Citation
Determination of percentage levels of palladium in alumina-rich matrices by fire assaying

Material Information

Title:
Determination of percentage levels of palladium in alumina-rich matrices by fire assaying
Creator:
Thomas, Michael Perry
Publication Date:
Language:
English
Physical Description:
xi, 155 leaves : illustrations, photographs ; 29 cm

Subjects

Subjects / Keywords:
Palladium -- Assaying ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 69-72).
General Note:
Submitted in partial fulfillment of the requirements for the degree, Master of Basic Science, Department of Geology.
General Note:
Department of Geography and Environmental Sciences
Statement of Responsibility:
by Michael Perry Thomas.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
16855275 ( OCLC )
ocm16855275
Classification:
LD1190.L56 1985m .T43 ( lcc )

Full Text
DETERMINATION OF PERCENTAGE LEVELS OF
PALLADIUM IN ALuMINA-RICH MATRICES
BY FIRE ASSAYING
by
Michael Perry Thomas
B.A., University of Colorado, 1972
A thesis submitted to the
Faculty of the Graduate School of the
University of Colorado in partial fulfillment
of the requirements for the degree of
Master of Basic Science
Department of Geology
1985 At


This thesis for the Master of Basic Science
degree by
Michael Perry Thomas
has been approved for the
Department of
Geology
by
Date
/f. /f3=
Collin Hightower


Thomas, Michael Perry (M.B.Sw Geology)
Determination of Percentage Levels of Palladium
in Alumina-Rich Matrices by Fire Assaying
Thesis directed by Assistant Professor Jeffrey Kurtz
This research investigated the feasibility
of the application of fire assaying methods for
the determination of palladium in alumina or in
neutral, acid ores and concentrates. Standard
fluxes and sample quantities were used in the
first experiment and found to be inadequate in terms
of fusion control and the recovery of palladium from
the dord beads.
In the final experiment, the fusions were
controlled with a potassium carbonate flux and a
fusion temperature of 975C. Silver inquarts up to
98.7 milligrams were used for the recovery of
palladium from one-tenth assay ton of sample
(2916.7 milligrams). The cupellation temperatures
were started at 950C and stopped at 825C for best
results.
A certain amount of lead was locked in the
dord beads. Therefore, an AA spectrophotometric
analysis of each dore bead was performed in order to


verify and/or modify the results of the gravimetric
determinations of the dore beads derived from the
IV
fire assaying method.
A 98 percent recovery of palladium was
achieved at a cost per sample of $7.50, including the
AA analyses.


DEDICATION
To my wife, Patricia, for her patience and under-
standing since I have spent the last two years in
the laboratory at the furnace, away from the family
on so many evenings and weekends.


VI
ACKNOWLEDGMENTS
It would be most unbecoming if no official
"thank you" were offered to the many who aided me
in this research project. Thus, I wish to express
my gratitude to my students in the Geology Labora-
tory at Mullen High SchoolMartin Robinson, Jean-
Clair Breyer. Christopher Clark, and Regan Wick
for their time and dedication in the promulgation
of this research.
I wish to extend a special, most gracious
appreciation to Dr. Warren Longley, Professor
Emeritus of Geology, University of Colorado, for
having enkindled the fires of enthusiasm in me many
years ago.
This particular research would not have
occurred if the members of the Analytical Department
at Marathon Denver Research Center had not posed
the problem to me two years ago. My greatest
appreciation to each of you. I hope that I have
helped to answer the question.


CONTENTS
Chapter Page
I. INTRODUCTION ................................ 1
Limits of Applicability of this
Research ................................ 3
Physical and Statistical Limitations
of this Research................. 5
Arrangement of the Thesis.......... 7
II. THE FIRST EXPERIMENT................. 15
The Objectives of the First Experiment 15
Experimental Methods for the First
Experiment....................... 15
Assay Set Number One................... 25
Assay Set Number Two................... 28
Assay Set Number Three................. 30
Assay Set Number Four.................. 33
Assay Set Number Five.................. 36
Results of the First Experiment. . . 41
III. THE FINAL EXPERIMENT
The Objectives of the Final Experiment 44
Experimental Methods for the Final
Experiment....................... 44
Assay Set Number Six................... 46
Assay Set Number Seven................. 48


Chapter
vm
Page
Assay Set Number Eight................ 49
Assay Set Number Nine............. 49
Assay Set Number Ten.................. 52
Results of the Final Experiment . 55
IV. RESULTS AND CONCLUSIONS OF THIS
RESEARCH............................ 5 8
The Experimental Data............... 5 8
Noteworthy Observations ............ 61
Conclusions......................... 65
BIBLIOGRAPHY ..................................... 69
APPENDICES.................................... 7 3
APPENDIX A RESULTS TABULATION ................ 74
APPENDIX B STATISTICAL TESTS ON EXPERIMENTS:
F TESTS, T-TESTS .............. 76
APPENDIX C A COMPARISON OF FIRE ASSAY
RESULTS WITH A.A. SPECTRO-
PHOTOMETRY RESULTS FROM ASSAY
SET NOS. 7,8,9, 10............................ 79
APPENDIX D TABLE OF FLUXES USED FOR THE
EXPERIMENTS...................... 82
APPENDIX E APPARATUS AND MATERIALS............ 84
APPENDIX F CHEMICAL STANDARDS............... 87
APPENDIX G OUTLINE OF ESSENTIAL STEPS FOR
CLASSICAL FIRE ASSAYING............ 89
APPENDIX H FLUX CALCULATIONS FOR CLASSICAL
FIRE ASSAYING.................... 99


IX
APPENDIX I
APPENDIX J
Page
- FIRE-ASSAY FLUX DESIGN AND
ANALYSIS FOR FOUR ROCK TYPES
OF THE STILLWATER COMPLEX. ... 108
-FIRE ASSAY FLUX TABLES FOR
STILLWATER ROCKS OF
APPENDIX I...................... 15 2


TABLES
Table Page
1. Flux tests and preliminary assays ... 8
2. Chart of experimental variables for
all assay sets............................ 14
3. Palladium and silver control assays . 24
4. Assay set number one....................... 27
5. Assay set number two....................... 29
6. Assay set number three..................... 32
7. Assay set number four...................... 37
8. Assay set number five...................... 39
9. Results of the first experiment .... 42
10. Assay set number six....................... 47
11. Assay set number seven..................... 50
12. Assay set number eight..................... 51
13. Assay set number nine...................... 53
14. Assay set number ten....................... 54
15. Results of the final experiment .... 56
16. Tabulation of results from the ten
assay sets................................ 59
17. Statistical tabulation of results from
the ten assay sets........................ 60
18. Average palladium recovered per assay
set....................................... 63
19. Time x temperature factor in fusion . 64


