DOE/MC/28162 - - 5008 uc-/07

Coolside Waste Management Research

Contract No: DE-AC21-91MC28162 Quarterly Technical Report

Report Period: January 1,1995 to March 31,1995

ASTER

Coolside Waste Management Research Technical Progress Report

Contract No. DE-AC21-91 MC28162 Report Period January I, 1995 to January 31,1995

Summary of Progress for the Period

Investigation continued into the similarities and differences between lab column and

field lysimeters and the potential thermodynamic considerations underlying those

differences. The field and column results generally supported the assumption that

initially, hydration of the metal oxides to the hydroxide form will result in an initial high

pH, followed at some time by calcite formation reducing the pH to around 8.5. Two of

the field lysimeters showed an unexpected initial pH dip. The two lysimeters exhibiting

this behavior were the least compacted. The third Coolside material lysimeter, the

most compacted, maintained high pH. Comparison of pH and mineral saturation

behavior of this field lysimeter with the only compacted laboratory column which gave

sufficient data revealed essentially identical behavior. The field and lab column fly ash

lysimeter results were also consistent. The thermodynamics indicate a large amount of

available carbonate-bicarbonate buffering present during the pH transient, but the

model results are suspect. The second difference between column and field lysimeters

is an indication of significantly greater periods of time during which ettringite formation

is favored in the field lysimeters. Thermodynamic data indicates that soluble aluminum

availability controls the formation of ettringite at high pH in the leachate system.

Thermodynamic Investigations of Mechanisms Controlling Leachate Chemistry

Investigation continued into the similarities and differences between lab column and

field lysimeters and the potential thermodynamic considerations underlying those

differences. The column results generally supported the assumption that initially,

hydration of the metal oxides to the hydroxide form will result in an initial high pH,

followed at some time by calcite formation reducing the pH to around 8.5. Two of the

field lysimeters showed an unexpected pH dip to around 8.5 followed by a rise to

around 12, at or above that expected by Ca(OH), equilibrium. The two lysimeters

exhibiting this behavior were the least compacted. The third Coolside material

lysimeter, the most compacted, maintained high pH. Comparison of pH and mineral

saturation behavior of this field lysimeter with the only compacted laboratory column

which gave sufficient data revealed essentially identical behavior. Comparison of field

and lab column fly ash lysimeters also gave no surprises.

The ionic strength of the solute appears to have a great deal of influence on the

pH behavior. A comparison of ionic strength over time for the four field lysimeters

shows that the lysimeters with anomalous pH behavior also have the highest initial

ionic strength (Figures 1 and 3). Figures 2 and 4 are comparative data for selected

column lysimeters. The three Coolside lysimeters approach the same ionic strength

and about the same pH over time but the two least compacted exhibit the pH lowering

during the high ionic strength period. The two possible reasons for the lack of

observed pH dip in the column lysimeters are a) the dip does not occur or b) the dip

occurred and corrected before leachate was initially extracted. Both scenarios could

be due to the smaller column size. Additional graphic illustration of the relationship of

pH and ionic strength (Figure 5) shows a transition in behavior between ionic strengths

of about 0.5 to I. The three data sets corresponding to the fly ash lysimeter, the most

compacted lysimeter and the group representing the two least compacted lysimeters

can easily be seen.

Ionic strength has a direct effect on the solubility of mineral species. The

saturation index log(lAP/kT) calculated by WATEQ compares the activity product of the

constituent ions (IAP) with that expected at equilibrium at a certain temperature (kT).

The activity of a constituent ion is its concentration (arrived at by mass balance

considerations iteratively starting with the observed concentration in the solute)

multiplied by an activity coefficient. The activity coefficient is 1 .O in dilute solutions. As

ionic strength increases, the coefficient initially rises above 1 .O, driving the equilibrium

condition' towards precipitation of the solid phase. As the ionic strength continues to

increase above about 0.5, the increased availability of ions in solution accelerates

formation of neutral aqueous complexes. The solid phase "sees" less ionic activity and

the equilibrium is driven towards dissolution of the solid phase. The effective ionic

strength decreases. As an example, the solubility of calcite (CaCO,) increases by a

factor of three as ionic strength increases from 0.5 to 2.0. Another effect of high ionic

strength is the reduction in activity of H,O because of polar bonding with species in

solution. That.portion of the water is no longer able to act as a solvent.

Ionic strength also affects alkalinity. WATEQ estimates available carbonate by

correcting measured alkalinity for non-carbonate alkalinity; primarily OH'. It is also

known that other species, notably silicate forms, introduce errors in measured

alkalinity. The WATEQ model predicts a significant carbonate-bicarbonate buffering

3

.

capability during the transition to lower pH (Figure 22). An additional illustration of the

inverse relationship between pC0, and pH is shown in Figure 23. This is by no means

certain. First, the measured alkalinity, and therefore calculated carbonate, may be

significantly in error at high ionic strength. Second, inspection of the calcite saturation

index reveals that the model also predicts that calcite remains supersaturated during

the transition, indicating precipitation except for brief periods initially (Figure 6). It is

unlikely that the carbonate is simultaneously extracted by calcite formation and readily

available for buffering, although carbonate may be available via equilibrium or non-

equilibrium dissolution. Additionally, the comparison of log(ionic strength) vs pH

(Figure 5), which shows a behavioral shift between ionic. strengths of 0.5 to 1 .O and

log(pC0,) vs pH (Figure 7), which exhibits striking linearity, would most easily be

explained by the direct dependence of carbonate on pH in the model calculations. A

plot of linear ionic strength vs pH ( Figure 8), two assumed independent entities, shows

a reasonable (imperfect) linear correlation. In both ionic strength vs pH plots, one can

easily distinguish the data points for the fly ash lysimeter, the compacted continuous

high pH lysimeter and a single group for the lysimeters which show the pH transient. It

must be concluded that the initial high ionic strength plays a major role in the initial

lowering of pH, perhaps in a manner similar to the observed high buffering capacity of

sea water, which has an ionic strength of approximately 1 .O. This capacity may be

enhanced by the diversity of ionic constituents available for the formation of neutral

aqueous species.

The major species readily available from the constituent waste are Nay CI, Ca

and SO4 (Figure 9). CI follows the form of Na closely, and is several thousand ppm

4

higher initially. Minor species are AI, OH, Mg and Si (Figure 21). CI does not readily

compiex with any species except the heavy metals (Pb, Cu, Zn for example).

