99-136

Embed Size (px)

Citation preview

  • 8/12/2019 99-136

    1/8

    Copyright 1999 by SME1

    SME Annual Meeting

    March 1-3, 1999, Denver, Colorado

    Preprint 99-1 36COLUMN STUDIES OF THE ADSORPTION OF GOLD CYANIDE COMPLEX BY GOLDSTRIKE

    ORE CARBONACEOUS MATERIAL

    P. A. Schmitz

    S. Duyvesteyn

    W. P. Johnson

    Univ. of Utah

    Salt Lake City, UT

    L. Enloe

    J. McMullen

    Barrick Gold Corp. Inc.Toronto, ON, Canada

    ABSTRACT

    The characteristics of gold leaching from carbonaceous

    matter by sodium cyanide, and sorption of gold cyanide to

    carbonaceous matter in autoclaved and non-autoclaved gold

    ore from Barrick Goldstrike Mines Inc. (BGMI), Nevada, was

    investigated. Much greater gold cyanide complex sorption

    was observed for the carbonaceous matter in high

    preg-robbing ores relative to the low preg-robbing ores. This

    observation corroborates previous reports that have attributed

    observed preg-robbing behavior to the carbon in the ore. Gold

    cyanide complex sorption was found to be reversible for

    carbonaceous matter from low preg-robbing ore, but was

    irreversible for carbonaceous matter from high preg-robbing

    ores. Autoclaving the ore appeared to enhance the sorptive

    behavior of the carbonaceous matter from both high and low

    preg-robbing ores.

    INTRODUCTION

    Barrick Goldstrike Mines, located within the Carlin trend in

    Nevada, ore can be subject to two forms of refractoriness. It

    has been estimated that more than 50% of the gold inGoldstrike ore is originally encapsulated within a sulfide

    phase (Chryssoulis et al., 1996). encapsulation of gold by

    sulfides requires oxidation to make the gold amenable to

    cyanide leaching. In the processing of the ore, oxidation of

    the sulfides is achieved through autoclaving, which releases

    elemental particles of gold that are sub-micron in size.

    Additionally, naturally occurring organic carbon is found in

    many of the gold deposits located within the Carlin trend in

    northeastern Nevada. The natural organic carbon has been

    implicated in a phenomenon known as preg-robbing

    (Hausen and Bucknam, 1985). Preg-robbing was firs

    described as the active adsorption of gold from pregnan

    cyanide solutions by Smith (196 8).

    Smith (1968) recognized that two types of carbonaceous ore

    occurred in the Carlin deposits: ores that would adsorb

    additional gold from the cyanidation procedure and ores that

    would not. Physical and chemical characteristics of the

    carbonaceous matter measured from X-ray diffractometry

    Raman spectroscopy, FTIR spectroscopy, and LECO

    oxygen/carbon analyses, show that the carbonaceous matter

    present in the ore is similar to that of commercial activated

    carbon (Nelson et al., 1986; Sibrell et al., 1990; Stenebrten

    and Johnson, 1998). This is significant because commercia

    activated carbon is used in the carbon-in-leach process used

    to recover the gold from these ores. It has been shown tha

    there is a rough correlation between the preg-robbing

    behavior of Goldstrike ore and the microcrystallite

    dimensions of the demineralized carbonaceous matter from

    non-autoclaved ore (Stenebrten and Johnson, 1998). This

    correlation suggests that the micro-crystallite dimension ofthe ore is related to its tendency to adsorb gold.

    The motivation for this study was to determine to what extent

    autoclaving during processing changes the physical and

    chemical characteristics of the carbonaceous matter and its

    Au(CN)2- sorbing behavior. Carbonaceous matter is the end

    result of a demineralization process used to remove the

    quartz, carbonate and some sulfide components of the ore

  • 8/12/2019 99-136

    2/8

    Copyright 1999 by SME2

    Two batch adsorption tests were performed on carbonaceous

    matter from autoclaved ores classified according to

    preg-robbing behavior, e.g. high preg-robbing (HPR) or low

    pregrobbing (LPR). The preg-robbing test is described in

    Stenebrten and Johnson (1998). The carbonaceous matter

    was contacted with a 0.5g.L NaCN solution containing 5ppm

    of gold. It was found, by measuring gold in solution after 24hrs, that while the HPR carbonaceous matter sorbed nearly all

    of the gold in solution, the solution exposed to the LPR

    carbonaceous matter resulted in an equilibrium gold

    concentration over 11 ppm. Presumably, gold that was either

    associated with phases destroyed during autoclaving (e.g.

    sulfides), or possibly originally associated with the

    carbonaceous matter, was leached by the cyanide ions in

    solution. The gold cyanide ion was then taken up strongly by

    the HPR carbonaceous matter but negligibly by the LPR

    carbonaceous matter.