XI
FIGURES
Figure Page
1. General flow sheet for the first
experiment......................... 16
2. Palladium spectrum photographed through
the Vreeland spectrometer ................. 22
3. Palladium spectrum, as above......... 22
4. Lead in a dore bead, #41, photographed
through the Vreeland Spectrometer . 23
5. Silver in a dore bead, #41, photographed
through the Vreeland Spectrometer . 23
6. Dord beads in contrast............... 34
7. Dore beads and cupels................ 3 4
8. A comparative graph of fire assay
results versus Atomic Absorption
Spectrophotometry Results ................. 62
9. Temperature-time function x 10^ . . 66


CHAPTER I
INTRODUCTION
This research has been conducted in order
to determine the feasibility of the application
of the Fire Assaying Method in the determination
of percentages of palladium in neutral, acid ores.
Palladium in an alumina matrix is one such neutral,
acid ore (Bugbee, 1940; Smith, 1979). The samples
used in this research are petroleum refining cata-
lysts. Although these samples are man-made materials,
they do possess the essential qualities of neutral,
acid ores.
This research was formulated after numerous
discussions with the staff in the analytical depart-
ment of Marathon Oil Company (Beazley, 1982, 1983;
Hoffner, 1982, 1983) and with mining assayers in
regard to the problems associated with palladium
determination (Scanlon, 1979; Shirley, 1983).
Controllability is the greatest difficulty with the
fire assay of materials (ores) containing upwards
of 0.1 percent palladium (Shirley, 1983).


2
Problems with precision and accuracy plague most of
the wet chemical methods employed in the determi-
nation of palladium in alumina (Hoffner, 1982, 1983).
It has been the focus of this research to
standardize a method whereby the greatest control-
lability and, therefore, the greatest precision and
accuracy would be attained in the determination of
palladium via a Fire Assaying Method. Since the
research deals with palladium in alumina (a neutral,
acid matrix), this work should prove to be useful to
both the refining industry and the mining industry
as a tool with which to initialize palladium determi-
nation via Fire Assaying in spite of any variations
in the matrix, concentrate, or ore with which each
might deal. In the case of a variation from a neutral,
acid matrix, which has been considered in this research
(see appendices on Stillwater, Appendix I), the assayer
need only follow the precepts given in this thesis and
vary the reducing power or acidity of the flux (see
flux tables in Appendix D and flux calculations in
Appendix H) as befits the matrix, concentrate, or ore
with which he must deal.


3
Limits of Applicability
Of this Research
Fire Assaying, as an analytical method, has
been used for the determination of palladium since
the middle of the nineteenth century, at least. So,
there is nothing new about the idea of performing a
fire assay in order to determine the amount of
palladium in an ore or in any material. However,
the determination of palladium in the range of 0.4
percent to 0.6 percent in neutral, acid matrices
(alumina with 0.542 percent palladium in this case)
is a subject which is not discussed per se in the
analytical chemistry literature. This research
addresses the determination of palladium in the above
range and implies that such determinations can be
made on materials which may contain up to one percent
of palladium by weight. This research differs in
scope and applicability, in this regard, from any
published work. Most analytical techniques address
the determination of palladium in terms of quantities
which are measurable in the range of one-hundredth
to one-ten-thousandth of one percent. A reasonable
assay of an ore might, for example, give a result
of two troy ounces per short ton, which is the same


4
as sixty eight parts per million, or 0.0068 percent
palladium. Almost any adequate fire assaying method
can handle such an assay without significant devi-
ations. If, however, the assayer were to attempt to
use the same methods on an ore which contains, for
example, 5,420 parts per million of palladium, then
the results would be predictably erratic since the
assay would be subjected to spitting, frothing, and
palladium loss in the slag (Shirley, 1983). An ore
containing such high quantities of palladium is unusual
in nature but is not unusual in ore concentrates or
in petroleum refining catalysts. Controllability of
the assay decreases as the percent content of palladium
increases (Shirley, 1983). This, then, is the heart of
the research: to determine a method using fire assay-
ing by which the sample is brought under maximum con-
trol throughout the assay process, giving the most
precise assay results.
Why should fire assaying be considered as a
legitimate method for the determination of palladium
or any other platinum metal in light of the recent
advances in other procedures, namely, Atomic Absorp-
tion Spectrophotometry, X-Ray Fluorescence Spectro-
photometry, Polarography, Mass Spectrometry,


5
Radioactivation Analysis, etc.? Most of these ana-
lytical methods require significant investments of
capital in order to procure the fundamental instru-
mentation. Certainly, each of these methods can
give accurate analyses. But each of these methods
is initially very expensive and beyond the reach
of all but the most solid corporations and universi-
ties. Fire assaying requires a good furnace, a
furnace controller, and a good analytical balance.
The initial investment difference is over one
hundred-fold.
The quantity of sample used in fire assaying
has an additional advantage over the other methods.
Fire assaying usually uses from 3 to 30 grams of
sample per assay. This helps to partially overcome
the problem of inhomogeneous distribution of precious
metals which is common in many ores.
Physical and Statistical Limitations
of this Research
There was some uncertainty about the homo-
geneity of the palladium catalyst used in this
study. According to the analytical department at
Marathon Oil Co., wet chemical methods gave an analysis
and accuracy of 0.542 percent palladium 2.5 percent


(Hoffner, 1982) palladium in petroleum refining
catalyst (palladium in alumina) for all of the
samples analyzed through May of 1982.
6
The average content of palladium in the samples
used for this research was taken to be 0.542 wt. per-
cent (Hoffner, 1982). This value of 0.542 wt. percent
was purportedly from American Cyanamid, but no certifi-
cate of assay was available.
The detection limits and sensitivity limits
for the instruments and balances were
1. Emission Spectrometer:
(0.1 g of sample was used in each case)
lead..............0.1 ppm (by experiment
with Pb on
carbon)
silver...........0.02 ppm (Crook, 1969)
palladium........0.05 ppm (by experiment
with Pd on
carbon std.)
2. Atomic Absorption Spectrophotometer:
lead..............0.03 ug/ml 1%
silver............ 0.002 ug/ml 2%
palladium.........0.025 ug/ml 2%


7
4. Ainsworth Analytical Balance:
0.00005 0.00001 g
5. Troemmer Analytical Balance:
(Pulp Balance)
*
0.001 g 1 0.00005 g
Sixty trial fusions and forty preliminary
assays were performed in order to determine the best
possible fluxes for neutral, acid ores including
alumina (see Table 1). The objective was
to sec whether the assay furnace and the cupellation
furnace functioned predictably and to test fluxes
for the assays to come.
Arrangement of the Thesis
The physical and statistical limits of the
thesis are considered in the introduction along with
the preliminary assays and calibrations for the
analysis of palladium in alumina. A set of first
objectives were derived from the preliminary
assays. These objectives constituted the focal
point of the first experiment. The first experiment
was accomplished in five sets of ten assays. Pro-
cedural changes were made after each assay set,
according to its outcome. Results and discussion
followed.


