-------
July 1979, a punch-plate top deck with
5-inch-iia.neter	holes on 5-3/8-inch
staggered centers was installed. Aluminum
loss was redc.ced, but at the cost of an
increase in large, flexible contaminants.
Little effect was seen in the minus-2-inch
material transferred to the glass recovery
plant.
In an attempt to reject large,
flexible materials and those in plate form,
6-inch-hiah	grizzlies o< 1/4-inch steel
plate were installed on the top deck. This
was generally ineffective, although sane
material traveled down the grizzlies.to be
re jected.
Several modifications were made to the
top deck, including adding a. static partial
flat plate deck with vibrating grizzly bars
above the top deck. In conjunction with
this top deck modification, a new distribu-
tor was fashioned which provided better
distribution of the material on both top and
bottom decks.
In early October 1080, both decks of
this screen were replaced. The top deck
was repi iced by a punched plate with 4-1/2-
inch-diameter openings on 4-7/8-lnch
staggered centers. The bottom deck was
replaced by a punched plate with 2-1/4-inch-
diameter openings on 2-5/8-inch staggered
centers. It has been determined that this
screen does not havp sufficient capacity
for this application, particularly the
bottom deck.
Air Concentrator
The purpose of this zig-zag type air
classifier (No. 17 in Figure 8) was to
remove any nonferrous metals from the over-
size material from screen Ho. 16 and the
eddy current reject stream. However,
composition analysis showed almost no heavy,
oversize metal in that fraction. In
addition, the device va« subject to
plugging. ¦ The unit was converted into a
chute.
Impactoz
This machine (No. 18 in Figure 8) was
designed to crush any friable material in
'the 4 by 2-inch screen (No. 16) fraction.
This was to be removed by screening for
further processing in the glass recovery
system. This also reduced the amount of
material in the 4 by 2-inch aluminum-rich
stream, which aided the recovery of
aluminum by tho eddy current separator.
Analysis showed the iir.pactor was
crushing large friable objects satisfac-
torily,- however, the unit would jam when
surges occurred. Further, little glass was
found in the 4 by 2-tnch stream. There-
fore, the process was streamlined by
removing the inpactor and discaiding the
glass that was contained in the undersize
material that was produced by screen No. 19.
Double-Deck Vibrating Screen
The original purpose of thif screen
(No. 19 in Figure 8) was to remove friable
material crushed by the impactor and any
residual minus-2-inch particles in the
aluminum-rich stream. However, it is
currently a supplement to screen No. 16,
which is too small.
The screen, as purchased, was fitte-1
with .single-screen wire cloth. Rectangular
openings were 2 inches wide by 6 inches
long. Aluminum cans flattened on one end
tended to wedge into the openings,
obstructing material flow and blinding the
screen. A 2-inch-square opening of woven
wire was tried, but soon blinded with paper
and textiles. Next, a 2-inch-diameter hole
punched plate was installed, but this also
blinded. In addition, the driven speed of
the eccentric shaft was increased from 600
to 900 rprn. Amplitude remained at
0.32 inch. No noticeable improvement in
screening was observed.
Experience indicated that a grizzly or
bar deck was needed to prevent blinding.
The punched plate deck was removed and
replaced with grizzly bars 2-1/2 inches
high, 1/8 inch thick, set on a 1-3/4-inch
spacing.
Based upon the insights gained in
several pre.'lous changes, the scrteen was
torn down and rebuilt. The rebuilt screen
utilizes two decks. The splash plate was
increased in length from 12 to 24 inches,
and two chevron-shaped flow distributors
were welded in place. Also, the screen was
raised.
The top deck is made of grizzly bars
24 inches long and 2 inches high with
2-1/2-inch spacing. The purpose of the
grizzlies is to catch flexible material,
such as rags. If the bars are longer than
1)0

-------
24 inches, the action of the screen causes
the rags to drop off the bars onto the
second deck. The lover end of the grizzlies
extends over a flat piste which completes
the top deck. The second deck consiatG of
a punched plate with 2-1/4-inch-diameter
openings. This configuration, shown in
Figure is currently in use. These
improvements work well, as the grinlies
remove a significant amount of flexible
material and the screen does not blind,
althooqh periodic cleaning is still
required. The 2-1/4-inch punched plate is
93% efficient in reooving the undersiza
material from the 2 by 4-inch fraction.
Drum Magnet
This secondary drum magnet unit (No. 20
in Figure 6) removes residual ferrous metal
(not removed by magnets Nos. 7 and 8) from
the 2 by 4-inch tract .on prior to the eddy
current separator. Magnetic items become
trapped in the eddy current separators for
a tine and cause blockages.
Bie principal difficulty of this unit
has been its inability to pass long, thin,
nonferrous objects, such as sticks. Items
as short ae 4 inches have caused jams. Two
changes were effected to alleviate the

r*KTC«,IM_
flow
IKfEtP iPLfliU
PLATE WITH Flow
.SSr. ©PEMINC
{TYPICAL }
II afACEis
iop otcx.
4- FLAT
PLATE
•SEC.ONP DtCK.
£.'/-<¦ D-
N
Figure 9. Equipment No. 19 Double-Deck Vibrating Screen plan and Side View
111

-------
problem. The infeed side of the housing
was enlarged. Also, the drum's original
3/4-inch-wide by 1/4-inch-high knockoff bar
was augmented bv adding two bars 1/8-inch
thick by 1-inch high. This prevents ferrous
metal froa sliding on the smooth drum and
not discharging.
Eddy Current Separator
The purpose of the eddy current
separator, or Al-Mag (No. 21 in Figure 8},
is to remove aluminum from the 4 by 2-inch
fraction, which appears in this feed mate-
rial at a concentration of 5%.
Problems with the Al-Mag centered on
the water coolant system. Each of the
eight banks of electromagnets contains
four iron-core copper-wire-wound magnets
that have copper tubing carrying coolant
water wound in the core. Once-through city
water is used. Overheating was experienced.
The in!rt water pressure was measured and
was at the lower limit allowed by the
separator's manufacturer. Pumps were
installed to boost the water pressure to
60 psi. This was not enough to prevent the
magnets from heating up from low coolant
flow. Magnet temperature causes an
increase in the coolant water temperature.
The coolant water routing is a series
arrangement such that the last bank of
magnets receives water heated from passage
through the other magnets. The heat
increase was sufficient to cause the
calcium carbonates to precipitate out of
the water and adhere to the balls of the
copper coolant tubes. This caused a
restriction which, in turn, caused the
coolant flow to decrease. The reduction in
coolant flow caused even more 'leat to be
transferred to the water, which accelerated
the mineral deposition.
The water coolant lines outside the
Al-Mag were modified to introduce a
flushing liquid and recirculate it until a
clean "coil condition was obtained. The
liquid selected is made by A. 0. Smith,
brara1-named "Unlime." It is circulated in
all coils as the liquid tlowraters are
nanitored to determine liquid flow rate
increase as the coils are cleared of the
mineral scale buildup.
Table 7 shows the recovery efficiency ¦
of the Al-Mag at various belt speeds for a
constant volume of feed material. Prepared
seed cans were used. Deformed cans, had
one end crushed. Note that the efficiency
is dependent upon can shape. The
efficiency was approximately 8* lower for
native (non-seed) aluminum of similar
shapes. Approximately lit .oose organic
contamination is contained ir» the Al-Mag
product.
TABLE 7. AI.-MAC EFFICIENCY
(In Percent)		
Belt
Speed 	
(fpp) Flattened Deformed ' whole Total
Can Type
300	97.7
400	92.9
500	90.7
98.4 100 ^8.4
98.4 100 98.0
95.0 66.6 94.5
The eddy current separator has shown
it is capable of satisfactory ssparatinq
efficiencies. Howevei, it requires more
attention than other materials recovery
equipment.
Air Classifier/Air Knife
The purpose of this two-stage device
(No. 23 in Figure 8) is to removes contami-
nants from the aluminum product of the
Al-Mag. The contaminants in this 2 by 4-
inch fraction are in two forms: (1) light
materials (paper and plastic film), and
(2) heavy items (such as grapefruit and
non^errous metal castings). These cumprise
14% of this material, sufficient to lower
th» price received for the aluminum. Tests
indicate that .90 to 100* of the light con-
taminants are removed and virtually all- of
the heavy contaminants.
This aluminum cleanup system'consists
of. a zig-zag air classifier, its attendant
cyclone and airlock, and an air knife with
blower, as shown in Figure 10. Taken
separately, no difficulties have been
encountered with' the zig-zag air classifier.
The cyclone that de-entrains the organics
from the zig-zag has been modified. This
was built to operate under negative pres-
sure with the butterfly damper and fan on
the downstream side. Light, filmy organics
occasionally passed through the cyclone and
snagged on the butterfly dampei gate. This
nas corrected by removing the gate from
inside the butterfly damper, and making a
1)2

-------
/VvOCiriEP
AIP, BYPAi>
E<^NICb
HCAVlEi
tic,wri
Figure 10. Aluminum Cleanup System
bleed gate and port on the air pipe above
the zig-zag air classifier to modulate the
air flow. The air. classifier is fitted with
a viewing window, allowing air tuning by
observing the separation.
The system was delivered with a
7-1/2 hp motor driving the cyclone blower.
It was retrofitted with a 10 hp motor to
obtain sufficient air flow. Testing on the
zig-zag shoved 97.2* of the aluminum cans
correctly dropped into the air knife, while
91% of the organic light material was
removed by the classifier and bl.~ m to the
cyclone for disposal.
The air knife separates light and
heavy material reporting as heavy from the
air classifier immediately above its
infeed. During shakedown, it was deter-
mined that the 8-1 '2 hp side channel com-
pressor did not produce enough air to
consistently blow aluminum cans acros3 the
knife to the lights side. It was replared
by a 10 hp blower. However, light aluminum
continued to report to the heavy chute.
Viewing window observations in the air
knife showed that light aluminum was being
initially deflected to the light material
side of the air knife, but then rebounding
back into the heavy material side. This
was greatly reduced bv installing two
screened openings to allow the air to
exhaust rather th.in deflect back into tive
air knife box.
Aluminum Hamaernull
The purpose of this hammenf-ill (No. 24
in Figure 8) is to reduce the aluminum to
less than 1-inch particle size. This
raises the bulk density to 15 pounds per
cubic foot. The increased density allows
for economical shipping. The shredding
also liberates moisture and some non-
aluminum. No problems have been
encountered with the mill operation.
13)

-------
12-Mesh Vibrating Screen
The purpose of this screen (No. 25 in
Figure 8) is to reoove minus-12-mesh mate-
rial from the aluminum product. The
purchase specification for this product
allows only 3* minus-12-inesh material. The
material removed 5* of the feed to the
screen; of this, 30% is Aluminum. The
screen has performed satisfactorily.
GLASS RECOVERY SYSTEM
market agreements at Recovery 1. ttiis
resulted in the decision to implement froth
flotation. Tests of this technology at the
U. S. Bureau of Kines' pilot facility in
College Park, Maryland, indicated that the
specifications could be met. Hcwover, it
vas recognized that the lack of color-
sorted products was a limiting factor on
pr'vJuct markets. Markets were secured for
the mixed color cullet iroa Recovery 1.
Glass Mass Balance
As originally designed, glass recovery
was to be done by screening, crushing,
electrostatic removal of conductors, and
optical sorting, all dry processes. This
was described in the Feasibility Study
published fry NCRR in 1972 112). A com-
• mitGiertt to purchase glass recovery system
equipment had to be made in 1974. At that
time, NCRR1s tests indicated thit none of
the optical sorters available could meet
the Glass Packaging Institute's specifica-
tions. which are incorporated into the
The froth flotation process flow was
designed by the NCRB staff, with technical
assistance from Bureau cf Mines personnel.
This initial fljw appears in the chart in
Figure 11.
The shaXedown effort made many changes
in the process. Some of these were small,
but some entailed removal of equipment
items. The modification and tuning effort
is still in progress. The process flow as
it exists at this tine is shown in
reace^t ADDCD
TO RKVT 	»
CiU FEED
OIL
REO-EANER. FLQKT
P£.0-EAKjE.R FLOAT
HYDKCJCYCLONIE.
VACUUM TABLE.
clEANE-U FV.0KY
Oku WE.A7ER
OKJ SITE. StLO
STOSkAC,e
"RftOquER.
FL&A.T
Figure 11. Initial Froth Flotation Process Flow (Cont'd)
1)4

-------
£
AIR CLASilflER
AIR CLA$Slf>EP.
MLAVtid
mpA
IT OR


I'/J." iOR.OPtMlNC
FLAT PECK. ICKJOsI
CRU5UCR
l* iCUARE OPEVUUQ
FLAX DECK. JCP£.ENJ
X
@
[ SURGE B>N»
	x—

&
ROD A1ILL
'/h~ PUAT
DECK SCRE.ENJ
S\±£
j V!ftRA.T\Mq fCEPEp"
©
J-K.' M,MERAL JVC,
OCCAM lO

ZO/MEiW FLAT DICfc.
:5CRE.I>J
»XO **i«
PC\NATtR>MC
.tREEM

ORCANO
REJECT
kv drocvclonL
-20 /nejH.
T©
UNDtRFLC#M
£CELL V?«tFV6KT
»WPB£>CVCLQMt
UNDERFLOW
Figure 11. Initial Froth Flotation Process Flow
(Continued)
133

-------
Figure 12. Also shown in the flow is the
mass balance for glass in the system.
Those data arc based upon samples t.iken
during operation in September 1980. No
formal testing of the glass system has been
conducted.
The mass balance for the glass Is
calculated at 100 tpd input of KSW, as this
was the operating rate during the testing
period. At this rate, the loss of glass in
the troaoel is estimated at 6». This is
much higher than the 1% loss found during
the tromnel testing at 62.5 tph, as
discussed in the Recovery Feed Preparation
section. AJso, there is a loss of tne
minus-l/4-inch glass (about 10V) in the air
classifier. The figure for the quantity of
minus-l/4-inch glass is based upon the
tromnel testing. A larger loss, cout 20\,
occurs on the vibrating screen (No. 16) .
Ttio changes to the screen discussed below
are expected to reduce this; however, some
of this loss may be shifted to vibrating
screen No. 28.

©
KOO MILL
©
VII3RAT»M<£|
.5Cfl,E.tN
AlR CLASSIFIED


'
r HEAVIER
S'.lV OoJOl£ aec.*.


-IV
<
IS
SOKC.E f.M
'
r
V AMttM
Vie»EATIh4C, SRCE W
r
¦*VM"
COKJVEYOU.
[
Q
PBWATEJUNm
3CH.E€M
-JO M»SH
l«0
Q>
HrDKO cycuowE
ONOE R FIjOVV
jctu. *^r.gcloa.T
ItT
©
©

gonmeyor
HVt3 E.C> CVC uow t
ONDtRFUSW
Figure 12. Current Froth Flotation Glass Recovery Flow and Mass Balance (Continued!
136

-------
RCAC.CMT ADDCD
TO FIP5T 	1
ROUCUtRCCU,FtSD
\H>7
3 CELL ROUGHER
FLOAT
0
CLE.AME.R PLOAT
?tt£UAN£U FIOAT
1 RE. CLEANER FLOAT
¦P
@
%
VACUUM



HVDROtVCLOMt

WATER,
fRCATMIMT
THtKMAU
PllH D^MEATEf*
¦©
l.Co^)
pryek
oiu flow
tfcb
©
o.-t SlTC £>n_©
STORAGE
T»iu wc. t>
O.xM
wA-re
TREATMENT
Pigure 12. Currant Froth Flotation Glass Recovery Flow and Mass Balance (Cont'd)
As can be seen In Figure 12, there is
a tmufccd flow rate change (4 to 1) at the
surge bin. This highlights the very impor-
tant function of this item in the glass
system operation. The hydraulic portion of
the system must oe run at a continuous
process. However, the front end of the
glass system operates only when waste is
being shredded. The shredding operation,
part of the recovery feed preparation, is
on a start-stop basis. The stops are caused
by lack of a landfill loadout trailer,
coffee and- lunch breaks, shift changes, and
equipment outages.
Tho wet portion of the glass systeo is
currently operated at about 40% of design
capacity. There are no specific known
bottlenecks which prevent increasing the
flow. However, insufficient feed naterial
is presented during the 8 to 10-hour glass
system operation to toake increasing the
flow worthwhile. The recovery in this
portion of the system is 80V. There are.
ho current plans to try to raise this yield.
The' current goal is the production of a
sufficient quantity of recovered glass to
allow a large-scale Belt teBt to be run in
a glass plant. Approximately 350 tons of
product have been delivered to the evens-
Illinois plant in Hew Orleans. Tests
Indicate that this meets the Glass
Packaging Institute specification.
Equipment Modifications
The glasB recovery systea is
essentially the last module in the recovery
equipment train. Because of difficulties
with upstream processing equipment, work on
the glaBs system was delayed for a number
of months. During attespts to operate the
systeo in 1977 and early 1979, it proved
impossible to achieve uninterrupted opera-
tion of sufficient duration to perform the
tests necessary for evaluation. Effort
was concentrated on the ferrous and
aluELinun systems. In mid-1979, effort was
again directed to the glass recovery
system. Since that time, the focus has
'.J7

-------
been on sequentially operating, modifying,
and evaluating the aluminum and qlass
systems.
The glass recovery technology employed
involves wet processing. Initial hydraulic
balancing problems caused water to overflow
the tanks and equipment, which contributed
significantly to the difficulties encoun-
tered. Overflow is now under reasonable
control. This is tho result of both equip-
ment improvements, which have led to better
operation, and to the installation of over-
flow control sumps and piping. This not
only improved housekeeping, but the
modifications made to control overflow
allow the operator more time to correct the
problem without having to shut town the
entire glass plant. Operating days wi'.h as
much as 10 hours of continuous operation
have been achieved. This has allowed for
test and evaluation and the beginning of an
optimization process.
Each piece of equipment shown in the
initial process flow (Figure 11) is dis-
cussed briefly below. The enphasis or each
is the modification made. Equipment items
which were eliminated are also discussed.
Two equipment items, the impactor
(Mo. 18) and the screen (No. 19) are shown
in the flow diagram in Figure B. These are
discussed in tha Aluminum Recovery System
section. The itrpactor, which was installed
to break large glass in that flow path, was
reabved when analysis showed very littlq
glass in that stream. Also, the underflow
froo the screen (No. 19) was rerouted to
the reslduei discharge.
4 by 2-Inch Vibrating Screen
This screen was discussed in some
detail in the Aluminum Racovery System
section.
Double Roll Crusher
This crusher (No.27 in Figure 11) was
intended to reduce the Bixe of all friable
material in the minus-2-inch undersize
fraction from the 4 by 2-inch screen
(No. 16).
Numerous janis occurred in the crusher.
These were caused by stones and metal items
lodging between the tocLhod rolls. Jams
could be prevented by increasing the dis-
tance between the two spring-loaded rolls;
however, less size reduction resulted when
this was done. A review was made of the
particle-size distribution of glass
reporting as trommel undersize. It was
determined that approximately f)0* was
minus-l-inch before crushing. Based upon
this data, the crusher was removed.
Ono-Inch Vibrating Screen
This screen (No. 28 in Figure 11) sizes
the material for the minerals jig. The
undersize. minus-l-inch, is discharged into
a surge bin for feeding to the jig. Studies
have shown one inch to be the practical
upper liirdt on the size of the jiq feed.
Initial calculations based on an infeed
rate of 12.3 tph indicated that 4 screen
area of 40 square feet was required. The
supplier, Vibranetics, submitted that a
screen with 16 square feet area would
suffice. This proved groscly inadequate.
Several remedies unre tried. The original
1-ir.ch (nominal) opening size screen cloth
was, in fact, enly a 0.875-inch opening.
This was chanced to a cloth with a 1-inch
clear opening, but no improvement w«is
observed. Two symptoms were noted:
(1) insufficient area, and (2) excessive
blinding. Washir; with high-pressure hot
water provided short-term relief. Constant
washing with ambient-temperature water was
better than no washing, but this increased
the moisture content of the undersize mate-
rial being stored in the surge bin, which
increased the difficulty of flow control
into tho jig. Also, washing separated the
fine dirt from the oversize material,
res-jiting in a mud buildup on the bottom
of tho screen.
As a method of alleviating the
blinding problem, a cloth with 2 by 1-lnch
rectangular openings was tried. But this
allowed particles which exceeded the jig
top size to pass. A 1-inch-diameter hole
perforated plate was installed, but the
open area was less than the l-inch-square
mesh. As a result, it could not process
material rapidly enough to keep pace with
the jig.
A temporary solution way to remove the
screen completely and make the 1-inch sizo
separation on screen No. 16 (sec Figure 11),
as mentioned previously. This change was
made, and sufficient throughput capacity
was obtained. However, sticks and other
long (3 inches or more) items entered the
jig and caused plugging in the jig dis-
118

-------
charge pumps. Also, a negative effect was
the introduction of an increased ainount of-
1 by 2-inch material into the middling
fraction. This additional material had a
detrimental effect on eddy current separa-
tor performance. A perforated screen wit.i
1 by 3-inch ovals was tried; this solved
ttw stick problem, but it also locked the
necessary capacity.
The supplier, Vibronotics, offered to
exchange the small screen for a
5	by 12-foot unit (moro than three timos as
large in area). However, this could no't be
easily accomodated into the existing con-
figuration, due to the sire and weight. A
price rc-iidtiG ..s fair by NCRF was
negotiated. However, it was determined
that a light-duty screen could be purchased
for less cool. Also, the liqht-duty 3cree»
would not have the space and structural
requirement as the one proposed by
Vi^ranctics. A 4 by lu-toot screen was
installed with a punched plate with 1-inch
diameter openings. A slotted punch plate
with 3/4 by 3-inch oval openings has been
ordered for the unit.
Surge Bin
This tin (No.-. 29 in rigure 11) was
constructed on site by the erection
contractor. Its purpose is to absorb surges
to that uniform feed to the jig can be main-
tained. The bin ic rectangular, with a
converging bottom. The top 7 feet has a
6	by 7-foot cross section. The bottom
7	feet converge cr< three sides to a rectan-
gular size of 2 Lv I feet. The nature of
the material and ti:i converging bottom
caused bin discharge problems from the
initial testing. Several remedies were
tried.
djgh-density, low-surface-friction
polyethylene was used to line the steepest
converging side. No apparent improvement
was seen. . Water injection on three sides
was tried with limited success.
Direct spray onto the upper surface
has been the jaosi. successful. The disad-
vantage of this method is the lack of
control on rate c.' discharge of the wet
material from the bin bottom.
A modification was also made to the
discharge of the bin. A manually adjustable
hinged opening was fitted which allows the
discharge from the bito be modulated.
AJso, because the sire of the bin discharge
openinn can be quickly doubled, incipient
plugging can bo avoided by opening the gate.
This provides additional modulation to the
variable speed which discharges the bin.
The disadvantage of the gate is that it
rcquiVes the attention of an operator.
Minerals jig
The purpose of the minerals jig INo. 30
in Fiqure 111 is to remove organic materials
from the qlasn-rich fraction. A secondary
purpose in to remove the small dense
objects, mainly nonferrous metals, from the
jig product.
This jig wao originally fitted with
BETZ water spray nozzles to fluidize the
incoming material in the vertical chute at
the feed end of the machine. These nozzles
did not prove to he practical, and were
removed.
The bed of the jig was Initially
operated with 1/4-inch-square cesh cloth
over l.'-l-iich-vide by 3/5-inch-lori^ slotted
hole* in steel plate. At the recomnendation
of the manufacturer, this upper screen
cloth vas removed. The jig was then
operated with punch plate steel and ball
raggi-g, 1/2-inch diameter in the first
half of the bed and 5/B-incn diameter in
the second. Frequent plugging of tho hutch
drains occurred frou large objects passing
the slotted holes. Therefore, a new
1/4-inqh-Gquare mesh stainless steel screen
cloth was installed. This prevented large
objects from passing into.the hutch, but
did not totally solve the hutch Urain
problem.
Early jig operation was donp with gate-
type valves fitted to each of the hutch
drains which wore difficult to clean. The
gate also contributed to plugging. These
valves were replaced by Bhort sections of
soft foam rubber hose fitted with circular
hose clamps, ttvis allows for a variable
orifice with s circular cross-section. Also,
the transition in size is smooth, so the
outlets are no longer £>o prorto to plugging.
Periodic disassembly end cleaning is
required. (Concentric valves are available
commercially, but are expensive.)
The jig was initially set with a 7"
angle of declination. This caused the
fluidizing of «ater4al to occur about half-
way down-the 16-foot-long bod. To r'luidlze
139

-------
the material closer to the feed end, the
jiq angle of declination va» decreased
to 5".
This aided the jigging process, but
still did not illow tot. ton water to pulse '
up through the bed of the first hutch.
Therefore, the fiist third of the bed was
covered with a .steel plate and horizontal
spray nozzles installed at Lite feed end to
sluice the material into the working
sections of thn bed. This reduced the
capacity, but proved workable at the
reduced feedrate.
In susanary, the jiq har. provcr. to be
the most sensitive'and difficult machine
to operate. Difliculties were encountered
with balancing trip and botton water and
naintainir.q proper bed heiqht at varied
infeed rates. Oreratinq procedures derived
from the ninorals beneficiation industry's
use of til a jig do not always apply to
iiqginq qlass frou organic waste.
Variability of the wasto requires adjust-
ments in the )iq control settinqs to assure
satisfactory material separation.
jonica Elevator in 7 fcreen
This vibratinq screen (No. 31 in
figure 11) was to receive the orqanic residue
discharge fro the jig arid dewater it. The
residue was to be discharged to landfill and
the water recycled.
The screen was initially fitted with a
60-mosh stainless steel woven wire cloth
which proved too tiqht a weave. The result
was no dewaterinq, with most of the water
passinq over the screen and beiriq discharged'
with the organic*.
To alleviate this problem, a 10-mesh
wire cloth was tried, with little improve-
ment. PinaTly. 16-nesh wire cloth (G.C18-
inch-diaacter wire) was installed and is
satisfactory. The penalty extracted by the
more open screen Is the addition of minus-
16-«nesh qrit and fines to the recycled
water. This increases the frequency of the
cleanout of the recycled water tank. The
oversize material is dewatored sufficiently
to be conveyed by canventiuial smooth belt
conveyors. Periodic washing of the screen
with high-pressure hot water is required to
prevent blinding by fatty materials which
close the screen openinqs.
1/4-Inch by 20-Mesh Vibrating Screen
The purpose of this screen (No. 32 in
Figure 11) was twofold: (1) to site the
ylass fraction for the froth flotation
cells at less than 20 nesh, and (2) to
remove any non-friable material larger than
!/4 inch frotn the crushing circuit.
This double-deck screer. was delivered
"ith a top deck of 1'4-inch square opening
wire screen cloth supported by 1/4-inch
rods longitudinal, and a bottom deck of
carbon steel with slotted openings 15-mesh
by S-mesh (nominal 20-mesh).
Two problems were encountered with
this arrangement. The carbon steel i iterial
rusted and beqan allowina oversize (plus-
20-neslil naterial to reiert to the under-
size (minus-20-mesh), thence to the froth
.flotation cells. Also, oversize, material
penetrated the seams in and around the
screens.
The rustinq problem was solved by
replacing the oriqinal cloth with 304 stain-
less steel screen cloth. The iir3t attempt
to solve the unsupported joint problem was
with two sections of 3creen cloth ¦» *-eet
wide by 10 leet lonq, with hooks on noth
sides, ft longitudinal searu appeared
feasible because thero was a flat bar
supported down the screen csnterlina which
could be used to retain the hoo&s. This
arnnqenent solved the support problem, but
placed excess tension on the hooks, which
resulted in the cloth tearinq away trom the
tension hooks on the sides.
After several attempted solutions,
discussions with the screen manufacturer,
CE-Tyler, resulted in replacing the bottom
deck with a two-piece stainless steel
{216 SS> 20-nesh screen (two 6-foot by
5-foot panels with 0.017-inch-diameter
wire). Backing cloth panels were installed
(B-nes!i, 0.02S-inch-diameter wire) to.
provide tensile strength and prevent the
liqhter-gauqe wire cloth above it frim
tearinq. These were installed ufing the
rubber wedges that provide seals along the
edge where the two panels abut and at
eitjior end. . problem was encountered with
the impingement nf the qlass on the top
screen. The result was wear and breaking
of the screen wires. Several :har7es to
the feed distributor in the -in it w.^re mad".
In addition a- third 8-nesh scr^e.n cloth
wjs installed over the 20-mesn deck in the
140

