United States                      EPA-600/R-99-109
             Environmental Protection
             Agency                         December 1999
X>EPA     Research and
             Development
             GREENHOUSE GASES FROM SMALL-SCALE

             COMBUSTION DEVICES IN DEVELOPING

             COUNTRIES: CHARCOAL-MAKING

             KILNS IN THAILAND
             Prepared for
             Office of Air and Radiation
             and
             Policy and Program Evaluation Division
             Prepared  by

             National Risk Management
             Research Laboratory
             Research Triangle Park, NC 27711

-------
                                 FOREWORD
The U. S. Environmental Protection Agency is charged by Congress with pro-
tecting the Nation's land, air, and water resources. Under a mandate of national
environmental laws, the Agency strives to formulate and implement actions lead-
ing to a compatible balance between human activities and the ability of natural
systems to support and nurture life. To meet this mandate, EPA's research
program is providing data and technical support for solving environmental pro-
blems today and building a science knowledge base necessary to manage our eco-
logical resources wisely, understand how pollutants affect our health, and pre-
vent or reduce environmental risks in the future.
The National Risk Management Research Laboratory is the Agency's center for
investigation of technological and management approaches for reducing risks
from threats to human health and the environment. The focus of the Laboratory's
research program is on methods for the prevention and control of pollution to air,
land, water, and subsurface resources, protection of water quality in public water
systems; remediation of contaminated sites  and-groundwater; and prevention and
control of indoor air pollution. The goal of  this research effort is to catalyze
development and implementation of innovative, cost-effective environmental
technologies; develop scientific and engineering information needed by EPA to
support regulatory and policy decisions; and provide technical support and  infor-
mation transfer to  ensure effective implementation of environmental regulations
and strategies.

 This publication has been produced as part  of the Laboratory's strategic long-
 term research plan. It is published and made available by EPA's Office of Re-
 search and Development to assist the user community and  to link researchers
 with their clients.

                        E. Timothy Oppelt, Director
                        National Risk Management Research Laboratory
                        EPA REVIEW NOTICE

     This report has been peer and administratively reviewed by the U.S.  Environmental
     Protection  Agency, and approved  for publication. Mention of  trade names  or
     commercial products does not constitute endorsement or recommendation for use.

     This document is available to the public through the National Technical Information
     Service, Springfield, Virginia 2216.

-------
                                                          EPA-600/R-99-109
                                                          December 1999
GREENHOUSE GASES FROM SMALL-SCALE COMBUSTION DEVICES IN
                     DEVELOPING COUNTRIES

           Charcoal-Making Kilns in Thailand

                                 by

                    Kirk R. Smith, David M. Pennise
      Environment Program, East-West Center, Honolulu, HI 96848 and
 Environmental Health Sciences, University of California, Berkeley, CA 94720

        Pojanie Khummongkol, Varisara Chaiwong, Kwanchai Ritgeen
   Department of Environmental Technology, School of Energy and Materials
            King Mongkut' s University of Technology Thonburi
                          Bangkok, Thailand

                            Junfeng Zhang
          Environmental and Occupational Health Sciences Institute
                        Piscataway, NJ 08854

                          Winai Panyathanya
                     Thai Royal Forest Department
                          Bangkok, Thailand

                           R.A. Rasmussen
            Oregon Graduate Institute of Science and Technology
                         Beaverton, OR 97291

                            M.A.K. Khalil
                        Portland State University
                          Portland, OR 97207

       U.S. EPA Cooperative Agreement CR820243 (East-West Center)
                 EPA Project Officer: Susan A. Thorneloe
                     Atmospheric Protection Branch
               Air Pollution Prevention and Control Division
                    Research Triangle Park, NC 27711

                            Prepared for:
                  U.S. Environmental Protection Agency
                  Office of Research and Development
                        Washington, D.C. 20460

-------
                                    FOREWORD

       This work is part of a multi-year international study of the greenhouse-gas implications of
small-scale combustion devices in developing countries. To date, measurements have been made
of household stoves in India and China and charcoal kilns in Thailand.  Although individually
small, these devices are so numerous and their emission factors per unit output are so significant
that, in total, they can have  an appreciable influence on  global and national  inventories of
important greenhouse gases.

       The charcoal  kiln measurements  described  in this  report took  place  at the Charcoal
Research Centre (CRC) in Saraburi, Thailand, which is operated by the Royal Forest Department.
In Thailand,  overall management was done by Khummongkol; Pennise, Chaiwong, and Ritgeen
took the samples and conducted lab analyses; and Panyathanya and his CRC team operated the
kilns. In the USA, Rasmussen, Khalil, and Zhang were responsible for lab analyses and quality
control.  Smith  was  overall  manager of the project.   Overall financial and  administrative
management of the project took place at  the East-West Center, funded though  a  Cooperative
Agreement with the USEPA, overseen by Thorneloe.
                                     ABSTRACT

       Airborne emissions from charcoal-making kilns commonly used in the developing world
were measured during typical operating conditions.  The kilns tested were of five types:  brick
beehive, mud beehive, earth mound, rice husk mound, and single (oil) drum.  These experiments
were carried out in Thailand, a nation which produces about 7.2 million tons of charcoal per year
of the estimated 26-100 million tons produced globally.  Emission factors for the production of
charcoal were determined for the direct greenhouse gases carbon dioxide (CO2), methane (CH4),
and nitrous oxide (N2O), the indirect greenhouse gases  carbon monoxide (CO) and total non-
methane  hydrocarbons (TNMHC), as well  as total suspended particulates  (TSP).   Charcoal
production efficiency (yield) and charcoal and fuelwood composition were determined as well.

       As is generally known to be the case for charcoal making, the conversion of wood carbon
to charcoal carbon was fairly inefficient, ranging from a low of 48% for the earth mound kilns to
a maximum of 57% for the more efficient brick beehive kilns.  Average emission  factors,
expressed as grams of pollutant per kilogram of charcoal produced, for the three runs of each of
the five kiln types ranged from 970-1600 for CO2,  13-58 for CH4,  110-340 for CO,  9-95 for
TNMHC, 0.017-0.084 for N2O,  and from 0.7-4.2 for TSP. Hence, a substantial fraction of the
original fuel carbon was lost as CO2 and other products of incomplete combustion (PIC). On
average, fuelwood carbon is approximately diverted as follows:  52% to charcoal,  24% to CO2,
and 10% to PIC.  Thus, due to the higher global warming potentials of PIC relative to  CO2 on
carbon basis, such kilns can produce rather large net greenhouse gas emissions, even when the
wood is harvested renewably.
                                          II

-------
                                    CONTENTS

                                                                      Page
Foreword                                                               ii
Abstract                                                                ii
List of Figures                                                          iv
List of Tables                                                           v
Glossary                                                                vi
Acknowledgments                                                       vi

I.  Introduction and Summary                                             1
II. Conclusions and Recommendations                                     3
IE. Methods
   A. Kilns and Wood Tested                                            4
   B. Parameters Measured                                              4
   C. Summary of Experimental Procedure                                 5

IV. Results
   A. Airborne Sampling Measurements                                   9
   B. Solid Product Measurements                                        9
   C. Carbon Balance and Calculation of Emission Factors                   16
   D. GC Analysis of Canister Samples                                    20
   E. Grab Samples for Determining Emission Ratios and Emission
      Factors                                                          20

V. Discussion                                                          28
VI. References                                                          35

Appendix A: Construction and Operation of the Charcoal Kilns                 37
Appendix B: Detailed Sampling Procedure                                  47
Appendix C: Analytical Methods                                           51
Appendix D: Follow-up Experiments to Determine Condensable Liquid
            Emissions                                                  52
Appendix E: Quality Assurance Plan                                       57
                                         III

-------
                                 LIST OF FIGURES

No.         Title                                                       Page

Figure 1.    A schematic of the hood and duct system used for emissions
            sampling in the 15 kiln experiments                            7
Figure 2.    Distribution of the original wood carbon in the products of the
            charcoal-making process                                      19
Figure 3a-c.  Comparison of net emission concentration ratios for KMUTT
            data and OGIST data                                         23
Figure 4.    Carbon cycle for charcoal making in the BBH kiln and the
            GWC resulting from renewable and non-renewable harvesting
            of the wood                                                 31
Figure 5.    Carbon cycle for charcoal making in the SD kiln and the GWC
            resulting from renewable and non-renewable harvesting of the
            wood                                                       32
Figure A-l.  Diagram of the brick beehive kiln (reproduced from
            Chomcharn, 1985)                                           40
Figure A-2.  Brick beehive kiln at the Charcoal Research Centre               40
Figure A-3.  Diagram of the mud beehive kiln (reproduced from
            Chomcharn, 1985)                                           41
Figure A-4.  Ignition of the mud beehive kiln in  the hood system at the
            Charcoal Research Centre                                     41
Figure A-5.  Mud beehive kiln inside the hood system at the Charcoal
            Research Centre                                             42
Figure A-6.  Diagram of the single drum kiln (reproduced from
            Chomcharn, 1985)                                           43
Figure A-7.  Single drum kiln at the Charcoal Research Centre                 44
Figure A-8.  Diagram of the earth mound and rice husk mound kilns
            (reproduced from Chomcharn, 1985)                           45
Figure A-9.  Ignition of the rice husk mound kiln at the Charcoal Research
            Centre (reproduced from Chomcharn, 1985)                     46
Figure B-l.  A schematic of the sampling system used in the 15 kiln
            experiments                                                 48
Figure D-l.  A schematic of the sampling system used in the follow-up
            experiments                                                 56
                                          IV

-------
                                  LIST OF TABLES
No.          Title

Table 1.      Sampling summary
Table 2.      Emission sample concentrations (from KMUTT mixing bag
             data) and ambient concentrations
Table 3.      Net emission ratios of gases and TSP to CO2
Table 4.      Solid product measurements
Table 5.      Carbon analyses and calorific values of solid products
Table 6.      Charcoal and brands yields (mass, carbon, and energy bases)
Table 7a-d.   Average emission factors (a. g pollutant/kg charcoal produced;
             b. g pollutant/kg dry wood; c. g pollutant carbon/kg charcoal
             carbon produced;  d. g pollutant carbon/kg wood carbon)
Table 8.      Percent distribution of the original wood carbon in the
             products of the charcoal making process
Table 9.      Comparison of emission ratios of CO, CH/i,  and TNMHC to
             CO2 determined using OGIST canister data and KMUTT
             mixing bag data
Table 10.     Comparison of emission factors determined using OGIST
             canister data and KMUTT mixing bag data (g pollutant/kg dry
             wood)
Table 11.     Comparison of grab sample method to mixing bag method:
             emission ratios of CO, CH4, and TNMHC to CO2
Table 12.     Comparison of emission factors determined using grab sample
             method and mixing bag method (g pollutant/kg dry wood)
Table 13.     Summary of previous charcoal-making kiln  emission studies
Table 14a-b.  Estimated annual air pollution emissions from charcoal
             production in Thailand, 1996 (Total and carbon only)
Table 15.     Estimated global warming commitments (GWC) from Thai
             and world charcoal production compared to  GWC from fossil
             fuel cycle
Table D-l    Data from follow-up experiments (determination of
             condensable liquid emissions)
Table E-l    Replicates: Two bags filled simultaneously through a y-probe
Table E-2    Storage test data
Table E-3    Analyses of duplicate small bags filled from the same large
             mixed bag
Table E-4    Field and laboratory blank filter data
age
10
11
12
14
15
17

18


22


24

25

26
30

33


34

54
59
60

61
61
                                          V

-------
                                    GLOSSARY
BBH
CRC
EM
EOHSI
FID
GC
GHG
GWC
GWP
HC
KMUTT
MBH
Mt
MtC
OGIST
PIC
ppb
ppm
RHM
SD
SM
THC
TNMHC
TNMOC
TSP
VOC
brick beehive kiln
Charcoal Research Center
earth mound kiln
Environmental and Occupational Health Sciences Institute
flame ionization detector
gas chromatography
greenhouse gas
global warming commitment
global warming potential
hydrocarbons
King Mongkut' s University of Technology Thonburi
mud beehive kiln
megatons (metric)
megatons of carbon (metric)
Oregon Graduate Institute of Science and Technology
products of incomplete combustion
parts per billion (by volume or mole)
parts per million (by volume or mole)
rice husk mound kiln
single drum kiln
sawdust mound kiln
total hydrocarbons
total non-methane hydrocarbons
total non-methane organic compounds
total suspended particulates
volatile organic compounds
                                 Acknowledgments

We appreciate the patient assistance of the kiln workers at the Saraburi Charcoal Research Centre
and the GC lab technicians at KMUTT during our extended use of their facilities and labors. We
thank Nancy Adams, Robert McCrillis, and D. Bruce Harris of USEPA as well as Auke
Koopmans (Regional Wood Energy Development Programme, FAO Office, Bangkok, Thailand),
Thomas Reed (The Biomass Energy Foundation, Golden, CO), and Alex English (Ontario,
Canada) for many useful comments on earlier drafts of this report.
                                         VI

-------
                        I: INTRODUCTION AND SUMMARY

       Biomass burning plays important roles in the global carbon cycle.  Although complete
combustion of biomass produces little more than CO2 and water, most actual combustion is done
in circumstances that result in substantial  diversion  of biomass carbon  into  products  of
incomplete combustion (PIC).  Indeed, current estimates are that biomass combustion accounts
not only for 25-45% of the annual global emissions of CO2, but also for 15-50% of CO, 3-10% of
CH/j, and 24% of total non-methane organic compounds (TNMOC) (Levine, 1990, Crutzen and
Andreae,  1990,  Andreae, 1991).   CO2  and CH4  are in addition the two most important
greenhouse gases  (GHG) and  CO  and TNMOC  indirectly affect  global warming through
atmospheric chemical reactions that  in turn  affect  GHG  levels.  A good characterization  of
biomass burning thus is  important for achieving scientific understanding of the potential for
human  activities  to  engender global  warming,  as well  as   informing  the international
political/economic  discourses about what GHG mitigation measures  are warranted and who
should pay for them.