Examination of the saturation indices for a sample of these shows a general trend for

. dissolution (Figure IO). Likewise, the saturation indices of common Na mineral forms

indicate undersaturation (Figure 1 1). Some Na-aluminosilicates show regions of

supersaturation but, as the equilibrium rates are low, no significant formation of these is

expected. It is apparent that the major ions Na and CI are leached directly into the

solute, a major portion of the observed high ionic strength. Major hydroxide forms also

show dissolution behavior (Figure 12), except occasionally at higher pH values, so that

precipitation of these are not causing a pH dip.

Thermodynamically, Ca and Mg behave similarly, with mineral forms consisting

of a solid solution with the substitution of Mg for Ca. The amount of-substitution will

initially depend upon the Ca/Mg ratio. Equilibrium diagrams show that the expected

equilibrium relation between Ca and Mg forms (Figures 13-16) are very comparable to

those seen in the laboratory lysimeters (Figures I 7 and 18). The fly ash lysimeters

have a lower Ca/Mg ratio and remain primarily within a dolomitic regime, while all

others show that the equilibrium favors calcite formation, with brief instance of slight

possibility of dolomite formation. The conditions under which dolomite forms are not

well understood. Attempts to precipitate Dolomite under laboratory conditions have not

generally been successful. It is anticipated that any dolomite seen via XRD will be due

to either the process addition of dolomitic limestone or a poorly formed solid solution

mixture of calcite and brucite,

Different forms of the same mineral have differing thermodynamic

characteristics. Often the least preferred thermodynamic form is the initial solid phase

formed, is more soluble and converts slowly to the favored form. Examples of this are

silica gel and quartz, aragonite and calcite, and anhydrite and gypsum (Figures 19 and

20). There are ranges of local environmental conditions which lead to more efficient

congruent dissolution of the least favored form and simultaneous precipitation of the

most favored. The primary SO4 mineral in arid Coolside waste is anhydrite. This will

spontaneously hydrate to gypsum with concomitant swell due to differing densities.

Comparison of saturation indices of gypsum and anhydrite illustrate this. Note that the

pH range is fairly narrow for the most efficient conversion, with anhydrite in dissolution

and gypsum in formation. Excess or insufficient hydration and the resulting pH (as

exhibited by the pH differences in the least and most compacted lysimeters) may

control the pace of this congruent dissolution/precipitation phenomena. When both are

thermodynamically likely to precipitate (log( IAP/kT) above 0) or dissolve (log( IAP/kT)

less than 0), equilibrium formation/dissoIution will still result in conversion, but at a

lower rate.

A clay-like deposit was reported during sample augering in the field lysimeters.

Inspection of kaolinite saturation indices indicate that clay mineral formation is possible

early in the lysimeter sample period (Figures 24,25).

While in the proper pH range, ettringite formation will be controlled by the

availability of calcium, sulfate and aluminum. The second major difference in

laboratory/field lysimeter behavior is the increased amount of time the thermodynamics

indicate favorability for ettringite formation in the field lysimeters, as opposed to the

laboratory columns (Figures 19, 20,24, 25). Indications of the availability of the

ettringite precursor forms, gypsum, calcite and diaspore, are shown in Figures 26 and

27, Gypsum and calcite have been discussed. Diaspore (AIOOH) is the basic oxide of

AI,O,formed by hydration. Further hydration in high pH solutions yield AI(OH)& the

major ionic carrier of available aluminum. In the lysimeter system, there is an

abundance of calcium, sulfate and OH-, so it is possible that ettringite formation is

actually controlled by the availability of soluble aluminum, with ettringite formation likely

at high pH under conditions of diaspore dissolution (Figures 19,20,24-27).

Many other mineral forms addressed by WATEQ show regions of

supersaturation but do not control leachate chemistry and are unlikely to be formed in

any but the most minor amounts. A sample of these (Figure 28) illustrate the typical

thermodynamic behaviors of these minerals. Albite is a sodium feldspar. Its saturation

index has an inverse relation with pH. Diopside is a Ca-Mg pyroxene which follows pH.

These are the most common behaviors seen. The mineral forms which must be

suspected of having control of the leachate chemistry are those whose saturation

indices remain not far from zero and do not show wild variability. This may sometimes

be deceiving, as illustrated by prehnite, Cafi12Si30,0(OH)2 The saturation index shows

these classic signs, but it has little to do with leachate chemistry and is unlikely to be

found. Its saturation index remains positive and shows little pH effect because of a

very low equilibrium rate, which makes it appear constantly supersaturated.

Effective Ionic Strength over time 2 I

L

91.5

0 0 20 40 60 80 100

samples lysimeterlevels

-1-3 +2-3 -3-3 -4-3

Figure 1

5 . q 2 , 5 2

0 10 20 30 40 50 60 collection

- Coolside 300, Static Saturated - Coolside 300, Static Saturated +CO2 + Coolside Pilot 1, Static WetlDry - Coolside Pilot 3, Static WetlDry - Coolside Pilot 1, Static Saturated - Coolside Pilot 1, Static WetlDry +CO

Figure 2

pH comparison 14

13

12

11

10

9

8

7

Ip

0 20 40 60 80 100

Figure 3

- Coolside 300 Static Saturated -I- Coolside Pilot 1 Static WetlDry -P- Coolside Pilot1 Static Saturated

+ Coolside 300 Static Saturated + CO: -cc- Coolside Pilot3 Static WetlDry -c Coolside Pilot 1 Static WetlDry +CO:!

Column Lysimeter pH l4 1

7 '1: : : : : : : : : : : 0 10 20 30 40 50 60

Effective ion strength vs pH 13 ,

0.01 0.1 1 effective ion strength

10

Figure 5

Calcite saturation indices 3

2

E iz - a 1 v cn 0 -

0

-1 0 20 40 60 80 100

samples

Figure 6

- - -

log pC02 vs pH

8

13

I I I 1 I I 1 I

12

11

Ip 10

9

-11 -10 -9 -8 -7 -6 -5 -4 -3 log pc02

Figure 7

Effective ion strength vs pH 13

11

Ip 10

9 It.

8 ' t

7 ! I 1 I 1 1 I

0 0.5 1 1.5 effective ion strength

2

Figure 8

Lysimeter I level 3 27000 -, (11.5 24000

21 000 18000

E 15000

12000 ::::I 3000 0

I n

10.5

10

9.5

1 1 -May-93 27-NOV-93 1 Wun-94 01Jan-95

- Ca -+ Na -SO4 *K - PH 1 Figure 9

10

0 E 2 -5 U CD 0 -

-1 0

-1 5

-20 07Jun-93 09-Dec-93 24-Mar-94

Chlorides Lysimeter 2 level 3

Ip

25Jul-94

Figure 10

. I_.-I

~~

-c Mirabilite -f- Natron - Thenardite + Themonatri +Analcime - pH

I O 14

13

12 5

0 11 g.