-------
upper half of the screen to serve as a
velocity brake Cor the glass particles.
This appears to be satisfactory.
Rod Mill
The purpose of the rod mill (No. 33 in
Figure 11) is to crush the glass fraction
prior to froth flotation. Sizing at 20 mesh
is done by the vibrating screen (No. 32).
It was noted that the slurry inside
the mill was seeping out on the feed end
near the large bull gear. This sl'iiry
contained small particles of very abrasive
glass that were beinq picked up by the
revolving bull and pinion gear and abrading
the faces of the gears. The leik was
stopped and the qaps between liners sealed
by applying a coatinq of Nordback, an epoxy
nvinul'acturod by Rexnord. The first layer
was of steel powder-impregnated epoxy,
followed by a Layer of cetamic bead-
impregnated epoxy „ror abrasive resistance:
No further leaks have occurred.
Several rod charges have beea used
since shakedown started. The size of the
charqe - nuinber of rods - and the amount of
water for a aiven amount of feed determine
tlic piiit.ii.le size of tne discharge.
Currer.tly, the nill charge is 37 rods
totaling 4,073 pounds. This is approxi-
mately 20\ of mill volume. The charge
varies as the rods wear. The particle size
distribution of the mill product is analyzed
to c tormine the need for rod replacements.
Hydrocyclone
The puipose of tl.is 10-inch
Hydrocyclone (ito. 34 in Figure 11) is to
remove the solids which are finer than
15C mosh fron the undersire mater'?1 (minus
20 .nesh) from Screen No. 32.
The initial configuration had a 3-1/2-
inch vortex finder and a 2-inch adjustable
apex. The underflow was observed and
attempts were made to optimize its opera-
tion by adjusting the inlet pressure. The
performance was not satisfactory; therefore,
the 6-inch-long spool section immediately
above the apex was removed. This improved
the performance. However, the most
critical parameter affecting cyclone opera-
tion was found to be pulp density, or
percent solids of.the slurry fed to the
hydrocyclone. Low pulp density resulted in
mo
-------
Increased reagent concentration
increases the yield, but does not recover
all of the fraction. Also, It increases
the cost. The system row recovers 87* of
the glass entering the first rougher cell
(initial flotation cell) . A slightly
smaller particle-size distribution might
improve the yield, but has not been tried.
Hydrocyclone (No. 38 in Figure 11).
The purpose of this 6-inch hydrocyclone is
to feed the vacuum filter and partially
dewater the slurry deposited on the vacuum
table.
This unit is located immediately ahead
of th«i vacuum table (Ho. 39) . Initially,
all of the underflow (apex discharge) was
put directly on the vaci'um table, with the
overflow water piped to the recycle water
¦system. To balance the cyclone, a tee was
fitted to the overflow and valved so that
one leg of the overflow was routed back
into the sump serving the pump which
supplies the cyclone slurry. The second
leg was routed to the waste water tank.
Vacuum Filter IKo¦ 39 in Figure 11).
The purpose of the vacuum filter is to
partially dry the glass product. Problems
experienced with this filter have been in
the area of appropriate-size filter cloth
weaves' and filter media.
Magnified examination of the original
top filter clot ¦. (400 SCFM rating) revealed
the individual polypropylene threads had
swollen, thereby reducing the open area.
All 20-panels were stripped from the table
and new- polypropylene cloth rated at
250 SCFM caulked onto the table above thn
existing backing cloth. Moisture removal
down to 5% is now possible, and no other
problems of dewaterinq were found.
The vacuum pump draws air through the
vacuum table to remove the free moisture.
It is constructed with carbor steel
impeller vanes that tend to r~ist and seize
to the vane housing if not frequently
rotated. In extreme cases, the pump must
be disassembled and cleaned to free the
vanes. during shakedown, several periods
of two to three vaekB with no activity were'
encountered. After the initial period of
inactivity, the impeller froze, but was
freed by rocking the impeller shaft. A
standard operating procedure was adopted to
start and run the vacuum pump for several
minutes once a week when an outage prevents
normal plant operation.
Dryer (Ho. 10 in Figut. 11). The
purpose of the Holoflite dryer is to reduce
the moisture level of the glass product
from 5 to 10% of the feed to less than 1%.
This device is essentially a twin-screw
conveyor which works by parsing a hot
(600°F) thermal fluid throuyh both the
jacketed trouqh and the hollow sci^ flites.
When the Holoflite dryer was purchased
used, both flites and jackets sustained a
hydrostatic pressure test. In addition, the
rotary valves were rebuilt by the equipment
supplier prior to purchase.
In. -ifficient moisture removal to
satisfy the product speci Mcation was a
problem from the b°ainning. Tests showed
the trough jacket was *:*ceiving sufficient
heat transfer oil to reach operating tem-
perature. However, the flites	not.
This was traced to improper valve operation.
The rebuilt valves had been used in a
different configuration and had been
improperly modified for this application.
Mew valves were purchased and installed,
and no further hcau iirjaiance was seen.
Althouqh the dryer passed hydrostatic
pressure tests, it developed two leaks in
the trough where rust fiad caused several
pitted are-s of thin metal. The conse-
quence of tht-
-------
test reports, and technical investiga-
tions, prepared as an aid to consulting
engineers, systems designers, and
others engaged In the design, con-
struction, jri implementation of
resource rec jry facilities: National
Center for Resource Recovery, Inc.,
Washington, D. C.
3.	National Center for Resource Reccvery,
In... »ew Orleans Resource Recovery
Facility Implementation Study - Equip-
ment. Econoaici.'. Environment; National
Center for Resource Recovery, Inc.,
Washington, D. C. (19*77) , 427 pp.
4.	Stephen E. Steimle, Solid Waste
Composition Study, City of N'ew
Orleans; prepared for the New Orleans
Department of Health (September 1972).
5.	National Center for Resource Recovery,
Inc., Trommel Performance at Nominal
Design Conditions. Test t.'o. 1.01,
prepared for the U. S. Environmental
Protection Agency, Cincinnati, Ohio,
under Contract No. 68-01-4423, Otarch
1980), 129 pp. Also available in
abbreviated form as Tronrvel Initial
OperaL+nq Report: Recovery 1 (TR-79-3) ;
National Center for Resource Recovery,
Inn., (October 1978), 26 pp.
u. A. F. Taggart, HandJaook of Mineral
Dressing, (New York: Wiley t Sons, 1964)
7. G. M. Savage, L. F. Diaz, and
G. J. Trezek, Performance Characteriza-
tion of Air Classifiers in Resource
Recovery Processing, presented at the
1980 National Waste Processing
Conference, Washington, D. C. (May 11-
14, 1980). Sponsored by ASHE Solid
Waste Processing Division.
8.	National Center for Resource Recovery,
Inc., An Investigation of Two
Approaches to Air Classification
(TR-80-3) , National Center for Resource
Recovery, Inc., (February 1980), 21 pp.
9.	I. Handler and K. Runyon, "Performance
and Testing of the Ferrous Metals
.Recovery System at kecovery 1,"
Proceedings of 1980 national Waste
Processing Conference (New York:
American Society of Mechanical
Engineers, 1980), pp. 173-188.
10.	National Center for Resource Recovery,
Inc., Magnetic Drum Sepaiator Perfor-
mance Scalping Shredded Tro.imel Over-
flow at nominal Design Conditions,
Test No. 4.03, prepareU for the U.'S.
Environmental Protection Agency, Office
of Solid Waste Management Program,
Contract No. 68-01-4423, Cincinnati,
Ohio (1980) , 45 pp.
11.	National Center for Resource Recovery,
Inc., Magnetic Drum Separator Perfor-
mance Scalping Tronrv;! Underflow at
Nominal Design Conditions, Test
No. 4.01, prepared for the I). S.
.Environmental Protection Agency,
Office of Solid Waste Management
Program, Contract No. 68-01-4423,
Cincinnati, Ohio (1980), 62pp.
12.	'National Center for Resource Recovery,
Inc., Materials Recovery System -
Engineering Feasibility Study, National
Center for Resource Recovery, Inc.,
Washington, D. C. (1972), 284 pp.
143
-------
SUMMARIES OF,COMBUST IONS OF REFUSE-DERI ID FUELS AND DENSIFIED FUELS
Carlton C. Wiles
U.S. Environmental Protection Agency
26 West St. Clair Street
Cincinnati, Ohio 45268
ABSTRACT
The Solid and Hazardous Waste Research Division has supported Investlaatlons of
the use of refuse-derived fuels (RDF) for onwer aeneratlon. Two projects used fluff
RDF in municipal power generating boilers. The St. Louis-Union Electric refuse fuel
project was the first demonstration plant in the United States to process rtw municioal
waste for use as a supplementary fuel In a utility boHer. A second such project was at
Ames, Iowa. Two other projects Investigated the use of denslfied RDF (d-RDF) $s a sub-
stitute for lump coal in smaller spreader stoke" bo-ilers.. AoproKiroately 2,SOD tons of
d-RDF were combusted with and without coal at plants located 1n Hagerstown, Maryland,
and Erie. Pennsylvania. Although other RDF combustion projects have been supported,
these four are summarized ir this paper.
INTRODUCTION
Recent concerns about th3 cost and
availability of energy have sparked great
Interest in the possibility of recovering
energy from refuse. Such energy recovery
would not only offset steeply r1s.1ng fuel
costs, it would also help solve the Na-
tion's solid waste disposal problem. The
U.S. Environmental Protection Agency (EPA)
assigned the Municipal Environmental Re-
search Laboratory (MERL) in Cincinnati,
Ohio, major responsibility for research
and development in the field of recovery
and use of municipal solid waste. This
paper describes four major EPA projects
that are aimed at the recovery of energy
from solid waste. Th®se projects combust
refuse either directl;1 for steam recovery
or in combination with fossil fuels for
power generation. The latter involve the
processing of the refuse to remove the
combustibles for use in a modified power
generation- boiler, usually 1n combination
with coal. The processed refuse is usu-
ally referred to as refuse-derived fuel
(RDF).
The RDF concept 1n the United States
was originally considered for major power
generating facilities that burned Dulver-
ized coal. However, the use of RDF need
not be limited to laroe users and may in
fact be more valuable to small power pen-
eratlnp facilities. Small Industrial and
institutional boiler owners may find RDF
an attractive and cheaper alternative to
fossil fuels, for which they receive no
quantity discounts, as do the laroe user;.
In addition, small users may have Increased
flexibility in neootiatlna contracts for
RDF (especially with regard to lenqth of
cormitment). Many snail power generators
are economically marginal because-their
boiler facilities are older, coal-burn1no .
models that require costly air pollution
equipment. The use of RDF may iielo such
facilities absorb the cost for such controls.
RDF prepared for large utility boilers
1s typically composed of the liqht fraction
of shredded refuse that has been air-clas-
sified, screened, or otherwise processed to
remove the noncombustlbles. In this fluffy
form, 1t can be pneumatically fed Into the
suspension utility boiler. For the smaller,
stoker-fed boilers, however, a denslfied
form of RDF is probably more desirable.
>44
-------
This densifled refuse-derived fuel (d-RDF)
nay approximate the physical uiaracteristics
of the stoker coal fed to the toiler. RDF
1n this form offers increased flexibility
in transport, handling, and storage, and
It can be mixed directly with the coal and
fed to the boiler with few if any modifi-
cations.
EPA realized that for fuel recovery
from refuse to be widely implement:;!, a
credible experimental program was needed
to establish the environmental acceptabil-
ity and economic and technical soundness
of such fuels. Of major importance to
users would be the effects on boiler phys-
ical facilities and operations, and on the
environment. A number of demonstration
projects were thus undertaken by EPA to
provide the necessary experimental data
and operational experience to interest
potential users in implementing the con-
cept.
Among these projects were the four
combustion studies that are the subject
of this paper. They include two demon-
strations of tne preparation and firing
of RDF (one at St. Louis, Missouri, and
the other at Ames, Iowa) and two demon-
strations.of the firing of d-RDF pellets
in spreader stoker-fired builers (one at
a Hagerstown, Maryland, institutional heat-
ing boiler, and one at a larger industrial
boiler in Erie, Pennsylvania). The com-
bustion studies are summarized as follows:
St. Louis/Union Electric Refuse Fuel Demon-
stration System
This project was the first demonstra-
tion plant in the United States to process .
raw municipal waste for use as a supple-
mentary fuel in a utility boiler. It was
designed as a simple, minimal investment
experiment to test the combined firing of
coal and RDF. Two separate facilities
make up the system—a processing plant
operated by the City of St. Louis, and an
RDF receiving, handling, and firing oper-
ation at the Union Electric Company's
Merartec plant near St. Louis. At the
processing plant, raw solid waste is mill-
ed to a nominal 38,1-nm (1-1/2-in.) par-
ticle size and air-classified into light
aad heavy fractions. The light fraction
(which accounts for approximately 80% to
85% of the incoming municipal refuse) is(
temporarily stored and then hauled 29 ktn
(18 miles) by transport truck to the
Meramec Plant. At the power plant, RDF
is unloaded from the transport trucks In-
to a receiving bin from which it is con-
veyed pneumatically to a surge bin. A
pneumatic feeder system conveys the RDF
from the surge bin through four separate
pipelines directly to the boiler.
Ames, Iowa, Solid Haste Recovery System .
The Ames project was the first con-
tinuous, full-scale solid waste recovery
system for the processing and burning of
municipal solid waste as a supplementary
fuel for power generation. Two facili-
ties make up the system: One is a 136-
Mg/day (150-ton/day) waste processing
plant where the RDF is received by means
of a pneumatic transport system stored
in a 454-Mq (500-ton) Atlas bin, and
fired into two small steam-electric
boilers. The 50-ton/hr processing plant
incorporates two stages of shreddino,
ferrous and nonferrous metal recovery,
and an air density separator.
Firing of Coal:d-RDF Blends in Spreader
Stoker-Fired Boilers at Hagerstown, Mary-
land
Experimental combustion tests of d-
RDF pellets were conducted at the State
of Maryland Correctional Institution for
Men. The d-RDF pellets used in these
tests were prepared by the National Cen-
ter for Resource Recovery (NCRR). The
boiler plant consists of a battery of
three ISO-psig Erie City boilers. Their
design steam ratings are 78,500, 60,000,
and 25,000 Ib/hr. Each unit is equipped
with Hoffman Combustion Engineering Fire-
rite spreader stokers to distribute the
lump fuel in the furnace. The large coal
pieces that do not burn in suspension are
combusted on the surface of Hoffman vi-
brating grates, and ash is discharged to
the front. The boilers have tub'e-and
tile furnaces.
Firing of Coal:d-RDF Blends in an In-
dustrial Spreader Stoker Boiler in Erie,
Pennsylvania
Long-term demonstration tests were
conducted at the General Electric Power
Plant in Erie, Pennsylvania, to determine
more accurately the general performance
of,d-RDF in a boiler representative of
those us?d throughout industry. Pellets
of d-RDF wc.-e supplied by Te-ledyne Na-
tional and NCRR. The demonstration con-
145
-------
sisted of four tests during which fuel
handling, boiler performance, and stack
emissions were monitored. The test boiler
was a 150,000 lb/hr steam generator con-
sisting of a Babcock and Wilcox two-drum
Sterling Boiler with a Detroit Rotograte
Spreader Stoker.
Other Related EPA Project
Several other projects were under-
taken to establish the economic impact of
burning d-RDF. Briefly, they include:
—	A research grant to tfcs National Cen-
ter for Resource Recovery (NCRR) to
study production requirements for d-
RDF;
—	A research grant to the University of
California to supplement the NCRR
grant in examining the theoretical
basis for producing and using d-RDf;'
—	A contract with Teledyne Matronal to
produce 2000 tons of d-ROF and provide
additional information on production
processes.
ST. LOUIS/UNION ELECTRIC REFUSE FUEL DEM-
ONSTRATION SYSTEM
In April 1972; the City of St. Louis
began 'to operate a 300-ton/day municipal
solid waste (MSW) processing facility de-
signed to produce RDF and to recover fer-
rous metals. The RDF has been burned and
tested by the Union Electric Company as a
supplement to pulverized coal in steam-
electric boilers. During the tests, the
RDF made up 0% to 27% of the heat input
to the boilers.
This project was partially funded by
an EPA arant to the City of St. Louis.
Technical, economic, and environmental
evaluations of the project were conducted
by Midwest "esearch Institute under con-
tract to EPA. Performance assessments
were made both at the processing plant
that produced the RDF anit at the power
plant where it was burned as a coal
supplement.
Processing Plant Evaluations
Process Description--
Residentlal and commercial waste is
first shredded in a horizontal hamnerml 11.
to a 1-1/2-in. particle size. A cyclone
particulate collection system over the
hannermill feed throat collects large
pieces of paper that blow back out of the
mill. The shredded material is then air
classified b> a vertical chute classifier,
where the lighter materials sucn as paper
and plastic films separated in a tur-
bulent air stream '•om heavier materials
sucn as metals, glass, rubber, wood,
textiles, and thick, dense plastics. The
light materials is de-entrained from the
air-stream in a cyclone and is conveyed
to a storage bin before it is trucked to
the power plant. The heavy fraction is
processed to recover ferrous metals for
use as a scrap charge in a steel mill.
The remaining residue is landfilled.
Plant Ooerations--
Processinq' rate--The overall process-
ing rate average for the 53-week test
period was 168 Mg/8-hr day (185.5 tons/8
hr. day) at 31.0 Mg/hr. (34.2 tons/hr.)
The plant was operated at maximum capacity
of 272 Kg/8 hr. day (300 tons/8-hr. day)
during the first- 2 weeks of the testing,
demonstrating that the plant could sus-
tain this rate for at least a short pe-
riod. The maximum 1-day average process-
ing rate was 45.8 Mg/hr. (50 tons/hr.).
Down time--Two major equipment, break-
downs occurred at the processing plant,
along with several plant shutdowns result-
ing from equipment maintenance outages at
the Union Electric power plant and from
repair of an electrical substation serving
the processing plant. Planned shutdowns
for normal maintenance also occurred.
Material balance—Plant material
balance by weight showed that plant out-
put averaged 7.6t less than Input. Scale
error and moisture and particulate loss
from the air classifier and dust collect-
ion system were identified to account for
1.6% loss, leaving a 6% error. Moisture
loss from the hammer-mill is thought to
be the major cause of this loss.
Product Analysis--RDF has approx-
imately 42% of the heating v.alue and 2.7
times the ash content of Illinois Orient
6 coal, but the refuse fuel has only some
12% of the sulfur and 35% of the nitrogen
content of the cOal. Ferrous metal re-
covered is a marketable byproduct used
as part of the scrap charye at a near-
146
-------
by steel mill. On the average by weight,
the RDF represents 80.6* of the process-
ed raw refuse, and the recovered ferrous
metal accounts for 4.5%. The plant re-
ject material, which is landfilled, has
very low energy content.
Operating costs--
Operatjng costs Increase rapidly
when the plant is operated below its de-
sign capacity. Total monthly operating
costs for the refuse processing plant
plus the receiving facility ranged from
$4.45 to $57.99/Mg (J4.04 to $52.6/ton).
Total average operating costs for the 53-
week period were $8.26/Hg ($7.49/ton).
Fnviromnental Assessments--
Air—Future plants using an air class-
ification system of the type used at the
St. Louis demonstration plant will need
an air emission control device to control
particulate emissions from the large de-
entrainment cyclone. Particulates in
the air exhaust to the atnosphece frora
this cyclone averaged 0.57 g/Wm (0.25
grains/ft ). Also, tnis exhaust air
contained total counts of bacteria and
viruses that were much higher than those
found in suburban ambient air.
Water -A small quantity of washdown
water from the paved area around the
plant is the only water emissions, and it
poses no pollution problem.
Noise--A sound survey of tiie plant
revealed several locations with noise
levels above 90 dBA, the maximum allow-
able for continuous 8-hr. exposure. But
since no worker is present at these loca-
tions for 8-hr. or more, no noise
exposure problem exists.
Leachate—Analysis was made of lab-
oratory-produced leachate from plant prod-
ucts that might be landfilled (RDF and
magnetic belt rejects). Results Indicated
that groundwater contamination could re-
sult if the dilution rate were not high
enough.
Power Plant Evaluations
Process Descrlption—
The processed RDF is unloaded at the
power plant from the transport trucks into
a receiving bin, from which it is conveyed
pneumatically to a surge bin. A pneumatic
feeder system conveys the RDF from the
surge bin through four separate pipelines
directly to the boiler. Unit 1, a corner-
fired pulverized coa' suspension boiler
wi.th e nominal generating rate of 125 Hw,
was the boiler used for this test program.
Plant Operations--
Ooerations at the Keramec power plant
using RDF as a coal supplement extended
over several months and demonstrated that
burning 5% to 201 RDF as a supplementary
fuel in a coal-fired boiler is a viable
concept. Shutdowns occurred for routine
mocification and maintenance, and many
short-term shutdowns or reduction in RDF
firing rate resulted from problems with
the pneumatic conveying lines and block-
ages of the discharge chutes from the
Atlas storage bin (surge bin). No major
equipment problems were encountered, how-
ever, and the burning of RDF had no dis-
cernible effect on boiler erosion/corro-
sion.
Leaks in the pneumatic conveying
lines to the boiler were a frequent prob-
lem. The erosion of these lines was
caused by the abrasive materials in the
RDF, which were initially present at high
levels because no air classifier was used
to remove metals and glass. But even
after an air classifier was added, some
metal and glass fragments remained in the
RDF, and erosion of the carbon steel pneu-
matic pipelines continued to be a problem.
Performance of the electrostatic pre-
cipitator decreased with increasing boiler
load. Above 120 Mw. the burning of coal
plus RDF did decrease ESP efficiency.
Note, however, that the boiler was oper-
ating in excess of design capacity above
120 to 125 Mw.
Costs--
Cost estimates for firing of RDF do
not Include any expense for purchase of
RDF from the city or any credit for the
coal saved by using the RDF. Costs for
the receiving building, which was owned
by the City of St. Louis, were included
as part of the operating cost for the
iiower plant. All other equipment was
owned and operated by Union Electric.
147
-------
Capital cost of the facilities at
the Meramec plant was $945,640. Of this
cost, $578,097 represents Union Elect-
ee's initial inxestiMent, and $367,643
represents the City's cost for the re-
ceiving building and associated equip-
ment.
Over ar< 8-month period (October 1974
to May 197c;), ope?-ating and maintenance
costs averaged $9.39/Mg ($8.2S>/ton) of
RDF. These costs ranged from $5.67/Mg
to il7.70/Mg ($5,14 to $16.05/ton) and
did rot include amortisation of equip-
ment. These costs are probably not
representative for such plants, however,
because the system usually operated
below design capacity and maintenance
costs were high because of the need
for frequent repair and replacement of
pneumatic conveying lines. Future well
designed plants should be able to oper-
ate rare economically than the Meramec
plant.
Environmental Assessments-
Potential emissions associated with
combined firing of coal and RDF include
(1) boiler bottom ash and fly ash from
the ESP, (2) boiler sluice water and ash
pond effluent, and (3) air emissions
(boO. particulate and gaseous) from the
boiler stack. To assess the potential
environmental impacts of these emissions,
comparisons were made of test results
when firing Orient 6 coal only and Orient
6 coal mixed with 5. to 101 RDF.
Bottom ;nd fly ash--A sevenfold in-
crease in boTler bottom ash occurred with
the burring of coal plus RDF. This in-
crease was accompanied by higher concen-
trations of Cu, Pb, Na, Zn, ?nd Cr, and
decrease Al, Fe, Li, and S. Landfilling
of bottom ash from combined firing could
create water pollution problems, but it
was not possible to assess them relative
to those.of the coal-only bottom ash.
Fly ash from combined firing had a
higher heating value than the coal fly
ash (2.361 compared with 1,551 kJ/kg).
The RDF fly ash also had higher concen-
trations of Sb, As, Ba, Cr, Cu, Pb, Hg,
Zn, Br, and CI, whereas the coal-only
fly ash had a greater iroii content. These
differences are not sufficient to pose
greater disposal problems >itn tne RDF fljr
ash, but they might create Teaching prob-
lems during landfilling. The relative
impacts are difficult to assess, however.
Boiler sluice water and ash pond
effluent—Levels of trace constituents
7n boiler sluice water are approximately
equal for the two types of fuels, but
the coal plus RDF produces higher con-
centrations of total dissolved solids
(IDS).
Ash pond effluents from coal-only
firing meet proposed Missouri guidelines
with respect to biological oxygen demand
(BOD), dissolved oxygen (DO), and sus-
pended solids (SS), but effluents from
coal plus ROI" dc.not. The latter are
lower in sulfates but higher -n amnonia,
B, Ca, chemical oxygen demand(CUD), Fe,
Mn, and total organic solids. Measure-
ments of 48 other parameters shrwed no
significant differences.
Conventional gaseous emissions--
Except for chloride emissions, {which
increased by some 30»), comoined firing
of coal plus RDF did not produce major
changes in the emission of gaseous
pollutants comoared with the firing of
coal-only. Average carton monoxide
emissions were slightly higher for' the
RDF fuel (89 compared with 82 ppm). Ho
apparent change occurred in NO^ emis-
sions, and the levels measured complied
with current Federal and State regula-
tions. Hydrocarbon levels did not
appear to be higher with the RDF fuel,
SO, emissions wovld tend to be lower
with the burning of RDF, which has a
lower sulfur content than coal; but
the decrease would not be sufficient to
meet Federal regulations. A shift to
lower sulfur coal or use of an SOj con-
trol system would be required. Emissions
of particulates from the ESP at loadings .
above 100 Mw could rjt >neet Federal
requirements, regardi v.s of the fuel m*.
Potentially hazardous ."'jrticulate scis-
sions (8e, Cd, Cu, Pb. r-j, T1, Zn, and
F) did increase when oal and RUF were
fired. CI, Br, Pb, may ?xceed accept-
able limits even when ccj alone is
burned, and RDF only compounds the prob-
lem.
AMES, IOWA SOLID WASTE RECOVERY STSTEM
The city of Ames, Iowa has been
comnercially operating a system for
materials and energy recovery from
148
-------
municipal solid waste since Novembar
1975. Constructed and operated solely
with municipal funds, it was one of the
first municipal resource recovery systems
and represented a giant step in this
field. As port of their resource
recovery research and development program
EPA implemented a 3-year detailed evalu-
ation of the system.
Processing Plant Evaluations
Process Description—Both industrial
and private vehicles deliver refuse to
the tipping floor where 1t is reduced
to a nominal.(6-in.) size by the first
stage shredder. After passing the mag-
netic separator for ferrous removal, the
shredded refuse is further reduced to a
nominal size (1 1/2 in.). The shredded
material is then air classified into a
combustible fraction (RDF) and the heavy
rejects, which are further divided into
ferrous and nonferrous material.
Characterization of Reruse-Oerived Fuel —
The Ames RDF aooears to be of hiqher
quality than that produced at St. Louis,
even though the processing systems are
quite similiar. The reason is probably
that the Aiw»'. waste contained a large
amount of cownercial waste with a high
percentage of paper, whereas the St.
Louis RDF was produced strictly from
residential refuse with a high proportion
of food waste and moisture.
The Anies RDF heat value averaqed
5,700 BTU/lb, the average moisture con-
tent was 22 percent, and the ash content
averaged 17 percent.
Processing Plant Performance—
The plant operated.regularly except
for * shredder bearing failure and two
fires at the processing plant. Average
production rate was 50 tons/hr. The air
classifier was a major maintenance area.
The second stage shredder wai discovered
to use nearly twice the electric power
required for the first stage shredder.
A major disappointment was the aluminum
recovery system, which produced only
minor amounts of marketable aluminum
scrap during 1976.
Costs -
lhe net cost of operating the refuse-
processing plant (net cost of refuse dis-
posal) for 1976 was $18.90/Hg of raw
refuse received. This total represents
the cost after credits are given for the
ROF, dump fees, and recovered metals.
Improvements in net cost can be achieved
by reducing operating expenses and in-
creasing the volume of raw refuse re-
ceived.
Power Plant Evaluations
Process Description--
Originally, PDF *as to be used as
a supplementary fuel vith an Iowa-Wyo-
ming coal mixture in « suspension-fired
steam generator. The supplementary
burning of the RDF in the stoker boil;rs
was to occur during snutdown of the.
pulverized coal unit. However, attempts
to fire RDF in the suspension system
revealed that continuous burning was
prevented by high dropout of unburned
wood, cardboard,, and large paper.
RDF is now burned in the Stoker-fired
boilers at an average rate of (4.5 to 5
tons/hr), or 50* input on the basis of
heat energy input.
Boiler Description—
Tiie two stoker-fired boilers used
at the Ames power plant were installed
in the 1950's and use cyclone collectors
(multicyclones) for particulates remov-
al from the exhaust gas to the atmo-
sphere. Both are traveling grate
spreader stokers. RDF is fed into the
boiler by a pneumatic conveying system.
Environmental Emissions—
Particulate emissions showed no
clear trends regarding the function of
RDF heat input, either before the par-
ticulate collector or in the stack par-
ticulate emissions to the atmosphere.
NO and SO emissions both tended to
delSryje with increased percentages of
RDF. Chloride emissions increased with
the percentage of RDF for all boiler
loads. The substantially higher.levels
of chloride emissions for coal plus RDF
appear to be a function of the chlorine
in the RDF. Formaldehyde, cyanide,, and
phosphate emissions were quite variable,
with no clear trends based on the per-
cent of RDF burned. No significant
hydrocarbon emissions in the C-j to C^
149
-------
raitge, were found, and many of the
heavy organic compounds in the stack
emissions were below the labora-
tory detection level. Most of the or-
qanics found wore in the stack gases-and
not in particulate form.
Boiler Perfomarvce--
RDF-combined with coal was suc-
cessfully fired In the stoker boilers
with some difficulty but wir.h no major
problems. The maximum RDF firing rate
was 50% of the heat input to -he boiler.
"Ine burning of RDF had no significant
direct effect on the measured boiler
thermal efficiency. The averaae heat
input leavinq as combustibles in the
ash was about 5° for both coal and RDF.
Secondary air (excess air) supplied was
increased by the RDF pneumatic feeders
and the additional overfire air required
to burn RDF. The increase in excess air
required to burn RDF reduced the boiler
thermal efficiency. The general con-
sensus arcong the boiler operators was
that more combustion air throuqh the
grate was necessary when f 1i9?F to
prevent slaqqinq and to maintain a
proper fire bed.
Ultimate foulinq occurred in the
super-heater section of one boiler.
Calculation of the fuel fouling index
correlates with this behavior. The most
significant influence is the hiqher
sodium content of RDF,which has a detri-
mental effect on the foulinq index. Soot
blowers would reduce this foulino, as
might an.alternate method of RDF injec-
tion.
At i"Ost boiler loads, bottom ash
tended tc increase somewhat and fly ash
tended to decrease with increasing per-
cent RDF.-
Ash fusion temperatures of RDF are
typically 60" to 110° C lower than for
coal, but no specific correlation of
boiler performance to ash fusion temper-
atures has been determined.
Boiler Corrosion Stud'es--
Examinations by metallography,
microbe, and chemical analysis were con-
ducted of waterwaJl tubes, superheater
tubes and their scales, and deposits.
Results show that during exposure to
firing of a mixture of 50t coal and 50S
solid waste for a period of 1,018 hr.,
the corrosion of the waterwall tubes
was virtually zero. Corrosion of super-
heater tubes, if any, did not exceed
approximately 0.025 mm. The scale on
the superheater tube contained up to
12"J to 18* sulfur. Whether this amount
which is present along with other ele-
ments constitutes a potential for cat-
astrophic corrosion is not known.
Chlorine in both weterwall and
superheater tube scales is present in
amounts below the limit of detection
of the analytical method used (600 ppm),
and it is thought to constitute a sig-
nificant factor in tube corrosion.
FIRING OF COAL:d-RDF BLENDS IN SPREADER
STOKER-FIRED BOILERS AT HAGERSTOWN,
MARYLAND
Introduction
After experience was gained with
the cofirtng of RDF and coal, little
information was available on the pro-
duction <".nd burning of d-RDf. EPA
therefore implemented parallel programs
to (1) determine to economics of pre-
paring d-RDF. and (2) assess the tech-
nical and environmental implications of
using d-RDF as a coal substitute. The
first part of the program involved the
three projects described earlier: Grants
to flCRR and the University of California
tQ study the production and use of d-RDF,
and a contract with Teledyne National to
produce 2000 tons of d-ROF and provide
information on production. The second
part of the nrogram involved a contract
with Systems Technoloqy Corporation
(SYSTECVP to conduct a comprehensive
technical and environmental test program
for coal:d-RDF coffring. The first
phase of The SYSTECH program was a
feasibility study with demonstration
tests carried out at the Maryland
Correctional Institute for Hen (MCI),
at Hagerstown, Maryland. The second
phase was a longer-term demonstration
test in an industrial sized boiler at
the General Electric plant 1n Erie,
Pennsylvania.
The MCI plant of Hagerstown had
three small institutional heatina
boilers that produced 3.1, 7.6, and
9.9 kg/sec. (25,000, 60,000, and 75,000
130
-------
lb/hr) of 1034 kPa (150 psi) saturated
steam.
The test was designed to combust
258.5 Mg (285 tons) of d-RDF during 236
hours of firing various blend ratios of
coal:d-RDF. These tests were conducted
in a series of burns with coal:d-RDF
volume ratios of 1:1, 1:2, and 0:1, and
with test durations ranging from 20 min.
to 132 hr. The cofiring tests were pre-
ceded and followed by a coal only test
with duplicate conditions. Because of
plant steam demand, the tests were de-
signed to ensure that d-ROF could be
safely burned without jeopardizing the
boiler's ability to meet the stean
demand.
d-RDF characterization
The (1/2-to 3/4-in.) pellets had
an average bulk density of 425 kg/m
(26.5 lb/ft ) and ranged from 400 to 465
ka/ra (25 to 29 lb/ft3). The material
density for intact pellets ranged from
1.22 to 1.34 * 10J kg/m (76 to 64 lb/
ft ), and that for,deteriorated pellets
averaged 0.98 x iO kg/m (6) lb/ft ).
The as received properties were 12.10
to 15.12 MJ/kg (5200 to 6500 BTU/lb).
201 tO-29* ash, 91 to 10% fixed carbon,
I21 to 131 moisture, 505 to 57S vo?-
atiles, and IWto 1152°C (2088° to
2105°F} hemispheric reducing fusion
temperatures. NCRP., who produced and
supplied the pellets, projected that
further processing of the shredded
refuse to remove glass and other incrts
could produce a pellet with a heat con-
tent of 13.1 MJ/kg (6200 BTU/lb) and
an ash content of 10% to 12%.
Haterial Handling
Throughout the field testing, 259
Hg (285 tons) of d-ROF was received,
stored, and conveyed to the boilers
without major difficulty or malfunction
of the fuel handling system. Difficul-
ties wtre limited to dusting and pellet
hang up in the feed hoppers.
Pellet Storage--
During successive periods, the
pellets were stored in (20-yd ) open
containers, in a warehouse (uncovered),
and on an outdoor concrete slab (tar-
paulin-covered).
Pellets stored in the open containers
tended to steam when received during the
winter, and eventually they froze in a
solid mass. But minimal rodding broke up
the frozen pellets, and subsequent han-
dling further restored the ind'vidual
pellet integrity without significant
degradation to the pellet.
Storage in ar. unheated warehouse
was the most effective method for main-
taining pellet integrity over extended
storage periods (2 months). Mild odors
and some fungus growth occurred, but
temperature Increases resulting from
composting were negligible since pile
depth was limited to 1.8 m (6-ft). Pile
temperature stabilized at 60® C (140°F).
The tarpaulin-covered pellets stored
1n a 1.8-m (6-ft) piles on an open con-
crete slab accumulated moisture under the
cover and caused pellets at the top to
deteriorate and cake. Swelling and
roughed edges also occurred in some
pellets because of water infiltration.
Pellet Feeding-
Coal and d-ROF were volumetrlcally
blended in various ratios by separately
feeding coal and pellets firm two hoppers
to a comnon bucket elevator, which sub-
sequently conveyed the mixture to a weigh
lorry. This feed system worked well
generally, but as the fines Increased from
excessive handling, the pellets would not
flow from feed hoppers without rodding.
The considerable dusting caused by these
fines throughout the plant was subsequently
controlled by installing a steam jet at
the conveyor transfer point.
Boiler Performance
Feeder Performance—
The (1/2- by 3/4-in.) pellets gen-
erally handled and fed well, with the
larger pellets traveling to the rear of
the grate and the fines falling close to
the spreader. During the initial com-
bustion tests with 100 percent pellets,
the spreader had to be adjusted to de-
crease the pellet trajectory by approx-
imately 0.3 in (12 in.). In addition,
because of volumetric fcedinq capacity
limitations, the maximum load that the
boiler could carry was 24,500 kg/hr
(54,000 lb/hr), or 70i of rating.
Ill
-------
Combustion of d-RDF—
The combustion of the various blends
of coal and d-RDF was generally as good
as that of coal only. But the length,
intensity, and volume of the fireball
grew as the proportion of of the d-RDF.
was increased. Flame temperature (1.5 m)
above the center of the grate also in-
creased from 1200°C (2192°F) for IOCS coal
firing to 1240°C (2264°F) for 100% d-ROF.
When the 1:1 blend and 100% d-RDF
were test fired, the fireball was kept
well away from the rear wall of the fur-
nace ~>y adjusting the overfire air. Once
these iets wee adjusted for minimum smoke
and ma. 
-------
the ash hopper revealed that the ash had
a taffy-like consistency. Under similar
conditions, when coal only was fired, the
bottom ash was much easier to break up
by rodding.
Tl;® bottom ash removal system mal-
functioned only during lOOi pellet firing.
The bottom ash was so fine that it would
not de-entrain properly in the cyclone. The
particles, which had been wetted by the
steam in the vacuum ejector, passed though
the cyclone and eventually plugged the
ejector.
Dust. Collector ftsh--As a greater
proportion of d-RDF was substituted for
coal, the fly ash particles became finer.
The size of the particles in the dust
collector ranged from 200 um for 100? coal
firing to 90 um (sizes at the 50th percen-
tile) for 100% pellet firing.. Also, the
carbon content of the fly ash decreased
significantly.with increasing d-RDF sub-
stitution.
Mass and Energy Balance—
Mass Balance—The mass balance indi-
cated that an unusually large amount of
tne fuel asn nad accumulated in tne col-
lectors. Subsequent analysis of the col-
lector fly ash revealed that the high
collector ash weights were due to the
presence of 50% to 70% carbcn in the col-
lector ash. Also, since 90% of the par-
ticles existing from the boiler were great-
er than 50 um in diameter, these large
particles were removed by the cyclone.
The carbon content of the stack fly ash
(not captured by the cyclone) was 301
to 40%. Analysis of the stick fly ash as
a function of blend revealed that its
carbon content decreased as the d-RDF sub-
stitution Increased.
Efficiencies--During the testing the
boiler efficiencies were extremely low
(551 to 60") primarily because of the low
boiler loads'(less than 30% of rating),
high excess a.ir (801 to 115%), and extreme-
ly high losses of combustibles in the
refuse (up to 25%). Results Indicated
that the coal-only and blend firing effi-
ciencies had no discernable differences.
However, this observation may be unique
to the boiler' instaliatinn ai: MCI, since
the large amount of jnburned combustibles
removed by the collectors is certainly
an anomally to expected boiler performance:
Environmental Performance
Data Normalization—Since the co-firing
tests spanned a 6-month period, the
propertius of the coal and d-RDF bu.*u.d
in the successive tests varied consider-
ably. To eliminate the effects of th.«se
variables, all the emissions data were
corrected to 50 percent excess air and
then normalized to a reference coal and
d-ROF composition. All the co-firing
emissions data were then statistically
compared with a coal-only baselines plot
of emissions concentration versus boiler
load. If the co-firing emissions data
fell outside the 90. confidence limits
for the coal-only emissions data, they
were considered to be "ignificantly
different.
Particulate Emissions—
Mass Concentration—The particulate
mass concentration in the 1:1 and 1:2
blend firings was slight'.y less than in
the coal-only firing, but the reductions
were not significant at the 90% confidence
level. The mass flux at a 40% boiler load
for 1:1 end 1:2 blend firinqs averaged
0.45 g/scm corrected to 12% CO^. lhe coal,
fired during these tests was a nominal
size of (1 1/4 x 1/4 in.,), with a max-
imum of 30% passing through a nominal
(1/4-ln.) screen.
Particulate Size—As more d-ROF was
substituted for coaTT the particulate
diameter decreased. In the May test, the
diameters for tne coal-only firings were
3 um, and those for the d-RDF-only firings
were 0.8 um (at the 50th percentile
point).
Particulate Resistivity—Because of
the unusually high carbon content in the
fly ash during the coal-only firing, tbe
resistivity was qenerally less than 10
ohm-cm. As d-RDF was substituted for
coal, the carbon burnout in the fly ash
improved,,and the resistivity Increased
to I x 10 ohm-cm for the 1:1 blend
firing.
Opacity--As d-RDF was substituted
for coa1, the overall opacity of the
piuiw, reduced significantly. At 40%
boiler load, the opacity for coal-only
firing was 16% (based on a 1.22-m-(4 ft)
diameter stack). At the same boiler load
and excess air, the opacity was only 10%
for d-ROF-only firing.
151
-------
Gaseous Emissions--
S02--Since the d-RDF had a sulfur
content of 0.4%, the SO, amissions de-
creased with increasing d-RDF substitution.
The decrease ws particularly significant
for the 1:2 coal:d-RDF blend and the 100%
d-RDF firings. At 40? boiler load and the
san>e excess air levels, the 50, dropped
from 1300 ppm for coal-only fifing to 250
pprt for d-RDF-only firing. This reduction
in SO, follows exactly the reduction in
sulfui* content of the fuel.
NO --No significant changes occurred
in NO e--As d-RCF was substituted
for roaT", the chlorh- in the e-issions
increased frora 60 ppm for "oal-cnly firing
to 650 ppra for d-RDF-only firing. There
appeared to he no appreciable change in
chlorine concentrations as the load
cnanged fron.70" to SOi of design capacity.
Fluorine—fluorine conctntrations
fllco	with ip'-rM';inr d-?!^ sub-
stitution. hut the concentrations were
very low (e.g., 8 ppi:> for coal-only firing
and 12 ppro for d-RDF-only firing at a 40%
boiler load and constant excess air
cijndi tiins J.
Hydrocarbons--No significant changes
occurred in hydrocarbon emissions, when
substituting d-RDF for coal. At a 401
boiler load, the toal hyorocarbons ranged
from 10 to 25 ppm. As the boiler load
increased, the hydrocarbon concentrations
decreased significantly. rlm reduction
fs probably attributable to the improved
cofrtusticr. conditions at higher boiler
loads.
Trace Organic and Inorganic Eraissions--
Organic Enissions--The overall emis-
sions of polycyclic compounds for coal-
only and blend firings were well below
the threshold lintls proposed by the Na-
tional Academy of Science., Typical mea-
sured values were 54J oq/m for anthracene/
phenanthrene, 100 ng/m for methyl
anthracene, and 137 ng/m for fluoranthene
(all at 1:1 blend firing).
Inorganic Ewissions--Fly ash analysis
for trace metals revealed that relative
to coal-only firing, the blend firing
enriched some metal but reduced others.
For examples, when a blend of 1:2 coal:
d-RDF was fired, compared the jrount of
lead.in the stack particulates was 8217
ug/m , with 230 i.n/in for coal-only
firing. Though d-RDF was the main con-
tributor of Br, tfn, Pb, and Sb, coal
was the primary source of As, Ni, and V.
Several elements, particularly As,
Ga, Na, and Sb, tended to concentrate
in small particles. In addition, as the
d-RDF substitution increased. Doth the
solubility of the fly ash and the quantity
of small particulates in the respiratory
range increased. Consequently, each of
these effects pose potential Hazards
from (1) respiration of heavy metals
associated witn aerosols, and f?) leach-
ing of hiqh- levels of heavy neuls in
landfills.
Conclusions
While the test was limited to firing
af reduced boiler loads, the preliminary
results fron these field tests indicate
that coal and d-RDF can be co-fired at
volu-vtr'tr c"! 1 ¦	"atios up to 1:2
with only minor adjustments to the boiler
and fuel handling systems. Subsequent
testinq should address the long-term
effects of corrosion and erosion on
bailer tubes.
FIR IMG OF COAL:d-RDF BLENDS !N AN INDUS-
TRIAL SPREADER STOKER BOILER IN ERIE,
PENNSYLVANIA
Introduction
Earlier field tests (Hagerstown and
others] involving the co-firing of coal
and d-RDF blends were typically of short
duration and *er« performed under less
then desirable boiler operating condi-
tions and specifications. The objectives
of demonstration tests conducted under
£PA contract to StSTECH at the General
Electric plant at Erie, Pennsylvania,
were to provide longer-term cofiring
t«sts in a boiler representative of those
used throughout industry. Sufficient
testing was to be done to determine (1)
wtiether or not d-ROF has any detriments
effects on the boiler system or its
performance, and (2) whether it can be
burnad within existing environmental
constraints.
154
-------
The test program was designed to de-
termine the technical and environmental
effects of cofiring d-RDF ar.d coal in an
industrial power plant. To determine
these effects, a (150,000 lb/hr) steam
generator equipped with a chain grate
spreader stoker was instrumented to
rronitor efficiency and emissions. The
denonstration consisted of four distinct
tests: (1) an initial coal baseline
test, (2) cofiring of coal and d-RDF using
pellets from NCRR, (3) cofiring of coal
and d-RDF using pellets from leledyne
National, and (4) a final coal baseline
test. Portions of the last two tests
were conducted without fly ash revnjection.
Throughout each test, fuel handling, boiler
performance, and the stack emissions were
monitored and evaluated.
Fuel Characteristics
Five different coals were used during
the blend firing and coal baseline tests.
Their sulfur contents ranged from 1.7"
to 6.8«, and ash contents ranged from 9.5»
to 18.2% (dry weight basis). Two differ-
ent d-RDF' s were tested. Both were formed
as (1/2-in) diameter cylindrical pellets.
The pellet produced by NCRR contained over
30* ash end had a heating value of 6,755
BTli/lb on a dry weight basis. Tfce pellet
produced by Teledyne National contained
14 percent ash and had a dry weight heat-
ing value of 8,123 BTU/lb. Moisture con-
tents of the d-RDF ranged from 14% to 34£,
and bulk,densities ranged between 30 and
35 lb/ft .
To ensure an adequate supply of
pelletlzed d-RDF, EPA contracted with
Teledyne National and NCRR to produce and
ship pellets by truck to Erie. Ultimate-
ly, 1702 tons of pellets (1256 tons by
Teledyne and 446 tons from f«CRR) w the
combustion of d-RDF. Calculations show
that cochu;tion of 100'S d-RDF requires
lcsc forced draft fan capacity, Er. more
induced draft fan capacity, and 3 1/2
times the volumetric fuel feed capacity
of coal. Grate speed was increased by
about 25" when combusting a 1:2 (coal:
d-RDF) volumetric blend.
Material Handling
The d-RDF was stored up to 6 months
at Erie, Pennsylvania in an open coal,
yard through winter and spring weather.
The piles of d-RDF formed a protective
crust (6 to 6-1n) thick. Wnile in
storage, the pellets increased in mois-
ture and fines content. Also, the aged
pellets expanded and formed serrated
edges that subsequently created handling
problems. While being conveyed to the
bunker, the d-RDF blended thoroughly with
the coal. But the low bulk density,
high elasticity, and fibrous shape of
the deteriorating pellets required con-
stant rodding for them to flow out of
the conveyor feed hopper. In the bunk-
er, the coal/d-RDF blend would "rat.
hole" and demonstrate angles of repose
in excess of 90°. Bunker vibrators and
air blasters did not eliminate the need
ISS
-------
to manually rod the fuel blend Into the
nonsegrating distribution chutes feeding
out of the bunker.
Compared with coal, the d-RDF blends
required more frequent ash removal because
o.c the increased ash content and decreosed
heating value of the fuel. Except for the
manual removal of the infrequent clinkers,
no ash handling problems were noted with
the pneumatic ash handling system.
Environmental
The combustion of d-RDF blends exhib-
ited the same range of particulate emis-
sions as the burninq coal only. Though
NCRS blends caused decreased particulate
size, firing the Teledyne blends resulted
in increased particulate size wnen com-
pared with coal only. Tlie ESP performance
was unchanged by the substitution of d-
ROF for coal. However. dgRDF raised the
f1y ash resistivity to 10 ohm-cm resis-
tivity of the coal-only fly ash. Pb emis-
sions increased by a.factor of six. Cd,
Zn, and Cr, emissions increased 50% to
1001 when firing d-RDF.
The substitution of d-RDF for coal
had no significant effect on NO , CO, or
hydrocarbon emissions. 3ut as exnected,
ci-ROF caused a 30» to 50« decrease in.SO
emissions and a 250% increase in chlorid$
emissions.
Fly ash reiniection from the r.iulticy-
dones was found to have no effect jn the
mass rate or size distribution of pjrticu-
lates at the ESP inlet when d-RDF blends
were fired. Fly ash reinjection was
found to decrease particulate size when
firing coal only.
Conclusions
Coal/d-RDF blends can bs combusted
in an existing spreader boiler at tne
same excess air levels as coal only. No
derating of the boiler was caused by the
d-RDF Substitution, and the decrease in
boiler efficiency was minimal. With a
low-ash d-RDF, no operational problems
(clinhering) were experienced. No ad-
verse impacts were measured on the plant
emissions as a result of firing d-RDF.
Materials handling problems were experi-
enced with the existing coal nandling
equipment, but such problem;, had not been
observed in earlier tests. They may
therefore be due only tc the aged and
weathered condition of the d-ROF.
Further study is warranted on '.he
long-term corrosion and erosion effects
of d-ROF combustion.
156
-------
SELECTIVE ENHANCEMENT OF RDF FUELS
Norman L. Uecht, Donovan S. Duvall
Anwar A. Ghazee, B. Lawrence Fox
University of Davton Research Institute .
Dayton, Ohio . 45469
ABSTRACT
Conversion of the orqanic fraction of municioal solid waste to "a pow-
dered fuel offers a number of advantages for improving both the quality and
marketability of the product. Cellulose embrittlement processes have been
developed to convert the organic fraction of 1SW to a powder. This project
concentrated on improving the embrittlement Process developed in an earlier
study and characterizino the properties of the powdered fuel obtained.
Over 363 Kg. (R00 lb) of Dowdered fuel were processed for characterization
studies. The Processing procedures used for convertinq the RDF frorti the
Americoloqy plant in Milwaukee is described. In the initial phase of this
project, the Physical, chemical and thermal properties of the oowdered fuel
were evaluated. Particle size distribution, morphology and density were
measured-. Proximate, ultimate and ash analysis were determined and TGA,
DTA and heat content were established. In the second phase of this project,
•the combustion c'ndL.icteristi.es of thu powdei alone and Fi>:ed with powdered
coal and oil were extensively evaluated. This work was conducted in ful-
fillment of Hrant No. R-806535010 bv the University of Davton Research In-
stitute under the sponsorship of the 'J.S. Environmental Protection Agency.
INTRODUCTION
The primary purpose, of this:
study w^s -to develop effective pro-
cesses to convert, the organic frac-
tion in MSW to a oowder. The Dro-
cesses l?or converting the organic
fractiort in MSW to a powdered mate-
rial offer a nun-iber of advantaaes
for improving both the qualitv an.i
marketability of refuse derived pro-
ducts. In the powdered form the re-
fuse is-a more effective fuel, caa
be used as a filler material in plas-
tic and -rubber products and can be
used as a feedstock in several bio-
mass conversion Processes (acid hy-
drolysis, oyrclysis, etc.).
This study concentrated on im-
proving the embrittlement process de-
veloped in an earlier studyd' and
better characterizing the properties
of the'powder obtained from embrit-
tling RDF from the Americoloqy Plant
in Milwaukee. Over 363 kg (800 lb)
of RDF were processed and the combus-
tion characteristics of RDF powder,
nowder/coal and powder/oil mixtures
were extensively investigated.
POKOER PREPARATION'
A major emphasis of this project
was the nroduction of large Quanti-
ties of powder from RDF for combus-
tion analysis. About 453.6 kg
(1000- lb) of RDF were obtained from
the Americology Resource Recovery
Plant in Milwaukee, Wisconsin to be
used as the feedstock for the embrit-
tlemeit treatments. During the
course of the project some 363 kg
(800 1 -a) of RDF were processed in the
UDRI Pilot Reactor Unit for the pro-
duction of 252 kg (555 lb) of minus
149y (100 mesh) powder for testing
and evaluation.
Batches of 2.3 kg (5 lb) of the
137
-------
RDF were embrittled in the reactor
unit at a treatment temperature of
149°C (300°F) using a 72 percent hy-
drogen chloride, 28 percent nitrogen
react?.nt gas mixture for treatment
times of 3-5 min. The embrittled RDF
was ball milled for two hours and
then screened in a sieve stack for
one hour. A tabulation of the oow-
der pr»eoaration conditions is pre-
sented in Table 1." The reactor unit
built by the University of Dayton
Research institute (UDRI) is shown
in Figures 1 and 2.
POWDER CHARACTERIZATION
Samples of the powder Produced
were obtained throughout the proqram
for chemical, physical, and thermal
analysis. Ir. additionsamples were
sent to the Commercial Testing and
Engineering Company and Battelle
Columbus Laboratory for analysis of
selected chemical and thermal pro-
perties.
•Physical and Chemical Analysis
Particle size distribution,
moisture content, morohology and den-
sity of the powder s.amDles were mea-
sured using a RoTap with quartered
powder samples and the results are
shown in Table 2. The data present-
ed include the moisture content mea-
sured for each size fraction. All of
the data reported reoresent the aver-
age values calculated from five or
more measurements.
Using a scanning electron micro-
scope, the mcrnhology of the minus
149y and 74vj (100 and 200 mesh) pow-
der were studied. Representative
electron micrographs are shown in
Figures 3 and 4.
The density of the loose powder,
lightly compacted, and compressed
was also measured. The loose density
of the minus 149u (-100 mesh) RDF
powder was measured by filling a 50
ml graduated glass cvlinder with pow-
der to a defined volume and then
weighing the cylinder contents. The
cylinder was then gently tapped to
remove air gaps, more powder was add-
ed to maintain the same volume ana
the new weight determined. The con-,
pressed density, of the powder was de-
termined by pressing 1.5 g (0.053 oz)
of oowder in a cylindrical die with
a hydraulic press to 1.55 x 10' kg/m*
(220 Dsig). Using this procedure,
several 1.27 cm x 1.27 cm (0.5 in.
x 0.5 in.) cylindrical pellets were
made- and the average density of the
pellets were measured. The results
of the density measurements are re-
oorted in Table 3. The data show
that a 400 Dercent increase in aen^
sity was obtained when loosely Backed
Dowder wf.s compressed to a pellet.
Chemical prooerties of the pow-
der were n-easured at both the UDRI
and the Coinmercial Testing and Engi--
neering Comoany, Chicago, Illinois.
Proximate and ultimate analysis, min-.
eral analysis, and fusion temperature
TABLE 1. POWDER PROCESSING CONDITIONS
Quantity of RDF
orocessed
Processing
temoerature
Processing time
HC1 flow rate
Nj flow rate
HC1 adsorbed
by RDF
Ball mill time
Pcreening time
2.3 kg (5 lb)
149°C (300°F)
3-5 min
439 cms/sec (0.93 ftl/min)
170 cm3/sec (0.36 ft3/min)
2% by wt
2 hr
1 hr
138
-------
SOUNOOVALVl—v y
n, now tcrccr
hci fto* ytitfl J
I J— pfttssuftt pegumqb
1 V IP"
n»w	I/——— ——¦"¦
IftOCH* J! ^
I L| ,5»m
i_	rij -l jr
*LVl
50liNO*0
lALVl
POL*
L1htcE»C
TURMC
N, Kt*Tl«

— MCI
toaidco wiaf .TtrLo*r"
5\«io_ rua#*.
1_WM4 KtAUfi
NoOH
SOLUTION -
y:*ueet» num
- wurot*
hO«


//!V\
sampie
PORT
Figure 1. Process flow diagram for pilot reactor.
of the ash in the powder were measur-
ed by the Commercial Testing and En-
gineering Company and the results
are presented in Tables '4 through 7.
A number o-f ash analyses were also
.conducted .by the UDRI on several dif-
ferent size fractions of powder
using ASTM'standard D271. The're-
sults obtained are presented in Ta- ¦
ble 8. As shown in Table 8 almost
40 percent' of the minus 345u (-45
mesh) powder from the ball mill is
non'combustible. The higher ash con-
tent found in the minus 74u (-200
mesh)¦fraction indicates that the
finer fraction of the powder contains
a greater 'fraction of inert materials
believed to be predominantly glass
particles.
Thermal Analysis
The heat content of the powder
was measured at the UDRI and the
Commercial Testing and Engineering
Company. At the UDRI the heat con-
tent was measured using 1 gm (0.035
oz) samples in a Parr Adiabatic Bomb
Calorimeter following ASTM D2015.
The results obtained are presented
in Table 9.
Thermogravimetric (TGA) and Dif-
ferential Thermal Analysis (DTA) me-
thods were used to further evaluate
the thermal properties of the RDF
powder. These studies were conducted
by Pattelle Columbus Laboratory. A
summary of the data from the thermo-
grams is presented in Table 10 along
with comparative data for coal.
POWDfR/OIL SLURRIES
Suspension of the RDF powder in
a fu£l oil offers, a number of advan-
tages for extending the commercial
potieritial of the RDF oowd«r. Powder/
oil suspensions should be easier to
transport and less subject to explo-
sions. In addition, suspension of
the powder in oil should.facilitate
introduction of the- oowder into con-
ventional oil fired boiler and fur-
nace units. During the course of
this program two different techniques
for preparing powder/oil slurries
were investigated and selected pro-
perties of the slurries prepared were
evaluated. The two methods were:
159
-------
(a)	qradual dispersal of minus
149u (-100 mesh) powder in diesel
oil by conventional blending proce-
dures ; and
(b)	coiTibined ball milling and
mixing of the treated RDF with die-
sel oil.
Unfortunately, the slurries
prepared by both methods did not dis-
nlay lonq term stability. Within B
hours the nowder was found to have
completely settled out of the diesel
oil. Dispersion agents were inves-
tigated for enhanced slurry stabili-
zation. Slurries of 2 3 and 37 weight
percent solids were prepared to eval-
uate settlinq characteristics, den-
sity, and viscosity.
Several 150 ml (40 x 10-^ gal)
Figure 2. Pilot reactor.
160
-------
Figure 3. 149u (100 mesh).material at 300X.
Figure 4. 74u (200 mesh) material at 300X.
161
-------
TABLE ?. PARTICLE SIZE DISTRIBUTION AND MOISTURE
CONTENT OF POWDERED RDF
Particle Size Quantity Moisture
u
(mesh) %
ft
+354
(+45) 18.1
5.86
-354
+ 149 (-45+100) 12.5
6.54
-149
+ 74 (-100+200) 31.1
5.91
-74
(-200) 38.3
5.03

100

TABLE 3.
DENSITY OF MINUS 149u (-
100 Mesh) POWDER*
Condition kg/m3
(lb'ft3)
Loosely packed 271
(16.9)
Compact 418
(26.1)
Compressed 1096
(68.4)

TABLE 4. PROXIMATE
ANALYSIS"*"
Property
Powdered RDF !%)
Midwestern Coal (%)

As Recv'd Dry Basis
As Recv'd Dry Basis
Moisture
7.3 3 xxxxx
11.07 xxxxx
Ash
24.91 26.98
12.26 13.79
Volati Le
51.83 55.93
34.96 39.31
Fixed Carbon
15.93 17.19
41.71 46.90

100.00 100.00
100.CO 100.00
TABLE 5. ULTIMATE ANALYSIS"1"
Property	Powdered RDF (%)	Midwestern Coal (%)

As Recv'd
Dry Basis
As Recv'd
Dry Basis
Moisture
7. 33
xxxxx
11. Q7
xxxxx
Carbon
32.70
35.29
59.49
66.89
hydrogen
4.30
4.64
4.15
4.67
Nitrogen
0.68
0.73
0.92
1.04
Chlorine
3.11
3.36
0.16
0.18
Sulfur
0.20
0.22
4. 39
4.94
Ash
24.91
26.88
12.26
13.79
Oxygen
26.77
28.88.
7.56
8.49

100.00
100.00,
100.00
100.00
+Data from Commercial Testing and Engineering Company
162
-------
TABLE P. FUSION TEMPERATURE OF ASH+
Property	Powdered RDF cC("F) Midwestern Coal °C(°?)

Reducing
Oxidizing
Heducinq
Oxidizing
Initial De-
formation
1065(1950)
1100(2010)
1093(2000)
1260(2300)
Softening Point
1218(2220)
1222(2230)
1182(2160)
1332(2430)
Hemispherical
1238(2260)
1250(2280)
1193(2180)
1343(2450)
Fluid, Tem-
perature
1318 (2400)
1360(2480)
1271 (2320)
1432 (2610)

TABLE
7. ASH ANALYSIS*

Powdered RDF	Midwestern CoaT
Mineral %
Weight, ignited
% Weight, Ignited
Silica, Si02
Alumina, A^Oj
Titania, Ti02
56. 40
9. 55
1 .10

47.52
17.87
0.78
Fernic, Oxide, f^Oj
Line, CaO
tfaanesia, MqO
Potassium Oxide, KjO
Sodium Oxide, Na^O
3.17
12.50
6. 30
1.50
6.88

20.13
5.75
1.02
0. 36
1.77
Sulfur Trioxide, SO,
Phos. Pent. P2°5
Undetermined
1.92
0. 30
0. 38

4.8
Silica Value
Base/Acid
T25o/p°iSGS
71.97
0.45
1291°C (2355°
F)
63.85
0.44
1299°C (2370°F)
TABLE 8. ASH DISTRIBUTION IN
RDF POWDER*
Particle Siz>
u (mesh)
Quantity
*
Ash
%

+ 354 (+45)
18.1
-

-354 + 149 (-45 ~
100) 12.5
35.1

-354 + 74 (-100 +
200) 31.1
37.5

-74 (-200)
38.3
43.6

"""Commercial Testing and
Engineering Company

*UDRI
161
-------
TABLE .9. HEAT CONTENT Of RDF POWDER*
Particle Size
u (mesh)	
-0.64cm (-1/4 in)
+354 - 0.64cm (+45 - 1/4 in)
+354 + 149 (-45 + 100)
-149 + 74 (-100 + 200)
-74 <-200)
Averaqe -354u (-45 Mesh)
(—
I Commercial Testinq &
engineering Co. ctata
¦•"measured at UDRI
slurry samples with 3? percent by
weight powder were prepared. The
slurries were prepared by gradually
mixing 50 q 11.9 M) of powder i.n
100 ml (26 x lb-3 gal) of dies-il oil
with a magnetic stirrer until uni-
form suspension was obtained. Slur-
ries of minus 149u (-100 mesh) and
"iir>us 7
-------
_ THERMAL
ANALYSIS OF POWDERED RIJF+


Thermogravimet.ric
Analysis
Property
Powder RDF
Coal
Ash %
44.4(e)
10.3
Temperature Range ("C)
130-380 (266"-716° F)
220-585 (428°-1085*F)
Maximum Rate of Weight Loss (mg/min)
37(81.6 x 10"6 lb/min)
17.5(38.6 x 10"6 lb/min)
Temperature at Maximum nate of Weight
Loss CC)
290 (554°F)
320 (6 0 8 0 F)

Differential Thermal
Analysis 'C'
Starting Exotherm CC)	222 (431°F)	233 (451°F)
Ignition Point CO	276 (529*F>	426 (799T)
End of ignition Exotherm	CC) 289 (552°f)	615 (11390F)
Secondary Exotherms ,CO	379-433 (714°- flll'F)			
Endotherm CC)	457 (855°F)	400-584 (752"- 1082°F)(d>
(a)	TGA performed with Cahn Electrobalance at 15°C/mir. (59"F/min) and air flow of
800 ml/min (.0035 gal/sec)
(b)	Temperature range over which most of wieglit Joss occurs
(c)	"DTA performed with Stone Model 202 at 15°C/min (59°F/inin) and dynamic gas flow of
0.0057 m3/hr (0.2 SCFH)
(d)	Endotherm range in presence of N2 gas. Three peaks in each range
(e)	RieBidue measured after combustion in the test burner
+Eattelle Columbu3 Laboratory
-------

TABLE 11
• PRELIMINARY
EVALUATION
OF OTL-PGWDGR
SLURRY


-«*¦

• .»»»»*- .x-«

» -x ,.,r«^


Part Ic !«¦
(mesh)
OiMnt ity ol iv>w-
»lr» in Slurry wt
%
l(y of
Slurry/ wl
1X10--4 qal)
P'ontity «>f ims-
j, » •: :r».j Arirp.t,'fn 1
(xio'' qaii
J'-oltlinq
1 T»r*c
hr
Set tling
(XIO*1 q-i!)
-149

150 (401
0
4
IflO
(76)
-149
(-1001
37
150 (40)
J.5 1.391
4

(G.fcl
"149
(-10U)

150 (401
1.5 (-J9I
20
iw
(?6f
-74
(-2D01
17
150 (401
0
1
ion

-74
(-30*1
M
150 (4QI
1.5 (-}¦»»
«
15
(1.91
-74
l-290t

lf.0 (-10)
1,5 1.191
;o
IftO
(?6 1
N'OTE * Above
screening was
conducted
in 2 50 mi (66 x
10~3 gal)
beaker.
TABLE 12. DFNSITY AND VISCOSITY OF OIL AND POWDER/OIL SLURRY
!At 24'C (70°f) i 10,336 kq/nr (14.7 osia)]
Density (c), q/cmJ (oz/in!.
Viscosity (u) ,• cp
Diesel Oil
0.84 (0.49)
10.b2
37v Powder/Oil Slurrv
1.01 (0.58)
32.00
NOTE: The Brookfield Viscometer Model 4RVT7 with number 1 spindle
was used for all viscosity (Measurements.
It should be noted that or.ly
preliminary slurry characterization
studies have been conducted, and
effective desiqn of transportation
and storaqe svstems for the powder/
oil slurry would require an extensive
investigation of the flow and possi-
ble corrosion characteristics of the
slurry.
INVESTIGATION OF THE BURNING CHARAC-
TERISTICS OF THE POWDERED RDF COAL
AND POWDER/OIL SLURRIES
In this cart r>f the study the
combustion characteristics of the
RDF powdor as a fuel was investi-
gated.. RDF powder, pawdc-red coal,*
^Illinois No. 6 Coal
and two RDF-coal mixtures were tired:
in Battelle's small pulverized coal
combostor units. The fuel feeder
and combustion unit used in this
studv are shown ir Fiqure 5. Fuel
is fed to the combustor by lifting
the snail particles from the solid
fuel :«rfuce with air jets and en-
training them in a 0.64 cm (0.25 in.)
O.D. tube, which leads tr> the burner
head, with this arrangement, the
Lombustor can burn pulverized coal
at rates from 0.454 to 1.13 kg/hr
(1 to 2.5 Ib/hr) and can maintain
maximum combustion temperatures ap-
proaching 14P2°C (2700°F)' for short
periods of time and lower tempera-
tures .loproaching [^13710C (^2500°F>]
for extended periods of time. The
166
-------
05)
TtlERKoCCA.'l'LE
WATFU-rriul^D
Dt'RNCD
UCATINC TACtS
cemteit DLAaitrr
' TMiCOTTIfcL
»stuu' (irMCTumr
arrtBCooLEu
TUKfUUCOUPLES
Ficjure 5. Combustion setup
dry pulverized coal is transported
to the combustor from a feeder by
primary air. This primary air com-
prises about 1/3 of the total air re-
quired. The remaining (secondary)
air can be introduced axially and/or
tangentially to the cohibustor to pro-
duce varying amounts of turbulence
and to alter flame shape. Gas resi-
dence time in the combustor chamber
ranges from about 0.7 to well over
1.0 sec depending on temoerature con-
ditions, fuel type, firing rate, and
excess air level.
The fuels and fuel mixtures used
in this study were as follows:
1. Illinois No. 6 coal*
2-. RDF powder
3.	25 percent RDF; 75 Percent
coal (by weight)
4.	50 percent RDF; 50 percent
coal (by weight)
The Illinois No. 6 coal was used as
+HIinois No. 6 Coal is a midwestern
bituminous coal
a referer.ce coal. Six combustion
rtiis were conducted with the various,
fuel combinations. A comoiiation of
the combustion runs and the results
obtained axe presented in Table 13.
¦ A slurry of RDF powder and dis-
tillate oil was fired in the Battelle
37 V.v (5C he) research boiler racil-'
ity' shown in Figure 6. This slurry
wa» prepared without the use of 3
dispersa.nt agent. ' The slurry burned
satisfactorily although fuel oiimDinq
problems were experienced when the
powder settled out of the distillate
oil-. The primary objective of these
co-nbustion tests was to demonstrate
the-feasibility of burning the RDF
powder in an oil slurry.
In these experiments, RDF pow-
der was co-fired wit'.i No. 2 fuel oil
in a 37 ku (50 hp) fire-tube boiler
with firigg r;ates between 1.3/ and
1.48 x 109 j/hr (1.3 and 1.4 x 106
BTU/hr). The slurry ignited easily ¦
and burned well. In order to feed
the powder/oil mixture into the boil-
er without plugging the nozzle, a
167
-------
TABLK	13. RESI'LTS Or THK COMDUSTOR EXPERIMENTS*
*»! . 		^	K» >
>#>	ffH (Mr ¦	¦~tl 1k('"	!iv |t«
»»> *,1*	*¦* 'Ml	Ini*|*	li» i vr. . i	Wi	»	»*
i ••	« S« It 11	»«~ MMV	.•* .'V«* »V1	1* ">	H.4 p J
j i»k4	•*.* |cj|	r»» ti*v»	i.»» kw!	a;o	i« « * t p • © *
:• . «	¦ * r ft	«'• »h»»i	•4 • »>< -• -	-- ics' i f
1 i «» *	1 M ill	-*l ,|tiM	v<» 'so ;.nx» -•	is » r- /«. 1 s .
i	+ t.M	*>	ft.*	ir» ISOO — If- « «. «.<~ *> 1
i» i • r*^vi
« • ^ i	I. * : 11	«V| JMISI	••» * t	j •» j w»	. - M « *	-» * .- »
*	' •*•>• '4.1	- t •) | Wt f ) V	I 1< > > 
-------
F'gure 6. Hattc!le-Columbus Combustion Research
I'ac lilt/
Table 14 summarized the results
of the combustion tests for the 15
percent by weight Dowder/oil mixture.
Visual examination of the flame in-
dicated good combustion with t[
generation of only a few "sparklers".
Smoke numbers could not be deter-
mi ->ed with any degree of accuracv
bocause of the presence of a yellow-
ish tint to the soot. Apparently
the combustion of the RDF oowder
somewhat "colored" the soot parti-
cles. This same Phenomenon has been
observed when firing residua; oils
containing trace metals. At excess
air levels above 6 percent, CO level."
were reasonably low. SOj and NO
levels are consistent with predic-
tions based on fuel analyses.
A particulate loader") of ftoo
mg/NmJ was measured for the powder/
oil mixture. This is relatively
high compared to the levels normally
recorded for residual oil (100 mg/
Mm' and distillate oil 20 mc'Vm").
Chemical analysis of the filter
catch indicated that less than on.*
percent of the oarticulate was un-
burned carbon. This result corfirms
the observation that the oowder/oil
slurry underwent comolste combustion.
These results also show that the high
inorganic content of the RDF Dowder
cause high particulate loading dur-
ing combustion.
As noted earlier, the slurry
with 30 nercent powder burned well:
however, sustained boiler ooeration
could not be achieved because of
fuel line and pump plugging problems.
Table 15 summarizes the data for the
30-minute run.
Based on the lim.'*-.ed tests con-
ducted it appears that the RDF pow-
der can be fired in slurry form.
169
-------
TABLE 14. SUMMARY OF 15% RDF POWDER IN DISTILLATE OIL CrvmUSTION
EXPERIMENTS. [BOILER FEED RATE WAS 36 KG/HR (80 LB/l!R)j
o2%
(dr/)
co2%
(dry)
CO
PPM
NO
PPM
S02
r>PM
Particulate
Load inn Mci-'N'rr
8.4
9
20
125
55
--
6.0
10.4
55
120
60
--
5.3
10.7
211
125
65
--
6.5
10.2
25
112
60
800
13.
SUMMARY OF 30* RDF POWDER TN IMSTILl.A-K Oil. COM-
BUSTION EXPERIMENT

0->%
(dry)
CO 2%
(dry)
CO
PPM
NO
PPH
so2
PPM

6.8
10.0
25
14"
70

5 . R
11 .0
2 35
140
7';

6.2
10.5
85
140
65
However, the hiqh inert content of
the povcer wi]l cause significant
particulate loadinc. In anJition, a
stabilizing agent is needcrf to main-
tain a homogeneous mixture of powder
and oil to inhibit settling in the
fuel handling system.
SUMMARY OF FINDINGS, CONCLUSIONS*
AND RECOMMENDATIONS
During the course of this study,
over 36 3 kg (600 lb) of RDF were pro-
cessed by embrittlement treatment for
conversion to a fine powder. The
major objective of this project was
the evaluation of" the RDF powder as
a fuel. The powder produced was
characterized and its combustion per-
formance with powdered coal and in an
oil slurry were measured.
Powder Characterization
The powder obtained by the em-
brittlement of RDF consists primarily
of short chopoy fibers minus 150u
(-100 mesh) in size. Disoersed in
the fibrous mix are•irreaular shaped
inert materials (predominantly sili-
cate glasses). The Dowder contained
5 to 7 1/2 rercent moisture artd 25 to
44 percent noncombustible material.
The high quantity of noncombusribles
("lass, metal and minerals) is a
characteristic of the RDF preoared
by Americologv in Milwaukpe and will
vary considerablv for different oro-
cesses and for different areas of
the country. From the proximate
and ultimate analysis it was found
that the Dowder contained abobt 52%
volatiles, had a carbcn content of
35*, and a chlorine content 'of 3%.
The powder had a loose density of
271 kg/m' (16:0 lb/ft5) and aa aver-
age heat content of 13025 Mj/kg.
(5600 PTU/lb). If an SDK with lewer
inert content were used, a hioher
heat rnnt-pnt v-mlrl hp obr-ai nprt. Thp
inert fraction of the powder was an-
alyzed to be a calcium, nagnesiiun,
sodium, aluminum silicate low in
iron. It had a fusion temperature
170
-------
above UeO'C (2480°F). Baseu ->n TG
and DT analysis i; is observed ti.
the RDF powder is easily ignited
(ignition temperature, 277°C (53D°F))
and burns readily.
Oompared to powdered Cf;al the
RDF' powder has a lower ignition tem-
perature, higher volati.le content,
much lower carbon content, aboiit .
half the heat content, twice the ash
content, one-tenth the sulfur, fif-
teen times the chlorine content ar.-i
is four-tenths the density. As a
fuel the RDF powder will ionitc and
burn more rapidlv than coal, but it
will generate less, thermal energy.
Combustion products will oresent
¦	about as nuch corrosion oroblems as
coal (CI versus S) but should pre;-
•sent greater handling problems be-
cause of a lower density and hiqn ash
content.
Powder/Oil Slurries
Suspension of the RDF powder in
~ fuel oil offers a number of advan-
tages for enhancing the use of the
TSF powder is l fuel (p isi'r 'r.ins-
cort and storage, greater -safety,
etc). The procedure for preparing
oowder/oil slurries and selected'
.properties were evaluated.
The powdered RDF can be easily
slurried in oil up to about 40
weight percent. The only problem-en-
countered was the sedimentation cf
the powder within four hours. How-
ever, more stable suspensions (in ex-
cess of 20 hr) were obtained with-
small (14) additions of a dispersion
agent (Rhectol). As would be expect-
ed the density and viscosity of the
powder/oil slurry is considerably
higher than the oure fuel oil (r* .
1 g/cm', „ = 32.00 cp)» . The stabil-
ized slurry does not appear to pre-
sent transport and handling diffi-
culties, however, long term exper-
ience with these fuel mixtures will
be required.
RDF Powder and ROT Powder/Coal Mix-
tures
The RDF powder and RDF powder/
¦	coal .Tiixtures were fired in a pul-
verized coal test combustor. The
RDF Dowder and mixtures of 25 and 50
woinht percent RDF powder with coal
were studied. \lthoyqh some hand-
ling problem.*! wre encountered with
the RDF courier .'je' to its lower den-
sity ar,j heat content, the RDF oow-
der/coal mixtures (particularly the
50/50 blend) handlei vt-v well a.*d
all ih'e corrDOsition.-; teste,: burned
well. Tn addition to its tiood hand-
line* characteristic? the 50'50 blend
proved to be the most effective fuel
nix studied. The lower ignition
temperature and higher quantity of
volatiles in the RDF aided the com-
bustion of the coal.
More comoletc- comoustxon is also
achieved ir the coal when it is nixed
with the RDF nowdor. The very low
sulfur and alkali content'of the RDF
oowder effectively, reduced SO-, in the
combustion gas emissiona. However,
the high chlorine content in the RDF
oowder may cause corrosion problems
and some environmental concerns. The
use of a lfess corrosive embrittling
agent *e.c., hno-j, H^PO.,
etc), if effective, could reduce or
eliminate "this nrohlen.
RDF Powder/Oil Slurries
The RDF powder/oil slurries
with ud to 30 vaiaht cercent RBF
burned well in the test furnace.
However, above 15 weight i>erce*jt there
were a number of problems enccantered
in transoorting the slurry. Plugging
of the Dunn and feed lines due to
powder sedimentation was the arimary
oroblem encountered during ti» com-
bustion tests. The use of a' iisT>er-
sion agent nhould alleviite this oro-
blem. The coTbustion of the cavder/
oil slurries oroduced considerably
more ash than is obtained wheat the
oil is burned alone. This causes
considerable handling problecs since
the conventional oil burning anits
are not designed lo process Xsrge
auantities of ash. The ash content
of the KDF powder is likely to be the
limiting factor for deterrrinisrj the
RDF powder to oil ratio fur r.J,iirrv
preparation.
Conclusions
An effective procedure Cor con-
171
-------
verting RDF to a fine powder by em-
brittlement treatment has been de-
veloped. The RDF powder aoDears to
be an effective fuel compatible with
coal and some oil burning equipment.
Althouqh the RDF powder can be burn-
ed alone or in combination with coal
and oil, its best performance was in
a mix with pulverized coal. The most
promising application for the RDF
powder is in a 50/50 mix with pulver-
ized coal. This fuel mixture burns
well (better than either component)
and results in lower SO2 emissions.
The major difficulties with the
use of the RDF powder as.a fuel is
the high inert content, (ash) and the
potential problems from th.a hiah
chlorine content. The use of screen-
ing and other classification oroces-
sing of the raw refuse should result
in a significant reduction of the
inert content and the use of differ-
ent embrittlement reagents (HNO3,
etc.) should eliminate poten-
tial .corrosion and environmental
problems. However, selection of ef-
fective unit processing procedures
to reduce inert content and selection
:f -x less corrosive er.be ittlemcnt
aqent will require additional re-
search studies.
During the course of the work
conducted on this project a number
of observations have be_n made per-
taining to the effectiveness of the
technology, the potential applica-
tions of the technology and the areas
needing further development. In this
work a technology for converting re-
fuse to a fine powdar has beon estab-
lished based on the use of cellulose
embrittlement techniques. However,
the mechanisms of the process are'
not completely understood and need
further elucidation. The identifica-
tion of effective alternative em-
brittlement reagents, particularly
reagents that would not leave resi-
dues which could be corrosive or de-
trimental in other ways is also need-
ed.
The RDF powder proved to be an
effective fuel when used in a 50/50
fWv	percent) rr.ijt with pulvsr-
ized coal. The RDF powder provided
considerable enhancemert for combus-
tion of the coal. However, more
needs to be known about the handlinq
(transport, storaqe, etc.) behavior
of both the MiF oowder and the pow-
der in a 50/50 mix with oulverized
coal. In addition, more needs to be
known about the emissions from com-
bustion of RDr'/coal mixtures.
It would appear that a varietv
of biomass materials, Darticularly
cellulose wastes from industrial and
agricultural sources (stalks, husks,
bark, wood and crou residue, ..straw,
etc.) could also be converted to a
powder for use as a fuel or as a
feedstock for biomass conversions.
However, more work is needed to de-
velop effective processing procedures
for powdering the variety of biomass
materials which might be available
for conversion to a powdered fuel.
Based on these observations the
following recommendations for future
work are proposed: (1) further elu-
cidation of the embrittlement mechan-
isms; (2) identification of alternate
embrittlement reagents; (3) further
characterization of the RDF Dowder
mixed with pulverized coal: and (4)
extension of the enuirittlement Dro-
cess to other biomass materials.
REFERENCE
1. Hecht, N. L., D. S. Duvall, A. A.
Ghazee, and B. L. Fox. 1980.
Obtaining Improved Products from
the Organic Fraction of Munici-
pal Solid Waste. EPA-600/P-80-
121, U.S. Environmental Protec-
tion Agency Cincinnati, Ohio.
172
-------
ASSESSMENT OF EPA1 s CELLULOSIC WASTE CONVERSION PROGRAM
Charles J. Rogers
L*. S. Environmental Protection Agency
Cincinnati, Ohio ^5268
ABSTRACT
Energy may obtained from municipal refuse by a number of processes, including direct
combust ion, cofiring vilh coal, pyrolysis, catalytic redaction or hydrogasification, acid
hydrolysis, ind byconversion the U. S. Environmental Protiction Agency (EPA) has sup-
ported research in these areas, but no single energy pxtraction process is 3 panacea.
In the early stages of EPA's resource recovery program, considerable emphasis was placed
on the use of microorganisms to convert ce1lulose into protein and chemicals. Research
the last 5 years has focused on the deve\opn*?nt of cellulose-to-glucose acid hydrolysis
technology for alcohol and chemicals production. This report summarizes research aimed
at the recycling of ccllulosic waste and discusses recent advances in acid hydrolysis
technology.
INTRODUCTION
An estimate 122 million metric tons of
solid waste are ppneratcd rach vear from
homes and commercial establishments (1).
A study completed in January 1979 (?) pro-
jected that gross solid wane discards
would total 142'million metric tons in
1
-------
bacteria used in the study ware mixed cul-
tures of Alcaligenes faecal:j and Cellulo-
monas. A microbial cell .oncentration of
12 g/liter and rates of cell synthesis of
1 g/liter pr hr were achieved in each
batch and occasionally in continuous fer-
mentation operations. A 552 protein pro-
duct was produced with an acino acid pro-
file tliat compared favorably with other
protein sources.
In another study, Leathervood (4)
studied tlie use of anaerobic bacteria to
ferment cellulosic wastes; strains of a
cellulolytic organism isolated from bovine
ruaen were evaluated. Products identified
included volatile fatty acids, soluble
carbohydrates, and microbial cells as a
potentially useful protein.
Undegraff (5) investigated the con-
version of waste paper co a protein sup-
plement for use as livestock feed. Only
one of the fungi evaluated, Aspergillus
fumlgatus, was capable of atcacking puri-
fied cellulose, ballmilled newspaper and
newsprint.
.'Joth the Leathervood and Undegrnff
studies were discontinued because of the
slow races of cellulose-to-protein conver-
sion by test microorganisms.
Pretreatment Processes
Physical ind chemical processes de-
signed to enhance the digestion of waste
cellulose during fermentation processes
were studied both in-house (f) and by other
wjrkers through EPA grants ind contracts.
The processes investigated ^re described
oriefly.
Sensitized Photodegradation
A 1967 patent of Schwartz and Rader
(7) revealed that polysaccharides such as
starch and cellulnsic materials can be con-
verted to saccharides of lower molecular
weight by irradiating them with light rich
in frequencies near 335 nm when the ma-
terial is in the presence of a water solu-
ble met.il or a nitrogen-based salt of
nitrous or hvponitric acid. On the basis
of this information, Frohnsdorff an!
TABLE 1. RESULTS OF A. FL'Ml LiATl'S CROW 111 XF.STS ON
MATERIAL FROM WASHED AND DRIED FILTER CAKES
Sample
Numbe r
Amount of
Amount of
Irradia-



