, ~A I/!!! The Engineering Society
`~ For Advancing Mobility
`~Land Sea Air and Space®
`INTERNATIONAL,
`
`400 COMMONWEALTH DRIVE, WARRENDALE, PA 15096-0001 U.S.A.
`
`SAE Technical
`Paper Series
`
`902119 ~
`
`Performance of Activated Carbon in
`Evaporative Loss Control Systems
`H.R. Johnson and R.S. Williams
`Westvaco Corp.
`Covington, VA
`
`S. A. Eo
`LIBRARY
`
`/c:;6>~ d
`
`Reprinted from SP-839 -
`New Directions and Developments in
`Automotive Emission Control
`
`International Fuels and Lubricants
`Meeting and Exposition
`Tulsa, Oklahoma
`October 22-25, 1990
`
`BASF-1026
`U.S. Patent No. RE38,844
`
`

`

`GGiobal
`
`Mobility
`JRDatabOse~
`
`The papers included '" th1s volume
`are abstracted and mdexed in the
`SA£ Global Mobtl1ty Database.
`
`No part of this publication may be reproduced in any form, in
`an electronic retrieval system or otherwise, without the prior
`written permission of the publisher.
`
`ISSN 0148-7191
`Copyright 1990 Society of Automotive Engineers, Inc.
`
`Positions and opinions advanced in this paper are those oft he
`author(s) and not necessarily those of SAE. The author is
`solely responsible for the content of the paper. A process is
`available by which discussions will be printed with the paper
`if it is published in SAE Transactions. For permission to
`publish this paper in full or in part, contact the SAE Publica(cid:173)
`tions Division.
`
`Persons wishing to submit papers to be considered for pres(cid:173)
`entation or publication through SAE should send the manu(cid:173)
`script or a 300 word abstract of a proposed manuscript to:
`Secretary, Engineering Activity Board, SAE.
`
`Printed in U.S.A.
`
`

`

`902119
`
`Performance of Activated Carbon in
`Evaporative Loss Control Systems
`H.R. Johnson and R.S. Williams
`Westvaco Corp.
`Covington, VA
`
`ABSTRACT
`
`BACKGROUND
`
`Although activated carbons have been used
`successfully for approximately 20 years
`for
`control of hydrocarbon evaporative emissions
`from motor vehicles, the correlation of
`funda(cid:173)
`mental activated carbon properties, molecular
`scale pore size/volume characteristics,
`to
`carbon performance has not generally
`been
`available.
`Improvements
`in wood-based carbon
`have focused on the pore size/volume character(cid:173)
`istics resulting in a doubling of performance,
`as measured by butane working capacity, over
`the past few years.
`In addition to
`relating
`the butane working capacity for different types
`of carbon to pore characteristics,
`laboratory
`studies using new carbons have demonstrated
`that the capacity for adsorption of vapors from
`gasoline is also related
`to this
`fundamental
`property.
`Furthermore, different
`types
`of
`in-use carbons from vehicle canisters have been
`evaluated to confirm that
`the key
`to carbon
`performance is pore characteristics and also,
`that laboratory evaluations of carbons, using
`proper conditions, can realistically simulate
`the
`performance of actual
`in-use carbons.
`
`Automotive evaporative emissions control
`systems were instituted in 1970 for cars sold in
`California as the result of regulations promul(cid:173)
`gated by
`the California Air Resources Board
`(GARB). Beginning in the 1971 model year, all
`cars sold in the United States were required
`to
`use evaporative control systems. These national
`evaporative emission control
`regulations were
`mandated by
`the U.S. Environmental Protection
`Agency and have become increasingly stringent as
`shown in Table 1.
`the
`require
`regulations
`Current U.S.
`control of diurnal and hot soak
`emissions.
`Although these
`and
`other
`regulations have
`achieved substantial
`reductions
`in vehicle
`emission levels, previously unexamined
`factors
`have uncoverered higher levels of evaporative
`emissions
`than were once
`thought
`to occur.
`These
`factors
`include high fuel volatility,
`higher ambient
`temperatures, multiple
`day
`
`Table 1. Summary of U. S. Evaporative Emmission
`Control Standards
`
`Hodel
`Year
`
`Test
`Method
`
`Limit
`(g/test)
`
`1970 (Calif.)
`
`Carbon Trap
`
`1971 (U.S.)
`
`Carbon Trap
`
`1972 (U.S.)
`
`Carbon Trap
`
`1978 (U.S.)
`
`1980 (Calif.)
`
`1981 (U.S.)
`
`SHED
`
`SHED
`
`SHED
`
`6
`
`6
`
`2
`
`6
`
`2
`
`2
`
`101
`
`