10

9

-1 0 8

-5

1 l-May-93 19-Aug-93 27-Nov-93 07-Mar-94 1Wun-94 234ep94 OlJan-95 Date

Lysimeter 2 Level 3

Figure 11

Lysimeter I level 3 10 4

c

12

10 Ip

-in J 8 2gJun-93 11-Nov-93 28-Feb-94 26-May-94 17-Nov-94

I+ Portlandite 4- Vocroite - pH - Brucite I Figure 12 13

A No. 1, Level 3

7 No. 1, Level 4

= No. 1, Level 5

Do I om'& [CaMg(C03)2] 1-

< Hunfite [CaMg3(C03)4] > \

-1 - -2 - Hydromagnesite> \,

\

\ -3 - Hydmrnagnesite \

[Mg4(C03)3(OH)2+3H20] '% 4

1 I 1 I I 1 1 I 1 I 1 - -12 -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1

pc02

~~~

Figure 13

5

4

3

2

1

0

-1

a 3

4

. A No. 2, Level 3

7 No. 2, Level 4

= No. 2, Level 5

Dolomite [CaMg(C03)2]

Brucite [Mg(OH)2] ' c Aragonite [ c a m s ) > -------. Hunfite [CaMg3(C03)4] >

\ . \

Hydromagnesite> . , \ Hydromagnesite \

[Mg4(C03)3(OH)23H20] '% I I - I 1 I 1 1 1 1 1

- I '

! -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 p c 0 2

Figure 14

Lysimeter 3 A No. 3, Level 2 .

\ No.3, Level 3

= No. 3, Level 4

No. 3, Level 5

\ Calcite [CaC03]

\ \w \ \ =% .. Dolomite [CaMg(C03)2]

\

< Hunfite [CaMg3(C03)4] > \ \ \ \ \ \

< Hydromagnesite> \, \ Hydromagnesite \

\

[Mg4(CO3)3(OH)29H20] 'A Magnes'h [ M > ~ Q \

-11 -10 -9 -8 -7 -6 5 -4 -3 -2 -1 pc02

I I 1 1 I 1 1 1 I 1 1 -

Figure 15

5

4

3

2

1

D

1

2

3

4

A No. 4, Level 1

A No. 4, Level 3

= No. 4, Level 4

Calcite [caco3]

[CaAMg(C03)2] A* A A

A

Aragonite [C~COS) > ---------------- < Hun& [CaMg3(CO3)4] > \

\ \ \ \ \

Hydmmagnesite> \, \ \ Hydromagnesite \

[Mg4(C03)3(OH)2'3H20] '-A Magnesite mimy - 2 -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1

I I I 1 1 I I 1 1 1 1 -

p c 0 2

Figure 16

A StaticSaturated 4 0 2 StaticWetlDry402 . Calcite [caco3] A A A

* ' % ' 4

9 4

Dolomite [CaMg(CO3)2] ' '

' \ \ Hydromagnesite> ',

\ Hydromagnesite \

[Mg4(C03)3(OH)2'3H20 '-A Magnesite [M;;c8~ ! -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1

1 1 1 1 1 1 1 I I 1 1 1 1 -

pc02

I I Static Saturated +C02 4 StaticWetlDly+CO2 Coolside 300

5

4

3

2

1

0

-1

-2

-3

4

Figure 17

5

4

3

2

1

0

-1

-2

-3

-4

T Israeli Sat 4 0 2

0 lsrael iWD402

A K y W D 4 0 2

+ KyW/D+CO2

Calcite [CaC03]

A T T

:+ Dolomite [CaMg(C03)2]

Huntite [CaMg3(C03)4] > \ \ \ \ I '

\ \ \I c Hydromagnesites \,

\ H yd m mag n esite [Mg4(C03)3(OH)2'3H20] '-A Magnesite th&ma,

'. -12 -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1

pc02

Figure 18

+Anhydrite -+ Gypsum -Aragonite - - i ~ Calcite -f- Ettringite - pH

3 I

14

t 13

12

11

10

9

8

I P

9

8 -3 1 ?-May-93 19-Aug-93 27-Nov-93 07-Mar-94 15Jun-94 23-Sep94 01Jan-95

Date Lysimeter I level 3

I Figure 19

I +Anhydrite + Gypsum -Aragonite 1 - 9 - Calcite + Ettringite - pH

-5 I t , 3 1 -Jam93 19-Aug-93

Lysimeter 3 Level 3

07-Mar-94 Date

23-Sep94

Figure 20 11

Lysimeter 1-3 minor species 50 11.5 I

40

30 -- .. E

Q Q

20

10

0 8.5 11-May-93 27-NoV-93 15Jun-94 01Jan-95

Figure 21

I -+ CaC03 aq- CaHC03 - pH -NaC03 I Figure 22

I Q

log(pCO2) and pH over time -3 -I=

-4

-5

-6

-7

-8

-9

-1 0 B

-1 1

time

I c 2-3 pC02 ea- 2-3 pH -3-3 pH +4-3 pCO2-4-3 pH

3-3 pC02

13

12

11 Ip

10

9

8

Figure 23

10

5 $ 0 3 cl) 0 -

-1 0

Lysimeter 1 level 3 12

11

10

3

9 11-May-93 27-NOV-93 15-Jun-94 01-Jan-95

1- Kaolinite - Ettringite - pH I Figure 24

IO

0

Lysimeter 2 level 3 112

-1 0 07Jun-93 28-Oct-93 27-Jan-94 24-Mar-94 16-May-94 08-DeG94

I - Ettnngite - Kaolinite - pH

Figure 25

2

1

-1

Lysimeter 3 level 3 13

12

11

I O

-2 9 1 I-May-93 27-NOV-93 1 5-Ju n-94 01-Jan-95

I - Gypsum - Calcite * Diaspore I -Quartz -pH

Figure 26

2

1

-1

Lysimeter 1 Level 3 12

i i

10

9 I I

Ip

t t -2 I I I ' 8 1 1 -May-93 27-NOV-93 15-Jun-94 - 01-Jan-95 - Gypsum - Calcite + Diaspore

4- Quartz - PH

Figure 27

Lysimeter 1 level 3 10

F - a 0 s U CT) 0 -

+ I1

10 I p

-1 0 1 I-May-93 27-NOV-93 15-Jun-94 01-Jan-95 - Prehnite - Diopside --Albite - pH I

Figure 28 a-I

Coolside Waste Management Research Technical Progress Report

Contract No: DE-AC21-91 MC28162 Report Period February 1,1995 to February 28 1995

Summary of Activity This report presents a summary of results from on going geotechnical testing. The long term swell data is found to be a function of prehydration, pre-aging and static loading with these factors decreasing swell. This data is congruent with considerations of the timing of mineralogic transformations, the most important of which are the formation of gypsum and ettringite. A total of 27 samples were collected during the month during five separate field excursions. Also initiated was the study of effect of Coolside

leachates on the permeability of natural liners.

Task 2 Materials Characterization Subtask 2.3 Geotechnical

Swe// Tesfs Monitoring of swell continued on the hydrated and non-hydrated FBC ash samples. Both types are still in the primary swell phase. Swell of the hydrated ash has been monitored for more than two years, while swell of the non-hydrated material has been monitored for 654 days. All samples were remolded near 95% of standard maximum dry density and optimum moisture content. The samples were remolded in CBR molds (6" diameter, 4.584" height) and swell was monitored under the following conditions:

approximately 12.5 Ib. surcharge, no age time approximately 2.5 Ib. surcharge; no age time approximately 12.5 Ib. surcharge, 7 days age time approximately 2.5 Ib. surcharge, 7 days age time

Time of aging represents the amount of time before the samples were placed in water.

Aged specimens were sealed at room temperature to prevent moisture loss before placement into the water tank.

There are three variables which are being investigated in the swell study: prehydration, surcharge and aging. Of these, prehydration appears to have by far the strongest impact. For example, the largest swell recorded in the hydrated samples was