Table l.--Flux tests iiiid pielminjiy assays
SdBple number date Sample Haul) Noleu Assay ilea(y) Flux type FI ox yuunL Dor A net l Pd I tea wt Furnace Slag notes
H-i I0/B2 Blank 20 Check furnace Ay llil Bo.O -0- 0.003 none Av. 1100* Clear
40 am.
hp2 20 " " Ay - Bb.U -0- 0.003 - - -
Ml'J " - 20 Ay - B6.0 -0- 0.003 - - -
Ml'4 - " 20 - Ay " U6.0 -o- 0.003 - -
MI'S - - 20 - Ay - Bb.O -0- 0.003 - - -
Ml'to n/1 j/ Pd 11) SO Pd Cdlalybl Pd a ox B6.0 20.Ib%
d2 Al iui na sample lO.OOly 0.004 0. 162 0.I5B 0.S42 T 1100* Cloudy with
(jOl.-UJ
T 1110* 2 3
* (MO streaks
Ml* 7 " Pd m ox 120.0 0.0045 none tai led T 1050
T 1100 no separation
y 40
Mi'U m " " I'd 11 ox 126.0 - 0.004 - failed - -
MI*4J " - - Pd - - - 0.004S talled -
MilO " ** - Pd - " " 0.004S - tai IcmI
CD


a>
Off .001 1 0011 \t. *X p* M n 1 T *00 *0 100*0* /.9t *6r 0*971 k n P.1 5* t- ID lAtrip.- pa 05 pm (am t p Ml p.| 70 /rt/n 1 MW
. 5ti 00 It .0011 - jx X P*M ipj *00 0 9MH
. sea oott oou * IX ?x pr,, |r, 5*0*0 51 dW
. 5te .006 .0*0! * t1 X P*t frj 0*00*0 *MH
lToq pup snioij 5te .006 .0*01 * Jx X pr>| tp| 9*1011 7*00*0 100*0* £9! *6? 0 *9? t xo 11 Fd ap!tnp? IrAipipo pj 05 pm t am (p ut p* 79 /7T/-T f 1 .?H
. 5Cf oou ORO! * IX fx t. 051 *0 5*00*0 *o tl Fd 05 7 MH
qn Aptiofa 5Ed OOU 0601 - *x J, t. on *o T *00"0 100*0* £9! *67 0*9"! X 1 M 7*t ID 1rA|pip3 pd 05 puinm|p ('! prl 79 /9I/I! TI.1H
KOIOM 6p|S r>piun^ 1* '*11 (ft) pd % i* 9 nyj (ft) (ft) i i P loci urnbui ftv (ft) 1* 9Jrlwr?5 tr>) IhpuIi *n|.4 f*!Al (9)W*lf Apr (pnnutlUfV)) I 5tqP,L


Tahiti 1 (Continued)
Sample number 8aB|il F*l' dale Sample Mesh Notea Assay 1 Lem(s) Flux type Flux yuant (m) Sample wt. W> Ay Inyuart wt. <>j) Dord wt Hi'IB 12/3/ 02 I'd l n i 1 uan ni bO I'd catalyst sumpl a Cl.- 145 I'd 11-OX I.'6.0 29. Ifa7 10.0041 0.001b none tailed t moo* T*-l100* 1 #30 Fiotlis and boi la
MII9 0.004b tailed T *IObO* t|iioo* #30
HiJl)* 0.004b tailed T I0b0* t 1100* 1 #40
Ml'2l n/2/ 04 I'd in alumina I'd catalyst sample Cl.-145 106.0 29.167 10.000b 0.0050 10.000b o. no 0. 12S 0. 429 T 1010* t 10101 #40 0^>ayue witli bands ol Fe^O.
MIJ2 0.146 0. 141 0. 449 T 1000* t 10JO* 30
Ml*-'J - - " - " - - - o. nby o. no 0.41 1 - -
Ml-'4 - - - - - - - 0. J4b 0. 140 0. 440 - -
Ml -'b 0. 161 0. lb 0. b4b T 9b0* tJ 1000* 1 #40
MI'.'u ** ** " 0.0060 1O.000b 0. Ib4 0. I4B 0. S07 **
Ml'27 - - - - - " - - - 0.14b o. no 0.442 - -
*l'y l4Ui.li: I J||J LDiiliul Ikjx iwljuulcil tiltul llila III:.ion.
o


Table 1 (Continued)
nuiubei: Sample Wl* date Suable Mesh Noted Assay itcm(s) Flux type Flux (|U.lllt <9l Sample WL (Ml *8 inquait WL. (8) Dore wt (9) Dole liet w) % Pd Item wt Furnace Slag notes
MP28 unt I'd in SO Pd catalyst Pd 11-L.X 106.0 28.167 0.0060 0.130 0. 124 0. 425 T 850* Opauue wil h
Ui a 1 uui iij sample 10.002 30.0005 1J 1000" bandit <. .f Fe O
GL- i 4S #30
MP28 - - - - - - - M - 0. 116 0. 110 0. 377 T * 1000* -
- 1040*
HI JO - " - - - - - - 0. 121 0. 1 IS 0. 384 T . 1000* -
- 10S0*
MP3I I/JO/ - - Pd Cdtdlysl - 11-OX 1 10.0 " O.OOoO 0.133 0. 127 0.43S T 10S0* Opaque with
tiJ sample 1 l 9 t! 1080* Fe O
GL- 145 f #30
Ml32 " - - - - " - - 0.0048 0. 1S3 0.118 0.50* T . 1080* "
Tf 1080
#35
MlJ 3 ** - - 0.0062 0. ISO 0. 144 0.484 -
M134 - " - * " - - O.OOSI 0. 156 0. ISI 0.S18 >
M1'35 - - - - - - - 0.0057 0.128 0. 123 0. 422 " -
Ml-JO i Pd in SO Pd catalyst Pd 1 OX 106.0 28.166 0.005J 0. 1JS 0. no 0.446 T * 1000* -
83 alumina sample 10.002 Tf . IOSO*
GL- 142 #35
Ml* 37 - - - - 0.0058 0. 14H 0. 142 0. 488 -
Ml* Jd - " - - - " 0.0048 0. 1 32 0. 127 0.4JS - -
Ml*38 - - - - - - - - 0.0052 0. ISO 0. ISI 0.S18 T B 1050* -
TI B 1080*
#35