    Alternatively, it must also be considered whether the CN -

    extractable gold content after autoclaving was equal for bothLPR and HPR ores, since an inequality in the amount of gold

    present would also result in apparent differences in sorption

    behavior. Although the LPR and HPR ores had similar opt

    values, variation in the sulfide content of the ore may have

    accounted for differences in gold amenable to leaching. This

    study was therefore designed to relate the gold adsorption

    behaviors of LPR and HPR carbonaceous matter, both

    autoclaved and nonautoclaved. This is a problematic task

    since the presence of any gold in the carbon interferes with a

    straightforward batch adsorption test.

    EXPERIMENTAL PROCEDURE

    Gold leaching with NaCN and elution with a 1: 1 mixture of

    NaCN and NaOH, followed by the adsorption of Au(CN)2-,

    and the subsequent desorption Au(CN)2- by NaCN, were

    studied sequentially in the carbonaceous matter of both

    autoclaved and non-autoclaved LPR and HPR ore. The initial

    task was performed to determine how much gold was

    available to cyanide leaching followed by sodium hydroxide

    elution from the carbonaceous matter of both autoclaved and

    non-autoclaved demineralized Goldstrike ore. This was

    performed with the recognition that gold associated with the

    carbonaceous matter may not be completely released into

    solution, but may instead be sorbed by the carbon, especiallyin the case of the HPR carbon. Exposure of the carbonaceous

    matter to a gold cyanide solution was then performed to study

    the tendency of the carbon to uptake Au(CN)2-. However, any

    gold not released during the initial leaching phase may have

    influenced the sorption of Au(CN)2- by the carbon. Lastly,

    NaCN and a mixture of NaCN and NaOH was used in an

    attempt to remove the sorbed gold from the carbonaceous

    matter.

    The ore was demineralized according to the procedure of

    Stenebrten and Johnson (1998). In all ores, HCI and HF

    were used to remove the carbonate and silicate minerals

    respectively.

    Approximately 80-90% of the sulfides were removed form theautoclaved ores during the oxidation process according to

    typical values determined by LECO analysis performed by the

    BGMI metallurgical services laboratory. An attempt to

    remove the sulfides from the nonautoclaved ore was made by

    heavy medium separation of the ore with sodium poly-

    tungstate. However, electron microprobe analysis of the

    demineralized non-autoclaved ores showed significant sulfide

    as inclusions in the carbonaceous matter. This demineralized

    fraction of carbon including variable concentrations of

    sulfides (depending on the success of heavy medium

    separation in nonautoclaved ores and the extent of oxidation

    in oxidized ores) is hereafter referred to as carbonaceous

    matter (CM). The carbon content of the non-autoclaved HPRCM used in this study was 52.5%, whereas the carbon content

    of the non-autoclaved LPR CM was 13.3%. The carbon

    content of the autoclaved CM was not known at the time of

    the study, but was estimated to be 75%, LECO carbon and

    sulfur analyses are currently being performed.

    Stock solutions of NaCN, Au(CN)2-, and a mixture of

    NaCN/NaOH were prepared as follows. The NaCN solution

    was prepared at a concentration of 0.5g/L, adjusted to a pH o

    10. 5 with 0. 1 M NaOH. The stock gold solution of 5ppm Au

    was prepared by adding gold standard to a solution of NaCN

    (0.5g/L) adjusted to a pH of 10. 5. A 1: 1 elute mixture of the

    NaCN solution and 0. 1 M NaOH was also prepared. The

    solutions were pulled through a mini-column prepared by

    packing 60 mg of CM in Teflon tubing (1/8"id). A glass woo

    plug was used to prevent the passage of the fine particulate

    of the CM. A 10 ml syringe was fastened to the top of the

    tubing to provide a fluid reservoir. Negative pressure to drive

    fluid flow was provided by a Digi-Staltic peristaltic pump

    placed on the effluent side of the column.