       Combustion of biomass harvested or naturally regrown on a sustainable  basis does not
cause a net increase of CO2 in the atmosphere.  Unfortunately, through deforestation and other
non-renewable practices, much burned biomass is not replaced. Even with complete recycling of
the carbon, however, a biomass fuel cycle  can produce a  net increase in  global warming
commitment (GWC) because of the emitted PIC,  which have, on average,  a higher global
warming potential  (GWP) per kilogram carbon than CO2.  As  a result of these two factors,
partially non-renewable harvesting and significant PIC production, there has been much work in
recent years to characterize biomass combustion of different kinds in different seasons around the
world (Levine, 1996).

       Most biomass combustion, whether natural or anthropogenic, is  done in circumstances in
which  access to  air is not greatly restricted.  Indeed,  for most human uses, better  combustion
efficiency is a distinct advantage. Thus, although not perfect, combustion efficiency is normally
relatively high, i.e., only a few percent of fuel carbon is diverted to PIC. There is  at least one
major exception to  this pattern, however, charcoal production, which is done basically by heating
the fuelwood in the absence of air, thus creating a higher quality fuel, but at the expense not only
of a significant loss of energy in the starting fuelwood, but also significant production of PIC.

       The Food  and Agriculture  Organization reports that about 26 million tons  (Mt)  of
charcoal were produced worldwide in 1995 (FAO,  1997) occurring largely in the developing
world.  This represents about 12% of worldwide fuelwood use (FAO, 1997).  Other estimates of
the amount of charcoal produced annually worldwide include  100 million tons (Rosillo-Calle et
al., 1996). The lower end estimates may not include  all of the production of charcoal in the non-
commercial  sector, explaining  some of the rather  large range  in the estimates.   Charcoal
production has increased at a rate  of approximately 3% per year over the period  1991-1995
(FAO,  1997).  Because of its large PIC production, it might be expected that charcoal's impact
on GWC is substantially greater than  its share of fuel demand. To date, however, the airborne
emissions from charcoal-making are poorly characterized  in existing greenhouse gas emission
databases.

-------
       The charcoal kilns used in the developing world are not easily monitored, because they
often operate in remote areas over many days or even weeks for a single run.  Emissions vary
dramatically over the run and may be released from a number of locations on the kiln, which can
be nothing more than a carefully constructed mound of earth many meters long.

       This project was designed to  characterize emissions from  the most common charcoal
kilns in the developing world. In this  phase, we worked at the Charcoal Research Centre (CRC)
in Thailand, the Asian country with the largest charcoal production.  Although  challenging
enough, conditions at the CRC were more controllable  than would  be the case at a  commercial
kiln  operating in the forest.  Thus, it provided a good location for developing and validating
methods to be used later in Africa and  Latin America.

       In addition, Thailand's charcoal production itself is substantial.  Some 7.2 million tons of
charcoal are produced per year (Thailand Department of Energy Development and Promotion,
1996).  Charcoal accounts for 9% of Thailand's total energy consumption, and 98% of use and
production takes  place  in  rural  areas  (Thailand Department of Energy Development  and
Promotion, 1996).  Thus, Thai kilns could potentially  account for  a significant part of overall
Thai GHG emissions.

       We found that average emission factors, as grams of pollutant per kilogram  of charcoal
produced, over three runs of each of the five kiln types, ranged from  970-1600 for CC>2, 13-58 for
CH4, 110-340 for CO, 9-95 for TNMHC, 0.017-0.084 for N2O, and from 0.7-4.2 for TSP.  This
means fuelwood carbon is approximately diverted as follows: 52% to charcoal, 24%  to CC>2, and
10% to PIC.  Put another way,  even  if the charcoal were burned  with absolutely  no  PIC
production during its final enduse, nearly one-fifth of the carbon available in the fuelwood has
already been released as PIC during charcoal production.

       Based on published  GWPs (using a 20-year time  horizon)  for CH4 and N2O only, we
estimate that 0.65-1.41 kg  C-CO2 (carbon as carbon  dioxide equivalents) is  emitted  per kg
charcoal produced. Based on reported distributions of production among kiln types, we estimate
that the total primary GWC of Thai kiln emissions is about 5.0 Mt C-CO2.  This is approximately
7.4% of total Thai  emissions from fossil fuel combustion. If our results accurately reflect kiln
performance worldwide, some 20 Mt C-CO2 comes  from  kilns (assuming an annual global
production of 26 Mt of charcoal), which equals about  0.3% of that from global fossil  fuel
consumption (IPCC, 1995).   (If GWPs above 1.0 are adopted for the other PIC  emitted during
charcoal making, these numbers would be higher.) Total emissions  from the entire charcoal fuel
cycle, of course, would also include the emissions from the stoves and other devices in which the
produced charcoal is burned.

       At the conclusion of this report, we compare our measured emission factors to the results
of previous studies and describe  how our methods might be applied to kilns in less-controlled
circumstances.

-------
                   II: CONCLUSIONS AND RECOMMENDATIONS

        Simple charcoal kilns of the types commonly used in developing countries convert only
some 43-47% of the energy  content of wood into charcoal.  Our measurements in Thailand
indicate that,  in  the process, Thai kilns  lose  40-50% of the wood carbon, approximately
distributed as follows: 20-25% as €62; 9-18% as gaseous products of incomplete combustion
(PIC-CO, CH4, HC); and the remainder as ash, aerosol, and brands (partially carbonized solids).
As a result of this significant diversion of carbon to PIC, which have higher global warming
potentials (GWP) than CO2, the overall global warming commitment (GWC) of the charcoal fuel
cycle is significantly higher than that from the CO2 alone. For example, if the wood is harvested
non-renewably (the carbon is not recycled), the GWC of the kiln is 1.9-3.6 times the GWC of the
charcoal burned at the stove,  assuming that the stove  charcoal is completely converted to CO2.
The GWC of non-renewable charcoal fuel cycle, therefore, is some 3.9-6.2 times that of a fossil
fuel  cycle producing the same  energy (assuming that the fossil-fuel GWC  previous to final
combustion  is 10% of the total for the fossil-fuel  cycle).   Even if the  wood is harvested
renewably (complete carbon recycling), the GWC of the  charcoal fuel cycle is still some 2-4
times greater than produced by burning an equivalent  energy content of liquid or gaseous fossil
fuel.  (These calculations account for the GWP of CH4, CO2, CO and HC. If GWPs of only the
first two are  counted, the values decrease by about 50%.)

       This  implies that charcoal  fuel  cycles are among the most  greenhouse-gas  (GHG)-
intensive in the world.  A natural gas fuel cycle, for example, would have to leak directly into the
atmosphere some 12-25% of the CH4 it delivered for combustion to emit as much GHG per unit
delivered energy as a non-renewable charcoal fuel cycle.

       These emission factors can be applied to other areas of the world where similar charcoal-
making methods are used.  This will allow for somewhat better global estimates of the inventory
of greenhouse gas and air pollutant emissions from the production of charcoal. More localized
emissions sampling is necessary, however, for accurate determination of emission factors for two
reasons. First, even given the same kiln type, there is great variability globally in both kiln sizes
and construction techniques.  Second, kiln behavior is largely dependent on operator tending
methods, which again vary greatly around the world and can even vary within the same  operator
across different charcoal making sessions.

       The results of this study are consistent with the default values listed by IPCC for charcoal
kilns in that  the IPCC values fall within the range of the kilns monitored here.  The IPCC values
are perhaps closest to those of the brick beehive kiln in Thailand.  Depending on the kiln type,
however, compared  to the  IPCC values our values range from -2x to + 1.6x for CO;  -1.4x to
+1.9x for CH4; and  -6.Ox to +1.9x  for TNMOC.   Thus, to accurately reflect actual  emissions,
there is need to specify the fraction of charcoal produced in different major kiln types.

-------
                                   Ill: METHODS

 A. Kilns and Wood Tested

       Based on knowledge of the charcoal kilns commonly used in Thailand, five types were
selected for testing (Chomcharn, 1985):

   •  Brick Beehive (BBH): beehive-shaped and charged with about 800 kg wood with a firing
     time of about 40 hours.  It is a popular, improved version of the MBH kiln.
   •  Mud Beehive (MBH): somewhat smaller using about 600 kg wood per run with a firing
     time of about 40 hours.
   •  Earth Mound (EM): layers of grass, leaves, and a final layer of dirt over a charge of about
     200 kg wood taking  about  20 hours.  This type is common throughout the developing
     world.
   •  Rice Husk Mound  (RHM): Similar to EM but with layer of rice husks plus 200 kg wood
     but taking only about  3 hours.  Most commonly used in agricultural areas.
   •  Single Drum (SD):  made from an oil drum and using about 80 kg wood with a firing time
     of about 4 hours. Portable and cheap.

On-site staff of the Thai Royal Forest Department built, operated, and maintained the kilns at the
Charcoal Research Centre (CRC) in Saraburi, Thailand. All of the kilns were built under a large
shed with a metal roof and open sides to allow research in all weather conditions.

     Two species of wood commonly used for charcoal production, Eucalyptus camaldulensis
and  Leucaena leucocephala, were chosen for these  charcoal  kiln tests.   Three runs were
conducted for  each kiln  type, where  runs 1  and 2 used eucalyptus and run 3 used leucaena.
Drawings, photos, and  other details  of the  construction and operation of the kilns are  in
Appendix A.
B. Parameters Measured

     Since the purpose of the sampling was to determine a detailed carbon balance for the kilns
as well as to quantify greenhouse-gas emissions, nearly all inputs and outputs were monitored.

1. Airborne Measurements: Concentrations of the following airborne species were measured in
   the kiln emissions as well as in the ambient air:
     •  Carbon dioxide
     •  Carbon monoxide (CO)
     •  Methane (CH4)
     •  Total non-methane hydrocarbons (TNMHC)
     •  Total suspended particulates (TSP)
     •  Nitrous oxide (N2O)

-------
2. Solid/liquid Measurements: The following fuel and solid/liquid product parameters were also
  measured:

    •  Wood (starting material): mass (amount used) and contents of moisture, carbon,  and
       energy.
    •  Charcoal (product): mass, carbon, and energy content
    •  Brands (partly carbonized wood product remaining in chamber) and ash: mass and carbon
       content
    •  TSP: mass and carbon content
    •  Condensable liquid emissions (condensables): mass and carbon content
    •  Deposited solids: mass

The wood  used in these  experiments came  from trees at the CRC  plantation in  Saraburi,
Thailand.  The trees were felled, cut, and the wood was allowed to air dry under a sheltered area
of the CRC. For each experiment, wood from  this stock was weighed and packed into the main
part of the kiln. The weight of the wood used  in the firing port of the kiln was also determined.
The firing port provides the heat to start the carbonization in the main chamber.
C. Summary of Experimental Procedure (Details in Appendix B)

       After pilot runs to polish the protocols, measurements were taken of three successful runs
of each kiln type, as shown in Table 1. Thus, the results in this study are based on 15 runs total.

       A large hood made of galvanized zinc sheet metal (3 m x 3 m at its base) was suspended
about 1 m above the ground over each tested kiln. A heavy plastic curtain was hung around the
base of the hood down to the ground.  The curtain was not placed in the area directly in front of
the kiln so as to not interfere with the kiln operators and to allow for proper circulation of air in
and around the kiln.  As soon as kiln firing began, a hood blower was  turned on along with the
sampling pump.   The blower drew the kiln emissions through the hood toward the sampling
probe via a duct system. The sampling probe was located inside of the duct parallel to the flow.
A schematic of the hood and duct system is shown in Figure 1 (also see Figure A-7 in Appendix
A).

       The sampling configuration consisted of a  sampling probe,  a TSP  sampling cassette
holding a  quartz fiber filter, a low-flow pump, and a  large (80-liter) Tedlar bag.  The pump
provided a constant sampling  flow rate  of 100-300  ml/min, depending on the kiln type.  If the
first large bag was filled before the firing process was completed, additional bags were used until
the end of firing process. One 80-liter bag (called the mixing bag) was created at the end of each
run by time-weighted filling  from the  multiple large bags used during continuous emissions
sampling. The mixing bag thereby represented a sample integrated over the entire firing process.
Small (1-liter) Tedlar bags were filled from the large bags and transported to the laboratory at
King  Mongkut's Institute  of  Technology Thonburi  (KMUTT)  in Bangkok, Thailand for  GC
analysis.  For a subsample, 850-ml stainless-steel canisters were also  filled and shipped to the

-------
Oregon Graduate Institute  of Science and Technology (OGIST) in Oregon, USA for more
detailed analysis.  Ambient air samples (large bag and TSP filter) were also collected at the CRC
before one run of each kiln type. Appendix C gives details of the analytic procedures used.
Table 1. Sampling Summary
Kiln Firing Sampling #Large Mixing Large bag Each Ambient Asc # grab
Expt. time (h) probe Bags & bag used? & filter large bag bag&filter canister samples
flow rate Filters replicate analyzed? sample filled? taken
(ml/min) used taken? taken?
BBH-1
BBH-2
BBH-3
MBH-1
MBH-2
MBH-3
SD-1
SD-2
SD-3
EM-1
EM-2
EM-3
RHM-1
RHM-2
RHM-3
40.7
34.0
38.2
38.5
41.2
60.7
7
4
5.5
21
20
24
5
4
5
100
100
100
100
100
100
250
250
250
200
200
200
300
300
300
4
3
4
4
4
5
2
1
2
4
4
5
1
1
1
Y
Y
Y
Y
Y
Y
Y
N
Y
Y
Y
Y
N
N
N
N
Y
N
N
Y
N
N
Y
N
N
Y
N
N
Y
N
N
Y
N
N
Y
N
N
Y
N
N
Y
N
Y
Y
Y
Y
Y
N
Y
N
N
Y
N
N
Y
N
N
Y
N
N
N
Y
N
N
Y
N
N
N
N
N
Y
N
N
Y
N
2
5
3
4
2
4
0
0
2
0
0
3
0
0
2
  Large bag - 80-liter Tedlar bag used for continuous emissions sampling
  Mixing bag - 80-liter Tedlar bag filled on time-weighted basis from large bags; 1-liter
  Tedlar bags were filled from the mixing bags and analyzed at KMUTT
  Filter - 37-mm quartz fiber filter used for TSP sampling
  Replicate - parallel sampling replicate taken for one integrated sample (80-liter bag and filter)
  Canister - 850-ml canister for gas samples; analyzed at OGIST
  Asc. - ascarite trap used in filling canisters (for analysis of N2O)
  Grab samples - Samples taken from directly over the top of the kiln into 1-liter Tedlar bags
       After  the end of the firing, each kiln is sealed and allowed to cool.   At this point,
charcoal, brands, and ash were removed and weighed.  Samples of each were collected and later
analyzed for carbon content and calorific value.