Solka-floc
NaNO,
tion'1 time


Protein''
(8)

(hr)
Addi tives
DP
(?.)
U-182
10
0-..100
19
c
lral acid
.336
9.8
V-183
10
1; 100
19
2ml acid
160
8.7.
W-187
10
0; 100
19
2ml acid
328
10.5
X-188
10
1;10G
19
3ml acid
149
9.1
Y-198
10
1; 100
19
-------
rookson (8) investigated the technical and
.'colonic feasibility of using the sensi-
tized ;>hotodegradation process fo enhance
the use of waste cellulose In recycling
processes.
When 2.72, 200 mesh cellulosic slur-
ries were Irradiated in a 13! sodium ni-
trate solution of 2537 Angstroo, the degree
<>t pjimeri'jtion (DP) was reduced from
1000 to 800 in less than 2 hr, and finally
to 125 after u8 hr. The number of scis-
sions calculated from the DP also increased
linearly.
Table 1 presents t!w growth response
of AsperpJ 11 us funlgatus on irradiated cel-
lulose under various conditions. Based
on protein yields, there appears to be no
relationship) between the DP and the diges-
tion of cellulose by this fungus. These
findings wore contrary to expected results.
Further tests wich other Fungi confirmed
the findings observed with A. fumlgatus.
Microwave
Johnson (9) reported that slurries
with starch contents as high as 752 could
be hydrolyzed to sugars and other decompo-
sition products with microwave energy at
180° to 200°C. To enhance the blodegrada-
tion, cellulose was treated with micro-
wave energy as follows:
TVenty millJ liters of a IS cellulose sus-
pension in water was placed in each of
several Pyre* tubes with 3/4-i.n. O-D. and
1/8-in. walls. These tubes were frozen,
sealed under vacuum, and then treated for
various lengths of time in 9 2450-mcga-
hertz microwave oven. During a period of
25- to 30-min, temperatures rose above
2C*°C and pressures above 400 psl were
attained. As a measure of cellulose modi-
fication, the filtered, treated solution
was tested for the presence of soluble
sugars, A 12 cellulose suspension in 21
hydrochloric acid was also subjected to
microwave treatment. After the liquid was
analyzed for sugar concentration, the re-
maining cellulose was washed, dried, and
subjectfd to enzvmolysis. Sugar produced
by cellulose action was taken as a measure
of cellulose structure modification by
microwave treatment. Under these experi-
mental cond_Ltons, microwave energy did not
improve enzymatic degradation of celluiose
(Table 2).
Laser Treatment
Klein (10) demonstrated that lig.10-
sulfonate could be photomodifled to improve
biological availability of this material.
Biological responses were measured by as-
says of substrate carbon before and after
growth, by fungal dry weights, and by the
presence of bacterial populations relative
to substrate photolysis treatment time.
The energy source used for llgnosul-
lotiate mouification was a medium-pressure,
500-watt, type A mercury vapor lamp, model
06.'3A-36. The power source vas a 020651-1
transfira^r (llanovia Company, Newark, SJ) .2
A Ha.'Ovla double-wall immersion well (model
119-i'*',) contained the lamp and allowed th"
coolant water to circulate. A Pyre* reac-
tion vessel was designed to fit externally
to the immersion well.
A IX U/V cellulose slurry (200-mesh
Solka-floc) In tap water was irradiated for
TABLE 2. EFFECT OF MICROWAVE ENERGY ON THE SUSCEPTIBILITY
OF CELLULOSE TO DEGRADATION BY ENZYMES3
Glucose produced at various enzyme reacrlon times

1,5 hr
3.0 hr
4.5 hr
6.0 hr
Control (unheated cellulose)
55.5
53.5
57.0
63.0
Cellulose heated 25 min In water
47.0
47.5
495
63.5
Cellulose heated 30 Bin in water
44.5
47.0
53.5
57.0
aCellulose (0.32 g) was suspended in 25-nl buffer (pH .4.0) and was incubated at
40°C. Alter Incubation, tne enzyme was in'tiiuiLeu l»> Iieatir^g ts S0°C for 10 iiir..
175
-------
I, 2, 16, and 20 hr with an air flow of
approximately 200 ail/per mixture.
When these laser-treated	samples were
. '.ested for increased rates of	biological
growth response, all showed a	marked im-
provement over the control.
Alkali
Cellulosic substrates were subjected
to an alkali-oxidation treatment (11).
The treatment sequence was designed to
swell the cellulose structure, break up
the llgnin physical structure, and modify
the cellulose into shorter molecular
chains. One gram of each alkali-treated
cellulose was tested to determine the bio-
degradation rate and compare it with that
of untreated cellulose. All samples showed
modestly improved rates of biodegradation
over the control (6).
High-Temperature Hydrolysis
A high-temperature, stirred chemical
reactor constructed of Type 316 stainless
steel was used to scudy the effects of
high-temperature hydrolysis on the liode-
gradation rates of cellulose. Thin treat-
cent process-was designed to remoV': some
crystalHnity and to provide some readily
fer.nentable sugars so the initial growth
of the molds would be accelerated. Theo-
retically, 10Z fermentable sugars could be
liberated under the conditions used. The
IZ cellulose suspension in the reactor was
renoved and adjusted tc pK 5. Biodegrada-
tion rates were then compared with that of
untreated cellulose. The treated samples
were less susceptible to biodegraaation
when comf> .t- ' with controls (6).
Electron irradiation
In an experiment designed to deter-
mine whether irradiated cellulose was store
susceptible to degradation than untreated
cellulose, two polyethylene b.igs (one con-
taining 10.0 of 40-oesh ground cellulose,
and the other containing 10.0 of 100—mesh
ground cellulose) were irradiated for 1
and 2 hr, rspectlvely. A dynamic electron
source (Dynamatron, NASA Lewis Laboratory,
Cleveland, Ohio), provided an exposure of.
10® rad/hr. The it radiated samples were
then used in biodegradation studies and
compared with nonirradiatcd cellulose -
samples of the same mesh sizes (6).
Evaluation of Pretreatment Processes
The degradation of untreated pure and
waste cellulosic substrates was compared
with that of celluloslc substrates previous-
ly treated by alkali oxiJatlon, high-tem-
perature hydrolysis, elect'on radiation,
and the nitrite photochemical process (8).
The mediun for these studies contained 1
each of Nl^Cl, K2HPO4, KH2po<<( MgS04, CaCli,
and yeast extract, and 40 ing of thiamine
combined with sterll water to a final volume
of 1 liter. This mineral salt medium was
adjusted to pH 5, and 100-ml portions were
placed in 500-ml shaker flasks, each con-
taining I g of cellulose. The flasks were
autoclaved at 121°C for 15 min. After
cooling, the medium was inoculated with
selected fungi, incubated at 35°C in an
environmental room on a reciprocal shaker
for 4 days, and observed for the disappear-
ance of cellulose and the growth of the
fungal mass. The inocula for these experi-
ments were obtained from 48-hr fungal cul-
tures grown in a mineral salt medium con-
taining glucose, cellobiose, and cellulose
in 0.1* concentrations. The dry weight of
each inoculum was approximately 10Z of the
weight of the substrate (6). Samples were
taken after 4 d.iys of incubation and ana-
lyzed for their protein value based on the
content of amino acids.
The .amino acid contents ot peanuts,
soybean meal and cellulolytic fungi grown
on cellulose are reported in Table 3. These
data confirm that fungi can produce protein
containing complements of essential amino
acid from cellulosic waste. The improve-
ment, identification, and development of
new, more effective pretreatment processes
will be required, however, if cellulose-
protein conversion Is to become a technical
and economically feasible reality.
Hydrolysis of Cellulose
Acid Hydrolysis
Porteous (12) proposed a new technique
for waste disposal that would use both agri-
cultural and municipal wastes as raw materi-
als. This proposal consisted of hydrolyzing
the cellulosic 'tastes to sugars.
Fagai: et al (13) explored this concept
Irt detail. The kinetics f-.r hydrolysis uf
cellulose in paper were experimentally de-
termined over a range of 180° to 240°C with
176
-------
TABLE 3. AMINO ACID CONTENT OK SELECTED FUNGI. PEANUT MEALS, AND SOYBEAN MEAL
(>;rarn amino ncid/lOO graas protein)		




J'rtch t-

(jffirii




A /vn.'£aifej
fumipixus
PfntciUium
Jrrnui
Choriorn»wn
rair/u/wn
Peanut
So\bcan
FAU
Amino acid
No. 3
No 6
Jp. No. 7
ttruJf ."«o ¥
>p. No. 10
No. II
hkjI*
me-l1
rcf* ¦
Alanine
5 90
5 60
5 90
5 l(p
6 60
6 00
42
3 30

Arginme
3 ¦'O
3 B0
7 80
4 00
a m
6 70
10 6
7 .V)

Aspariic acid
8 N*
H 30
8 90
- 7 70
10 70
ft 40
15 1
3 70

C)>(inc, half

0 40




1 60
1 9
1 2
Glutamic acid
II 00
i: 30
lJ 30
9 YO
11 70
U JO
17 40
If 40

Cilycrhr
¦ > JU
3 SO
¦ 5 70
5 90
5 6U
4 2o
5 (10
4 Oil

KblidtiK
2 :o
\ w
3 30
t bO
4 70
J 20
2 io
2 W

Im? leucine
* .10
1 20
4 50
4 00
S 60
J 90
4 00
6 00
4 I
Leucine
8 80
10 90
5 70
5 10
6 GO
6 70
6 70
S 00
4 K
Lysine
J.40
; 60
5 70
4 40
5 60
5 30
3 00
6 80
4- 2
Methionine
7 3*1
6 50
1 141
1 U)

: 70
I (W
1 70
2 1
Phenylalanine
6 10
5 30
4 ;u
5 50
6 10
7 40
5 10
5 30
2 &
Praline
; yo
3 40
3 60
3 70
- 2 B0
3 20
s :o
5 Ou

Serine
¦ 5 50
4 )0
4 50
I 60
4 :o
5 IX)
ft 60
4 20

Threonine
7 00
4 40
5 60
4 40
-1 so
7 40
I 60
3 90
2 8
Tryptophan








I 4
Tyrosine
3 W
3 00
6 10
3 70
4 20
4 20
4 40
4 0
2 3
Valine
2 NJ
5 30
6 10
i BO
6 60
|2.4(}
4 40
5 3
4 2
4 Bi.Kk.jnJ Bolltntil^l).
* Ck*m\CJ}end	.\f »J iWl.
f Food jnJ Agriculture Oifcuu/^tronO?}?)-
0.2Z to l.GZ acid In a batch leactor. The
sugar solution obtained (com refuse hydro-
lysis can be concentrated to 121 with the
use of a multi-effect evaporator, and it
can He used In commercial fermentation
processes.
This hydrolysis study further'"indicated
that development of continuous acid hydroly-
sis technology is needed to convert large
tonnages of cellulose to a readily, fermen-
table product for use in the production of
fuels, foods, and chemicals.
Contin-ious Acid Hydrolysis
The EPA-supported experimental inves-
tigations of tlie dilute acid hydrolysis of
waste cellulose to a glucose have been on- .
going at the Department of Applied Science
of New York University (SKU) over the past
5 yeaTs (14). The waste cellulose feedstock
employed In these studies was primarily used
newsprint.
Initially, the hydrolysis experiments
NYU were carried out in a 1-liter stirred
autoclave equipped vlth appropriate acces-
sories, Including electrical heating units
and a quick-discharge ball valve for remov-
ing the reaction mixture from the autoclave
after acid hydrolysis. The data obtained
with the 1-liter stirred autoclave rsartnr
experiments were analyzed to determine the
glucose yield at various reaction conditions.
This work was followed by additional test-
ing in a 5-liter stirred autoclave reactor.
The batch-scale hydrolysis experiments
showed that glucose yields up to 502 more
of the charged available cellulose values
could be obtained. The optimum reaction
conditions were found to be temperatures
of 220° to 230°C and reaction times of less
than 30 seconds, with about IS of sulfuric
acid by weight. These conditions agree
rather well with the results of the kinetic
rate studies previously reported by Fagan
et al (13).
.The NTH work includes investigation of
continuous processing technology for indus-
crial-scale conversion of waste cellulose
to glucose. Fro* this part of the study
emerged a continuous waste-cellulose-to -
glucose pilot plant with a capacity'of
1 ton per day. This pilot plant utilized
waste paper and sawdust as cellulose feed-
stock, and a reactor device for continuous-
ly reacting cellulose at suitable elevated
temperatures.
The Werner & Pflelderer ZDSDS3 (53-no)
twin-screw extruder was selected because of
its capacity for conveying, mixing, and
extruding the required amounts of cellulo-
sic feedstock. This machine allows accu-
rate control of £e£pcr<*i.uit: urting inten-
sive mix. int.
177
-------
For continuous processing, the extru-
der must be coupled vi^h a feeding mectw-
lsm for cellulose slurries arid a discharge
system for reacted materials while mai..::ln-
lng pressure and temperature in tl.e reac-
tion zone. A steam-jacketed cramner-feeder
made by Werner & Pileiderer was integrated
with the twin-screw extruder to maximize
throughput with preheating as required. An
i.itensive-rservice 2-ln. ball valve (Kamyr
Valve Company, Glens Falls, New York) was
selected as the major component for the de-
sign of the discharge system. Other ancil-
lary equipment includes a high-pressure
steam generator for supplying energy to the
reactor, an acid pump capable of high-pres-
sure injection, and a sluTry pump for intro-
ducing feedstock into the crairmer-feeder.
The continuous acid hydrolysis optimi-
zation studies are continuing. Even though
the study has achieved a 50Z conversion of
cellulose to glucose, further 'experiments
are underv-iv to optimize reaction conditions
for maximum glucose yields.
Two-Stage, Hydrolysis System
A proposed two-sta(;e hydrolysis system
development based on the VYl' work (Figure 1)
would consist oi two extruder reactors —
one operated at 336°K for pentosan hydroly-
sis, and the second operates a- A50°F for
•hexosan twJrolysis. The pentosan is hydro-
lyzed at lowi^r temperature and lower acid
concenttirion than hexosan. The feedstock
is hydrolyzed in the first-stage reactor
and discharged into slurry tank that con-
nects the two hydrolyzers. The feed ma-
terial containing the hydrolyzed pentosan
and unhydrolyzed cellulose is mixed with
water in the slurry tank and pumped into the
second reactor. The pentoses are recovered
by the action of the twin screws at the de-
watering drain, and the cellulose (tree of
pentosan) is conveyed through the second
unit and hydrolyzed to glucose. The pre-
hydrolysls of pentosan is expected to im-
prove further Che 50Z to 60Z cellulose-glu-
cose yields now being achluved with the
l-con/day unit.
In the past, little or no consideration
wap given to recovering pentoses. Acid
hydrolysis realtor developnert studies fo-
cused primarily on converting hexosan to
biuCwSC. Ditto Oh c'leuiiA-dx diijuyws oi agri-
cultural residues and wood (Table 4) re-
vealed that pentosan constitutes up to 29.SZ
of some agricultural residues (15). The
pentosan fraction is Important because It
has been demonstrated that both pentoses
and hexoses can be fermented to ethanol by
yeast (14). If maximum feedstock utiliza-
tion is to be achieved, acid hydrolysis
technology must be developed to recover
both the five- and six-carbon sugars
efficienfly.
Yield Projections
Investigators have used diluted and
concentrated sulfuric acid in two stages
to yield 953! conversion of pentosans to xy-
lose and 90! conversion of hexosan to glu-
cose (15). Optimization of the two-extruder
process is expected to achieve up to 95?
conversion of pentosan tj- xylose and 80J
recovery of the sugar. With the removal of
pentosan by prehydrolysis, we further ex-
pect to achieve values greater than 60Z con-
version of celluose to glucose.
The combined two-stage process has the
advantage of (a) having a design potential
for recovering up to 100Z more sugars than
the single-stage process, (b) reducing de-
gradation products of pentose that are dis-
charged as a glucose mixture, and (c) sup-
pressing glucose fermentation.
Table 4 presents sugar yields that
could be expecfed if the two-st^ge extruder
reactor were optimized to recover 60S
IMlrOVMM MAtmitf ASaCkJLrUAAl

•(•ini i
1
IWK» *
Con }tmr
ir.o
11 6
lt.Q
Con C**»

71.f
19.4
*Ml UW
ii.).
J).4
i«o
UK* Strw


19.9
eau miii -
»»i


HfitM
u.>
M.I
lt.t
Mm
Oi*
S.I
!».7
11.4
tt.t
N.I
178
-------
pentose and 60% glucose from agricultural
residues ~nd wood.
In the United States, an estimated 271
million metric tons of agricultural resi-
dues from food production are unused (14).
If this tonnage were used to produce ethyl
alcohol, a considerable savings In corn
could be achieved. At projected rates of
gasoline consumption, 10 billion gal of
ethanol will be needed by 1990 to achieve
universal use of gasohol.
To calculate the amount of Eugar avail-
able from the unused 271 million metric tons
of agricultural wastes, consider the follow-
ing. Total sugar pocentlally recoverable
from the acid hydrolysis of agricultural
residues ranged from 31.12 to 44.42
(Table 5). Thus, if we assume only a 301
sugar recovery from the hydrolysis of agri-
cultural residues, 271 million metric tons
of waste would yield 81.3 million netric
tons of sugar.
In the project of alcohol yields, It
is assumed that 15 lb of waste-recovered
sugar (as opposed to approximately 14 lb
of corn-recovered sugar) are required to
produce 1 gal of alcohol. Ethyl alcohol
yields, from -71 million netric tons of agri-
cultural residues are therefore calculated
are follows:
271 million-metric tons x 301 «
81.3 million metric tons of sugar
Gallons of ethyl alcohol
m 81.3 pillion metric tons x 2.000
15 pounds of sugars
¦ 6.0 x 10^ x 7.0 x 10^ of alcohol
- 12.0 x 109 gal
Thus ethyl alcohol requirements for
gasoiiol can be theoretically produced from
unused waste agricultural residues.
MMi-.iMtnrncoviMiiiiiAtnrm
AMCU11UU4 MUQuU MO OOOO
1 Tom MM t
n.o	n.o
n.t	«.•
».«	».«
ai	r.:
lot	».«
K.I	>1.1
]t.a	v.l
n.t	u.>
The corn equivalency is 4.4 billion
bushels of com: Divide the projected al-
cohol yield from agricultural residues (12
billion gallons) by the number of gallons of
190° proof alcohol derived from 1 bushel of
corn (2.7). In 1979, about 7.8 billion
bushels of corn were s'.own in this country.
Developing and implementing hydrolysis
technology could contribute greatly to meet-
ing our energy requirements by savin? the
energy needed to cultivate an adddltional
119 million hectares of grain for the 10
billion gal of ethanol. Savings would also
be effected on fertilizers and chemicals,
and a more efficient use of total crop
would be provided. Considering the poten-
tial of hydrolysis technology, It is incun-
bent that nothing less than a national
effort be committed to the development and
widespread implementation of this technology.
Conclusions
The widespread use of grain to produce
alcohol as automobile fuel will likely
cause a surge in food prices. It, therefore,
appears that Alcohol produced from grain is
not, therefore, likely to be cost competi-
tive with gasoline derived from oil. The
key to usin^ ethyl alcohol as a gasoline
extender depends on a technological develop-
ment that provides for the conversion 6f
cellulosic materials into simple sugars for
alcohol production. Currently, research is
under way to optimize hydrolysis conditions
with the extruder reactor for oaxlim
-------
2.	Forecasts of the 'Quantity and Conposi-
tlon of Solid WAt:e, (In Press)
Contract ff(>8-03-:'649, U.S. Envlroniucn-
tal Protection A.ency, Cincinnati, Ohio
45268.
3.	Callihan, C.D. 1968. Construction of
Continuous Chemlcal-Hicrobial Pilot
Plant to Produce Edible Proteins from
Celluloslc Waste. Contract SPH-86-68-
152. U.S. Environmental. Protection
Agency, Cincinnati, Ohio 45268.
4.	Leatherwood, J.M. 1968. Utilization of
Fibrous Wastes as Sources of Nutrients.
Grant No. EC-00274-03. U. S. Environ-
mental Protection Agency, Cincinnati,
Ohio 43268.
5.	Undegraff, D.N. 1967. Degradation of
Wastepaper to Protein. Crant No. EC-
00271-02. U. S. Environmental Protec-
tion Agency, Cincinnati, C!iio 45268.
6.	Ropers, C.J. et al. 1972. Production
of Fungal Protein from Cellulose and
Waste Cellulosics. Environmental Sci-
ience Technology, Volume 6, Number 8,
(August).
7.	Schwartz, A.M., and Rader, C.A. 1967.
Method of Degrading Polysaccharides.
Using Light Radiaticn and a Water-Solu-
ble Metal, or Nitrogen Base Salt,
Nitrous, or Hyponitric Acid. LI. S.
Pat. 3,352,773, Nov. 14.
8.	Frohnsdorff, C., and Fooltson, A. 1971.
Investigation of the Sensitized Ph'ito-
degradation of Cellulose and Wastepaper
to Useful Products. Contract No. ETA-
68-03;-0006. U.S. Environmental Protec-
tion Agency, Cincinnati, Ohio 45263.
9.	Johnson, J.A. 1971. Starch Hydrolysis
in Water with Microwave Energy. Paper
No. 44. 56th Annual Meeting Am. Aisoc.
of Cereal Scl. Today 16, 1972.
10. Klein, D.A. 1968. Laser-nediated Lignln
Solid Waste Fermentation. Grant N-j.
EC-00278-02. U. S. Environmental Pro-
tection Agancy, Cincinnati, Ohio 45268.
12.	Portrous, A. 1967. Toward a Profitable
Means of Refuse Disposal. ASME publi-
cation 67-WA/PTD-2, 17P,
13.	Fagan, R.D., Gethleln, H.E., Converse,
A.O., and Porteoub, A. 1971. Acid
Hydrolysis of Cellulose lc Refuse,
Environmental Science and Technology
5, 545.
14.	Brenner, W. and Ruftg, B. (1975-1980)
Continuous Twin Sciew Acid Hydrolysis
Reactor Development for 1-ton/day
Celltilose-Clucose Pilot Plant, EPA
Coooerative Agreement No. 805239. New
York University, N.Y.
15.	Sitton, C., December 1979. Ethanol
fron Agricultural Residues. Chemical
engineering Processing.
16.	Dunning, J.W. ; nd Lathrop, E'.C. 1945.
Industrial Engineering Chemical 37 (1)
24.
11. Dunlap, C.E. 1969. Single-cell Protein
Production from Bagasse. PhD disserta-
tion, Louisiana brace University, B_ucm
Rouge, LA.
-------
ADVANCES IN THE RECOVERY AND UTILIZATION OF LANDFILL GAS
Stephen C. James
U. S. Environmental Protection Agency
26 W. St. Clair Street
Cincinnati, Ohio 45268
ABSTRACT
Natural decomposition processes within a landfill produce gas consisting prinarily
of methane and carbon dioxide. The generation of this gas within a landfill poses many
potential problems due to migration through the. surrounding soils. In order to .ontrol
migration, collection systems have been installed. The gas collected from tne system had
traditionally been flared. Now, however, methane gas generated in landfills is being ef-
ficiently collected, processed, and used as a source of energy. Eight utilization systems
are currently on-ljrie, with several others scheduled for start-up within 1981.
INTRODUCTION
The amount of recoverable pnerny avail-
able from landfill gas is unknown; however,
it appears that approximately 100-200 bil-
lion cubic feet of raw landfill gas is pro-
duced each year from disposal of municipal
solid waste. While this is a very small
amount in terns of U. S. natural gas needs,
landfill gas can play a significant role in'
.-.meeting industrial and utility energy reauire-
nents when considered on a case-by-case basis.
PRODUCTION
Decomposition - The Basic Process
¦¦ The process "»f landfill gas formation
is mainly an anaerobic process, not unlike
that of a sewage sludge digester. Aerobic
digestion takes place initially because
la>-ge quantities of air are entrained in the
waste during placement. The oxygen is
quickly consumed and the process becomes
anaerobic shortly after refuse placement.
Anaerobic refuse decomposition is a
continuous pricess that stabilizes the
organic wastes and results in the production
of methane. The organic material, such as
leaves, paper, and food waste, is used as
food tor tne acia-rorming Dacteria. Ihis
organic material is then chamed by the
bacteria to simple organic material, mainly
organic acids. Methane-forming bacteria
then use these organic acids as food aril
produce carbon dioxide and methane gas.
When the waste stabilization process
proceeds normally, approxinvately 40 to 60
percent of the gas produced will be methane.
The remainder will primarily be carbon
dioxiue.
Methane fonrers grow quite slowly com-
pared to acid formers since they obtain
very little energy from their food. This
results in the methane formers being' very
sensitive to slight changes in the environ-
mental factors. The acid formers are rapid
growers and are not so sensitive to environ-
mental conditions. Thus the production of
landfill gas is largely dependent upon
maintaining optimum conditions for the
methane-forming bacteria.
Environmental Factors Affecting The Decom-
position Process
Unfortunately, the landfill decorposi-
tion process is different from a sewage
sludge digestion process because critical
environmental factors (tenperature, pH,
and moisture) cannot be easily controlled.
Temperature control i* a kpy fartnr
for successful anaerobic stabilization of
organic matter because sudden tenperature
changes greater than two degrees centi-
181
-------
grade will result in losing the buffering
capacity and possibly incapacitating the
digester. The temperature should also be
controlled in the range of 29 to 37 degrees
centigrade so that the optimum gas production
may be achieved. Although the temperature
1n the landfill cannot be controlled, It has
been determined that the internal teirperature
of many landfills tails within the optimum
temperature range for gas production. It
has also been observed that the core tem-
perature of deeper landfills is not affected
by diurnal temperature fluctuations.
The ffoi.sture content required for op-
timum anaerobic decomposition has been
reported to be greater than 60 percent.
This, again, often occurs in landfill situ-
ation, although many landfills with far
lower percentages of moisture have been found
to produce large quantities of gas. Mois-
ture addition at landfills has been pro-
posed to enhance gas production.
The optimal operating range for pH is
from 6.8 to 7.2. Many landfills report
lower pH levels but still produce signifi-
cant quantities of gas. It is believed
that the pH within a landfill does not fall
below 6.2 when methane is produced.
The factor which is probably most cri-
tical to the landfill stabilization process,
particularly when methane gas recovery is
anticipated, is air infiltration. Whenever
methane gas is removed from a landfill,
there is a tendency for air infiltration
due to leakage through the recovery wells
and landfill surface. Air is toxic to the
methane-forming bacteria and thus will stop
the production of methane gas. Here the
typical configuration (depth) of the landfill
becomes an important factor because the oxy-
gen in the infiltrating air is consumed in
the upper portion of the landfill and does
not hinder the anaerobic process at the
bottom of the landfill. Depths greater
than 100 feet are ideal for landfill gas
recovery. Depths as low as 30 to 40 feet
are suitable for gas recovery, but more
control over minimizing air infiltration
is needed.
RECOVERY
Control and Recovery Methods
In the last 10 years, landfill opera-
tors, owners, and engineers have become
increasingly aware of the potential hazards
caused by the methane components of landfill
gas. Methane migrating through soils
adjacent to the landfill has on occasion
collected in nearby structures and ignited,
resulting in structural damage, injuries,
and even deaths. This has resulted in the
technology for the collection and recovery
of landfill gas.
Actual recovery of landfill gas for
methane resulted from efforts to stop the
migration of gas to adjacent properties.
The first control methods used to prevent
landfill gas migration were peripheral
trenches filled with porous media or peri-
pheral vent pipes which allowed gas to vent
to the atmosphere. These control methods
were found to be generally ineffective.
Recently, the technology has advanced
to the point that most new control systems
were power exhaust vent systems Composed of
wells and a header connected to an exhaust
blower. This advance in the technology
coupled with the Impending natural gas
shortage was the catalyst necessary to
launch the landfill gas recovery industry.
In general, recovery of landfill gas
for utilization employes one of the following
two options. First, raw landfill gas can
be used with minimum processing. Removal
of moisture and some compression are neces-
sary for initial upgrading. In this form, .
landfill gas can be used to generate elec-
tricity and fire boilers for space heating
end industrial processes. The other option
is carbon dioxide and, if necessary, nitro-
gen removal. This produces a gas which can
be directly injected into an existing natural
gas pipeline. These options are currently,'
in operation at landfill methane recovery
systems.
ON-LINE LANDFILL METHANE RECOVERY SYSTEMS
Mountain View Landfill, Mountain View.
California
This is a demonstration project started
1n 1975 and jointly funded by Pacific Gas
and Electric Company and U. S.' Environmen-
tal Protection Agency. This facility recov-
ers landfill gas from a 30 acre site by
means of 33 ve'ls. Since this is a shallow
landfill (3C - 40' deep), a greater number
of wells are necessary due to the lower
rate per well. The gas is upgraded by
compression- and adsorption to remove water
and carbon dioxide. The resultant 700 6tu/
scf gas is then injected into a PG&E main
transmission line.
182
-------
Palos Verdes Landfill. Rolling Hills Estates.
Cali fornia
This is the first landfill methane re-
covery system^ The plant is owned and oper-
ated by Getty Synthetic Tuels, Inc., and
has been in operation since 1975. Severe
corrosion problems experienced in the pre-
treater regeneration system adversely af-
fected the system's operational reliability.
A plant modification program was completed
in 1976 and the system has operated reli-
ably since January 1S77. The product gas
is sold to Southern California Gas Company
and has a higher heating value of approxi-
mately 1.000 Btu/scf. In May 1979, further
plant modification increased the capacity to
two million cubic feet per day inlet capa-
city.
Azusa Landfill, Azusa, California
Since 1973 Azusa Land Reclamation
Company, Inc. has been recovering and sel-
ling low Btu landfill gas to Reichhold
Chemical Company. A well system consisting
of 20 wells recovers 400,000 cubic feet per
day of raw gas. After water removal and
compression, the treated gas (500 - 520
Btu/scf) is transported via a 4,000' pipe-
line to Reichhold Chemical Company for use
in their boilers. The concept of on-site
electrical generation is being investigated.
ftscon Landfill, Wilmington, California
This system is owned and operated by
Getty Synthetic Fuels, Inc. and represents
an advancement over their Palos Verdes opera-
tion. This operation consists of dehydration
and carbon dioxide removal to produce 4,000,000
cubic feet per day of pipeline quality gas
at maximum capacity. The pipeline quality
gas is sold to Southern California Gas
Company for injection into their transmission
line.
City of Industry Landfill. City of Industry.
California
This system was designed by SCS Engineers
to supply medium Btu gas as fuel for the
boilers at the Industry Hills Convention
Center. The system consists of dehydration
and particulate removal to produce 500,000
cubic feet per day of treated gas.
Sheldon-Arleta Lan> ,11. Los Angeles.
California
Completed in November of 1979, this
project supplies low Btu gas to the Valley
Generating Station for use as boiler fuel
for the production of electricity. Four-
teen recovery wells provide gas for the
dehydration and compression system. A 1.8
mile 10-inch pipeline carries the gas to
the generating station. At the power
station, the landfill gas is used as a
supplemental fuel supply during normal
boiler operation. However, the landfill
gas can also be used for ignition and boiler
start-up when natural gas is not available.
Cinnaminson Landfill, Cinnaminson, New Jersey
Started in 1979; Public Service Electric
and Gas Company of Newark, New Jersey, recov-
ers raw landfill gas for the purpose of sel-
ling it untreated to a nearby sponge-iron
factory owned by Hoeganaes Corporation.
Initial quantity of gas delivered to the
plant was 230,000 cubic feet per day. How-
ever, consumption has increased to 600,000
cubic feet per day with plans to deliver
1,000,000 cubic feet per day in the future.
Systems Nearino Completion
Getty Synthetic Fuels, Inc. is cur-
rently startlng-up systems in Calumet City,
Illinois, and the San Francisco Bay Area
(San Leandro). Both of these systems are
patterned after the Honterey Park Landfill
recovery system.
Brooklyn Union Gas in conjunction with
Mew York City Office of Resource Recovery
and Waste Disposal and Hew York State ERDA
wlll.be conducting a one-year test on a
landfill gas electricity-generating faci-
lity supplying 100 KM of electricity for
on-site us®.
In addition to the above projects,
many studies are being conducted at various
landfills to determine the technical and
economic feasibility of recovery and uti-
lization of landfill gas. These studies
are concerned with the direct use of low
and high Btu gas as well as the conversion
to electricity.
ON-GOING RESEARCH
Corrosion Studies
Pacific Gas & Electric Company is
currently conducting a study to determine
the corrosion potential of landfill gas at
183
-------
the Mountain View Landfill. Corrosion pro-
blems were observed in the radiation cooVfiS
tubes prior to the injection of a corrosion
Inhibitor. Corrosion monitoring probes
have been installed and this problem ii cur-
rently being studied.
Industrial burner tests at the Fresh
Kills Landfill have shown the recovered
landfill gas to be a non-corrosive fuel for
use in conventional natural gas designed
equipment.
Controlled Landfill Process
Pacific Gas $ Electric Company and
Southern California Gas Company are conduct-
ing joint studies at the Mountain View
Landfill concerning enhancing gas production
and improving gas recovery. Initial studies
have been performed and a field demonstra-
tion program is being developed. Factors
such as moisture content, nutrient additions,
buffering agents, and the effect of density
on gas production will be investigated In
the enhancement project. In the area of
improving gas recovery, the effectiveness
of different extraction systems and the
effectiveness of different coyer systems
will be investigated.
Development of Models for Gas Production
Currently, one major problem involved
in landfill gas utilization 1s the deter-
mination of production rates. Production
rates determine the size of the landfill
gas processing plant and are the guidelines
for economic success. Presently, Dr.
Charles Moore of Ohio State University is
developing computer programs which will
allow design engineers to assess the varying
system parameters to optimize production.
HaxBlanchet of Pacific Gas & Electric
Company has also developed a model for pre-
dicting gas production over and beyond the
operation life of the landfill by examining
the various decomposition rates of the com-
ponents of municipal refuse.
NEEDED RESEARCH
Gas Production
Further testing and development of
models for gas production are needed, the
need is not just for evaluation In conjunc-
tion with recovery systems, but also In
conjunction with gas migration problems.
In reference to migration problems, there
is a definite need for prediction of pro-
duction rates over a time period for com-
pliance with RCRA.
Pollution Aspect
There are generally two waste streams
from landfill gas treatment technology.
The first 1s a condensate from the dehydra-
tion process. The second is a low quality
Btu blowdown stream from the carbon dioxide
removal system. Currently, the condensate
1s recycled back to the landfill and the
blowdown is vented to the atmosphere. Both
of these waste streams should be further
analyzed to determine their pollutional
aspects.
Other Aspects
Further studies on corrosion including
impurities 1n the pipeline need to be per-
formed. Utilization of by-products, such
as carbon dioxide, need to be further inves-
tigated, especially in the area of markets.
Methods for reduction of air intrusion into
the landfill also need further investiga-
tion, especially at shallow landfills.
In addition to the technological
aspects, the following economic and insti-
tutional aspects need to be further inves-
tigated:
o marketing difficulties
o risk assessment
o utility regulations
CONTACTS
The following Is a list of contacts
for the projects described in this paper.
Max Blanchet
Pacific Gas & Electric Company
245 Market Street
San Francisco, California 94106
(Mountain View Landfill)
Fredrick Rice
Getty Synthetic Fuels, Inc.
2750 Signal Parkway
iignal Kill, California 90806
(Palos Verdes, Honterey Park, San
Leandro, and Calumet City)
IM
-------
Hike Miller
L.A. Bureau of Sanitation
Room 1410, City Hall East
los Angeles, California 90012
(Sheldon - Arleta Landfill)
Anthony Giuliani
Brooklyn Union Gas Company, Inc.
195 Montgue Street
Brooklyn, New York 11202
(Fresh Kills Landfill)
Douglas Nielsor
Public Service Electric & Gas Company
of Hew Jersey
80 Park Place
Newark, Hew Jersey 07101
(Clnnaminson Landfill)
Joseph Seruto
Watson Energy Systems, Inc.
Suite 207
22010 S. Wilmington Avenue
Carson, California 90745
(Ascon Landfill)
Ralph Rule
Southeastern Portland Cement Company
3055 Uilshire Boulevard
Los Angeles, California 9CC10
(Azusa Landfill)
Robert Stearns
SCS Engineers
401' Long Beach Boulevard
Long Beach, California 90807
(Industry Landfill
Dr. Charles Kcare
Department of Civil Engineering
Ohio State University
2070 Neil Avenue
Columbus, Ohio 43210
(Methane Generation Rates)
-------
A REVIEW OF EPA-SUPPORTED RESEARCH ON PTTROUTIC OILS
Charles J. Rogers
D. S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
Summaries are given of two ETA-supported research projects to characterize pyrolytic
oils. One project alms to separate pyrolytic oils Into fractions that each contain a
broad class of chemicals, and the other project Is directed toward Identifying the speci-
fic chemicals within chose fractions. Separation (extraction) techniques show the grpatest
promise for processing the oil Into the large fractions, and the distillation process ap-
pears to be the best method for processing an oil fraction into more highly refined and
purified products.
INTRODUCTION
This report summarizes the results of
two EPA-Bupported research projects on py-
rolytic oils. One study was being conducted
at Geotgla Institute of Technology (EPA 600/
2-28-122) to develop methods for the separa-
tion of pyrolytic oils Into broad classes of
chenlc&ls (I), and the other Is at Atlanta
University (Grant #RS06440-02) to identify
specific chemicals that make up pyrolytic
oils (2).
Large quantities of agricultural, for-
estry, and municipal wastes are produced
each year in the United States. The proper
utilization of these materials is extremely
important if they arc to become a resource
rather than create disposal and environmen-
tal problems. Fyrolysls is one approach to
UBlng these materials that has received a
great deal of attention in the past several
years. Pvrolysis of llgnocellulosic or
celluloslc iraterlal produces char, pyroly-
tic oil, water-contalning/water-aoluble or-
ganic substances, and noncondenslble gases.
The char is primarily carbon that can be
used as a fuel or converted to activated
carbon. The major components of the noncon-
denslble gases are hydrogen, carbon aonox-
ide, and methane along with slnor amounts
of the other hydrocarbon gases. These
gases can be used onslte as a clean burning,
low-BTU, gaseous fuel. The pyrolytic oils
are clean burning, with heating values that
are approximately two-thirds of those for
fuel oils. In addicion to their value as
fuels, these oils have great potential as
a source of industrial chemicals or as a
chemical feedstock. Such uses may have
greater economic value than the use of these
oils as fuels. Also, the use of pyrolytic
oils as a souice of chemicals would reduce
the demand for tne petroleum now used to
malte chemical feedstock. To realize the
potential of pyrolytic oils au a source of
chemicals, the processing technology must
be developed to produce refined fractions
of the oils.
Comparison of the CeorRla Pyrolvsts Piocesa
with the Original Wood Distillation Process
Fyrolysls is an old process that has
been used industrially in the past on a
batch basis to produce charcoal, pyroligne-
ous liquor (mostly water with dissolved or-
ganic compounds), Insoluble tars, and non-
condenslble gases. The process was used
during and after World War I in this country,
and It was known as wood distillation. With
the use of petroleum as a chemical feedstock,
however, the pyrolysis process became uneco-
nomical and U3S discontinued. Various as-
pects of wood distillation and Its products
have been discussed lr. the literature
(3, 4 and 5).
The distillation of wood was generally
carried out as a batch process in a retort
1(6
-------
with external heat. The significant and
Important difference between the Georgia
Institute of Technology pyrolysls process
(1) and the old wood distillation process
is that the Georgia process is self-sus-
tained and continuous. The difference is
significant because the pyrolytic oil pro-
duced in this manner from a given feed
material under specific operating conditions
Is a reproducible produc with definite phy-
sical and chemical properties. The product
therefore h«:s potential as a chemical feed-
stock Cor making other- products an a com-
mercial scale.
Potential Yield and Demand
for Pyrolytic Oils
Pyrolytic oil from various wastes re-
presents a potential feedstock for the
chemical industry. Of all oil consumed in
the United States, about 6Z (or 50 million
metric Ions annually) is used to derive
feedstock for the chemical industry (6).
The yield of pyrolytic oil frcm lignocellu-
losic materials, varies from 15Z to 25Z,
depending on feed materials and operating
conditions (7). Consequently, it would re-
quire 181 to 299 million metric tons of dry
lignocellulosic material to supply pyrolltlc
oil In quantities comparable to the petro-
leum used by the chemical Industry. This
observation does not i*aply that pyrolytic
oil would be processed In the same manner
as petroleum feedstock or that 1 ton of py-
rolytic oil is equivalent on a feedstock
basis Co 1 ton of petroleum.
Accurate estimates of wastes produced
by different source? are difficult to obtain.
Based on Inquiries (particularly with the
U.S. Forest Service), the amount of fores-
try waste in the United States is estimated
to be 91 million metric dry tons annually
(Heyvood T. Taylor, Forest Service, Private
Conmunicatlon, June 1976). This quantity
could yield pyrolytic oil tonnage equal to
33Z to 50X of the petroleum tonnage now used
by the chemical Industry. The significance
of these data is that the potential exists
for pyrolytic oil f rot forestry wastes alone
to sake a significant contribution as a
source of chemical feedstock. Anderson In
197.2 (8) estimated that 1.1 billion barrels
of oil potentially could be produced from
the total organic wastes gererated each
year In the Uni ed States. Tillman (9) has
recently reported that approximately 1 bil- '
lion dry tons of cull or rough trees and
solvable dead trees presently exist in the
United States. The important fact is that
large quantities of vasce materials exist
that have the potential of being converted
to pyrolytic oils to help meet the Nation's
energy needs.
Analysis and Characterization of
Pyrolytic Oils
The chemical and physical properties
of pyrolytic oils havs been determined by
standard analytical techniques. The oils
are dark brown to black with a burnt, pun-
gent odor. Their boiling range is about
100° to 200°C, above which thermal degra-
dation beings to occur. The heating values
of these clean-burning oils are approximate-
ly two-thirds of those for petroleum fuel
oil. Pyrolytic oils are acidic and exhi-
bit some corrosive characteristics. They
are composed of many oxygenated compounds
with a wide spectrum of chemical function-
ality.. Results of this study showed pyro-
lytic oils to contain phenolice, polyhy-
droxy neutral compounds, neutral compounds
of a high degree of aromaticlty, aid vola-
tile acidic compounds.
Because the oils obtained froji the
pyrclysls of llgnocellulosl*. materials are
complex mixtures of organic compoundJ and
usually contain some water, their charac-
terization requires a variety of analytical
and testing techniques. Properties that
are of inters- j., characterizing pyrolytic
oils include (but ar<; not necessarily limi-
ted co) density, water rontent, heating
value, acidity, flash point, corrosiveness,
filterable solids, ash, solubility in.
various solvents, distillation range, vis-
cosity and elemental c mtent (particularly
carbon, hydrogen, nitrogen, sulfur, and
oxygen) (see Table 1).
Sources of Oil Samples
Samples of pyrolytic oils for this
study were obtained from two tuijor sources:
(1) the 45 dry metric ton/day field demon-
stration pyrolysls facility at the pilot
plant of the Engineering Experiment Sta-
tion, George Institute of Technology, which
In operated on campus. Some Samples of
oil were produced in a 6-in. tube furnace
fitted with a condensation train and gas
collection system. A complete description
of this apparatus and the pyrolysls proce-
dure has been reported on by Knight (1).
in
-------
1ABl> I PRIVlHIlHSOf WOOD OJI.S FROM IU H AIR «« U«Y TON D«Y FAr||jn

Dr«(t P«b Oil
taMtiy
llittr conceal
(wltht t|
laartti v«lw
(w*l 6«lla)
At 14 e
rusb relet
Ml(•ritlt 9olt4»
IvtttM X)
Coppri btrlp
Corram I on
(wl|M t)
four fntit
teb	Z)
Mm lUation
nr*c OtOP
1J1 Peine
U! EaApelnt
>11 Wpolat
ScloblUt*
(wrifbt T)
Ufim
i.ni it(• 1.10'
«
(«.I«0 Stu/lb)
J\ m tOHJt
ill r
HllV»
Sk.Vt
(>o*n
WC
lore
24.* ri/14
uo.wo iiu/)b>
i:rc
?*.rc
itC'H
101'C
!0>*C
.'dS'i
isth n *s->o
AaIW 0 H0-4A
IX m 1 ditpiriM
In 
0.#
SlUhtly
Slfffell*
(Vt
i.e
Methods for tdentyfying Chemical
Species and Compounds
The chemical species and compounds
present must be identified along with rela-
tive quantities If methods are to be devel-
oped for using pyrolytic oils In applica-
tions other than as a fuel oil. Among the
most useful techniques for obtaining these
data are gas, thin-layer, and liquid chro-
matography (LC); gas chromatography)/mass
spectroscopy (GC/MS); and Infrared and UV
spectroscopy. The two methods used In the
studies described here were LC and GC/MS.
Liquid Chromatography
Pyrolytic oils are heat sensitive and
reactive, and they contain a relatively
large number of organic compounds. A tech-
nique was needed for analyzing the oil frac-
tions obtained by the dif'erent processing
uethods *hat would net change the chemical
character of the fractions. LC appears to
be the oxsthocl of choice, because it Is car-
ried out it ambient temperatures, it is
capable 01 high resolution of cmplex mix-
tures, anl component detection is nondes-
tructive. In addition, the pyrolytlc oils
are soluble in organic-aqueous aolvent sys-
tems that are very useful In LC.
The variables that were studied to find
satisfactory LC conditions were LC columns,
UV wave length, solvent gradient ant* sol-
vent flow rate. The first two variables are
discussed briefly here.
LC Columns
To select the most,suitable LC coluran,
several columns were tested with the raw
wood oil using a 1-ml/mln flow rate and a
UV detector at 254 run. The chromatographic
columns and conditions tested and'the re-
sults are given below in the order In which
the testing was carried out.
A. Vydac adsorption silica gel 30 .
column. Solvent, OX to 1001 2-propanol in
lsooctane, 20-mln gradient, 20 concave.
Results; No resolution obtained; only one
large peak.
8, Partlsil adsorption silica gel S
column. Solvent, 51 to 30* 2-propanol in
lsooctane, 20-mln gradient, linear.
Results: Resolution of only eight peaks.
188
-------
C.	Partlsll PAC 5 column. Solvent,
OZ to 1001 2-propanol In ltiooctsne, 30-min
gradient. 35 concave. Results: Resolu-
tion of 12 to 20 peaks.
D.	Partlsll ODS S column. Solvent,
101 to 100Z acetonltrlle In water, 30-min
gradient, 35 concave. Results: Resolution
of 30 to 40 peaks
E.	Partlsll 0QS 5 coluim. Solvent,
10Z to 100% acetonltrlle In water, 10Z to
402 with 20-min hold, then 402 to 100Z 35
concave gradient. Results* Resolution of
47 to 50 peaks. Total run tine: 60 mln.
F.	Partlsll ODS 5 column. Solvent,
10Z to 100Z acetonltrlle in water, 30-oin
linear gradient. Results: Better overall
presentation of chroraatogram and better
resolution of later peaks without excess-
ive run time.
The resolutions obtained with the
conditions given in item D above are very
suitable for our survey chromatograms; the
conditions in E and F yield even greater
resolutions.
UV Wavelength
UV wavelength of 200, 220, 254, 280,
300, 320, 360 nm were selected, and LC runs
¦were made using constant conditions (G
above) other than wavelength. The results
were as follows:
a.	Many component responses appeared
or. disappeared with the change in wave-
length;.
b.	Ho single wavelength was entirely
satisfactory at'the shorter wavelengths of
200-220 nm peak resolution;
c.	The longer wavelengths of 300 to
360 urn produced sharply resolved peaks,
but only a small total number of peaks
actually appeared;
d.	The most satisfactory results for
our purpose were attained at 280 nm, with
254 nm being the alternative choice.
The sets of LC conditions for obtain-
ing chromatograras of the oil samples are
given In Polk (2).
Gas Chromatography/Maaa Spectrometry
Qualitative analysis of pyrolytlc oil
distillate by CC/MS analysis was performed
on s DuPont Model 21-490 GC Mass Spectro-
meter. The GC conditions vere as follows:
Column ... 6ft* 1/8 in. Se-30 Stain-
less steel with 10 ml/mln
flow rate
He initial temperature 	 -25°C
Programming rate 	 4°C/mir
Pinal temperature 	 -200°C,
detector flame ionization
Some of the mass spectral results are
listed in Table 2. Details of the quali-
tative analysis of pyrolytlc oils are re-
ported by Polk (2).
Techniques for Separating larae
Oil Fractions
The broad classes of chemical sub-
stances In raw pyrolysis oil are phenollcs,
aromatic neutral compounds (neutrals of
high aroeatlcity, NHA), acidic compounds,
and a group of substances with sugar-like
characteristics that are termed polyhydroxy
neutral compounds (PNC). The emphasis in
the separation experiments has therefore
been to focus on obtaining fractions of the
oil that contain essentially one of the
general classes of chemicals. The five
major extraction techniques tested were:
1.	Extraction o£ oil sequentially
with water at 25°, 50° and 95 C.
2.	Extraction of oil with sodium sul-
fate solution .(saltlng-out effect).
3.	Extraction of oil simultaneously
with an organic colvent and water
(three-phase system).
4.	Extraction of sodium hydroxide
soluble fractions of pyrolysis oil.
5.	Extraction of organic solvent so-
lutions of pyrolysis oil with water.
Results of a number of extraction and
separation experiments on a batch basis with
raw and vacuum-stripped pyrolytlc oils,¦
189
-------
TUU* Gf/HJOF CONDfMfcl OIL hCSTILLAIt
•/• tR*uti«« ittw«nr V
fiktao), F«rfuj*J.
KetaUeepfe
ruwi
Iucm. Hapuael
n««i
I -Oc taaoi
•<*M p-Crvsoli
Auiuol
Vtnirtlt
0MUri(*Mei.
MfKUIIIK
X*pktk4-
lm, :,!•
DuaifertpfcMftt


*«*). U. 4®(4J.
W<4). 94(471. M(MI. 9'(Ji
»(«$]. ttim. i|(IM). 4M4.M UfHK
tJlWJ. SMll. «D1
»|4*l.	41(100), 4»(in. SS<50I.
•t{si. tin), ioocm. m<«>. n». m<»i.
4*441), 5J(»1. 941 SCO), »i(S)
4MT|>. ««<100). 41(13),.
"nr. toot4)
»fSM. 43(H). 41(91): 44tiM. SfcfN).
tri.H,. ?"<>4);
IIOCOI
40(MK 4i(M>. 44(14), U(IM.
U ;ui. Mull. tfl(IJ), 7TC'],
10*190). 104(1001. HO(I)
40tJ?). 41(41], 54(75). !:[»?).
109(100), 124(11*. I iM4|
JSM9). 40131). 4UIU0). 44|4V). M(UI.
9S(?»),	imii»
«0(M), <11*71. «(»3). 5«UI)
91(19). 1W7[90). UI(4J). U3U0OI.
l/l l?4)
29419) 40(!Ot. 41(31). 44(11) ,
«. 7!|U, i0J(e), 137(11). IM(IOS).
I»<9»
.^14^). 40ISI), 41{IJ]. 4.'(4J|. ^(rt).
??{J4), «IU*l)a IC'IIM). 1^(49).
US(«l, 147(5)). 141(11). J<9»1
404I9|. 41(111. 42(|S). S2(IIJ.
S4(7), S*(H). 61(10).	6*h?).
W4(fcJ, |K|1), lMtlM). 15119)
M;U), 40(j:>. 4l("|. 44(11 |, 7TJJJJ
91121'. IOIiJ?), |B»!»1. ItSfttl.
iM(lJ).	142(100). IS^(M)
Indicated that vaeuum-sttipped oil gave bet-
ter results than the row oils. The vacuus
stripping provides for the removal of the
volatile organlca and most of the vater in
the oil with potential subsequent recovery
of these orgaulc compounds. the analysis
shoved that the major organic component in
the volatile fraction ir acetic acid. Our
preliminary separation techniques are there-
fore based on using vacuum-stripped oil. A
100-g sample of crude pyrolysia oil would
yield 82.lg of vacuum-stripped oil, 10.8g
of water, and 7.1g of acid when processed
at 2nm and ambient temperature (see Figure
1).
Extraction of Oil Sequentially with Water
at 25°C, 50°C, and 95°C
A sample of vacuum-stripped oil was
extracted sequentially with water at 25°,
30° and 95°C in an effort to separate the
more vater soluble substances. Figure 2
illustrates this separation process and
the recovery of the different factions are
given in Table 3.
The significance of these results is
that the oil can be separated Into water
soluble and water Insoluble fractions that
offer the opportunity for recovery of use-
ful fractions of aromatic compounds. The
water Insoluble fractions, based on our
analysis, are composed of phenolic9 and
neutral aromatlcs. The separation of this
fraction into a highly concentrated pheno-
lic fraction and a highly concentrated
fraction of aromatic neutra.1. compounds
could probably be accomplished by either
fractional distillation or ex:ractlon with
alkaline solution. The aqueous phases
could be combined and subjected to a sepa-
ration of the components wltb an aqueous
Bait solution tc yield a fraction with-main-
ly phenolics and another fraction with main-
ly polyhydroxy neutral substances.
Other Extraction Techniques
Results from this evaluation of other
extraction techniques tested are presented
in Knight et al, (1).
Conclusions
The yield of pyrolytic oils from lig-
nocelluloslc materials vary from 155 to
25Z, depending on feed materials and opera-
tional conditions. -An estimated 1.1 bll.lion
barrels of oil could be produced from the
Crude PyrolyBls Oil. 100 g

Vacuum stripped at
2 mm and ambient
temperature



Vacuum Stripped Oil, 82.1 g
Volatile Fraction
Water 10.8 g
Acids 7.1 g
Figure 1. Removal of volatlles from pyrolytic oil.
190
-------
fmttmm *irtii-»7.n
for Continuous Processing. U. S. Envi-
ronmental Protection Agency, Cincinnati,
Ohio, 45268.
2.	Polk, K;D., 1977-. Development o£ Me-
thods for the Stablizotion of Pyrolytic
Oils. U. S. Environmental Protection
Agency, Cincinnati, Ohio 45268
3.	Kagglund, E., 1952. Chemistry of Wood.
Academic Press, ~nc., New York, N.Y.
4.	Hamn, A. J., and E. E. Harris, 1953.
Chemical Processing of Mood. Chemical
Publishing Co., Inc.," New York, N.Y.
ttWOT •« «'C-
*1wr
i kwntlw
5. Ravley, L. F.,.and L. E. Wise, 1926.
The Chemistry of Wood. Chemical Catalog
Co., New York, N.Y.