    In the initial leaching phase, a total of 30ml of the NaCN

    solution was introduced to the column (0.3ml/min) and

    collected sequentially in 5ml aliquots. Following the NaCN

    NaCN/NaOH eluent was passed through the column(0.3ml/min) and collected in 5ml aliquots. Both the

    autoclaved and non autoclaved HPR CM were exposed to 15

    ml of the NaCN/NaOH eluent, however both of the LPR CMs

    were contacted with only 10 ml of the NaCN/NaOH eluent. A

    third step in the leaching process involved contacting the CM

    with I pore volume of stock NaCN solution for 24hrs. The

    CM was eluted with additional NaCN solution to a total

    volume of 5ml. In the subsequent adsorption experiment, a

    total of 30ml of gold cyanide stock solution was introduced to

  • 8/12/2019 99-136

    3/8

    Copyright 1999 by SME3

    the column (0.3ml/min). The column effluent was collected in

    5ml aliquots. This was followed by the equilibration of 1 pore

    volume of gold cyanide stock solution with the CM for a 24hr

    period. The pore volume was washed with additional gold

    cyanide to a total volume of 5ml. The column was the rinsed

    with 5 ml of deionized water. The desorption stage of the

    experiment was performed by equilibrating the CM with onepore volume of the NaCN solution for 24hrs. The pore

    volume was rinsed with additional NaCN to a total volume of

    5ml. All solutions were analyzed for gold using ICP-AES

    against 1 ppm and 5ppm gold standard solutions. The

    solutions were compared at the 242.795nm wavelength using

    a Perkin-Elmer Plasma 4000 Emission Spec ICP. Gold

    leaching and desorption into NaCN was measured directly,

    while gold adsorption was measured by calculating the

    amount of gold lost from the initial 5ppm solution.

    EXPERIMENTAL RESULTS

    The results of the leaching and elution of gold from thevarious CMs are shown if Figures 1-4. The first 5 ml of

    leachate run through the autoclaved LPR CM showed almost

    10 ppm of gold (Figure 1). Subsequent volumes of NaCN

    eluent showed exponentially decreasing gold concentrations.

    A final aliquot of NaCN contacted with the CM for 24 hrs

    leached greater gold concentrations into solution, indicating

    kinetic limitations in leaching under these conditions. Figure

    2 shows almost an identical curve for the non-autoclaved

    LPR, the major difference is that the overall concentration of

    gold leached into solution is lower. This difference is

    presumably due to differing amounts of gold amenable to

    leaching between these samples. Removal of the sulfides in

    the second sample during heavy medium separation reduced

    the gold content of that sample. Furthermore, the remaining

    gold was largely non-amenable to leaching due to

    encapsulation in the sulfides. In contrast, autoclaving of the

    first sample released gold from sulfides, and it appears that

    the released gold remained with the carbonaceous matter

    during demineralization.

    Figure 3 shows there was a significant amount of gold

    leached from the autoclaved HPR CM during the first 5 ml of

    NaCN exposure. However in subsequent aliquots, the

    concentration of gold in solution was near zero. Although the

    NaCN/NaOH eluent did not show any measurable gold insolution, there was a significant amount of gold released into

    NaCN after 24 hrs contact time. In figure 4, the non-

    autoclaved HPR shows there was significant gold leaching

    into the first 5 ml aliquot of NaCN and no elution of gold by

    the NaCN/NaOH. However, unlike the autoclaved HPR, the

    non-autoclaved HPR showed a gradual decrease in gold

    concentration during the NaCN leach. Also, the leachate from

    the non-autoclaved HPR differed from the autoclaved HPR in

    that no gold was measured in the NaCN leachate after 24 hr

    contact time.

    The amount of gold leached and eluted from the CMs was

    calculated by multiplying the concentration of gold by the

    volume of each aliquot. In the autoclaved HPR, 55% of the

    total amount of gold leached and eluted from the CM wasreleased during the 24 hr contact with NaCN, indicating a

    significant kinetic limitation. Leaching of gold from the LPR

    CMs also showed kinetic limitation with 5% gold recovery

    during 24 hr leaching from the autoclaved LPR, and 15%

    gold recovery during 24 hr leaching from the non-autoclaved

    LPR. In both the autoclaved and non-autoclaved LPR CMs

    elution of gold by the NaCN/NaOH resulted in detectable, but

    insignificant amounts (less than 1%) relative to the tota

    amount of gold obtained from the CM.