       Measurements were made of the emissions that deposited onto the hood and inner kiln
surfaces in order to determine the total amount of condensable material ultimately emitted from
each kiln.  Attempts were  made in these experiments to measure the amount  of condensable

-------
                                       flexible duct
              metal
              duct
metal
duct      protective
          filter
                            vented to atm.
                                                                        sampling
                                                               U-tube    probe
                 hood
                 blower
                                                      heavy plastic
                                                      curtain
Figure 1.  A schematic of the hood and duct system used for emissions sampling in the 15 kiln experiments.

-------
liquids emitted from the kilns by running the air flow through a U-tube.  This proved unworkable
and, therefore, a set of two follow-up kiln experiments (BBH and EM kilns) were performed in
order to better quantify the condensable liquid emissions (condensables).  See Appendix D for a
detailed description of the follow-up experiments.

       A number of activities were performed as part of the quality assurance plan, including
sharing of common  laboratory standards and cross-laboratory  comparisons of split samples.
Appendix E provides the details.

-------
                                    IV: RESULTS

A. Airborne Sampling Measurements

       The net concentrations of the airborne kiln  emission samples were  determined by
subtracting the ambient concentrations found at the Charcoal Research Centre.  The emission
sample concentrations for each of the fifteen kiln  runs, as determined at KMUTT from the
mixing bag samples, are shown in Table 2. The ambient sample concentrations are also shown at
the bottom of Table 2 (A-BBH-1, A-BBH-2, etc.).  Table 3 shows the net emission ratios to CO2
on a carbon basis (molar or mass) for CO, CH4, TNMHC, TSP, and total products of incomplete
combustion (PIC  = CO  + CH4 + TNMHC  + TSP).   The  ratios in Table  3,  not the net
concentrations themselves, are used in the calculation of emission factors for each of the airborne
species as shown in section C below.

B. Solid Product Measurements

       The total mass and mass of carbon in the solid species involved in each of the fifteen
charcoal  making experiments (total wood, charcoal, brands, ash, condensables, and TSP) are
shown in Table 4.  Total wood is the combination of the wood loaded into the kiln (wood in kiln)
and wood used for external firing of the kiln (wood for fire).  The wood moisture fraction was
calculated on the dry basis: moisture fraction = (wet wood mass - oven dry wood mass)/dry  wood
mass. An average value of 5% moisture content was  applied for the charcoal produced. We were
not able to determine the total mass of condensable species ultimately emitted from the kilns  in
the original 15 experiments.  Therefore,  the emission factor for condensables was  determined in a
set of two follow-up experiments using BBH and EM kilns. Carbon in the condensable liquid
emissions accounted for 2% of the original wood carbon in the follow-up BBH kiln experiment
and 4% of that in the follow-up EM experiment, as discussed in Appendix C. An average factor
of 3% of the original wood carbon was applied to estimate the amount of condensables carbon
emitted in each of the 15 original kiln experiments shown in Table 4.

       The mass  of emissions that deposited onto the large hood and inner kiln surfaces was
determined in each of the 15 experiments.  The solids  deposited on the inside of the large hood
were assumed to be 50% carbon, while those that deposited on inner kiln surfaces were assumed
to be 70% carbon (carbon analyses were not performed on  these products).  The total mass  of
carbon in these  deposited substances ranged from 0.03-0.74% of the original wood carbon for
the 15 experiments, with an average of 0.24%.  Because of their small contribution to the total
carbon outflow from the kilns, these results were not included in Table 4.

       Table 5 presents the results of the carbon content and calorific value analyses of the solid
species from  the  15 original kiln experiments.  These  carbon content values were used  in
calculating the carbon masses shown in Table 4.  Two batches (labeled "a" and  "b") of each  of
the two wood types were  analyzed for  carbon content.  The charcoal yields, or  kiln conversion
efficiencies, were  determined for each  experiment.  These data are shown in Table 6.  The dry
basis charcoal yield is the mass of charcoal produced divided by the total dry mass of wood used
in the kiln run (wood inside kiln plus wood for fire). The wet basis charcoal yield  is the ratio of

-------
Table 2. Emission sample concentrations (from KMUTT mixing bag data) and ambient
        concentrations (labeled "A-")

BBH-1
BBH-2
BBH-3
MBH-1
MBH-2
MBH-3
SD-1
SD-2
SD-3
EM-1
EM-2
EM-3
RHM-1
RHM-2
RHM-3
A-BBH-1
A-BBH-2
A-MBH-1
A-SD-1
A-EM-1
A-RHM-1
(KMUTT)
A-RHM-1
(OGIST)
C02
(ppm)
24970
15860
27770
6437
5411
7263
10680
12800
15570
2060
3286
3504
10070
13500
11930
671
618
611
551
492
490
416

CO
(ppm)
6666
4556
6102
1422
981
1147
3995
5150
4200
493
907
737
1343
1056
1376
1.12
0.58
0.35
NA
NA
NA
1.4

CH4
(ppm)
2353
1429
2212
300
223
328
1095
1410
1456
119
162
186
225
224
364
2.65
1.63
2.01
7.54
NA
NA
2.0

TNMHC
(ppm)
2285
1361
2980
266
249
410
1494
2568
1793
718
398
215
301
153
207
5.38
12.2
NA
29.4
24.9
12.4
NA

TSP
(mg/m3)
77.8
115
8.96
4.91
11.0
3.54
66.6
18.7
85.0
7.21
6.22
12.1
29.7
0.980
4.74
0.825
NA
NA
0.539
0.480
0.170
NA

N2O
(ppb)
591
572
809
475
359
380
528
NA
NA
341
484
764
591
968
801
NA
NA
NA
NA
NA
NA
335

  NA = not analyzed
charcoal produced to the total wet mass of wood used.  The carbon yield is the mass of charcoal
carbon produced  divided  by the total mass  of  carbon in the  original  wood  used.   Energy
conversion to charcoal is the ratio of the total energy content of the charcoal product to that of the
wood input.  As expected, the  better insulated, more manageable brick kilns (BBH) had the
highest average yields, for both charcoal yield and charcoal carbon yield. The earth mound kilns
(EM) had the lowest average charcoal carbon yield.  Also shown in Table 6 are the brand yields,
using both wet and dry mass bases and an energy basis. Brands  can be considered a secondary
product of charcoal production, because they are often sold as cooking fuel, commanding a price
                                           10

-------
greater than that of raw wood, but less than that of charcoal.  Brands are also often reloaded into
a kiln and fully converted to charcoal in a subsequent firing of the kiln.
Table 3. Net emission ratios of gases and TSP to CO2 (ratio of carbon; dimensionless)

BBH-1
BBH-2
BBH-3
MBH-1
MBH-2
MBH-3
SD-1
SD-2
SD-3
EM-1
EM-2
EM-3
RHM-1
RHM-2
RHM-3
CO/CO2
0.2729
0.2975
0.2241
0.2413
0.2016
0.1707
0.3943
0.4203
0.2796
0.3258
0.3313
0.2491
0.1410
0.0815
0.1209
CH4/CO2
0.0962
0.0931
0.0811
0.0504
0.0452
0.0484
0.1078
0.1148
0.0968
0.0764
0.0582
0.0618
0.0233
0.0171
0.0317
TNMHC/
C02
0.0929
0.0878
0.1089
0.0423
0.0478
0.0586
0.1458
0.2082
0.1183
0.4640
0.1393
0.0670
0.0298
0.0105
0.0167
TSP/CO2
0.00349
0.00824
0.000299
0.000810
0.00230
0.000447
0.00535
0.00122
0.00462
0.00368
0.00174
0.00324
0.00250
0.0000376
0.000312
PIC/CO 2 a
0.4655
0.4866
0.4144
0.3349
0.2969
0.2781
0.6532
0.7446
0.4993
0.8698
0.5305
0.3812
0.1967
0.1090
0.1696
N2O/CO2 b
1.28E-05
1.50E-05
2.42E-05
3.02E-05
8.02E-06
1.30E-05
1.77E-05
NA
NA
1.81E-06
3.17E-05
1.30E-04
3.70E-05
6.33E-05
5.17E-05
  NA = not analyzed
  a PIC = products of incomplete combustion = CO + CH4 + TNMHC + TSP
   units: molecular ratio
                                           11

-------
Table 4. Solid product measurements

BBH-l
BBH-2
BBH-3
MBH-1
MBH-2
MBH-3
SD-1
SD-2
SD-3
EM-1
EM-2
EM-3
RHM-1
RHM-2
RHM-3
Wet wood in
kiln (kg)
831.0
755.0
759.0
561.5
575.0
551.0
74.5
81.0
71.0
200.0
200.0
200.0
200.0
200.0
200.0
Wet wood for
fire(kg)
42.5
47.0
38.0
40.0
50.0
50.0
5.1
2.4
7.1
13.3
7.3
7.3
8.2
7.4
7.8
Total wet
wood (kg)
873.5
802.0
797.0
601.5
625.0
601.0
78.5
83.0
77.0
211.0
206.0
206.0
206.5
206.0
206.0
Wood moisture
fraction "
0.172
0.145
0.182
0.161
0.176
0.171
0.267
0.204
0.183
0.205
0.215
0.210
0.269
0.235
0.299
Total dry
wood (kg)
745.4
701.4
674.3
518.0
531.3
513.4
62.8
69.3
66.0
177.0
170.6
171.3
164.2
168.0
160.0
Total wood C
(kg)
330.0
310.5
294.6
229.3
235.2
224.3
27.8
30.7
28.8
78.4
75.5
74.8
72.7
74.4
69.9
Charcoal
(kg)
249.5
238.0
219.0
157.0
171.5
152.8
19.0
22.4
16.9
50.0
48.5
56.0
50.0
40.6
54.9
Charcoal C
(kg)
192.3
169.9
168.4
114.1
125.0
113.2
13.8
17.8
13.3
35.7
32.8
41.7
39.5
33.5
41.7
Brands
(kg)b
37.0
77.0
57.0
68.0
73.6
34.0
1.0
1.0
0.60
6.0
31.0
30.0
21.0
17.0
12.0
Brands C
(kg)
20.9
46.5
36.7
41.1
44.5
20.6
0.59
0.59
0.35
2.9
15.4
15.2
10.1
8.27
5.91
                                                                                                          (continued)

-------
Table 4. (continued)

BBH-l
BBH-2
BBH-3
MBH-1
MBH-2
MBH-3
SD-1
SD-2
SD-3
EM-1
EM-2
EM-3
RHM-1
RHM-2
RHM-3
Ash
(kg)
1.0
1.5
1.5
1.0
2.3
2.5
1.0
1.5
0.90
3.0
4.5
4.0
8.0
11.0
6.0
Ash C Condensables C
(kg) (kg) c
0.093
0.12
0.15
0.061
0.20
0.16
0.14
0.20
0.12
0.17
0.22
0.15
0.56
1.30
0.99
9.9
9.3
8.8
6.9
7.1
6.7
0.83
0.92
0.86
2.4
2.3
2.2
2.2
2.2
2.1
TSPC
(kg)d
0.254
0.469
0.0170
0.0408
0.104
0.0292
0.0401
0.00782
0.0438
0.0733
0.0282
0.0365
0.0426
0.000986
0.00512
  a Determined on a dry basis; moisture fraction = (wet mass-dry mass)/dry mass
   The mass of brands produced in the BBH-3 run was set as the average of the masses of brands produced in the other two BBH runs
  0 Set equal to 3.0% of wood C (as determined in the BBH and EM follow-up experiments)
  d TSP was determined from its ratio to CC>2 in the airborne emissions samples

-------
Table 5.  Carbon analyses and calorific values of solid products

BBH-1
BBH-2
BBH-3
MBH-1
MBH-2
MBH-3
SD-1
SD-2
SD-3
EM-1
EM-2
EM-3
RHM-1
RHM-2
RHM-3
Wood Samples
Eucalyptus (a)
Eucalyptus (b)
Leucaena (a)
Leucaena (b)
Charcoal
% carbon
80.97
75.00
80.79
76.34
76.59
77.81
76.52
83.47
82.50
75.00
71.02
78.24
83.00
86.60
79.74
Wood
% carbon
44.52
44.02
43.67
43.70
Charcoal
calorific
value (kJ/g)
29.70
28.38
30.57
28.77
29.17
29.74
NA
30.59
30.14
NA
NA
NA
NA
NA
NA
Wood
calorific
value (kJ/g)
20.41
19.82
18.88
19.62
Brands
% carbon
59.28
NA
67.72
NA
NA
NA
61.49
NA
61.41
51.48
NA
53.18
50.43
NA
51.74

Brands
calorific
value (kJ/g)
23.07
NA
26.54
NA
NA
NA
24.02
NA
23.93
NA
NA
NA
NA
NA
NA

Ash
% carbon
8.83
7.87
10.17
6.11
8.91
6.58
13.67
NA
12.96
5.82
NA
3.86
7.03
NA
16.52

TSP
% carbon
NA
54.1
46.7
NA
51.7
NA
NA
40.2
NA
NA
40.0
NA
NA
NA
NA

  NA = not analyzed
                                            14

-------
Table 6. Charcoal and brands yields (mass, carbon, and energy bases)