~l •
Sww yr»ctloo
rhrnollcs - cTR
F«lyf>Td ro»y o*utrii»o./Z
6. Farmer, R. H., 1967. Chemistry in the
Utilization of Wood. Pergamon Press,
Oxford.
Rpse : Euractn of «tf •qantaltii nl m(«t « .Yf, VT[, «ad oj*<\
total organic wasces generated annually in
the United States. The extraction techni-
ques reported here show promise of having
the greatest potential for processing py-
rolytic oils into fractions containing
specific classes of chemical coopounds.
Further study wlH be required to deter-
mine if any of the oil fractions cnn be
used as chemical feedstock lit industrial
applications.
7.	Hawley, L. A. 1923. Wood Distillation.
The Chemical Catalog Co., Inc., New
York, N.Y.
8.	' Anderson, L. L., 1972. Energy Potential
from Organic Wastes. -Circular 8549.
U. S. Department of the Interior.
9.	Tillman, D. A., 1978. Wood as an Energy
Source. Academls Press, Inc., New York,
H. Y.
REFERENCES
Knight, J.A., L.W. Elston, D.R. Hurst,
R.J. Kovsc, 1980. Pyrolytic 011s-
Characterizatlon and Data Development
TABLE 3.Y1ELDS OF FRACTIONS FROM WATER EXTRACTION OF OIL
Water Insoluble
Fraction
Water Soluble Fractions
25*
50°
95°
Total
Phenolles
Aromatic
neutrals
Poljrhydroxy
aiiutrols
Totals
10 g
39.2 g
49.2 8
6.7 g
1.7 g 0.6 g 9.0 g
15.4 g 5.9 g 2.6 g 23.9 g
22.1 g 7.6 g 3.2 g 32.9 g
191
-------
STANDARDS FOR REFUSE DERIVED MATERIALS
P. Aar"? Vesilirid
Department of Civil Engineering
' Duke University
Durham, North Carolina 27706
John Love, Jr.
Department of MerSanical & Aerospace Engineering
University of Missouri
Columbia, Missouri
ABSTRACT
The American Society for Testing and Materials, through a grant *rom the U. S. Environ-
mental Protection Agency, is developing standards and specifications frr the reclamation
of materials and energy from refuse. The need for such standards and four examples of
typical tasks in this project is discussed.
Introduction
The pressures for the Implementation
of resource recovery are Intensifying.
Sanitary landfills, long considered to be
the ultimate solution to the problem of
solid waste disposal are frequently Indict-
ed as sources of ground water pollution
and often represent a necessary nuisance
to the community. Further, tne cost of
land close to the centers of solid waste
production are forcing new landfi:is fur-
ther away from communities, thus increas-
ing transportation costs. In some com-
munities, there simply are no further
landfill sites, and alternate methods of
solid waste disposal are needed.
One historically important alterna-
tive is incineration, or thermal volume
reduction. Due to the extremely strict
air emission standards now required by
many states, and increasing value of
energy from ever decreaslngly reliable
sources, the use of incinerators only
as volume reduction processes has become
economically questionable.
The only other alternative to the
landfill 1s the recovery of energy and
materials from municipal solid waste.
This option not only provides a solution
to the problem of disposal of refuse, but
has the added advantage of providing
secondary materials for industry so as to
preserve our virgin sources and nlake us
less dependent on foreign supplies. In
addition, resource recovery contributes
useful energy to our increasingly energy-
limited society. In short, resource
recovery solves two problems.
•	disposal of sclid wastes
•	savings in energy and materials.
The promotion of this alternative is
thus in all of our interest.
192
-------
The Need for Standards
Although the secondary materials
industry is an old one, the large scale
recovery of metals, glass and organic
fractions from municipal refuse is a
fairly new problem.* The three major sec-
tors in this process—the public, the re-
fuse processor, and the user of the re-
covered materials—all have conflicting
interests. The public simply wants the
refuse to disappear since it (by defini-
tion) no longer has value to the individ-
ual. The processor would like to have
the ^blic perform separation of the waste
materials, but since this represents a
bother'to the public, such source separa-
tion programs, with a few notable! except-
ions, have not caught on.
The refuse processor thus must sepa-
rate the mixed components into individual
fractions. His objective, however, is to
produce as dirty a product as possible and
still be able to market it to secondary
materials users. . Finally, the purchaser
of these materials is seeking maximum
purity since pure materials require the
least cost for processing and subsequent
manufacturing, in short, the three major
groups involved with resource recovery all
have different objectives, and yet need
each other in order to make resource re-
covery successful.
One means of promoting and facilita-
ting such cooperation is to develop
standards for refjse derived materials
and resource recovery processes. Quanti-
tative and well-defined definitions of
secondary materials, properties and the
processes used to produce them will act
as a catalyst in the promotion of resource
recovery—to the benefit of all cf us.-
This recognition led the U. S.
Environmental Protection Agency Office of
Research and Development to fund a program
to assist *he American Society for Testing
and Hateri.iis in the development of stand-
ards for refuse derived materials. This
project, administered through ASTM's E-38
*0ne of the fir't la*ge scale resource
recovery prjj,.cts vas started by Colonel
Waring, New York City, in 1880. The
facility reportedly paid for itself In
recovered materials, but was nevertheless
closed within a year.
Committee on Resource Recovery, has been
ongoing for four years and has signifi-
cantly assisted in the promulgation of
such standards. The objective of ihis
paper is to rerort on the progress of
this program.
Organization
E-38 Committee on Resource Recovery
is organized into a number of subccrrit-
tees. In,addition to such service sub-
committees as editorial, terminology, and
research and planning, the li.ie subcom-
mittees (those charged with the deve'op
ment of standards) include the following:
E-38.01 Energy
E-38.02 Ferrous Metals
E-38.03 Non-ferrous Metals
Et38.0". Paper and Paperboard Products-
E-38.05 Glass
E-38.06 Construction Materials
E-38.07 Health and Safety
E-38.08 Processes and Unit Operations
These subcommittees meet bi-annually
and by the process of consensus and com-
promise, develop the standards as needed. -
As a need for background information
such as round robin testing and commercial
laboratory analysis are identified, re-
quests are made to the Et38 Management
Board to assist in the funding of these
tasks. If the need is judged to promote
the development of standards, the task is
funded.
At the present time E-38 has 21 stan-
dards in various stages of approval and
development, of whicn all are financially
assisted through this project. Ter. of
these standards will be published in the
near future.
In this short presentation It is
obviously not practical to discuss all of
the ongoing projects. We thus have chosen
only four tasks to illustrate the type of
work being performed and the potential
193
-------
benefit from such studies.
Four Typical Tasks
Development of Test Methods and Specifi-
cations for Recovered Glass in Structural
Clay Products
This project is under the guidance
of E-38.0? Subcommittee on Construction
Materials. The objective is to develop
test methods and specifications for vari-
ous resource recovereo glass-rich frac-
tions for use in structural clay products.
This can be achieved by determining
chemical and physical properties of glass-
rich fractions as related to their effects
on maturing temperatures, shrinkage, water
absorption and compressive strength. The
properties which require specifications
include, for example, glass content, solu-
ble salts, and fluxing effect.
The program involved a two-step ef-
fort: laboratory studies and production
testing. The first phase of the work was
successfully completed and the results
published in a report "Development of Test
Methods and Specifications for Resource-
Recovered Glass in Structural Clay Pro-
ducts", presented at the I1TR1 Conference
in August 1979. The next phase of the
study is now on-goinq.
Structural clay products encompass a
wide range of materials such as building
brick, roofing tile and pipe. The build-
ing brick industry produces the largest
volume of structural clay products. The
initial commerical application of the
proposed test methods and specifications
are therefore to be tested in the pro-
duction of building brick.
This phase involved three phases:
1) characterization of the glass-rich
fraction by Bureau of Mines, 2) brick
production tests to be performed by a
manufacturer, and 3) testing by a com-
mercial laboratory.
The Effect of Aluminum and Iron on the
Heating Value of Refuse Derived Fuel
This task originates in E-3B.01 Sub-
committee on Energy, and addresses the
problem of obtaining heating values for
a refuse derived fuel which contains me-
tals such asaluninum and iron. If ^
sample of RDF with these metals is
combusted in a bomb calorimeter (the
coirmon method of measuring heating value
of solids.) the results will include the
heat generated by the exothermic oxidation
of the metals. This oxidation, however,
does not occur at the much lower tempera-
tures experienced in RDF combustion
chambers, and the calorific values can
thus overestimate the jctual heating v«.lue
achieved in the combustion of RDF.
The objective of this task is to con-
duct round robin tests to establish what
correction factors, if any, are necessary
for describing the useful heating value
of RDF when it contains significant
amounts of aluminum and iron.
Method for Preparing a Gross Laboratory
Sanple from a Gross Lot of Refuse Derived
Fuel
Refuse is a nighly heterogeneous
material, and thus any RDF produced from
. refuse can be expected to retain consid-
erable heterogeneity. A shredded and air
classified light fraction (comtonly known
as RDF-3) can have pieces of paper on pla-
stics as large as 5 cm ( 2 in). Labora-
tory samples for such analyses as heating
value must, however, be as snail as one
gram. The statistical problems involved
in obtaining a representative sample are
thus immense.
This task, under the leadership cf
E-38.01 Subcommittee on Energy, addresses
the problem of how a truly representative
sample can be obtained from a highly
heterogeneous mixture such as RDF-3. The
test hes been developed, and the task is
presently engaged in a series of round
robin analyses to develop its precision
. and accuracy.
Development of a Test for Measuring Refuse
Derived Fuel Pellet Characteristics
The organic fraction in refuse,
separated out and used as a source of
energy, can take rr^ny forms, depending on
the processes used. One potentially use-
ful fuel is pelletized or densified refuse
derived fuel (dRDF). The organic fraction
is first separated out using processes
such as shredding and air classification.
The material is next processed in a pelle-
tizer which, under high pressure, extrudes
or cuts the ROF into dense chunks. These
pellets have significant advantages over
194
-------
other forms of RDF, particularly If the
fuel Is to be burned on grates. Pelleti-
zlng also enhances the ease of handling
and increases the fuel stability.
The objective of this task is no
develop and evaluate a test for determi-
ning the characteristics of pellets, in-
cluding physical properties and pellet
integrity.
The E-38 subcomnittee on Processes
and Unit Operations is developing a
standard for evaluating the characteris-
tics of pellets as well as pelletizer
performance. In this task, a series of
appropriate tests for measuring such
characteristics are being developed.
This project is in three phases:
a)	review of existing information on
pellet tests and testing procedures,
b)	development of specific test procedures
and precision and accuracy statements,
and c) round robin of pellet tests, using
the test procedures developed.
Conclusion
The promotion of resource recovery is
a well-established public benefit. Re-
source recovery can be facilitated by the
development of standards for both the
production as well os use of the secondary
materials. The American Society for Test-
ing and Materials through a grant from
the U. S. Environmental Protection Agency,
is developing such standards using a
broad-based voluntary consensus mechanism.
Significant progress has occurred to date,
and the project is continuing. The E-38
Committee on Resource Recovery is a purely
voluntary organization, open to all
interested persons. Assistance and parti-
cipation is welcomed and encouraged.
Acknowledgement
This project is supported by a grant
from the U. S. Environmental Protection
Agency EPA 68-03-2528 Carlton Wiles, Pro-
ject Officer. Donald Mihelich is the
chairman of E-38 and Pickett Scott is the
chairman cf the Management Board which
oversees the project in behalf of ASTM.
J9S
-------
ENVIRONMENTAL ASSESSMENTS OF WASTE-TO-ENERGY CONVERSION SYSTEMS
H. M. Freeman
Industrial Environmental Research Laboratory, U.S. EPA
Cincinnati, Ohio 45268
K. P. Ananth, M. A. Colcmbiewski
Midwest: Research Institute
Kansas City, Missouri 45268
ABSTRACT
This paper discusses the re&tilts of environmental assessments at five different
types of vaste-to-energy conversion cystems; refuse pyrolysis; municipal incinerator fired
with HSU; power plant boiler fired with wood waste and fuel oil; steam boiler fired vith
coal and d-RDF; and power boiler fired with RDF.
Introduction
Increased emphasis on energy and Date-
rial recovery and the need for alternatives
to solid waste disposal in Landfills have
generated growing interest in waste-as-fuel
processes. The processes include, on a ge-
neric basis, watervall incinerators, pyrol-
ysis systems, combined fuel-fired systems
(coal plus refuse derived fuel (RDF), RDF
plus municipal sewage sludge, coal plus
wood waste), and biochemical conversion of
waste to methane.
For thepast three years, Environmen-
tal Protection Agency's (EPA's) Industrial
Environmental Research Laboratory (IERL)
in Cincinnati has sponsored a program at
Midwest Research Institute (KR1) to conduct
environmental assessments of some ol the
above waste'to-energy conversion" processes.
The overall objective of this program is
to evaluate the potential multi-media en-
vironmental impacts resulting from using
combustible waste as an energy source and
thereby identify control technology needs.
As part of this program, KRI has undertaken
extensive sampling and analysis efforts at
the following waste conversion facilities.
K 181.4 tig/day (200 ton/day)
refuse pyrolysis. system
A 108.8 Hg/day (120 ton/day)
municipal incinerator fired with
municipal solid waste (MSV)
A 10 KW power plant boiler fired
with wood waste and No. 2 oil
A 31,752 kg/h (70,000 lb/h) steam
boiler fired with coal aud densi-
fied refuse-derived fuel (d-RDF)
A 20 MV power plant boiler fired
with RDF
A description of tbe facility, the
sampling and analysis methods used, and the
results obtained are individually presented
below for each of the above facilities
tested.
Refuse Pyrolysis System
The Union Carbide refuse pyrolysis
system (PUROX) at South Charleston,
West Virginia, was designed to pyrolyze
181.A Hg/day (200 tons/day) of RDF. The
refuse fuel was produced by shredding MSW
to a 7.6 cm (3 in.) size and then removing
magnetic materials from the shredded waste.
The PUROX system is a partial oxidation
process that uses oxygen to convert solid
wastes into a gas having a higher beating
value (HHV) of about 14.5'MJ/mJ (370 Btu/
scf).
Raw refuse is received by truck in
tbe plant's storage building. It is moved
and stacked in tbe storage area by a front
end loader. Tbe same loader picks up the
stored waste, veighs it on a platform, and
-------
dunps it on a ronveyor leading Lo the
shredder, where it is shredded Lo a 7.6 cm
(3 in.) size. Ferrous materia] is removed
by a magi.elic recovery system.
The refuse fuel is fed into the top of
the reactor, the principal unit on the pro-
cess, by two hydraulic rams. There are
three general zones of reaction within the
reactor (tirying, pyrolysis, and combus-
tion). The reactor is maintained essen-
tially full of refuse, which slowly de-
scends by gravity from the drying zone
through the pyrolysis zone iuto the com-
bustion zone. A counterflow of hot gases,
rising from the combustion zone at the hot- .
ton, dries the incoming, moist refuse. As
the material progresses downward it is py-
rolyzed to form fuel gas, char, and organic
liquids.
Oxygen is injected into the bottom
hearth section at a ratio of about 20% by
weight of incoming refuse. The oxygen re-
acts with char formed from the refuse to
generate temperatures of 1370 to 16jO°C
(2495 to 3002°F) in the lower zone, which
converts the noncombustihles into a molten
residue. This residue is discharged into a
water quench tank where it forms a slag.
The hat gases from the hearth section
are cooled as they rise through the zones
of the reactor. After leaving the reactor,
the gases are. passed through a recirculat-
ing water scrubber. Entrained solids are
separated from the scrubber water in a
solid-liquid separator, and recycled to the
reactor for disposal. The water product
discharged from the separator system is
sent to a plai.t treatment system. The gas
leaving the scrubDer is further cleaned in
an electrostatic precipitator (ESP) and
then cooled in a heal exchanger prior to
combustion in a flare' coobustor. During
the tests the gas was burned in a package
toiler.transported to the site for these
tests. The fuel gas consisted of about 40%
CO by volume, 23% CO , 5% CH^, 26% H , and
the rest being Nj, C^H^, etc.
Sampling at the Purox facility was di-
rected to the three effluent streams; slag,
scrubber effluent, and gaseous emissions
from a boiler when fired with PlIROX gas and
when fired with natural gas. As overview
of the sampling and analysis scheme is
shown in Figure 1. As car be seen in this
figure, sampling and analysis of each
streca was ratber complex, being directed
to conventional pollutants but including,
among others, priority pollutants in water
samples and sampling of both liquid and
gaseous emissions for most of the analyses
prescribed under EPA's Level 1 environmen-
tal assessment protocol. Particulate emis-
sion sampling in the boiler stack was
conducted according lo EPA Method 5, but
usiug a High Volume Sampling System (HVSS)
because of the expected low particulate
loading. Boiler slack sampling also in-
cluded use of the Level 1 Source Assess-
ment Sampling System (SASS1 train.
Water samples also underwent analysis
for priority pollutants, but the data are
too lengthy for inclusion in this paper.
The results of these analyses showed that
few of these pollutants were present aL
detectable levels in the scrubber effluent,
but that the Unox system did effectively
reduce their concentrations.
Results of the testing effort showed
that, of the criteria pollutants, only NO
and particulate emissions increased when
burning Purox gas as compared to natural
gas. NO and particulate levels were of
the order of 350 to 400 ppm and 0.0046 to
0.011 g/scm (0.002 to 0.005 gr/scf), re-
spectively.' St>2 emissions averaged 70 to
100 ppm. Particulate and SO^ emissions
¦iere below present standards, whereas NO
required further reduction. Also, analy-
sis for metals and other pollutants indi-
cated that these should not present any
problems.
Because of the difficulty involved in
interpreting ouch of the data collected in
this test, especially the Level 1 analysis
results, the environmental .assessment work
was extended to include application of the
methodology known as tbe Source Analysis
Model (SAM/1A) developed by EPA. Basi-
cally, this model compares the measured
concentrations of pollutants with approxi-
mate emission concentration guidelines
known as minimum acute toxicity effluents
(HATE) values. These KATE values have been
tabulated for several compounds or classes
and there is a specific HATE concentration
for each compound and for each type of ef-
fluent stream (solid, liquid, or gaseous).
The HATE values are used to compute the
ratio of the measured concentration to the
HATE concentration, and this ratio is
termed the "degree of hazard." The "de-
gree of hazard" for each pollutant is then
suned to provide tbe "degree of hazard"
for the effluent stream under considera-
tion. This value, when multiplied by the
197
-------
M	
t lowotoi
D*l#f4»ii«d 1>y	Mulo't
|U^ut( & I* wlol rtio*n(4)
lot# 4 gitd> lonflfi ot I fi'. Il*«
«ii & qualtt to ebteln OfV
Aml/tlii HjO.f'o*. &UUIevt« IBIV ml.
Mr'ali (I2>y
Anl^m (CI. f. !•)
PCI'i
WMS
I Ivwiulfl
• / llo»»i«4»*/ lo ta imhI lit Irfltl
o**r« *tolri)it ®.
toVpUl loiari It. twitualci)
ImjJU; 2 p«i ilof.	lot
m|of A Miv ynwt ty (. - Mwiliad 7	dn, )
fotni prMflcu'u'r
ol 0.5 tin tm 6 lui.
enljr), Wrljli & SSMS On,il^l*l,
plui AA ouilji"*''/!®' Ilixftali^'
V/
t*J Mlnw A iKHf'nloll \>y AA enolytlt
pj. fp. A. »fc. n,. c., Co. Mn. Cd, HI. *«)
b/ AoilpJi pi>il««i»wd anip |l tunpt* qmntlly •« •ulfiitrnl
AtMHfJ*«*
$k.«fcfed
_ 1 w(fro|y
JAM & fci'i
SSMS
rWiatii	isV*4l'(lani l»»tl
Sjb *	0» ill*
Soiipllng: | liAuf iffiii Kmpl«| dvting Ititlng,
AfJcntfr.r ,t'\Sf)4 )
Moilljr	{ fAlt. ft ft,
**luli, ft*illflJrt. He.) &
t*»rl I ofi n»H"
*«~•» OuQ|Jlr (600. COD. DO.
ISS. luibSd'tp pll.
ell & 9*««rW )
JL
I loortQl*; film liavftif .n po*t o' IFVSS (Mithod ))
^iirflliif
A Ar«al pili- a. C o»i|rK«t>ti «»olj'i*>i - ()J, NO» , IIC« CO
(HVJv 6)
b. Opo<)*y • Meiled 9
e , AlJ»|t^d«f ((. sikcn^li ) * 0.5 ilr* In IS min.
(itAC Iiitin), "? pn doy
d M#irury I.liit (ilg ) >7 p«« iinj», Alu . OnolpI*
» OM H l"»r» '« C{'. f\ tt', CM*
». O.tai - Cy. ro?. CO 
-------
efllueul fiovrdle, ill specific units (e.g.,
liters ,itr second), establishes the "tonic
unit discharge rate" (TUDR) foe the stream.
The SAH/1A methodology, as described
above, was utilized to analyze the data ob-
tained for each of the three primary efflu-
ent streams from the furox process (slag,
scrubber effluent, and boiler stack gas;.
Based on the SAH/IA methodology, the scrub-
ber effluent had the highest "degree of
hazard," being considerably greater than
the "degree of hazard" for the input river
water. - However, the slag stream had the
highest "toxic unit discharge rate." The
bailer flue gas effluent had the lovest
"degree of hazard" and- the lowest "toxic
unit discharge rate." Both of these values
were comparable to the baseline values com-
puted for boiler flue gas vhes burning
natural gas.
Municipal Incinerator Fired With MSW
The Braintree municipal incinerator
(Braintree, Massachusetts), is a mass-burn
facility consisting of twin vatervall com-
bustion units, each with a design capacity
of 108.8 Mg (120 ions) of MSV for a 24-h
period.- A portion of the steam produced
(20 to 35%} is supplied to neighboring man-
ufacturers and the remainder is condensed.
Each furnace is equipped with an ESP and
both ESP's exhaust to a comoa stack..
The Riley Stoker boilers are of the
single pass design, each having a rated ca-
pacity of 13,608 kg (30,000 lb) of steaa/h
at 204.4°C (400°F) and 1,723.7 kPa (250
psig). The ESP units ire single field, 12
passage precipitators with a specific col-
lection area of 0.0068 o2/acah (125 ft2/
1,000 acfa); each has a design collection
efficiency of 93%.
Eovironoental assessment of the incin-
erator facility was conducted using EPA
approved sacrpliag and analysis procedures
similar to these identified in Figure !.
Results and conclusions of the testing ef-
fort at*- summarized below.
Of the criteria pollutants, SO^, H0 ,
and hydrocarbon emissions were low. How-
ever, CO levels vere high and could not be
explained considering the large quantities,
of excess air that were used. The average
particulate concentration was 0.55 g/dscn
(0.24 gr/dscf), corrected to 121 CO.. This
levrl exceeded the federal and state
regulations. However, subsequent tests for
compliance had an outlet particulate load-
ing of 0.17 g/dsca (0.074 gr/dscf), which
showed coapliance.
Elemental analysis of the glass- and
metal-free bottom ash revealed an overall
increase in the elemental concentrations
when compares to me reluse feed. The col-
lected fly ash contained levels of chlo-
rides, sulfates, and some trace metals
which nay be of concern. PCB's were not
detected in the collected fly ash; 4 PAH
co-pounds were identified.
Levels of BOD, COD, oil. and grease,
TS5, and TDS in tbe bottom ash quench water
ao not appear to be of concern. The phen-
olic content was found to be <0.1 og/liter
in all samples.
Levels of gaseous chlorides and other
haiides were low. Presence of PCB's was
confirmed only in the OASS train XAD-2
resin at .a concentration of 3-6 pg/a*.
Results of tbe &AH/1A environmental
assessment procedure showed the incinerator
stack emissions to have the highest appar-
ent degree of health hazard. . Further anal-
ysis is needed to determine the exact com-
position of the organic cooponents of the.
stack emissions to better ascertain the
hazard poLenLial. SAM/1A also showed that
the bottoa ash effluent had the largest
toxic unit discharge rate due primarily to
tbe abundance of phosphorus and aetals in
this stresa.
Power Plant Boiler Fired With Woodwaste
and Fuel Oil
The No. I unit at the Burlington
Electric Plant (Burlington, Vermont) was
originally a coal-fired boiler which was
aodified to fire wood chips with supple-
Bent try No. 2 fuel oil. Because of the
high moisture content of tbe chips, the
boiler cannot provide the desired steam
output on vood alone. Therefore, No. 2
fuel oil is used. Stean production is
rated at 45,360 kg/h (100,000 Ib/h), which
powers a 10 HV turbine generator. Residual
ash froa the boiler is discharged at the
end of the grate into a hopper and is then .
pneunatically transported to an emission
control system consisting of two, high ef-
ficiency mechanical collectors in series.
For a flue gas flow rate of 101,940 acoh
(60,000 acfa) at 165.6®C (330°F), the
199
-------
Collet.Lors were designed lor an overall
pr?ss'ire drop of 1.6 kPa (6.5 in. H_0)
and a collection efficiency of 97.75V
Stapling and analysis was based on
the matrix shown in Figure 2. Major re-
sults and conclusions of the tests are as
follows:
On a heat input basis, wood accounted
for 801 of the boiler fuel, and oil the
remainder. The heat of coobustioo of wood
was 13.65 HJ/kg (5,870 Btu/lb) as received,
and for oil, the beat of combustion was
45.36 KJ/kg (19,500 Btu/lb).
Bottom ash analysis indicated that
most elements were core concentrated in
the asb relative to the input fuels. No
PCB's were detected in bottom ash but one
PAH compound, phetianthrene, was present at
a concentration of 0.89 Jig/g. Prioary and.
secondary collector asj contained no PCB's
but several PAH compounds were identified
in the secondary asb, with one sample con*
taining 10 (Jj/g of phenanthrene.
Particle sizing at the collector
inlet and outlet could not be established
due to constant plugging of the optical
counter's dilution system. Stack concen-
tration of particulates averaged 0.18
g/rtscm (0.08 gr/dscf) and the collector
had a particulate efficiency of 96.21.
NO,, and SO, concentrations averaged 66
and 138 ppo, respectively. CO averaged
213 ppo and hydrocarbons 9 ppm. Analysis
of Method 5 particulate indicated concen-
trations approaching 100 pg/dsca> for Pb,
Ba, Sr, Fe, and Ti in the stack gases.
PCB and PAH tests of the stack gases were
negative.
EPA's SAH/1A analysis indicated that
the secondary collector asb contained the
highest degree of hazard although all
three ash streams were similar in the oag-
nitude of their hazard values. Stack eais-
sions showed a low degree of hazard. The
prioary collector ash had the highest toxic
unit discharge rate.
Steam Boiler Fired With Coal and d-RDF
Emission tests were conducted on the
CSA/Pentagon facility's No. 4 boiler in
Arlington, Virginia during a test burn pro-
gram coordinated by the General Services
Adainistrati'u (CSA) and the National
Center for Resource Recovery (NCRR). The
No. 4 unit is an underteed-retort stoker
boiler with a rated steam capacity of
31,752 kg/h (70,000 Ib/h) at 861.9 kPa (125
psig) and 176.7°C (350°F). During the
tests, the boiler was equipped with a oul-
ticlone collector for removal of particu-
lates from the exhaust gases.
The test burn program i^r'wded three
fuel firing modes: 100% coal (baseline
conditions), 20% d-RDF + 80% coal, and 40%
d-RDF + 60% coal. Samples of coal, d-KDF,
and the coal/d-RDF mixtures were collected
hourly by NCRR and analyzed for moisture,
ash, heating value, and chemical composi-
tion. Several daily samples of bottom ash
were also collected by NCRR and analyzed
for loss-on-ignition and chemical composi-
tion. NRI condr ted sampling and analysis
of the stack effluent. Parameters measured
included particulate concentration, gaseoua
criteria pollutants (SO^, NO^, CO, and to-
tal hydrocarbons), and chlorides. The par-
ticulate samples were further analyzed for
lead content.
Results of the emissioa tests showed
that:
Particulate emissions were re-
duced from 22 to 38% when d-RDF
was blended with the original
coal fuel. Filterable particu-
late ecissions were lowest when
using the 20% d-RDF blend and
rose again when the proportion
of d-RDF was raised to 60%. This
finding may not be conclusive,
however, since the boiler load
was held steady duriag the 20%
RDF firing but not during the 60%
¦ode.
The amount of particulate lead
eaitted when burning d-RLF with
co»l is substantially higher than
that from combustion of coal
alone (an average of 1,000 pg/o3
with 20% d-RDF, and 2,260 jig/n1
with 60% d-RDF, versus 330 pg/o3
with coal only).
Chloride emissions showed no def-
inite trend which could be used
to correlate chloride eaissions
with RDF modes, though slightly
higher concentrations of HC1 were
observed in two of the samples
collected during combustion of
the 60% d-RDF blend.
200
-------
Wood r««d
SumpliiKji fuU Unas	SumpU*
foch Ogy. Mia ond Ulfocl
o l-Lil«r Composite.
Afftjlyiiti I)«l«rinin* I I/O Cotrtcnf
HIIV, Pio^inuli/Ullinitiii
LlvrnvtUul Anulpii by
VjMS
Oil fv*d		 	
Sunf)liiiui Tola One o* lwwO.5-
lil«rf Satnplvt
Anulytlt: MMV, PiobIdnjI*/Ullimai*
ll«n«iilal Analpit by
Bottom Adi
bunpliiiQi li^a flifaa I ky
tacit Day Mix and (itiQil
I ty CoiftfMtil*
Aftolyiit: tl«Mn-«iiul AjKjIytii by
KB/KAII
( ult«CtO( |lll*t
^umplliKj and Aiyiljrtiu
u. Metl»ad b Putliiululi - I P«r (Juy
lltimnlul AitalykU by
l>. O.wl (Oj & COj)
c. PMilivU	- I Plrl Ouy
Cotle«Jo< Outlet
Wiikrt poilti
fli.OOO ftn/la
Willing and Anqlytit;
it. Melliod 5Pailicul Je - I Ktx> P«i l>«iy
tl«meitlul Ajulyvt by SbMS*
b.	Oiuii {O; & COj )
c.	PiJtticIc	- I Run Pci Duy
d.	(Melliod V) • |wo I lit Velll/IViy
«. Co4itinuoui AjKilyicii (Oy. NOg,	IK.. CO)
I. PCB/ PAH willi Motitil liaiit - 2 Kuni
0. SAVj » I Ki>n
AiHilyti Per Lewi I R«.-(juiieffl«iilt
PiiatUfy &	y CoMecloi Adi
Sampling: foA** A I ky Gid>	liM.lt Ilom -
Mil & L«ha«.l I kij CouygiiU
Aiwl|tiii ^MS
p< b/PAii
Piinui/ jttomlxy
Cj(Ii i loi C olJrcloi
Ad.	AJ>
•Aiojiiu. ALtorpfioti AituJytit May tbr C
-------
Concentrations of sulfur dioxide,
sitrogen oxides, and carbon mon-
oxiJe all appeared to decrease
slightly when the RDF was used
with coal. Because of the very
low sulfur content of d-RDF. SO,,
emissions were reduced progres-*"
sively as the proportion of d-KDF
with coal was increased. How-
ever, the reduction in NO and CO
levels, may or may not have been
the direct result of burning
d-RDF since they are highly de-
pendent on boiler combustion
conditions.
Power Boiler Fired With RDF
The Hempstead Resource Recovery Plant
(Long Island, New York) receives NSW, pro-
duces a RDF and converts the fuel to elec-
trical power. The facility consists of
two distinct segments: a refuse process-
ing operation, utilizing the Black Clawson
Hydrasposal system; and a power house,
which contains two steam boilers and two
20 MW electrical turbine generators, plus
the associated control equipment.
Tests were conducted by KRI on the
No. 2 unit of the power house, vhich is an
air-swept spreader stoker, watervall boiler
with a nominal capacity of 9C,720 kg/h
(200,000 lb/h) of steam at 4,309.2 kPa (625
psig) and 398.9°C (7S0°F). The boiler was
fired with 100% RDF, although auxiliary oil
burners were used for start-up and during
fuel feed interruptions. Air pollution
controls for the boiler consisted of a
bank of 12 mechanical cyclones followed by
an ESP.
The purpose of the assessment was pri-
marily to investigate organic constituents
of the stack gases and to quantify odorous
components. However, other tests were also
included. Emission streams evaluated in-
cluded the boiler bottom ash, cyclone ash,
ESP ash, and the stack effluent gases.
Samples of the RDF were also collected and
analyzed for moisture plus chemical acH
elemental composition. The three ash
streams were analyzed for elemental compo-
sition. Stack emissions were continuously
monitored for S0_, NO , CO, 0,, and total
bydrocarbcn concentrations, and were also
tested to determine levels of vaporous mer-
cury and aldehydes. In addition, a sample
was collected using the EPA SASS for analy-
sis, under El'A's Level 1 protocol.
Initial results of the test program
did not indicate any pollutant emissions of
major concern. Stack gases contained rela-
tively low concentrations of SO,, NO , and
hydrocarbons. Carbon monoxide levels were
slightly greater than anticipated.
Emission of carbonyl compounds (alde-
hydes) were detected at a maximum level of
7 ppm (2.95 kg/h).
Mercury vapor concentrations n the
stack effluent vere very low (<0.12 mg/n1),
and it appears that mercury 'evels are
greatest in the fl.y ash collected by the
ESP. The concentration of mercury in sam-
ples of the RDF was constant at about
3 Mg/g-
Several trace metals were <>tected in
the stack gases at relatively high concen-
trations. Of these, lead, autimony, chro-
mium, and arsenic were most notable. Their
respective concentrations in the SASS
sample were 580, 460, 640, and 560 pg/m3.
Elemental analysis of the bottom ash. cy-
clone ash, and ESP ash streams also indi-
cated that many of the more volatile ele-
ment1- were associated with the smaller
sized particles.
Organic analysis of the SASS sample,
using EPA Level 1 and additional gas chro-
matography /mass spectrometry (GC/NS) ana-
lytical techniques, shoved a variety of
organic constituents. No single compound
grcup appeared to predominate, although
several polynuclear aromatic hydrocarbons
were detected. All organic results were
qualitative.
Compounds consistently observed in all
SASS component extracts included naphtha-
lene, fluoranthene, aceoaphthylene, pyrene,
phenanthrene/anthracene, bis(2-ethylhexyl)-
phthalate', and diphenylaoine. The majority
of additional compounds were found in the
KAD-2 resin extract and included two chlo-
rebenzenes, hcxachlorobenzene, fluorene,
and dibutylphthalate.
Detailed reports for all of the proj-
ects discussed in this paper are in various
stages of publication. Requests for these
reports should be forwarded to Incineration
Research Branch, IERL-Ciocinnati,
Cincinnati, Ohio 45268.
202
-------
WASTE-TC-ENERGY FACILITIES:
A SOl'RCE OF LEAD CONTAMINATION
Ad-a C. S.ilas
York Services Corporation
Stamford, CT 06906
David F. Lewis
C.A.F. Corporation
New Jersey
Donald A. Oberacker
U.S. Environmental Protection Agency
ABSTRACT
Resource recovery activities are rapidly growing in the United States and among the most
actively pursued technologies are those that recover energy from Xcnicipal ?olid Waste
spirable size range, with aerodynamic diameters usually less than 2 Urn, and the
location of the facilities typically adjacent to densely populated areas require that
the subject be given careful attention.
INTRODUCTION
Almost four Founds of NSW are gener-
ated per person per day in the United
States, amounting to an annual 140 million
Kg (160 million tons). If the rate of in-
crease in waste generation maintains its
historic g.-owth of 1.5* a year, by the end
o1." this century the U.S. will be faced
with the problem of diseasing re-re than'
180 million Mg (200 millior, tons) of
garbage annually (12).
Sixty percent of that waste is gener-
ated in the ISO major urban centers, popu-
lated by close to 140 million people.
These areas have to dispose of almost a
quarter million Mg of garbage daily (12)
in an efficient cost effective and ervir-
omentally acceptable manner.
In the'e major urban areas, landfills
which have in the past provided an
imnediate solution to the HSU disposa.
problems have become scarce, expensive
and its use restricted by stringent
environmental control regulations.
Ironically, at the same time the MSW dis-
posal problems faced by these urban .en-
ters increased in magnitude, the U. S.
entered a period of restriction on the
availability of inexpensive fuels, which
fostered the development of new sources
of energy.
203
-------
The ability to solve the MSW dis-
posal problem of the cities, while simul-
taneously contributing to the solution
of the erergy situation became a challeng-
ing qoal. ConsequentI5' wasre-to-energy
projects proliferated, using various
technological approaches, most of which
had in common combustion as the final
fate for all or part o: the MSW.
At thv s.ino time, concern rapidly
surfaced about the er.viroruiient.al sound-
ness of these alternatives relative to
convention:*! disposal. The release to
rhe atnosphere of pollutants cither
present in the MSW or formed during com-
bustion were possibilities that had to
be carefully investigated.
Lead is one of the toxic elements
found to exist in MSW. its deleterious
effects on human beings h-ivc been
known for centuries.(101 (13) Hence, the
onqin and the amount of that pollutant
ir: M.'K arc! rLf-se derived fuel? (RTF),
its fate uunng combustion and the degree
oi air contamination that could result
froo the pro 1iferaticn of energy recovery
facilities," have been the subject of an
investigation sponsored by *:he fc'PA. (35)
OCCURRENCE OF LEAD IN KSW
Lead exists in both the combustible
and the non-corabustible fraction ot" MSW.
in the non-coobustible fraction it may
be found in materials such as piping,
cable sheathing, ceramics, glass, solder,
etc. (23). In the .ombustible fraction
it is primarily found in the pigments
used in printing inks, in paper ;oods,
and in the stabilizers used for plastics
products (20) (21) (24).
Little information about lead analy-
sis performed in representative samples
of MSW is available, perhaps because of
the enormous difficulty in obtaining
such a sample. Nevertheless, results of
analyses performed in different fractions
Of MSW indicate that between 1000 and
5000 ppm by weight o. lead can be
exp
Weight %
Reference
2 30
110-1300
(37)
265
220- 300
(17)
12
7- 1-i
(15)
1^4
16- 124
(14)
390
85-lbO0
(4)
400
	