`

`vapor
`fuel permeation and
`parking events.
`migration.
`In January 1990, EPA published a
`proposed rule for evaporative emission control
`that would not only tighten existing standards,
`but also would
`include the control of hydro(cid:173)
`carbon
`losses while
`the vehicles are
`run(cid:173)
`ning.(l)* The EPA proposal relies on
`test
`procedure changes that
`influence purge volume
`and a design review process to control
`running
`losses. General Motors Corporation has pro(cid:173)
`posed a procedure
`that
`includes
`real-time
`diurnal testing as well as the direct measure(cid:173)
`ment of
`running losses. (2) An'other new
`test
`procedure
`from GARB
`also
`includes
`direct
`measurement of
`running
`losses and
`real-time
`diurnal testing.(J) Regardless of
`the final
`form of a new national
`test procedure,
`it
`appears that
`future evaporative
`loss control
`systems will have to perform at higher efficien(cid:173)
`cies.
`Evaporative loss control systems primarily
`consist of an activated carbon canister for the
`adsorption and desorption of
`hydrocarbons,
`control hardware and
`the associated piping.
`Although evaporative emissions can be influenc(cid:173)
`ed by control hardware, purge systems and
`materials of construction,
`this paper will
`focus on the design and
`testing of activated
`carbon products for the automotive application.
`Activated carbon used in
`the automotive
`canisters
`in
`the early 70's was
`typically
`coal-based granular carbon with a
`butane
`working capacity
`(BWC)(4)
`in
`the
`range
`* Number in parentheses designates reference
`at end of paper.
`
`the mid 70's. wood-based
`In
`of 5 g/lOOml.
`introduced in
`the United
`granular carbon was
`States
`for
`the
`automotive application
`and
`rapidly gained a large market share as a
`result
`of product advantages. The initial wood carbons
`also had a typical BWC level of 5 g/100ml.
`The butane working capacity of wood carbon
`has more than doubled to 11 BWC since 1975
`in
`response to more stringent regulations
`(Figure
`1).
`Improvements have also provided for more
`flexibility in the canister geometry and size.
`Further wood carbon performance improvements are
`now in the plant
`trial stage with an objective
`of 13 BWC er higher.
`To date, wood carbon has
`been used
`in over 150 million vehicles
`for
`control of evaporative emissions.
`improvements
`The strategy for wood carbon
`has focused on defining the opL1mum pore size
`characteristics
`that are
`important for
`the
`automotive application,
`and developing
`labora(cid:173)
`tory test methods
`to accurately simulate
`the
`characteristics and
`performance
`of
`in-use
`carbons.
`In activated carbon terminology, pore size
`characteristics refers
`to
`investigation
`of
`carbon properties on a molecular scale. Optimum
`pore size characteristics have been defined
`for
`gasoline fuels and are currently being
`investi(cid:173)
`gated for alcohol blends. Evaluation of
`in-use
`carbons have confirmed that pore size distribu(cid:173)
`tion is
`the key parameter for
`the automotive
`application and that laboratory test procedures
`are available
`for realistic simulation of a
`carbon's performance in an actual canister.
`
`12
`
`10
`
`E
`
`0
`0
`........
`C'l
`
`~ 8
`CJ < j
`(!) z
`~
`n:::
`0
`3:
`L&.l z
`g
`
`6
`
`4
`
`2
`
`0
`
`m
`
`n:::
`:5
`~
`z
`~
`
`(!)
`0
`0
`0
`3:
`1976
`
`hi
`~ a..
`
`0
`0
`0
`3:
`
`n:::
`:5
`~ z
`~
`(!)
`0
`0
`0
`3:
`
`1978
`
`1986
`
`1981
`MODEL YEAR
`Figure 1. Performance Improvement History For Wood Carbons
`
`102
`
`