BBH-l
BBH-2
BBH-3
MBH-1
MBH-2
MBH-3
SD-1
SD-2
SD-3
EM-1
EM-2
EM-3
RHM-1
RHM-2
RHM-3
Charcoal yield
(wet basis) "
0.286
0.297
0.275
0.261
0.274
0.254
0.242
0.270
0.219
0.237
0.235
0.272
0.242
0.197
0.267
Charcoal yield
(dry basis) b
0.335
0.339
0.325
0.303
0.323
0.298
0.303
0.323
0.256
0.282
0.284
0.327
0.305
0.242
0.343
Charcoal
carbon yield °
0.583
0.547
0.572
0.498
0.532
0.505
0.498
0.580
0.460
0.456
0.434
0.557
0.544
0.450
0.596
Energy conversion
to charcoal
0.471
0.456
0.491
0.413
0.446
0.438
—
0.468
0.382
—
—
—
—
—
—
Brands yield
(wet basis) "
0.0424
0.0960
0.0715
0.113
0.118
0.0566
0.0127
0.0120
0.00779
0.0284
0.150
0.146
0.102
0.0825
0.0583
Brands yield
(dry basis) b
0.0496
0.110
0.0845
0.131
0.139
0.0662
0.0159
0.0144
0.00909
0.0339
0.182
0.175
0.128
0.101
0.0750
Brands carbon
yield0
0.0633
0.150
0.125
0.179
0.189
0.0916
0.0211
0.0191
0.0122
0.0375
0.204
0.203
0.139
0.111
0.0846
Brands conversion
to charcoal
0.0542
—
0.111
—
—
—
0.0181
—
0.0108
—
—
—
—
—
—
  a wet basis yield = product mass/wet wood mass
  b dry basis yield = product mass / dry wood mass
  0 carbon yield = product carbon mass / wood carbon mass

-------
C. Carbon Balance and Calculation of Emission Factors

       Starting with a carbon balance on the system, we calculated the emission factors for each
of the species of interest (CO2, CO, CH4, TNMHC, TSP, N2O). We will present the emission
factors in four different forms, each having its own utility depending on the desired end use:
       a.  grams of pollutant emitted per kilogram of dry wood loaded into the charcoal kiln
       b.  grams of pollutant emitted per kg of charcoal produced
       c.  grams of pollutant carbon emitted per kg  of wood carbon
       d.  grams of pollutant carbon emitted per kg  of charcoal carbon produced.

       The carbon balance for the charcoal making process can be  written as  follows,  on a
carbon basis:

  wood = charcoal + brands + ash + condensable liquids + CO2+ CO + CH4 + TNMHC + TSP,

dividing through by CO2 and rearranging yields:

  1 = (wood - charcoal - brands - ash - condensables)/CO2 - (CO + CH4 + TNMHC + TSP)/CO2.

We define (CO + CH4 + TNMHC + TSP)/CO2 = K.

The total amount of CO2  emitted (still in terms of carbon) was found by solving the above
equation:

                CO2 = (wood - charcoal - brands - ash - condensables)/(l + K)

The absolute CO2 emission factor was then found by dividing this total amount of CO2 emitted
by the total amount of wood used in the kiln run  or charcoal produced in the process.  The
emission factors for the other species of interest were found via their emission  ratios to CO2
(again on a carbon basis).  All four forms of the emission factors for all airborne species  were
calculated, and  the average values for the three runs per kiln  are shown in Table 7a-d.  The
coefficients of variation were calculated for the first set of emission factors and are shown in the
lower half of Table 7a.

       The distribution  of wood carbon for each of the runs and the means  for  each of the five
kiln types is shown in Table 8.  Figure 2 shows the  balance for the averages for each of the five
kilns. They range from the best performance in BBH in which  only 290 g C is diverted to PIC
and CO2  to the worst, SD, in which 435 g is lost to airborne species.
                                          16

-------
Table 7a. Average emission factors, grams of pollutant per kilogram charcoal produced, and coefficients of
        variation (CV)

brick beehive (BBH)
mud beehive (MBH)
single drum (SD)
earth mound (EM)
rice husk mound (RHM)

BBH
MBH
SD
EM
RHM
C02
966
1235
1517
1140
1570
Coefficients
0.10
0.25
0.34
0.32
0.44
Table 7b. Average emission factors, grams

brick beehive (BBH)
mud beehive (MBH)
single drum (SD)
earth mound (EM)
rice husk mound (RHM)
CO2
322
378
434
334
443
CO
162
158
336
226
106
CH4
31.8
21.7
57.7
27.7
12.7
of variation (CV)
0.16
0.19
0.15
0.43
0.19
0.16
0.27
0.25
0.44
0.16
TNMHC0
29.7
19.9
71.5
95.3
8.53
TSP
1.90
0.693
4.19
2.25
0.807
PIC
226
200
470
351
128
gases+TSP
1192
1436
1987
1491
1699
N2O
0.0166
0.0212
0.0259
0.0458
0.0843
for above average emission factors
0.15
0.41
0.08
1.11
0.36
of pollutant per kilogram dry
CO
54.2
48.4
98.1
65.7
30.9
CH4
10.6
6.63
16.6
8.09
3.71
TNMHC1
9.88
6.08
20.9
27.2
2.51
0.92
0.60
0.70
0.55
1.42
wood
TSP
0.640
0.217
1.17
0.660
0.248
Table 7c. Average emission factors, grams of pollutant C per kilogram charcoal C

brick beehive (BBH)
mud beehive (MBH)
single drum (SD)
earth mound (EM)
rice husk mound (RHM)
C02
350
460
539
441
536
Table 7d. Average emission factors, grams

brick beehive (BBH)
mud beehive (MBH)
single drum (SD)
earth mound (EM)
rice husk mound (RHM)
CO2
199
234
269
206
274
CO
92.9
92.6
188.3
137.6
57.4
CH4
31.7
22.2
56.3
29.3
12.0
TNMHC
33.8
23.2
79.5
115
9.21
TSP
0.380
0.131
0.604
0.345
0.112
0.13
0.20
0.14
0.59
0.17

PIC
75.3
61.4
136.8
101.6
37.4
produced
PIC
159
138
325
282
78.7
0.10
0.24
0.29
0.38
0.42

gases+TSP
397 0
439 0
571 0
435
480

gases+TSP
509
598
864
723
615
0.27
1.43
0.23
0.99
0.19

N2O
.00551
.00645
.00775
0.0144
0.0232

N2Ob
0.0120
0.0158
0.0192
0.0345
0.0572
of pollutant C per kilogram wood C
CO
52.7
47.1
95.4
63.8
30.0
CH4
18.0
11.3
28.3
13.7
6.31
TNMHC
19.2
11.8
40.6
52.6
4.86
TSP
0.779
0.249
1.07
0.599
0.224
PIC
90.6
70.5
165
131
41.4
gases+TSP
290 0
305 0
435 0
337
315
N2O°
.00683
.00798
.00318
0.0179
0.0287
  PIC = products of incomplete combustion = CO + CH4 + THMHC + TSP
  gases+TSP = CO2 + CO + CH4 + THMHC + TSP
  a Assuming a per carbon molecular weight of 14
  b In units of N atoms per 1000 atoms of charcoal C
  0 In units of N atoms per 1000 atoms of wood C
                                             17

-------
Table 8. Percent distribution of the original wood carbon in the products of the charcoal making process

BBH-1
BBH-2
BBH-3
MBH-1
MBH-2
MBH-3
SD-1
SD-2
SD-3
EM-1
EM-2
EM-3
RHM-1
RHM-2
RHM-3
Averages:
brick beehive
mud beehive
single drum
earth mound
rice husk mound
Charcoal
58.3
54.7
57.2
49.8
53.2
50.5
49.8
58.0
46.0
45.6
43.4
55.7
54.4
45.0
59.6

56.7
51.1
51.3
48.2
53.0
Brands
6.33
15.0
12.5
17.9
18.9
9.16
2.11
1.91
1.22
3.75
20.4
20.3
13.9
11.1
8.46

11.3
15.3
1.74
14.8
11.2
Condensables
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0

3.0
3.0
3.0
3.0
3.0
C02
22.1
18.3
19.3
21.9
19.1
29.2
27.0
20.9
32.9
25.4
21.5
15.0
23.4
35.3
23.5

19.9
23.4
26.9
20.6
27.4
CO
6.03
5.45
4.33
5.29
3.86
4.98
10.6
8.77
9.21
8.27
7.12
3.74
3.30
2.87
2.84

5.27
4.71
9.54
6.38
3.00
CH4
2.12
1.71
1.57
1.11
0.866
1.41
2.91
2.40
3.19
1.94
1.25
0.929
0.546
0.602
0.745

1.80
1.13
2.83
1.37
0.631
TNMHC
2.05
1.61
2.10
0.929
0.914
1.71
3.93
4.35
3.89
11.8
2.99
1.01
0.698
0.370
0.392

1.92
1.18
4.06
5.26
0.486
Ash
0.0281
0.0380
0.0518
0.0266
0.0871
0.0733
0.492
0.650
0.404
0.223
0.288
0.206
0.773
1.75
1.42

0.0393
0.0623
0.516
0.239
1.31
TSP
0.0770
0.151
0.00578
0.0178
0.0440
0.0130
0.144
0.0255
0.152
0.0935
0.0374
0.0487
0.0586
0.00133
0.00733

0.0779
0.0240
0.107
0.0599
0.0224
  Note: any differences between averages of the displayed sample values and the displayed averages are due to rounding.

-------
 rice husk mound
    earth mound
     single drum
    mud beehive
    brick beehive
              0%
20%
     40%              60%
Percent of Original Wood Carbon
80%
100%
                   I Charcoal  D Brands  • Condensables   DC02  ICO  1CH4  1TNMHC   BAshandTSP
Figure 2. Distribution of the original wood carbon in the products of the charcoal-making process.

-------
D. GC Analysis of Canister Samples

       The data on gaseous emissions generated from the analyses of the canister samples (from
each of the 15 kiln experiments) at OGIST (as a part of the Quality Assurance Plan - Appendix
E) were treated in the same fashion as the gaseous emissions data generated from analyses of the
mixing bag samples at KMUTT.  Emission ratios  of CO, CH4,  TNMOC (total non-methane
organic compounds), and the sum of those three species to CO2 were determined using the
OGIST GC data.  These  emission  ratios were  compared to  the ratios  determined from the
KMUTT  data set (Table 3),  as shown in Table 9.  The percent difference between the two values
is given,  using the KMUTT data as  the denominator.  Please note that KMUTT measured total
non-methane  hydrocarbons (TNMHC),  while OGIST measured total non-methane  organic
compounds (TNMOC). Although it is not possible  to make  an exact  comparison between
TNMHC  and TNMOC due to the difference in the  analytical techniques, we did display these
numbers, because the  comparison is still of some  value. This difference in the analytical
techniques  explains the greater variability (higher percent differences) between  KMUTT and
OGIST in this case compared to the inter-laboratory differences seen in the cases of CO2, CO,
and CH/j, as seen in Table 9.

For comparison,  the emission ratios  from the OGIST data were plotted against those from the
KMUTT  data, shown in Figures 3a-c along with the slope  and  correlation coefficient (r2). Note
that the r2 ranged from 0.91 (methane) to 0.44 (hydrocarbons).

The OGIST emission ratios were also used to calculate the emission factors (in units of grams of
pollutant  per kilogram wood) for each of the 15 experiments. A comparison of the OGIST-data
emission  factors  with the earlier calculated KMUTT-data  emission factors (from  Table 7) is
shown in Table  10.  Note  that because of the insensitivity of the carbon balance method to
variations in individual concentration measurements,  most differences in emission factors are
relatively small, except for TNMHC where for some single runs, they differed by a factor of two
or more.  In the first Earth Mound run, the difference reached nearly a factor of six.

E. Grab  Samples for Determining Emission Ratios and Emission Factors

       Since building of a large hood for  continuous monitoring would be impractical for kilns
in remote locations, we wished to test how well grab sampling above the chimneys of the kilns
would replicate continuous sampling in the hood.  From 2 to 5 grab samples were taken during
each of 9 kiln runs. These  grab samples were analyzed at KMUTT along with the mixing bag
samples (primary method).   Time-weighted averages of the net emission concentrations found for
each of the grab samples for a given kiln run were calculated. These grab sample average values
were used to calculate emission ratios and, subsequently, emission factors. Comparisons of the
emission  ratios and absolute emission factors determined  from the grab  sample data to those
determined from  the mixing bag data (from Tables 3 and 7) are shown in Tables 11 and 12. The
percentage differences are also given, using the mixing bag method as the denominator and are
highest for hydrocarbons.  Note that runs where only 2 grab samples were taken account for the
majority  of the instances where the percent differences were the largest (e.g., MBH-2 where there
was  a factor of  three difference for hydrocarbons).  More  grab samples  allow for  better
                                          20

-------
characterization of the  true gaseous emissions.  Although the  grab  samples  do not exactly
duplicate what was found in the mean hood samples, the variation is similar to that among hood
runs.
                                          21

-------
      Table 9. Comparison of emission ratios of CO, CH/j, and TNMHC to CC>2 determined using OGIST canister data and KMUTT mixing
              bag data

Sample name
BBH-1M
BBH-2M
BBH-2B-R1
BBH-3M
MBH-1M
MBH-2M
MBH-2B-R1
MBH-3M
SD-1M
SD-3M
EM-1M
EM-2M
EM-2-R2
EM-3M
A-RHM-1
RHM-1M
RHM-2M
RHM-3M
CO/CO2
KMUTT
0.267
0.287
0.321
0.220
0.221
0.181
0.187
0.158
0.374
0.270
0.239
0.276
0.305
0.210
0.000
0.133
0.0782
0.115
OGIST
0.252
NR
0.296
NR
0.237
0.230
0.241
0.243
0.206
NR
0.211
0.176
0.252
0.145
0.00337
0.105
0.0889
0.116
difference
-5.7
—
-7.9
—
7.3
27
29
54
-45
—
-12
-36
-18
-31
—
-21
14
0.17
CH4/CO2
KMUTT
0.0942
0.0901
0.107
0.0796
0.0466
0.0412
0.0332
0.0452
0.103
0.0935
0.0578
0.0493
0.0654
0.0531
0.000
0.0224
0.0166
0.0305
OGIST
0.0934
NR
0.0687
NR
0.0465
0.0493
0.0400
0.0624
0.0653
NR
0.0673
0.0422
0.0665
0.0449
0.00481
0.0234
0.0197
0.0255
difference
-0.90
—
-36
—
-0.20
20
21
38
-36
—
17
-14
1.6
-15
—
4.7
19
-17
TNMHC/CO2
KMUTT OGIST a
0.0343
0.0312
0.0559
0.0390
0.0147
0.0164
0.0217
0.0207
0.0510
0.0418
0.127
0.0440
0.0253
0.0222
0.00918
0.0110
0.00412
0.00652
0.0285
0.0387
0.0509
0.0467
0.0298
0.0293
0.0222
0.0347
0.0284
0.0509
0.0192
0.0130
0.0296
0.0194
0.00130
0.0135
0.00880
0.0149
difference
-15
24
-8.8
20
103
79
2.4
68
-44
22
-85
-70
17
-12
-86
22
114
129
(CO+CH4+TNMHC)/CO2
KMUTT
0.395
0.408
0.484
0.338
0.282
0.239
0.242
0.224
0.528
0.405
0.424
0.369
0.396
0.286
0.00918
0.167
0.0990
0.152
OGIST difference
0.374 -5.3
NR
0.415 -14
NR
0.313 11
0.308 29
0.303 25
0.340 52
0.299 -43
NR
0.298 -30
0.231 -37
0.348 -12
0.210 -27
0.00948 3.2
0.142 -15
0.117 19
0.156 2.3
to
to
       Note: any differences between percent differences of the displayed values and the displayed percent differences are due to rounding.
       NR = not reported
       a OGIST measured TNMOC (total non-methane organic compounds), while KMUTT measured TNMHC (total non-methane
        hydrocarbons).