-13% (Figure I ) for a non-aged low surcharge sample, while the equivalent non- rehydrated sample swelled -38% (Figure 2). The second strongest factor is age time. For the hydrated samples, the non-aged samples expanded by factors of 50-1 00%

more than the aged samples. Surcharge did have a noticeable impact on the non-aged samples; however, it did not have a strong impact on the aged samples. The maximum difference between the samples is almost by a full order of magnitude, -4% for the aged rehydrated, high surcharge samples to almost 40% for the non-rehydrated, non- aged and low surcharged samples.

Mineralogic Reacfions These observations are congruent with research relevant to hydration and mineralogic transformations which take place in the samples. Two phases of hydration and mineralogic transformation are recognized. The first reaction is the formation of ettringite,

(1) 6 Ca2+ + 2 AI(0H)i + 3 SOP+ 4 OH' + 26 H20 =* Ca~12(S04)3(0H),2-26H20

This reaction is thought to occur rapidly. Schwiete and Niel investigated the growth of ettringite crystals from solution and reported that ettringite formed as early as 30 seconds after hydration, and needle-like crystals could be identified after only 4 minutes of hydration.' Mikhail and Abo-El-Enein suggested, based on XRD analyses of early ettringite, that after 1 day of hydration, the intensities of ettringite peaks increased, while those of the reactants tricalcium aluminate and CaSO, decreased.2 However, at higher reaction temperatures (60°C), characteristic peaks for ettringite appear as early as 15 minutes after hydration. Muhammad, et al. found that a highly

defective, but clearly identifiable ettringite formed within 3 minutes of hydration, and a

23

0

-5

-10

-1 5 1

FBC Ash (hvdrated)

+-+9-4 "-'--.-.+--g ............-

10 100 Time (days)

* No age, 1230 Ib. surcharge ....$ .... No age, 250 Ib. surchqe

- 7 days age, 1230 Ib. surcharge -Q .... 7 days age, 230 Ib. surcharge

Figure 1. Swell versus Time, Hydrated FBC Ash.

0,001 0.01

* No age Time, 252 Ib. surcharge

- 7 Days Age Time, 2,50 Ib, Surcharge

Non-t rdrated a

0.1 1 10 Time (days)

100 lo00

+ No age Time 12.62 Ib. surch;rrge

+ 7 DAYS& Time, 12.37 Ib. surcharge

Figure 2. S w e l l versus Time, Non hydrated FBC Ash.

fully developed ettringite in as little as 80 minutes3 Our own research, as yet unreported, verified that ettringite forms rapidly; however, significant formation was found to still occur for a period of several days.

ability to cause expansion is a function of where it forms, with nucleation on mineral

surfaces causing expansion.

Ettringite can form either in the material voids or upon mineral surfaces. Its

The second reaction is the formation of gypsum from anhydrite,

(2) CaSO, + 2 H20 - CaSO4-2H,O

This transformation is accompanied by a substantial increase in the molar volume. Gypsum has a molar volume of 74.2 cm3/mole, anhydrite has a molar volume of 52.3 cm3/mole. Thus this reaction results in substantial molar volume increase (+AV) of

42%. The interpretation of the swell data in the light of these two equations is

straightforward. The prehydration and aging of samples before they were molded allowed ettringite to form and also allowed the hydration of anhydrite to proceed. Thus these samples had the lowest expansion, due only to the completion of the hydration of anhydrite to gypsum. The non-prehydrated, non-aged samples had the highest swell because ettringite and gypsum formation took place primarily in the molds for these ,samples. Surcharge has the effect of forcing ettringite formation to take place interstitially. Surcharge only affected samples where ettritigite formation took place in the molds, i.e. the non-aged non-prehdyrated specimens. Surcharge had no effect on the rehydrated aged samples (Figure 1) because the ettringite had been fully formed before the samples were molded.

Unconfined Compressive Sfrengfh Unconfined compressive strength tests were

performed on remolded Coolside specimens. The specimens were aged for approximately 730 days in sealed containers. No excess moisture was present during the aging process. Samples tested were remolded near: 1) one hundred percent of

standard and modified dry density; 2) ninety-five percent of standard, modified, and low energy dry density; and, 3) ninety percent of standard dry density. Unconfined compressive strengths of specimens sealed and aged at room temperature begin to increase after about 14 days age time as seen in Figures 3,4, and 5 with data obtained previously.

The data demonstrates two unusual features relative to the geotechnical characteristics of the Coolside materials. The first is the importance of precompaction

on the strength development. The lowest degree of compaction (Figure 3) achieved strengths of only -750 psi which can be compared to highest degree of compaction (Figure 5) where strengths of -2500 psi have been achieved. The second feature is the length of time which the compressive strengths of the materials are observed.

Significant strength gains are observed for samples in the second and third year of observation.

Permeabilify Permeability tests were performed on remolded Coolside specimens. The specimens were aged for approximately 730 days in sealed containers. Samples tested were remolded near: 1) one hundred percent of standard and modified dry density; 2) ninety-five present of standard, modified, and low energy dry density; and, 3) ninety percent of standard dry density. The permeability of the materials declined most rapidly during the first 60 days of the test program achieving values approaching I O " cm/sec for the 90% and 95% compacted samples (Figure 6,7) and exceeding I O "

in the samples compacted to 100% of optimum moisture (Figure 8). Percolation values which are slower than 1 0" can be considered practically impermeable for most purposes.

Unconfined compressive strength tests were performed on some specimens after permeability tests were completed. The tests were performed to determine the