`

`EXPERIMENTAL
`
`of
`Evaluation
`EQUIPMENT DESCRIPTION
`fuel
`carbon performance
`for. adsorption of
`test
`vapors was conducted using
`the canister
`fully
`system shown in Figure 2.
`The system.
`instrumented and computer controlled, has
`the
`capability of automatically
`testing
`carbon
`canisters for adsorption/desorption performance
`with a variety of fuels.
`The fresh and spent fuel tanks are located
`remotely.
`Both
`tanks have a capacity
`of
`approximately 40 liters. Fuel is delivered via
`nitrogen pressure from
`the fresh fuel storage
`tank to
`the fuel volume charge vessel.
`The
`charge vessel is equipped with a level sensor
`so that a precise volume of fresh
`fuel
`is
`delivered
`for each adsorption cycle.
`Fuel
`flows from
`the charge vessel
`to
`the vapor
`generator on demand and once the temperature in
`the vapor generator is below a set point. Heat
`transfer fluid is then pumped into the water
`
`bath to heat the fuel to a desired temperature.
`After
`the desired
`fuel
`temperature
`is
`reached, clean, organic-free, dry air is bubbled
`through the fuel
`in the vapor generator.
`The
`air flow
`rate
`is controlled by a mass
`flow
`instrument with the actual air rate predefined
`to give the desired vapor generation rate and
`vapor concentration. The air/fuel vapor passes
`through coalescing filters to remove any
`liquid
`droplets prior to the carbon canister.
`internal
`During
`the
`adsorption cycle,
`canister temperature and the canister weight are
`continuously measured and recorded. The hydro(cid:173)
`carbon content of the air exiting the canister
`is also continuously monitored by a Beckman
`(THA).
`Model 400A Total Hydrocarbon Analyzer
`When breakthrough from the canister reaches
`the
`equivalent of 0.57. volume butane, the adsorption
`cycle is stopped by
`terminating airflow to
`the
`vapor generator and closing
`the appropriate
`valves.
`
`N2
`
`PRESSURE
`
`AIR
`
`-
`
`t
`
`WATER
`
`BATH
`
`REMOTE
`LOCATION
`
`....
`
`LABORATORY
`
`PURGE
`
`AIR
`
`PURGE
`
`TO
`
`EXHAUST
`
`EXHAUST
`
`CARBON
`CANISTER
`
`TOTAL
`HYDRO-
`CARBON
`ANALVZER
`
`INPUTS OUTPUTS
`
`• t
`
`PROCESS CONTROL COMPUTER
`
`Figure 2. Westvaco Automotive Canister Cycle Test Equipment
`
`103
`
`

`

`the
`After a set time delay, purging of
`canister is
`initiated. The purge air,
`regu(cid:173)
`lated by a mass flow controller, flows
`through
`the canister counter-current to the fuel vapor
`flow path. The purged hydrocarbon vapors are
`directed to a vent line that enters a
`labor(cid:173)
`atory hood.
`Internal canister temperature and
`canister weight are continuously monitored
`during the purge cycle.
`Two different prototype canisters have
`been used. Both canisters had carbon volumes
`of 375 ml. Prototype canister "A" had a carbon
`bed
`length
`to diameter ratio equal
`to
`1
`(L/D=l). Prototype canister "B" had a
`carbon
`bed L/D=2. The test system is also designed to
`accommodate production canisters.
`(PLC)
`A programmable
`logic controller
`monitors all detectors and regulates the mass
`flow controllers and control values. Data
`collected by the PLC is sent to a computer
`for
`archiving, report generation and analysis. The
`computer displays
`include:
`canister weight,
`canister temperature, and THA output
`trending,
`process overview
`(Flow
`rates, valve states,
`vessel fill
`levels) and
`remote
`tank
`fill
`levels.
`Commands
`to
`the PLC are
`entered
`through the computer keyboard.
`PORE SIZE VOLUME MEASUREMENTS - The pore
`size/volume characteristics were determined for
`new carbons using a Micromeritics ASAP 2000
`porosimeter.
`In the first step of the proce(cid:173)
`dure the carbon
`sample is degassed under a
`vacuum while maintaining the sample at 125°C
`for 2-3 hours. After "cleaning" the pores and
`surface of the carbon by degassing, the pore
`characteristics are
`then defined based
`on
`nitrogen adsorption data
`taken at the boiling
`point of
`liquid nitrogen
`(77.4°K).
`Pores
`inside
`the carbon particles
`that are 1000
`angstroms in size and smaller are
`initially
`filled with condensed
`liquid nitrogen at a
`pressure near one atmosphere. The pressure
`is
`then
`reduced
`in
`discrete
`increments with
`nitrogen desorbing
`from a given pore size
`range. The volume of nitrogen desorbed
`is
`measured for each discrete step
`in pressure
`reduction, and the pore size range from which
`desorption takes place is related to the actual
`pressure level by the Kelvin equation. Uti(cid:173)
`lizir:g the Kelvin equation and nitrogen desorp(cid:173)
`tion volume measurements, the pore size/volume
`characteristics are calculated
`for a given
`carbon.
`~OISTURE ADSORPTION MEASUREMENTS - Adsorp(cid:173)
`tion of moisture for new and in-use carbons was
`performed using one of two methods. Under
`the
`first procedure, dried carbon samples of known
`weight were placed into sealed desiccator
`jars
`containing various sulfuric acid concentrations
`to give selected constant
`relative humidity
`atmospheres.(S)
`The
`carbon
`samples were
`contained on trays made of 60 mesh stainless
`steel wire to ensure circulation. The desicca(cid:173)
`tors were maintained at room temperature until
`
`an equilibrium weight was obtained, usually
`after three to five days.
`The
`second procedure utilized a McBain
`balance.
`In this procedure a carbon sample
`is
`suspended from a quartz spring inside a
`temper(cid:173)
`ature controlled vessel. The carbon is first
`outgased under vacuum, and
`then water vapor
`is
`admitted in measured doses. The weight adsorbed
`is measured by the extension of the spring and
`correlated with the vapor pressure measured with
`a capacitance manometer.
`BUTANE WORKING CAPACITY PROCEDURE - Butane
`working capacity (BWC) in this paper refers to a
`Westvaco Test Procedure.(4) Butane
`is passed
`through a bed of activated carbon of known
`weight at 25°C until saturation occurs~ After
`noting the weight of butane adsorbed (the Butane
`Adsorption Capacity), the carbon bed is purged
`with dry, organic-free air at 25°C. The
`total
`purge is approximately 645 bed volumes.
`The
`amount of butane desorbed is noted and
`reported
`as the butane working capacity in grams butane/
`lOOml carbon.
`
`RESULTS AND DISCUSSION
`
`PORE VOLUME DISTRIBUTION - Activated carbon
`is a form of carbon with an irregular, highly
`porous structure
`that provides
`for efficient
`adsorption of organic compounds. A 1 mm
`carbon
`particle can contain over 1 billion pores. Most
`of these pores have diameters in the molecu!ro
`size range, 10-100 angstroms (1 angstrom = 10
`meters). There are also
`larger pores that are
`useful for molecular diffusion.
`The size and
`volume of the smaller pores can determine
`the
`suitability of an activated carbon for a partic(cid:173)
`ular application. For
`the automotive applica(cid:173)
`tion,
`the pore volume distribution can
`be
`correlated directly to the carbon's performance.
`For the purposes of this paper, micropores
`are defined as those pores with diameters
`less
`than about 20 angstroms, mesopores are
`those
`with a diameter of approximately 20-1000 ang(cid:173)
`stroms, and macropores are those with diameters
`greater than 1000 angstroms.
`the
`During adsorption,
`fuel vapors enter
`carbon canister and diffuse into the carbon pore
`matrix. Attracted primarily by physical
`forces
`known as Van der Waals Forces,
`the fuel vapor
`molecules are adsorbed
`into
`the pores.
`The
`strength with which these molecules are held
`in
`the pores
`is related
`to the ratio of
`their
`diameter to
`the pore diameter. The closer a
`pore diameter is
`to a molecular diameter,
`the
`stronger the physical force holding the mole(cid:173)
`cule. For pore diameters
`larger
`than about
`10-20 times the molecule's diameter, adsorption
`is inefficient.
`the fuel
`that hold
`forces
`The physical
`the carbon pores are rela(cid:173)
`vapor molecules in
`tively weak
`for all but
`the smallest pores.
`Therefore,
`the molecules can
`be
`desorbed
`effectively by using an ambient air purge. For
`
`104
`
`