-------
               0.35
                                      CO/CO2 Ratios
               0.00
                  0.00     0.05    0.10     0.15    0.20     0.25
                                           KMUTT Data
                                                              0.30
                                                                     0.35     0.40
Figure 3a. Comparison of CO/CO2 emission ratios from KMUTT data and OGIST data
               0.10
                                     CH4/CO2 Ratios
           co
           8
               0.00
                  0.00
                           0.02
                                     0.04
                                               0.06
                                            KMUTT Data
                                                         0.08
                                                                   0.10
                                                                             0.12
Figure 3b. Comparison of CH4/CO2 emission ratios from KMUTT data and OGIST data
                                 TNMHC/CO2 and TNMOC/CO2 Ratios
                            0.01
                                      0.02       0.03       0.04
                                        KMUTT Data (TNMHC/CO2)
                                                                    0.05
                                                                              0.06
Figure 3c. Comparison of TNMHC/CO2 emission ratios from KMUTT data to TNMOC/CO2
          ratios from OGIST data
                                            23

-------
      Table 10. Comparison of emission factors determined using OGIST canister GC data and KMUTT mixing bag GC data (g
              pollutant/kg dry wood)


BBH-1
MBH-1
MBH-2
MBH-3
SD-1
EM-1
EM-2
EM-3
RHM-1
RHM-2
RHM-3
CO2
OGIST KMUTT
data data
365 358
340 356
285 311
412 467
529 438
561 412
418 348
261 241
385 379
558 572
370 376
CO
OGIST KMUTT
data data
59.7 62.2
55.8 54.7
47.4 39.9
71.4 50.8
71.9 110
84.7 85.4
50.9 73.5
27.1 38.2
27.2 34.1
32.8 29.7
28.4 28.9
CH4
OGIST KMUTT
data data
12.7 12.6
6.22 6.54
5.77 5.12
10.4 8.23
13.0 17.2
15.3 11.5
6.92 7.38
4.75 5.42
3.44 3.22
4.13 3.56
3.56 4.35
TNMHC"
OGIST KMUTT
data data
9.31 10.6
9.64 4.80
8.30 4.73
14.0 8.72
13.7 20.3
10.5 60.9
5.13 15.5
4.95 5.14
4.84 3.61
4.45 1.91
5.03 2.00
TSPb
OGIST KMUTT
data data
0.784 0.631
0.185 0.152
0.563 0.377
0.211 0.122
1.78 1.59
0.636 1.03
0.288 0.413
0.515 0.532
0.904 0.648
0.0185 0.0147
0.0993 0.0800
PIC
OGIST KMUTT
data data
82.5 86.0
71.9 66.2
62.1 50.1
96.0 67.8
100.4 149.0
111.2 158.8
63.2 96.7
37.3 49.2
36.4 41.5
41.4 35.2
37.1 35.4
gases+TSP
OGIST KMUTT
data data
448 444
412 422
347 361
508 535
630 587
672 571
481 445
298 290
422 421
599 608
407 412
to
       a Assuming a per carbon molecular weight of 14
       b TSP was not determined at OGIST. The "OGIST" TSP emission factors were determined using the KMUTT TSP data along with
         CO2 data from OGIST canister analyses.

-------
     Table 11. Comparison of grab sample method to mixing bag method: emission ratios of CO, CH/i, and TNMHC to CO2

BBH-l
BBH-2
BBH-3
MBH-1
MBH-2
MBH-3
SD-3
EM-3
RHM-3
# of
grabs
taken
2
5
3
4
2
4
2
o
J
2
CO/CO2
grab mixing %
bag difference
0.302 0.273 11
0.317 0.298 6.6
0.231 0.224 2.9
0.259 0.242 7.1
0.306 0.203 51
0.120 0.171 -30
0.356 0.280 27
0.288 0.251 14
0.0822 0.121 -32
C7VC02
grab mixing %
bag difference
0.125 0.0963 30
0.123 0.0933 32
0.134 0.0812 65
0.0521 0.0508 2.7
0.0770 0.0456 69
0.0303 0.0487 -38
0.153 0.0970 58
0.0778 0.0627 24
0.0129 0.0318 -59
TNMHC/CO2
grab mixing %
bag difference
0.121 0.0930 30
0.150 0.0889 69
0.0837 0.109 -23
0.0645 0.0452 43
0.242 0.0513 371
0.0301 0.0597 -50
0.262 0.119 121
0.0700 0.0685 2.2
0.0111 0.0170 -35
(CO+CH4+TNMHC)/CO2
grab mixing %
bag difference
0.549 0.463 19
0.591 0.480 23
0.449 0.415 8.3
0.376 0.338 11
0.624 0.300 108
0.180 0.280 -36
0.771 0.496 56
0.435 0.382 14
0.106 0.170 -37
to
       Note: any differences between percent differences of the displayed values and the displayed percent differences are due to rounding.

-------
       Table 12.  Comparison of emission factors determined using grab sample method and mixing bag method (g pollutant/
                 kg dry wood)
CO2
grab
method
BBH-1 338
BBH-2 276
BBH-3 302
MBH-1 345
MBH-2 247
MBH-3 506
SD-3 445
EM-3 231
RHM-3 398
mixing
bag
358
297
309
356
311
467
527
241
376
%
difference
-5.5
-7.0
-2.4
-3.1
-20
8.3
-16
-4.1
5.7
CO
grab
method
65.1
55.9
44.3
57.0
48.2
38.5
101
42.2
20.8
mixing
bag
62.2
56.3
44.1
54.7
39.9
50.8
93.8
38.2
28.9
%
difference
4.6
-0.57
0.53
4.2
21
-24
7.3
11
-28
CH4
grab
method
15.4
12.4
14.8
6.55
6.94
5.59
24.8
6.54
1.87
mixing
bag
12.6
10.1
9.13
6.54
5.12
8.23
18.6
5.42
4.35
%
difference
23
23
62
0.24
36
-32
34
21
-57
TNMHC"
grab
method
13.1
13.2
8.05
7.10
19.1
4.85
37.2
5.15
1.40
mixing
bag
10.6
8.32
10.7
4.80
4.73
8.72
19.9
5.14
2.00
%
difference
24
59
-25
48
303
-44
87
0.24
-30
to
                                                                                                                (continued)

-------
Table 12.  (continued)
TSP

BBH-1
BBH-2
BBH-3
MBH-1
MBH-2
MBH-3
SD-3
EM-3
RHM-3
grab
method
0.450
1.11
0.0548
0.148
0.499
0.0731
1.88
0.755
0.0814
mixing
bag
0.631
1.24
0.0541
0.152
0.377
0.122
1.65
0.532
0.0800
%
difference
-29
-10
1.4
-3.1
32
-40
14
42
1.7
PIC
grab
method
94.0
82.6
67.2
70.8
74.6
49.0
164
54.7
24.2
mixing
bag
86.0
75.9
64.0
66.2
50.1
67.8
134
49.2
35.4
%
difference
9.3
8.8
5.0
7.0
49
-28
23
11
-32
gases + TSP
grab
method
432
359
369
416
322
555
609
285
422
mixing
bag
444
373
373
422
361
535
661
290
412
%
difference
-2.7
-3.8
-1.1
-1.5
-11
3.7
-7.8
-1.5
2.5
  Note: any differences between percent differences of the displayed values and the displayed percent differences are due to rounding.
  a Assuming a per carbon molecular weight of 14

-------
                                   V: DISCUSSION

       Table 13 presents a summary of earlier charcoal kiln emission studies.  The study by
Brocard et al. utilized a sampling probe located directly inside of the tested kiln (for some of the
samples), and, hence, the results may not so accurately reflect what is ultimately emitted from the
kiln. Note that Missouri kilns (USEPA, 1995) are industrial kilns much larger (17.6 tons or more
of charcoal produced per cycle) than those commonly used in developing nations, where almost
all of the world's charcoal is produced.

       As shown in Table 13, the results of this study roughly validate default values listed by
IPCC for charcoal kilns in that the IPCC values (IPCC, 1997) fall within the range of the kilns
monitored here.1  The IPCC values are  closest to those of the brick beehive kiln in Thailand.
Depending on the kiln type, however, compared to the IPCC default values, our values range
from -2x to + 1.6x for CO;  -1.4x to +1.9x for CH4; and -6.0x to +1.9x for TNMOC.  Thus, to
accurately reflect actual emissions, there is need to specify the fraction of charcoal produced in
each major kiln type.

       One way of looking at the greenhouse-gas implications of charcoal-making can be seen
in Figures 4 and 5 for the two most extreme kilns tested, BBH and SD, respectively.  Here, two
types of global warming commitments (GWC) are applied.  The first, called "primary GWC"
assumes GWPs only for CO2 CH4, and N2O (respectively, 1, 23, and 290 by mole - IPCC, 1995).
The second, "total GWC," applies GWPs also for CO and hydrocarbons, which are less certain
(respectively, 4.5 and 12 - IPCC, 1990).  For each, the GWC is calculated under two  extreme
assumptions: that the kilns  rely on a completely renewable wood supply,  i.e. all  carbon is
recycled back into trees, compared to the case in which there is complete deforestation, i.e., no
carbon recycling to biomass. In the latter case, of course, the GWC includes the full complement
from CO2 (GWP=1.0).  In the former, no  contribution from CO2 is included.

       We estimated the amount of charcoal produced in  each type of kiln in Thailand.  The
amounts were  based upon data on  charcoal  consumption  in  Thailand in 1996  (Thailand
Department of Energy Development and  Promotion, 1996). The fractional use of each kiln type
throughout Thailand is shown in the left side of Table 14a-b.  Note that sawdust mounds  are
kilns used to produce charcoal with wood waste from sawmills.  Production of charcoal from
sawdust mounds has risen rapidly in Thailand during the past several years. Although we did not
perform tests on sawdust mound kilns, we believe their emissions patterns to be similar to  the
rice husk mound because of their  similarity of design. Using the annual amount of charcoal
produced  in Thailand,  the fractional use  pattern, and  the  experimental  emission  factors
determined in this study, the total amounts of greenhouse gases emitted from charcoal production
in Thailand were calculated.  This is shown in Table 14a-b as megatons of pollutant (Mt) and as
megatons of pollutant carbon (MtC).
1 Oddly, however, the default charcoal conversion factor used by IPCC, 21%, is substantially lower than we
observed. This may be due to its use of a different definition of "conversion efficiency," which the IPCC document
unfortunately does not explain.


                                           28

-------
       Finally, using 20-year global warming  potentials (GWPs), the total global warming
commitments (GWCs) for Thailand's and the world's charcoal kilns are estimated in Table 15.
This makes the extreme assumption that the mean emissions for Thai kilns can reasonably be
applied to all the world's kilns.  Shown for both Thailand and the world is a two-by-two matrix
of GWCs, divided as discussed above for Figures 4 and 5.  The global GWC numbers are based
on the assumption that all of the charcoal produced globally comes from the following mix of
kiln  types: 50% EM  (roughly representing the  mobile informal-sector2 kilns of Africa), 25%
BBH (roughly representing the large commercial kilns of Brazil), and 25% from the mix of kilns
as in Thailand (roughly representing Asia).  Also shown for  comparison are the total estimated
fossil-fuel contributions to the GWC of Thailand and the world. To pinpoint these values more
precisely, however, we are now proceeding to measure emissions  from African and Brazilian
kilns.

