuel/oxygen  equivalence ratio
 (See  Figures  V-13 and V-14), its rate of decrease  is  not as  rapid
 as  that  of the volatiles.
      Also of  great interest is tho effect of temperature on  the
 volatile/char nitrogen partitioning and conversions to NO of fuel
                                                          J\
 nitrogen.   Comparison of Figures V-13 and V-14 for char oxidation
 at  1750K and 1250K suggests that the fractional  conversion  of  char
 nitrogen to NO  is independent of temperature. (The data of  Figure
               J^
 V-12  at  1500K,  obtained much earlier in the studies,  is thought
 to be of lower  reliability and to exhibit more scatter.)  The results
 of Figure  VI-4,  however, indicate that the fractional  conversion to
 nitric oxide  of the  volatile nxLr jen (r^)  decreases with increasing
 temperature.   This result is consistent with the general  observation
 that  increases  in fuel nitrogen concentration result in  a decrease
 in conversion of fuel nitrogen to nitric oxide.  As the  temperature
 is increased,  the fraction of the fuel nitrogen devolatilized
 increases  and the concentration of fuel  nitrogen in the  gas phase
must  also  increase.   Interestingly, however,  the results of

-------
                                                                  168
Figure VI-*4 also indicate that with increasing temperature the

fractional contribution to total NO  by volatile  I ttnA\ increases.
                                   X              I    " I
                                                  V n* )
This, as inspection, again, of Figure V-6 reveals, is not incon-

sistent with the observations above in that with increasing

temperature, a much larger fraction of the fuel nitrogen is evolved

as volatiles.  The results of Figure VI-rA merely imply that the

rate of increase of volatile evolution with increasing temperature

exceeds the complementary rate of decrease in conversion efficiency

to NOX.  These two effects tend to compensate and partially explain

the small dependence on temperature of the net conversion of coal

nitrogen.  In this study, a small net decrease is observed (See

Figure V-15).  In a separate study, Pershing (1976) has found

practically no effect of temperature on the conversion of fuel

nitrogen to nitric oxide.

-------
                                                                   169




VI-4.  CONCLUSIONS AND  IMPLICATIONS FOR CONTROL OF NO  EMISSIONS
     Major conclusions of  this work may be summarized as follows



     (1)  Evolution of nitrogen  from coal is kinetically



      controlled  for  conditions  encountered in typical



      pulverized  coal flames.  Although carbon forms a



      stable char through  condensation reactions, no com-



      parable  stabilized nitrogen  structures are formed.



      Consequently, nitrogen  can he completely eliminated



      from the char at temperatures exceeding 1750K.




     (2)   When coal  is  slowly heated under an inert



      atmosphere  in a crucible,  most of the volatile matter



      evolution occurs at  temperatures below that of the



      ASTM Proximate  Analysis test (1023K), whereas most



      of  the nitrogen is evolved at higher temperatures.




     (3)  The  early stages of pyrolysis of nitrogen from



      coal can be empirically correlated, for the two coals



      studied,  by a pseudo-first order rate constant of




      value:





                1; - 9.T  x  103     -22,700/RT) sec"1




     (4)  Little  nitrogen  is  lost, under pyrolysis conditions,



      until 10-15% of the  weight of the coal has been devola-



      tilized.  After this initial induction period, most



      likely associated  with  loss  of side chains and scission



      of  aliphatic bridges in coal, the rate of fractional



      nitrogen loss is found  to  be proportional to the rate

-------
                                                           170
 of fractional total weight loss.
(5)   Conversion of fuel  nitrogen  to  nitric  oxide  in
 pulverized coal flames  decreases with  increasing
 fuel/oxygen ratio, falling to negligible values  at
 equivalence ratios greater than  1.5.   However, at
 temperatures of 1500K  or  lower, some  of the  coal
 persists  as unburned  char  which may contain a
 significant fraction  of the original fuel  nitrogen.
(6)   Conversions to nitric  oxide of  fuel nitrogen
 during char burnout are 2-3 times lower than  the
 corresponding values  for coal.
(7)   During coal burnout, fuel nitrogen is  partitioned
 between volatiles and char, higher  temperatures
 favoring  a larger fraction of volatiles.   The effi-
 ciency of conversion  of the volatile nitrogen to
 nitric oxide  decreases with increasing fuel/oxygen
 ratio and decreasing  flame temperature.  The  effi-
 ciency of conversion  of char nitrogen  to nitric
 oxide also decreases  with  increasing fuel/oxygen
 ratio, albeit, more  slowly than  does volatile
 nitrogen, but is, however, rather insensitive to
 temperature.  Carbon  burnout, of course, is enhanced
 at higher temperatures.
(8)   Oxidation experiments  on chars  support the view
 that there is no selectivity between nitrogen and
 carbon loss during oxidation  but that the char

-------
                                                                 171
     nitrogen  may  undergo further  pyrolysis  in parallel
     with  the  oxidation.
     The above results  may be used to  guide  the development of
control strategies.   In a single stage combustor, conversion of
nitrogen to  nitric oxide  may be reduced by the increase in the
fuel/air ratio,  to the  limit imposed by the  emission of unburned
carbon, CO,  or soot.  Because of the compensating effects discussed
above, little  change  in fuel nitrogen  conversion to nitric oxid<=-
can be expected with  changes in combustion temperature but the
contribution of thermal fixation of nitrogen will decrease with
decreases  in combustion temperature.
     In a  staged combustor, the conversion of  fuel nitrogen to
nitric oxide can be decreased by increasing  the  fuel/air ratio,
which results  in decreases in conversion  efficiency of the fuel
nitrogen to  nitric oxide  for the nitrogen content of both volatiles
and char.  Since char oxidation is slow and  nitrogen retained in
the char may be oxidized to nitric oxide  in  a  second stage, it is
advantageous to operate the first  stage at high  temperatures to
devolatilize the coal nitrogen in  the  locally  fuel rich environment.
Although these strategies may be evident, the  result of the present
study provides  a quantitative framework to guide  the development of
contro-1 strategy.  For  example, the kinetic  data on the pyrolysis
of coal can  be used to  estimate the temperature-time combinations
required to  devolatilize  the nitrogen  in  the first stage of a staged
combustor.   In addition,  the kinetic data on char oxidation may be
used to size the second stage of the combustor.  The staging,  can
of course, be  achieved  either by introduction of the combustion
air at different positions  in  a furnace or by aerodynamic  means.

-------
                                                                  172
                           REFERENCES
 1.  Anson,  D., Moles,  F.D.,  and Street,  P.J.,  "Structure and
     Surface Area of Pulverized Coal During Combustion,"
     Combustion and Flame,  16,  265 (1971).

 2.  Axworthy,  A.E., Jr.,  "Chemistry and  Kinetics of Fuel
     Nitrogen Conversion to Nitric Oxide",   AIChE Symposium
     Series  No. 148, TL' 43 (1975).

 3.  Axworthy,  A.E., Schneider, G.R., Shuman,  M.D.,  and
     Dayan,  V.H., "Chemistry of Fuel Nitrogen  Conversion  to
     Nitrogen Oxides in Combustion", U.S. EPA  Report EPA-
     600/2-76-039 (February,  1976) .

 4.  Badzioch,  S., Gregory, D.R.,  and Field, M.A.,  "Investiga-
     tion of the Temperature Variation of the  Thermal Conductivity
     and Thermal Diffusivity of Coal", Fuel 43,  267  (1964).

 5.  Badzioch,  S., and Hawlesley,  P.G.W., "Kinetics  of Thermal
     Decomposition of Pulverized Coal Particles",  Ind. Eng.  Chem.
     Process Des. Develop.  9_,  521 (1970).