`

`the micropores, this process
`the fuel held in
`is inefficient, and part of the fuel remains on
`the carbon.
`This hydrocarbon
`that
`remains
`after purging is referred to as the hydrocarbon
`heel. The hydrocarbon
`that is removed during
`purge is referred to as the working capacity.
`Twelve new carbon samples were selected to
`study the effect of pore volume distribution on
`carbon performance.
`The carbons
`selected
`included commercially available wood and coal
`carbons as well as several developmental wood
`carbons.
`Pore
`characteristics and
`butane
`working capacity were measured for each carbon
`according to procedures previously described.
`Gasoline breakthrough adsorption capacity was
`also determined for each carbon. The gasoline
`adsorption breakthrough capacity will be
`re(cid:173)
`ported in terms of grams hydrocarbon adsorbed
`per liter of carbon. Adsorption breakthrough
`capacity will be used rather
`than working
`capacity since
`the measurement of adsorption
`capacity more closely simulates the SHED certi(cid:173)
`fication procedure.
`In actuality, the break(cid:173)
`through capacity and working capacity
`are
`essentially equal after
`the carbon has been
`through several adsorption/desorption cycles.
`The following conditions (Table 2) were
`used with
`the canister test system for
`the
`determination of
`the gasoline
`breakthrough
`capacity:
`
`the results of the butane
`Table 3 gives
`and gasoline adsorption capacities
`(GAC),
`as
`well as the· relative pore volume contained
`in
`the selected pore size range,
`for the carbons
`tested. The pore volume was determined on
`a
`(cc/Liter of carbon) and
`then
`volume basis
`normalized such that the 9.9 BWC wood granular
`sample is assigned a value of 1.00
`for each
`range.
`The relationship between pore volume and
`GAC was
`examined by determining correlation
`coefficients
`(R Square)
`for
`the "best fit"
`straight line when GAC was plotted against pore
`volume
`in defined
`pore size
`ranges.
`The
`highest correlation coefficient, R Square
`0.91, was obtained when the gasoline adsorption
`capacity was correlated with the pore volume
`in
`the small mesopores (Figure 3). This desired
`pore size range for GAC
`can also be related
`to
`the molecular size of the molecules in gasoline
`vapors. Several studies
`(6,7,8) have examined
`the composition of gasoline vapors from evapora(cid:173)
`tive emissions. The primary constituents of the
`vapors are c4 , c5,
`and c6 alkanes and olefins.
`There are also smaller amounts of c6 and higher
`alkanes and aromatics.
`Van der Waals
`radii
`information shows
`that butane, pentane
`and
`hexane all have molecular diameters in the 5.0
`angstroms size
`range.
`Therefore,
`the above
`relationship indicates
`that
`the desired pore
`sizes for gasoline adsorption capacity
`are
`several times the approximate molecular diame(cid:173)
`ter.
`
`TABLE 2.
`
`ADSORPTION CYCLE
`
`Gasoline
`
`Fresh Gasoline per cycle
`
`Gasoline Temperature
`
`Airflow to Gasoline Vapor Generator
`
`Vapor Flow Rate to Canister
`
`Carbon Volume
`
`Breakthrough Concentration
`
`Purge Volume
`
`Purge Rate
`
`Number of Cycles
`
`PURGE CYCLE
`
`105
`
`Amoco Super Unleaded
`RVP
`11.0
`
`300 ml/cycle
`
`150 ml/min
`
`0.6-0.8 g/min
`
`375 ml
`
`Approx. 0.57. Volume as
`Butane
`
`200 Bed Volumes
`
`40 Bed Volumes/minute
`
`25
`
`