       A full analysis of the charcoal fuel cycle for comparison with other fuel cycles would not
only require evaluation of the final end use, e.g., cookstoves, as well as the kilns, but also the
alternate fate of the wood input if it were not used for charcoal making.  For example, in some
locations part of the fuel may decay anaerobically in the environment  leading to considerable
CH4 emissions. In such a case, the  accounting should allocate only the net change in emissions
to the charcoal fuel cycle.
2 An informal-sector charcoal kiln is one that is not registered with the authorities, usually made of temporary
materials, and is often constructed by one or a few households only in those times of year when agricultural work
demand is low.
                                            29

-------
     Table 13.  Summary of previous charcoal-making kiln emission studies
Study

present study
present study
present study
present study
present study
Brocard et al.
(1996)
EPA AP-42
(1995)
IPCC Reference
Manual (1997)
Shah et al.
(1992)
Kiln Type

brick beehive
mud beehive
single drum
earth mound
rice husk mound
African earth
mound
Missouri
(world average)
metal partial-
combustion kiln
% Charcoal yield
(charcoal mass/
dry wood mass)
33.3
30.8
29.4
29.8
29.7
27.6

—
20.8 (d)
32.7
Emission factors
CO,
966
1235
1517
1140
1570
1593

550
—
1192
CO
162
158
336
226
106
254

145
210
336
(g of pollutant per kg
CH4 TNMHC N2O
31.8
21.7
57.7
27.7
12.7
39

55
30
(e)
29.7 0.0166
19.9 0.0212
71.5 0.0259
95.3 0.0458
8.53 0.0843
7.2 0.11
(asC)
80 (b)
51
72 (f)
of charcoal produced)
NOX TSP
1.90
0.693
4.19
2.25
0.807
0.24 14
(asC)
12 (c)
0.3
—
Condensables
(tars and oils)
58
63
66
65
65
	

155 (c)
—
133
(g)
Acids
...
...
...
...
...
0.72
(as C) (a)
—
—
71
OJ
o
       (a): sum of formic and acetic acids only
       (b): derived by subtracting CH4 from the emission factor given for volatile organic compounds (VOC)
       (c): TSP was included in the condensables
       (d): Converted to a dry basis by assuming 20% moisture in wood.
       (e): CH4 included in the TNMHC column
       (f): sum of CH/i, ethane, and ethene
       (g): sum of tars, phenols, and furfurals

-------
0.15kgBrands-C (4.3 MJ)

 0.040 kg Condensables-C

       0.00053 kg Ash-C

        0.0011 kg TSP-C
                                             3 kg Dry Wood
                                             (1.3 kg Carbon)
                                               (59.5 MJ)
                                1 kg Charcoal
                                (0.75 kg Carbon)
                                (28.1MJ)
      No CO2 Recycled
Primary GWC = 0.79 kg C-CO2

Total GWC = 1.40 kg C-CO2
0.26 kg CO2-C

0.068 kg CO-C

0.023 kg CH4-C

0.025 kg TNMHC-C

0.000017 kg N2O
     All CO2 Recycled
Primary GWC = 0.53 kg C-CO2

TotalGWC=1.14kgC-CO2
Figure 4. Carbon cycle for charcoal making in the BBH kiln and the GWC resulting from renewable and non-renewable
        harvesting of the wood.

-------
        0.025 kg Brands-C  (1.1 MJ)

          0.045 kg Condensables-C

                  0.0078 kg Ash-C

                  0.0016 kg TSP-C
                                                     3.4 kg Dry Wood
                                                     (1.5 kg Carbon)
                                                       (67.4 MJ)
     SDkiln
>_.
'
1 kg Charcoal
(0.77 kg Carbon)
(29.0 MJ)
OJ
to
                No CO2 Recycled
          Primary GWC = 1.41 kg C-CO2

          Total GWC = 2.79 kg C-CO2
0.40 kg CO2-C

0.14kgCO-C

0.042 kg CH4-C

0.061 kgTNMHC-C

0.000026 kg N2O
     All CO2 Recycled
Primary GWC = 1.00 kg C-CO2

Total GWC = 2.38 kg C-CO2
         Figure 5. Carbon cycle for charcoal making in the SD kiln and the GWC resulting from renewable and non-renewable
                  harvesting of the wood.

-------
Table 14a. Estimated annual air pollution emissions from charcoal production in Thailand,
           1996 a
Kiln type
BBH and MBH
EM and RHM
Sawdust
Mound
TOTAL
fractional use
in Thailand
0.26
0.21
0.53

1.0
Table 14b. Estimated annual air
Thailand (1 996) a
Kiln type
BBH and MBH
EM and RHM
Sawdust
Mound
TOTAL
fractional use
in Thailand
0.26
0.21
0.53

1.0
C02
(Mt)
2.1
2.0
6.0

10.1
pollution
C02
(MtC)
0.56
0.56
1.6

2.8
CO
(Mt)
0.30
0.25
0.41

0.96
emissions
CO
(MtC)
0.13
0.11
0.17

0.41
CH4
(Mt)
0.050
0.030
0.048

0.129
(as carbon)
CH4
(MtC)
0.038
0.023
0.036

0.097
TNMHC
(Mt)b
0.047
0.078
0.033

0.157
TSP
(Mt)
0.0024
0.0023
0.0031

0.0078
N2O
(Mt)
0.000036
0.000098
0.00032

0.00046
from charcoal production in
TNMHC
(MtC)
0.040
0.067
0.028

0.135
TSP
(MtC)
0.0012
0.00092
0.0012

0.0034
N2O
(MtN)
0.000023
0.000062
0.00021

0.00029
  Note: any differences between summations of the displayed values and the displayed TOTAL
  values are due to rounding.
  a Using a total production of charcoal in Thailand of 7.2 Mt
  b Assuming a per carbon molecular weight of 14
                                          33

-------
Table 15. Estimated global warming commitments (GWC) from Thai and world charcoal
         production compared to GWC from fossil fuel cycle


CC>2 recycled
CC>2 not recycled
GWC from fossil fuel use °
Thailand
primary GWC

2.26
5.02
67.4
Thailand
total GWC
(MtC as CO 2
5.72
8.49
78.9
World
Primary GWC a
equivalents)
11.8
20.1
7887
World
total GWC a

37.4
45.6
9749
  a Assuming global charcoal is produced from the following mix of kiln types: 50% EM
   (Africa), 25% BBH (Brazil), and 25% from a mix of kilns as in Thailand (Asia)
   Using 20-year global warming potentials
  0 For Thailand, primary GWC includes CO2 and CH4 only and total GWC includes CO2, CH4,
   and CO only
                                          34

-------
                                 VI: REFERENCES

Andreae, M.O., 1991, "Biomass burning:  its history, use, and distribution and its impact in
environmental quality and global climate," In Global Biomass Burning, J. S. Levine, editor. The
MIT Press, Cambridge, MA, pp. 3-21.

Brocard, D., Lacaux, C., Lacaux, J.-P., Kouadio, G., and Yoboue, V., 1996, "Emissions from the
Combustion of Biofuels in Western Africa," In:  Biomass Burning and Global Change, Vol. 1.
Joel  S. Levine,  editor. The MIT Press, Cambridge, MA,  pp. 350-360.   See also, "Emissions
Atmospheriques des Combustions Domestiques," Doctoral thesis  of D. Brocard,  L'Universite
Paul Sabatier de Toulouse, #2477, Nov. 1996.

Chomcharn, A., 1985, Charcoal Production Improvement  for Rural Development in Thailand,
Forest Products Research Division, Royal Forest Department, Bangkok, Thailand.

Crutzen,  PJ. and Andreae, M.O., 1990,  "Biomass Burning  in  the Tropics:  Impact on
Atmospheric Chemistry and Biogeochemical Cycles," Science 250:  1169-1178.

FAO, Food and Agriculture Organization,  (1997), Yearbook  of  Forest Products  1995, FAO
Forestry Series No. 30, Rome, Italy.

IPCC, Intergovernmental  Panel on  Climate Change,  1995, Climate  Change 1994: Radiative
Forcing  of Climate  Change. J.T.  Houghton et al., editors.  Cambridge University Press,
Cambridge, UK.

IPCC, Intergovernmental Panel on Climate Change, 1997, Guidelines for National Greenhouse
Gas  Inventories: Reference Manual (revised 1996), Volume  3. J.T. Houghton et al., editors.
Bracknell, UK.

Johnson, R.L., Shah, J.J., Gary, R.A., and Huntzicker, J.J., 1981, " An automated thermal-optical
method  for the analysis  of carbonaceous aerosol."  In ACS  Symposium Series, No.  167,
Atmospheric Aerosol: Surface/Air  Quality Relationships.  Edward S. Macias  and Philip K.
Hopke, editors.

Levine,  J.S.,  1990,  "Global  Biomass  Burning: Atmospheric, Climatic, and Biospheric
Implications," EOS 71(37): 1075-1077.

Levine,  J.S., ed.,  1996,   Biomass Burning and Global  Change:  Vols.  1  & 2,  MIT Press:
Cambridge, MA.

Rasmussen, R.A.  and Khalil, M.A.K.,  1980, "Atmospheric  halocarbons:  measurements and
analyses of selected trace gases." In: Proceedings of  the NATO Advanced Study Institute on
Atmospheric Ozone: Its Variation and Human Influences. A.C. Aiken, Editor. U.S. Dept. Of
Transportation, Washington, DC, pp 209-231.
                                         35

-------
Rasmussen, R.A. and Khalil, M.A.K., 1981, "Atmospheric methane: trends and seasonal cycles,"
J. Geophysical Res. 86: 5172-5178.

Rasmussen, R.A., Khalil, M.A.K., and Chang, J.S., 1982, "Atmospheric trace gases over China,"
Environmental Science and Technology, 16: 124-126.

Rosillo-Calle, F., de Rezende,  M.A.A., Furtado, P., and Hall,  D.O.,  1996,  The  Charcoal
Dilemma, Intermediate Technologies Publications, London, UK.

Shah, N., Girard, P., Mezerette, C., and Vergnet, A.M., 1992, "Wood-to-charcoal conversion in a
partial-combustion kiln: an experimental study to understand and upgrade the process," Fuel, 71:
955-962.

Thailand Department of Energy Development and Promotion, 1996, Thailand Energy Situation
1996.

USEPA, 1995,  Compilation of Air Pollution Emission Factors, Vol. 1,  Supplement A (GPO
055-000-005-001),  AP-42, Environmental Protection Agency, Office of Air Quality Planning
and Standards, Research Triangle Park, NC.
                                          36

-------
             Appendix A: Construction and Operation of the Charcoal Kilns

       All kilns were constructed and operated by the CRC staff, which has had long experience
with these models and wood species.

Brick Beehive Kiln (BBH)

•  Construction: The brick beehive kiln used was a typical commercial charcoal kiln, but slightly
   smaller in size. The raw materials for constructing the BBH kiln were bricks, clay,  and sand.
   An area of 4 m by 4 m was cleared for the kiln. The foundation was dug out to a depth of three
   layers of brick. When  three brick layers of foundation were constructed, spaces for four
   chimneys, a loading port, and a firing port were left before the base wall was begun.  The brick
   arrangement used for the walls from  the foundation to the top of the kiln was the same as in
   normal brick construction. The cementing material was made from soft clay and sand. A space
   for an accelerator hole was left in the wall at 0.8 m above the kiln floor opposite the space for
   the firing port. Four chimneys were constructed after completion of the kiln wall. Drying and
   curing of the binding materials was carried out by burning 1-2 armfuls of firewood  inside the
   kiln for 3-4 hours. A diagram and a picture of the BBH kiln are given in Figures A-l and A-2,
   respectively.

•  Operation: The kiln was loaded by  vertically stacking smaller pieces of wood on the kiln floor
   and placing larger pieces above the  smaller pieces and near the firing port.  The wood  was
   stacked as close as possible in order to produce more charcoal per kiln run. Areas adjacent to
   the chimney outlet holes were left  clear of wood for at least a 10 cm by 10 cm area. The wood
   was loosely packed near the firing  port. The kiln was lit by combusting firewood in the firing
   port. The kiln  temperature  was  increased  slowly after  firing  via  the  hot air  from the
   combusting wood in the firing port. The  hot air passed into the kiln by way of the accelerator
   hole and chimneys, thereby replacing the cooler air inside the kiln. The accelerator hole  was
   closed when the smoke temperature  was approximately 120 °C so that the chimneys could
   perform their normal function  thereafter. The kiln temperature was raised slowly until  it
   reached 180-200 °C. The smoke then became thicker and whiter indicating the beginning of
   carbonization in the front and top of the kiln. The hot combustion air from the firing port was
   then limited by reducing the area of the firing port opening to about 6  cm by 6 cm. The firing
   port size was thereafter monitored and adjusted to allow for proper cold air inlet by  observing
   the emitted smoke until complete carbonization took place. The smoke turned bluish when the
   kiln temperature was greater than 450  °C and then became  more clear at the end  of the
   carbonization process, when the temperature was higher than 500 °C  . Each chimney  of the
   kiln was sealed when the tar inside the outlet of the chimney was dry and hardened. The run
   was completed when all chimneys were sealed. The kiln required about 2 days for cooling.

Mud Beehive Kiln (MBH)

•  Construction: The MBH kiln used in  these experiments was typical of the charcoal kilns used
   in rural areas of Thailand. The material used for construction of the MBH kiln wall was a soft
   lateritic mud or  sandy clay. The kiln  floor was dug below ground level. All of the wood  was
                                           37

-------
   placed in the pit. Small pieces of wood, grass, and leaves were placed on top of the bundle.
   This pile was then covered with 10-15 cm of mud. The kiln wall was compacted by pounding
   with a flat piece of wood. The soft mud wall dried and hardened after a few days and was then
   self-supporting. Cracks in the dried mud wall were sealed with a slurry of mud and water. The
   design also included a firing port, one accelerator hole, and two chimneys. See Figures A-3,
   A-4, and A-5.

•  Operation: see Brick Beehive Kiln (BBH) Operation (above)

Single Drum Kiln (SD)

•  Construction: The single drum kiln used in these experiments was made from a typical 200
   liter oil drum. It came from the Philippines where it was primarily used to produce charcoal
   from coconut shells. Because of its small size and durability, it is a mobile kiln. The drum was
   90 cm high with a 57 cm diameter.  There were four air inlet holes in the base of the drum,
   each 2.5 cm in diameter. Air inlet holes of the same diameter were also placed in the side of
   the drum 5 cm, 35 cm, and 65 cm up from the base of the drum. See Figures A-6 and A-7.

•  Operation: The operation of the SD kiln began with  combustion and was followed by
   carbonization with direct draft. The fire was set at the bottom of the kiln. The wood was then
   loaded into the top of the drum.  The kiln air inlet holes were plugged when the  smoke emitted
   from the chimney became more clear, indicating that the transformation to charcoal had been
   completed. The cooling period was about 6 hours.

Rice Husk Mound Kiln (RHM)

•  Construction: This kiln was  made by simply digging a shallow pit in the ground. The  wood
   was loaded horizontally into the pit and covered with rice husks. The RFDVI kiln  did  not
   require a chimney. See Figures A-8 and A-9.