 6.  Bartok, W., Crawford,  A.R., and Piegara,  J.J.,  AIChE Symp.
     Series  No. 126, 68, 66 (1972).

 7.  Benson, S.W., and O'Neal,  H.E., "Kinetic  Data on Gas Phase
     Unimolecular Reaction",  NSRDS-NBS-21 (1970).

 8.  Blair,  D.W., Bartok,  W.,  and Wendt,  J.O.L.,  "Devolatilization
     and Pyrolysis of Fuel Nitrogen from  Single Coal Particle
     Combustion", Sixteenth Symposium (International)  on
     Combustion (1976).

 9.  Bronshtein, A.P.,  Makarov, G.N., and Silivinskaya,  I.I.,
     Thermal Degradation of Duznetzk Gas  Coals,  Coke Chem.
     (USSR)  Vol. 5_, 12 (1970).

10.  Davis,  J.D., and Parry,  V.F., "Carbonizing of Subbituminous
     Coal from Puritan Mine,  Dacono, Weld County, Colo.",
     U.S. Bureau of Mines Rept. of Invest., 3428 (1939).

11.  Department of Transportation, "Air Quality,  Noise,  and Health-
     Report of a Panel of the Interagency Task Force on Motor
     Vehicle Goals Beyond 1980",  Publication  No. TAD-443.1 (1976),

12.  De Soete,  D., Revue de L'Institut Francais du Petrole, 28,
     95  (1973); Institut du Petrole, Reports IFP 21, 309, and
     21, 381 (1973) .

-------
                                                                   173
13.  Dryden,  I.G.C.,  "Chemical Constitution and Reactions of Coal",
     in H.H.  Lowry, ed., Chemistry of Coal Utilization, Supplementary
     Volume,  p.  232,  John Wiley and Sons, Inc., New York (1963).	

14.  Edwards, H.W., AIChE Symposium Series No. 126 68, 91 (1972).

15.  Fenimore, C.P.,  "Formation of Nitric Oxide from Fuel Nitrogen
     in Ethylene Flames", Combustion Flame, 19, 289 (1972).

16.  Field  (1970),  Combustion and Flame MA Measurements of the
     Effect of Rank on Combustion Rates of Pulverized Coal,
     Combustion  and Flame 14-237, 1970.

17.  Fieldner, A.C.,  and Davis, J.D., "Gas-, Coke-, and By Product-
     Making Properties of American Coals and Their Determination",
     U.S. Bureau of Mines, Monograph 5  (1934) .

18.  Furusawa, T., Kunii, D., Oguma, A., and Yamada, N. , "Kinetic
     Study of Nitric  Oxide Reduction by Carboneceous Materials",
     paper submitted  to Kagaku, Kogaku  (Chemical Engineering Japan)
     (1977) .

19.  Can, H., Nandi,  S.P., and Walker, P.L., Jr., "Nature of the
     Porosity in American Coals", Fuel 51, 272 (1972).

20.  Gibbs, B.M., Pereira, F.J., and Beer, J.M., Sixteenth Symposium
     (International)  on Combustion  (1976).

21.  Gray, J.A., Donatelli, P.J., and Yavorsky, P.M.,  "Hydrogasifica-
     tion Kinetics of Bituminous Coal and Coal Char", ACS Division
     of Fuel  Chemistry, Preprints. 20_ (4), 130 (1975).

22.  Heredy,  L.A., and Fugassi, P., "Phenanthrene Extraction of
     Bituminous  Coal", in Coal Science, Advances in Chemistry Series
     55_, p. 448, ACS, Washington, D.C.  (1966).

23.  Hill, W.H., "Recovery of Ammonia Cyanogen, Pyridine, and Other
     Nitrogeneous Compounds from Industrial Gases", in H.H.  Lowry,
     ed., Chemistry of Coal Utilization, Volume II, p. 1008,
     John Wiley  and Sons, Inc., New York (1945).

24.  Howard,  J.B., and Essenhigh, R.H.,  "Pyrolysis of Coal Particles
     in Pulverized Fuel Flames", Industrial Engineering Chemical
     Process  Design Develop., 6, 74 (1967).