`

`R Square = .91
`200 Bed Volumes Purge
`All Carbons ke Wood Unless Otherwise Noted
`
`•
`
`•
`
`European Pellet a
`
`2
`
`1.5
`
`Ill
`UJ
`0::
`0
`D..
`0
`Ill
`UJ
`::::l:
`
`0.5
`
`Commerclol
`Carbons
`
`Developmental
`Carbons
`
`0
`
`20
`
`25
`
`40
`35
`30
`GASOLINE ADSORPTION CAPACITY, G/L
`
`45
`
`50
`
`Figure 3. Relationship Between Pore Volume and Gasoline
`Breakthrough Adsorption Capacity
`
`Table 3. Relationship Between Pore Volume And Carbon Performance
`
`Butane Working
`Capacity g/lOOml
`
`Normalized Pore Volume In Selected
`Size Range, (Volume Basis)
`Small
`Mesopores Micropores
`
`Gasoline Breakthrough
`Adsorption
`Capacity, g/L
`
`Hydrocarbon
`Heel, g/L
`
`9.9
`9.9
`10.3
`10.8
`11.8
`11.8
`12.4
`13.4
`14.3
`
`4.4
`9.3
`
`8.3
`
`1. 00
`1.18
`1.04
`1. 29
`1.40
`1.45
`1.38
`1.59
`1. 82
`
`0.27
`1.13
`
`0.82
`
`1. 00
`0.91
`1. 03
`1. 04
`1.05
`1. 03
`1.16
`1.37
`1.57
`
`1.77
`1. 65
`
`2.06
`
`39.4
`36.4
`37.6
`39.3
`41.4
`42.7
`44.1
`47 .• 4
`46.6
`
`23.2
`38.8
`
`31.0
`
`49.3
`49.6
`49.6
`58.4
`53.6
`57.6
`60.8
`68.5
`66.9
`
`80.8
`84.8
`
`74.4
`
`Carbon
`
`Wood Granular
`Wood Pellet
`Wood Granular
`Wood Granular
`Wood Granular
`Wood Pellet
`Wood Granular
`Wood Granular
`Wood Granular
`
`Coal Granular
`Coal Granular
`
`Pellet (European)
`
`Gasoline Breakthrough Adsorption Capacity is the average value for cycles 21-25.
`
`Hydrocarbon heel determined after 25 cycles.
`
`106
`
`