•  Operation: The fire was lit at the bottom of the piled wood. The entire wood pile was  allowed
   to burn with free access to air until the wood bark was in flames. This required about 1  hour.
   The wood pile was then covered with about 6 inches of rice husks. The supply of air to the
   carbonization reaction was achieved via the porous nature of the rice husk pile. At the end of
   the carbonization process, the smoke became more colorless. The kiln was then sealed with a
   covering of soft, sandy soil  or a metal sheet and sealed with clay. Two or three days were
   required for cooling.

Earth Mound Kiln (EM)

•  Construction: This kiln was also made by simply  digging a shallow pit in the ground. The
   wood was loaded horizontally into the pit and covered with grass or leaves and then  earth. The
   layer of grass or leaves was thick enough to prevent the earth from directly touching the pile
   of wood. An opening was left in the earth mound covering for a chimney. See Figure A-8.
                                          38

-------
•  Operation: The operation of the EM kiln was similar to that of the RHM kiln. The entire wood
   pile was covered with grass or leaves and then earth, except for one area required for lighting,
   located opposite of the chimney. The fire was lit at the opening with a small amount of
   firewood. The pile was then allowed to burn with free access to air until the wood bark was in
   flames. At this point the flaming end was covered with earth. Reverse draft and control of the
   air supply was achieved through the chimney.  At the end of the carbonization  process, the
   chimney was removed and the hole  was covered with earth.  The kiln was then sealed in the
   manner as the RHM kiln.

Age of Kilns

The beehive kilns used in these experiments were mature, having been fired by CRC many times
previous to our use of them. They thus represent a more typical case than would newly made
kilns, which take some time for their performance to stabilize. The single drum kiln was a newly
made device, but of an old drum as is the common practice in the field. Rice husk and earth
mound kilns, of course, are only used once since they are destroyed in the process of unloading
them.
                                          39

-------
fcv
r\


1_U!
4ff— 	 - • 	 	 fThlmiiey ., .
fri 7f*$
•s/ / ,,,., --
, — -





M_

t











<

Ghligixsy hole




l^m.iM-tiM
                                                    Chimney cross—sectioa
        Bottom crosa-seccion
                                              T
                                             .SO m
.ccelerating
      hole
   C/07X.07 m)
                               Loading port


                               Firing port
         Front
  rl  U07X.O? m)
m 1	I	—Chimney
  L_L       hole
       (.04x,04 m)
                                                          Side viaw
Figure A-l. Diagram of the brick beehive kiln (reproduced from Chomcharn, 1985).
         Figure A-2. Brick beehive kiln at the Charcoal Research Centre.
                                          40

-------
                                                                         Accelerating
                                                                            hole
                                                                            0.08 m
                                                                         Ch.ism.ey
Firing port


Front view
                                                           Side view
   Figure A-3. Diagram of the mud beehive kiln (reproduced from Chomcharn, 1985).
Figure A-4. Ignition of the mud beehive kiln in the hood system at the Charcoal Research Centre.
                                           41

-------
Figure A-5. Mud beehive kiln inside the hood system at the Charcoal Research Centre.
                                     42

-------
         Handle
 Air Inlets
      0 2.5 cm
                                                             si
                                                             o
                                                                     Side view
     Air inlets
               0  2.5
                                                                  Bottom view
        Chimney


         0 10 cm
                                                                  Top view


                                                            (Lid Cross-Section)
Figure A-6. Diagram of the single drum kiln (reproduced from Chomcharn, 1985).
                                        43

-------
Figure A-7. Single drum kiln at the Charcoal Research Centre.
                           44

-------
                 -©-•
               150 cm
                                150 cm
                                 1:25
                        4

                                                      -fc^.
                                                      jr^Cfor'• SM&EM only)
     rice husk
     or sawdust
     or earth

\
                                      0
                                                 'L
                              layers  of leaves
                              
-------
                                                          rfV,sJf
                                             "-': ,"•'; ' ''.. .  :;~~ '-.r.5' '':"'f j^^^g'^-4l^?'^-: l'i''
                                              "  -^ ..^   C^iJfS:^---:'.'^ .1?^
Figure A-9. Ignition of the rice husk mound kiln at the Charcoal Research Centre (reproduced
            from Chomcharn, 1985).
                                               46

-------
Appendix B. Detailed Sampling Procedure

The sampling configuration, from upstream to downstream, consisted of a metal sampling probe,
a TSP sampling cassette holding a 37-mm diameter quartz fiber filter (Whatman),  a low-flow
pump (SKC Aircheck Sampler, model  224-PCXR7), and an 80-liter  Tedlar bag (SKC West,
Fullerton,  CA).  These parts were connected with Vi"  Teflon tubing.  Figure B-l  shows  a
schematic of the sampling system used in the 15 kiln experiments.

80-liter Tedlar Bag Sampling: The duration and sampling flow rate of filling each large bag was
recorded. An 80-liter bag (the mixing bag) was created by filling an empty 80-liter Tedlar bag
with a time-weighted fraction of air from each filled sample bag. When only one 80-liter bag was
required for a given run (when the firing time was less than 5 hours), that bag was treated as the
mixing bag. The sampling parameters and mixing bag creation method  for each kiln type are
shown below:

  • MBH and BBH: The firing time of these kilns was from 38 to 61  hours.  A sampling flow
    rate of 100 ml/min was used, resulting in the filling of 4 or 5 large (80-liter) Tedlar bags per
    run. The duration of filling each bag was 12 hours. To create the integrated sample, one liter
    per two hours of filling was taken from each large bag and placed in the clean, empty 80-
    liter mixing bag (please see Table 1  for a sampling summary).

  • EM: The firing time of the EM kiln was from 20 to 24 hours.  At a sampling  rate of 200
    ml/min, 4 or 5 large bags were needed for each run.  Each bag was filled for 5  hours. Two
    liters per hour of filling was taken for the mixing bag.

   •  SD and RFDVI: The firing time of the SD and RFDVI kilns was from 4 to 7 hours. Only one
     or two 80-liter bags were needed for each run.

Use of small (1-liter) Tedlar bags: These were used to capture samples from filled large Tedlar
bags. Two small bags (replicates) were filled per mixing bag. These 1-liter bags were taken to the
laboratory at KMUTT for GC analysis.  For runs 1 and 3 for each kiln, 1-liter bag samples were
only taken from the mixing bags (1 bag per kiln run).  For run 2 of each kiln, samples were taken
from the mixing bag and all of the 80-liter Tedlar bags that were filled throughout the run.

Replicate Sample: A parallel sample (replicate)  was taken once for each  kiln (during run 2).
When conducting the  parallel sampling, a "Y" shaped probe was  used  to allow for a pair of
identical sampling trains running simultaneously. The purpose of the parallel  sampling was to
evaluate the method precision. See Table E-l.

Ambient  Sample: A probe, filter cassette, pump, and 80-liter Tedlar bag was also  used for
ambient air sampling during one run of  each kiln (run 1). This sampling was conducted at least
head-height off the ground, upwind of the kiln area, and was completed before the firing of the
kiln. A flow rate of 3 liter/min was used for 25 minutes to fill the large bag. With cassette alone,
however, the sampling continued for another 2 hours to collect a large enough amount of TSP.
                                          47

-------
                      duct (connected to hood)
                               I
kiln emissions
from hood
                                    7
                     0.25" i.d. metal tubing
                0.25" i.d. Teflon tubing	3

                               TSP filter
                               and cassette
 hood blower
X
                                                low-flow
                                              sampling pump
vented to atm.
                     SOL
                   Tedlar bag
 Figure B-l. A schematic of the sampling system used in the 15 kiln experiments.

-------
Canister Protocol: From the mixing bag of each of the 15 runs, a pre-cleaned and evacuated 850
ml stainless-steel canister was filled to 2 atm using a battery-operated pump. In addition, during
run number 2 for each kiln, one other canister was filled with an ascarite trap placed upstream of
the canister. Use of the ascarite trap allows for more accurate N2O analysis.

TSP Sampling  Protocol:  Whatman quartz fiber filters (37 mm diameter) were employed  for
collection of TSP from the charcoal kiln emissions. Before sampling, the filters were baked at
700°C in a furnace for at least 2 hours and then placed in a desiccator for at least 24 hours before
weighing. The filters were weighed in a 5-place balance immediately after being taken out of the
desiccator. Each weighed  filter was placed  in a labeled cassette, sealed with plastic caps, and
transported to the site (CRC).

1-liter Tedlar Bag Grab Samples:  A set of grab (one-time, non-integrated) hood samples were
collected in 1-liter Tedlar bags for 9 of the 15 total kiln experiments. A Vi" Teflon tube attached
to a wooden pole was used as the sampling probe. The inlet of the Teflon tube was placed about
0.5 m directly above the  top  of the kiln.  This area above the kiln was well-mixed  and full of
smoke due to the large hood which contained and directed the emissions.  (Without the hood,  the
smoke from all the kiln outlets would not be well-mixed and at relatively high concentration.)

The 1-liter grab samples were analyzed at KMUTT, as were the mixing bag, individual large bag,
and ambient samples. The purpose of these  grab  samples  was to investigate and develop a
sampling  method much simpler  than  the more complete (and cumbersome) integrated  hood
measurements. The similarity of the grab  sample analytical results  to those from the integrated
measurements was evaluated (see results section).  Such extended grab sampling may be useful
for future field work (kiln sampling),  obviating the difficult tasks of building  and  handling a
large hood and duct system and making long duration measurements.

Deposited Solids Emissions Measurements: In order to quantify the  amount of material from  the
emissions that deposited onto the inner wall surfaces of the hood and kiln, two patches made of
galvanized zinc (3" x 3") were placed on each of these surfaces. Weight and surface area of each
patch were measured before kiln firing. After kiln firing was completed, each patch was removed
and weighed at KMUTT.  The total mass  of material deposited onto  the hood surface and kiln
walls was calculated based on the net weight increase of the patch and the ratio of the total hood
or kiln surface area to the patch area.

Condensable Liquid (Condensables) Emissions Measurements: A metal U-tube was built into  the
emissions sampling ductwork system downstream  of the  large hood, prior to the gaseous
sampling system to collect condensable species. This proved to be an unworkable method  for
estimation of the total  amount of condensables ultimately emitted from the kilns. Therefore, two
follow-up kiln experiments (one BBH  and one EM) were performed in order  to better quantify
the condensables emitted.  In these  experiments, the sampling probe was placed directly over a
kiln chimney and led to a series of two flasks immersed in water baths in order to more directly
trap the condensable liquids.
                                           49

-------
       A summary of the filter, bag, canister, and grab samples taken for each kiln run, along
with firing times and sampling probe flow rate, is shown in Tablel.
                                          50

-------
                           Appendix C: Analytical Methods

Solid Samples: Calorific value and carbon (C) content were determined for two samples of each
of the  wood species  (Eucalyptus camaldulensis and  Leucaena  leucocephald).  Weight and
moisture content were measured for each of the fifteen batches of wood used. Wood moisture
content was determined by weighing several cross-sectional samples before and after oven drying
(several hours at 110 °C).

      After  kiln firing was complete, unused firing wood, charcoal,  brands, and ash were
removed from the kiln and weighed. Samples of the charcoal, brands, and ash were removed and
later analyzed  for C content. Calorific value of the charcoal and  wood was also determined.
Carbon content of the  TSP was analyzed using an automated thermal-optical method at Sunset
Laboratory, Forest Grove, Oregon, USA (Johnson et al., 1981).

Gaseous  Samples: Samples collected in 1-liter Tedlar bags were transported from the CRC to the
laboratory at KMUTT for gas chromatography (GC) analysis of CO2, CO, CH4, and TNMHC. A
Chrompack CP9001 GC  along with a methanizer system interfaced to a  Chrompack Flame
lonization Detector (FID) was employed for analysis of CO, CO2, and CFLj.  In this system,  a
Carbosphere (Alltech Co., USA) packed column was used to separate these three compounds.
The separated CO and CO2 were converted by the methanizer to CH4 which was then detected by
the FID. Total hydrocarbons (THC) were measured by the Chrompack GC/FID system equipped
with a blank stainless-steel column (the air peak was subtracted). TNMHC was then determined
indirectly by subtracting the CH4 content from the THC peak. The GC was calibrated before each
use via a calibration curve (at least 5  points) generated from locally made  CO2, CO, and CH4
standards. The GC calibration was checked with a standard canister gas mixture of CO2, CO, and
CH4 prepared by Scott Specialty Gases, Inc., USA

     The filled stainless-steel canister samples were shipped to OGIST  and analyzed for CO2,
CO,  CH4, TNMHC, and N2O. OGIST used a GC/FID/methanizer method to analyze CO2, CO,
and CH4. The canisters were analyzed for N2O using a  GC interfaced with  an electron capture
detector (Rasmussen and Khalil, 1980, 1981; Rasmussen et al., 1982).

     Information about the  stability of gases in the bags over time is presented in Table E-2.
Analyses of duplicates (two small bags taken from one large bag) are presented in Table E-3.

Filters: After sampling, the filter cassette  was re-capped and transported to  the laboratory at
KMUTT. Before weighing, the caps were removed, and the filter and cassette were placed in a
desiccator for at least 24 hours. The  filter was weighed immediately after removal from the
desiccator. Weight, date,  code, and balance number were recorded  in a sampling sheet. The
weighed  TSP filter was then place in a labeled petri dish which was sealed in a plastic bag and
stored in a refrigerator until being sent out  for carbon analysis. The cassette was cleaned by
wiping and could be reused.  Table E-4 shows the weighings of laboratory and field blanks.
                                          51

-------
    Appendix D: Follow-up Experiments to Determine Condensable Liquid Emissions

       The results of these experiments are summarized in Table D-l.

Kilns Tested: One run each of BBH and EM kilns was performed using eucalyptus wood.