25.  Howard,  J.B., and Essenhigh, R.H.,  "Mechanism of Solid  Particle
     Combustion with  Simultaneous Gas-Phase Volatiles Combustion",
     Eleventh Symposium (International)  on Combustion (1967).

26.  Hurel, C.D., and Simm, J.I., "Pyrolytic Formation of Arenes. III.
     Pyrolysis of Pyridine, Picolines,  and Methylpyrazine ,  J.  Amer.
     Chem. Sci.  84, 4519 (1962).

-------
                                                                   174
27.  Jodidi, S.L. ,  "Organic Nitrogeneous Compounds in Peat Soil",
     J.  Amer. Chem.  Soc.  32, 396 (1910).

28.  Jonke, A.A.,  Carlis,  E.L.,  Jarry,  R.L.,  Anatasia,  L.J.,  Haas,  M.,
     Paulik, J.R.,  Murphy, W.A., Schoffstau,  C.B., Vargo,  G.N. ,
     "Reduction of Atmospheric Pollution by Application of Fluidized
     Bed Combustion",  Annual Report,  Anl/ES-CEN-1002, Argonne
     National Laboratory  (1970).

29.  Kimbur, G.M.,  and Gray, M.D.,  "Rapid Devolatilization of
     Small Coal Particles", Combustion  and Flame 11,  360 (1967).

30.  Kimbur, G.M.,  and Gray, M.D.,  "Measurements of Thermal
     Decomposition of  Low and High  Rank Non-Swelling  Coals at
     M.H.D. Temperatures", B.C.U.R.A. Document No. MHD 32  (1967).

31.  Kirner, W.R.,  "The Occurance of  Nitrogen in Coal", in H.H.
     Lowry, ed.,  Chemistry of Coal  Utilization, p. 450, John  Wiley
     and Sons, Inc., New  York (1945).

32.  Kirov, N.Y.,  "Specific Heats and Total Heat Contents  of  Coals
     and Related Materials at Elevated  Temperatures", The_  British
     Coal Utilization  Research Association, Monthly_BuTletin  ^9  (2),
     February-March, Part I(1965).

33.  Kobayashi, H.,  "Devolatilization of Pulverized Coal at High
     Temperatures",  Ph.D.  Thesis, M.I.T., Cambridge,  MA (1976).

34.  Kunle, v., Brennstoff- CHEM. 1928  9, 295.(1928).

35.  Lowry, H.H.,  Landau,  H.G.,  and L.L. Naugle, "Correlation of
     the Bureau of Mines-American Gas Association Carbonization
     Assay Tests with  Coal Analyses", Trans.  Am._Inst.  Mining,
     Met., Pet. Engrs., 149, 297 (19421"!

36.  Mandel, G.,  "Gasification of Coal  Char in Oxygen and  Carbon
     Dioxide at High Temperature",  S.M. Thesis, M.I.T.  (1977).

37.  Marsh, H., "The Determination  of Surface Areas of Coals  -
     Some Physicochemical Considerations", Fuel 44, 253 (1965).

38.  Martin, G.B.,  and Berkau, E.E.,  AIChE Symposium  Series No.  126,
     68, 45  (1972).

39.  McCabe, L.C.,  and Boley, C.C., "Physical Properties of Coals",
     in H.H. Lowry,  ed.,  Chemistry  of Coal Utilization, p. 310,
     John Wiley and Sons, Inc., New York (1945).

40.  Nsakala, N., Essenhigh, R.H.,  and  Walker, P.N.,  Jr.,
     "Rapid and Slow Devolatilization of North Dakota Lignite
     (PSOC-246) , OCR Quarterly Progress Report, Penn. State
     (May 22, 1975H

-------
                                                                  175
41.  Padia, A.S.,  The Behavior of Ash in Pulverized Coal Under
     Simulated Combustion Conditions", Sc.D  Thesis  MIT
     Cambridge, MA (1972).             ~	'       "

42.  Paulson, L.E., Beckering, W., and Fowker, W.,  "Separation
     and Identification of Minerals from Northern  Great Plains
     Province Lignite", Fuel 51, 224  (1972).