\
3
gs*o*o
/.too
WO x o?
%s/.**o x *ot
__________ N 0* "1 T£
N 0* oa |C
*srff!nirs tnj
ss|HrS m.j
*V*I
V=U i|)in onbndo sr OfiOl OSOl * J*1 X 06to CM *0 6H 0 9S000 TOO 0? 991 *6.7 O'OOf yo-| Pd TH-10 9|dni**R 3A|f3Pn Od piittun(p OS ti r p,i rn /fir/i OMW
S9inu 6* IS r>tM3nt| Pd % (t>) inn ei'ri <*> 1* JMOfJ (*> jjrnbuf *v <*> 1* 0(riwr*c; IMrtih *"IJ mIXi Aprby BWIOM i^p rll t>1 "|dwts jnqwn'l 0(dwrq
(}ontifitinr)) i pfrin


13
The objectives for the second (final) experi-
ment were developed throughout the first experiment.
Five sets of ten samples each were analyzed for this
experiment. All of the procedural changes, thought
to be indicative of the best results, were incorpo-
rated into the final experiment.
A chart of expected experimental variables
was drawn up as a guideline for the variations in
both experiments (see Table 2 following).
The conclusions of this research are based
upon the results obtained from one hundred fire assays.
Only ten assays (assay set 2) failed because of equip-
ment malfunction.
The results and conclusions for the thesis
are given following the final experiment. Appendices
regarding this research and preliminary research on
rocks of the Stillwater Complex are included after
the main body of the thesis.


Table 2.-- Chart of expel imental variables for all assay sets
Assay set no. Sample lucp. Sample wt. Range of inyuuit wt. Flux typo Flux quunt. Fusion temp. Fusion time Litharge wash Cupellation start temp. Cupellation stop temp. See table no.
1 * 29X67 my. Kxc.i 2, 3, and 4 3.0 to 6.5 mg. 11 OX. * 107S*C 40 min. None 1000C 825C 4
2 * 29167 mg. (1 A.T.) 4.1 50 4.3 mg. tfl ox. * 1095"C 45 min. None 1000*C 825C 5
3 * 29167 mg. (i A.T.) 3.7 to 4.B mg. M3 ox. * 1080C 35 min. 50 g. 1020C 815 C 6
4 roast ed **90OC diy g 105 C 20167 mg. (1 A.T.) 3.9 to 5.2 mg. #1 ox. * 1060C 40 min. None 1000C 800"C 7
5 t ousted <$ 900 C dey y 105C 29167 mg (1 A.T.) 16.6 to 20.1 mg. M3 ox * 1080C 40 min. 50 g. 1020C aisc B
6 iousted y oooc J*y 22c 2916.7 mg. (1/10 A.T. ) 47.3 to 62.7 mg. #4 ox. 79 g 1030C 30 min. None 1000C 850C 10
7 Iousted 0 900C dry 22C 2916.7 mg. (1/10 (A.T.) 41.0 to 61.2 mg. MS ox. 79 g. 97SC 35 mrn. None ioooc 825 C 11
B t ousted fe* 900 C diy t 22 C 2916.7 mg. (1/10 A.T. ) 47.0 to 71.0 mg. M5 ox. 79 g. 975C 35 min. None 950C 825-C 12
9 e oooic diy 0 22C 2916.7 my. (1/10 A.T. ) 48.5 to 74. 1 mg. 5 ox. 79 y. 975C 30 min. None 950C 825C 13
10 rousted y 900 c diy y 22C 2916.7 mg. (1/10 A.T.) 80.5 to 98.7 mg. MS ox. 79 y 975 C 35 min. None 950C 825C 14

J lent as given lit llie section on experimental proceduies


CHAPTER II
THE FIRST EXPERIMENT
The Objectives of the First Experiment
The general objective for the first fifty
assays was to generate sets of data on the palladium
content of the refining catalyst (palladium in
alumina) using one assay-ton samples (29,167 milli-
grams) per assay: one assay-ton is a traditional
quantity in fire assaying and was a sufficient
sample size for initial results in the research.
Additionally, the assay sets were to give indica-
tions as to the best flux, time, and temperature
combinations in order to maximize the recovery of
palladium from each assay.
Experimental Methods for the
First Experiment
The general flow sheet for the first experi-
ment is given on the next page (see Figure No. 1).
The following is a list of the experimental methods
and procedures in terms of the time sequence for
each assay set:


General Flow Sheet for the First Experiment
Assay Sets.
Figure 1. General flow sheet for the first experiment.
CT


17
1. Sample of 400 grams of catalyst was
dried at 105C for 24 hours.
2. Sample was cooled and rolled in thin
rubber mat to homogenize (heavy particles
concentrate toward the bottom of the
sample).
3. Sample was split on the thin mat with a
long, dull spatula into cuts of about
30 grams each.
4. Each 30-gram sample was scraped from
the thin mat into a preweighed weigh
boat.
5. The sample was placed upon an analyti-
cal balance and reduced to a net of
29.167 milligrams by removing sample
from bottom to top with a small spatula.
6. The sample was transferred to an assay
crucible. The average loss of sample
in transfer of ten samples from weigh
boat to crucible was deducted from the
sample weight so that an average of
29.167 milligrams 0.1 milligram was
recorded for each assay.
7. Sample weight was recorded in the log
of the experiment.


18
8. A silver inquart from 3 to 7 milligrams
was weighed out on the analytical
balance and placed in an assay
crucible and recorded. Assay crucibles
were numbered with cobalt markers on the
outside.
9. The analytical balance was recalibrated
after measurements for each assay set.
10. The flux for each assay set was prepared
and thoroughly mixed. Fluxes were placed
into the 30-gram assay crucibles. There
was roughly 100 to 125 grams of flux
per assay; this depends upon the flux
components.
11. The flux and sample were thoroughly
mixed in the crucible; mixing was done
with caution so that sample would not
be lost.
12. The fire assay furnace was preheated to
a starting temperature.
13. Starting temperature and time for the
fusion were decided on the basis of
preliminary assays or preceding assays.