    The total amount of gold leached from the CM (Table 1) can

    be related to the amount of gold that was originally present in

    the bulk ore prior to demineralization. The amount of goldleached and eluted from the CM was divided by the carbon

    content of the CM, and multiplied by the % carbon of the

    bulk ore to back-calculate the gold content in ounces per ton

    In the autoclaved LPR CM, 99.7 g of gold was leached. This

    accounts for 57.8% of the gold originally present in the ore as

    determined by fire assay (Table 1). In contrast, 2.2% of the

    gold, based on original opt values, was leached from the

    non-autoclaved LPR. However, when the total amount of gold

    leached by the nonautoclaved LPR CM in subsequent phases

    (adsorption and desorption) was included, the non-autoclaved

    LPR CM released 6.8% of the gold content of the ore (by fire

    assay). Approximately 5% of the gold of the original opt

    value was leached from both the autoclaved and

    non-autoclaved HPR CM.

    The results of the adsorption of gold on the LPR CM are

    shown in Figure 5. The autoclaved LPR CM samples showed

    sorption in the initial aliquot of gold-cyanide solution

    However, successive aliquots indicated varying amounts of

    sorption and desorption. The autoclaved LPR CM showed

    significant additional adsorption of gold onto the carbon after

    subsequent equilibration with gold cyanide solution for 24

    hrs. The total sorption of gold cyanide on the autoclaved LPR

    CM was 0.021 g Au/mg CM, which corresponds to a value

    of 0.028 g Au/mg C (after division by carbon content of theCM) (Table 2). The non-autoclaved LPR CM showed initial

    sorption of gold onto the CM, followed by significan

    leaching of gold into the gold cyanide solution. The net resul

    of the adsorption process on the non-autoclaved LPR CM

    was the leaching of 0.083 g Au/mg CM. Figure 6 shows

    gold cyanide sorption onto the HPR CM. Both autoclaved and

    non-autoclaved samples showed continued sorption of gold

    throughout the experiment with significant additiona

  • 8/12/2019 99-136

    4/8

    Copyright 1999 by SME4

    sorption during the subsequent 24 hr contact with gold

    cyanide solution. There was slightly less total sorption of gold

    cyanide per gram of carbon (1. 175 g Au/mg CM vs. 1.240

    g Au/mg CM) onto the autoclaved HPR CM relative to the

    non-autoclaved HPR CM (Table 2).

    The removal of the sorbed gold by desorption into NaCN(Table 3) showed that both the autoclaved and non-autoclaved

    LPR CM desorbed significant amounts of gold back into

    solution. The autoclaved LPR CM desorbed an amount of

    gold equivalent to what was adsorbed during the step where

    the CM was contacted with gold cyanide. In contrast, gold

    was leached from the non-autoclaved LPR during the

    adsorption stage of the experiment, and an additional 2.6

    g of gold (almost 30% of the total amount of gold leached

    from the nonautoclaved LPR throughout all stages of the

    experiment) was leached during the desorption stage. As

    expected, both the nonautoclaved and autoclaved HPR CMs

    showed limited desorption of sorbed gold; however, the

    non-autoclaved HPR showed slightly higher recovery (3.2%)relative to the autoclaved HPR (2.1%).

    ANALYSIS AND DISCUSSION

    Autoclaved LPR CM released significantly more gold than

    the other CMs in the initial NaCN leaching phase of the

    experiment. Since the low preg-robbing behavior of LPR ore

    has been theorized to be associated with the low sorbing

    behavior of the CM, and since autoclaving releases gold from

    the sulfides, high concentrations of gold released into solution

    would be expected from the autoclaved LPR. If it is assumed

    that autoclaving the HPR ore released similar amounts of gold

    from the sulfide phase as in the LPR, the extremely low

    concentrations of gold released by the HPR ore corroborates

    the theory that the sorbing behavior of the CM is responsible

    for preg-robbing; that is, the gold solubilized by NaCN was

    sorbed by the HPR CM and therefore not released into

    solution. The percentage of gold released during the initial

    leaching, relative to initial opt values determined by fire

    assay, was 17% and 10.4% for the non-autoclaved LPR and

    HPR, respectively (Table 1). Assuming that the HPR and LPR

    samples initially had similar NaCN extractable gold contents,

    the greater percentage of gold released from the LPR CM

    relative to the HPR CM supports the theory that pregrobbing

    is due to sorption of gold onto the CM upon initial contactwith NaCN.