(26)
152
T 1 IO
I rt 1 \
46&
	
(161
956
447-1749
(9) (22)
500
———
(41)
The components of the combustible
fraction of MSW, their contnoution to
its lead content, the oriain of that
lead and the ranges of concentration are
presented in Table 2.
It is evident that HSW components
bearing-pigments and stabilizers carry
the ipajor responsibility for the presence
of lead in the combustible fraction cf
MSW. Certain printing inks, probably
rich in lead chromatea. may have to
20,000 ppm of lead. (17)
ENERGY RECOVERY1 TECHNOLOGIES
Several technologies have been
developed to recover energy f-om MSW.
The ultimate goal of these technologies
is to burtj MSW or its combustible
fraction to generate steam and/or
electricity. (42)
The technologies can be grouped in
two basic categories:
(1)	Separation processes, which produce
different types of RDF.
(2)	Thermal processes, which generate
steam an-J/or electricity through
the thermal processing of HSW or
RDF.
The pieparation of RDF can be
achieved by the separation vf the various
204
-------
TABLE ?. LEAD CONTENT OF THE COMBUSTIBLE FRACTION OF MSW IN THE U.S.
Component
Lead Component
LeoJ Content-ppm
Corrugated board
Newspapers (26)
Paperboard
Paper packaging (26!
Office paper
Magazines, books (20)
121) (26)
Tissue paper, towels
Paper plates, cups
Other
Yard waste
Food Waste
Plastic (16) (21J
Kood
Leather, rubber
Textiles
Pigments-Printing Method Recycle
Pigments
Pigment*;
Pigments, Glazes
(26) From recycle
Pesticides, Fungicides
Stabilizers, Pigments
Paints, Primers
Stabilizers
C
1
40
2
50
69
500
-10,000
10
6 - 3,600
2 - 150
10 - 1,000
components of MSW into specific fractions,
one of which is a combustible fraction,
usually organically rich in cellulosic
and plastic materials. According to the
separation process employed, the RDF
produced may be classified as one of
four different types: (33) (42) (45) (47)
(1)	Fluff-RDF (f-RDF), essentially
plastic and paper products, three
inches or less in size, obtained
by a combination of trommeling,
shredding, air classification and
magnetic separation of MSW.
(2)	Densified RDF (d-RDF), obtained as
f-RDF, shredded again and pellctized
in a pellet mill.
TABLE 3. TYPES OF THERMAL PROCESSES
Systems
Material Burned
Katerwall
Incinerators
Co-firing Systems
Modular Coobustors
Pyrolysis Systems
MSW
Shredded MSW
RDF
RDF & Fossil Fuel
MSW
MSW
Shredded MSW
RDF
LEAD EMISSIONS
Analyses performed in sediments and
materials of biological origin indicate
that the natural concentration of lead
in the atmosphere should be about
0.0006 (ig/m3. (1)
(3)	Powdered RDF (p-RDF), prepared as
f-RDF and then chemically treated
and put through a hot ball mill,
resulting in a very fine powder.
(4)	Wet RDF (w-RDF), where the shredding,
and separation of the combustible
fraction is accomplished in an
aqueous medium, the resulting fuel
typically having about 50% moisture.
The thermal processes can also be
divided into four categories, as present-
ed in Tiiblo 3.
However, the actual concentration is
considerably above that level, due essen-
tially to the widespread use of lead for
a multitude of purposes.
The lead content of the ambient air
varies with location and that variation
can be related to the activities of the
particular areas, i.e.,
o Suburban areas, about 0.1 Ug/m^ ^
o Cities, on the urder of 1.0 lig/m
o Near highways, during heavy
traffic, up to 20 ug/m3
o Near poorly controlled smelters,
up to 300 Vig/m3
203
-------
The major source of lead contamina-
tion in the U.S. is the antiknock addi-
tives used in gasoline. They are
responsible for 88% of the lead pollution
in the air. (1) The restriction in their
use is expected to reduce lead emissions
from mobile sources from 130,000 Mg
(144,000 cons) per year, as they were in
1975, to less than 40,000 Kg(44,000 tonj)
per year by 1985. (1)
The second largest contributors to
atmospheric lead are waste motor oil
combustors and municipal incinerators,
with combined emissions of about 12,000
Mg (13,200 tons) of lead per year.(1)
These contributors and other
stationary sources, in particular coal-
fjred power plants which in 1976 were
estimated to release approximately 640 Mg
(706 tons) of lead to the atmosphere(8)and
MSW/RDF burning facilities will become
even more important as their numbers
continue to increase.
Emission of Lead from ctationary
Combustion Sources
Lead is emitted from stationary
combustion sources such as coal, oil,
and refuse burning facilities mainly as
particulates, although the possibility
exists of some fraction leaving the
stacks in the vapor form. (161 (19) (22)
(31) (39)
The concentration of lead in the
fly ash generally increases as the
particle size decreases, a phenomenon
explained by vaporization-condensation
theory (19) (20) (21) (24). Due to the
high vapor pressure of lead and some of
its compounds, much of the lead is
volatilized during the combustion prwess
and subsequently deposited on the rurface
of particulate matter by condensa.icn in
the cooler sections of the system. The
smaller-sized particles, with their
higher surface-to-mass ratio can be
expected to have a higher proportion of
lead than the 1irger-sizcd particles.
Figures 1 to 4 illustrate this phen-
omenon and are based upon data from emis-
sion tests performed at the St. Louis
demonstration plant and at a coal-fired
power plant respectively. (16) (7)
The characteristics of thes :
emissions indicate that more than one-
half of the lead emitted may oe concen-
trated in the particulates with aero-
dynamic diameters less than 2 Urn (25).
IWWCTOR •IMC Iweee*
>)
FIGURf 1.
1 80
40
UNCONTROLLED PARTICULATE
COLLECTED WITH BRINKS
IMPACTOR AT ST. LOUIS
DEMONSTRATION PLANT
t	I
900	1000
LEAD too*)
FIGURE.2. RETAINED FLY ASH COAL-FIRED
POWER PLANT
6000"
4000*
200P
9 8 7 6 5 4 3 ? I
AWCRSOtt MPACfOR 8IA0C	ttj* 	9
FIGURE 3. SUSPENDED PARTICULATE
COLLECTED WITH ANDERSON
IHPACTOR - ST. LOUIS
DEMONSTRATION P] ANT
206
-------
12-
4-
-T—
1000
1600
2000
LEAD tam)
FIGURE 4. .SUSPENDED PARTICULATE -
COAL-FIRED i'CWER PLANT
When analyzing data from different
sources it L-ecame evident that rarely
do the mass' balances close reasonably.
It would appear that some of the lead
simply cannot be accounted for. In the
Ames facility, for example, 445 g of lead
per Mg of fuel burr.cd were expected in
the emissions according to the mass
balance, however only 203 g per Mg of fuel
were detected by emission testing. In St.
Louis, 43 g of lead per Mg of fuel was
calculated tc be released in the stack,
however only 3.6 g of lead per Mg of fuel
were found in the emissions. This situa-
tion has also been reported for coal-fired
power plants (27) and tor waste motor oil
conibustors.
Although in the case of KSW and RDF
burning facilities socne errors can be ex-
pected because ;of the heterogeneity of the
materials to be sampled and hence, the
difficult task of obtaining a representa-
tive sample, it seems puzzling that almost
always, a "loss" of lead occurs. Further-
more, in the case of coal or waste oil,
the lack of homogeneity excuse does not
seem valid.
The problem could also be in the
sampling of the bottom ash. This is
suggested by the fact that a wide range
of lead concentrations (from less than
1 ppm up to several hundred ppn) have
been found in the bottom ash at the
various plants.'
The possibility exists also of lead
exiting, the stack in the vapor form or as
very.small particles, undetected by
standard particulate collector trains.
In any case, further investigation should
be performed to determine the f-.te of
that "missing" lead.
Particulate Emissions from Stationary
Combustion Sources
Host of the lead that leaves the
combustion zone is carried on the surface
of particulate entrained in the combus-
tion gases, with the relative concentra-
tion of lead on these particles a function
of their size. Hence, the amount of
particulate emitted and the size distri-
bution of these particulates are impor-
tant factors to be considered.
The amount and characteristics of
uncontrolled particulates produced by
stationary combustion sources depends
upon several factors, including:
o Type of the fuel burned
o Mode of combustion (suspension
firing, stoker, etc.)
o Degree of mixing
o Combustion chamber design
o Boiler parameters (load, air supply,
combustion zone temperature profile,
heat rate, etc.).
Hence, substantial differences in
lead emission levels can be expected for
different wastc-to-energy facilities.
Table S and Figure 5 present parti-
culate emission rates for several station-
ary combustion sources and their effect,
on particle size distribution.
CONTROL OF LEAD EMISSIONS
The elimination cf lead from KSW
prior to burning does not appear
econoaitally feasible at this time.
Separation of the inorganic fraction
containing a portion of the lead is
feasible ard routinely performed in
the manufacture of RDF. This, of
course, will reduce the amount of lead
per unit weight of the combustible
(organic) fraction. Nevertheless, the
RDF still contains a average 3S0 ppm of
.\ead, as previously discussed. That
lead is inherently attached to the very
materials that form the combustible
fraction and the ilDF in the form of
pigments and st£l>i lizers. Although
several methods cxint for "de-inkinq"
paper, the use of such techniques to
207
-------
TABLE 5. PARTICULATE EMISSION RATES
Type of riant
Emission Ka;cs
(Kg of Part./Mg of Fuel)
Municipal Incinerators
without heat recovery (391
Municipal Incinerators
with heat recovery (5)
Pulverized Coal
U01 ash coal) (5}
F.toker
UOt ash coal) (5)
Cyclone (5)
Cofiring-Tangential Firing
<7\ M5F f S.6* ash coal) (L4>
Cofiring-Spreader Stoker (19)
13
10
80
SO
10
46
45
(801 of	total	:oal	ash)
150% of	total	coal	ash)
(lot of	tota!	coal ash)
(42\ of	total	fuel	ash)
(12% of	total	fuel	ash)
50-
«o
30-
TO
10-
|>in
con
\
/
/
/
/ /
y
u 25 50 TS 100 125 ISO
Mnncte SUE MmI
rismus 5. TYPICAL PARTICLE size distri-
butions OF VARIOUS rOA', COHEUSTIOK
SYSTEMS
remove lead from a material that will
ultimately be u&ed as a fuel is not
economically attractive at this tine.
Nu other technology that would be used
to eliminate lead prior to conbustion
has yet been commercially demonstrated.
No information about methods tc
retain, the lead durir-g the combustion
process were found, although limited
data from the combustion of MSW with
sewage sludge suggests that a reduction
in liead emission Day be possible with
co-disposal, another area deserving
further investigation. (6)
At the present time it appears that
the reduction of lead emissions from
MSW/RDF burning facilities is restricted
to particulate control in the stack gases.
The difference between various
control systems in their capacity to
retain lead is directly related to
their capacity to control fine particu-
late. This is evident from emission
data obtained from three different
municipal incinerators for which it is
valid to assume similar composition of
the ?!SW burned. These data are presented
in Table 6.
The difference between the total
efficiency of particulate control 
-------
TABLE 6. LEAD EMISSION RATES FROM INCINERATORS WITH
DI-FEREHT AIR POLLUTION CONTROL EQUIPMENT
Incineraf
Particulate Control
	Equipment	
Lead Emissions
q/Mq of MSW burned
Nicosia
Alexandria
SWRS #1
Plate (water, counter-flow) scrubber
Spray (baffle) scrubber
Spray quenchinq/multiclone/ESP
280
272
35
Plant
TABLE 7. OVERALL AND LEAD REMOVAL EFFICIENCIES
Particulate
Control Device
Overall
Efficiency ~
Lead Removal
Efficiency %
St. Louis (coal/RDF)
Ames (coal/RDF)
Braintree (MSW)
ESP
Multiclone
ESP
90-99
82-92
74
47
. 7
S3
be obtained making certain simplifying
assumptions. Table 8 identifies the
plants that have been tested for lead
emissions, the type of fuel burned, the
air pollution control equipment used
and the lead emission factors based upon
the combustible fraction of MSW. When-
ever possible, emissions have been
calculated through mass balance.
All the existing and planned waste-
to-er.ergy facilities in the United States
can be grouped into four different
categories represented by the plants
in Table 8.
It can be assumed that all the plants
in each category will behave, with
respect to lead emissions, as the one
that represents that category, indicated
in Table 8. Then, knowing the nominal
capacity of ecch plant (12; (36) (42)
and using the lead emission rates from
Table 8, the total lead emissions that
can be expected, if all the plants become
operational at full .capacity can be
calculated.
This esitmate assumes that all f-RDF
and p-RDF produced will be cofired in
facilities similar to those at Ames and
TABLE 8. PLANTS THAT REPRESENT DIFFERENT CATEGORIES OF KAS7E-TO-ENERG/ SYSTEMS
Plant
(Category)
Fuel
St. Louis (14)
(co-firing)
Ames (19)
(co-firing)
Arlington (12)
(co-firing)
Braintree (13)
(w.w. incinerator)
Particulate
Control Device
Stack Lead Emission
Factors (g lead/Mg MSW)
Combustible Fraction
Coal/f-RDF (8*)
Coal/f-RDF (30*)
Coal/d-RDF
MSW
North Little Rock (11)
(MCU)	MSW
Baltimore (22)
(pyrolysis)
MSW
ESP	305*
Multiclone	896*
Mult*clone	82
ESP	52
None	95
Wet Scriibber/Dehumidifier	9
* Calculated through mass balances.
209
-------
St. Louis; that d-RDF will be fired as in
the Arlington plant; chat water wall
incinerators will produce emissions
similar to those produced by the Braintree
facility and that all modular cotnbustors
will behave at the North Little Kock Unit.
Table 9 identifies the number of
facilities of each .category which are
planned to become operational by 1985,
(10) (32) (37) (39) the total MSW thit
will be processed and/cr burned and the
expected lead emissions.
The figures indicate that if 30
million tons of KSW are processed and
burned there is a potential for the
emission of 1500-6000 ng of lead pec-
year and.as it has previously stated,
much of this emission will be in the
form of very fine particulate matter, in
the respirable size raiye.
If resource recovery activities
advance At the expected rate (10), it
is projected that by 1990, i0» of the NSW
generated in the 150 largest urban areas
in the U.S. will be processed in
resource'recovery facilities. This
amounts to about 34\ of the t^tal MSW
generated in the whole country. The
technologies that are nest likely to be
.pursued are those that allow for recovery
of both materials and .energy. Therefore,
it can be readily appreciated that the
potential emissions of lejd from MSW/RDF
combustion could reach 12,000 Mg/year
by 1990.
The figures for the potential emis-
sion of lead can be put in perspective by
comparing them to estimated atmospheric
lead emissions from other sources. Thus,
gasoline combustion is estimated to have
been responsible for the atmospheric
emission of about 140,000 metric tons of
lead in 1975 (1). Projected lead
emissions from MSW/RDF combustion are
small compared to those of automobiles,
they are similar to the 1975 estimate of
10,000 Mg of lead emissions from waste oil
combustion (1), and are far greater than
the 640 Hg emitted from coal combustion
in 1975. (8)
POSSIBLE LFKECT ON THE AMBIENT AIR
One of the criteria by which emission*,
of lead from RDF/MSW combustion will be
judged, is certain co be the effect of
these emissions upon the national ambient
air quality standard. The standard of
1.5 lig of Pb/m3 is set on the basis of a
time weighted average calculated over a
calendar guarter.
Some simplified calculations were
performed to determine the instantaneous
effect of emissions from the St. Louis,
Ames and Braintren facilities. (35) (46)
The calculations were intended to show
tno maximum emission rates that can be
tolerated without having the instantaneous
maximum, ground level, downwind lead con-
centration in the ambient air exceed
1.5 Mg/m-5
The values obtained must be viewed
witji caution. In the first place, the
calculations assume that the emission
source is sitting on flat, featureless
terrain with a normal, rather than invert-
TABLE' 9. NUMBER OF PLANTS, CAPACITY AND POTENTIAL LEAD EMISSION FOR EACH CATEGORY
No. of Plants in	Total Nominal
U.S. (Operational, Capacity Kg of	Potential Lead
in Construction and MSW Processed Per	Emission
Type of Plant	Planned (12) (36) (42) Day (12) (36) (42) (H«j/Year)
Waterwall Units 34 46,668	497
Modular Combustion Units 16 2,123	38
RDF Plants 29 33,100	925-4935
Pyrolysis 4 1,800	4
Undecided (RDF) 	1_ 3,000	83- 447
TOTAL 84 86.691	1547-5«21
Possible Dnissions were calculated for an average of 50 weeks a year at 6 days a
week of operation, using emissions factors shown in Table 7.
210
-------
ed, temperature gradient. Local topography
and atmospheric conditions could, in a
real case, give rise to results which may
be dramatically lower, or hiqhe.' than those'
that have be^n calculated. Secondly, it
must be noted ttiat at any site, it is
almost impossible that steady state condi-
tions will occur for extended periods of
time. In the natural meteorological pro-
gression, the emissions from a ^iven
source will be spread over 160° according
to the local wind rows. Accordingly, the
three month average of the lead concentra-
tion at any given site can be a.fraction
of the instantaneous values that have
been calculated.
The results of these calculations
indicated that for the St. Louis
facility there is a broad set of condi-
tions where the actual emissions from
the plant are at a level that could
cause the instantaneous, maximum, ground
level concentration to be at least 25\
of the standard. For the Ames and
Braintree sources there are a number of
conditions that would give rise to
instantaneous, maximum downwind, ground
level concen*rations of lead that are
35\ to- 45* of the standard.
The standard is, of course, based
on a three month average, and it is
certain that changing atmospheric con-
ditions will., in any three month period,
tend to disperse the emissions over a
much broader area than that considered
in the calculations.
It is, therefore, highly improbable
that the emissions from any of the
sources studied in themselves would
cause a violation of the national ambient
air standard. On the other hand, allow-
ing for changing wind directions and
meteorological conditions, it seems
possible that all of th? sources could
emit sufficient lead to consume a signi-
ficant fraction of that allowed by the
national ambient air quality standard.
The calculations made are of a type,
which serve to make preliminary estimates
of concentration of species in the
ambient air. However, the results do
indicate that the effect of a given RTF/
MSW combustor upon lead in the ambient
air, cannot be dismissed lightly. It
would seem prudent to extenJ the calcu-
lations in a more rigorous manner prior
to making any definitive statements.
CONCLUSIONS
o Lead mainly in the 'orm of pigments
and. stabi 1 izers, is an inlierent
part of much of the combustible
fraction of municipal solid waste.
o Lead in the combustible fraction of
municipal solid waste, as well as in
refuse derived fuel, is present at
about 350 ppo.
o Good mechanical separation of the
combustible and non~comt>ustible com-
ponents of municipal solid waste will
help to reduce the amount of lead in
MSW and RDF. However, it will not
solve the problem.
o At present there are no economically
feasible, commercial methods for dis-
associating lead and its compounds
frott the combustible fraction of
municipal solid waste.
o When MSW or RDF is combusted, much
of the lead is emitted with particles
that are extremely Email and difficult
to retain effectively in conventional
particulate control devices. Multi-
clones are very inefficient in the
retention of lead and ESPs have a
better performance although they
retain only about 50%. Baghouses
probably could be the solution to
the problem.
o The asKessaent of lead emissions via
the collection of particulate
oaterial on a 0.3 Um filters, may
be underestimated. Since a signifi-
cant proportion of lead is likely to
be associated with submicron particu-
late natter, or may even be in the
vapor phase. It is possible that
seme proportion of the lead may
not be captured in the sampling train.
o The studies of lead emissions published
in the literature are not consistent
in two critical areas - the collection
of representative samples and the
analysis of these samples. These
deficiencies make for difficulty in
the comparison of data.
2tl
-------
o Emission of lead from the combustion
of MSW or PDF may be responsible for
the emission of 1500-6000 Mq of lead
per year into the atmosphere by 1985.
The figure could approach 12,000 Mq/
year by 1990.
o At least SOI of the lead will be
emitted pn particles with aerodynamic
diameters below 2 I'm.
RECOMMENDATIONS
o The suspicion that emissions of lead
may be underestimated as a result of
-lead passing through the sampling
device, warrants further investigation.
o Although particulate control devices
can capture a portion of the lead-
bearing particulates, there is
considerable room for improvement.
Effort in this direction would be well
advised.
o Sei"<"ted MSW or RDF combustion units
should be studied in a specific, care-
fully designed, well thought manner,
for the purpose of accurately deter-
mining lead emissions.
o Atmospheric modeling calculations
should be-performed to p-ovide realis-
tic estimates of the effect of indi-
vidual emission sources upon the
concentration of lead in the ambient
air.
o The idea of curtailing the use of lead-
ed pigments in paper should be studied.
References
1.	Air Quality Criteria for Lead - ErA
600/18-77r017 U.S. Environmental
Protection Agency, Washington, DC.
297 pp.
2.	Ananth, K.P., L. J. Shannon and
M.p. Shrag. 1977. Environmental
Assessment of Waste to Energy Process-
es - Source Assessment Document. EPA
600/7-77-091, D. S. Environmental
Protection Agency, Cincinnati, Ohio.
81 pp>
3.	Black, P., 1976. Energy Recovery
from Solid Haste: A Review of Current
Technology. NTIS PB-260-633. U. S.
Department of Commerce. 68 pp.
4.	Campbell, W.J. 1976. Metals in the
waste we burn? Environmental Science
and Technology, 10(5):436-439.
5.	Compilation of Air Pollution Emission
Factors. 1960. EPA AD-42 Third
Edition. (J. S. Environmental
Protection Agency, Research Triangle
Park, NC.
6.	Cross, F.L., R. J. Draqo and 11. E.
Fr-ancis, 1970. Metal And particulate
emissions from incinerators burning
sewage sludge and mix refuse. In:
Proceedings of 1970 National Inciner-
ator Conference ASME Cincinnati, Ohio
pp. 183-195.
7.	Davison, R.L., D.F. Natusch and J.R.
Wallace, 1974. Trace elements of
fly ash, dependence of concentra-
tion and particle size - Environ-
mental Science and Technology 8
(13): 1107-1113.
8.	Drehmel, D.C. and J. H. Abbott, 1978.
Control of particulates from com-
bustion- In: Enerqy Environment
III. EPA' 19- 7P-022 L\ S. Enviion-
tnental Protection Agency, pp. 323-335.
9.	Elo, H. K. and F. R. Rhodes. 1976.
Solid waste fuel pellets provided
fuel supplement. Pollution Engineer-
ing.
10.	Engel, R.E. et al, 1971. Environ-
mental Lead and Public Health U.S.
Environmental Protection Agency-
Air Pollution Control Office.-
Research Triangle Park. NC 34 pp.
11.	Evan, J.C. et al, 1977. Evaluation
of the Ames Solid vtaste Recovery
Systems, Part I. EPA 600/2-77-205,
216 pp.
12.	Franklin, W.E., M.A. Franklin and
H. E. Weber. 1979. Solid Waste
and the Paper Industry. Executive
Suxmary and Overview - A report
prepared by Fran)-li.n Associates, Ltd.
and Technology Corporation for the
Solid Waste Council-of the Paper
Industry - American Paper Institute.
279 pp.
13.	Frounfelker, R. 1979. Small Modular
Incinerator System with Heat Recover
A Tecljiical. Environmental and
212
-------
Economic Evaluation. Executive
Siiranary - U. S. Environmental
!-rotection Aqency, Cincinnati, Ohio.
65 pp.
14.	Golembitwsky, M.A. 1979 Environmental
Assessment of a Waste to Energy
Process - Boiler Co-fired with Coal
and Densificd Refuse Derived Fuel.
Virginia Heating and Refriqerating
Plant Arlinqton, 'Jk. Draft Final
Report prepared by Midwest Research
Institute, Kansas City, Missouri for
U.S. EPA, Cincinnati, Ohio. €0 pp.
15.	Golejiibiewsky, f*.A. ft al. Environ-
mental Assessment of a Waste to
Energy Process- Braintree Municipal
Incinerator-Final Report. EHA
Contract No. 68-02-2160.
16.	Gorman, D. C. et al 1977. St. Louis
Demonstration Final Report: Power
Plane Equipment, Facilities and
Environmental Evaluation. EPA 600/2-
7-155b. C. S. Environmental Protec-
tion Aaer.cy, Cincinnati , rhio. 40^ pj;
17.	Gorman, D. et al 1978 - Environmental
Assessment of a Waste to Energy Pro-
cess - Union Carbide Purox System -
Piwi Report - EPA Contract No.
6b-02-2i6. 201 pp.
18.	Green. V. A. et al. 1976 Lead
poisoning. Clinical Toxicology
9 (1) pp 33-51.
19.	Greenberg, R.R. 1976. A itudy of
Trace Elements on Particles from
Municipal Incinerators. Ph.D.
Thesis, University of Maryland,
30} pp.
20.	Grecnberg, R.R. et al 197B. Composi-
tion c£ particles emitted from the
Nicosia municipal incinerator. En-
vironmental Science and Technology
12 (12): 1329-1332.
21.	Greenberg, R.R., W. H. Zoller and
G.E. Gordon, 1978. Composition
and size distribution oC particles
released in refuse Incineration.
Environmental Science and Technology
12(5): 566-573.
22.	Hall, J.L. et al 1978. Evaluation of
the Ames Solid Waste Recovery System-
Part UI-Vol. II: Appendices EFA
Contract Ho. P.-6039 03-01-0 to the
City of Ames, Iowa. u. S. Environ-
mental protection Agency, Cincinnati,
Gnio.
23.	liankin, L. , G. H. Heichcl and R. A.
Hotsford. 1973. Lead Poisoning frcm
Colored Printing Inks - Clinical
Pediatrics 12 (11): 654-655.
24.	Haynes, B.W., S.L. Law and W. J.
Campbell. 197ti. Scurccs of Metals
in the Ccnbustibie Fraction of Muni-
cipal Solid Waste. U. S. Bureau of
Mines. R.I. 8319 - U. S. Department
of Interior. 16 pp.
25.	Haslbtck, J.L. and 6.C. McCoy 1978.
Source Emission Tests at the Balti-
more Demonstration Pyrolysis Facility
EPA 600/7-78-232 ". S. Environmental
Protection Agency, Cincinnati. Ohio.
80 pp.'
26.	Jacko, R. B. and D.VI. Neuendorf, 1977.
Tr.ice retail particulate emissior. test
resul' from a number of industrial
and r^inxcipal point sources - Journal
of the Air Pollution Control Associa-
tion 27 (10): 989-994.
27.	Jackson, J.W. 19'6. A Bioenviron-
mexital Study of Caissions from
Refuse Derived Fuel. NTIS AD/A-024-
661 - U. S. Department of Conner Co.
117 pp.'
28.	Joselow, M.M. and J. D. Bogden 1974.
Lead content of printed media.
(Warning: Spitballs might be
hazardous to your health), AJPH.
64 (3): ."38-240.
29.	Kah - Hock, L., R. A. Lane and
B. Walker. 1975. Toxic materials
in paper indicate limits should be
set. Journal of Environmental
Health 3B (3): 168-169.
30.	Kaakinen, W. et al. 1975. Trace
element behavior in coal fired power
plant - Environmental Science and
Technology 9 (9); 862-86°.
31.	Kaakinen, J.W. and R.N. Jorden. 1973.
Determination of trace element mass
balance for a coal-fired power plant-
In: Proceedings of the First Annual
213
-------
NSF Trace Contaminants Conference-
Sponsored by the National Science
Foundation, pp 165-184.
32.	Lav, S. and c~. £ Gordon -1979 Sources
of metals m municipal incinerators
emissions - Environmental Science
and Technology 13 (4): 432-43P.
33.	Lav, 5., U. Haynes and W. J. Canpbell
197fl. Kreburn separation should
limit metal emissions. Waste Aae.
Sept. 1978 pp 51-^9.
3-1. Lav, S.L. , B.W. Haynes and W. J.
Omptw^i. 197B. ttirect incineration
of municipal waste versus separation
of combustibles - In: Proceedings
of the Sixth Mineral Waste Utilisa-
tion Svrvosium co-sponsored by the
U. S. Bureau of ttineb and the IIT
Institute, Chicago, 1L. pp 164-170.
35.	Lewis. D-F. anJ A- Salas 1980.
Removal of Lead frcm KJK. Final
Draft Report, EPA Contract
2742 and reierenevs neroin.
36.	McEvert, L.B. 1977. Waste eduction
and Resource Recovery Activities-
A Nationwide Survey - ii,\ Publication
SW-142. L5. S. Dwironrcental Pro-
tection ^qency - Washington, DC.
78 pp.
37.	Narr IT I, H.E., S.L. Law and W.
Campbell. 1976. Concentration and
sourc* of trace elements in the
combustible fraction of urban refuse-
In: Proceedings of the Fifth
Mineral Waste Utilization Symposium,
co-sponsored by the U. S. Bureau of
Mines and ITT Research Institute.
pp 251-252.
38.	Harr III, H.E., S. L. Law and D.L.
Sleyban. 1975, Trace elements in the
combustible fractions of urban
refuse. In: Proceedings of the
International Conference or. Environ-
mental Surveying ami Assessment.
39.	Korey, 3. 197C. Inorganic; resource
recovery and solid waste preparation
from municipal trash. In: Proceed-
ings of the Fourth Mineral Waste
Utilization Symposium co-sponsored
by the U. S. Bureau of .Mines and
117 Research Institute, pp 85-94.
40.	National Frimary and Seconda*-y
•tabient Air Quality Standard £or
load - 40 CFR 50.12
41.	Natusch, D.F. 5., Wallace, J.R. and
.Evans, C.A. »'r. Toxic trace elements,
preferential concentration in respir-
able particles - Science 1S3 (t):
202-204.
42.	NCCR Bulletin. The Ocurnal of
Resource Recovery, i960. 10 (3):
57-65.
43.	Resource recovery activities: A
Status Report 1978. Resource Re-
covery and Energy Review. 5 (1):
PF 13-16.
44.	Rinujci, G.K. et al. l?~s, An
Evaluation of ISiission Factors for
Waste-to-Energy Systems. Executive
SuCT»ary prepared by Monsanto Research
Corporation, Dayton, Ohio, for the
U. S. Environmental Protection
Aaency, Cincinnati, Chio. 23 pp.
45.	Sussman, D. R. and S. L. Levy. 1979.
Recovering energy from municipal
solid vaste - A review of activity
in the U.S. in: Proceedings of the
Fourth Japanese-Australian Conference
of Solid Haste Management- Washing-
ton, DC.
46.	Turner. D. B. 1969 - Workbook of
Atmospheric Dispersion Eatimates-
U.-S. Department ot Health, Education
and Welfare - Public Health Service
Publication No. 999-AP-26-84 pp.
47.	Vance, T.P. and D. L. Powers. 1980.
Resource Recovery Systems- Part.3
Solid Haste Management. 23 (5): 26-34.
46. Vauqham, D.A., et al. Summary Report
or. Environmental Effects of Utilizing
Solid Waste as a Supplement Power
rlan Fuel - EPA Research Grant No.
R 804008-02.2.
ACKNOWLEDGEMENT
This research vas supported by-
the U.S. Environmental Protection Aaency,
Municipal Environmental Research Labora-
tory under Contract No. 68-03-2742,
Mr. Donald OteracVcr, Projoct Officer.
:u
-------
HEALTH AND SAFETY ASPECTS OF RESOURCE RECOVERY
E. Joseph Duckett
Schwartz o Connolly, Inc.
Washington, D.C. 20C06
ABSTRACT
Health and safety concerns have been addressed by all major sectors of the resource
recovery community. To date, there is no evidence that resource recovery pcoes any
significant and unmanageable risks to health or safety but there nave been concerns raised
about"dusts, airborne, microorganisms, noise, explosions and fires that may be encountered
during refuse processing. The response of the resource recovery community to these
concerns is reviewed for the examples of microbiological aerosols and explosion protection.
EPA, particularly the Municipal Environmental Research Laboratory, has played a major
positive role in Identifying and helping to resolve health and safety problems. Important
roles have also been played by other governmental agencies, by private companies and by
non-profit organizations.
Although some issues have been mishandled and some remain to be resolved, the response has
generally been straightforward, technically sound and marked by an unusually high degree
of government-industry cooperation.
Introduction
Not long ago, a headline In a major
U.S. newspaper began with the phrase,
"Health Hazards Beset Trash-Tired Plant."
The rest of the article then oroceeded to
describe a variety of problems allegedly
associated with this plant but, buyond the
headline, there was no further mention of
any specific health hazards or health
effects attributable to the plant. The
purpose of this illustration Is not to
comment upon the issue of whether health
problems really did exist at thic plant,
but simply to point out that health hazards
make headlines.
Several years ago, a major U.S.
magazine included in an article on problems
facing resource recovery the reported
emission of bacteria from a demonstration
facility in concentrations of up to one
alllloa (per cubic fact cf exhaust air).
The implication of the article vas that
lar^e numbers of airborne microorganisms
meant large risks to human health. No
attempt vas made In the article to quantify
the risks or even to explain that the
measured emission concentrations did not
represent human exposures. The purpose of
this illustration Is not to criticize the
magazine but only to indicate the difficulty
of expressing health rls!s in stralghtfor- .
ward, quantifiable terms.
Both because of the attention it
attracts an^ the confusion it can create,
the subject of health and safety aspects of
resource recovery Is difficult to address.
The middle ground between alaralsm and
Pollyannalso is often difficult to find,
not Just when writing news articles but,
more Importantly, when attempting to
identify and control whatever health or
safety hazards may exist.
¦ Generally, the resource recovery
community—governmental agencies and
private companies alike—have acted
responsibly to protect the health and
safety of those who work in the Industry
and those who live near resource recovery
facilities. This is not to" say that every
potential hazard has.been remedied
213
-------
lnnediateiy or that there have not been a
few false alarms raised, but on balance,
the record is a gooJ one. To date, there
Is no evidence that resource recovery poses
any significant and unmanageable threats to
health or safety.
I A historical rev.ev of the handling of
health and safety lssjes in resource
recovery suggests that there has been a
pattern to the way in vhich these issues
have been dealt with. The pattern has
often begun in an atmosphere of anxiety and
confusion but has often ended In a condition
of reasoned response. In between, a process
has beer evolving thac nay well signal two
encouraging trends. One trend is towards
the application of state-of-the-art science
to clarify the nature and extent of health
or safety hazards at •¦esu-rce recovery
facilities anil co develop metnoiis Cor
controlling chese hazards. The second
trend toward increasing government-
induscry cooperation In Identifying and
ccrricilng; potential health or safety
problems.
As a means of illustrating hew this
process seeas to be working, it is
worthvhile reviewing two examples. These
examples are chosen because they cover two
of the three major health and safety areas
that are of concern to the resource
recovery cunnunity. These two areas are:
a? microbiological aerosols generated
during refuse processing; and b) explosion
protection, especially during refuse
shredding. A third major area, air
emissions from refuse combustion faculties,
has already been covered in two preceding
papers.
Mlcroblc '.ORlcal Aerosols
It Is obvious to any observer of a
refuse-processing facility (whether
processing it for resource recovery or not)
that the processing generajes airborne
dusts. Not as obvious, but of concern, Is
that, because these dusts origlna.e in
municipal refuse, they nay contain chemical
compounds and/or microorganisms of potential
health significance.
The first published reports in which
airborne dusts from a resource recovery
plant wre sampled and analyzed for
mlcroliological content were based on
testing of the St. Louis Resource Recov.ry
Demonittit Ion Plant.This plant was
constructed and operated as an EPA-
supported demonstration of the processing
of refuse to prcluce refuse-derived fuel
(RDF) and to recover ferrous metals. Over
the period 1974-1975, EPA's Municipal
Environmental Research Laboratory (MERL)
sponsored a series of tests at the St.
Louis plant, Including an environmental
evaluation.
The 1974-75 environmental evaluation
of the St. Louis plant wa9 designed to in-
clude noise sampling, wastewater sampling
and determination of total particulate
emissions. What attracted mojt attention
and raised serious concerns, however, was a
small series of bacterial and viral
analyses conducted on samples of dust
emissions froo the exhaust ducts of the air
classifier and the RDF storape bin.
By EPA's oun description, these early
measurements of microbiological aerosols
at the-St. Louis plant were "cursory" and
were not Intended to be representative of
conditions at the St. Louis plant itself,
much less al 1 refuse proce^siii'j facilities.
Cursory or not, the reported bacterial
concentrations of one million per cubic
foot and the reported detection of viruses
aroused serious and understandable concern
within the resource recovery community and
Identified the need for a much more
thorough assessment of microbiological
aerosols at resource recovery plants.
The limitations .and inadequacies of
the 1974-75 tests of microbiological
aerosols at the St. Louis plant were
acknowledged by EPA and Included:
a)	lack of background data on
microbiological concentrations in
ambient air or in other types of
waste-processing facilities;
b)	lack oi sampling in areas where
workers or the public could actuilly
be exposed to the microbiological
aerosols;
c)	absence of standard ne'.hods for
sampling airborne bacteria or
viruses;
d)	sampling confined to only one plant,
a demonstration facility not
designed for long-term operation
and therefore not equipped with a
dust cleaning system;
e)	lack of quantified data on the
concentrations of specific
pathogens that stay have been
present among the large reported
216
-------
counts of total microoganisms;
f)	lack of confirmatory data on the
presence of viruses; and
g)	lack of health effects data on
which to Judge the potential
significance of measured
concentrations.
Because of these acknowledged
deficiencies In the Initial St. Louis
report, EPA-MERL supported a second series
of tests at the St. Louis plant. The tests
took place over the l?76-77 period.* For
this second series of tests, sampling,
protocols were carefully designed and
reviewed, ambient concentrations (both
upwind and downwind) were measured, samples
were taken In areas where workers might be
exposed, analyses for specific pathogens
and for viruses were performed, and the
effect of dust cleaning was evaluated.
Importantly, the tests were extended beyond
the St. Louis plant to Include an
Incinerator, a landfill, a transfer station
a sewage treatment plant and the back of a
refuse collection truck.
The results of the 1976-77 tests
suggested that bacterial concentrations
recorded in and around the St. Louis plant
fell or the high side of the range of
concentrations measured in and around other
waste treatment facilities.8 The
quantitative differences in concentrations
between resource recovery plants and other
facilities were hovn to be much smaller
for specific pathogens than for total
organisms.- No viruses were detected In any
of the facilities studied. A small number
of baghouse sysf.m tests revealed that
conventional uust cleaning equipment was
effective in removing airborne micro-
organisms.
This second series of St. Louis tests
did not overcome all of the shortcomings of
the initial series but, both directly and
indirectly, the second series represented
significant progress in the assessment of
the "microbiological aerosol problem."
Among the direct accomplishments were:
a) flrst-of-a-kind comparison of
•Actually, by the time this second series
of tests were conducted at the St. Louis
plant, the demonstration program had ended
and the plant had been closed. Therefore,
the plant had to be returned to operation
solely for the purpose of conducting
the microbiological aerosol tests.
microbiological concentrations in
ln-plant and ambient air at several
type:: of waste processing facilities
using a fixed set of sampling and
analytical protocols;
b)	reporting of concentrations nf
potentially pathogenic organisms as
well as total and indicator
organisms;
c)	acknowledgement that tlye previously
reported detection of viruses could
not be confirmed and may well have
been In error;
d)	illustration of the large degree of
temporal and locatlonal variability
among reported concentrations of
airborne Secterla and of the
imprecision of the sampling and
analytical protocols;
e)	review of the literature on
infective dosages in an attempt tq„
gu'.iic the Interpretation of the
health significance of reported
concentrations; and
f)	empirical demonstration of ihe
ability of conventional dust
cleaning equipment (I.e., fabric
filters) to remove airborne
bacteria.
The 1976-77 tests at St. Louis were
still based on sampling at a plant without
dust controls and did not resolve such
issues as the effects of employing a dust
control system In a full-scale plant or the
health significance of exposure to airborne
bacteria. The St. Louis tests did, however,
stimulate a number of efforts to better
understand and control microbiological
aerosols at solid waste processing
facilities.* Government-sponsored efforts
included: EPA-MERL support for testing at
refuse processing pilot plants at Richmond,
California and Houston, Texas;^'*R Energy
Research and Development Administration
(ERDA, row Department of Energy or DOE) and
EPA support for testing at the Ames, Iowa
resource recovery facility;12 and DOE
support for development apd evaluation
of improved sampling nethods at
•It Is noteworthy, that although this
paper focuses on solid waste processing,
there has been considerable research on
the nicrobiologlcal aerosols generated
during sewage treatment and spray irriga-
tion.1, 2» 10>25» 26 Much of this research
has been supported nnd/or stimulated bv
EPA and has complemented the solid-waste
related research. '
21?
-------
the Ames-DOE Laboratory.^ Industry-
sponsored efforts Included testing of dusts
at the National Center for Resource
Recovery's pilot plant in Washington, D.C.,®
and several -unpublished studies of
microbiological aerosols at full-scale
resource recovery plants. Interest In
airborne microorganisms extended to other
countries and sanpllng programs were con-
ducted at waste processing facilities
in Lngland, Sweden, Switzerland, and
Brazil.	An action taken '.ndependent-
ly of governmental aBctlcies and specific
companies was the formation of a health and
safety subcommittee (desigr.ated ~s E-38.07)
within the American Society for Testing and
Material (ASTM) Committee £-38 on Resource
Recovery:^ Functions of this ASTM
Subcommittee are discussed later in this
paper. The results of the St. Louis
studies and the other efforts cited above
were a better understanding of the
microbiological aerosol Issue and a
refinement In the technique employed to
address the issue.
Builaing upon the experience of these
earlier efforts, the National Institute for
Occupational Safety and Health (KIOSH) is
currently sponsoring a series of industrial
hygiene surveys of selected full-scale
resource recovery facilities. The NIOSH
contractor for these surveys is the .Midwest
Research Institute. Included within these
surveys Is a microbiological aerosol
sampling and analysis program which Is more
extensive, more technically advanced, and
ttore carefully controlled than any previous
program of its type. The NIOSH-sponsored
program Includes quantitative analysis of
specific pathogens, size fractionation of
airborne bacteria, detailed quality control
procedures and an experimental attempt to
detect viruses. It Is interesting that
sunlclpal and corporate operators of the
selected resource recovery facilities have
voluntarily agreed to have their plants
surveyed and, In some cases, have actually
requested that'their plants be Included in
the survey.
- The NIOSH. survey Is still in progress
and no final "reports have yet been
published. Preliminary data has been
presented, however, and these suggest
several findings. ^ For example,
conventional dust control systems, as
Installed in most full-scale resource
recovery .facilities, have been shown to
reduce airborne concentrations of
microorganisms to levels 10 to 100 times
below those recorded in pi lit plants not
equipped with dust controls. Reported
concentrations from the full-scale plants
are still generally higher than
concentrations reported for ambient air
or for air within manufacturing plants
but are at or belou the levels reported
for other types of waste-processing facil-
ities. Some specific pathogens have not
been detected at all (e.g.. Salmonella,
Shigella. Mocardia ast'eroides) and others
have been detected sporadically and in
low concentrations (e.g.. Staphylococcus
aureaus, Mycobacterium sp.).. Some micro-
organisms (e.g., Klebsiella pneumoniae
and Aspergillus fumlgatus) have been re-
covered In as many as one-third of the
duat samples collected but in concentra-
tions that, on a dally basis, could account
for no mote Uivv W' of the reported in-
fective dose. ,1J,i No viruses have been
recovered In any of the samples to date.
The final report of the NIOSH-sponsored
study is scheduled to be available by
Jun1981.
Explosion Protection
Interest in explosion protection
arises.from two features of refuse
processing plants. First is the fact that
the solid waste stream can contain
flammable and/or explosive materials,
including gasoline, solvents, propane, and
even discarded military ordnance. Second
is the opportunity, during various stages
of processing, for these materials to
accumulate and mix with air to form
explosive concentrations before
encountering 3 source of ignition.
Shredders, dust hoods, pneumatic ducts,-
cyclones, and storage bins are examples of
places where explosions have ccurred in
refuse processing plants.
Refuse processing Is not unique in
providing an opportunity"for explosions.
Dust explosions during the milling of grain
have been reported since early In the
1900's. Coal dust explosions within ball
mills continue to be of great concern to
utilities and othtrs using pulverized coal
as a boiler fuel. Explosions involving
gasoline, solvents and propellants hove
been reported among scrap processing plants
and other Industrial facilities.
Vithin these other Industries and
218
-------
within the refuse processing community,
most of Che concern for explosion protection
has been focused on shredders, or more
generally, size reduction equipment. This
is understandable for several reasons.
Shredders are often the first st^p in the
process flow through a refuse processing
facility. In some cases (e.g., shied and
landfill operations>, shredding may be the
only processing step, and many resource
recovery plants that recover materials
employ some type of size reduction
equipment. Also, shredd«rs can provide Che
conditions (enclosed space and metal-to-
metal sparks) that can lead to accumulation
and ignition of cxplosible gases or vapors.
The flT3t detailed Investigation of
explosions within refuse shredding plants
was sponsored by ERDA (now, DOE) in 1976.
Just as in the case of the initial St.
Louis tests for microbiological aerosols,
the ERDA-sponsored study of explosions cane
about almost as an afterthought. During
ERDA*s review of plans for a demonstration
plant for biological conversion of refuse
to fuel gaff (the Pompano Beach, Florida
project), questions were raised about the
safety of shredding refuse '.n preparation
for biological digestion, io answer theBe
q .estions, ERDA commissioned a survey of
explosions that had occurred in refuse
shredders.
The ERDA-sponsored study documented
over 100 explosions that had taken place at
refuse processing facilities (primarily
shred-landfill operations) over a three-
year period.27 The study excluded such
Incidents as "pops" due to aerosol cans and
reported that three of the 100 explosions
had resulted In injuries (none fatal) and
five had caused more than $25,000 in damage.
The study determined that most shredder
explosions are due to flammable vapors and
gases rather than dusts. An important
feature of the study was the identification
of approaches to preventing and controlling
explosions in refuse processing plants.
These Include; removal of explosive
material** from the waste stream; provision
of a venting system to relieve pressures;
use of chemical explosion suppression
agents; and the isolation or shielding of
plant personnel from likely explosion areaa
The reaction to the ERDA-sponsored
survey of shredder explosions was more
muted than had been the reaction to the
EPA-sponucred St. Louis tests. The
explosion survey did not Immediately launch
a aeries of follow-up studies nor did it
attract as much attention in the news media.
There were, however, several follow-up
activities that did occur. A subcommittee
on exploalc/u safety was formed within the
Waste Equipment Manufacturers Institute of
the Natural Solid Waste Management
Association. Several, articles appeared in
the trade journals of the solid waste
Industry. ",23 Approaches to controlling
shredder explosions were developed by •
shredder manufacturers and at least one
operator of a refuse shredding. planc20,21
but the interest In such approaches had not
yet peaked within the industry.
Unfortunately, and peculiarly, the
event which probably accoun..d for renewed
Interest in refuse shredder explosions was
a highly publicized explosion that did not
occur in a refuse shredder. This was the
explosion that occurred In November 1977 at
the East Bridgewater, Mass., resource
recovery plant. The explosion killed one
worker at the plant. Apparently, the
explosion took place In a cyclone designed
to de-entrain pulverized refuse-derived
fuel from a pneumatic conveyor. The causes
of this explosion were never conclusively
determined.
The renewed interest that followed the
tragic accident at East Brldgewater was
sustained by several other explosions at
resource recovery facilities and eventually
stJtaulated a major EPA-supported
investigation of explosion venting systems
for refuse shredders. This study, which is
still in progress, is significant not only
because of its technical subject but also
because of Che way it was originated and Is
being managed.
Technically, the explosion venting
study is the first rigorous (and non-
proprietory) investigation of explosion
protection for refuse shredders. Venting
is viewed by many as an essential form of
passive explosion protection for industrial
equipment in which explosions may occur.
The current study is designed to apply
generally recognized principals of
explosion venting to the specific
circumstances of a refuse shredder. This
is being accomplished by actual explosion
testing of various vent configurations on
a mock-up of a refuse shredder. Details of
this testing program are described In a
paper presented earlier In this, conference. '
219
-------
The results of Che shredder venting tests
are to be available later this year and
should be useful as a guide for the design
of venting systems.
The Initiation and wangenf-*: of the
current shredder venting Investigation
represents an unusual degree of cooperation
among all parties involved. As noted,
there had been considerable industry
Interest stimulated by the ERDA-sponsored
survey and by the explosions that had
occurred. This Interest was noted by
another agency, EPA-MERL, and, in late
1978, EPA tentatively decided to support a
follow-up study In the genera) area of
explosion protection. Before deciding on
the specific scope for the study, however,
EPA sought the opinions of refuse
processing system designers and operators
familiar with the explosion protection
problem. The major vehicle for seeking
such opinions was a newly-formed explosion
protection task group within the
aforementioned ASTM subcommittee on the
Health and Safety Aspects of Resource
Recovery (E-38.07). This task group
contains representatives of shredder
manufacturers, engineering firms,
governmental agencies, research
organizations and other explosion
protection specialists. The group
recommended that EPA focus Che study on Che
design ani testing of venting systems and
offered Its services for periodically
reviewing the results of the study. |1o
avoid any potential conflicts of interest,
the task group was not Involved in selection
of an EPA contractor to conduct the study.)
As the study has progressed, the EPA
contractor (Factory Mutual Research
Corporation) has reviewed its research
plan with the ASTM task group and has
presented formal progress reports at the
regular (semi-annual) meetings of the
group. Current plans call for the task
group to be the vehicle through which the
results of the EPA-sponsored study will be
Incorporated into an ASTM standard for the
explosion venting of refuse shredders.
Conclusion and Research Needs
Despite some dissimilarities, the
microbiological aerosol and explosion
protection examples seem to fit a pattern
that has also been followed for such topics
as stack emissions from refuse-combustion
facilities and physical and chemical
characteristics of dust concentrations
within refuse processing plants. The
pattern begins with an initial (sometimes
inadvertent) finding or event that usually
generates concern and controversy, but not
understanding. The intprmedlate steps In
the pattern tend to document the nature of
the concerns. The later stages employ
advanced techniques of sampling, analysis
and control to provide sound advice on how
best £o reduce any risks to health or
safety.
Obviously, It- Is oversimplification to
suggest that all safety and health problems
in resource recovery have been handled
smoothly and according' to a fixed pattern.
It is fair to "v, for example, that EPA
mishandled somi f the early publicity on
the St. Louis microbiological aerosol
results and that private- companies were
slow to respond to the early warnings on
shredder .explosions. On balance, however,
the pattern has been a positive one and has
been marVed by honest concern, careful
analysis, informed application of control
tecnnology and productive industry-
government cooperation.
One vehicle for Industry-government
cooperation has been the above-mentioned
ASTM E-38.07 subcommittee. ASTM is the
world's largest voluntary consensus
standards-setting organization and it
provides an excellent setting for
cooperation of the type needed on health
and safety topics. The membership of the
E-38.07 subcommittee is drawn from all
major segments of the resource recovery
community. Among the subcommittee's
current activities arc the development of a
standardized method for sampling
microbiological aerosols, the compilation
of explosion protection guidelines, and
the preparation of recommended practices
for health record-lteeping in resource
recovery facilities. Both N10SH and EPA
have used the subcommittee as a mechanism
for independent and interested review of
the progress of studies of health and
safety issues in resource recovery.
It is encouraging to note the progress
that the resource recovery community has
made in addressing health.and safety issues.
It Is now almost conmon practice for the
design of a resource recovery facility to
include dust cleaning systems and shredder
explosion venting systems. EPA has played
a major—and generally positive—role in
enabling resource recovery to make such
220
-------
progress. NIOSH, DOE Especially the Ames
!-uls
Demonstration Final Report: Refuse
Processing Plant — Assessment of Bac-
teria and Viral Emissions. U.S. Envi-
ronmental Protection Agency, Cincin-
nati, Ohio.
9.	Henry, M. C., J. Ehrlich, and W. Blair.
1969. Effect of nitrogen dioxide on
resistance of squirrel monkeys to
Klebsciella Pneumonia infection'. Ar-
chives of Environmental Health, 18:
604-614.
10. Hickey, J. L., and P. C. Relst. 1975.
Health significance of airborne micro-
organisms from wastewater treatmertt
processes—Part two: health signifi-
cance and alternatives for action.
22)
-------
10.	(cont'd.)
Journal WPCF, 47 (December 1975): 2758-
2773.
11.	Hollovay, J. R. 1976. EPA resource re-
covery demonscradon: summary of air
emissions analysis. Vaste Age, (August
1976): 50-52.
12.	falsely, R. K. 1979. Microbiological
air quality of the Ames municipal sol-
id waste recovery system. In: Quarter-
ly Report of the Anes-OOE Laboratory
for the Period January 1-Harch 31,
1979. Aaes Laboratory, Ames, [ova.
13.	Lacey, J. 1975. Potest'al haiards to
animals and man from microorganisms
in fodders and grain. Tr3ns. British
Mycology.Soc., 65: 171-184.
14.	Ledbeeter, J. 0., L. X. Hauck, and R.
Reynolds. 1973. Health hazards from
wastewater treatment practices. Envi-
ronmental Letters, 4: 225-232.
15.	Leanbice, L. L., and ». U. Knlscly. 1981.
Precision of the all-gjass inpinger
and the Anderson tmpicior in a munici-
pal. solid waste prrcessir.g facility.
Applied and Environmental Microbiology.
16.	Mansdorf, S. Z., et al. 1981. Environ-
mental health and occupational safety
aspects of resource recovery. In: Pro-
ceedings of the Seventh Annual Mineral
Waste Utilization Symposium. Illinois
Institute of Technology Research Insti-
tute, Chicago.
17.	Microbiological Contamination of Ref-
use Handling. 1977. Unpublished report.
PLJ^-Sellbergs, Stockholm.
18.	Midwest Research Institute. 1977.
Evaluation of Fabric Filter Performance
at Browning Ferris Industries/ Ratheon
Service Company Resource Recovery
Plant in Houston, Texas. Draft of final
report. U.S. EPA Municipal Environment-
al Research Laboratory, Cincinnati,
Ohio.
19.	National Fire Protection Association.
1954. Explosion Venting. SFVA No. 68.
NFPA, Boston.
20.	Kolltt, A. R., and E. T. Shervln. 1979.
Ad approach to energy attenuation of
explosive wastes in processing equip-
ment. In: Proceedings of the Sixth
Mineral Waste Utilization Symposljm.
Ed. £. Aleshln. I'.S. Bureau of Mines
and ITT Research Institute, Chicago.
21.	Nollet, J.R., and E. T. Shervin. 1979.
Causes and alleviation of exclusions
in solid waste shredders. Presented
at: 1979 Convention of the GRCDA.
San Diego. August 1979.
22.	Robinson,. V. D. 1979. Solid waste
shredder explosions: -nat do they have
in conmun. Solid Wastes Management,
22 (May 1979): 46-46.
23.	Sandoval, M. 1978. Shredders' explo-
sion potential: more awareness needed.
Waste Age, 9 (April I978J: 103-110.
?4. Shannon. 1, J., D. E. Flscus, and P.C,
Coroan. 1975. St. Louis Refuse Proces-
sing Plant: Equipment, Facility, and
Environmental Evaluations, EPA-650/2-
75-044. Environmental Protection Agen-
cy, Washington, D.C.
25. Sorber, C. A., and K. J. C-uter. 1975':
Health aid hygiene aspects of spray
irrigation. American Journal of Pubr
lie Health, 65 {January 1975}: 47-52.
2b. Sorber, C. A., et al. 1976. A study of.
bacterial aerosols at a wastewater ir-
rigation site. Journal WPCF, 48 {Octo-
ber 1976): 2367-2379.
27. Zaloah, R. G., S. A. Weiner, and J. L.
Suck-ley. J97£. Assessnent of Explosion
Hazards in Refuse Shredders. Report Jfo.
ERDA-76-71. U.S. Department of Energy,
Washington, D.C.
222
-------
VERHICOMPOSTING OF ftUNIClPAL SOLID WASTES
John F. Donovan, Albert B. Pincince, and Ira D. Cohen
Camp Dresser & McKee Inc.
ABSTRACT
This paper is sn engineering and scientific assessment of municipal solid waste mana-
gement by vermicomposting. Vermicomposting is the conversion of waste materials by
earthworm consumption to castings which may be used as a soil amendment. The cost of
vermicomposting is estimated to be high compared to those of alternative methods.
For this reason, and because of the lack of established market for the castings, it
is believed that conrcunit'les have available to them technologies that are more
attractive than vermicomposting.
Introduction
Vermicomposting is the degra-
dation, of organic wastes by earth-
worm consumption. Some species of
earthworm (Eisenia foetida and
Lumbrirus rubellus) thrive in
managed conditions on £ diet and
substrate composed almost entirely
of organic matter. They feed on
the wastes, consume a portion of
the- organic matter and expel the
remains as feces, or castings.
After the worms have fed on
the waste anc converted it Into
castings, they are usually
separated from the castings. Worms
can be recycled into new vermicom-
posting beds or, possibly, marketed
in some form. Castings, once dried,
have properties that might make
them a desirable soils amendment.
A portion of the wastes is usually
not biodegradable and simply remains
as residue for disposal. The end pro-
ducts of vermicomposting, there-
fore, are worms, castings, and
solid-waste residue.
The role of earthworms in
nature has been recognized since
ancient times and was studied
extensively by the biologist
Charles Darwin in the late 19th
century. Despite this awareness,
and despite the fact that success-
ful culture of earthworms need
involve no new technology, the
practices of raising worms
(vermiculture) and using them for
waste management (vermicomposting)
have not been advanced until
recently.
There are now numerous indivi-
duals, gardeners and entrepreneurs
raising worms in a soi1/peat/manure
bedding in indoor bins or outdoor
plots; most of those practicing
vermiculture depend heavily on the
ba1tK3rrc market to realize some
income from the business.
Only since 1.97.0 has the ver-
micomposting of wastes (including
municipal solid waste) been
attempted at more than backyard
scale. Pioneering efforts conmonly
223
-------
mentioned in the literature
(3, 4) include a demonstration in-
ject at Hollands Landing, Ontario
(Canada), which was begun in 1970
and has since been operated under
private ownership. A pilot-scale,
one-time denonstration of vermicom-
posting ixinicipal solid waste
(WSW), was conducted in 1975 at
Ontario, California. Neither
operation was conducted under tlie
controlled conditions that would
yield reliable engineering design
parameters.- The Hollands Landing
facility has vernicomposted small
amounts of manure, food-processing
wastes, and sludge (Klauck, per-
sonal communication)the Ontario,
California, project involved the
vermicomposting of municipal refuse
in a ; -jram jointly conducted by
the City and Worth American Bait
Farms Inc. The short-duration pro-
ject involved a total of 10 tons of
mixed municipal refuse, which was
hand-picked to remove glass, netal,
plastics, and rubber. The
remaining 9 tons were windrowed
adjacent to established earthworm
beds at a wornbreeding farm.
Reportedly, consumption of organics
was 90-percent comolete in 68 days.
Only castiiigs, bulky materials such
as tree limbs, and inorganic resi-
dues remained.
In a latt- experiment at the
same facility, a shredded and air-
classified organic fraction of
municipal solid waste was
transported from a local resource-
recovery facility. Reportedly,
consumption of the wastes was
faster, and more complete than in
the earlier experiments, even
though the process was hampered by
large quantities of plastic present
in the. waste and by very dry con-
ditions resulting from the
windrow's direct exposure to
sunlight. No documentation is
available on the amount of castings
produced during the Ontario tests
or their ultimate use.
Another demonstration of
solid-waste vermicomposting was
carried out in 1978-79 by the
American Earthworm Company (AEC) in
Florida (Wesley Logue, personal
conmuni cation). This project
involved vermicomposting, over a
12- to 18-month period, of about
500 tons of runicipal solid waste.
Wastes were truckea to a ]-acre
vermicomposting facility in
Sanford, Florida, where cans, large
objects, and newsprint were removed
by hand. The remainder was fed to
a hammer-mill shredder, which
reduced the wastes to a 3-in par-
ticle size. The shredded waste was
placed in windrows about 6 in deep,
which were irrigated to increase
moisture. Moisture levels achieved
and other pertinent data are not
known. AEC used approximately one
ton of earthworms in the vermicom-
posting operation. Reportedly,
some of the finished castings,
which contained glass, were uti-
lized by a local nursery; The.
facility is no longer in operation.
Other work has been carried
out in Japan, where some pulp and
food processing industrialists have
turned to vermicomposting tech-
niques for management of sludges
and waste byproducts (3).
Information was obtained through
two sources in the vermiculture
industry; Aoka Sangyo Co, Ltd and
Toyohira Seiden Kogyo Co. The Aoka
Sangyo Co. reports tney have' three
1,000-ton-per-month plants pro-
cessing wastes from pulp and food
processing companies (Shizuro
Aobuchi, personal comnunication).
The operation appears to be
labor-intansive, and the economics
appear to depend heavily on dispo-
sal fees paid by the industry.
Reportedly about 400 tons of
casting arid 10 tons of earthworms
»i»e produced per month. The earth-
worms are freeze-dried and sold as
fish feed. Worm castings are also
sold.
The Toyohira Seiden Kogyo Co.
reports that rice-plant straw,
municipal sludge, sawdust, paper-
nraking wastes, food-processing
wastes and manure are vermicom-
posted (Katsumi Yamaguchi, personal
conmunication). They estimate that
about 20 private companies with
224
-------
monthly capacities of 2,000 to
3,000 tons are in operation. An
additional 3,000 individuals may be
vermicomposting 5 to 50 tons of
wastes per month. However, these
estimates are only approximate as
the enterprises are not well-
organized.
In Europe, no cornnercial-scale
vermicomposting operations have
been reported in the literature. A
demonstration facility was recently
established in Modena, Italy,
however (Carla Chiesi, personal
connunication). Reportedly, a
screened and composted municipal
refuse is fed *.o Zisema foetida.
In several other curopean
countries, university laboratory
research in waste vermicomposting
is underway (5, 7).
Ogden, Utah Verm composting Pilot
Proqran
When this study was conducted, the
vermiconroosting pilot facility
operated at the Ueber County Refuse
Disposal Facility in Ogden, Utah,
by Annelidic Consumption Systems,
Inc., was the only operating ver-
micomposting facility in tne United
States exclusively utilizing muni-
cipal solid waste.
At the Weber County Refuse
Disposal Facility, which is
operated by Teledyne National,
about 350 tpd of nixed, residential
and conmercial waste; are burned,
before combustion, the wastes are
shredded to a nominal size of 6 to
8 in, and ferrous metals are
removed by a magnetic pulley.
In early August 1979, opera-
tors at the Weber County Facility
used front-end loaders to transfer
approximately 47 cubic yards of
shredded waste to three windrows in
an area prepared for a sanitary
landfill operation (Figure 1).
Based on a field measured density
of approximately 240 Ib/cu yd,
about 5.6 tons of wastes were
windrowed. The windrows, each
measuring 20 to 35 ft long and
about 9 ft wide, were spaced 10 ft
apart. The maximum depth of each
windrow was approximately 3 ft.
The windrows were watered to
increase the moisture content.
After several days, tem-
peratures began to increase in the
windrows due to bacterial breakdown
of organic matter (composting).
The elevated temperatures continued
through 27 August; windrow tem-
peratures on that day were measured
at about 32°C.
The windrows were wet down on
27 August, in order to drop tem-
peratures within the range at which
vermicomposting car. begin. (After
wetting, the windrows had a tem-
perature of 24°C). At this point,
some 1,450 lb of earthworms were
applied to the windrows. The worms
reportedly infiltrated the windrows
within a day.
On three occasions during the
next 35-day period, 1-ft increments
of shredded and unshredded wjstes
were added at the top of e.'.ch
windrow. About 36 cu yd [4.4 tons)
were added during this period; in
all, a total of 83 cu yd (and 10
tons) of solid wastes were ver-
micomposted. For the next 1-1/2
months, the earthworms converted
the waste to castings.
On 31 October 1979, four addi-
tional windrows were constructed
between existing windrows. The new
windrows contained about 13 tons of
solid wastes and occupied about
1,050 sq ft.. No new earthworms
were added to these windrows. Some
of the earthworms migrated from the
first thn>e windrows to the second
set of windrows in search of more
attractive food sources.
Facility workers began har-
vesting portions of the first three
windrows on 14 December 1979, 100
days after earthworms were first
added and about 130 days after
wastes were first windrowed. In
the judgment of Ronald E. Gaddi of
Annelidic Consumption Systems,
225
-------
Inc., 90 percent of the wastes had
been converted ai this point.
Based on the screening results
obtained later, however, a much
lower percentage had been con-
verted.
Harvesting Has accomplished using
an inclined cylindrical, rotating
screen driven by a small motor, a
type of device coiimonly used in
vermiculture (Figure 2). In normal "
operation, this type of screen
separates castings (which fall
through the screen) from earthworms
and residual so)id waste (which
travel the length of the screen and
are discharged). Because of the
large amour.l of plastic in the
waste stream, however, and perhaps
because not all wastes had been
shredded properly prior to
windrowing, the harvesting screen
did not work very well. It was
apparent that less than 50 percent
of the material fed to the device
was screened out as castings. Most
¦of the material either had to be
removed by hand from the front of
the.machine, which clogged
repeatedly, or was discharged
through the length of the harvester
as residue. Very few earthworms
were separated. Figures 2 also
shows the castings recovered and
the residue discharged. Figure 3
is a^closeup of the discharge end
of the screen.
In the spring of 1980, new
windrows of prepared solid wastes
werd: placed adjacent to the
existing rows. It was anticipated
that the worms would migrate into
the ne material and begin producing
castings there. The area was
struck with heavy'rains at about
that time, and the worm populations
migrated away from the piles. At
that point, the project was aban-
doned.
In the literature, estimates
of worms' performance in *ermicom-
posting vary widely. Entrepreneurs
working in the vermiculture
Industry routinely report that
worms will consume one-half to
twice their weight in waste each
day. To provide a rational basis
for the design of facilties, we
calculated d consumption rate based
on.the experience at Ogden.
At Ogden, 10 tons of wastes
were processed during a 110-day
period using 1,450 lb of earth-
worms. This time does not include
3 weeks of composting prior to the
introduction of earthworms, even
though this preliminary composting
is essential in ordor to maintain
optimum temperatures during ver-
ni composting. The calculated con-
version rate of vermicom^osting at
Ogden is, therefore, about 0.13 lb
waste processed/lb of earthworms
per day.
Area reouirements for vpr-
micomposting are determined from
the ratio of the weight of earth-
worms utilized per unit of area.
At Ogden, 1,450 lb of earthworms
were applied to about aOO ft of
windrows. The resulting worm:
area ratio of 1.8 Ib/sq ft compared
with ratios used in vermicomposting
of wastewater sludge of 0.4 to 2.3
lb/sq ft. (I).
Based on the conversion rate
(0.13 lb wastes/lh earthworms per
day) and area requirements (1.8 lb
earthworms/sq ft of wastes)
observed at Ogden, a loading rate
of 0.23 lb/sq ft per day is calcu-
lated. Stated another way, each
ton/day of waste requires about 0.2
acre, plus about 20 percent to
allow for composting before worm
addition.
In practice, earthworms would
be added to the initial set of
windrows. After 130 days, new
windrows would be constructed bet-
ween the "existing windrows. This
allows the earthworms originally
added to the first set to migrate
to ..the second set of windrows.
Then the initial set of windrows
are remo\,d and screened. In this
way. the land is used in 130-day
cycles and earthworms are reused.
226
-------
figure 1:
(left) Oaden, Utah, Vermicompocting
Pi lot Facility
Moure 2:
Firure 3:
(left) Rotary Screening Device
(above) Solid Waste Residue
-------
ACS speculates that under more
favorable climatic conditions,
and/or with an earthworm thit is
acclinated to solid wastes, oily 7U
djys would be required for conver-
sion [Gaddie, personal
co
-------
PRE-
PROCESSING
r^1 I ir—
TV
I Afc.')Hl | 
-------
vbi.A1ii.i2tD
MUDS
SIT
VftHtCO^OSTNG
FIGURE 5 t/ERMJCOMPOSTING MATERIALS BALANCE
i KS£0 0* rSTi.*ArF$orvti0PfOftrAiimfLi0(C CONSuUMtO* trs-fv; IKC 1
•OSWC LOSS
AMD
VOiATiLllCD
SOU OS
WtDC^C '
U
* 'tww r
| ltt«OVk '
?Vt«MiUaiP05T*G '
^ JtT
• «esioc£
#«OOOCT
FIGURE 6. VEAMlCOMPOSTiNG MATtRIALS BALANCE
230
-------
remaining 8 tpd would consist of
ferrous material recovered for
sale. (Ferrous recovery may range
between 6 and 10 percent, depending
on local waste composition and
facility design and operation.)
The vermiconposting fe;tluy would
also receive the 92 tpd of shredded
wastes with ferrous material
reinoved from the 1.2Q0 tpd faci-
lity.
Of the 92 fxJ diverted to ver-
micomposting, about 31 tpd --
including glass, rubber and wood
wastes -- would be landfilled. The
amount of castings produced would
be 70 percent of the remaining por-
tion, or 43 tpd. The other 30 per-
cent (18 tpd) would be volatilized
or evaporated..
It is important to note that
none of the abo.ve estimates has
been vorified at actual demonstra-
tion facilities. The rotary
screening operation at Oyden has
not yielded tne quantities of
castings used for this report. As
noted above, however, much of the
work at Ogden has been carried out
at less than optimum conditions and
with some unshredaed wastes pre-
sent. The materials balance is
optimistic in order to provide the
most favorable economic analysis.
Preprocessing facilities
Preprocessing facilities can
be subdivided into units for
receiving, shredding, magnetic
separation, and storage of solid
wastes. Figure 4 shows prepro-
cessing,' vermicomposting and resi-
due disposal facilities for the
100-tpd facility. For the 1,200-
tpd facility, separate prepro-
cessing would not be required.
Shredded wastes would be introduced
directly to the windrows.
Collection and transfer of
municipal solid waste would ter-
minate at a receiving building,
where refuse would enter the
preprocessing system. There, the
transfer-haul trucks would tip
their solid waste onto the floor,
where a front-end loader would
stockpile it into the center of the
building. Large bulky items, such
as tree stunps aid white goods,
would be sorted out of the process
stream by front-end loader, prior
to dumping onto conveyors. Hand-
sorting could be performed during
the conveyor stage to eliminate a
large portion of other unpro-
cessable items.
Refuse then would pass into
the shredder building. Particle
size of the shreddod product could
be varied to meet the requirements
of vermicomposting. A shredder
would be expected to reduce refuse
to a nominal 4- to 6-inch size.
Annual maintenance costs for the
shredder ai;d building would be high
in comparison to other equipment,
due in part to the relatively fre-
quent occurrence of damaging explo-
sions.
For the 100-tpd facility, only
ferrous removal could be economi-
cally justified. Aluminum recovery
may or may not be justified, but
its exclusion from this analysis
will not affect the cost of vc-
roicomposting relative to other .
methods. Further preprocessing,
including the mechanical removal of
jiass and paper is not economically
feasible at a snail facility.
In the smaller facility (100
tpd), the shredded product would be
bunkered in a passive, three-walled
tipping floor arrangement. The
waste would then be fed to trucks
by front-end loader and transported
to the vernicomposting facility.
For the larger, l,20C-tpd RDF
facility, the processed waste (RDF)
would pass from the shredder
building via conveyor to a storage
bin. This facility would incor-
porate live bottom hoppers to feed
conveyors which, in turn, would
feed semi-suspension boilers' or
transfer-haul vehicles. Prior to
reaching the storage bin, approxi-
mately 92 tpd of processed waste
231
-------
would be redirected via a conveyor
into trucks dedicated to hauling
the processed waste to the ver-
rciccmposting facility.
Vemicomposnrtq and Residue
Disposal Facilitics
The 92-tpd remaining after
removal of ferrous material would
yield a volume of about 900 cu
yd/day at a density of approxima-
tely 200 lb/cu yd.
Figure 4 shows the major steps
in vermicomposting: windrowing and
screening. Shreddeb wastes would
.be trucked to nearby windrows and
spread by a second smaller loader.
This method would be less expensive
than utilization of loaders alone
because of the distance and tiine
required to travel to the windows
in a .typical operation.
To determine the area required
for vermicomposting, the assunp-
tions presented above are followed.
Total residence tine would be 110
days, resulting in a loading rate
of 0.23 lb waste/sq ft per day. An
additional 20 percent is added to
the area to provide space for pre-
liminary composting. For a daily
average loading of 92 tpd, the area
requirement would be 23 acres,
including land for an access road.
One possible site arrangement
would be a configuration approxima-
tely -1.000 ft square. About 9S
5CQ-ft-long windrows could be
accomodated on each side of a
central access aisle. Each windrow
would be 10 ft wide and 3 ft deep.
Initially, wastes would be
windrowed in every other row.
About three 5U0-ft windrows would
be constructed daily, and, during
the first month after earthworm
addition, three additional top-
dressings of wastes would be
applied to each windrow, as at
Ogden. After one-half of the total
area for windrows was constructed,
shredded wastes would be windrowed
on the alternate rows. The site
would be equipped with a sprinkler
system for initial moistening of
the windrows to achieve the proper
noisture content.
Based on pilot-plant
experience, earthworms would be
added to one-half of the windrows
(480,000 sq ft) at a rate of 1.8
Ib/sq ft; a total of approximately
430 tons of earthworms would be
required for this initial earthworm
addition. Earthworms recovered
during harvesting would be reused
for new windrow construction. Some
of ttiose working in the field (6) have
suggested that excess earthworms
could be produced during vermicom-
posting, but more research is
needed to determine the rates at
which earthworms will breed under
conditions of vermicomposting. In
order to reduce capital costs,
several vermicomposting operators
have suggested that inly one-fourth
to one-third of the total required
stock night be purchased initially,
with the process phased into full
operation over a period of several
months as excess earthworms are
produced. For this report, we have
assumed that one-half of the total
area would be stocked with
purchased earthworms because of the
lack of data on eartl *orm produc-
tion during vermicomposting. Even
if minimal costs were included for
earthworm purchase, the net costs
of vermicomposting would be relative-
ly unchanged with respect to other
alternatives.
After a total residence time
of 130 days, the wastes would be
recovered from the windrows by a
front-end loader. At this point,
nearly all of the biodegradable
portion thai can be converted by
earthworms would have been con-
sumed. Approximately 450 cu yd (74
tons) of material would be removed
daily. The wastes would be
transported by front-end loader to
a movable surge hopper located
above a central collecting con-
veyor, which would move material tc
the screening area.
Two rotary harvesting screens,
each approximately 6 ft in diameter
233
-------
and 12 ft long, would be employed;
one screen would serve as a
standby. The screens would be fed
directly by the variable-speed con-
veyor. Castings would fall through
the screen to a product storage
pile. The operation could be
expected to produce approximately
300 cu yd {43 tons) of castings per
day. Residual waste would be
discharged at the low end of the
Inclined screen to a conveyor,
which would remove the material to
another storage pile. These wastes
--totilling about 150 cu yd, or 31
tons of residue per day could
then be recovered by a front-end
loader and trucked to the landfill.
The harvesting screen would
also be able to collect earthworms
for reuse. The compacted residue
volume would be about 19 acre-ft
per year. Based on a total lift of
about 10 ft, about two acres of
landfill-would be required each
year. For a 20-yr period, a 38-
acre landfill site would be
required. The site might be devel-
oped in 5- to 10-acre modules.
Costs
The costs of vermicomposting
municipal solid wastes can be
divided into three components:
preprocessing, vermicomposting, and
residue disposal. The costs are
shown in Table 1.
Preprocessing costs include
receiving, shredding, and storage
facilities. About two-thirds of
the capital costs of vermicom-
posing is for purchase of the ini-
tla1 stock of earthworms.
Amortization of capital costs
accounts for about 60 percent of
the total annual costs of vermicom-
posting. Costs for residue dispo-
sal Include development and
operation of sanitary landfill.
The net costs of vermicom-
posting include the credit for
revenues generated by the sale of
ferrous metals and earthworm
castings. Ferrous metals can
currently be sold for $30 to $50
per ton. Based on a recovery of
2,500 tons per year, the revenue is
between $75,000 to $125,000 for the
100-tpd vermicoiaposting facility.
There are no existing markets
for the sale of earthworm castings
derived from municipal solid waste. .
Several vermicomposting operators
have suggested potential inarket
prices. Estimates range from $0 to
$15 per ton, reflecting the lack of
marketing experience, eased on
13,400 tons of castings recovered
per year, up to $200,000 in annual
revenues might be. accrued. Taking
Into account potential product
revenues, the total costs of'ver-
mi composting are about $750,000 to
$1,000,000 or approximately 524 to
$32 per ton processed.
For a municipality producing
100 tpd of solid waste, three
alternative methods of solid waste
management are a sanitary landfill,
combustion in e, modular combustion
unit and windrow compostiny. Costs
of these methods compared to ver-
mi composting are shown below:
Approximate Net
Method	Cost ($/ton)
Sanitary Landfill	6
Hodular Combustion
Unit	15
Windrow Composting	24-28
Vermicomposting	24-32
Thp ..micomposting process is
at the high end of the scale. In the
case where a. portion (100 tpd) of
the solid waste from a large com-
munity (1200 tpd) would be ver-
micomposted, vermicomposting would
not be attractive, primarily due to
substantial lost revenues from
reduction in the refuse/fuel
stream.
Findings
Vermi composting of municipal
solid waste is not an economically
feasible technology. Although the
technique appears to offer the
233
-------
TABLE 1. TOTAL COSTS OF VERMICOMPOSTING 100 TPD
MUNICIPAL SOLID WASTES