. unconfined compressive strength of the Coolside material in a saturated condition. The aged Coolside material has high compressive strengths, even if saturated, after aging (Figure 9,10,11). The strength data was found to show a strong inverse relationship

-+ W 00 00

t w CD 4 I 0 €

0

A

8

Unconfined Compressive Strength (psi)

E Q

UNCONFINED COMPRESSIVE STRENGTH 95% Compaction

2500

2000

1500

500

0 1 10 100 _ _ -

Aging Time (days) * Sample 1 Standard + Sample 3 Standard -c 388 Modified -.y..” 397 Low Energy

Figure 4. Unconfined Compressive Strength versus Time, 95% Compaction.

I c

d

I-L

8

Unconfined Compressive Strength (psi)

\

--\. --\

L.

0.04

0.03

0.02

0.01

0

PERMEABILITY 90% Compaction

0 100

* 1040 Standard

200 300 400 500 600 700 800 Aging time (days)

+ 388 Modified -c 397 Low Energy Figure 6. Permeability versus Time, 90% Compaction.

E t e

0.07

0.06

0.05

0.04

0.03

0.02

0.01

0 0

PERMEABI LlTY 95% Compaction

200 400 600 800 lo00 120( Aging time (days)

* Sample 1 Standard -c 397 Low energy

+ Sample 3 Standard + 388 Modified

Figure 7. Permeability versus Time, 95% Compaction.

3%

0.015

0.01

0.005

0 0

PERMEABILITY 100% Compaction

100 200 300 400 500’ 600 700 800 Aging time (days)

* 1040 Standard + 397 Low Energy -c 388 Modified Figure 8. Permeability versus Time, 100% Compaction

UNCONFINED STRENGTH AFETR.PERMEABlLlTY 90% Compaction

2000

1000

1

* 1040 Standard

10 Aging time (days)

+ 300 Modified

100 1ooC

- 397 Low Energy F i g u r e 9. Unconfined Compressive Strength, after Permeability Test, 90% Compaction.

- 8

0

Unconfined Compressive Strength (psi)

Y

-.-.. 8 0

\

UNCONFINED STRENGTH AFTER PERMEABILIT

2000

lo00

0 L

100% Compaction

I I

1 7 14 28 56 200 365 730

t

X

1 7 14 28 56 200 365 730

* 1040 Standard Aging Time (days)

+ 397 Low Energy -c 388 Modified Figure 11. Unconfined Compressive Strength versus Time, after Permeability Test, 100% compaction.

with permeability, with decreasing permeability correlating with increasing strength. Presumably this is a function of cementitious and hydration reaction products filling and blocking pores.

Task 3 Field Lysimetry

Subtask 3.2 Field Lysimeter Monitoring

The field lysimeters were sampled on the 2nd, 9th, 15th, 20th and 23rd of February. A total of 27 samples were taken, primarily from the leachate transport tubes located at the 10 foot depth. Temperatures in the bottom of the lysimters were at their annual low of 6 to 8" C, while recorded surface soil temperatures were in the 1 to 3" C range.

Task 4 Laboratory Leaching Studies

Subtask 4.1 Clay Liner Testing Natural clay liner tests have been initiated with three clays and two FGD by-products. The soils being used are naturally occurring clays which meet most states' criteria for a

landfill liner; that is, they have permeabilities less than 1 O 7 cm/sec when compacted. The FGD by-products are a FBC material from an electrical generating station in Pennsylvania and Coolside material from CONSOL's pilot plant, run #2. The tests are being performed by compacting the soils and FGD by-products in 3-inch diameter by 12-inch length PVC cylinders near 95% of standard maximum density and optimum moisture content. Soil # I is a residual clay collected in Fayette County, Kentucky. Soil #2 is a glacial till collected at the Franklin County Landfill just south of Columbus, Ohio. Soil #3 is a residual clay collected from a landfill in Kentucky. Three tests consist of water moving through the compacted 12 inch soil specimens. Six tests have water leaching through six inches of compacted FBC and Coolside material and six inches of compacted clay. The tests are configured so that each clay will be subjected to leachate from each FGD material (Figure 12). Permeabilities of the clays in these six tests will be determined at the end of the leaching period (approximately 1.5 years) and will be compared to values obtained before leaching. Hopefully, this will indicate any effects that water leaching through the FGD materials into compacted clay soils has on

Figure 12. Schematic of C l a y Liner Leaching Tests

I Pressure @ OylinderTop: - bS . -- E' x 16.44 f - 7.1 24 psi

Elusnt 144

7.1 24 + Hp"(.43333) psi

Figure 13. Schemat ic of w a t e r tubing i n C l a y Liner Leachate Tests.

the permeability of the compacted soils. Two tests have 12 inch FBC and Coolside pilot plant #2 specimens.

Water is supplied from a reservoir, mounted above the specimens, through 114 inch OD nylon tubing connected to the samples. Silica sand is used to fill the annular space in

the end cap between the tubing connection and the sample. Three way valves are located in-line between the reservoir and samples. The valves allow the water supply to be cutoff and the water remaining in the nylon tubing to be vented to atmospheric pressure. Falling head permeability tests can be performed by measuring the water elevation change with time. Permeability tests will be performed at various intervals to determine changes, if any, of the compacted materials. Details of the leachate testing setup are shown in Figures 12 and 13.

References 1. Schwiete H.E. and E. Niel, 1965, Journal of the American Ceramic Society, V48, p. 12-14.

2. Mikhail, R Sh., S.A.. Abo-El-Enein,S. Hanafi and R. Good, 1981, Cement and Concrete. Research, V11, p. 665.

3. Muhammad, M.N., P. Barnes, C.H. Feniman and D. Hauserman, et. al., 1993, A Time- Resolved Synchrotron Energy Dispersive Study of the Dynamic Aspects of the Synthesis of Ettringite During Minepacking, Cement and Concrete Research, V23, p. 267-272.

COOLSIDE WASTE MANAGEMENT RESEARCH Technical Progress Report

Contract No: DE-AC21-91MC28162 Report Period: March 1,1995 to March 31,1995

Statement of Project Objectives

Produce sufficient information on the physical and chemical nature of Coolside waste to design