`

`Figure 4 also shows a strong relationship
`between pore volume in the small mesopores and
`the butane working capacity. ·This plot con(cid:173)
`tains points for several carbons not
`found in
`Table 3. These include a coconut, coal granu(cid:173)
`lar and several wood carbons. As with gasoline
`
`this
`working capacity,
`that the micropores are
`ciently purged;
`and,
`mesopores are
`too large
`tion.
`
`indicates
`relationship
`too small
`to be effi(cid:173)
`the medium-to-larger
`for effective adsorp-
`
`2
`
`R SCjuare = .96
`All Carbons Are Wood Unless Otherwise Noted
`
`w
`25 0.5
`
`D.. ! 0::
`
`0 2
`
`Commercial
`Carbons
`
`Developmental
`Carbons
`
`12
`10
`8
`6
`BUTANE WORKING CAPACITY, G/100 U..
`
`H
`
`16
`
`Figure 4. Relationship Between Pore Volume and Butane Working Capacity
`
`in
`Figure 5 is a plot of the pore volume
`the micropores and
`the hydrocarbon heel after
`Cycle 25.
`The
`relationship exhibited here
`again indicates
`that a
`large amount of pore
`volume in the micropores is inefficient for the
`automotive application. The molecules in these
`pores are held to
`the carbon surface with
`too
`
`much force to be effectively purged with ambient
`air. The point for the European pellet does not
`appear to follow this relationship as closely as
`the other carbons.
`The unusually
`large pore
`volume in
`the smaller micropores possessed by
`this carbon may have an effect on
`the above
`relationship.
`
`2.2
`
`2
`
`1.8
`
`Ul
`w
`t5
`&
`5
`5!
`w
`i=
`~ 1.6
`w
`::::!
`::J d 1.4
`>
`w
`25
`
`R Square = .88
`After 25 Cycles with 200 'i!V Purge
`All Carbons Are Wood Unless Otherwise Noted
`
`• European Pellet
`
`•Coal
`
`..
`
`1.2
`
`D.. !
`
`0::
`
`0.8 40
`
`50
`
`70
`60
`HYDROCARBON HEEl., G/L
`
`80
`
`90
`
`Figure 5. Relationship Between Pore Volume and Hydrocarbon Heel
`
`107
`
`

`

`The relationship between butane working
`capacity and gasoline adsorption capacity
`is
`examined in Figure 6.
`It is evident that there
`exists a
`strong relationship between BWC
`and
`GAC. This relationship would be expected since
`
`both BWC and GAC have excellent linear correla(cid:173)
`tions with
`the pore
`volume
`in
`the
`small
`mesopores. The butane working capacity
`test
`used by carbon manufacturers for quality control
`is therefore a good
`indicator of how well a
`carbon will perform with gasoline vapors.
`
`50
`
`R Square = .95
`·
`All Carbons Are Wood Unless Otherwise Noted
`
`~ .. r
`
`~ -45
`u §
`z
`0
`ti:
`~
`Ul
`~ 30
`w z
`d
`~ 25
`
`J5
`
`Coal •
`
`• European Pellet
`
`Commercial
`Carbons
`
`Developmental
`Carbons
`
`20 2
`
`4
`
`12
`10
`8
`6
`BUTANE WORKING CAPACITY, G/100 1.1..
`
`16
`
`Figure 6. Relationship Between Gasoline Adsorption Capacity and
`Butane Working Capacity
`
`the mid
`in
`introduction
`their
`Since
`1970's, the butane working capacities of wood
`carbons have dramatically
`increased from 5
`to
`over 11 g/100ml.
`As previously
`shown,
`this
`corresponds
`to a
`significant
`increase
`in
`gasoline working capacity. Figure 7 demon-
`
`in butane
`time
`improvement with
`strates the
`working capacity as it relates to the percentage
`of the total pore volume in the small mesopore
`size range. A significant effort continues
`to
`increase the amount of pore volume
`in
`this
`critical range so
`that even higher capacity
`carbon is available.
`
`AUTOMOTIVE PORES
`PERCENT OF TOTAL
`
`85%
`
`47%
`
`1981
`MODEL YEAR
`Figure 7. Pore Volume History for Wood Carbons
`
`1988
`
`199?
`
`1976
`
`1978
`
`108
`
`