Capital
Annual
Preprocessing Facilities
Vermicomposting Facilities
Residue Disposal
Sub-Total (rounded)
SI,800.000
3,150,000
575,000
$5,500,000
$ 450,000
520,000
95,000
SMOO.OOO
Revenue:


Ferrous Metals
Earthworm Castings

S 75,000-125,000
0-200,000
Net Cost (roundeo)
$5,500,000
il,000,000-750.000
Cost per Ton

$32 - $24
advantage of resource recovery, it
has several major disadvantages
compared to other management
methods. Briefly, some of the
disadvantages are:
o Expensive preprocessing
required (shredding)
o Large land area requirements
(23 acres of windrows for a
100 ton per day (tpd) facility)
o High capital costs <55-5
million for a 100 tpd facility)
o High net unit costs ($24
$32 per ton)
o An unknown product market
(for castings)
. Acknowledgement
This paper is based on a study
for the U.S. Environmental Protection
Agency (Contract No. 68-03-2803, Office
of Research And Development, Municipal
Environmental Research Laboratory) \2).
Laura A. Rinyenbach was the project
manager for EPA.
References
1. Camp Dresser & MctCee Inc.
1980. tngineering assessment
of vermicomposting municipal
wastewater sludaes, Municipal .
Environmental Research
Laboratory, Office of Research
and Development, U.S.
Environmental Protection
Agency, Cincinnati, Ohio (in
press).
2.	Camp Dresser S McKee. 19»0.
Compendium on solid waste
management by vermiconposting.
EPA-600/8-80-033, U.S.
Environmental Protection
Agency, Cincinnati, Ohio. 61
pp.
3.	Carmody, F.X. 1978.
Vermicowposting: An
assessment of the state of the
art as of May, 1978. Privately
published manuscript. 27 pp.
4.	Gaddie, R.R., Sr., and D.E.
Douglas. 1977- Earthworms
for ecology and profit, Vol
11. Bookworm Pub)ishing
¦ Company, Ontario, CA. 254 pp.
5.	Gaff, 0. 1980. Preliminary
experiments of vermicomposting
of different waste materials
using Eudrilus euqeniae
Hinberg. Proceedings of
Research Needs Workshop on the
Role of Earthworms in the
Stabilization of Organic
Residues. Kalamazoo, Michigan,
April 9-12, 1980.
ZJ4
-------
6.	Grobecker, D.W. (for Annelidic
Consumption Systems, Inc.).
1978. A feasibility study:
the earthworm conversion of
municipal solid waste.
Bookworm Publishing Company.
Ontario, CA. 23 pp.
7.	Hunta, V. 1980. Results of
preliminary experiments
culturinq Eisenia foetida on
different types of sewage
sludge, animal and human
excreta mixed with low nitro-
gen organic materials.
Proceedings of Research Needs
Workshop on the Role of
Earthworms in the
Stabilization of Organic
Residues. Ka'amazoo, Michigan
April 9-12, 1980.
8. f(ew York Times. Sept. 1977.
Japanese industrialists and
fanners look to earthworms for
help.
235
-------
ENERGY AND MATERIALS RECOVERY FROM
HUKICIPAL SOl.'D WASTE
Joan D. Finch
Building Research Advisory Board
ilational Research Council
Washington, D.C. 20418
ABSTRACT
In this study, selected resecrch and development documents dealing with
energy and materials recovery from municipal solid waste are reviewed and
synthesized by a Committee of the Building Research Advisory Board
(BRAB). Eleven review documents represent projects sponsored by the
U.S. Environnerita J Protection Agency and other federal agencies over a
live year period. In its Teport, the JRAB Committee presents a summary
of the documents reviewed, a discussion on the stste of the art of
resource recovery implementation, impediments to implementation, on-going
and needed research, and conclusions reached by the Committee as a result
of its study.
INTRODUCTION
As a result of tne Resource
Conservation and Recovery Act of
1976 (PI 94-580, 1976) numerous
project have been sponsored by
various agencies of the federal
government to.prorotc the pro-
tection of health and the envi-
ronment and to conserve valuable
material and energy resources.
Considerable attention has been
given to the investigation of
existing and emerging technol-
ogies ioi recovery of resources
from solid waste and to the
potential for implementation of
recovery systems. As is usually
the case, investigations were
undertaken by a variety of inves-
tigators under individual con-
tracts or grants with several
federal agencies and dissemi-
nation of-the results of these
studies was limited and ad hoc in
nature. Thus no collective know-
ledge was available to assess the
elate of the art and the conclu-
sions and recommendations arrived
at through these research,
development and demonstration
projects. The Environmental
Protection Agency Selieved that a
review and synthesis the results
of these studies would provide a
valuable information service and
asked that .fhe Building Research
Advisory Board undertake the task
Conduct of the Study
Under contract to the
Environmental Protection Agency,
an ad hoc conmittee was appointed
in accordance with the policy am.'
procedures ot the Rational
Research Council to review and
synthesize eleven research and
development documents selected by
the sponsor. The Tepnrts
selected for review were:
1.	Energy Conservation Haste
Utilisation Research and
Development Plan.
2.	Energy and Resource Recovery
from Solid Waste
236
-------
3. Materials and Energy from
. Municipal Waste.
A. . National Recycling Research
Agenda Pro ject.
5.	Waste Resources as a
Fotential Topic for.
Integrated Basic Research.
6.	Unit Operations in Resource
Recovery Engineering.
Resource Recovery.
J. Present Status and Research
Needs in Energy Recovery from
Wastes .
8. -Study of Processing Equipment
for Resource Recovery Systems,
Volume 1 - State of the Art
and Research Needs.
Study of Processing Equipment
for Resource Kecovory Systems,
Volume II - Magnetic
Separators.Air Classifier and
Ambient Air Enissir~s Tes t s.
Viocessing Equipment for
Resource Recovery Systems,
VoHn:' 111 - Fie Id 1 est
Evaluation of Shredders.
5. Resource Recovery Research,
Development and Demonstration
P1 an.
10.	Wsste-to-Energy Technology.
11.	Fuela and Feedstocks from
Solid Waste.
The review documents were
reviewed by the Coonrittce,
abstracts of each were prepared
and the Cocrmi11ec 's report was
developed. Included in the
report is n discussion of' the
rationale for energy and
materials recovery fron> tiunicipal
solid waste; a discussion of the
state of the art in terms of
implementation of existing
techniques; impediments
to resource recovery; and
research and developments needs.
The Committee also presents its
findings and conclusions based on
the documents reviewed and the
collective expertise represented
by the Committ»c.
Committee on Energy and Materials
Recovery from Municipal Solid
Waste
MICHAEL B. BEVER, (Chairman),
Massachusetts Institute ot
Technology, Cambridge, XA
LUCIAN C, BIEL1CKI, President,
Huni Recovery Systems, Inc.
White Plains, NK
JACK DUNN, Black S. Veach,
Bethesdo, MD
HARVEY D. FUNK, Vice President
and Director, Henningson,
Durham & Richardson, Inc.,
Omaha, NB
CHAR1ES A. JOHNSON, Technical
Director, National Solid Waste
Management Association,
Washington, DC
CHA6LF.S0. mZY, Tresider.t,
Charles R. Vel?y Assoc., Inc.,
Artoonk, NY
237
-------
RECYCLING IN THE UNITED STA ITS:
TOE VISION AND THF. REALITY
Oscar V. Albrecht
U. S. Environmental Protection Agency-
Cincinnati, Ohio 45268
Ernest H. Manuel, J-., and Fritz W. F.faw
Mathtech, Inc.
P. 0. Box 2392
Princeton. N. J. 08540
ABSTRACT
Nationa.' attention on the growing solid waste problem was first evidenceu by cpnctmcnt
of the Solid Waste Disposal Act in 196S. Subsequent legislation relating to soliU vastc
management and resource recovery has been passed by the U.S.-Congress, the roost recent
being the Energy Security Act in 1980. Public anticipation that recyc 1 ing. ;>i 11 eventually
solve the waste management and conserve limited natural resources ha« persisted for more
than a decade now, but progress toward these goals has fallen considerably behind initi.il
expectations. Various reasons tor the lag in recycling have been offered, and federal
research efforts have concentrated particularly on solution of the technological difficul-
ties encountered by recycling plants.
The lini.ted States Environmental Protection Agency in 1978 contracted with Mathtech,
a division of Matheisatica, to evaluate the impediments to recycling and their economic
effects- Facilities selected for indepth study represented a broad cross-section of re-
cycling technologies, rwnersh;p types, geographical locations, and recovered products.
Adjustments for differences in accounting procedures and noriral i:at ion of data were made
and econosic performances for eight plants compared on a net cost of disposal basis.
Performance cased on cost of disposal permitted comparison with alternative modes of
disposal, mainly landfills.
Based on the recent financial data, none of the facilities were found to be economi-
cal in the sense of providing the l^ast cost mode of disposal for municipal solid waste:
all facilities experienced net losses in operations. Indications were that performance
for the projected year (1979) would improve somewhat allowing at least one facility to
realize a slight profit. A major deterrent to recycling was the predominance of lower-
cost landfills. Competition from landfills is expected to diminish in some areas when
new environmental regulations for land disposal of solid waste become effective. Results
of the Mathtech study have the usual limitations inposed by sampling, missing data, and
possibly computational errors. The results, however, provide useful information to
policy-makers and decision-makers having responsibility for efficient management of munici-
pal solid waste.
Introduction
In 1970 Richard M. Nixon, President of
the United States, wrote "We can no l'nger
afford the indiscriminate waste of our nat-
ural resources; neither should we accept as
inevitable the mounting costs of waste re
noval. We must move increasingly toward
closed systems that recycle what now are
considered wastes back into useful and
238
-------
productive purposes ."(1) Soo:i afterwar-is,
in October 1970, the U.S. Confess enacted
the Resource Recovery Act of 1970 which in-
cluded among its purposes the demonstration,
construction, and application of solid
waste management and resource recovery sys-
tems . (?)
A number of issues concerning recy-
cling ..?source recovery systems for
nuniciial solid waste arose in the Oecade
that followed. The federal government
sponsored numerous symposia, funded a long
list of research and demonstration projects,
encouraged industry, and gave orders to ¦
federal agencies to implement recycling
activities. Th«? Resource Conservation and
Recovery Act, enacted in 1976, established
a Resource Conservation Committee comprised
of representatives from various departments
and agencies of the federal government.1-')
The reci-nt passage of the Energy Security
Act of lf>S0 which authorizes federal price
supports.and loans for waste-to-energy
activities constitutes vet another step in
the continuing efforts to increase recy-
cling.(4)
This paper is limited to an appraisal
of the problems with recycling municipal
solid waste. In particular, it discusses
the findings of recent case studies con-
ducted by Mathtech under contract to the
U.S. Environmental Protection Agency. The
discussion of these findings will hopefully
clarify many issues on recycling and in-
crease the understanding of local officials
and others interested in resource recovery
programs for municipal solid waste.
Optimism Changes to Pessimism
The initial expectations were that re-
cycling would reduce or elimir.ate the
mounting solid waste prob'ci" :»nd also con-
serve valuable resources. v5) Newspaper
articles containing captions such as
"American Trash Could Be Bonanza As Source
Of Ore" led the public to believe that re-
cycling was an immediate and forthcoming
answer(6)Carlsen in Environmental Affairs
stated that "recycling will soon become an
increasingly important phase of economic
activity. It represents the solution to
both the problems of increasing solid waste
with its attendant insatiable demands of
land and those associated with the growing
scarcity of natural resources." (7)
Not all prognosticators agreed with
the initial optimism expressed by some.
Early in 1971 the privately supported
National Center for Resource Recovery cau-
tioned that managing municipal wasle
through resource recovery might be 5, 10,
or even 20 years away.(3) The Institute
of Scrap Iron and Steel estimated that at
the end of 197S unrecycled ferrous scrap
including that discarded by households
could cover the Nation's capital 100 feet
deep.(9) Observers became more pessimis-
tic. Sylvia Porter in the Sew York Post
questioned whether recycling was in a tail-
spin. 110) The Solid Waste Management/
Refuse Removal Journal, a trade publica-
tion for the solid waste industry noted
that the initial experiences consisted of
"delays, breakdowns, cost overruns, mis-
conceptions". fl 1)
From Materials to Energy Recovery
The energy crisis in late 1973 shifted
attention away from materials to recover-
ing the potential energy value in munici-
pal solid waste. Some facilities were
subsequently designed solely for energy
recovery (Table 1). As with materials
recovery,reports of success were soon
dampened by reports of failure. The May
1978 issue of .American City and County
published the article "Despite higher costs
than expected, this Iowa City's refuse-to-
energy project is still going strong after
18 months of operation".(1 -) The New York
Times reported "L.I. Town a Tioneer in
Recycling Its Garbage Into Electric
Energy". (15") Meanwhile, a private consul-
tant to the industry declared at an annual
meeting "...our dreams have taken us into
what now seems to be the clouded skies of
resource recovery - of refuse to energ/ -
of burning to earning. Ke are not facing
up to the truth about the expense, -in-
efficiency and impracticability of the
waste recovery projects that make Monday
morning headlines..." (li)And Dr. Rocco A.
Petrone, president of the National Center
for Resource Recovery, wrote "even now,
with 30 plants operating, under construc-
tion, or well past commitment, resource
recovery has not reached the operating
stage." (15) Three years after the multj-
million dollar waste-to-energy recovery
facility began construction in 19~6 in the
city of Chicago, Hasterbrbok observed that
the SSFF (Southwest Supplemental Fuel
239
-------
TABLE 1. CHARACTERISTICS OF PLANTS RECYCLING
MUNICIPAL SOLID WASTE IS U.S.*
Resource Recovery Activitv

Start ing

Investment
Throughput'
Materials
Energy
Facil ity
Date
Status
(mil's)
¦"Design
Oper
Design
Operat.
Design
Opera
Altoona, PA
1963
op
..
SO-
..
X
X


Ames, IA
19"5
OP
6.3
400
200
X
X
X
X
Ansonia, CN
1975
op
--
200
250
X
X


Baltimore City, MD
19T5
op
30.1
1000
600
X
X
X
X
Baltimore Co., MD
1976
op
10.0
1500
SSO.
X
X
X
X
Blytheville, ARK
197S
op
0.8
SO



X
X
Braintree, MA
'371
op
3.3
240
186


X
X
Chicago(Crawford) IL
1977
st
20.0
1000

X

X

Chicago(SW), IL
1971
op
23.0
1600
1000
X

X

Chicago(SN), IL
1963
cl
6.8
1200

X
X
X
X
Crossville, TEN'S'
1978
op
--
60
— *


X
X
East Bridgewater, MA
1976
op
12.0
1200
S50 .
X
X
X
X
Franklin, Oil
1971
op
3.4
150
50
X
X


Croveton, NH
1975
op
0.2
30
--


X
X
Hsrrisburg, PA
1972
op
14.1
720
430
X
X
X
X
Hempstead, SY
1979
sh
81.1
2000
...
X

X

Lane Co., ORE
1978
ih
5. U
SCO
"
X
X
X
X
Leviston, ME
19"
op
1.5
130

X



Madison, KIS
1979
op
2.5
400
200
X

X

Menlo Park, CA

cl

100
--
X
X
X
X
Miami, FLA
19S6
cl
5
900
--


X
X
Milwaukee, h'IS
1977
op
28.9
1600
640
X
X
X
X
Monmouth Co., NJ
1976
op
3.6
400
375
X
X


Mt. View, CA
1978
op
0.8
1
0.5


X

Nashville, TENS'
1974
op
24.S
1060
400


X
X
New Castle, DEL
1972
op
2.4
800
300
X
X
X

Sew Orleans, LA
1976
op
9.1
700
650
X
X


Norfolk, VA
1967
op
2.2
360
360


X

No. Litt!-. Rock, ARK
1977
susp
l.S
100
62


X
X
Oscelo, AKk'
1978
op
--
25
--


X
X
Pompano Beach, FLA
1978
sh
3.6
75

X

X

Portsmouth, VA
1976
op
--
160
115


X
X
St. Louis, MO
1972
cl
:.s
300
--
X
X
X
X
Salem, VA
1978
op
1.9
100



X

San Diego Co., CA
1977
susp
15.0
200
0
X

X

Saugus, MA
19'6
op
50.0
1200
950
X
X
X
X
Siloau Springs, ARK
1975
op
0.4
20
16


X
X
Taeoma, WASH
1978
op
2.5
500
—
X
X
X
X
Source: . V^^ous government reports, trade publications and newspaper articles.
* Excludes plants in planning or construction stage in 1978.
S In.standard tors except Mt. View facility is in million standard cubic feet daily(nscfd).
*• Op = operating; sh = shakedown stage; cl = closed'; " st.= starting; susp = suspended.
. + Dollars expended in year of investment.
240
-------
Facility) was still in ' •jhaVcdowu1 , and
"the city may be better off without it."
(16)
¦Late in 1977 and early 1978, a t'.S.
congressional comnittee visited 8 recovery
sites to observe the state of the art in
recycling. The committee subsequent!/
issued a report that (1) claims made in
behalf of recycling were overstated and
costs understated, f2) technologies needed
further deveJopaent, and (3) the economics
for recycling would become more or able
as the costs of disposal rise. (1"]
Conflicting reports on the success or
failure 6f recycling materials and energy
have local officials and the public in
general greatly confused over what projects
should be undertaken. The confusion stems
from issues over lack of demand and avail-
able markets for reclaimed materials, in-
adequate and undependable supply of waste,
conflicting public policies such as tax
laws 3nd transportation regulations favor-
ing extraction and use of virgin materials,
institutional impediments, and the failure
of markets to fully reflect environmental
externalities iri land disposal of waste
arid increasing scarcity of mineral de-
posits. These issues form a mixed bag of
social, economic, technological and in-
stitutional problems that inhibit the
growth of recycling.
EPA-ltothtech Case Studies
The U.S. Environmental Protection
Agency (Office of Research & Development)
initiated research in 1978 to identify and
evaluate the impediments to recycling
nunicipal solid waste. (13) Answers were
sought to three basic questions: 1) Are
recycling facilities in operation economi-
cal? 2) If not, what are the impediments
to economical operation? and J) Khat
changes are needed for economical opera-
tion?
The context in which the term "eco-
nomical" was applied differs from the
theoretical definition which includes the
consideration of social costs and benefits.
As employed in the Mathtech study, the
tern was limited to an accounting compare
ison of private revenues and costs,
synonymous-with successful operation from
a private market perspective. The evalua-
tion c>f "economical" was restricted to the
above definition because of the many
difficulties involved in quantifying social
costs and benefits.
Selection of Facilities
Evaluating the financial performance
of facilities required accessibility to
the financial records and operating ex-
perience of sufficient duration for rep-
resentativeness of probable long-term
results. The facilities selected for case
studies were screened from 3 total number
of plants believe! to have progressed
beyond the "startup" or "shakedown" stage.
This greatly l.mited the number available
for studv. Privately owned plants -and
those with operating difficulties tended
to be mon1 reluctant to provide access to
data. Table 1 lists the facilities
initially designed for.energy and materials
recovery and indicates whether these in-
tentions were realized at the time of the
study. Whether a plant was in "startup"
or "shakedown" stage appealed to be arbi-
trary in some cases. Capital investment
included initial and subsequent capital
costs, including expenditures for pollu-
tion control equipment.
The selection of eight facilities was
based on criteria requiring at least one
year of operating history and accessibi-
lity to adequate financial records. These
facilities were selected to obtain a valid
representation of technologies and actual
financial experiences in recycling, given
the selection restrictions previously
mentioned. About half of the facilities
were situated'in the densely populated
Northeast where intense land use and
greater waste disposal problems exist.
Half of the facilities were owned and.
operated hy ounicipalities, representing
the type of government managing the major
portion of residential waste in the United
States. The remaining facilities included
ownership or operations by a county, a
large private corporation, an investor-
owned utility, and a joint venture by one
county and small private corporation. All
of the sample facilities were constructed
within the past 10 years and included a
variety of technologies currently in use.
Plant capacities ranged from 100. to 1600
tons per day (tpd), a range generally
considered adequate for the disposal re-
quirements of communities with populations
between SO,000 and one miliion.
241
-------
Significance of Impediments
Impediments to recycling were grouped
according to effects on: 1) quantity and
composition of solid waste supply, n cap-
ital cost and facility design, 3) operating
costs and efficiency of operations and
J) market demand for energy and materials
reclaimed. The impediments within these
broad categories are listed (not necessaii-
ly in the order of their importance) in
Table 2. The significance of these im-
pediments are discussed in the following
sections.
At most facility Locations, land dls-
po>al was a lower-cost alternative to re-
cycling. The few exceptions to this gener-
al rule included one location where all
landfills within the county had been
closed, and several locations where an
adequate inflow of waste was assured
through contractual arrangements. The
majority of recycling facilities were un-
able to effectivel) compete with nearby
landfill charges. Tipping fees were
generally below actual operating costs in
order to attract sufficient quantities of
waste; even so, a number of plants were
unable 10 attract sufficient quantities to
match the design capacity for the facility.
In the United States waste collection
is mostly organized by local jurisdic-
tions. i(19) This dispersion of authority
resulted in less waste available for re-
TABLE 2. CATEGORIES OF IMPEDIMENTS ACCORDING TO EFFECTS
ON RECYCLING OF MUNICIPAL SOLID h'ASTL
1.	Restrict quantity and/or composition of solid waste supply.
a.	Pricing land disposal of solid waste below full social cost.
b.	Source separation programs.
c.	Mandatory deposits for beverage containers.
d.	Fragmentation of local solid waste management/authority.
e.	Seasonal and business cycles.
2.'	Increase capital cost and/or restrict facility design.
a.	Inadequate existing technology.
b.	High prices for equipment and/or parts.
c.	Limited facility design experience.
d.	Restrictions on muriicipil financing.
e.	Low bid approach to public contracting.
3.	Increase operating costs and/or decrease efficiency of operations.
a.	Limited managerial and/or operating experience.
b.	Limited availability of spare parts.
c.	Shortage of skilled labor.
d.	State and federal regulations for air quality.
e..Regulations	for disposal of ash and residue.
4.	Diminish the market potential for reclaimed energy and/or materials.
a.	Facility siting restrictions,
b.	Franchise restrictions on energy sales
c.	Tax subsidies for extraction of virgin materials.
d.	Regulated prices for oil and gas.
e.	Lack of future markets for reclaimed (secondary) materials.
f.	Unpriced social costs for extraction and processing of virgin materials.
g.	-Rate regulation of electric utilities.
h.	Restrictions on use of ash, residue, and shredded materials.
i.	Product labeling restrictions.'
j.	Government procurement policies.
242
-------
cycling than desired at several of the
case study locations. The effect of re-
duced volume on cost of operations is
illustrated in Tabic 3.
Community programs involving source
separation and mandatory beverage container
deposits appeared to have only a negligible
effect on facility performance. These
programs were varied but potential losses
of revenues were generally insignificant.
The fraction of aluminum recover^, was
generally small. Seasonal fluctuations in
waste supply caused minor rescheduling of
winter operations at one plant but did not
significantly affect operations at the
other facilities.
Recycling operations were hampered by
technological problems in at least half of
the facilities studied. Technology prob-
lent, including facility designs and limit-
ed operating experience, are usually
symptomatic of a new activity as it pro-
gresses from pilot plants and demonstra-
tion projects to commercial sire opera-
tions. Facilities designed for RDF
(refuse-derived fuel) and recovery of
materials had relatively greater tech-
nological difficultie. , mainly because the
materials handling processes were more
complex than for other recycling systems.
Environmental regulations also tended
to impede recycling efforts. Conversions
from coal to oil during the 1970's due to
air pollution regulations resulted in a
greatly reduced number of boilers avail-
able for co-firing RDF with coal, parti-
cularly in the industrial	of the
Northeast. Several states also had re-
strictions on the reuse or ash and second-
ary materials as an alternative to dis-
posal. Uncertainty over whether existing
environmental regulations vould continue
or be revived was found to produce a wait
and see attitude and consequently delayed
recycling decisions.
TABLE 3. ESTIMATED UNIT OPERATING COSTS
FOR SHALL AND LARGE RECYCLING FACILITIES
Breakeven-tipping fee*
Facility Type**	Small design***	Large dealer***
Modular Incineration	$ 8.83	$ 3.91
Mass burning vatervall Incineration	$ 5.79	$ 3.06
Seal-suspension vatervall Incineration	$ 4.68	($ 0.25)
with ferrous recovery
Refuse derived fuels with ferrous, aluminum	$16.43	$10.70
and glass recovery
Ferrous, aluslnum, and glass recovery	$11.25	$ 8.30
Source: U.S. EPA Contract No. 68-03-2761. Draft reoort. Appendix A, Table 5.
* In 1977 dollars per standard ton.
** Incinerator systems assume sale of steam.
*** Small design assumes 50 tone per day (tpd) for modular Incinerators, 500 tpd for all
others; large design assunes 100 tpd for modular Incinerators, 1500 tpd for all other.
243
-------
The federal tax code contains certain
provisions relating to thi extraction of
virgin materials. These provisions in
effect amount to subsidies which lower the
supply price for virgin materials. It was
estimated that the likely increases in
revenues fro® recycling a ton of municipal
solid waste would be small if the tax ad-
vantages were removed. These estimate*,
however, were based on short-run elastici-
ties of supply and demand; the long-run
investment patterns and resulting elastic-
ity night provide more fjvor.'iM* price
relationships for recycled jnaterials. i?'J)
Regulation of prices'for domesticallj
produced oil and natural gas can in cer-
tain circumstances result in retail prices
helow what they would otherwise be, and
lower than *orld equilibrium prices.
These artificially maintained low prices
can consequently result in lower decanJ
for energy from solid waste as appears to
be the situation in the United States to-
da.». Regulated utility rates w!,ich pro-
vide for "pass through" incremental fuel
costs aie generally assumed to weaken the
demand by utilities for enerpv from solid
waste; however, six of the eijsht facili-
ties were recovering energy from waste
and utilities were actively involved at
four sites. Franchise restrictions on
steam and electric sales did not appeaT to
adversely affect revenues for the facili-
ties. Perhaps because of the increasing
fuel prices, revenues froci sales of energy,
are lesL. affected by impediments than re-
claimed materials. Of the eight facili-
ties, five had undertaken uaterials re-
covery b\»t sales of reclaimed materials
were gcntrally small compared with revenues
from energy.
Aluminum recovery was limited to hand
picking of.trjlky items at the several
plants where it was undertaken; recovery of
aluminum cans was not generally practiced
because the technology for separation is
not well dewloped rather than lack of
markets. Lack of uniformity in product
standards caused abandonment of some
materials recovery in two instances wr.ile
institutional restrictions on the use of
reclaimed materials curtailed plans for
marketing certain products at three faci-
lities. '
The difficulties involved in siting
facilities ire often asserted to I* a
major problem in expanding recycling; the
study found this impediment existed in
one case. Siting considerations were
found to be important, however, in delin-
eating the market area for products rela-
tively costly to transport.
Other factors frequently described as
impediments to recycling, such as restric-
tions on interstate movement of waste and
restrictions on financing and contracting
by public authorities were found to have
generally minimal effects on recycling ir
the selected facilities. Uncertainty over
future recycling policies in state govern-
ments contributed to reduced capital in-
vestment at two pl.ints and a high sensiti-
vity to risk exhibited by public officials
in one community resulted in a decision
favoring private ownership and operation
of the facility.
Financial Experiences of Selected
Faciliti es:
Anaiysis of the financial performance
of Selected recycling facilities indicated
that none ot tile case study facilities
were successful from an ac:ounting or mar-
ket sens? for the fiscal year 19?ft (Table
4). Net losses ranged from SO.60 to over
Sab.00 per ton in fiscal year 19"S. Kith
two exceptions, the projections were that
financial performances would improve during
fiscal year 1979.
The comparison of financial statements
encountered considerable difficulty due to
differences in accounting systems, financ-
ing techniques, and peculiar circumstances
at individual plants. Financial data were
adjusted for these differences and in sev-
eral instances adjustments were also made
because operations were temporarily sus-
pended for mechanical difficulties and
failure to meet air pollution regulations.
About half of the plants kept their records
on the basis of tons processed; these were
converted to tonnage received to maintain
comparability among plants.
Performance Based on Net Disposal Cost:
The financial performance for each
facility was computed as the total opera-
ting cost less revenues from sale of re-
claimed materials and energy. This was
defined as the net disposal cost incurred
by t.ie recycling plant. The computatio;.
permitted comparisons of financial perform-
ances on a per ton basis and also compari-
244
-------
TABLE 4. FINANCIAL PERFORMANCES OF SELECTED FACILITIES
IN FISCAL YEARS 1978 AND 1979

Total
Cost+
Profit(Loss)
Facility*
1978
1979
1978
1979+
Ames (RDF)
S30.S9
$24.83
($17.49)
($10.60)
Bralntree (WI)
46.91
22.24
( 36.17)
(10.89)
Karrlsburg (VI)
18.02
21.17
( 5.90)
( 2.60)
Milwaukee (RDF)
24.31
26.70
( 9.<>7)
( 9.1tl)
Monmouth (S)
15.85
11.98
< 11.94)
( 7.9b)
Mountain View (LG)
8.26
8.60
( 1.04)
0.76
New Castle (S)
7.77
9.81
( 0.60)
( 1.94)
N. Little Rock (MI)
19.27
20.61
{ 6.10)
( 8.17)
Source: U.S. EPA Contract No. 68-03-2761.
*Legend: WT^waterwall incinerators, MI-modular incinerator, S"shre«
-------
TABLE 5. RECYCLING NET DISPOSAL COST
BASED ON ORIGINAL AND CAPITAL REPLACEMENT
COMPARED WITH LANDFILL DISPOSAL*
Net Disposal Cost*
Vacllitv
Original-
Replacement
Landf 111
1. Ames, Iowa
$11.40
$20.50
$ 8.00 - $12.00
2. Braintree, Massachusetts
19.00
28.33
17.00 - 20.00
3. Harrlsburg, Pennsylvania
14.03
20.42
4.50 - 7.50
4. Milwaukee, Wisconsin
22.78
25.27
8.00 - 12.00
5. Monmouth Co., New Jersey
11.95
12.56
2.50 - 7.50
6. Mountain View, California
5.88
6.45
6.84
7. New Castle Co., Delaware
9.81
10.92
N.A.
8. No. Little Rock, Arkansas
11.46
14.02
3.33
Source: II. S. EPA Contract No. 68-03-2761.
~Projected for fiscal year 1979 based on partial year's data.
^Dollars per standard ton computed as total costs less revenues from resource recovery
and excluding transportation costs to the facility.
Excessive amounts of impurities and con-
taminants caused problems to users of the
recycled materials.
Another major impediment to success-
ful financial performance was t!>-* inaccu-
rate projections. Projections of the
quantities available for markets and
revenues to be received failed to develop
to expected levels in most cases. . These .
projections were particularly important for
energy which accounted for the major por-
tion of recycling revenues.
ftie to faulty projections and over-
optimism, excess capacity was ohserved in
over half of the facilities studied.
Underestimation of capital and annual op-
erating costs were additional causes of
poor performance; these errors caused high-
er recycling costs. Expenditures for cap-
ital structures and equipment were 40 per-
cent higher than original engineering csti-
246
financial records were examined. The rel-
ative importance of this and other imped-
iments in terms of frequency of occurrence
is 'hown in Table 6. As indicated earlier,
facilities were reluctant to coKect tipp-
ing fees equal to their net operating costs
for fear that solid waste would be diverted
to landfills. Two facilities had an ad-
vantage in not having to compete directly
with landfills; one facility had received
prior commitments for the wastes from the
communities which it served and in another
instance the county had closed the nearby
landfills.
Technological difficulties coupled
with limited experience in recy:ling solid
waste resulted in excessive costs and thus
also- impeded successful financial perform-
ance. The technological difficulties in-
cluded corrosion and abrasion problems,
explosions, difficulties with the aluminum
separation, and over-sired or bulky wastes.
-------
TABLE 6 RELATIVE SIGNIFICANCE OF SELECTED IMPEDIMENTS
TO RECYCLING FACILITIES
Impediment	Occurrence
Inability to compete with landfill disposal fees	xxxxxxx
Technological difficulties and limited experience	xxxxxxx
Unrealized projections of revenue frotv product sales	xxxxxxx
Government Intervention In market pricing and resource
allocations	xxxxxxx
Overtleslgn of plant capacity	xxxxx
Underestimation of initial capital cost9	xxxx
Environmental regulations	xxxx
Fragmentation/conflicts of governmental authorities	xxxx
Underestimation of 0 & M costs	xxx
Overesclmation of materials recovery	iocx
Overestlmation of available waste quantities
Pricing energy below market replacement value
Public opposition .to siting
Local politics and labor-aanagement problems
Source-s--?rated waste programs
mates in one ca?*: in another instance they
were 60 percent higher. The increased cap-
ital costs added as much as $5.00 per ton
to a facility's net disposal cost. At one
facility, start-up and minor equipment
costs were nearly five times above the
initial estimate.
Environmental regulations affected re-
cycling facilities in several ways. In the
Northeast a major shift from coal to oil
and natural gas during the 1970's in re-
sponse to air quality regulations greatly
limited the availability of boilers for
utilizing refuse derived fuel (RDF). ' En-
vironmental considerations also restricted
Teuse of residue from energy and materials
recovery. Decentralization of solid waste
B&nagement was an impediment in obtaining a
sufficient supply of waste in about half of
the cases. Local jurisdictions are fre-
quently reluctant to relinquish their au-
xx
XX
XX
XX
X
thority in order to establish regional
solid waste management. In one case,
however, federal and state activities Crea-
ted a climate of uncertainty which later
resulted in government competing directly
with an existing recycling facility.
Failure to properly price energy ob-
tained from recycling reduced the revenues
for several facilities. Existing contracts
had inadequate provisions for adjusting
steam prices as energy costs rose. Lack of
information about the cross-elasticity of
demand for diflfMi' through .ecycl ing vis-
a-vis landfilling may have played a role in
incorrect pricing of the recycling disposal
service; in one instance, however, a facil-
ity had a monopoly position for the dispo-
sal service but yet failed to impose a tip-
ping fee equal to its incremental operating
cost.
247
-------
Public opposition to siting impeded
recycling in at least one case. Local pol-
ices anil labor-management disputes result-
ing in downtimes and delays caused addi-
tional costs at several facilities. Source
separation programs, although active in
about half of the locations involved, were
not extensive and had no significant effect
on financial performance. Mandatory bev-
erage container deposits programs also had
little effect except at one facility where
it may have contributed to the discontin-
uance of the r.lumin-ira separator. Restrict-
ions on interstate movement of solid waste
did not appear to adversely impact on re-
cycling facilities, in several case.-, these
restrictions actually helped reduce compe-
tition from nearby out-of-state la.n'fills.
Seasonal variations in waste quantities
generally had little effect on financial
performances of the facilities.
limitations of Results:
The FPA-Mathtech study on financial
performances of actual recycling facili-
ties represented a pioneering effort in
recycling .inalysis. Past studies on re-
cycling were mostly concerned with techno-
logical performance, comparative economic
evaluations were usually based on engi-
neering cost estimates and hypothetical
operations. While the small saraple sire
and choice of facilities may have unin-
tentionally biased the results, the sample
was limited by the small number of oper-
ational plants in the United States and
cooperation of owners or managers. A
number of plants had not yet reached full
scale operations at the time of the survey.
Limited research funds also dictated the
sample size. The facilities selected for
indepth study thus represented the broad-
est cross-section of technological pro-
cesses, scale of operat-ons, ownership
arrangements, marketable products, and
geographical locations obtainable under
the circumstances.
As in most studies of an economic
nature, data gaps necessitated interpola-
tion and extrapolations. Financial data
for the various facilities were standard-
ized to a common ' 'counting base, and ad-
justments made where necessary. Uncer-
tainties exist with recpect to the true
values for the universe. And the nature
and direction of biases was generally rec-
ognized but not quantified in the study,
nor was the simultaneous effects of the
various impediments fully evaluated.
The comparison of net disposal costs
assumed that recycling facilities compet-
ed primarily i;t the waste disposal market:
consequently, recycling was viewed as an
alternative disposal mode. Recycling can
also be viewed as competing with virgin
materials in the products market. From
the perspective of influencing prices,
however, for the forseeable future recy-
cling is likely to have greater potential
for in the waste disposal market and will
probably continue to be a price taker in
the material products market.
Recycling might be viewed not as a
separate activity, but as one component of
a total systems approach to solid waste
management. In this context Sullivan
suggests for efficiency reasons the pri-.
vate sector should be given the responsi-
bility for recycling as private firms are
better able to evaluate its feasibility
in conjunction vith the collection service,
ui> Ultimate disposal, it is suggested,
is more likely to produce externalities
and therefore should be the responsibility
of the public sector.
It should bt mentioned that the cur-
rent emphasis on waste-to-energy recycling
probably constitutes * short-run perspec-
tive of conservation. Energy '.-an be pro-
vided from renewable resource , while the
non-combustible naterials in solid waste
represent non-renewable resources. While
no credit was give.T to disposal costs for
materials conserved, a lone-run oersncctive
of recycling suggests that perhaps greater
emphasis should be given to the value of
conservation nonrenewable resources.
Recycling Municipal Solid h'aste
What's Ahead?
In spite of the early optimism for
recycling in the United States, the record
todatc has been considerably below general
expectations. The results obtained from
the recent EPA-Mathtcch study indicate
that recycling of municipal solid waste
todate has not been the least cost means
of disposal. Lo.-cost competition from
landfills has been a major obstacle to
successful recycling operations. Land-
filling costs, however, may increase sharp-
ly if the regulations proposed under Sub-
title D of RCRA become effective. Land-
fill disposal costs presently average
slightly more than $1.00 per ton. (22) An
increase of 88 percent in costs as project-
248
-------
ed by Hart Associates would provide a con-
siderable impetus for recycling, particu-
larly in areas where land use is intensive
and considerable investment would be need-
ed to comply with the new regulations.(23)
The U. S. Congress recently passed the
Energy Security Act Which provides loan
guarantees and price supports for waste to
energy systems . Critics of the legislation
note, liovcver, that the principal problem
in energy fron waste is insufficient
demand rather than the supply siuc of the
market.
It remains to be seen whether the
future economic climate for recycling will
improve and whether new facilities now
under construction or beginning operations
will benefit from experiences of the
earlier plants. Local community officials
and waste manners arc interested primar-
ily in disposing of solid waste in the
least-cost manner. Recycling need not be
economical to the extent that revenues
exceed costs, however, net disposal costs
incurred through recycling need be compe-
titive with any environmentally r.cceptable
disposal alternatives unless addi:ional
credit and subsidies are included for
energy and materials conservation. In
testimony for a Senate subcommittee
recently, Abert predicted that only about
one-third of the available waste in the
U.S. would 'je recycled by the turn of the
century. (24)
REFERENCES:
1.	Nixon, Richard M.. The President's
Message in Environmental Quality:
First Annual Report of the Council on
Environmental Quality. Washington,
D.C.. August, 1970. p.xiii.
2.	U.S. Congress. Public Law 91-512
3.	U.S. Congress. Public Law 94-580.
4.	U.S. Congress. PubUc Law 96-294.
5.	Recycling and resource recovery are used
interchangeably in this paper, although
some would include the total process
of waste recovery and reuse of consumer
goods in recycling.
6.	Cincinnati Enquirer. December 12, J971.
7.	Carlsen, Eric. "The Economics of
Recycling." Environmental Affairs.
Spring 1973, pp.653-666.
8.	National Center for Resource Recovery.
Builetin: Octoher. 1971, p. 13.
' 9. "The Horn Of Plenty Keeps Overflowing."
Phoenix Quarterly. Fall 1977, p.ji.
10.	Porter, Sylvia. "Recycling In A
Tailspin?" New York Post. Mjrch 12,
1976.
11.	Solid Wastes Management/Refuse Removal
Journal. .July 197b, p. -14.
12.	Holloway, Robert. "Despite higher
costs than expected, this Iowa City's
refuscrto-encrgy project is still
going strong after 18 months of
operation." The .American City anJ
County. May 1978, p. 57.
13.	"I..I. Town a Pioneer in Recycling Its
Garba;, . Into Electric Energy."
Hew York Times. .January 14, 1979.
14.	Pollock, Eugene L.. "Resource Recovery:
Fact or Fiction?" National Solid
Wastes Management Asscciation Reports.
July 1979, p.4 ff.
15.	Petrone, Rocco A.. NCRR Bulletin.
March 1979, p.3.
16.	Easterbrook, Gregg. "The Great Garbage
Fiacco." Chicago. December 1979,
pp. 164 ff.
17.	U. S. Congressional Research Service,
Library of Congress. The Status Of
Resource Recovery: A Report Of Site
Visits. Washingtor, D.C. U. S.
Government Printi..,; Office. 1978
pp. CRS-27 ff.
18.	U.S. Environmental Protection Agency
Contract no. 6S-03-2761.
19.	Savas, E.S. . Lexington. 'las:achusetts:
D. C. Heath and Company, 1977.
Chapter 4. The Organization and
Efficiency of Solid Waste Collection.
20.	Andersen, Robert C. and Richard P.
Spiegel. Impact of the Federal Tax
Code on Resource Recovery. EPA 600/
5-76-009. December 137fc.
249
-------
21.	Sullivan, Arthur M.. Govcrr.mcnt
Interaction in Municipal Solid Waste
Markets. '.'npublished manuscript (n.d.)
22.	U. S. Environmental Protection Agency
Contract No. 68-03-286S. (Research in
progress).
2.V Fred C. Hart Associates. Draft
environmental Impact Statement On
The Proposed Guidelines for the
Landfill Disposal of Solid Waste.
U. S. Environmental Protection Agency,
Office of Solid Kaste. March, 1979,
p. 134.
24. Abcrt, Janes G. . Testimony before
the Subcommittee on Energy Conserva-
tion and Supply of the Committee on
Energy and Natural Resources...
February 27, 1980.
250
-------
OPTIONS FOR RESOURCE RECOVERY AND DISPOSAL OF SCRAP TIRES:.
A REVIEW OF TECHNOLOGIES AND ECONOMICS
llaynes C. Goddard
Industrial Environmental Research Laboratory
United States Environmental Protection Agency
Cincinnati, Ohio -I526S
ABSTRACT
This paper summarizes the results of a stud)1 of the scrap tire problem sponsored by
the United States linvifonmental Protection Agency. An overview of the problem is presen-
ed, and the technology and economics of retreading, rubberized asphalt and energy recovery
a're briefly discussed.
INTRODUCTION
Each year approximately 200 million
automobile and 40 million truck tires are
removed from service, a total of some 4
million tons, most of which find landfills
as final resting places. Although various
new resource recovery, options (mainly ener-
gy recovery! for the use of these tires
have appeared at least superficially at-
tractive, the economics of these processes
have permitted only a very slow growth- in
the utilization of them. Traditional in-
dustries for the reuse of these scrapped
tires, rubber reclaiming, tire splitting
and retreading have all experienced zero
to negative growth in recent years, with
the result that an increasing proportion
of these tires are placed in landfills.
There are a number of impacts associ-
ated with this situation that tend to make
uie costs of managing waste tires higher
than the mininum necessary. Basically,
there are two such impacts:
I. Resource Impacts. Tires are nearly
wholely produced from petroleum deriva-
tives. Since a large percentage of this
petroleum is imported, there is an obvious
negative impact on the nation's balance t f
payments, especially in a world of cartcl-
ized oil prices. Further, these tires
displace volume in landfills that in many
communities is increasingly scarce in
supply, raising disposal cost. This is a
problem primarily in densely populated
areas, where the distribution of scrap
tires closely parallels the population den-
sity. Reducing the size of the tires by
splitting or shredding in order to save
landfi 11; vol iime is costly. Charges imposed
to cover the cost of shredding creates an
incentive for tires to be illegally dumped
or littered, causing negative esthetic im-
pacts.
2. Health and Safety Impacts. These im-
pacts are principally the greater risk of
fire and disease from stockpiled and/or
littered tires. Tires have been implicated
in mosquito borne encephalitis cases in at
least one community.
The causes of the scrap tire problem
are to be found in the interplay of tech-
nological and economic factors. On the
technical side, such factors as lack of
simple processes to reclaim high quality
rubber and/or constituent materials such
as carbon black from used tires means that
such products, if obtained, are costly and
not competitive with virgin materials, al-.
though the rise in virgin feedstock prices
should begin to reduce the cost differen-
tial. Technical difficulties that American
companies encountered with th~ construction
of steel-belted radial tires have made them
virtually unretreadable, probably leading
to .1 higher rate tire scrappage than would
have otherwise occurred.
251
-------
The fact that a tire problem exists
indicates that the market for tires is not
functioning properly. This means that the
tire market, as an institution, is impro-
perly guided by the various resource prices
that affect its operation - in particular,
tires are improperly priced, since the
costs they impose on individuals and com-
munities are not reflected in the prices
consumers pay for them. Corrections of
these institutional shortcomings would do
much io solve this problem.
One of the manifestations of this im-
perfection can be seen in the proliferation
of tire sizes, which, apart from demand
considerations, leads to a induction in the
percentage of tires that arc rerrcadcxl.
This percentage is reduced because retread-
ers do not find it profitable to maintain
a variety of molds to match all the varia-
tions in tire size that tire manufacturers
offer. This proliferation in sizes is
caused by two factors: (1) the profit mo-
tive leads tire companies to exploit the
various dimensions of demand in the tire
market, 3nd (2) the auto companies have
been integrating tire design into the de-
sign of automotive suspension systems. The
latter source rill probably dimmish in
importance as the auto manufacturers reach
a new equilibrium with respect to smaller
auto sizes, but this will have little to
no impact on the incentive that tire com-
panies have to increase the murber of tire
sizes.
TECHNICAL OPTION'S FOR TIRE REUSE
At the moment, there are four techno-
logies ir various stages of current use and
development that lead to reuse of scrap
tires in some form, and may lead to greater
future use. These technologies are re-
treading, shredding, rubberi:ed asphalt and
energy recovery. Of these, we discuss only
retreading, rubberized asphalt and energy
recovery. Shredding is a fairly well de-
veloped technology, and is an intermediate
processing'step for disposal, rubberized
asphalt and energy recovery.
Retreading:
Currently, approximately 13 million
truck-tires and SI million auto tires are
retreaded annually; the figure for passen-
ger auto tires has been declining through-
out the decade of the seventies, and stood,
for example at J6 million in 1974 . The
state of the art is ;uch that well-made
retreaded tires arc produced with relative
ease, but that the variable performance
that is observed is due mainly to the many
small retread shops that do not have the
proper equipment for good retreading and/or
seek to avoid the cost of producing high
quality retreads.
Also, it is true that the shift to
radials is having a significant negative
impact on the rate of retreading. In 197S,
for example, the retread ratp for bias-
belted tires produced in 1976 was 38®; for
radials, it was 6°o. Unless American tire
manufacturers can solve the technical pro-
blems of radial tire construction that
havmade their radials unretreadable, and
unless retrcadcrs get the equipment in
place to retread these radials, there will
be a significant further decline in retread
rates, leading to a-substantial increase
in the number of tires discharged for dis-
posal .
Rubberized Asphalt:
The concept of adding rubber to as-
phalt is an old one, although the use of
scrap tires as the source of the rubber
component in the mixture has received seri-
ous attention in only the last 10-15 years.
There are two very similar processes in
use, both developed in Phoenix, Arizona.
Rssentiallv, the process consists of add-
ing crumb rubber derived from a rubber
reclaiming process to hot asphalt in an
asphalt distributor truck and spraying it
on the road surface in the usual fashion,
covering it with stones ("chips"). The
anticipated benefits in road applications
(potentially the largest use for the mate-
rial) are two: prevention or retardation,
of the rate at which cracks reflect
through new asphalt courses that have been
overlaid on older and failing pavements,
and as a waterproof membrane, for use on
bridge decks, for example. The most ef-
fective use may be in prevention of pave-
ment failure resultant from* expansive
soils, such as clays, which stimulate
"al1igatDring", so named because of the
dense and interconnected nature of the
cracks, "ihe effectiveness of rubberized
intcrlayers in tvo other major types of
pavement failure, lateral and transversal
cracking, is more problematical. Lateral
cracking stems from weather caused expan-
sion and shrinking of concrete pavements,
and transversal cracking from pavement
252
-------
and/or base failure, often caused or ex-
acerbated by excessive weights in tehicles.
The technical potential for rubberized
asphalt to prevent anil/or retard each of
these sources of reflective cracking is
currently under study in a multi-year set
ol" test projects being conducted by the
Federal Highway Administration with sup-
port from EPA. The analysis of the tech-
nical and economic data will begin next
fiscal year, and hopefully will provide
mere precise'answers to the technical and
ecor.cnsic questions that .ive been posed
about the use of this material.
For example, rubberized seal costs are
about 70'a (S.45/vd') more costly than the
conventional non-rubberized treatments.
Table 1 shows the discounted payback
periods implied by various conbinalions of
discount rates :>nd cost savings on a square
yard basis. As yet, there does not exist
good information on what the savinps in
fact are. In the C3se of the stress-
relievins interlavcr (placed between The
old pavement surface and a new, thicker
overlayer or finish course), some prelim-
inary evidence from one application of the
material in Arizona indicates .in annual
TOintenance cost savings of 5.26/yd-,
suggesting a discounted payback of four to
five years at rates of discount of six to
ten percent.
Another important use of rubberised
asphalt is in crack and joint seal com-
pounds, The preliminary evidence suggests
that the rubberized scaler is technically
superior to conventional sealers, which
oftev, fail in less than u year, and the
cost premium is only IP". Thus, thr like-
lihood of the cost effectiveness of the
material is high, although ag.iin, the
answer to this question of cost effective-
ness is not known with precision,
Energy Recovery
Essentially, two basic technologies
are being considered by a number of firms
as methods of recovering the energy con-
tent of tires: direct conbustion and
pyrolysis, both of which normally require
shredded tires as feedstock. Neither of
these technologies in the various forns in
which thry appear arc nuch bcvorid experi-
mental or pilot stages, and many cases
have not even readied the pilot stage.
The economics of the processes for the
most part present an obstacle vet to be
surmounted.
Oirrct combustion techniques have
taken tires, whole or shredded, and burned
there either singly nr mixed with other
fuels, especially coal, typically for
steaa production. There does not exist
any comprehensive information on the air
pollution impacts of burning tires, but
past experiences with the process suggest
that proper feed rates and standard emis-
sions control equipment will be able to
deal with tire related residuals.
TABLE 1. DISCOUNTED PAYBACK PERIODS FOR AN ASPHALT-RUBBER SEAL COAT
j	(years} 	
Annual Maintenance
Savings {Cents
Discount	Pct Square
Rate	Yard