19
14. Assays were fused in sets of two.
15. Assays were checked at 5 and 10 minutes
into the fusions for evidence of rapid
boiling.
16. At the termination of the fusion, the
crucibles were taken from the furnace
and the contents were decanted into a
slag mold and left to cool.
17. The slag of each assay was separated
from the lead button; the button was
lightly hammered into a cubical form.
18. The cupellation furnace was preheated
to a starting temperature and the cubi-
cal, lead button of each assay was
placed on a sheep bone ash cupel and
placed in the furnace (cupels were
preheated to furnace temperature).
19. The lead of the cubical, lead button
was oxidized until only a remnant bead
was left in the cupel of each assay.
This bead is called the dore bead. The
cupellation process was terminated
when the remnant bead (dore bead) of
each assay gave off its latent heat of
fusion, i.e., a flash known as a "blick."


20
20. The cupel and dord bead of each assay
were removed from the furnace and allowed
to cool.
21. The weight of each dord bead was deter-
mined on an analytical balance and
recorded.
22. Portions of the slags, crucibles, and
cupels were analyzed by emission
spectrometer in order to determine the
presence of palladium or silver in these
items for each assay.
23. The dore bead of each assay was placed
in a 10-dram vial, labeled, and stored.
24. The silver inquart was subtracted from
the dore bead and the net palladium
recorded.
A more in-depth discussion of the fire assay-
ing process is given in the appendix titled "Outline
of Essential Steps for Classical Fire Assaying"
(see Appendix G).
Item number 22 in the above procedures should
be amplified. The cupels, slags, and remnants of
the fusions in the crucibles were pulverized to 100
mesh and placed upon ceramic hearths, 0.1 gram per


21
sample, for inspection with the Vreeland Model #7
Emission Spectrometer. The presence of small
amounts of palladium and silver could be detected
in this way (see section on Physical and Statistical
Limitations of this Research). Some palladium
was discovered on the tops of the cupels of some
assays and in the slags. The amounts were so small
that the palladium appeared as flickers of the
spectrum in the yellow and green wavelengths and
these appeared with insufficient intensity to be
quantified (Gerlach and Schweitzer, 1929; Seath and
Beamish, 1953). Two sets of photographs showing
positive results from the spectrographs of palladium
are shown in Figures 2 and 3. Spectrographic photo-
graphs for Pb and Ag are shown in Figures 4 and 5,
respectively. Luminous intensities from palladium
in the region of 0.001 percent concentration are
sufficient to be reproduced on a photocopy of a
3000 speed polaroid photograph of the emission
spectra. The discussion of these results follows
where appropriate with each assay set in the
experiment.
Four calibration assays were run prior to
the beginning of the first assay set (see Table
No. 3). A known amount of palladium with carbon


22
Figure 2. Palladium spectrum photographed
through the Vreeland spectrometer.
Sample is from Eastman Standard,
5% palladium with carbon. 5-83
Figure 3. Palladium spectrum, as above.
Sample from dord bead, assay
No. 10. 5-83


23
I
Figure 4. Lead in a dor<§ bead, #41, photo-
graphed through the Vreeland
Spectrometer. 5-83
Figure 5. Silver in a dore bead, #41, photo-
graphed through the Vreeland
Spectrometer. 5-83


Table 3.--Palladium and silver control assays
Sample Sample prep Assay FI ux Flux Sample Ag inquart Core A-A Spec A-A Spec Hecovery Hecovery Total
number date Sample Mesh Notes i tern (s) type quant wl. wt. wt Analysis Analysis grams grdms Recovered
(9) (9) (9) (9) Pd PPM Ag PPM Pd *9 9
SI
2/2/
S3
Pd
A1 0
2 3
100
Pd, O.05g
10.001 on
carbon and
Pd
Ag
II ox 106.Og 29.166g 0.005: 0.054 490500 48000
10.002g 10.0001
10.05g
Pd
0.04905 0.00480 0.005385
S2 M . " 0.005 10.0001 0.052g 466000 476000 0.04660 0.00476 0.05136
S3 2/20/ 83 Fd t ai23 100 Pd, 0.150g 10.001 on carbon and Al23 Pd Ag 11.ox 104.0 No flour 29.166g 10.002g + 0.150g Pd 0.005 10.0001 0. 149 453020 32148 0.1420 0.00479 0. 1468
S4 0.005 10.0001 0. 152 971050 29805 0. 1476 0.00453 0. 1521
Note: IS1, S2, S3 Dare beads contain some lead, verified by A-A.
Notes: Furnace temp Pd % Ag %
________*c___________recovered recovered
si 1050C av. 98% 96%
for 35 min.
S2 93% 95%
S3 1040C av. 95% 96%
for 35 min.
S4 - 98% 91%
ro


25
was salted into 29,167 milligrams of aluminum oxide
(taken as alumina) for each calibration assay. The
average percent recovery of palladium from four
assays was 96 percent by weight. This was lower
than the expected recovery which was expected to
be in excess of 98 percent (Fraser and Beamish,
1954; Plummer and Beamish, 1959).
Assay Set Number One
A calculation of the flux needed for a sesqui-
silicate slag (see Preliminary Assays, Table 1
and Flux Calculations, Appendix H) indicated that
the #1 Ox. flux which had been used for the prelimi-
nary assays was in the correct range of acidity
(Bugbee, 1940). The amount of palladium returned
from the preliminary assays (0.460 wt. percent) was
unacceptable since the best known value was 0.542
wt. percent (see the section on Physical and Statis-
tical Limitations of this Research). It was neces-
sary to determine what effect the fusion tempera-
ture might have on the recovery of palladium, so
the temperature was set at 1075C instead of 1040C
which was used in the last of the preliminary assays.
Also, the fusion time was increased to 40 minutes.


26
The ten samples for the first assay set were
prepared as outlined in the experimental methods sec-
tion.
The fusions were performed as described in
the outline. Sample Nos. 2, 3, and 4 were spilled
while they were being taken out of the furnace.
An attempt was made to replace these assays with
fresh assays using sample which was left from the
original split. There was not enough for three assays,
so it was decided to run two h A.T. (assay ton)
samples and one 28,903 milligram sample (slightly
less than one assay ton) as replacements.
Sample Nos. 1, 2, 3, 4, and 8 were difficult
to control during the fusion; frothing and spitting
occurred from all of these samples; this possibly
caused some sample loss. There was a glaze on the
outside of the crucibles and some glaze on the floor
of the muffle from these samples.
There were some traces of palladium in the
slags and cupels after these assays (see Table 4).
The first assay set yielded an average of 0.515
wt. percent palladium with a standard deviation of
0.0392 wt %. Three values in the set were rejected
at the 99 percent confidence level on the basis of
the t-test (See Table No. 4 and Appendix B).