and construct physically stable and environmentally safe landfills.

Summary of Progress for Period.

Preliminary data from the laboratory lysimeters packed with Coolside-waste materials were

obtained during the past reporting period. Details of the construction, packing, and test matrices

for the majority of these columns were described in the November 1994 monthly report.' In

addition to the eighteen columns described in that report, four additional columns have been

constructed and add& to the study since that time. These four columns (LC#33-36) were

included in an effort to examine the effect of prehydration on leachate chemistry in accordance

with suggestions made at the January 1995 Coolside contractor's meeting in Morgantown WV.

Relevant information for all the laboratory columns packed with Coolside wastes, including those

added in the hydration study, is summarized in Table 1. With the exception of columns LC #33-

36, additional information is available in a prior report.'

i I I

I

i

I i

I

I

t

i i I i

5. 0

1 I

1’0 PP2 bbsa X I 0 REPEAT I 2.138 40.4 24.0 49.0 11 Pp2 loose X 2.5 2.138 40.4 24.0 49.4 12 PP2 b s a X 5.8 2.138 40.4 24.0 49.0

ADDITION OF THE HYDRATION STUDY.

Changes in leachate chemistry that can be attributed to variations in the amount of prehydration

water added to the Coolside-waste samples prior to packing in a lysimeter is unknown.

Therefore, columns LC #33-#36 were added to the study in an effort to determine if leachate

chemistry is measurably affected by this parameter.

The four hydration-study columns were packed with waste material from Coolside Pilot-Plant

run #2 (PP2). Lysimeter construction and sample hydration and loading were conducted in a

manner similar to that previously described.' Equal amounts of dry PP2 sample (2.138 lb) were

added to each column. Rehydration water was added at 0,15,30, and 45 wt% (dry waste basis)

to LC #33, #34, #35, and #36, respectively. A loose packing density (49 lb/f+) was targeted for

each of the four columns. However, the physical (mudlike) consistency of the sample in LC #36

containing 45% prehydration water was such that loose compaction was not possible as the

density was initially greater than 49 lb/ft?. Therefore, proctor density (65.3 lb/ft?) was chosen

for this column instead. Achieving proctor density required only light tamping of the column.

On the other hand, Column #33 which contained no prehydration water had to be tamped quite

vigorously in order to obtain the targeted, loosecompaction density. It should be noted that even

though the packing densities are different and different levels of energy were required to attain

the packing densities shown in Table 1, all four columns contain equivalent amounts of dry PP2

3

sample and will receive the same amount of water each week.

Water addition to the hydration-study columns. Because columns #33-36 were several weeks into

the study, the initial addition of water to these columns lagged that of the other columns by

exactly six weeks. Furthermore, since this group of columns was added for the purpose of

examining of the effects of prehydration on chemistry and flow rates, a decision was made to

accelerate this particular phase of the study by doubling the amount of water added each week

over that added to the fixed feed columns, i.e., 93.0 vs 46.5 mL/week.

..

PRELIMINARY RESULTS

Due mostly to the manner in which water has been added to the laboratory columns, little

analytical data are available at this time. Unlike the preliminary column-leaching study in which

the columns were maintained under saturated conditions from the start, for this study, a

predetermined amount of water is being added to each column each week beginning with

week #l. What this’means is that for most of the columns, breakthrough of the water being

added did not occur for several weeks into the study. Even now, 15 weeks into the study

(9 weeks for columns #33-36 at double additions), several of the columns have yet to pass a

significant amount of leachate water. In addition to the problem of not having leachate samples

available for several weeks into the study, the sheer volume of data that has since been generated

has generated in a backlog of data awaiting processing and computer entry. This is in spite of

4

the fact that the analyses are being conducted within a reasonable period (1-2 days) for those

columns which are now flowing.

Column flows.

As described in the November, 1994 report,' water was added to the laboratory lysimeters in one

of h e manners. The majority of columns referred to asmed-feed columns received a fixed

amount of water weekly, 46.5 mL. Water was added to a second set of columns (Ws 3,4, 13,

14, and 15) in amounts which matched the rainfall measured at a weather station located near

the field lysimeter site. At the end of a twelve month period, the cumulative amount of water

added to these rain simulation columns should be equal to the amount of water added to the

fixed-feed columns. The third set of columns (LC #33-36; hydration stuby) receive exactly twice

as much water per week (93.0 mL) as do the fxed feed columns.

Flow rates through the fixed-feed columns. The rates of water flow through the fixed-feed

columns are shown in Figure 1. Flow through those columns that were loosely compacted are

shown separately (top) from that of the moderately compacted columns (bottom). As can be

seen, none of the columns exhibited flow until at least the second week of collection. Two of

the loosely compacted columns, #8 and #16 have yet to exhibit significant flow. However, the

majority of these loosely-compacted, fixed-feed columns have begun to flow freely thereby

providing sufficient sample for the full suite of analyses. Although all three of the moderately-

compacted columns have begun to flow, the water passing through these columns tended to lag

5

Flow through Fixed-Feed, Loosely Compacted Columns