.10
.15
.20
.25
.30
.40
64
6
4
3
2
2
2
10
7
4
3
2
2
2
IS
8 •
5
3
3
2
2
20
IS
S
4
3
2
2
25
50
7
4
3
3
2
2»
-------
Large scale combustion of tires be-
comes dependent on an adequate supply of
tires for the process - as such it is
quite sensitive to the cost of collection,
especially transportation costs, and to
prices to be paid (positive or negative)
for delivery of tires to a facility. Pro-
cessing costs do not seeo to be an impor-
tant consideration, at least in terms of
plants of 30 tons per day or mere. A
study of a hypothetical facility in New
England suggests that the process would be
profitable if as few as six percent of
tires in Sew England were collected and
delivered to the plant. As of today,
there are still no important facilities
for shredding or otherwise processing
tires for energy recovery. The cost of
collection and the insufficiently high
prices of alternative fuels seem both to
serve to make energy recovery from tires
uneconomical at present.
Pyrolysis of whole or shredded tires
is a process tfiat has attracted and still
attracts the attention of many chemical
engineers, .iuch companies as Firestone,
Goodyear, Tosco and others have made sub-
stantial investments in the past seeking
to recover fuel oil, carbon black and
gases from pyroliied tires. Typically the
quality of the carbon black, a major in-
gredient in tires, is of insufficient qual-
ity to make it competitive with virgin
blacks without further processing. Again,
economics proves to be the hurdle that
remains to be surmounted before energy
recovery from tires becomes profitable.
Wiy the-rise in relative energy
prices has not led to a more favorable
economic environment for energy recovery
from tires is not really understood. Fed-
eral prices control and entitlements pro-
grams surely have something to do with
this problem, and it may also be that the
process is itself sufficiently energy in-
tensive in the collection phase to leave a
margin between revenues and costs too
small to be attractive, except at very
high prices for -i. tentative fuels.
SUMMARY AM) CONCLUSIONS
At present, the only technologies in
use for reusing tires in any significant
fashion are.retreading, rubber reclaiming,
tire splitting and rubberized asphalt. To
¦an unknown, but'probably small extent,
there is direct combustion occurring in
various plants and shops around the country.
None of the trends suggest that there
will be any significant increase in scrap
tire recovery in the near future - indeed
the trends are distinctly downward, save
the small and slowly increasing use of
tires in rubberized asphalt application.
Since collection costs are relatively high,
this aspect of the used tire cycle will
likely remain relatively unorganized and
possibly inefficient until l;irge scale
tire recycling be;ofne= more profitable.
A significant risf in the relative price
of petrolerm is a necessary condition for
creating a strong demand which may in turn
stimulate the development of a more effi-
cient collection system. In the absence
of this strong demand, the disposal and
littering problems that exist may require
some external mechanism such as product
charges to ensure that tires are disposed
of in an environmentally acceptable manner.
Such a charge on the order of two cents
per pound of tire would probably be suffi-
cient to finance a nationwide system to
stimulate tire collection and appropriate
disposal. Little is known in detail,
however, about how such a mechanism might
work.
REFERENCES
Deese, Patricia. w. al., J>80. Options
for Resource Recovery and Disposal of
Scrap Tires, U.S. Environmental
Protection Agency, Cincinnati, Ohio
144 pp. Available from National
Technical Information Service.
254
-------
KCRA STUDY OF GLASS WD PLASTIC RESOURCE RECOVERY
Tom Archer
Pacific Environmental Services
1930 14th Street
Santa Monica, California 90404
Jon Huls
Secondary Resources Development Consultants
115 So. Patrick St. "304
Alexandria, Virginia 22314
ABSTRACT
The research program was initiated as an objective of the Resource Conservation and
Recovery Act. It had the overall objective of assessing and evaluating State-of-the-Art
for recovery of glass and plastic resources from solid wastes. Currently, labor-
intensive source separation of glass and plastics predominate, although mechanical and
themal recovery will achieve greater importance in the years anead.
Literature was gathered from numerous sources, contacts were made with industrial and
recycling organizations, and questionnaires were distributed among applicable firms
involved in glass and plastic recovery. Data derived from literature was collected,
reduced and evaluated for technical, economic, and environmental content.
This report was submitted in fulfillment of	Contract Ho. 68-03-2703 by Pacific Environ-
mental Services, Inc. under the sponsorship	of the .U.S. Environmental Protection Agency,
Stephen James, Project Officer. The report	covered the period May 1978 to January 19S0,
and work was completed as of July 1, 1980.
INTRODUCTION
The objective of this RCRA mandated
study was to define the state-of-the-art
for recovery of plastic and glass resources
from waste as determined from available
literature. Resource recovery technolo-
gies, both mechanical and labor intensive,
were assessed for municipal and industrial
waste source?. Where data was available,
these technologies were discussed in terms
of technical, economic, environmental, and
social aspects. Current trends in plastic
and glass waste recovery practices outside
the United States were assessed. Fast and
present research efforts were identified,
and research needs to enhance recovery of
resources were addressed. Study findings
are discussed following.
MANUFACTURING AND INDUSTRIAL BACKGROUND FOR
PLASTICS AND GLASS
Plastics Manufacturing and Plastics
Industry
Plastics is a generic term describino
strong, durable, light, easy to fabricate,
fairly inexpensive materials derived from
petrochemical feedstock. Plastics are
available in ever 40 "families" or material
types with a broad range of performance
characteristics (1). Plastics are a rapid-
ly increasing ssgment of the economy, and
new and variable uses and markets make
industry characterization difficult.
All plastics are either themosettino
or thermoplastic. Thermosetting plastics
255
-------
are set into permanent shape by the appli-
cation of heat and pressure and on reheat-
ing, cannot be reshaped. Thermosets
account for over 20 percent of the total
U.S. polyner production ar.d are often used
for durable goods such as counter tops,
pot handles, knobs, highly engineered
applications, and do not significantly add
to the municipal solid waste stream (1).
Thermoplastics soften upon reheating
and harden, upon cooling. Ease of use of
thermoplastics, plus specific resin char-
acteristics enhance their use. Thermo-
plastics are often found in the municipal
solid waste stream, (1) and they account,
for approximately 80 percent of polymer
production (2).
Plastics manufacturing is a diversi-
fied and complex operation.' From ihe raw
material input to the final consumer prod-
uct, the various operations within the
plastics industry are integrated into
various segments. Figure 1 shows the in-
terrelationship among the various opera-
tions involved in the manufacture of
plastics (3). Integration of operations
within the plastics industry is extensive;
thus, one company can be a resin producer,
compounder, and fabricator; and a nanu-
facturer/packager can sometimes operate
as fabricator and converter. As a plastic
product is nade, starting from the resin,
it normally passes through manufacturing
facilities that progressively become
smaller in size, and more dispersed geo-
graphically. Ihe wholesaler/retailer and
consumer segments are dispersed according
to population density and end use markets.
Glass Manufacturing and Glass Industry
Glass has the following characteris-
tics: chemically inert, impermeable \o all
liquids and gases, sanitary and odorless,
can be made transparent, and versatile and
adaptable in that it can be molded to al-
most any shape and size (4). The manu-
facturing process is usually a fully in-
tegrated one step process which begins with
raw material feedstock and a finished
product at the same location. Basic raw
materials include soda ash, limestone, and
sand. Limestone and sand are cheap and
abundant. Cullet, or waste glass, can be
used in lieu of soda ash, which is in
demand.
As of 1980, there were 12T- primary
glass producing companies which altogether
operated 240 individual plants (5). These
glass manufacturing facilities are located
throughout the United States and are
usually situated near the markets they
serve. Plants are found in 34 states with
the majority located in the followinq 10
states: California, Illinois, Indiana,
New Jersey, New York, Ohio, Oklahoma,
Pennsylvania, Texas and West Vii-qinia.
STATE OF THE ART FOR PLASTICS RESOURCE
RECOVERY
Plastic Waste Generation
Plastic waste is generated from indus-
trial-manufacturer, conmercial and munici-
pal sources. The amount of plastic wastes
generated in 1977 and projected for the
years 1980-1990, is presented in Table 1
ccmsum a
Figure 1. Interrelationships among various operation;
in the manufacture of plastics.
256
-------
and was scaled from 1974 data. They pre-
dict a steady increase of plastics. This
would occur principally due to expanded
uses of plastics as a substitute for other
Items. Substitution irises from transpor-
tation costs accrued during distribution
and the need to conserve materials,
wastes as received at landfills, the break-
down by source is 84 percent, for house-
holds, 10 percent for conmercial/institu-
tional and 6 percent for industrial. As a
fraction of the municipal waste stream,
plastics represent a small portion of
approximately 4 to 5 percent (6). Plastics
as a component of municipal refuse has
increased about 50 percent in the last
decade.
in the municipal waste stream are normally
plastics packaging. No hard data exist to
indicate exact quantites of plastics
recovered frotn waste streams. Estimates
indicated that of the 7,500 Gg (16,500
million lbs) generated annually from all
sources, about 2,200 Gg (4,850 million
lbs) were recovered, primarily through
industrial recycling (1). Solid wastes are
produced at essentially every step in the
manufacture of plastics, with the post-
consumer segment accounting for the
majority of wastes.
TABLE 1. ESTIMATES AND FORECASTS OF PLASTICS WASTES
GENERATED AND RECOVERED
Utegory
Quantity by year (Billion torn tnc tg)*,0
TTT7
*i?Tj
"W
W

WT
1
MT
sq
IP

~WT' ~
tq
lotai solid Mite
MO
127
160
145
180
163
2Jo
161
Ibnicipal generation
6.9
6.3
8.4
7.6
11.2
10.1
13.4
12.1
Cosaer^ial generation
O.B
IU
0.9
0.3
1.2
1.1
1.4
1.3
Industrial generation
O.i
0.5
0.7
0.6
1.0
.9
1.2
1.1
Recover-'
1.4
1.2
1.6
1.4
2.4a
2.7
2.b"
2.S
Total vattt it generated
6.9
6.3
8.4
J.t
n.o
9.9
13.2
11.9
Percent plastic 1n
sited aaites .
4.9
4.9
S.J
6.3
6.2
6.2
6.6
6.0
Plastics recovery is a
S of plastic aastes
(•unicipal) for energy
recovery
0
0
4.2
4.2
11.4
13.4
24.3
2'..i
To'.al «astes as disposed
6.9
6.3
8.0
7.2
9.6
6.7
lb.b
9.0
*	- AsMM no variation In lnflustrial*»unicipel, comcctal ratios of ger;ration
*	- Composite of Mio«est Researcn institute a no PCS estmates.
*	- Recovery it composite of source t«p«r«tion «rw energy recovery.
*	- Incorpcr«t»j p£T recycling at 26 percent efficiency.
Plastics production in 1977 totaled
15,411 Gg* (33,948 million lbs) (1). Of
that amount, approximately SO percent were
thermoplastics, whicn are amenable to re-
melting and, thus, refabrication-, to a
certain extent. The largest single end-
use for plastics is in packaging, although
most plastics are utilized in long-term
uses. As a result, plastic wastes found
•	q
* Gg is the metric abbreviation for 10
grams.
Plastics Waste Resource Recovery
Due to the tremendous qrowth in the
use of polymers or plastics, especially in
short-term packaginq usaqe, increasinq
attention has been focused on its recovery.
However, the recovery of plastic, from
municipal refuse within the United States
is basically embryonic. Currently only
257
-------
specific plastics which are uncontaminated
and segregated from other polymers and
wastes have potential for recovery. PET
bottles, PVC scrap, polyethylene con-
tainers, and HDK film are currently spo-
radically recovered for recycling. As a
result, enet'gy derived from combjstion in
waste to energy plants most likely repre-
sents future prevalent elastics "recycling."
A less familiar but equally important
area is that of "pre-consumer" wastes,
those generated by producers, processors
and fabricators of products. While
recovery of plastics from municipal refuse
1*! not extensive, industrial, and to a
certain extent, coirmercial recovery is
quite extersive. Essentially, scrap recov-
ery has long ceased to be an afterthought
in most plastics processing operations.
Scrap handling has the potential of being
as important a plastics processing opera-
tion in its own right as processing virgin
polyn.ers, since the rising costs of feed-
stocks Hakes even small loses significant.
There are fewer and fewer operations that
cannot justify either regrinding equipment
or recovery of off-spec resir. for sale (7).
Post and present recovery programs
are listed in Table 2.
Reuse strategies have shown that
clean and single material plastic waste
streams derived from municipal waste (PET,
for example) can be collected and recycled.
However, this is limited and is useful
only for beverage packaging.
Except on such limited basis, plas-
tics materials recovery from the mixed
municipal waste stream appears to be tech-
nically or economically infeasible at
present. The greatest potential for suc-
cessful plastics waste recovery seems to
be (1) the derivation or recovery of ener-
gy from cortbustior. of a mixed plastics/
organics waste fraction in the municipal
waste stream, or to enhance volume reduc-
tion through various forms of thermal
treatment by utilizing the high energy
value of plastics, and (2) selected source
separation.
In the former, the presence of plas-
tics enhances combustion due to a hich BTU
consent. As waste contains a number of
nonconbustible items and significant
quantity of .noisture. plastics can be an
important offsetting combustible fraction.
Thermal Treatment
Thermal treatment can be grouped into
three general categories:
•	Large scale and modular incinera-
tion (with and without energy
recovery)
•	Pyrolysis
•	Preprocessing for refuse-derived
fuel
For each of these methods, prooonents
desire the hiqh enerqy content of plastics
to enhance the overall eneray content of
the solid waste. Plastics found in MSU
have heating values in excess of 42 kJ/g
(19,000 BTU/lb). Refuse heating values
range near 11 kj (5,000 BTU/lb). As a
comparative point, coal has. a typical
energy content of 28 U/q (12,000 ETll/lb).
An additional benefit of the themal
treatment systems is the potential for
volume reduction of solid waste by as
much as 90 percent.
Polyester-polyethylene Terephthalate (PET)
Eottles
Recovery and re^yclinq of post-con-
sumer (municipal level) PET is established
and growing. Prompted by bottle deposit
laws, anti-litter movements and the need
¦to conserve costly raw materials, re-
cycling of PET is regarded as fully
commercial. Both DuPont Co., Delaware and
Goodyear Tire and Rubber Co., Ohio, have
Started up pilot plants for recycling.
Industry estimates that PET recycled in
1979 amounted to approximately 3,499,090
kg (3,849 tons). This is estimated to
almost double in 1980 (8). Owens-Illinois,
a container corporation (.as published a
guide to PET recycling which gives urqent
attention to the recovery of PET in markets
which have container deposit systems. In
this guide, bottle fillers are recomended
to include a reclamation svstem. Most
bottling facilities can accomodate the
necessary equipment. The major component
is the granulator. It grinds all materials
into scrap particles. Accessories to the
granulator include variable feed hoppers,
conveyers, air evacuation systems, and
scrap shipping gaylords. A mifiimum of 400
cn::.ire feet is required for the processing
system, with additional space needed for
transfer and storage operations (9).
236
-------
TABLE 2
PAST AND PRESENT RECOVERV FACILITIES AND RECOVER* PROGRAMS
AT MANUFACTURING OPERATIONS



fr|p»
i.
(M«l (tHMI Utft'll*
1.
tr««4 ciittu «v vo*« rr»i«c«wil.
I.
0^ lit

UIMV »' 3 lltilHl #•* IliwttiOf.
j.
Df/Vkf torpofiiu*
).
• Mr 1( «nO W <0«» K»M,
¦»I|V H l«lf lvl0Vb>l« Mil
*,
trtM, IK.
1.
Iwill K'M to kMUllWK,
I.

I.
liflfM fpttlimi »•» •' IMf* l'«# >llll »l
K'W-
i.
Sw 4HMt*l
4,
ItfMl «•*> If Ml ii,
eou, »U.
t.

?.
Ikmm at)
I.

».
UltKli tomma
vtf'l IV '*(/(<«» M to (»t( 1
tin M '«-«»•* l« »Kt<|»n| Wtno^).
».
fe)4 FlttlKt Si*»tCti
t.
Mlt'n i« »<»*«¦
io. we en cmp**/
10.
|a<**por*t(v icw »n»
pT0C*f l t*a gt t'*k* t«fk.
1). «|far
11.
Vtft tflKllM tOlwttOA to *K»if HI llillU
Ur«i.
11.
«('tf PlHtKI
U. otti tt<«# t« piitiit ce«'«it
inp»»Hn.

I|. rft » t**»-

ficttgi*! iMiitnn
l«.
twuru <4to »£**» 10 mr*«l9 «i.
1). fell tip* t—*»/
H. lia*l k»C« frt ftlMf,
II. tK (JfffMllM
U. « Mlv«
If. C
I>.
C«**«rU *n»d bo1vw*Ikj«* la pu ijcIk.
It.
¦ IH-
It.
Iims »'#>».
It.
%)» »
»».
1*19 •'~'I'M! fCC/U
n.
(ttftl IPlWIH
JO.
icri m x^f


71.
Nt »«»«r
«.

it.
•olffdOyl***
Reclaimers are interested In high
volume usage and purchase truck load
quantities 22 to 24 gayloads which each
are 4*4 foot square fiberboard con-
tainers. An average purchase price for
PET is estimated at SD.03 per lb.
Markets for Recovered Plastics
Harkets for plastics recovered from
municipal refuse is limited. Most poly-
mers will not find markets for reuse. The
emeraing market is for PET. Specifica-
tions vary according to market.
PET
Recycled PET currently finds its
largest end use 1n strapping and fiber
fill; insulation forwinter clothing, car-
pet backing, thermoform sheeting for clear
packaging. As of February 1979, PET re-
cyclers were documented in literature (8)
and are listed.
Building Components, Inc.
Van Nuys, CA
St. Jude Polymer Corporation
Mahoney City, PA
E.I. DuPont de Nemours & Co.
Wilmington, DE
Midland Processing, Inc.
Pomona, NY .
Plastic Recyders
Richmond, CA
Plastic Recycling Incorporated
Dallas, Tji
239
-------
Plastics Development Corporation
Los Angles, CA
Three M - 3M
St. Paul, MN
Willman Industries, Inc.
Johnsonville, SC
Pure Tech Industries, Inc.
Pine&rook, NJ
Ralco Industries, Inc.
Cumberland, RI
STA'.E Of THE ART GLASS WASTE RECOVERY
Source Identification
Waste glass generation in the United
States stems from three primary sources:
industrial, commercial, and municipal.
Industrial waste glass f?r this analysis
is assumed to be any glass waste generated
during the manufacturing of glassware.
Commercial waste glass is assumed to be
any glass w.iste generated from sources
where glass is used as an integral part of
the establishment's product line. For ex-
ample, waste commercial class can emanate
from bottle tilling operations, the food ¦
packaging industry, the-construction in-
dustry, food and beverage service industry
(including bars), the automotive industry,
and about any establishment that uses glass
for their products. It is noted that
cotmiercial glass waste finds its way into
the municipal waste stream or directly
into landfills. Municipal glass waste is
assumed to be that glass which 1s discarded
after the useful life of the product has
ended. Examples include-beverage con-
tainers, food containers and windows, etc.
Industrial waste may be included in munici-
pal solid waste and can be disposed at
landfills. Retail outlets and bars may
contribute to these wastestreams.
Figure 2 presents an idealized drawing
summarizing the major components of the
industry, with sources of generation and
recycle material flows (10).
Quantities- of Glass Haste
The total glass production in 1978 was
estimated to be abojt 18 Tg (20 million
tons). About 70 percent of this glass was
container glass. However, the amount of
container glass found in municipal waste
is reported to be about 90 percent (11).
This is expected since the useful life for
container glass is relatively short when
compared with other glass types such as
flat glass and fiberglass, in the absence
of reuse systems. According to the latest
available statistics, glass is reported
to comprise up to 10 percert of the total
municipal wasteload (12) of 134 Gg
(148 x 10° ton). Table 3 presents waste
glass generation estimates based on avail-
able data and inference.
Estimates of future quantities of
glass waste in the municipal waste stream
are numerous. Future projections of any
sort are based on previous trends -nd many
factors such as marketing conditions and
competition. Table 4 presents a proiection
of glass wiste and the amounts recovereo
from mixed municipal waste for the period
1980 to 1990, incorporating such factors,
and beginning with the base year of 1972.
Review of Table 4 indicates that the
total quantity of olass waste generated
will increase slowly through 1980 while
the percentaae of glass in the total waste
load will, decrease. This is based on the
assumption'that glass waste recovery
through 1990 will increase due to advances
in recovery technology and wider comnunity
recycling efforts. This estimate could
change if reuse programs and container
deposit legislation efforts are success-
fully implemented on a wide scalp.
Glass Waste Recovery
The recovery of glass from municipal
waste within the United States today is
more representative of an emerging "tech-
nology, rather than an age-old oractice.
Nonetheless, a secondary materials industry
does exist, and methods for recovering
materials from municipal waste are
achieving new levels of sophistication and
success.
Within the recycling "closed system,"
three defined segments exist: (1) glass
manufacturing and secondary materials
users; (2) cullet dealers; and, (3)
municipal and private collection programs.
Glass manufacturers are the principal
actors. Raw material users have tradition-
ally utilized glass cullet derived from
off-spec olass, etc. *lost recycled glass
from Dost consumer sources has been used
by glass container manufacturers to produce
260
-------
r
i
i
*-*
r*
i
i
L
Cloil Conhi'rwn
lo* Mof

tato AiK
liwng*
Off*
om( Htoflf*
Cwlkt
ChcnitaN
Oirttl NFCjrtl iriy
*r«uatf -ttowrt /Other
u
fce* Motor..-1*
Production
]nr>owse
So »»d
f;k..

Wo A*
0«lv

O'S*


CwUti


:jt
	1
•—+
ftow Mo
I
I
~n
11• 1
' hesun'to
~!
I
	j
	»
F/cheitd G^i
Mirrori, IIcWdyOmeme#i»«d
tit.
CemuMtf
Co"w*»*ft'o1
IMuilrittl
Moiiih
I
LJ
l^4vilriel
MeiUH
tiqurp 2.
fi'ass InduS'.r-' recycle flnwci'.irt *nd <,ourro' of '.ulitl wa^t<
-------
TABLE 3. ESTIMATED WASTE GLASS GENERATION 8Y SOURCE (1977)
Total	Residential	Coiwcial Industrial
Category Haste ? RT Eg" ~i MT 5g ~X *TT Eg"
148,000	-- - 		 --
Glass waste 14,800	82 12,135	10,983 16 2368 2143	2 296 267
Glass container
waste 13,220	82 10,840	9,610 16 2115 -19'i4	2 264 239
Noncontainer
glass waste 1,580	82 1.295	1,172 16 252 228	2 31 28
a90 percent of glass waste
TABlE 4 . SUMMARY OF GLASS WASTE ESTIMATES, PROCESSING PhD RECOVERr
	FOR MUNICIPAL WASTE, 1972-1985 (1,000 ton)"	
Category	1972	1975	1980	1935 1990
Total col id waste	130,000	140,000	'60,000	1BO.OOO
Glass available	13,200	14,500	16,400	16,600 16,90j
Percent glass of
total waste	10.1	10.5	10.3	9.3
Glass processed for
recovery4	0	20	170	540 66u
Glass recovery
-	source separation
collection	175	180	225	225 225
-	cullet dealers	100	85	50	. 50	SO
-	waste recovery
plants	0	10	100	350 600
Total resource
recovery-glass	275	275	375	60C 85u
Percent recovery of
tota'-glass	2.8	1.8	2.3	3.6 5.0
"Processed in central facility with glass.subsystem
^Estimates by Midwest Research Institute
262
-------
new containers. Recently there has been a
shift to composites of glass, plastic and
fibers. These new secondary uses promise
gla',5 recycling an e>panded cullet capacity
with reduced specification levels. Addi-
tionally, economic problems exacerbated by
inflation and energy shortages have "in-
proved" the economics of smaller scale
enterprises. It has been theorized that
small scale, local industrie will fce more
apt to utilize locally-derived cullet.
thereby eliminating hign transfer costs
(13).
Cullet dealers represent a second seg-
rent. As intermediate processors, they
provicie the important function of aggrega-
tion and Quality control. Cullet dealers
are, however, a diminishing segment of the
industry. Less than 20 dealers exist
today (conversation with glass cullet
dealer).
finally, the delivery or collection
system, represented by grass roots re-
cyclers, municipalities, and small busin-
esses form the third regment. They often
deal through intermediate processors, al-
though larger pronrams niay se'l directly
to a manufacturer.
Industrial Source Glass'Haste Recovery
The glass siarufactu. ing industry has
traditionally recycled its in-house waste
materials (factory cullet) for various
economic advantages. Cullet added to the
glass furnace assists tJie melting process
of the virgin batch by lowering its nelt-
ing temperature, and speeding up its melt-
inn time in proportion to the percentage
change ("4). As a result of heat reduction
in the. fjrnace, the life of the refractory
furnace linings is extended and fuel con-
suir.ptior is reduced (H). Sorte manufac-
turers have also met air quality regula-
tions jy increasing cullet usage. Users
report a significant reduction pf particu-
lates and hign temperature related
emissions'(14).
In-house cullet is usually available
from off-specificatim glassware and re-
jected or broken glass fromwithin the
plant. Generally, 20 percent of the batch
material -will end up as "in-house" cullet
(lij. Socie additional Cullet (foreign
cullet) may be purchased from external
sources such as oottling plants, cullet
dealers, and municipal recovery programs
to increase the quantity of quality cullet
input to the furnace where Quality control
is stringently exercised (15). The max-
;nun percent of cullet that can be intro-
duced to a batch without altering the
quality of the finished oroduct is greatly
dependent on the quality of the input
cullet and how compares to th? batch
recipe. Several furnaces in Eurooe and
elsewhere are currently using nore thar
50 percent cullet in each batch, and some
furnaces have been.operated successfullv
on 103 percent cullet for short periods
under special circumstances (16).
Municipal Glass Waste Recvery
Municipal waste, in general, i; con-
sidered to be tnat discarded oost-consjner
material whicn is collectea and disposed
in the muricinal solid waste svster. !n
addition, industrial and conmeixial estab-
lishments discard wastes a^ona witn con-
sumer waste, so that municipal solid waste
can be a combination of a 1three types.
Glass waste in the municipal solid
waste stream, as previously noted, repre-
sents about 1C Dercent of the total waste
load. In eddition, aaout 92 oercent cf
the glass waste oriainated from container
glass.- As such, the container industry
has been interested in recovery of this,
material octh from an economic and ouDlic
relations perspective.
The glass recovery system actually
incorporates three basic ooerations: (1)
collection or delivery; (2) processing
and (3) recycling. All three, whicn to-
gether comprise the closed recycling sys-
tem, are interrelated and interdependent
upon each other.
Glass can be recovered fron -unicipal.
waste by source separation systers, n-echan-
ical recovery systems or by various reuse
strategies. Source separation is the
simplest and oldest method of glass waste
recovery, and requires separation of dis-
carded glass from other solid waste.
Source separation is- generally categorized
as a labor-intensive endeavor.
Mechanical separation involves the
application of mineral extraction and
separation techniques to nuniciDal solid
waste to extract glass from mixed refuse.
Mechanical systems for glass recovery are
at prtsent emerging or experimental, and
261
-------
are usually fcund as a subsysten and not in
an independent mode.
Tie recnanisr. cf recycling is very
sirilar to virgin material systens. Con-
ceitrat;pf\ ourification, and manipulation
O'" r.aterial characteristics occur. Once
materials are collected, they usually "lust
be aggregated, processed, and otnerwise
brougnt up to specification levels accept-
able to secondary r-aterial 5 users. Gullet
uealers aererally reuresent this oortion
cr' the sys'.en.
''e system ts closeH *he ourchase
cf collet by ranufflcturers, usually glass
container i^rfustries. ''ar'>.e: dynamics, as
v.i'1 be discussed later, acrear favrra-jie
*-p\ar,' use of glass cul'et in secondar.
rrcc-cts, wnere s?ecificat:;ns are less
st'-irceni.
fcjrce cerjrj-ion
5ru'-ce separation is r>aiic to the
-ran/ ecoloaically Totivated conmunity re-
cycling efforts throuohout fh.e nation.
Source separation actually b-Mns municipal,
industrial and comet-cial recycling, how-
ever. it is difficulty of mecnanically
extracting saleable material fron rixed
v.aste tnat nas sparked renewed interest in
source separation,
other public and private aycncfes and
industry (17).
In the following, discussion will
for.us on collection centers, cullnt
dealers and separate collection procrans.
'¦lost people are familiar with the
corrr-unity recycling center. Inert are
over 2,000 conriunity recycling centers
..•h'ch have proliferated across tne nation,
for reasons of community involvement,
small cedita1 exoense and on th? strenoth
of a good cere of participants (IS).
Additionally, rcost centers accept multi-
(r.aterials (all grades of pepor, netals .
and glass;, and as a result, are better
able to weather rarket fluctuations, an
all too cordon occurrence.
Internediate olass processors or
cullet dealers represent a kev link iri
the effort tc recover glass through separa-
tion. Industrial, conrercial and runicinal
programs are increasincly dependent on the
services that internediate processors pro-
vide. -'is a secondary materials cealer,
they do not produce new alass, but instead
act as t'ie purcnase agents for a number of
nlass manu'acturers.. There are fewer than
2? cullet dealers in the United States.
Thev arp located in "iew inland, rlorirt /J i a ~ ft	C oriiQ
as nixed recyclable purchasers for re-
source separation is a traditional
practice that accourts for nearly all of
tie al. ">s "-esource recovery currently con-
ducted. Sojrce separation involving
e-mer curpside collection or collection
centers "as grown ever the last decade.
Recycling technology is increasing in soph-
istication, ¦ level cf efficiency and work-
ability as rev. enuicrent, procedures, and
processes nave Seen exclusively developed
for separate collection and processing.
There has been an effort to standardize
Procedures for such programs that ensure a
¦"eliaDlp and ouality rrocuct frorr. these
CDc-atior.s. In part tnis success is
^	^	^	A ¦* r*< • b	K.,	I" ^
Environmental Protection .Agency (EPA) and
Reuse programs generally involve a
tax or deposit O'i ..aste containers, which
is redeerable upon return. The vast rajor
itv 3f reuse yrooraiKS disappeared as the
rarket economy favored one-way containers,
t-o.vevei', energy and envirorrental consid-
erations have spurred renewecs interest in
reuse strategies.
264
-------
cycling groups. In Los Angeles, Califor-
nia, one cullet dealer is the sole pur-
chaser of recyclable material, including
cullet, from the Downey DART source separa-
tion program. In that operation, collec-
tion vehicles transfer glass cullet and
other material collected directly to the
glass processor. The cullet dealer
separates metals andpaper from the scrap
glass, and.sells all components.
Most of the recent research and suc-
cessful glass recovery has cone from multi-
category source separation schemes in-
volving curbside collection. Curbside
collection programs generally operate in
residential areas. There are approx-
imately 220 such efforts on-line in the
nation (19). In c typical program, resi-
dents routinely set out for collection
using barrels, bags, ard boxes, recyclable
fraction(s) segreoated from ref,,<;e.
Either separate trucks or integrated
-collection vehicles collect the recyclables
and/or refuse. Material is normally taken
to a processing station where, i* glass
materials are already segregated in glass
colons or type, minimal processiny is con-
ducted. Where recyclables are mixed, hand
or mechanical sorting is required. Long
term storage may or miy not occur depending
o;i volume.
Source separation programs collecting
glass are presented as Table 5. Design
variables are indicated.
Mechanical Separation
High technology recovery systems are
emerging state-of-the-art for glass recov-
ery. While these s>stems do recover other
materials as well, jnly the subsystems
applicable for g".<;s recovery are addressed
!n detail. While no subsystem has as yet
been proven on more than an experimental
or pilot basis, extensive activity contin-
ues towards achieving viable and cost-
effective mechanical separation.
Froth Flotation for Glass Recovery
Froth flotation is ar, emerging tech-
nique for glass waste recovery. This
technique has been extensively tested by
the Bureau of Mines, and by the National
Center for Resource Recovery (NCRR) at its
full-scale operation in flew Orleans called
Recovery I. The test results indicate low
refractory particle content cullet is
recovered (20). It does not, however, meet
industry specifications for qlass con-
ta iner manufacture (20). Froth flotation
is a technique utilizinq differences in
the chemical properties of fine ground
glass and the contaminants to achieve
material separation. The glass and con-
taminants are mixed with a physiochemical
reagent, which absorbs preferentially to
the surface of the qlass. The coated
glass attaches to bubbles formed by agita-
ting the mixture by air. This glass-rich
froth rises, is swept off the top, and is
washed. Commercial glass-sand operations
and other reprocessing operations have
been using the froth flotation principle
for decades to separate silica sand or
other ores from unwanted minerals.
Normally a series of froth flotation cells
are used where progressively more and more
of the contaminants in the glass are re-
moved .
Optical Sorting for Glass Recovery
Optical sorting is designed to remove
any foreign material fron a glass-rich
fraction of a waste strean and to separate
the qlass by color. This method of
separation is commonly used in the food
processing and other industries and :ias
teen modified fcr the purpose of q'ass
recovery. It is considered a new-tech-
nology for glass recovery.
Considerable research has been con-
ducted on the oropietary Sortex marHne.
The Sortex machine consists of a series of
photocells which separate the opaque
particles from the transpaient particles
by matching the intensity of light trans-
mitted through the particles with a fixed
shade background. In the process, olass-
rich fragments are charged to a Sortex
machine via high speed belts. When the
particle does not match the correspondina
background, a jet of air i^ automatically
released, and the particle is deflected
into the appropriate receiving bins. The
transparent narticles, comprised of
primarily glass particles, are also color
sorted in the photocells by the similar
mechanism.
This method of separation is most
effective when the particle size of the
feed stream is larger than 6 mm (1/4 in)
since.the particles are examined indiv-
iuudlly ds liiey pass "sii'yle file" through
the sorter.
265
-------
TA3LE 5 . SOURCE SEPARATION PROGRAMS COLLECTING GLASS WITH DESIGN VARIABLES
Collection mum1
Site R T CV SI	Haterul contract Hand. ord. Scavenging oro'.
East Lroe. U	- *"
He«ugton, CI	*
Mlerfeury, CI	* *
Maltnein, HA	si
Anoover, HA	x *
Bedford, HA	ft
He-ton, HA	k *
Soo»r*»lle.	i i *
flarblenejo, n*	* *
Hamilton, HA	*
Tiverton, RIO	i
SuonU, NJ	* a *
west Orange* ftj	a 11
Bound Brook., It J	an *
Ithaca, *\t	a *
Bo*>l, MO	i «
Albi/tqlOn, PA	a a	a
C)>ft»r> Heists, PA	*
AtUfVU.	*
Wdlbdr»fi» IN	a
Boulder, CO	*
ClO«ney, CA	a a
Fresno-do* is, CA	* *
QdVlS. CA	a a	*
S^n Luis Obispo, CA	a ¦
Modelto, CA	a *
£1 CerritO, CA	a a
SantA *OS4, CA	a l 		
* R ¦ Rack, t » Trailer, botn ere Integrated collection; CV • CflopartBental uea venide, $T •
Seoarator True*.
b Ton program ^s tne onl> one to sMMy collect «Usi curt>s>0* caHmiun. Progr«ai
generally col lecl newsprint 4nd «euis in waryi*9 conoinatioft*.
tote: This listing is not inclusive.
A series of tests were conducted by'
EPA at a resource recovery plant in Frank-
lin, Ohio (21). Initial findings indicated
that contamination levels of refractories
were excessive. Flint glass averaged si*
refractories per pound, and the color
mixed fraction (green and brown) con-
tained 25 refractories per pound (21).
Systems for Concentrating Glass Wastes
Several othtr preprocessing methodolo-
gies can be used to produce glass-rich
fractions from which glass can be separ-
ated. These are usually used alone or in
conjunction with other units to provide
suitable fractions for.subsequent froth
flotation or optical separation systems.
They include:
• Air classification - normally
used as a preprocessing step for the com-
plete solid wastu recovery system* 7'•ere
are two basic way- -n which air can be
injected' into the system to achieve the
separation of waste materials tiy weight.
The first way of separation involves air
flowing horizontally throuih a fallinq
stream of solid waste material. Heavy
fractions including nlass of the waste
stream are unaffected by the air flow and
fall to the bottom of the classifier. In
the second method of air classification,
shredded solid waste is introduced into
the side of a vertical tube with a risinn
air flow (22). Light parties are
carried out the top of the tube by an air
stream, while heavy particles settle out
at the bottom and are conveyed to sub-
systems for additional separation.
• Risinq current separator - Prior
to allowina the material to enter the
separator, the incoming refuse is first
preprocessed. The particle size of the
separator's feed stcck is between 0.6 and
5 cm (1/4 inch and 2 inches). In the
separator, water is continuously pumned
thrOuyli irie system. As a result of the
rising water current, light organics re-
maining in the heavy fractions are carried
to the top and removed. Heavy.fractions
266
-------
at the bottom of the separator consist of
mixtures of glass, rock, aluminum and
other nonferrous metals, which can be
further processed to obtain individual
species (23).
•	Heavy media separation - Ihe sys-
tem is based on different specific gravi-
ties of the incoming material. The separ-
ator is basically a tank consisting of •
heavy "liquid" {suspension of a mineral
in water) which acts like a single fluid
with high density. As the mixture of
glass, aluminum and other nonferrous
roetaiS is fed to the separator, glass and
aluminum (of lower density) float, while
the other metals sink. The glass and
aluminum are skinned off for further
urocessing.
•	Shredders, screeners, and jiggers
- often are used tc augment the above
described components for concentrating
glass waste fractions.
Current Hiqh Technology Resource Recovery
Programs in the 'Jnited States
Table 6 presents a Dartial listinq of
current resource recovery programs tr.at
recover glass as one product (24). These
resource recovery programs do not now
render a saleable glass prcduct.
tainers at .he point of consumer Durchase.
This was the traditional system that was
essentially'uoplanted by the one-way con-
tainer system In theory, the consumer
returns the container to the retailer to
reclaim the deposit. Once returned, the
retailer stores the container until it is
returned to the bottler. There are bottle
deposit systems now in place in several
states including Michigan, Vermont, Conn-
ecticut, and Oregon.
In response to deposit systems, it is
noted herp that industry has advocated a
litter tax program. Such programs levy
manufacturers, retailers, etc. a small
annual tax that is collected by state
agencies and later parceled out to re-
cyclers througn the foifi of qrants or
loans. This svstem has been implemented
in California and Washington ano has beer,
seemingly successful, both in danpening
enthusiasm for bottle bills and in en-
couraging recycling.
Another form of reuse is character-
ized by the ENCORE! bottle washing opera-
tion which originated in Alameda County
in 1"75 (25). Recconizino that the 74
miiliGn gallons of wine consumed annually
in California reauire over 110 thousand
tons of glass "throwaway" bottles, ENCORE!
attempted to demonstrate to wineries.
TABLE 6. RESOURCE RECOVERY ACTIVITIES WHICH RECOVER GLASS
Capital
LOCItlOn	Proc«tft	Uitpwt	Capacity V*i	Statu!
BaHioor« County, NO
air
moftfflrjr proOwd
flatt
600->Utt)
8.4
operational
BMd9ep©rt, CT
frotft flotation
giaiv culltt
18UU
S.J
not 10 operation
Daor County. Ft
•fc/droDoval"
gtau cul let
3000
I6i.
1981 startup
*ii*«u«ee, w!
air
9'av^j «97req«le
I6UU
IS.
in operation
County, V
fr-jtn flotation
•ueo glau
2000
so.«
tW'ttd up 1979
Or 'eani, 1*
from flotation
91 ast cwllet
'UU
9.1
operational prooiMi
Mevptieat. w
•*70rjoo*«l%/%ortei
color toned 9»aw
2000
71.
S*atedow«
Uil«mq«en, DC
troth flotation