    It can be assumed that the total amount of gold released from

    the non-autoclaved LPR represents gold initially associated

    with the carbon or left behind from the demineralization

    process. With this assumption, it can be presumed that the

    much greater amount of gold released into solution from the

    autoclaved LPR represents release of gold from sulfides

    during autoclaving. The difference in gold leached between

    the non-autoclaved and autoclaved LPR indicates over 50% of

    the gold in the ore is associated with the sulfides

    corroborating the work of Chryssoulis et al. (1996).

    Because there was additional gold present in the autoclaved

    CM due the release of encapsulated gold during the oxidationof the sulfides, it can be presumed that the autoclaved HPR

    CM was exposed to a higher Au(CN)2-concentration relative

    to the non-autoclaved HPR during the initial NaCN leaching

    Given this, it is strange that the non-autoclaved HPR CM

    actually released more gold (0.033 g Au/mg C) relative to

    the autoclaved HPR (0.020 g Au/mg CM). It is hypothesized

    that autoclaving increased the sorption of Au(CN)2- by HPR

    CM, possibly due to a change in functional group conten

    during oxidation. This has been tentatively corroborated by

    the appearance of additional peaks (1720 cm-1) in the FTIR

    spectra data of HPR CM after autoclaving (manuscript in

    preparation). This peak was not found in the FTIR spectra

    data of non-autoclaved CM (Stenebrten and Johnson, 1998)This may also explain why autoclaved HPR CM showed

    similar sorption of gold relative to non-autoclaved HPR CM

    during the adsorption stop despite the probable higher

    gold-cyanide ion concentrations experienced by the

    autoclaved HPR CM relative to nonautoclaved HPR CM

    Presuming a higher gold-cyanide loading on autoclaved

    relative to nonautoclaved HPR, one would expect a lesser

    tendency of the autoclaved CM to sorb Au(CN)2-during the

    adsorption step. Since this is not observed, one can infer an

    increased sorptivity of the carbon due to autoclaving. The

    same effect may also be observed in the LPR CM, where the

    autoclaved LPR CM shows slight uptake of Au(CN)2-whereas

    nonautoclaved LPR CM shows release of gold into solution

    even during the adsorption stage of this study.

    During the leaching stage, the LPR CM released much more

    gold (99.7 g autoclaved, 3.8 g non-autoclaved) than the

    HPR CM (0.88 g autoclaved, 1.0 g non-autoclaved)

    During the adsorption stage, LPR CM sorbed minor

    amounts (1.2 g autoclaved, -4.8 g nonautoclaved) o

    gold-cyanide complex relative to HPR CM (51.7 g

    autoclaved, 37.4 g nonautoclaved). This indicates HPR CM

    had much higher gold loading than LPR CM prior to the

    desorption phase, yet the HPR CM showed only 2-3%

    recovery of gold, whereas the autoclaved LPR CM gave 100%recovery. Desorption from non-autoclaved LPR CM could no

    be evaluated due to lack of gold sorption. The above result

    indicate that gold cyanide complex sorption onto LPR CM is

    readily reversible, whereas gold cyanide complex sorption

    onto HPR CM is irreversible under the conditions of the

    experiment. For both the LPR and HPR CMs, desorption of

    gold cyanide complex differed between the autoclaved and

    non-autoclaved CMs. Non-autoclaved CMs showed higher

  • 8/12/2019 99-136

    5/8

    Copyright 1999 by SME5

    recovery of gold relative to the autoclaved CMs indicating

    that autoclaving the CM is deleterious to gold recovery.

    Sorption of gold to the non-autoclaved HPR CM was almost 2

    orders of magnitude greater than sorption to the

    non-autoclaved LPR CM, assuming that both of these CMs

    had similar original gold content. If sorption is related to thephysical characteristics of the CM, this difference may help to

    explain the correlation observed between the physical

    characteristics of the non-autoclaved carbon and the %

    pregrobbing behavior of the whole ore (Stenebrten and

    Johnson, 1998). Furthermore, although autoclaving appears

    to increase the sorption tendency of both the LPR and HPR

    CMs, the autoclaved HPR CM remains much more sorptive

    than the autoclaved LPR CM. This may explain why physical

    properties of the non-autoclaved CM also correlated with %

    recovery (Stenebrten and Johnson, 1998), despite the ore

    being autoclaved during the recovery test.