0 2 4 6 8 10 12 14 16 Week

VLC1 pcLc2 LC7 ELC8 * L a LClO * LC11 e LC12 E LCi 6

Flow rates through e moderately compacted columns.

0 2 4 6 8 10 12 14 Week

w

&i

LC5

LC6

LC15

- Feed 3s

Figure 1. How rates through columns receiving a fixed amount of water (46.5 mL) each week. Top-ff ow through loosely compacted columns; Bottom-flow through moderately compacted Columns.

6

behind that of the loosely compacted columns both in terms of the initial breakthrough and the

average rate of flow following breakthrough. Due to the relatively high concentrations of ions

in the leachate waters that are present early in the study, the small sample size has generally not

been a problem in terms of generating sufficient sample for analysis. However, as ion

concentrations decrease with time, we will soon be compelled to combine the leachate from two

or more weeks collection to obtain adequate sample if the column flow do not improve.

Flow through the Coolside 3000 series samples is shown in the top of Figure 2. As shown in

Table 1, these columns represent both loose and proctor compaction as well as both fixed-feed

rates and rain simulation. An analogous plot for LC #7 and #8, packed with Coolside lo00

samples, is shown on the bottom of Figure 2. These latter columns were prepared as duplicates

in an effort to examine the reproducibility of the results from the laboratory lysimeters.

However, flow through these two columns has, to this time, been radically different. LC #7

exhibited breakthrough at week #11 which was followed in week #12 by a relatively large slug

of leachate water. In contrast, LC #8 has exhibited only minimal flows thus far. As additional

data becomes available, it will be interesting to follow the trends in ion concentrations for these

two columns to determine if such trends follow similar patterns despite obvious differences in

the interval of breakthrough and the overall volume of flow.

Rain simulation flow rates. Water flow through the five rain-simulation columns is shown in

Figure 3. The first breakthrough for this set occurred for column 3 during the third week of

collection followed by column 4 the following week. Both of these columns are packed with

7

140 120

5 80 60

$ 40 20 0

Columns 1-6; Flow rates Coolside 3000 series.

0 2 4 6 8 10 Week

12 14

v-

* LC 1

LC 2

LC 3

LC 4

LC 5

- LC 6

355

e

ffft

. .

200 I

5 100 U 1 E 50

0

Columns 743; Flow rates Coolside 1000 series.

Y

A A A A u--

0 2 4 6 8 10 12 14 16 Week

tL- LC 7 24- LC 8

Figure 2. Flow through the lysimeters packed with Coolside 3000 (top) and Coolside IO00 samples (bottom).

8 Y7 - . . .

Flow rates through rain-simulation columns. +

* * 3 s ’

49-

*

LC3

LC4

LC13

LC14

LC15

FEED 0 2 4 6 8 10 12 14 Week

Figure 3. Water flow through the rain-simulated columns.

Flow Rates; Columns 33-36 Pilot Plant run #2.

100

0 2 4 6 .8 i o Week

Figure 4. Water flow through the hydration-study columns.

9

-f

* -mi-

%

LC 33

LC 34

LC 35

LC 36

Feed

Coolside 3000 samples and both are tracking water feed reasonably well at this time. The

remaining three columns, which are packed with FGD wastes from Pilot Plant run #3 (PP3), have

exhibited no or minimal flow to date.

Flow rates in the hydration-study columns. Flows through the four columns (LC #33-36) that

were added for the hydration study are shown in Figure 4. Even though these columns receive

double the amount of water as the fixed-feed columns, breakthrough did not occur until week 7

for LC35 (30% prehydration water) and week 8 for LC34 (15% prehydration water). Columns

33 (0%) and 36 (45%) have yet to flow. Due to the relatively high rate of water feed, a standing

head of water can be observed above all four of these columns with the height of this water

column increasing steadily from LC 33 to LC 36.

In addition to columns 33-36, columns 25 and 26 were included in the original set of lysimeters

in an effort to examine the effects of hydration. These two columns are also packed with wastes

from PP2 (as are LC#33-36) but receive only half the amount of feed water each week

(46.5 mL) than LC #33-36. The FGD waste loaded to LC #25 was not prehydrated whereas

LC #26 was packed after addition of prehydration water to 40% by weight (dry basis). Leachate

flow through these two columns is shown in Figure 5. Although small quantities of water were

collected during week 2 for both columns, no water flow has been measwed since week 5. For

LC #26, the water collected is believed to originate from excess prehydration water. The 0.2 mL

collected from LC #25 during week 2 likely originated from the sand layer at the bottom of the

column which was wetted prior to packing. As with LC #36, there is a significant water covering

10

1.4 i.2

Y a i 5 0.8 5 e 0.6 .I E 0.4 0.2

0

Flow Rates; LC25-26 Pilot Plant run #2.

0 2 4 6 8 IO 12 14 I6 Week

Figure 5. Water flow through LC25 (no prehydration) and LC26 (40% prehydration).

over LC #26. LC #25 appears to be simply soaking up the added water at this time.

ANALYSES OF LEACHATE WATERS

DH, Thus far, pH values have typically ranged from about 11.6 to 12.8 for all columns packed

with Coolside wastes with the exception of those columns (LC 13-15) packed with materials from

Pilot Plant #3. These latter have ranged from about 9.5 to 11.5. However, due to the low flow

rates through these latter columns to date, it is uncertain if the pH values will remain relatively

low or increase as larger volumes of water pass through these columns. pH values are shown

in Figm 6 for columns packed with FGD wastes from Coolside 3000 (LCl-6), PP run #2

(LCI0-12), and PP run #3 (LC13-15) wastes.

11

Columns 1-6; pH Coolside 3000 Series

f

14-

a-

?s

LC 1

LC 2

LC 3

LC 4

Week

Columns 10-12; pH Pilot Plant run #2

11.8 I 0 2 4 6 8 1 0 1 2 1 4

Week

Columns 13-15; pH Pilot Plant run #3

11.0

g10.5

10.0

9.5