43.  Pereira, F.J., "Nitric Oxide Emissions from Fluidized Coal
     Combustion", Ph.D. Thesis, Univ. of Sheffield (1975).

44.  Pereira, F.J., and Beer, J.M., "NO Formation from Coal
     Combustion in a Small Experimental Fluidized Bed",
     Second Europeon Combustion Symp., Orleans, France (1975).

45.  Pershing, D.W., "Nitrogen Oxide Formation in Pulverized
     Coal Flames", Ph.D. Thesis, Univ. of Arizona (1976).

46.  Pershing, D.W., and Wendt, J.D.L., "Pulverized Coal Combustion:
     The Influence of Flame Temperature and Coal Composition on
     Thermal and Fuel NOX", Sixteenth Symposium (International)  on
     Combustion (1976).

47.  Pohl, J.H., "Fate of Fuel Nitrogen", Sc.D. Thesis, M.I.T. (1976)

48.  Pohl, J.H., and Sarofim, A.F., "Devolatilization and Oxidation
     of Coal Nitrogen", Sixteenth Symposium (International) on
     Combustion (1976).

49.  Reteofsky, H.L., and Friedel, R.A., "Spectra of Coals and
     Coal Extracts: Proton Magnetic Resonance Spectra of Pyridine
     and Carbon Disulfide Extracts",  in R.A. Friedel, ed.,
     Spectrometry of Fuels, p.70, Plenum Press, New York  (1970).

50.  Rossin, P., and Rammler, E., "The Laws Governing the Fineness
     of Powdered Coal", J. Inst. Fuel 7, 29 (1939).

51.  Sarofim, A.F., Williams, G.C., Modell, M., and Slater, S.M.,
     "Conversion of Fuel Nitrogen to Nitric Oxide in Premixed
     and Diffusion Flames", AIChE Symposium Series No. 148, 71,  51
     (1975).

52.  Shaw, J.T., and Thomas, A.C., Proc. Seventh International
     Conference on Coal Science, Prague (1965).

53.  Shaeklock, R.W., and Drakeley, T.J., "A Preliminary Investiga-
     tion of the Nitrogeneous Matter in Coal", J.  Soc. Chem.  Ind.
     4, 478T (1927) .

54.  Song, Y.H., Beer, J.M., and Sarofim, A.F.,  "Fate of Fuel
     Nitrogen During Pyrolysis and Oxidation", Paper presented at
     Second Annual Symposium on Stationary Source  Combustion,
     New Orleans, La. (1977).

-------
                                                                  176
55.  Sternling, C.V., and Wendt, J.O.L., "Kinetic Mechanisms
     Governing the Fate of Chemically Bound Sulfur and Nitrogen
     in Combustion", NTIS Report No. PB-230895,  U.S. Department
     of Commerce, Springfield, Va.  (1972).

56.  Suuberg, Eric M., "Rapid Pyrolysis and Hydropyrolysis of Coal",
     Sc.D. Thesis, M.I.T. (1977).

57.  Swann, P.O., Aceardice, D.J.,  and Evans,  D.G.,  "Low Temperature
     Oxidation of Brown Coal.  1.  Changes in Internal Surface Due
     to Oxidation",  Fuel 53, 85 (1974).

58.  Tingey, G.L., and Morrey, J.R., "Coal Structures and
     Reactivity", Battelie Energy Program Report (December,  1973).

59.  Turner, D.W., Andrews,  R.L., and Siegmund,  C.W., AIChE
     Symposium Series No. 126, 68,  55 (1972).

60.  Van Kreueler, D.W., Coal; Typology-Chemistry-Physics-
     Constitution, Elsevier Publishing Co., Amsterdam (1961).

61.  Wendt, J.O.L.,  and Schulze, O.E., "On the Fate  of Fuel
     Nitrogen During Coal Char Combustion:  AIChE Journal 22,
     102 (1976).