Table 4.--Assay set number one
Sample No. Sample Wt. (mg) Dore Wt. (mg) Ag InquarL Wt.(mg) Correction W t. (mg) Pd in Sample, WL.(mg) Pd %
1 29167 139.6 3.0 0 136.6 .4683
2 14584 71.6 4.5 " 67.1 .4601
3 14584 63.1 5.0 M 58.1 .3984
4 28903 171.0 6.5 M 164.5 .5691
5 29167 179.1 4.0 175.1 .6003
6 M 162.0 4.1 M 157.9 .5414
7 M 159.0 4.5 M 154.5 .5297
8 M 153.5 4.8 M 148.7 .5098
9 M 174.5 4.4 M 170.1 .5832
10 M 158.0 4.1 153.9 .5277
Statistics:
Highest % Pd in Sample 0.6003
Mean % Pd in Sample 0.4994
Lowest % Pd in Sample 0.3984
Average % Pd in Sample 0.5151
Values 13,5
Conditions:
and 9 were
rejected at
Standard Deviation
the 99% confidence level
0.0392
per t-tests.
Date of Hun: 10-82
Sample Lot No. GL-142 Sample Description: Pd in Alumina, Petroleum Refining
Catalyst
flux; 1 Ox. Furnace Temperature: 1075C Fusion Time: 40 minutes
Cupellation Temperatures: StarL 1000C, stop 825C
Slag and Crucible Analysis: Trace Pd in slag
Remarks Concerning Assay: Spitting, Vaporization from Crucibles No. 1,2,3,8,
Frothing, Boiling Nos. 1, 2, 3, 4. Id weighLs in
samples corrected for loss on ignition.
to
-4


23
Assay Set Number Two
In order to increase the recovery of pal-
ladium in this assay set, the furnace temperature
was set to an average of 1095C and the time was
extended to 45 minutes.
The preparation procedure followed the out-
line in the experimental methods section.
Samples Nos. 14 to 20 boiled over at about
30 minutes into the fusion process. Samples Nos.
11 to 13 showed some glazing on the outsides of the
crucibles. The fusions were terminated early and
the crucibles were inspected after the remaining
contents were decanted. The crucibles had been
attacked by the assay charges; even the floor of the
muffle had been partially decomposed. It appeared
that the charge had acted like a very strong base
at the 1095C fusion temperature. Alumina has been
found to exhibit such activity at high temperatures
(Shirley, 1983).
The average palladium netted from the three
remaining samples, Samples Nos. 11 to 13, was 0.4956
wt. percent. The standard deviation was 0.0561 wt.
percent (see Table No. 5). The whole assay set was
rejected owing to the sample losses from each sample.


Table 5.-- Assay set number two
Sample Do re Ag Pb Pd i n
Sample Wt. Wt. Inquart Correction Sample, Pd
No. (mg) (mg) Wt.(mg) Wt.(mg) Wt. (mg) %
11 29167 156.0 4.1 0 151.9 .5208
12 M 160.1 4.1 M 156.0 .5349
1 3 M 130.1 4.2 II 125.8 .4313
14 M II M II 0 0
15 H " 1 1 0 0
16 W " 1 II 0 0
17 M II I 0 0
18 II H " 0 0
19 N II 0 0
20 M " N 0 0
Statistics:
Highest % Pd in Sample: .5349 Lowest % Pd in Sample: .4313
Average % Pd in Sample: 0-4956
Standard Deviation: 0.0561
Conditions:
Date of Run: 11-82
Sample Lot No. GL-145 Sample Description:Pd in Alumina, Petroleum Refining
Catalyst
Flux: 1 Ox. Furnace Temperature: 1095C Fusion Time: 45 minutes
Cupellation Temperatures: Start 1000C, stop 825C
Slag and Crucible Analysis: Trace Pd; less than 0.05 ppm Pd.
Remarks Concerning Assay: Pyrometer Failure, Samples 13 through 20 Spit &
boil Over, Muffle Destroyed. Samples 11 through 13
frothed and boiled; some sample may have been lost.


30
Assay Set Number Three
The flux used in the previous assay set was,
in part, to blame for the failure of the assays. It
was decided to use the #3 Ox flux for assay set
number three since this might dilute the slag and
add some acidity to it (see Flux Table, Appendix D
and Flux Calculations, Appendix H).
Since there was such rapid vaporization in
the previous assays (Assay Set Number Two), the fur-
nace temperature of this set was lowered to 1080C
and 35 minutes was allowed for the fusion time.
Since there were traces of palladium in the
slags of sample nos. 11 to 13 in tfce previous assay
set, some litharge was added to each assay, about
50 grams, ten minutes before the end of the fusions
in order to wash the crucibles. This reduced the
probability that any palladium would be caught in
the slags and reduced the viscosity of the slags
for pouring.
The cupellation procedure was started at
1020C and stopped at 815C in order to inhibit any
palladium or silver from draining into the cupel at
the termination of this procedure; some palladium
and silver had been discovered on the tops of cupels
from Samples Nos. 11, 12, 13 in the previous assay.


31
The dord beads recovered from this assay set,
with the exception of Sample No. 22, weighed over 1.3
times more than the theoretical maximum for 0.542 wt.
percent palladium. This prompted a more detailed
study of the contents of the dore beads (see Table
6) .
A preliminary colorimetric analysis of the
dord beads was attempted using methods described
by Fraser and Beamish (1954) and also the Hach Chemi-
cal Company (1977). All ten of the dord beads of
assay set number three were digested and prepared
for analysis in a single solution, since very low
concentrations of Pb were expected. The high weights
of the dord beads were thought to have been due to an
error in the weighing of the Ag inquarts. Colori-
metric determination of the solution netted an
average of 13.3 percent of Pb locked in the dore
beads which was much higher than expected.
The problem with Pb being locked into the
dore beads was not unheard of, but there appeared to
be very little about this phenomenon in the litera-
ture except for the work of Havinga (1973). Some
lead-palladium structures were known to exist and
had been examined by x-ray diffraction. Pb^Pd^ and