Parameters Measured:

•  Airborne Measurements: CO2, CO, CH4, TNMHC

•  Solid/liquid Measurements:
    - Condensables (liquid): mass, carbon content
    - Wood: mass
    - Charcoal: mass
    - Brands and ash: mass

Summary of Sampling Procedure:

•  Dual sampling line  system: Two  sampling lines were utilized.  The first line,  called the
   condensables sampling line, consisted of 1A inch i.d. metal tubing, followed by two 2-liter
   Erlenmeyer flasks(capped with two-holed rubber stoppers), followed by a large protective
   quartz fiber filter, followed by a SKC Aircheck Sampler pump venting to the atmosphere. The
   sampling end of the metal tube was placed directly above  one of the chimneys of the kiln.
   This sampling line pump ran continuously throughout the kiln experiments at 1.5 1/min. The
   large, protective filter was checked and replaced when visually full to prevent blockage of the
   line. The two Erlenmeyer flasks were immersed in water baths in order to keep the flasks cool
   to promote  condensation. The bath water was changed periodically with fresh, cool stream
   water. The second part of the sampling line, called the gas sampling line, was connected via a
   "Y" connection to the condensables sampling line after the second Erlenmeyer flask and just
   before the large, protective filter.  This sampling line  consisted of 1A inch  metal tubing,
   followed by a 37 mm  Whatman quartz fiber sampling filter, followed  by a SKC Aircheck
   Sampler pump,  followed by  a  80-liter Tedlar bag. The gas sampling line pump was
   programmed to collect emissions for 2 minutes out of every hour of kiln firing time. The gas
   sampling line pump flow rate was the same as that used in the condensables sampling line (1.5
   1/min). The TSP sampling filter was checked periodically for particulate buildup to prevent
   line blockage. The TSP sampling filter was replaced if a large amount of particulate buildup
   was seen on the filter. Figure D-l shows a schematic of the sampling system used in the
   follow-up experiments.

•  Handling of the  condensables (liquid): When the kiln  firing was completed, the condensable
   liquids were weighed in the field (at the CRC). The condensables were transferred to a glass
   jar,  sealed, and brought back to KMUTT for subsequent carbon content analysis as were the
   solid samples of the fuel, charcoal, and brands.
                                          52

-------
•  Collection of kiln emission gases in Tedlar bags: At the end of the kiln run, three 5-liter
   metal-reinforced Tedlar bags (replicates) were filled from the large 80-liter Tedlar bags. These
   small bags were filled to a maximum of 3 liters to help minimize bag leakage. In addition, one
   ambient air sample was collected in a 5-liter metal-reinforced Tedlar bag for each of the two
   kiln experiments. The ambient samples were taken before kiln firing began, at a height of at
   least 1.5 meters above the ground, and away from human  activity. These 5-liter bags were
   filled to a maximum of 3 liters. The bags were wrapped individually and placed in a hard case
   for shipment. GC analyses of the 5-liter bags were performed at EOHSI.
                                           53

-------
Table D-l.  Data for follow-up experiments (determination of condensable liquid emissions)

Total wet wood mass, kg
% wood moisture (dry basis)
Dry wood mass, kg
Charcoal mass, kg
Brands mass, kg
Ash mass, kg
C fraction of wood *
C fraction of charcoal *
C fraction of brands *
C fraction of ash *
Wood C, kg
Charcoal C, kg
Brands C, kg
Ash C, kg
C in gases and condensables, kg (by difference)
Charcoal C yield
Charcoal mass yield (dry basis)
Firing time, hr
Condensables sampling flow rate, L/min
Mass of condensables collected, g
Condensables emission concentration (g/m3)
Carbon fraction of condensables
Condensables emission concentration (gC/m3)
CO2 (ppm)
C02 (gC/m3)
CO (ppm)
CH4 (ppm)
TNMHC (ppm)
CO/CO2
CH4/CO2
TNMHC/CO2
condensables/CO2
K (sum of above 4 ratios)
CO2 (C) emitted (kg)
CO (C) emitted (kg)
CH4 (C) emitted (kg)
TNMHC (C) emitted (kg)
Condensables (C) emitted (kg)
BBH
786
20.2
651
230
5.0
1.5
0.443
0.789
0.635
0.089
288
182
3.2
0.10
104
0.63
0.35
39
1.5
13.8
3.93
0.68
2.68
61300
30.1
16280
5122
9633
0.266
0.0836
0.157
0.0889
0.595
64.9
17.2
5.42
10.2
5.77
EM
210
19.4
175
53
5.0
4.0
0.443
0.748
0.523
0.048
77.6
39.6
2.6
0.20
35.1
0.51
0.30
19
1.5
2.59
1.51
0.68
1.03
16120
7.9
3503
696
954
0.217
0.0432
0.0592
0.130
0.449
24.2
5.27
1.05
1.43
3.14
                                                (continued)
                                          54

-------
Table D-1.  (continued)

Percent of wood carbon in products:
charcoal
brands
condensables
ash
CO2
CO
CH4
TNMHC
Total
BBH

62.94
1.10
2.00
0.0463
22.51
5.98
1.88
3.54
100.0
EM

51.10
3.37
4.05
0.247
31.25
6.79
1.35
1.85
100.0
  * Average of values from the three original kiln experiments (carbon analyses not performed in
   these follow-up experiments)
                                          55

-------
0.25'y. d. stainless steel tubing
                                                                                          protective
                                                                                          filter         pump
                                                                                                     (continuous)
                                                                                                sampling pump
                                                                                                (intermittent)
                                                                                         SOL
                                                                                      Tedlar bag
                                                                                                                   vented
                                                                                                                   p
                                                                                                                   to arm
Figure D-l. A schematic of the sampling system used in the follow-up experiments.

-------
                         Appendix E: Quality Assurance Plan

Pilot Study: Pilot work was conducted to: a) check the temperature at the sampling probe, b)
check for leakage around lip of hood, c) check TSP build-up on filters, d) test the patch system
and e) check for flow rate changes over the entire run period. Short-firing-time kilns (SD  and
RHM) were used in the trial runs.

For Each Kiln Run:

•  Two  sampling pumps were prepared before sampling by  a) calibrating the pumps using a
   bubble flow meter, b) checking internal protection filters  (they were changed if broken or
   discolored), and c) cleaning tubing, if discolored, with water and alcohol and drying.

•  The Tedlar bags were cleaned by flushing with clean air at least three times before each use.
   The bags were checked to assure that there was no leakage.

•  Several parameters were checked throughout the runs, including sampling flow rate, hood
   blower flow rate, and hood leakage.  These parameters were checked near the beginning  and
   end of each kiln run, as well  as every 6 hours of firing for MBH and BBH, every 4 hours for
   EM, and every 2 hours for SD and RHM.

Labeling: Filters, large bags, small bags, etc. were  carefully labeled.  Sampling  sheets were
carefully filled out and maintained. The following labeling rules were used:

•  The basic identifier on sampling sheets was "kiln type-run#", i.e., MBH-1, MBH-2, MBH-3,
   BBH-1, BBH-2,  etc.
•  Large Tedlar bags: LB-01, LB-02, etc. (the first large bag from run#l  for each kiln type was
   the ambient sample)
•  Pumps: P-01, P-02, P-03
•  Filters (code on cassettes and petri dishes): F-001, F-002, etc. (the first filter from run#l for
   each kiln was the ambient sample)
•  Small Tedlar bags:
    - MBH: MBH-1M (mixed bag from MBH run #1), MBH-2M (mixed bag from MBH run
         #2), MBH-2A (first integrated bag from MBH run #2), MBH-2B (second integrated
         bag from MBH run #2), MBH-2B-R (replicate of the second integrated bag sample
         from MBH run #2), MBH-2C, MBH-2D, MBH-2E (labels A through E represent each
         of the integrated 80-L Tedlar bag samples), MBH-3M (mixed bag from MBH run #3)
     - BBH: BBH-1M, BBH-2M..., the same as for MBH.
     - EM: EM-1M, EM-2M, EM-2A, EM-2B, EM-2B-R, EM-2C, EM-3M
     - SD: SD-1, SD-2, SD-2R, SD-3
     - RHM: RHM-1, RHM-2, RHM-2R, RHM-3
     - Ambient Samples: A-MBH-1, A-BBH-1, A-EM-1, A-SD-1, A-RHM-1.
     - Grab Samples: BBH-1-G1, BBH-1-G2, BBH-1-G3, BBH-2-G1, etc.

•  Canisters: C-01, C-02, etc.; keyed to the kiln and run number in the sampling sheet
                                          57

-------
Method and Analytical Precision:

•  One parallel, replicate sample (80-liter Tedlar bag and filter) was taken for one run of each
   kiln to assess the precision of the measurement method. These replicate analyses are shown in
   Table E-l.  Three of the 1-liter mixed bag samples were run through the GC a second time
   after a one-week storage period in order to see the effect on the mixture of major gases. These
   data are shown in Table E-2.  Also, two duplicate small Tedlar bags were filled from the large
   mixing bag for five of the kiln runs. This allowed for duplicate GC analyses at KMUTT. See
   Table E-3  for these data.

•  In the local laboratory at KMUTT, two or more injections were  made into the GC for each
   sample until a relative standard deviation (RSD) of less than 10% was obtained for all peak
   areas of interest. All GCs were calibrated before each use with at least five calibration points.
   The calibration curves for all measured compounds had a linear regression coefficient (R2)
   greater than 0.99. The concentrations of unknown samples fell into the calibration range for
   all species  other than CO.  In some instances, due to very  high  sample concentrations, the
   maximum CO calibration point was exceeded.

•  After ten TSP filters were weighed, the first was reweighed. If the difference was more than
   0.1%, all ten filters were reweighed. In addition, two laboratory blank filters  and two field
   blank filters were used for each kiln run  in order to check for any contamination during the
   transporting and handling processes and/or any changes in the analytical balance over time.
   The average field blank filter mass gained was subtracted from each  of the  sample filter
   masses. See Table E-4 for the blank filter weighings.

Cross Laboratory Checks: Results obtained from the local GC  analyses (at KMUTT) were
compared  with results of canister samples analyzed by OGIST (see Results section). Although
the KMUTT  laboratory and the OGIST laboratory used their own locally  made  calibration
standards  on daily operation basis, each institution was provided with a calibration  mixture of
CO2, CO,  CH4 from the same source (Scott Specialty Gases, Inc., USA).  Thus, the locally made
standards  were calibrated with a common  standard and calculations could be corrected based
upon the common standard if there was any difference between a local standard and the common
standard.
Back-Calculation of Volumetric Emission Rates:

       The emission factor determination method used in this study was not designed to require
knowledge or measurement of the volumetric emission rate of the kilns. The hood and blower
system used in these experiments was designed  to capture  as much of the kiln emissions as
possible, but can tolerate  some leakage since the primary measures are the CO2  ratios. It is
important, however, that the flow rate is constant throughout.  The flow rate of the single-speed
hood blower, which was measured several times throughout each kiln run, was found to be quite
constant. The range of average measured flow rates of the hood blower (for three runs of each
kiln type) was 0.022-0.027 m /sec.
                                          58

-------
Table E-l. Replicates. Two bags filled simultaneously through a y-probe.

BBH-2B-I-1
BBH-2B-r2
MBH-2B-rl
MBH-2B-r2
SD-2M-rl
SD-2M-r2
EM-2A-rl
EM-2A-r2
RHM-2M-rl
RHM-2M-rl
C02
(ppm)
21485
21038
4100
4115
12801
11756
6747
6587
13497
13898
CO
(ppm)
6897
6839
766
755
5150
4141
2039
2010
1056
1106
CH4
2300
2314
136
131
1410
1124
438
431
224
248
TNMHC
(ppm)
3307
3153
237
228
2568
2899
459
458
377
394
                                          59

-------
Table E-2. Storage test data.
Sample/
date of analysis
BBH-2M-bl
5/13/96
5/19/96
MBH-lM-bl
5/13/96
5/19/96
MBH-2M
5/13/96
5/19/96
CO2
(ppm)

16267
12157

6541
4743

5411
4172
CO
(ppm)

4818
4369

1436
1281

981
911
CH4
(ppm)

1403
1277

305
274

223
216
TNMHC
(ppm)

1365
1438

243
285

249
232
All Species
(ppm)

23853
19241

8525
6583

6864
5531
CO/CO2

0.296
0.359

0.220
0.270

0.181
0.218
CHJCO,

0.086
0.105

0.047
0.058

0.041
0.052
TNMHC/CO2

0.084
0.118

0.037
0.060

0.046
0.056
(CO+CH4+TNMHC)/
C02

0.466
0.583

0.303
0.388

0.269
0.326

-------
Table E-3.  Analyses of duplicate small bags filled from the same large mixed bag
Name
BBH-lM-bl
BBH-lM-b2
BBH-2M-bl
BBH-2M-b2
BBH-3M-bl
BBH-3M-b2
MBH-lM-bl
MBH-lM-b2
MBH-3M-bl
MBH-3M-b2
C02
(ppm)
23352
26587
16267
15459
26742
28801
6541
6333
6986
7540
CO
(ppm)
6730
6602
4518
4594
6241
5962
1436
1407
1151
1143
CH4
(ppm)
2338
2367
1403
1454
2187
2236
305
295
329
327
TNMHC
(ppm)
2382
2187
1365
1357
3027
2934
243
289
423
403
Table E-4. Field and laboratory blank filter data. The average of the field blank values was
subtracted from the real TSP samples. The laboratory blank data were not used for correction.
Field (CRC) Blanks
BBH-1
BBH-2
BBH-2
BBH-3
BBH-3
MBH-1
MBH-1
MBH-2
MBH-2
MBH-2

Laboratory Blanks
BBH-1
BBH-2
BBH-2
BBH-3
MBH-1
MBH-1
MBH-2
MBH-2

Filter #
F-047
F-012
F-013
F-052
F-053
F-016
F-017
F-031
F-034
F-035


F-041
F-001
F-002
F-054
F-014
F-015
F-032
F-033

Mass Change (mg)
0.010
0.045
0.055
0.175
-0.210
0.120
0.055
0.040
0.025
0.020
mean = 0.034

-0.015
0.065
0.060
-0.025
-0.020
0.150
-0.035
0.060
mean = 0.030
                                           61

-------