62.  Zeldovitch,  Ya. B., Sadovnikov, P.Ya., and Frank-Kamenetskii,
     D.A., "Oxidation of Nitrogen in Combustion", Acad. Sci. U.S.S.R.,
     Inst. Chem.  Phys., Moscow-Leningrad (1947).

-------
                                                                          177
                                TECHNICAL REPORT DATA
                          (Please read Instructions on the reverse before completing)
1. REPORT NO.
 EPA-600/7-78-165
                           2.
                                                      3. RECIPIENT'S ACCESSION NO.
4.T.TLE ANO SUBTITLE Combustion Research on the Fate of
 Fuel-Nitrogen Under Conditions of Pulverized Coal
 Combustion
                                5 REPORT DATE
                                  August 1978
                                  ERFORMING ORGANIZATION CODE
7. AUTHORIS)
Y.H.Song
          J.M.Levy, J.H.Pohl, A.F.Sarofim, and
                                P. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING ORGANIZATION NAME AND ADDRESS
 Massachusetts Institute of Technology
 Department of Chemical Engineering
 Cambridge, Massachusetts  02139
                                10. PROGRAM ELEMENT NO.
                                EHE624A
                                11. CONTRACT/GRANT NO.

                                Grant R803242, Task 2
12. SPONSORING AGENCY NAME AND ADDRESS
 EPA, Office of Research and Development
 Industrial Environmental Research Laboratory
 Research Triangle Park, NC 27711
                                13. TYPE OF REPORT AND PER.OD COVERED
                                Task Final; 8/74-8/77	
                                14. SPONSORING AGENCY CODE
                                  EPA/600/13
  . SUPPLEMENTARY NOTES TRL-RTP project officer is John H. Wasser,  Mail Drop 65,
 919/541-2476.
16. ABSTRACT
          The report gives results of an experimental investigation of coal pyrolysis
 and oxidation, and char oxidation to  determine the effects of temperature and fuel/
 oxygen equivalence ratio on the conversion of coal-nitrogen to NOx.  Experiments
 involved a laboratory laminar flow furnace under conditions  representing pulverized
 coal combustors; i.e. ,  heating rates of 1000 to 1 million K/sec, temperatures of 1000
 to 2100 K, and residence times of 2 to 1000 milliseconds. The devolatilization experi-
 ments showed that no nitrogen loss occurred until 10 to 15% of the coal had been devol-
 atilized,  consistent with the hypothesis that coal-nitrogen is contained mostly in het-
 erocyclic rings. Nitrogen was completely removed from the char by prolonged heat-
 ing at above 1750 K, implying that nitrogen does not form condensed stable structures
 at pulverized flame temperatures, as does carbon. After initial fracture, loss of
 nitrogen and total weight loss are linearly correlated with a nitrogen-to-carbon
 slope of 1.25 to 1. 5. Volatile nitrogen compounds accounted for the major fraction of
 NOx produced from coal-nitrogen, especially at high temperatures and low fuel/oxy-
 gen equivalence  ratios.  The results suggest that low NOx emissions from pulverized
 coal combustion are favored by a two-stage design: the first stage operated fuel-rich
 at high temperature; the second,  fuel-lean at low temperature.
17.
                             KEY WORDS AND DOCUMENT ANALYSIS
                 DESCRIPTORS
                                          b.lDENTIFIERS/OPEN ENDED TERMS
                                                                   c. COSATI Held/Group
 Pollution
 Combustion
 Research
 Coal
 Pulverized Fuels
 Nitrogen
Pyrolysis
Oxidation
Nitrogen Oxides
Pollution Control
Stationary Sources
Fuel-Nitrogen
Char
Devolatilization
13B
21B
14B
21D

07B
07D
07C
18. DISTRIBUTION STATEMENT

 Unlimited
                     19. SECURITY CLASS (ThisReport)
                     Unclassified  ______
                    2O. SECURITY CLASS (This page)
                     Unclassified
                             187
                        22. PRICE
EPA Form 2220-1 (9-73)

-------