Table 6.--Assay set number three
Sample No. Sample WL. (mg) Dore Wt. (mg) Ag Inguart WL. (mg) Pb Correction* Wt. (mg) Pd in Sample Wt.(mg) Pd 4
21 29167 215.2 4.8 28.7 181.7 .6230*
22 M 136.1 4.5 17.9 113.7 .3898*
23 H 216.8 3.9 29.0 183.9 .6305*
24 " 189.5 4.3 25.2 160.0 .5486
25 M 208.9 3.7 27.9 177.3 .6079
26 H 197.5 4.0 26.3 167.2 .5733
27 H 180.1 4.7 23.8 151.6 .5196
28 192.3 3.9 25.6 162.8 .5582
29 II 184.5 4.2 24.5 155.8 .5342
30 ** 185.6 4.0 24.7 156.9 .5379
Statistics:
Highest % Pd in Sample: 0.6305 Lowest % Pd in Sample: 0.3898
Average 4 Pd in Sample: 0.5542
Standard Deviation: 0.0293
Value nos- 21, 22, and 23 rejected at the 99* confidence level based on L-tests.
Conditions:
Date of Run: 1-83
Sample Lot No.: GI.-145 Sample Description: Pd in Alumina, Petroleum
Refining Catalyst
Plux 3 Ox. Furnace Temperature: 1080C Fusion Time: 35 minutes
Cupellation Temperatures: Start 1020C, Stop 815C (av.)
Slag and Crucible Analysis: Slag contains high 4 Pb, red-brown coloration on
Crucible; no Pd.
Remarks Concerning Assay: Near boil-over, high 4 Pb locked in Dore bead (see Havinga,
1973), hydration? Samples washed with 50 g. PbO 10 minutes
before crid of fusion. Id weighLs in samples corrected for
loss on ignition.
= 1 3 3-;.
Av. Pb in each boro
OJ
M


33
PdPb2, for example, have structures, owing to the
catenation tendency of lead, which form at tempera-
tures near the fusion temperatures of the fire
assays. Figures 6 and 7 show the dore beads with
and without the lead-lock problem. The dore bead with
lead locked in it is relatively brittle and has a
coating of what is presumed to be an oxide of pal-
ladium on the outside. What causes the formation of
such lead-palladium structures is not known. It was
decided that the washing of the charges with litharge
may have contributed to driving the Pb from the PbO
of the fusion reaction toward complex formation with
the palladium in the lead button before the pouring
took place (see Figs. 6 and 7).
A spectrometric analysis of the slags,
crucibles, and cupels detected no palladium in these
items.
The average palladium returned from assay set
number three was 0.5542 wt. percent and the standard
deviation was 0.0293 wt. percent.
Assay Set Number Four
A high percentage of water in the samples was
thought to be part of the problem with the fusions in
assay set number three. In order to avoid the problem


34


Figure 6. Dorg beads in contrast, a photograph.
Oxidized Pd on the dor£ beads, left. Non-
oxidized Pd in the dord beads, right.
Figure 7. Dore beads and cupels, a photograph.
Oxidized coating PdO on the dore bead in
a cupel, left. Non-oxidized Pd dore bead
on the right.


35
of rapid vaporization, the water content of the
samples was reduced. The sample preparation pro-
cedure was adjusted in the following manner:
1. 300 grams of sample were roasted at
900C for one hour and then placed
in a drying oven at 105C in the
presence of CaSO^ for 24 hours.
2. The sample was weighed out after being
cooled to room temperature in a desic-
cator and found to be 282 grams; this
represents a weight loss of 6 percent
owing to dehydration.
The remainder of the procedure for the prepa-
ration of this assay set was outlined in the section
on experimental methods.
In order to avoid the lead-locking problem
experienced in the previous assay set, a compromise
furnace temperature of 1060C was chosen with a
fusion time of 40 minutes. The flux content of PbO
was reduced from 47.9 percent (#3 Ox flux) to 58 per-
cent (#1 Ox Flux) since there appeared to be an
excess of Pb in the dore beads from the previous
assay set (Assay Set Number Three).
The cupellation process was started at 1000C
and finished at 800C. This was done since the 1000C


36
temperature was considered sufficient to start the
program yet low enough to avoid the possibility of
spitting during the cupellation process. No litharge
wash was used at the finish of this assay set.
The apparent palladium content of the samples
averaged 0.5609 wt. percent and the standard devi-
ation was 0.219 wt. percent. Four of the determi-
nations were rejected at the 99 percent confidence
level based on a t-test (see Table No. 7 and
Statistics, Appendix B).
The emission spectrometric examination of
the dor£ beads from this assay set revealed that
seven of the ten dor£ beads (Nos. 34 to 40) contained
some Pb, although this could not be quantified.
There were also traces of palladium in the crucibles
and slags of Samples Nos. 34 to 40.
Assay Set Number Five
The samples for this assay set were prepared
as in the previous assay set except that the Ag
inquarts were increased about four times in weight
to help collect the palladium (Fulton and Sharwood,
1929). The dehydration of the samples appeared to
be beneficial in the control of the fusions. (It
should be noted that loss on ignition corrections of


Table 7.--Agsay set number four
Sample No. Sample Wt. (mg) Dore Wt (mg) Ag Inguart Wt.(mg) Correction* Wt.(mg) Pd in Sample, W t. (mg) Pd %
31 29167 149.8 3.9 0 145.9 .5002*
32 M 152.7 4.1 II 148.6 .5095*
33 H 151.0 4.3 " 146.7 .5030*
34 M 173.8 4.8 M 169.0 .5794
35 M 156.7 4.6 H 152.1 .5215
36 N 171.5 4.7 M 166.3 .5719
37 M 169.7 4.9 M 164.8 .5650
38 M 165.3 4.7 II 160.6 .5506
39 M 173.6 5.2 M 168.4 .5774
40 it 181 .0 5.0 1 176.0 .6034 *
Statistics:
Highest % Pd in Sample: 0.6034 Lowest % Pd in Sample: 0.5002
Average % Pd in Sample: 0.5609
Standard Deviation: 0.0219
Values rejected at the 99% confidence level based on t-tests.
Conditions;
Date of Run: 1-83
Sample Lot No.: GL-145
Sample Description:
1) Pd in Alumina, Petroleum Refining
Catalyst
2) Dehydrated at 900C 600C 1 hour,
cooled, weighed out.
Flux: 1 Ox. Furnace Temperature: 1060C
Cupellation Tempera Lures: Start 1000C, stop 800C
Slag and Crucible Analysis: Trace Pd; less than 0.05 ppm Pd
Fusion Time: 40 minutes
Remarks Concerning Assay: Pb locked in samples 34 through 40 dore beads per emission
spectrometer.
to


38
6% were made to Assay Set Nos. One, Two and Three
since these samples were not dehydrated prior to
fusion.)
Since there was some evidence of Pb in the
dore beads in Assay Set Number Four, the furnace
temperature for the fusion process was raised to
1080C and 40 minutes were allowed.
The change from #3 Ox flux to #1 Ox flux
(see Flux Table, Appendix D) appeared to have no
effect on the lead-locking problem. So the #3 Ox
flux was used for this assay set. It was thought
that this might lower the viscosity of the slag and
acidify it slightly.
The cupellation start temperature was raised
to 1020C and the finish temperature to 815C (see
Table 8) in order to help overcome the lead-palladium
complexing problem.
Since there was palladium in the slags of
Assay Set Number Four, it was decided to attempt to
wash the fusions with litharge once again. The silver
inquart had been increased to help to gather the
palladium in the fusion process, but it was thought
that additional litharge toward the termination of
the fusion would help to collect the remaining
palladium from the slag. About 50 grams of litharge