`

`ADSORPTION CAPACITIES OF IN-USE CARBONS
`To provide additional confirmation
`that pore
`size characteristics is
`the key parameter
`for
`defining performance for the automotive appli(cid:173)
`cation and that butane working capacity can be
`used as a meaningful
`indicator of
`carbon
`performance, several samples of actual
`in-use
`carbons were evaluated.
`The
`in-use carbons
`were from evaporative emission control canis(cid:173)
`ters removed from different types of vehicles
`operated in
`two different geographic areas:
`Covington, Virginia, and Charleston,
`South
`Carolina. After removal from the vehicles, the
`canisters were disassembled with
`the carbon
`removed in three separate portions designated
`as the carbon nearest the air purge
`inlet,
`middle of the canisters relative to flow path,
`and nearest the hydrocarbon vapor inlet.
`The
`in-use carbons were
`then evaluated for perti(cid:173)
`nent characteristics,
`including
`adsorption
`breakthrough capacity, using
`laboratory
`test
`procedures.
`
`FUEL-INJECTED VEHICLES - For evaluation of
`gasoline adsorption breakthrough capacity, 125
`ml portions of the carbon from each section of
`the canister were placed
`into a 375 ml
`test
`canister in
`the
`same order
`that
`the carbons
`were removed from
`the actual in-use canisters.
`Using the new canister
`test system and
`the
`adsorption cycle conditions presented in Table
`2, the in-use carbons were evaluated with
`the
`following test sequence:
`1. Adsorb to breakthrough and purge with
`200 Bed Volumes (BV) of dry air, 15 L/min for 5
`minutes, for a total of six adsorb/purge cycles.
`Record average
`of
`breakthrough
`adsorption
`capacity for Cycles 2 through 6 (Table 4).
`2.
`Increase air purge volume
`to 500 Bed
`Volumes, 15 L/min for 12.5 minutes and continue
`test sequence
`for 25 additional adsorb/purge
`cycles.
`Record average of breakthrough adsorption
`capacity for Cycles 21 through 25 (Table 4).
`
`Table 4. Breakthrough Adsorption Capacity For Carbons
`Removed From In-Use Fuel Injected Vehicles
`
`Carbon
`Ty2e
`
`Wood Granular
`Wood Granular
`Wood Granular
`Wood Granular
`Wood Granular
`
`Wood Pellet
`Wood Pellet
`Wood Pellet
`Wood Pellet
`
`Coal Granular
`Coal Granular
`
`Vehicle
`
`'86 Chev. Celebrity SW
`'86 Olds Calais
`'88 Chev. Corsica
`'87 Chev. Blazer
`'87 Olds Delta 88
`
`'86 Ford Tempo
`'87 Ford Taurus
`'88 Ford Tempo
`'90 Ford Taurus
`
`1 84 Ford LTD SW
`'85 Ford Thunderbird
`
`Pellet (European)
`Pellet (European)
`
`'84 BMW 325
`'85 VW Jetta
`
`Note: Test Canister A: L/D
`Test Canister B: L/D
`
`1.0
`2.0
`
`Mileage
`
`Test
`Canister
`
`Breakthrough
`Adsorption Capacity, g/L
`Gasoline Vapors Butane
`533BV
`500BV
`200BV
`Purge
`Purge
`Purge
`
`84,600
`80,600
`55,500
`31,800
`11,500
`
`75,800
`65,000
`26,000
`10,000
`
`82,900
`73,600
`
`80,000
`65,000
`
`A
`A
`A
`A
`A
`
`B
`B
`B
`B
`
`B
`B
`
`A
`A
`
`43.7
`38.4
`
`32.6
`31.5
`39.0
`31.1
`32.0
`
`30.9
`37.7
`36.3
`36.3
`
`28.3
`29.5
`
`36.2
`35.1
`45.9
`36.1
`38.0
`
`38.1
`42.9
`43.2
`42.7
`
`34.0
`35.2
`
`25.8
`
`31.3
`
`27.7
`35.5
`
`109
`
`