    Since LPR CM showed good recovery of the adsorbed gold,this suggests that the Au(CN)2

    -sorbing mechanism for LPR is

    readily reversible. This correlates well with the high values of

    gold released during the initial leaching of gold in the first

    phase of the experiment. Although the LPR CM is unlike

    commercial activated carbon used in the CIL process because

    it shows little ability to sorb gold from solution, it is similar to

    commercial activated carbon in terms of readily reversible

    aurocyanide sorption. The difference in sorbing behavior

    could be due to a significantly smaller surface area of the

    naturally occurring LPR carbon compared to that of

    commercial activated carbon, and future studies will attempt

    to bear this out. It is evident that the aurocyanide sorption

    reversibility of HPR CM is quite different from that of both

    LPR CM and commercial activated carbon. Physical and

    chemical analyses have indicated strong similarity between

    commercial activated carbon and Goldstrike ore CM

    (Stenebrten and Johnson, 1998). However, the relative

    irreversibility of gold cyanide complex sorption onto HPR

    CM indicates a fundamental difference between the

    mechanism of gold cyanide complex sorption on HPR CM

    and the mechanism of gold cyanide complex sorption onto

    commercial activated carbon.

    CONCLUSIONS

    There was a significant difference in the amount of gold

    cyanide complex leached from the carbonaceous matter from

    HPR and LPR Goldstrike ore despite there being some kinetic

    limitation on the leaching process. After initial gold leaching,

    the HPR CM adsorbs an amount of gold cyanide complex

    which is almost two orders of magnitude greater than the

    amount of gold cyanide complex adsorbed by the LPR CM.

    Additionally, whereas the gold cyanide complex adsorbed by

    the LPR CM is readily reversible, the gold cyanide complex

    adsorbed by the HPR CM is not reversible under the

    conditions studied. Finally, autoclaving increases the

    sorptivity of the CM: there was more gold sorbed by the

    autoclaved CM from both the LPR and HPR ores, and the

    recovery of sorbed gold was higher for the non-autoclaved

    CM. It is suggested that there is a mechanistic difference in

    the sorptive behavior of HPR and LPR CMs and that theoxidation of the CM due to autoclaving has an effect on the

    chemical properties of the CM.

    REFERENCES

    Chryssoulis, S., Weisner, C. and Wong, C., 1996

    Deportment of gold in composites HPR- and LPR-l o

    Goldstrike, Barrick Goldstrike, unpublished company report

    10 pp.

    Hausen, D.M. and Bucknam, C.H., 1985, Study of

    preg-robbing in the cyanidation of carbonaceous gold ores

    from Carlin, Nevada, Applied Mineralogy, Proceeding ofthe Second International Congress on Applied Mineralogy

    Park, W.C., Hausen, D.M., Hagni, R.D. eds., AIME

    Warrendale, PA, pp. 833-856.

    Sibrell, P.L., Wan, R.Y. and Miller, J.D., 1990

    Spectroscopic analysis of passivation reactions for

    carbonaceous matter from Carlin trend ores, Gold '90, SME

    symposium, Salt Lake City, UT, pp. 355-363.

    Smith, G.C., Feb. 20, 1968, Discussion of refractory ore,

    Carlin Gold Mining Company unpublished report.

    Stenebrten, J.F. and Johnson, W.P., 1998

    Characterizations of Goldstrike ore carbonaceous material

    1. Chemical Characteristics, submitted for publication, 33

    pp.

    Stenebrten, J.F. and Johnson, W.P., 1998

    Characterizations of Goldstrike ore carbonaceous material

    2. Physical characteristics, submitted publication, 47 pp.

  • 8/12/2019 99-136

    6/8

    Copyright 1999 by SME6

    Table

    3.DesorptionrecoveryofAdsorbedGol

    din

    HPR

    andLPR

    CM

    T

    able2.AuadsorptiondataforHPR

    and

    LPR

    CM

  • 8/12/2019 99-136

    7/8

    Copyright 1999 by SME7

  • 8/12/2019 99-136

    8/8

    Copyright 1999 by SME8