1 uw
>5.
w^aer construction
Reuse Strategies
There are other methods for recovering
glass materials from the waste stream.
One of these is deposit legislation.
Another is the collection of bottles
thrrvunh hutf-SArk r\r(\nvamt for »wa< h* «¦»«
-•«- —j	r • " s' 	 '»•
reuse.
In the former, a legal mechanism is
initiated which places a deposit on con-
recycling centers,.restaurants, stores,
concerned groups and individuals that
empty wine bottle* could be collected,
washed, and reustj on a large scale.
Used wine bottles ?re collected and
returned tc c central j5rt'in3 Warch'jUSc in
Berkeley. There, they are washed and
sterilized in a hydraulic bottle washer-
custom-designed for ENCORE! and incoroora-
267
-------
ting special energy saving techniques. Al-
so, ENCORE! is currently heating water by
uti'izing solar energy. The "revitalized"
bottles are then distributed to participa-
ting wineries.
ENCORE! maintains the strictest qual-
ity control standards and insets all state,
local and Federal hcolth regulations,
identification ot such areas around the
United States where such programs could be
implemented has not been conducted. It is
also noted here that GPI has not" promoted
this form of reuse due to the opinion that
health and safety risks with reusing the
washed bottles outweigh the potential con-
servation benefits.
Markets and Specifications
The waste glass recovery cycle is
closed when materials are ultimately
delivered to and utilized by a market.
Markets or users of glass derived from
post-consumer so'iid waste, though, are not
as well develooed as for other recyclable
materials such as fibers and netals.
Reasons include the extreme nature of
specifications, the need to color sort
glass and the ready availability of raw
resources for container manufacture. .
There -is one geographic region which en-
joys an excellent "mixed" cullet market-
California, primarily becaur.e of the wine
bottling industry. New England is the
recipient of an excellent cullet market,
and there is some mover.jnt toward incor-
porating nixed cullet purchase.
On a national basis, it appears that
new secondary product applications such as
fiberglass insulation will sparV growth in
glass recycling over the near an«j long
term. A reason is these products require
less stringent specifications on color
ard contamination.
Specifications
The specifications or standards are
dictated by the particular product or
apolication being considered. ' In the case
of containers, the standards are rather
well-defined and quite rigid. On the
other hand, for the vast majority of ether
oroducts and potential products which
could utilize secondary glass, the
standards arc cither very broad, vayue, or
.essentially nonexistent (10).. Glass manu-
facturers have to keep close control over
the batch of raw materials to maintain the
quality of the finished product. The
cullet extracted from mixed municipal
solid waste generally consists cf foreign
particles and chemical compounds used in
coloring container glass. These contam-
inants must be removed to a level which
is acceptaole for use in buyers batch
recipe-, or else.rasped effectively. Spec-
ifications noraully concern color and con-
taminants. Exact specifications are
provided in ASTM specification No. C708-79,
ASTM Book Part 41, 1980 issue.
Future Trends
Glass recovery from municipal solid
waste is limite.d at oresent. EPA estim-
ates place recovery at a 4 oercent rate
(26). There is some impetus beinq oiven.
to increase glass cullet usaoe through
national energy and resource conservation
efforts, and stringent air quality control
regulations. Stringent specifications,
the lack of a national mixed cullet mar-
ket, manu'acturer reluctance to maximize
cullet usaqe in batches, and readv avail-
ability of appropriate silica sand
deposits tend to mitigate growth trends.
Several trends have been identified
from conversations with acknowledged in--
dustry experts and current literature.
These ?re listed below:
•	trend emergirin toward smaller
scale operations (economy of scale)
•	reduced distribution lines
•	more enphasis on fibernlass
production, insulation and plastic
composites
•	effort to increase mixed cullet
recycling
•	reuse program resurgence
•	source separation has emerged as
the dominant recovery strategy
t continued inroads by plastics into
traditional glass packaging mar-
kets will spur glass manufacturers,
into secondary products and the
use of metals and plastics.
ENVIRONMENTAL AND ECONOMIC EVALUATION
In' the commercial and i i-iufacturinn
segment', resource reco/ery activities
have been straightforward. The economics
are based on tne material being of known
composition and quality, and free of con-
tamination. In particular, the economics
268
-------
of the plastics industry is very much de-
pendent on the recycle of scrap (waste)
internally or by sale. "Scrap" is usually
reintroduced into the production stream
either directly or "downstream" of the
resin manufacturers. Through the recovery
of plastic and glass wastes, adverse en-
vironmental and economic impacts ere
mitigated and beneficial impacts are
realized.
In contrast, plastics and glass
wastes from municipal sources are mixed
with other wastes and are contaminated.
They must than be separated from other
solid wastes or at least concentrated in-
to suitable, fractions, homogenized, and
decontaminated prior to any successful
utilization. It is apparent that, at the
present time, recycling from municipal
sources is limited. For both plastic and
glass cases, there exists a paucity of
environmental and economic information.
As a result, environmental and economic
impacts are difficult to assess. More-
over, no existing commercial recovery sys-
tem, other than certain pilot mechanical
and sourcp separation systems, recover-
plastics or glass fron MSVI as a sole
procuct. Consequently, Identification of
specific impacts and costs is, at best,
a most -difficult proposition.
Environmental Impacts of Plastics Waste
Recovery
Comoustion
The environmental impact of recover-
ing energy values frcro plastics contained
in municipal solid waste is ^.-iranly
limited to atmospheric emissioi.i. The
impact of plastics upon the overall at-
nospheric emissions from burning refuse is
difficult to quantify. Experimental
evidence has indicated that burning of the
three most widely used plastics «- poly-
ethylene, ;r]ystyrene, and polyvinyl
chloride -- contributes insignificant
emissions under properly maintained com-
bustion conditions (27). Polyethylene
melts early in the burning process, is
completely consumed in any properly
operated plant, and leaves minimal resi-
dues. The only byproducts are carbon
dioxide and water. Polystyrene emits
biact. imute c-a«*t'ic>cS into the 3tn,osph»re
wnen burned in the open air. However, in
properly operated incinerators and boilers,
tnis smoke is reduced and any particles
would be captured effectively by control
devices such as fabric filters, scrubbers,
or electrostatic precipitators.
Some concern has risen over potential
hydrogen chloride and vinvl chloride
emissions from burning waste that contains
chlorinated plastics such as polyvinyl
chloride (PYC| and vinyl cnloriae. Both
hydrogen chloride and vinyl chloride
emissions could be potential health prob-
lems. The chloride emissions could also
erode the metal surface within the furnace
unit. In as much as new installations
have not faced serious problems with
this pollutant, it can be said that the
impact of hydroqe.n and vinyl chloride
emissions is minimal and can be effectively
contained throunh proper equipment desinn
and control systems, i.e. wet scruobers.
Secondary Products
Of interest here is whether secondary
products made from plastic waste represent
potential pollutant sources, whether in
ths process or product. Host of these
products, such as fenco posts, tiles,
plasticizers, paint, extenders, etc., are
too new to have any definitive data as to
their impact. The processes tnemselves
are a different matter. Fcr exaracle,
there have been some problems related to
workplace hazards and plastics recycling.-
Especially pyrolysis of polyurethare will
yiold tolidine diisocyanate (TD1), a
deadly toxic substance (28). Literature
noted that plastic granulators used to
reqrino off-spec resin and other scrap
have be°n developed which mount the entire
cutti1 :hamber assembly on shock absorbers
to is .-:e the cuttinn impact from the
frame of the machine and decrease noise (7).
Economic Impacts of Plastic Recycling
To ensure that maximum profits tan be
realized from purchased materials, manu-
facturers strive to reouce waste to a min-
imum. Some plastic scrap which is rot
reprocessed internally by primary polymer
producers and fabricators is sold to scrap
dealers or processors. The scran is
attractive from an economic viewpoint,
since it creates new business opportunities
and reduces raw material costs. For ex-
amole. one company is reported tc purchase
fairly clean scrap polyethylene tnn and
bags, and rework it. The recovered
granulated pellets a-e added to virgin in-
269
-------
put in the ratio cf five to ten percent
scrap. The cost of recovered polyethylene,
including the purchase of scrap and
processing, is said to be less than 50
percent for the virgin material (29).
While inplant recycling is an estab-
lished fact, once the olastic waste be-
comes part of the municiual waste stream,
the economic incentives for recovery are
generally insufficient to overcome the
cos:s of recovery, except in the case of
consistent scrap or PH.
Environmental Impacts of Glass Recovery
Glass waste from the glass manufac-
turing industry and in many commercial
establishments is routinely collected and
used in the manufacturing process. Ad-
verse environmental impacts associated
with these practices are minimal.
In fact, environmental impacts
associated with glass waste recovery from
municipal refuse are mostly beneficial..
Major impacts are listed below:
•	lessened impacts from extractive
industry pperat iors
•	diverted landfill voluire
« reduced air emissions from glass
manufacture and the potential
offset from cullt't reuse
•	energy conservation from increased
cullet usage in batches
By increasing recyclable volumes in-
to batch processes, a considerable reduc-
tion could indirectly occur in environ-
nental emission. Th.-ough virgin materials
extraction, ano virgin materials processing
considerable amounts of energy are ex-
pended and environmental degradation
occurs (30).
The introduction of an inert material
such as glass into landfill systems poses
no environmental pollution problem for
around water quality. However, by con-
tinuing its introduction into scar:e land-
fill, especially those classified for
hazardous waste, glass acts as a potential
competitor for limited space.
There have been investigations,
arrnn'inn In literature f 10") , that con-
• • - -	^	i.l
firm the relationship between cullet rein-
troduction into production and lowering of
air emissions. Raw virgin materials used
270
in glass manufacture go through complex
chemical reactions resulting in the release
of gases and particulates. Cullet, havinn
once undergone these reactions, will not
add to emissions (31)' A'so, it has been
documented that cullet lowers the neltinq
teirerature. Gy doing so, emissions
associated with higher temoeratures (?uch
es N0X) are mitigated. One glass manu-
facturer in California has reported that
increasing cullet usage has allowed his
operation to meet air quality regulations.
t!ith record to enemy, cullet intro-
duced into tlie bat:!i at a controlled rate
can reauc.. Iti.iQ energy by about 1/4 to
1/3 of a percent per 1 percent of cullet
added. This fonvjla is applicable at
charges up to iO percent cullet. Antici-
pated energy savinos could ranne as hiih as
11-12 percent per batch (32-35). While
detailed energy studies have been conducted
for glass container manufacture in Europe,
and limited studies in the U.S., no studies
have been performed for other qlass produc-
tion sectors.
The EPA in the 5-volume study on the
larblenead and Somerville. Massachusetts
source seoarition programs conducted
analysis cf energy consumption-savinns for
three alternative scenarios: landfill,
transfer station, and a ccmtination system
involving source separation. Of the
alternatives, the combined system 'iad the
highest enerqy return (36).
Iripacti of reuse systenis would apoear
tc b? mainly focused in the economic sector,
thougn bottle washing should increase
emissions in wastewater (primarily food
stuffs and organics), decrease landfill
burden, and save enemy (37).
Economic Impacts r>f Glass Recovery
Costs can "center" ov*;r specific
operational modes. For example, trans-
portation of products to market constitutes
cne mode. This cost center is slightly
different for intermediate processors and
conmunity recycling centers. For all
proqrams, though, generalized cost centers
include:
•	collection
•	processing and storage
•	transportation
•	marketing
•	administration
-------
Cost elements involved in these cost
centers are factors that change from area
to area, an1 fro type nf technology to
marketing conditions and products. Cost
elements include, and are uniform, among
all cost centers for specific technologies:
•	labor
•	utilities and fuel
•	capital expense and amortization
e maintenance
» overhead (insurance, etc.)
•	building modifications
•	publicity
While no specific study has detailed
all these centers and elements in a com-
prehensive economic analysis, tjiere have
been individual reports detailing one cr
more of selected factors.
Representative costs for dropoff sys-
tem structures were prepared usinri estim-
ated costs per ton developed on a com-
parative basis. Llsinq two studies and two
estimates, costs were assigned to citizen
preparation, public payment, processing
eauipment, laDor, construction, storage,
transportation and administration. Costs
ranged from	« S37/ton. The two
largest elements were labor and transpor-
tation.
A true economic picture of recycling
centers is difficult to obtain. "Hidden"
costs can include volunteer or low-paid
labor (both 'r source generators and at
the centers), free materials, free land-,
government grants, etc. As noted earlier,
some programs have access to CETA employ-
ees (38). On the other hand, there are
intangible benefits which cannot be in-
cluded or a ledger including the benefits
in education and diverted disposal. There-
fore, in Sn economic study done by EPA,
the costs per ton of recycled material
varied from S6 per ton after disposal
credit t& a net S169 per ton excluding land-
fill credit (39)..
In a concurrent analysis o'f curbside
collection programs (40, tl) representa-
tive costs were developed based on all
recyclable* collected. It showed a net
gain in the Marblehead, MA program, but a
net loss in four other programs. These
rAttt woro aco'nnoH «n f-ho hac-ic	art-nil
processing per ton, disposal savings per
ton, a.id revenue per ton.
In a study performed by CPA of the
Franklin, Ohio resource reco/ery plant, the
economics of olass recovery wv»e ess^ssed
at the 50 TPD level (21). Projections were
then made at the 500 TPD and 1000 TPD
levels using 1975 dollars. Cist projection
included facility amortization, operating
and maintenance costs, ana olass, ferrous,
and aluminum sales. The study concluded
that combined sales held economic viability,
but operating the entire subsystem (heavy
medii' separation, electrostatic separation,
eoloc sorting, etc.) for glass recovery
alone was not feasible. The relative sig-
nificance of the revenues potentially
available are also important in this re-
spect. One-third of the projected reven-
ues would be attributable to qlass. Yet
over half the costs associated with the
plar.t are assoc-'ated with glass recovery.
This is because color sorting and nlass
processing is highly capital and enerqy
intensive regardless of whether aluminum
ferrous is recovered. All pre-processino
steps are necessary for qlass recovery.
A critical assumption of the study was
that at larger installations, ceramic con-
tamination could be reduced"to meet
stringent industry specifications or that
the glass industry would accept a higher
level of contamination than was accepted at
the time. (These specifications have not
been relaxed sufficiently. (42))
SPECIFIC ECONOMIC ISSUES
There are a few economic issues im-
portant to a full treatment of plastic and
glass waste recovery. These include the
following:
•	employment and other socio-economic
impacts
•	litter tax and reuse strategies
•	obstacles to increased recycling ¦
•	economic development
•	diversion credits
Employment and Other Socio-Econoriic Impacts
Recovery of waste materials has posi-
tive employment and social impact. In the
case of collection centers, such programs
directly involve citizens in solid waste
management activities. As a result, it can
holr tho nuKl^r nnrlorrf artH tho r»«*r»h"lAmr ftf
•.«•••» -¦ •w . r- '¦**' 1 • **	-- ¦- r* • ^- .. ~ ¦
sclid waste management and achieve certain
levels of conservation and litter a'jate-
271
-------
rent (18). An important -onsideration is
that source separation-collection centers
are labcr-intensive. Many programs are
initiated with the aoa. being to hire the
handicapped or difficult to employ person
(12). Another aspect is that collection
centei-i assist in raising revenue to under-
take corarnuMt> beautification rrograns (IS).
It has been theorized that recycling
industries, basically "conservation of
energy ir.djstri.js" which also reduce pol-
lutants generation and material's usage,
viill serve as major foci of investment and
urban redevelopmer' strategies in the near
future (1j). Not only will employment as-
pects be served, but additional market
capacity will encourage further reeove";'
of waste materials. This has a direct im-
pact on the quantities of materials being
landfilled or otherwise disposed.
Litter Tax and Reuse Strategies
In sonie states, "bottle bill" oppon-
ents have supported "litter tax" measures
designed to tax litter generators or those
who produce items t-nat berome l'tter, a
small amount. This levy is then collected
.and composed into a fund, that may bp racie
available to anti-litter programs and
"recycling" programs. In Calif'rnia, a
"litter tax" program has been i" effect
for two years at tnis writing. Over ?5
million has been awarded to waste recovery
programs for developmental activities. The
program will he operational for another 3
years (43). This litter tax program has '
effectively defused bottle bill proponent
efforts.
Waste reduction activities have had a
degree of impact on the composition of the
waste stream. According to data from
Oregon, metal containers have been reduced,
and glass has become a larger component of
the waste stream. In another study in.
Michigan, it wa; found that metal contain-
ers were initially reduced, but within a
year had increased although not to levels
prior to bottle bill enactment (44).,
Reuse measures do provide increased
recovery opportunities. For example, in
Michigan PET 2-liter containers are being
returned tnrough the deposit system (44).
PET !'	rCCVClSd Koranco tho roiiCC)
system ensures a consistent and clean
recyclable component. As PET is the only
pla'itic beverage container, there is
currently no major or.jblem of compati-
bility of plastics. The majo- problems
arise with the PET bottle contaminants,
e.g., aluminum caos, paner labelino, themo
formed bases, and different colors used
(45).
Obstacles to Recycling
Current obstacles exist whicn inhibit"
increased glass and plastic recvclinn. One
obstacle is the general Drice differential
between virgin and recycled materials.
Virgin materials have been cheaper in the
U.S. because natural resources have been
plentiful; bec.;ise public policies favor
virain materials; and because'environr-en-
tal and other social ensts (externalities)
have been omitted from t^e orice (fl6).
Federal Land Us^ge
One obstacle is oublic policy on rad-
eral land usage. Virgin material ex-
tractors, for example, gain competitive
advantage frcr. the resources and technical
and scientific assistance from a number of
supportive Federal agencies (47).
Tax Structures
Tax structures r'.-nerally favor viroin
material processors. Over the years, the
Federal government has developed tax
policies that favor extractive industries.
For example, caoital costs it.juried in ex-
ploring and bringing mineral deposits
(glass silica) into production may be de-
ducted as current expenses rather than
amortized over the useful life of the
property. Also, the costs of developnent
are deductible after a commercial mineral
is established. While it is trje that at
one time it was necessary to quickly arid
comprehensively exploit our resources,
it is not necessar''.y true today.. Incen-
tives to explore and develop virgin mater-
ials retard demand for investment in re-
cycling (47).
Railroad Freight Rate Discrimination
Transportation typically accounts for
a very large fraction of the delivered
cost of materials. Ideally, all materials
would be charged costs that relate to the
arfual rnst. of hauling. However, in prac-
tice. different conwodities and shippers
are charged rates that differ relative to
the reguired costs incurred by the carrier.
272
-------
a condition referred to by many as freight
rate "discrimination" (30). Substantial
and systematic rate differentials have been
held by Congress, the U.S. EPA, the secon-
dary materials industry and the U.S.
Supreme Court to contribute to inefficient
relocation of resources and to work against
resource conservation. Currently, the
Interstate Comerce Commission (ICC) is
under litigation to revise their rates
relative to scrap materials; to, in effect,
equalize freight rates. The ICC recently
estimated that glass freight discrimina-
tion amounted to Si.70 per ton (30).
Economic Development
Traditionally, the primary use for
reclaimed glass is in the manufacture of
new containers. However, this is not al-
ways feasible or practical especially
where extensive mixing of cullet arA con-
taminants occur (necessitating1 high costs
of up' vading) and where transportation
costs inhibit purchase of cullet. In all
cases, where the cost of the cullet ex-
ceeds the cost of virgin materials, the
primary markets will utilize virgin mater-
ials.
'The need for development of secondary
product markets indeoendent of, ir. many
cases, artificially low-cost, virgin
materials, is becoming increasingly urgent
as solid waste management enters an era of
large-scale salvage. While many processes
are technically feasible, only recently
have economics improved to a point where
secondary product development provides a
competitive return on investment.
Economy of Scale
An area of importance is the apparent
reversal of trend from construction of
larger, regional-type orojects toward
smaller, local and community-scaled indus-
tries (13). The impetus behind this is
partially explained by the following:
•	the n*ed to provide loc3l employ-
ment
•	the ready availability of waste
resources for utilization
•	with local cor.trol of these two.
aspects, a lessening of impact by
~icrn economic facto" (including
inflation and recessions)
•	less capital is necessary for
start-up, research, operation and
construction, hence smaller com-
munities and small businesses are
able to afford such systems.
6PI evaluated the potential for the
commercialization of qlass rubble buildinn
panels (48). The' constructed flat panels
produced by vibro-i mpaction techniques
measure up to 10 feet x 4 feet x A feet,
and weiah up to 1,900 lbs. The finished
product would typically be composed of 94
percent ground waste glass, and the re-
mainder a composite of clay and demolition
rubble. Various complex sizes and shapes
can be made, with ultimate usage as decor-
ation or structural application. The
immediate competition is with brick and
please concrete panels for structural
application. Decorative panels must com-
pete with rock, marble, mosaic, stucco,
and glass-plastic composite panels (rt8}.
Under economic analysis, it was
determined that investment and development
of thpse panels as an industry is closely
keyed to "carefully defined (oeographic)
areas at a carefully determined scale of
ooeration." Manufacturing costs range for
snail-scale operations costs of S'/sq
ft to S0.3C for larger operations; bott.
are competitive orices. Wall panels.made
from "virrjin" materials are currently
being produced in more than 200 different
plants within the U.S.. reflecting a
strong local nature of the business. Most
of the operational and successful plants
are located within 7G miles of their
imnediate market (48).
• t'nder a simil. investigation, mineral
wool insulation uti; zing waste glass was
evaluated for technic*! and economic
feasioility (49). Ust -s for up to 50 per-
cent cullet change include insulation
batts and blankets, blowing wool, and high
temperature felt insulation. Benefits to
the manufacturer include:
•	permits significant shortcut in
manufacturing by bypassing the con-
version of silica sand and chem-
icals to the glass feedstock state.
I useful for up to 1,200 F tempera-
ture, which represents a 50 per-
cent increase in use
•	costs less to manufacture
However, extensive contamination
(stones, etc.) can significantly impair a
mineral wool process by slugging the "spin-
27J
-------
ner.
Estimated capital requirements (incor-
porating o TOO percent inflation factor
for 1971 figures) for an 18 ton/day facil-
ity is shown in Table 7.
excreta
•	economics of manufacturinn terrazo
with waste lass aggregate
•	economics of manufacturing foamed
glass construction materials made
with waste glass and animal
TABLE 7-. ESTIMATED CAPITAL REQUIREMENTS
FOtl 18 TON/DAY GLASS rfOOi PLANT
Category	J
A.	Amortized investment
Engineering, research and
development	100,uuu
Start up	200,000
Subtotal A	3uj,0uo
B.	Fixed investment
Structures and improvements	260,000
Macninery and equipment	580,000
Subtotal 8	840,000
C.	Recoveraole investment
Land	80,0uu
Working capital	4jp.OOU
Total recoveraole investment	J 530,000
Total capital requirement	Ji.oio.ouO
D.	Capital requirement per ton/day
capacity*	$ 92,777
*2f>0 days per year operation
Tl« total manufacturing cost for the
18 ton/day operation is between S1S7.50
and S300.00 per ton. This range is accept-
able within the limits of current glass
wool prices (49).
Secondary Materials Development Opportun-
ities.
Regardless of the HI
has investigated several potential waste
utilizing enterprises, however, all have
been related to low specification uses
which could utilize glass derived from
resource recovery enterprises. These
enterprises iriCiuue. (jj-jj);
t economics of manufacturing ceramic
t'le with waste glass and animal
excreta
• economics o' manufacturing slurry
seal with waste alass agorenate.
Diverted Disposal Values
Materials diverted by source separa-
tion activities have a diverted disposal
value. Alt lough not necessarily credited
to a center, the value shoild be consid-
ered when assessing program viability.
Savings in diverted solid waste dis-
posal costs are dependent on whether the
municipality in which the proqran is
located operates its own facility or
franchises out to contractors. In another,
secondary sense, the savings value varies
n i cm lite vuic ui me	i itikiifu.
Sanitary Landfill
274

-------
Benefits of source separation on land-
fill operations include a decrease in the
rate of use of remaining landfill space and
a decrease in landfHl equipment use.
Based on operating costs ranging from SI
to 58 per ton, an average of $4 per
ton is estimated. Land costs are assumed
to represent SO.50 of the total cost
based on the disposal of 10,000 tons per
acre and a net land cost of $5,000/acre.
Therefore, diversion of recyclables can be
assumed to potentially save SO.50 per ton
in land costs at the landfill and up to
S3.b0 per ton in operating cost savings
(54).
Incineration
The diversion of-materials from incin-
eration through source separation activi-
ties can be expected to reduce equipment
usage and residue disposal requirements.
Further, there is a net benefit to energy
efficiency when noncombustible recyclables
are removed. Incineration costs range
from S20 to $30/ton with an average of
about $20/ton.
In addition, ash residue must be
hauled for final landfill disposal. Resi-
due transport costs vary with many factors,
but. can be assumed to average $0.50 per
ton of residue (54). Total costs can be
assumed to be equivalent to the costs in
ti.-e preceding landfill discussion.
A 95 percent reduction of weight.of
material can be assumed for paper wastes.
No such corresponding weight reduction can
be assumed for glass and metals, both non-
combustible, if processed through an in-
cinerator. An average of Sll per ton can
be assigned to source separation as a
result.
STATE-OF-THE-ART PLASTICS AND GLASS WASTES
RECOVERY ABROAD
The study of technologies for plastic
and glass wastes recovery and recycling in
other countries has limited but worth-
while application tn the United States.
In most foreign countries, capital is
scarce and labor usually plentiful. There-
fore, emphasis is most often placed on
labor-intensive materials extract'.on rather
than energy-capital intensive extraction
techniques as used here in the United
States-. In the industrialized nations of
Europe and Japan where the situation is
analogo'-'S to the United States, maximizing
human energy is becoming increasingly
important as fossil fuel-derived energy
costs soir. It is expected that this will
be true, if not already, for the United
States.
Material; and enetgy recovery from
solid wastes has traditionally been prac-,
ticed around the world and in the United
States. In some countries, recycling
practiceis have retrained relatively static
for thousands of years although materials
have changed. Sone countries are more ad-
vanced than others depending on the degree
of technological sophistication and indus-
trial-commercial organization. These rely
more on state-of-the-art techniques.
In Cairo, Eqypt waste naterials have
been recycled for thousands of years by
so-called "refuse people" who live outside
Cairo in their own "refuse city." Today
they collect and process sourte separated
wastes in much the same manner.as in the
past: oxen drawn carts are used for
collection of refuse. Of course, smelting
of aluminum, and plaitic reclamation are
relatively new (55).
In Japan, plastics reclamation from
municipal waste is more widely practiced,
especially by technological methods, than
elsewhere. A significant reason behind
this is that plastics represent a sweater
percentage of the Japanese waste stream,
energy is at high oremium, and recycling
is highly institutionalized. In fact,
Japan passed a law recently nandating
nationwide recycling (56).
In Europe, there are found similari-
ties with the United States in issues,
objectives, and technologies. Europe is
slightly advanced in some areas, especially
market develc:nent for secondary nlastic
products.
RESEARCH ON PLASTICS AND GLASS WASTE
RECOVERY/REUSE
Plastics Wastes Recovery Research
Basic plastic waste recovery research
programs generally focus on the site
specific needs of manufacturers. These
include: (1) processes for the chemical
or mechanical separation of various blends
ot plastic waste, [2) processe* oi auui-
tives which improve the bonding character-
275
-------
isties of mixed plastic types, (3) devel-
opment of specifications to both aid con-
sumers ir. ident fying plastics and to en-
hance recyclab1' tity and, (4) processes and
systems to upgrade segregated plastic
scrap types normally uniformly contanin-
ated (e.9., PVC nolded around copper
w i re).
Less research has been devoted to
recycling plastics from mixed municipal
refuse due to many factors, including cost-
effectiveness, lack of markets, low vol-
ume, and lack of demonstrated need. In
Europe, the Flakt system concentrates
plastic wastes, hut this is currently rot
recovered. Rather, it is removed as a con-
taminant from paper f'bers (57). In the
United States, the research efforts focus-
ing on municipal refuse as a sourc-; of
plastic for recovery are combustion-energy
recovery operations, which favor the high
Btu content (42 kj/g (10,000 Btu/lb)) of
plastics, selected solvent separation,
cryogenics, source separation, air separa-
tion, electrodynamics, sink flotation,
and research re)3ted to PET bottles.
Glass Waste Recovery Research
Research efforts for recovery/reuse
have been concerned with mechanical separ-
ation, source separation, new secondary
products ai'J reuse programs.
foremost, a market for the recovered
glass must exist. Presently there are
only linited markets. One area of re-
search that has been promising for glass
waste recovery is its use in secondary
products. Products such as glasphalt and
glass foam insulation demonstrate the tech-
nical feasibility of using glass waste for
secondary products.
An area of some interest is reuse of
products.. The ENCORE! system, a wine bot-
tle washing operation, depends on free
market forces. It is both profitable a-id
effective. Although there is a question
of safety, no serious problems have been
encountered to date.
CONCLUSIONS
The following conclusions were devel-
oped based on the State of the Art:
Plastics
•	Industrial anc comrercial sources
efficiently recycle using simple,
proven technology. The main
reasons are waste materials are
concentrated, relatively uncontam-
inated and usuallv of known quality
and conposition.
•	No proven conmercial scole recovery
system sinqularlv effects recovery
of waste. Rather, such materials
are recovered as one component of
an over-all recovery-collectior.
approach.
•	Secondary products, on the whole,
have not had specifications dev-
elooed on product reuse. , This has
acted as a barrier to increased
utilizations since reuse processes
have not necessarily been stan-
dardized.
•	Combustion and energy recovery
nold the greatest premise for re-
covery of the bulk of the plastics'
fraction of the solid waste stream
due to the number oi different
types of plastics and th» diffe'rino
degrees of degradation of compon-
ents.
•	Source separation from the indus-
trial. to the residential levels
constitutes the only significant
recovery of was.te iron municipal
waste soruces.
•	For the immediate future, indus-
trial and commercial sources will ~
comprise the majority of recycling
activity. Recovery fron post-
consumer wastes must overcome-sig-
nificant market, institutional,
technical, transportation, and
specification barriers in c-der to
compete successfully with virqin
products.
Glass
4 Glass manufatturers claim that 25
percent of the post-consumer waste
stream could be recycled right now.
Tr",3purta11 OH auj col ltft.1 fori/
delivery problems and contaminant
levels mitigate aoainst this.
•	Industrial and cotmercial sources.
176
-------
efficiently recycle using simple,
proven technology. The main
reasons are waste materials are
concentia ted, relatively uncon-
taninated and usually of known
quality and ccicosition.
e Municipal sources of wastes are
most often nixed with other com-
ponents of refuse; hence, recover}
is difficult with poor economics;
also, the ease of obtaining raw
materials, prevents a significant
. ecovery incentive.
•	No proven commercial scale recovery
system singularly effects recovery
of glass. Rather, such materials
are recovered as one component of
an overall recovery-collection
approach.
•	Scarce separation often lacks in
collection equipment dr.d efficient
piocessing; hence, recovery is
inhibi ted.
•	Secondary products, on the whole,
have not had specificati J devel-
oped on product reuse. This has
acted as a barrier to increased
utilizations as reuse processes
have not necessarily been standar-
dized.
« Mechanical recovery systems for
glass wastes nave primarily orig-
inated from ot^ec industries s'jcfc
as nining. They laqfc proven usage
in waste separation where moisture,
composition, physical prof^ties,
and economics vary widely.
•	A national market for mixing color
glass cull-el could significantly
enhance recovery of jlass wastes
from municipal sources by simli-
fying collection and processing.
•	Source separation from the indus-
trial to the residential levels
constitutes tre only significant
recovery of waste from municipal
waste sources.
t For the immediate future, indus-
trial and commercial sources will
comprise the majority of recycling
activity. Recovery-from post-
Consumer wastes must overcome sig-
277
nificant market, institutional,
technical, transportation, and
specification barriers in order to
compete successfully viith virgin
Droducts.
REFERENCES
I.	National Center for Resource Recovery,
Inc. (NCRR). Plastics Fact Sheet,
October 1973. 2 pp.
?. Chester U'. Ma>-yno*ski, Disposal of
Polymer Solid Waste by Primary Polymer
Producers and Plastics Fabricators.
EPA-PA 36-68-160, U.S. Environmental
Protection Agency, Washington, DC,
1972. 92 pp.
3.	Hilqiom, Jack. Incentives for Recy-
cling ttnd Reuse of Plastics. EPA-SW-
41C-72, U.S. Environmental Protection
Agency, US EPA PS 214-045, 1972.
?i'6 pp.
4.	Hutchins, J.R. and R.V. Harrinnton.
Glass. Coming Glass Norks. Re-
printed frort encyclopedia of Chemical
Technology, 2nd edition, vol. 10,
pages 533-604. 1966 by John Wiley and
Sens, Inc.
5.	Anon. Glass Industry Hcurnal, 59(12),
December 1979.
6.	Anon. Recycling and the Consumer.
U.S. Environmental Protection Agency,
Washington, !)C, 1973.
7: Anon. Scrao Recovery: Csn Voo Afford
.'lot To? Plastics 'World, flarcri 1979.
S Technical communication, T.D. Uharton,
Jr. to Lad Thanka, Ph.8. February 12,
1980.
9.	Owens-II liioii. A Guide to PE1 Plas-
tic Beveracje Container Recycled, 1980.
10.	Oarnay, Arsen and vj t. Franklin.
Salvage Markets for Materials in jcOid
Wastes. CPE69-3, U.S. Environnental-
Protection Anency, 1972.
II.	Dirckett, E.J. Glass Recovery From
Municipal Solid Haste. National Center
for Resource Recovery, Inc.,' Wash--
I....* in-rr*
1'iyvyti^ uv1 uutic uty.
12. Anon. Fourth Report to Congrsss on
-------
Resource Recovery and Haste Reduction.
SW 600, U.S. Environmental Protection
Agency, 1977. N2 pp.
13. Seldnan, '.eil, Richard Anthony, Jon
Huls, Mindv Kershner, Jervl Snecter,
Jir. Sullivan, National Recycling Re-
search agenda Project. NSF-OPA 79--
17013, National Science Foundation,
Washington, DC, IScO. 100 pp.
K. Technical Communication with Bob
Drake, Technical Director, Glass
Packaging Institute, Washington, DC.
15.	Abrahams, Jr., J.H., Recycling Con-
tainer Glass - An Overview. Glass
Container Manufacturers Institute,
Inc. Proceeding of the Third Mineral
Waste Utilization Symposium, 1972.
16.	Cook, R.F. The Collection and Recy-
cling of Waste Glass (Cullet! in
Glass Container Manufacture, A paper
presented at the First World Recycling
Congress, Basle. Switzerland, March
1978.
17.	l-.'oodyarc!, John. Reduce, Heu?e,
•Recycle: Conference Proceedings,
Presentation on Source Separation.
California Resource Recovery Associa-
tion, 1978.
ISi O'Brien, J.R.	Source Separation...
The 'Recycling Center' Approach.
NCRR Bulletin,	V1II(2), Spring 1978.
pp. 31-37.
19.	Co*>en» David. National Survey of
Separate Collection Programs. 5W
776. U.S. Environmental Protection
Agency, July 1979. 100 pp-
20.	Morey, 8. and Cunr»ngs, J.H. Glass
Recovery from Municipal Trash by
Froth Flotation. Proceedings of the
Third Mineral Waste Utilization Sym-
posium, 1972.
21.	Garbe, Y. Color Sorting Waste Glass
at Franklin, Ohio. Resource Recov-
ery and Energy Review, 2(6), November-
December 1976. p. 20
22.	New Orleans Resource Recovery Facil-
ity Implementation Study Equipment,
Economical Environment. National
Center for Resource Recovery, Ire.
Washington, DC, September 1977.
23. Resource Recovery Systems. National
Resource Recovery, Inc., Bulletin
No. 3, October 1, 1°74.
?£. NCRR Editorial Staff. Resource Recov-
ery Activities, A Status Report.
NCRR, Washington, DC, January 1379.
25.	ENCORE! Program Leaflet, June 1977.
26.	'lilnrom, J. Waste Glass. Resource
Recovery and Energy Review. Vol. 2,
--7, 1975.
27.	Pamohlet from the Public Affairs
Council of the Society,of tne "las-
tics Industry, Inc., New-York,
28.	Or. Hanna Klaus, Owner of Proprietary
Plastics Waste.Recovery System,
September 9, 1979.
29.	International Research and .Technolooy
Corp. Recycling Plastics - A Survey
and Assessrent of Research and Tech-
nolcoy. 1972 Studies for The Society
of the Plastics Industry.
30.	Resource Conservation Committee.
Choices for Conservation. Final
Report to tne President and Congress,
July 1979. U.S. Government Printinn
Office. 130 pp.
31.	Vaunhan, O.A., et si. An Analysis of
Current Impact of Plastic Refuse
Disposal Upon the Environment. NERL
P8 238-654, U.S. Environmental Protec-
tion Agency, Cincinnati, Ohio, 1974.
57 pp.
32.	PEOCo Environmental. Background
Study in Support of New Source Per-
formance Standards for Industrial'
Boilers. Departraent of Eneray, Kay
1979.
33.	Anon. Energy Analysis of Secondary
Material Use in Production Manufac-
ture. State of California. Resource
Recovery Division. Draft Report.
1979.
34.	Personal correspondence with Sob
Drake, Technical Director, Glass .
Packaging Institute, Washington, DC.
July 1979. .
35.	Personal conmunlcation with Guardian
Industries, Kingsbury, California,
278
-------
June 17, 1979.
36.	RPA. :-'.ultinaterial Source Separation
in Marbleheed and Somervi'i le, MA.
Energy Use ,ind Savings from Source
Separated Materials and Cther Solid
Waste Management Alternatives for
Marblehead. Vol IV, SW 324, Decem-
ber 1979. 22 pp.
37.	Huh, Jon. Solid Waste Management
in Sncialist Countries. Unpublished
pa pet , November 197S.
33. RCA. Multinaterial Source Separation
in Marblehead and Sotiervil le, .HA.
Composition of Source Separated
Materials and Refuse. U.S. Environ-
rental Protection Aaency November
1979. 89 pp.
39. Staff. How to Starta Recycling
Center. California Solid Waste Man-
agement Board, 1979. 37 pp.
10. Stearns, R.P. and Robert H. Oavis.
Economics of Separate Refuse Collec-
tion. Reprint from Waste Age,
May-June 1974.
41.	Prepared.by Steve Howard for Glass
Packaging Institute, January 1979.
42.	Systech. Tecnnical, Environmental
and Economic Evaluation of Glass
Recovery Plant at Franklin, Ohio.
68-01-2211, U.S. environmental Pro-
tection Agency. 112 pp.
43.	Review of press releases on funding
awards for SB 650 program.
44.	Sharkey, Leo. The Michigan Experience.
Beverage World. Nov. 1979.
45.	Anon. PET Bottle Update: Technology
and Market Trends. Industry News.
January 1980. pp. 111-117.
46.	Seldman, Neil, Ph.D. and Daniel Knapp.
Waste Knot: The Politics of Garbege
and Recydinq. To be published
Surmer 1931.
47.	League of Women Voters. Recycle,
1972. 40 pp.
48.	MRI. Economics of Manufacturing
Glass Bubble Building Panels. Pre-
pared for Glass Containers Manufac-
turers Institute, 1971. 26 pp.
49.	MRI. Economics of Manufacturino Glass
Wool Insulation. Prepared for Glass
Containers Manufacturing Institute,
1971. 20 pp.
50.	MRI. Economics of Manufacturing
Ceramic Tile with Haste Glass and
M.iinal Excreta. Prepared GCMI, 1971.
20 pp.
51.	MRI. Economics of Manufacturino
Terrazo with Waste Glass Aggregate.
Prepared for GCMI, 1971. ?6 pp.
52.	MRI. Economics of Manufacturino
Foamed Glass Construction Materials
Made with Waste Glass and Animal
Excreta. Prepared for GCMI, 1971.
26 pp.
53.	MRI. Economics of Manufacturing
Slurry Seal with Waste Glass Agare-
qate. Prepared for GCMI, 1971. 26
pp.
54.	SCS Engineers. Analysis of Source
Separate Collection of Recyclable
Solid Waste -- Collection Centers --
EPA-68-01-0789. U.S. Environmental
Protection Agency, August 1974. 50
PP.
55.	Gurderin, Charles. Presentation to
World Bank on Recycling in Cairo,
Egypt. Presented November 7, 1979.
56.	Anon. Resource Recovery and Energy
Roundup. Resource Recovery and
Energy Review. November-December
1976.
57.	Ceduholm, Curt. Using Air Technology
to Recover Resources from Soli a Vfste.
Reprint from ASFA International 19/9/
1.
279
-------
FORMALISM VERSUS REALITY IN ECONOMIC FORECASTING
Stedman B. Noble
Consultant
Washington, D.C. 20003.
David van Tijn, Consultant
United States Metric Board
Arlington, Virginia 22209
ABSTRACT
This paper originated in a review of two reports which predicted the composition of future
national levels of solid waste. This involves associating technological information with
the formalisms of economic analysis. The paper defines this problem in a detailed context.
1. INTRODUCTION
This paper originated in a review of
two reports which predicted the com-
position of future national levels of
solid waste. In particular, the assign-
ment was to predict the future volume of
sixteen materials in this waste stream.
Predicting the waste stream Involves
at least itapllcitely the prediction of
the future use of end items in which the
materials are embodied. One study used
very little technique beyond an explora-
tion and use of available data sources.
The other study used an input-output
based model to explicitely predict the
usage of end items. Neither approach,
however, was successful in producing the
best estimates that detailed tech-
nological analysis, married to predic-
tion of usage on a micro-scale, can cane
up with in particular cases.
The approach of each study had cer-
tain advantages, but each study encoun-
tered major difficulties. We found that
the two approaches are much more dif-
ficult to reconcile than it might
appear.. Since this difficulty—how to
reconcile a knowledge of reality with
the current economic models—Is a wide-
spread problem, this paper focuses on
this: point. It defines the problem in
precise context of the interface between
two professions.
2. TWO REPRESENTATIONS OP Ml ECONOHY
The two studies reviewed started by
examining production data, and then trans-
formed that information first into final
products, and then into a waste stream. In
both, the first and key step is to estimate
the (potential waste-) material content of
articles put into service, i.e., the flow of
materials Into products. There are two wayB
of performing this estimation! the iftput-
outpuc method and the flow method.
a. The Input-output Method for Finding
Flows.
An input-output table .'s essentially a
doubly balanced accounting model among
'industries' and final users. The elements
are the purchases by each sector from
others, and the sales of each sector to
others. There are balancing accounts for
such destabiliiing factors as inventory
changes, depreciation, and'profit.
An Input-output table can be used to
Identify material flows. When this is
done, the model for many materials is found
to have a simple structure in several
sensess the largest flows are short, and a
few fl.Tws account for the preponderant part
of output. To Illustrate, the three larg-
est flows for aluminum to households are:
280
-------
Al—* rotor vehicles	>192,000 B. tons
hi * trailer coaches	>122,000 «. tons
Al——* metal stampings i motor vehicles
70,000 o. tons
The quantities represent the amount of
aluminum fiat is used (according to the
input-output table.) In two cases, the
only multiplication is a single coeffi-
cient, aluminum purchased per vehicle
tines final demand for motor vehicles or
trailer coaches. In.the other case, two
coefficients in the model are multiplied
(aluminum per metal-stampings tines
metal stampings per motor vehicle], and
the product is multiplied times final
demand for motor vehicles.
Each of these lines is called a path
(a technical term of the mathematics of
directed graphs.) A previous report*
studied the paths associated with the
16 materials whose use results in muni-
cipal waste. It analyzed the paths that
it called "significant," that is, that
accounted for 3'< of the output of each
material. For this definition of
significance, it was necessary to have
213 paths, an average of 13 1/3 paths
per naterial.
Two aspects of structure are signif-
icant: (a) how many parts of the eco-
nomy affect demand foe each naterial.and
(b) how indirect is ibe connection. The
number of significant paths for each
aaterial ranges fron 2 to 35 for various
materials. The spread in the lengths of
the paths la the following:
Nuabei of paths:	213
Single step	91
Two steps	107
Three or four steps	15
The analysis must be taken further.
Bach of the paths ends with a final pro-
duct purchased by consumers, tt must be
multiplied times demand to obtain the
quantity of potential waste. This
demand, in turn, depends on aggregate
household income which, in long-term
economic forecasting, is assumed to
depend upon a set of factors (the
population, the proportion of population
that seeks work, the politically-acceptable
unemployment rate, and productivity) which
are f recast exogenously. A path, then, is
illustrated by ths following:
several exogenous forecasts
household intone
I	part of
glass >soft	> demand for—>glaEs
drinks soft drinks in waste
streani
What began as a system of simultaneous
equations has become reduced to a set of
analyses which are not at all formidable
oenputat iona 11 y.
The simplicity of the structure of an
input-output table is a result oi the pro-
cedures used in tabulating and presenting
data. Data are grouped into "industries."
An industry may include several steps,
e.g., makers of automotive parts are part
of the industry "motor vehicles." Th& data
reflect the tendency towards vertical
Integration in plants, and the large effi-
ciency of combining various steps as In
steel mills. Lt is no surprise that the
steps, in a model of such large units, are
short.
It should be emphasised that this 'con-
version of part of the matrix into a small
number of short flows holda because an
input-output model has certain empirical
properties. These are reflected in a lack
of significant Interdepondence in the ma-
trix. It does not hold for those parts of
an economy where there is substantial
interdependence {e.g., transportation,
energy), and it does not account for all of
the output of materials. (An attempt to
account for all1flows of a material
generally results in ari extremely large
number of paths.)
* Stedaan B. Noble, "Forecasting the Composition and Weight of Household Solid Hastes
Dsing Input-Output Techniques," 2 Vols., 1975. OTIS, P8257 499/4GIj and PB257
S00/9G1. Also see Martin Hershkowlti and Stedaan B. Noble, "Finding the Inverse and
Connections of a Type of Large Sparse Matrix," Nava; Research Logistics Quarterly,
Idol. 12 (1965), pp. 119-132.
281
-------
b. Flow Method.
The Input-output table rep-resents
transactions between sectors of Che
economy, and can be used to identify
material flows Into final ub-j. An
alternative awthod Is to identify these
flows directly. Moot of th-s infor-
mation is readily available from such
sources as
(1)	the Census of Manufactures,
(2)	other government data such as
by Bureau of Mines,
(3)	data produced by trade
associations and fade press.
Why would an analyst go directly to
.Jata when an input-output table is
available? Mainly because the table
may f>ot satisfy his needs. For exanpj.e,
the cauDositlon of solid waste depends
upon the weight of the inpuu>, where an
input-output table is given in valuf.-o.
Largely for this reason, each of the
studies being reviewed developed
historical information on waste by going
to.published sources. Because they did
this, they provide information on tht
difficulty of a direct method.
Our procedure is the following.
Osing the 213 flows already defined, we.
have determined whether the necissary
information is or Is not available
directly in print. Table 1 sucnarizes
this information.
The nonferroua metals poae special
proble&B, so putting theo aside, we can
focus on steel and nonaetaln. Foe
these, alnoet half of the paths are
directly in print. Being "in print"
Involves two considerations! (a) that
the material per end use be published
(e.g., plywood purchased by the house
trailer Industry), and (b) that the pro-
portion selling to final demand can be
known. Since quantities on alloctlon to
final demand are not qenerally published
(except In input-output tables), a path
Is listed in Table 1 as *in print" when
sales of Its final industry are 85* or
aore to households. If the information
in print is incomplete the input-output
furnishes an allocation.
* "Textile Organon," Nov. Issue each year.
" Stedaan B. Noble, op. elt., vol. X, Table
In a further number of paths, the
published data can be used because the
categories are adequate or even mote than
adequate for the analysis of solid waste.
Textiles provides an example. Here, the
input-output model distinguishes knitted
goods from woven apparel, a distinction of
considerable importance for makers of tex-
tile machinery. For the study of soJid
waste, what is wanted \s the composition,
by fabric type, of types of cpparel, house-
hold textiles, etc. This information is
available directly in print.*
The direct availability of data in
print gives the analyst considerable flexi-
bility in constructing his flews. We have
already seen one example where this
mattered: the direensionality of the data.
For solid waste, the appropriate dimension
is weight. The analyst could obtain this
directly, or use an available dimension
(e.g., board fee*) that is more readily
converted than are values. If each user
paid the same price per pound, values would
be acceptable. Investigations reveal wide
divergence in the average price per round
paid by different users.**
There are many other examples where
direct use of data affects the way flews
are constructed. For example, the largest
use of rubber is for tires. An input-
output model allocates soeie tires to new
cars, sane to households, and sooe to com-
mercial establishments for replacement.
The analyst can afford to ignore this
d;-tinction, using the unit all "tires for
passenger cars *nd light trucks," and
giving special attention to retreading of
tires since this represents a diversion
frcn the waste stream- The distinction
between users made in the input-output
table is not necessary for analysis of
solid waste, and the "retreading" path does
not appear in the input-output table.
"tie most important reason to go
directly to the data, however, is the abil-
ity to use more detailed information than
la available in the input-output table.
For example, the input-output table coo-
bines lawn mowers with farm machinery
Whereas the Census of Manufactures iden-
tifies the metal purchased by *ach of
these. The Flow Method can reaoily
4.19, p. lilt Vol. II, pp. 1-48 ff.
282
-------
TABLE 1.
SIGNIFICANT
INFORMATION THAT
IS OR IS NOT
AVATLABL2 IN PRINT.




Involves no



Categories
aggregation

Number of
Paths not
in print
error in

significant
directly
are not
input-output

paths
In print
acceptable
aodel
Lumber
5
1
0
0/0
Paper
14
ii
7
0/7
Building paper
7
6
6
5/6
Textiles
22
10
0
0/0
Plastics
35
21
15
6/15
Rubber
4
4
0
0/0
Leather
2
0
-
0/0
Glass
5
2
1
0/1
Steel
_14
	6
3
_2£'
SUBTOTAL
loe
61
32
14/32
Copper
20
15
13
3/13
Lead
20
20
18
3/18
Zinc
35
35
•>1
1/3
Aluaimra
13
10
8
0/R
Other netals
12
11
11
Uii
TOTAL
213
152
U6
22/116
accomodate the level of detail that is
required to trace th actual physical
flows. 'Much of this detail Is hidden
(aggregated away) in the actual con-
struction of an input-output table, and
the structure of an input-output aodel
does not lend itself to putting the
detail back In. The significance of
this point for two flows is illustrated
in Figure 1.
In the first case, data at the
4-diglt level in the Census of Manufac-
tures shows that • note restricted path
Is called for, and that the largest path
generated by the input-output model is
Inaccurate. In the second case, the
most detailed Information in the Census
of Manufactures delimits the flows, but
not sufficiently. Going to technological
descriptions of the uses of copper*, the
special properties' of copper are used in
furniture hardware, marine hardware and
sane specialized builder's hardware.
Automotive hardware is not owntioned, indi-
cating that the largest path from hardware
In the input-output table which is t.ie one
that begins 'copper hardware*, is not valid
Aaerican Society for Metals, Metals Handbook, 9th ed.. Vol. 2, pp. 395-439.
28}
-------
?ath in 'nput-output model:
zinc.
zinc-	>electric lighting	-^household
and wiring equipment	lighting
Path according to data In Census of Manufactures:
none*	current-carrying wire devices
.0603'
noncurrent-carrying wire.
.112OS-
devices (e.g., fuse boxes)
residential lighting fixtures,
.ccnraercial lighting fixtures
-^construction
-^households
Path In input-output model:
copper.
copper.
.^cutlery, hand tools,
and har const ruction
motor vehicles
ships and boats
Figure 1. Paths in the Input-output aodel and in	reality.
* Allocations using aoet detailed data In Census	of Manufactures. Cnlts lbs per $.
** Allocations using technological information on	uses of copper. "None" neans not men-
tioned In source.
for copper. Since no data exists at the
appropriate level of detail, allocations
of copper must be est luted.
Fiji many nateiials, trade associa-
tions provide data that avoid preratlng .
errors. They do this by techniques
parable to those used in Figure 1. Tar
some extent, they tabulate data for indivi-
dual plants to avoid errors. More
generally, they ask their respondents to
answer different questions froro those
284
-------
aslced by the Census of Manufactures
which underlies the input-output toble.
They do not ask Who did you sell to? but
rather What was the material ultimately
embodied in? The respondents, being
familiar with the technology of use and
it3 association with the specific
materials being produced, can answer or
li le£St make educated guesses.
How pervasive are aggregation
errors? The fourth columi of Table 1
enumerates the proportion of paths that
were accurate out of those checked. The
complement therefore indicates the inci-
dence of serious aggregation errors- in
the input-output table. A largn propor-
tion of the paths are lnaceuriie for
the nonferroapparel
chemical3
in fact, lead is used foe tetraethyl
"lead fx gasoline and in pigments for
paint used in construction. It is not
used significantly in the dyes used in
fabrics. The long flow of zinc alacst
always involve s&ies of zinc to steel,
resulting in paths such as
tine > steel—ymetal	> motor
stampings vehicles
In fact, zinc Is used to galvanize (coat
netal, especially wire) in the steel
indvstry. It is not used for other uses
or steel. A matrix connects each input
to each output, being unable to
recognize the selective uses of these
materials. It turns ojt that the aggre-
gation error resulting .from the use of
only an Input-output table and model to
trace flows of 'jaste are both pervasive
ar.d large.
It is important V> emphasize the
formal similarity between the paths
constructed freehand and those generated
from an input-output model, because this
shows that the analysis can be broken into
a set of distinct stepj:
(a)	create coefficients, step-by-step,
(b)	study whether the coefficients change
over time,
(c)	study the demand by households for
the end product,
(d)	separately forecast total household
income.
The difference between ttv two approches is
their use of categories. In an input-
output table, the categories are defined in
advance. In the flow analysis, the cate-
gories emerge during the analysis.
3. THE INPUT-OUTPUT AND FLOW METHODS
00MPARB3
The preceedlng section has dlready
identified some of the relative strengths
and weaknesses of the two kinds of model in
use for tracing the physical relations be-
tween sectors of an economy. In this
section, we _nnarize sane of this
discussion arid appraise the contribution of
each approach. The section begins with a
short theoretical discussion of the nature
of model3.
a. Reflections on Models.
The act of modeling Is the assertion
of a correspondence between features of
. ths economy and certain sets ol formal
symbols. If in addition we poetulat*
characteristics of the symbols and rela-
tionships between them, then uhe act of
modeling asserts that thc3e characteristics
and relationships hold "approximately* be-
tw**n the phenomena in the economy which the
s> ~x>ls ere supposed to represent.
For example, the rows and columna of
an input-output matrix represent sectors of
the econcny, the coefficients of the matrix
represent the r«tlo of one sector's pur-
chases from another to ltd total purchases.
Using these coefficients implies that what
c«>e sector sells another is passed to the
The data foe metals, published by trade associations and Bureau of Mines, are not pub-
lished in adequate detail for this study. Therefore, Table 1 is not able to treat
-his as "in print" and is based upon other sources. The available data do demonstrate
that these data collectors could provide indirect data at -.dequate detail If thty
wished to do so.
285
-------
second sector's customers In proportion
to the values of the second sector's
sales.
Thii therefore assumes that the set
of inputs of _ny industry is strict
complements, vl.wed as flow*, It is
assumed thfti each input is us«d in the
same proportion whatever the pattern of
(tales of the Industry. The example of
the uses of lead to lnduJtrlal chealcalx
to clothing shows how wrong this
interpretation can be.
The correspondance between rvallty
and symbols is called interpretation,
and the issue Is whether sconoolc Models
are interprotable, and whether the
interpretation i? part of the analysis.
Interpretation of physical flows depends
upon knowledge of technology.
Interpretation becomes particularly
important in forecasting future events.
The use of computer models in fore-
casting tmpllritely assu<*es a course of
future developatnt of the economy.
These assumptions nay or say not be In
accordance with the best technological
estimate that can be made at present.
Only if these implicit assumptions are
Bade explicit, and checked for their
verisimilitude. I.e., only if the opera-
tion at the model is interpreted, can.
the prediction carry sufficient convic-
tion to provide a basis for policy
doveloputnt.
b. Dm of the Flow Method.
The strength of the flow method is
that it ties, directly in the best
available technological Information. In
to far as the netted cons lets of tracing
the physical flow of aateriala via pro-
ducts and on to waste, it is already in
physical terns. I.e., interpreted.
Purthormoce, it can take' account of- the'
aost current information in the
industry.
Ma haw already seen that a flow
analysis may Introduce new categories
.such as "retreaded tires* or it may
merely replace data., e.g., copper by
vqy of har
-------
What t»ie Input-output oodel provi-
des, is a inherent franscwork for GNP
calculations and distribution of income
over industries. The complementarity
iissustption however nay cause large
errors in this distribution of incone
over industries. The flaw method pro-
vides. trie technical Insight to identify
such errors, but present aysteas have
not developed nethods to correct the
results.
4. THE WO STUDIES OF SOLID WASTE
a. The Franklin Study.
Let js not* turn to the two studies
being reviewed. One of these, the study
by Franklin Associates*, is now easily
describes). It uses the flow not hod,
uning data directly in print, rather
¦uch ns described, no use was made of
data fron an input-output table.
The Franklin Study had-nany einor
errors, e.g., it did not always one the
beat available data sou-ces. More
relevant, tor out purposes., is its pre-
diction method. It obtained data frcn,
or estimated, the.material per end use
and then extrapolated that. It slsoat
never broke this down into separata
steps. The analysis was veekirc-d by not
dolnq ttit*. To Illustrate, for household
appLianctja, data were Obtained on the
total naterials conposition of a nunber
of specific prototype products, e.g., tha
total copper, rubber, etc.. In e speci-
fic dishwasher. The assumption was made
that this caBpooitirn will be constant
over tin*. However, data fro* tha
Census of Manufactures shows that the
purchase of aluainua in that Industry
has been growing relative to steel. . if
coefficients had been identified, and
the aSBuaotion of Constance had bean
checked, a different forecast of
material use would have beer-seen to be
warranted.
Further, tho Franklin Study did not
review studies of coruuner expenditure and
use these as part of their projections.
Their projections are attached to tinse
series of the use of materials in end
products, and are projections fo. the
future use of naterials in end products for
which no rationale is provided.
b. The HUT Study.
is
The approach of the IJUT Study
almost completely the reverse, of the
Franklin Study in its heavy reliance upon
other st.iiiea. The study uses the flow
method to develop estimates of potertlal
vaote, drawing 1- part on precursors o£ the
Franklin StuJy a..d in part On the inpuf4-
output model. The use of the input-output
model to generate paths-without checking
the reality of the flow led to errors. For
instance, I1UY estimates that municipal
waste has twice as nuch aluninoa and over
four tlacs as nuch copper, lead and zinc as
Is claimed by Franklin. Since nonlertouD
metals in aunicipal waste are a major
incentive for resource recovery, the dif-
ference has important implications.
A coo pari eon with Bureau of Mines,-
both its data and its anecdotal descrip-
tions of the uses of these ratals, does riot
substantiate IRfcT's clains. IRtT provided
flow diagram of this analysis, and these
include many paths that are generated by an
input-output model but are in fact spurious
(e.g., zinc to household lighting, copper
to motor vehicle*). Apparently, reliance
on the input-outjut model to generate paths
led to tha error.
For its projections of future waste
stream, the IRA? stuiy uses an Input-
output model. He have seen that this model
represents a portion of the National.
Accounts. If GNP is specified In commodity
detail, then the Input-output model solves
for the production levels of all industries
in the econcny, taking account of their
Franklin Associates, Ltd., Post-Consumer Solid Mast* and Resource Recovery Baseline,
Working Papers, Nay 16, 1979.
International Research and Technology Corporation, Forecasts of the Quantity and
Composition of Solid Waste, U.S., environmental Protection Agency, 1979. Also se*
International Research and Technology Corporation, Impacts of Economic Growth on tho
Quantity ind Composition of Solid Waste, January, 1980.
287
-------
sales co on* another.
Several ptograms have carried the
analyst farther by closing the model,
that is to Also provide equations by
which the eleoents of GNP can be
forecast. One such effort, producing
forecasts about ten y»ars ir^o the
future, and upJated every several years,
is the Econdmie Growth Model developed
by the Bureau of Labor Statistics.
Another systea is INTOPUM. developed and
maintained by Alaon*. BJTORUM is a ccn-
puterlied systen which is kept up to
date by continued statistical analysis
o£ time aerioa a* updates are published.
It is very large: aproximately 2,000
equations generate over IS,000 nuabers
for each projected year.
One characteristic ot iMroROH is
that almoot any number in it can b»
"overridden," that is,-revised by the
user. It is designed for the com-
putation of Menarioe. These overrides
are also used for updating th»* nodel.
*9 new information becunes available,
Alston con paxes it with his projections
and tests whether his systea accurately
predicted t^e new nuobers.. When it did
not, new r.lationships can be tried
which provide a better fit, and their
consequences can be studied. In addi-
tion to forecasting eleaentsof GNP,
forecasts are also nade of each Input
coefficient based upon past trends, and
oorrected for trie ability of the coef-
ficients in a row to predict production
data in eocf) Annual Survey of
Hanufacturei. Since there is always a
lag in collecting and producing data—a
lag of nany years for the best data such
aa the Census of Manufactures—INFORUM
la never entirely up to date.
The use.of a computer Model devel-
oped by one analyst and used by another
analyst, and for a different purpose,
raises the question of "portability."
The IUT study is really involved in
double portability: HUT used a version
of XNTORUM that had been transferred to
another aodel—SEAS—where it had been
permitted to get out of date. This ver-
sion retained the override features of
INFbRUM, but did not have the continual
updato of the time series that drives the
aodel to bo updated.
The real problea of portability is
salting a fit between technological infor-
mation and the use of the input-output
model, including the processes of "closing"
the aodel. The difficulties are three-
fold:
(1)	input-output modal8 use a fixed scheaa
for partitioning the econony, with its con-
ooranitant assumption of caaplenentarity.
Even INFORUM, vith ail its tlexiM 1 ity, has
no built-in procedure for partitioning
Industries. It has no method to assist in
discovering aggregation errors, or for
making adjustments, if they are discovered.
(2)	In general, the updating of the cloned
nodel3 is dor.e by statistical analysis of
tine-series, for instance, Alaon updates
INFOR'JM yearly using the newest nunber# of
oonsuaer expenditure. He found that expen-
diture on tires was risintj faater than his
equations suggested. Accepting this as a .
permanent change in the situation, he
corrected his forecasts upward.** The data
used'were expenditure of households without
regard for the nunber or tires being
purchased. The reason for the rise,
however, was the increasing acceptance of
radial tires which are oore expensive per
tire.. Radial tires last cnich longer,
however, so that there will t-e a period of
increasing acceptance when aggregate
expenditure increases. Once ownership !•
saturated, demand will fall sharply. This
illustrates the difficulty. A time series
of expenditure nay contain the present'
effect of current technological changes,
but seldom contains the information to
interpret it.
O) The nodels lock adequate "interpre-
tation", that is, they lack a description
in technological teras o! the nuabers and
relations embodied In the aodel. In the
exaaple lust mentioned, IR*T was not avara
that the onsuaption ot' tires had been per-
manently raised. Ail is unfortunate
because IR*T ia a flra whose aaln expertise
* Copper Alaon, Jr. and others, 1935i Interindustry Forecasts of the American Economy,
1974.
** Alaon, op. clt., p. 34.
288
-------
is relating technological Knowledge to
economic data and analysis. Its ana-
lysts were familiar with the tech-
nological situation in the tire
industry. Had there been a warning,
they probably would have caught what was
happening,.
No conclude that the available in-
put-output mode13 are not portable in a
technological environment. They are
not progi/aaned to accept technological
intonation, nor is the information pro-
vided from which the user night >m the
necessity of transforming technical
knowledge into the kind of input the
model? will accept. HUT was a vietln
of this lack of portability; the analy-
sis never really managed to incorporate
all the technical information developed
by the flow net hod.
5. CONCLUSION
The review of the two attempts to
predict future waste flows has involved
an extensive review of two sets of ana-
lytical tools) analyses of aaterial
flows and input-output nodeIs. It must
be concluded that it is not at present
possible to adopt.Input-output based
models to accomodate Jetalled technical
knowledge On a wide scale.
The reason for this difficulty is
~.hot input-outp-.t models, in otOer to
model the entire eooncny which is their
rationale, deal with technology schema-
tically and in fixed categories.
Technological analysis on the other hand
will uncover singularities and change.
In principle nest of these can be accom-
modated into an input-output schema by
disaggregation. In practice, an input-
output model can be manually adjusted to
accept a few irregularities. But in
present nodols such manual adjustments
do not oeea practicable when a large
number of Interrelated changes are to b«
made.
The problem Is a general one. Two
professions have each developed their
analytical tools for thblr own vises,
from different points of view. Input-
output models contain the reality of
fiscal trar&actions, but physical tran-
sactions and technology are only foc-
isally embocUod in the model. Plow,
analysis will produce the reality ot
phyolcal'relations. Ex po.*t facto, for-
malism and tealicy canr.ot b« mashed.
ACKNOWLEDGEMENT
The work upon which this paper is based
was supported by the Municipal Environmental
Research Laboratory of the Rnvlronoantal
Protection Agency. The authors are par-
ticularly indebted to consent and support by
Oscar H. Albrecht.
2S9
-------
-------