`

`adsorption
`to measuring
`In addition
`capacity with gasoline vapors,
`a few of
`the
`in-use carbons were also evaluated for Butane
`Breakthrough Capacity. The Butane data was
`developed with separate, 375 ml, portions of
`the in-use carbons using pure Butane at a
`flow
`rate of 0.5 grams per minute
`to
`the
`test
`canister. The test sequence was:
`1. Flow butane to test canister to break(cid:173)
`through.
`2. Purge with dry air at 20 liters per
`minute for 10 minutes; 533 Bed Volumes of
`Purge.
`3. Flow butane to test canister to break(cid:173)
`through and report weight gain for this cycle
`(second adsorb cycle) expressed in terms of g/L
`carbon as adsorption breakthrough
`capacity
`(Table 4).
`The average in-use capacities from Table 4
`are summarized in Table 5, along with
`typical
`pore characteristics and BWC's for new carbons
`of the grade and type used in original manufac(cid:173)
`ture of the canisters. Also included in Table
`5 is the average residual organic heel for
`the
`in-use carbons (to be discussed later).
`As discussed earlier, based on
`laboratory
`tests with new carbons, the key carbon charac(cid:173)
`teristics for performance
`in
`the automotive
`application are the amount of pore volume
`in
`the small mesopores and the distribution of the
`pore volume
`relative to micropores. The new
`wood pellet and wood granular carbons have
`relatively
`high
`pore
`volume
`in
`the
`small mesopores and
`relatively low volume
`in
`the micropores.
`These pore size character(cid:173)
`istics result
`in relatively high adsorption
`capacity and low residual organic heels for the
`in-use wood carbons.
`In contrast,
`the coal
`granular and European pellet carbons have more
`pore volume in the smaller pores resulting
`in
`high residual organic heels and low adsorption
`capacity for the in-use carbons.
`
`that
`The data in Table 5 also demonstrates
`is a
`butane working capacity for . new carbons
`meaningful indicator for the expected perfor(cid:173)
`mance of actual in-use carbons. The SAE paper
`by Urbanic et al.(9) stated that butane working
`capacity is not a meaningful measure for carbon
`performance. However, this conclusion is based
`on data generated for carbons that were
`loaded
`with water heels,
`including liquid water,
`to
`unrealistic levels. Since actual in-use carbons
`have much lower water heels (discussed
`later),
`this conclusion relative to the significance of
`butane working capacity
`is not supported by
`in-use data.
`reported
`also
`reference(9)
`This
`same
`working capacity data
`for in-use carbons based
`on a
`laboratory Cyclic Evaporative Loss Test
`Instrument Concept
`(CELTIC)
`test
`procedure.
`However, the in-use carbons were first subjected
`to 100 adsorb/desorb cycles with 60% RH purge
`air. After 100 cycles, which most
`likely
`condensed liquid water on to the in-use carbons,
`the measured capacity was
`reported as
`the
`working capacity of in-use carbons. It is well
`known
`that
`liquid water
`reduces a carbon's
`working capacity.
`In contrast, the work
`re(cid:173)
`ported
`in
`this paper presents
`the working
`capacity of the in-use carbons based on adsorp(cid:173)
`tion for the second through sixth test cycle.
`This procedure more realistically assesses
`the
`condition of in-use carbons since fewer
`labora(cid:173)
`tory cycles are used and liquid water is ex(cid:173)
`cluded.
`the
`The data in Table 4 also demonstrates
`influence of air purge volume. The breakthrough
`adsorption capacity is higher by an average of
`17.4% with 500 Bed Volumes Purge than with 200
`Bed Volumes purge. This trend was expected amd
`confirms that
`increased purge volumes can be
`beneficial in meeting
`future, more stringent
`regulations.
`
`Table 5. Comparison of In-Use Carbon Characteristics to Properties of New Carbons
`
`Adsorption
`Capacity (Average)
`For In-Use Carbons g/L
`
`Residual Organic Heel
`(Average) For In-Use
`Carbons g/L
`
`Typical
`Properties of New Carbons
`Relative Pore Volume
`In Select Size Ranges
`Small
`Meso pores
`
`BWC
`g/lOOml
`
`Micropores
`
`35.3
`
`33.2
`
`28.9
`
`25.8
`
`44.4
`
`60.6
`
`110.6
`
`113.2
`
`1.25
`
`1.00
`
`1.00
`
`0.88
`
`1.00
`
`1.00
`
`1.60
`
`2.00
`
`10-11
`
`9-10
`
`8.5-9.5
`
`8-9
`
`Carbon Type
`
`(1) Wood Pellet
`
`(2) Wood Granular
`
`Coal Granular
`
`Pellet (European)
`
`( 1) Typical properties of new carbon is for Nuchar BAX 950.
`
`( 2) Typical properties of new carbon is for Nuchar WV-A 900 Series.
`
`110
`
`

`

`to
`In addition
`CARBURETED VEHICLES
`from
`fuel-injected
`evaluating in-use carbons
`vehicles, carbons from carbureted vehicles have
`also been tested over
`the years. As an exam(cid:173)
`ple, in-use carbons
`from 1976-1980 vehicles
`were evaluated several years ago
`to compare
`wood granular and coal granular carbons
`(Table
`6). After
`removal
`from
`the vehicles, each
`in-use canister was first subjected to a
`room
`air puge of 40 cubic feet. Using pure butane
`at a flow rate of 1.5 grams per minute,
`the
`butane breakthrough adsorption capacity of
`the
`canister was then measured.
`Working capacities for the five canisters
`with wood granular carbons ranged from about 22
`g/L carbon to 31 g/L, while two canisters with
`coal granular tested at 22-24 g/L. These data
`demonstrated
`that
`the working capacity
`of
`in-use wood granular carbons at the time was at
`least as good as the capacity of coal granular
`carbons. Note that
`the wood granular carbons
`in use in vehicles manufactured in the mid
`to
`late 70's had initial butane working capacities
`in the
`range of 7-8 g/lOOml. Wood carbons
`provided today have been improved significantly
`and have initial butane working capacities of
`9-12 g/lOOml.
`Improvements
`in this
`type of
`coal granular carbon are not known.
`More
`adsorption
`recent
`information on
`capacities of
`in-use carbons
`from carbureted
`vehicles is also
`included in Table 6.
`For
`these tests, the carbon was first removed
`from
`the
`in-use canisters
`in
`three sections as
`discussed previously.
`Smaller portions
`(125
`ml) of carbon
`from eac