~~~

LC 15

0 2 4 6 8 10 12 14 16 Week

-.. . . .

Figure 6. pH values for leachate waters from columns packed with Coolside 3000 (top), PP run #2 (middle), and PP run #3 (bottom) wastes.

12

Selected results for columns packed with Coolside 3000 and Pilot Plant 2 wastes.

Similar to the results from the preliminary laboratory-leaching studies, sodium and potassium are

the major cations and chloride and sulphate are the major anions being removed from the

Coolside wastes at this early stage of the study. Calcium concentrations are relatively low at this

time (400 ppm) but appear to be increasing as the concentrations of Na and K decline. These

ions are plotted on the following pages for columns 1-6 (Coolside 3000) and for columns 10-12

(PP #2) for which the most complete sets of analytical data are currently available. . -

Sodium is the most prevalent cation for both sample sets ranging up to 3% of the leachate water

for the PP2 columns versus about 2% for the Coolside 3000 samples. The concentration of the

second and third most prevalent cations, K and Ca, are about twice as high in the leachate from

the Coolside 3000 columns than from the PP2 columns. Sulfate (SO4=) is by far the dominant

anion in the leachate from the PP2 columns. This anion ranges up to about 48,000 ppm which

is about four times as high as in the leachate from the Coolside 3000 columns. Chloride is the

most prevalent anion 'in the leachate waters from the Coolside 3000 columns ranging from about

10,000 to 38,000 ppm. This is in contrast to an approximate range of 400 to 1,400 ppm C1'

measured in the leachate from the PP2 columns.

Alkalinity and conductivity data are plotted for the same two sets of columns in Figure 11 and

Figure 12, respectively. The alkalinity of the leachate from the PP2 columns is substantially

13

Columns 1-6; Sodium Coolside 3000 series.

0 2 4 6 8 1 0 1 2 Week

2000 J I 0 2 4 6 8 1 0 1 2

Week

- f

14-

* = e

tEl-

LC 1

LC 2

LC 3

LC 4

LC 5

LC 6 -

Columns 1-6; Calcium Coolside 3000 series.

0 2 4 6 8 1 0 1 2 Week

- -

Figure 7. Concentrations of the major cahons (Na-top, K-center, &-bottom) in the leachate from columns 1-6 (Coolside 3000).

14 53

Columns 10-12; Sodium Pilot Plant run #2.

30000 i I

+ A-

-#

LC 10

LC ll

LC 12

rn O J I 0 2 4 6 8 1 0 1 2

Week

Columns 10-12; potassium Pilot Plant run #2.

A 3000

6 2000 6 1500

+so0 Y

0

1000 0

+ -a-

* LC 10

LC ll

LC 12 - _. mJ. I

0 2 4 6 8 1 0 1 2 Week

Columns 10-12; Calcium Pilot Plant run #2.

0 2 4 6 8 1 0 1 2 Week

f

14-

(P-

LC 10

LC 11

LC 12 -

Figure 8. Concentrations of the major cations (Na-top, K-center, Ca-bottom) in the leachate from columns 10-12 (Pilot Plant run #2).

15

Columns 1-6; Chlorides Coolside 3000 Series

40000 I 35000

p. '25000 6

0 5 20000

15000 10000

0 2 4 6 8 18 12 Week

f

LC 1 -&

Lc 2 * LC 3

LC 4

LC 5

- LC 6

S G

e

€3-

Columns 1-6; Sulphate Coolside 3000 Series

12000 1 1 1000

3 10000 v 9000 4 8000 6 7000

6000 5000

0 2 4 6 8 Week

10 12

~

f

7+

LC 1

LC 2

LC 3

LC 4

LC 5-

LC 6

* .g

4 s

El-

-

Figure 9. Concentrations of the major anions (Cl-top, SO4-bottom) in the leachate from columns 1-6 (Coolside 3000).

16 54

f ' LC 10

LC 11

LC 12

2%

a-

Columns 10-12; Sulphate Pilot Plant run #2.

-f

* *

LC 10

LC 11

- LC 12

Figure 10. Concentrations of the major anions (Cl-top, SO,-bottom) in the leachate from columns 10-12 (Pilot PIant run #2).

17

Column 1-6; AI kalinity Coolside 3000 Series

n 12000 1 1

0 ' I

0 2 4 6 8 10 12 Week

+ A-

#a-

LC I

LC 2

LC 3

LC 4

LC 5

LC 6

* -@-

4 3

Column 10-12; Alkalinity Pilot Plant run M.

A I8000

8 6000 8 4000

2000 0 2 4 6

Week 8 10

-f

7%

-m-

LC 10

LC 11

LC 12

Figure 11. Alkalinity of the leachate from columns packed with Coolside 3000 (top) and Pilot Plant #2 (bottom) wastes. .

18

Columns 1-6; Conductivity Coolside 3000 Series

110 1 1

100

'js 0 80 70

0" 60

.I E 90 >

c

50 40

0 2 4 6 8 Week

10 12

-f

7 3

LC 1

LC 2

LC 3

LC 4

LC 5

LC 6

E

6%

Ezl- -

Columns 10-1 2; Conductivity Pilot Plant run #2.

100 I

20 '

f

-8-

*

LC 10

LC I1

LC 12

0 2 4 6 Week

8 io

Figure 12. Conductivity of the leachate waters from columns packed with Coolside 3000 (top) and Pilot Plant #2 (bottom) wastes.

19

greater than for the waters from the Coolside 3000 samples. Both alkalinity and conductivity for

both data sets appears to be dropping rapidly as one would expect as highly soluble anions and

cations are rapidly depleted.

SUMMARY

In addition to the data presented in this report, additional analyses are being conducted including

analysis of 23 additional cations and dissolved solids (on selected samples). Suffice it to state

that none of the additional data collected thus far demonstrates a reason for alarm. These data

will be summarized in a later report. It is still too early in the study to attempt to understand the

full impact of compaction, method of water feed, or the addition of gaseous CO, to selected

columns.

The most significant finding presented in this report is that the majority of the 22 laboratory

lysimeters packed with Coolside waste have begun to flow. Thus, at least one significant hurdle

in this particular phase of the Coolside study appears to be no longer a reason for concern.

1. Coolside waste management research Technical Progress Report Contract No: DE-AC21- 91MC28162 Report Period: March 1, 1995 to March 31, 1995.

20

\

This report was prepared as an account of work sponsorai by an agency of the United States Government Neither the United States Government noi any agency themf,.nor any of their empioyets, makes any warranty, expnss or implied, or assumes any legal liability or responsibility for the acnuacy, completeness, or use- fulness of any information, apparatus, product, or pnxxss disclosed, or represents that its use would not infringe privately owned righu. Refenndherein.to any spe- cific commercial product, process, or service by trade name, trademark, manufac-, turcr, or otherwise does not necessarily constitute or imply its endorsement, raxm- mendzition, or favoring by the United States'Governmcnt or any agency thereof. The views and opinions of authors utprcsiezi'herein do not nectssarily state or reflect those of the United States Government or any agency thereof.

I

,

.