Table 8.--Assay set number five
Sample No. Sample Wt.(mg) Dor 6 Wt. (mg) Ag Inguart Wt.(mg) Pb Correction* Wt.(mg) Pd. in Sample Wt.(mg) Pd %
41 29170 215.6 16.8 39.0 159.8 .5478*
42 M 214.9 16.6 38.9 159.4 .5465*
43 M 204.8 20. 1 36.2 148.5 .5091
44 H 204.5 18.5 36.5 149.5 .5125
45 M 206.5 19.8 36.6 150.1 .5146
46 M 206.5 19.8 36.7 150.7 .5166
47 M 204.4 18.0 36.5 149.9 .5139
48 M 203.8 19.0 36.2 148.6 .5094
49 M 213.0 18.5 38.1 156.4 .5362
50 M 209.3 19.3 37.2 152.8 . 5238
Statistics:
Highest l Pd in Sample: 0.5480 Lowest % Pd in Sample: 0.5091
Average % Pd in Sample: 0.5170
Standard Deviation: 0.0090
* Value nos. 41 and 42 were rejected at the 99% confidence level based on t-tests.
Conditions:
Dale of Run: 3-83
Sample Lot No.: GL-97
Sample Description: 1) Pd in Alumina, Petroleum Refining
Catalyst
2) Dehydrated at 900C to 600C
1 hour, cooled, weighed out.
Flux: 3 Ox. Furnace Temperature: 1080C Fusion Time: 40 minutes
Cupellation Temperatures: Start 1020C, Stop 815C
Slag and Crucible Analysis: No Pd
Remarks Concerning Assay: No boil-over, high % Pb locked in each sample Dore bead;
by A-A Spectrometer. Each sample washed with 50 g PbO
10 minutes before end of fusion process. **
** Average of 18.1% Pb locked in each sample.


40
were added to each assay about ten minutes before
the end of the fusion process.
The dore beads from this assay set had values
which were 1.3 times the known average value of the
samples; the known average value was 0.542 percent
(see section on Physical and Statistical Limitations
of this Research). Lead was again suspected in the
beads.
Each dore bead was dissolved in hot aqua
regia and brought up to volume in 100-ml volumetric
flasks. The aliquot from each dord bead was run on
the atomic-absorption spectrophotometer against a
standard. The values were converted to milligrams
and recorded (see Table 8, Pb corrections); the
colorimetric methods for the determination of lead
were shelved following the acquisition of a lead lamp
for the atomic-absorption spectrophotometer.
The average palladium content in the samples
was 0.517 wt. percent and the standard deviation was
0.009 wt. percent. Only two samples from this assay
set were rejected on the basis of the t-tests at the
99 percent confidence level (see Statistics, Appen-
dix B) .


41
Results of the First Experiment
Assay Set Number Two was rejected for this
summary since seven of the samples boiled over and
the remaining three showed glaze on the outsides of
the crucibles, an indication of possible sample
loss during the fusion step of the assay (see
Table 9) .
The weight percent values of Pd in Assay Set
Nos. 1 and 3 were corrected for loss on ignition since
the samples were found to contain an average of 6%
h2o.
The remaining assay sets gave a return of
0.537 wt. percent palladium for the four assay sets,
after the rejection of outliers (see Table 9). The
standard deviation for these sets was 0.0241 wt.
percent.
Some sample from the assays was lost during
the fusion step; the assays were inspected during
this process and rapid vaporization and boiling
occurred in 21 of the 40 assays. The controllability
of the fusions was a problem in spite of the sample
preparation, the changes in flux, time, and tempera-
ture. Although Assay Sets Nos. 4 and 5 showed
improved controllability, there was too much boiling
during the fusions, resulting in possible sample loss.


Table 9.Results of the first experiment
Set no. Sample wt. (mg) Dord bead wt. (mg) Silver inquart wt. (mg) Pb Correction Wt. (mg) Palladium wt. (mg) Palladium wt. %
1 189322.0 1014.7 31.5 -0- 9383.2 0.5151
3 204169.0 1338.4 28.8 178.0 1131.6 0.5542
4 175002.0 1010.6 28.9 -0- 98o.7 0.5609
5 233336.0 1652.8 152.3 294.0 1206.5 0.5170
Aver. 200457.2 1254.1 60.3 118.0 1075.7 0.0240
Note: Assay Set No. 2 was rejected since the assays failed.


43
The discovery of lead in the dord beads,
particularly in those assays which were washed with
litharge, complicated the gravimetric determination
of palladium: the lead contents had to be determined
before the percent palladium could be determined for
those assays.
All of the fusion temperatures for the first
experiment were between 1060C and 1095C. The
crucibles from these assays were inspected after the
last assay set and over 40 percent of them had been
attacked by the slags; it is commonly known that
strong basic slags attack clay assay crucibles. The
combination of the litharge and the alumina appears
to have acted as a strong base at those fusion
temperatures. The behavior of alumina in slags
appeared to change with the temperature; as the
temperature increased, the alumina acted more like
a base in the slags. Bugbee (1940) addressed the
problem of alumina in slags as follows:
Metallurgists have never entirely agreed about
the behavior of alumina in slags; some regard
it as an acid, others as a base, and ye~ other
authorities believe that it remains uncombined
with any other oxide in slags, and therefore
exists as a neutral substance held either in
solution or in suspension in the liquid slag.


CHAPTER III
THE FINAL EXPERIMENT
The Objectives of the Final Experiment
The fusions had to become controllable in
order to maximize the recovery of palladium in the
assays: the correct flux, fusion time, and fusion
temperature had to be found.
The cupellation temperature ranges needed
to be adjusted in order to help minimize the amount
of lead locked in the dore beads and in order to
avoid sample loss by spitting during this process.
Experimental Methods for the
Final Experiment
The experimental methods given for the first
experiment were the same for the final experiment
with the following exceptions:
1. All of the samples were dehydrated at
900C for one hour and then cooled to
room temperature in desiccators.
2. The dore beads from each assay were
analyzed by AA spectrophotometry and
Pb corrections were made.


45
3. Lower fusion temperatures were used
for each assay set in order to reduce
the probability that the alumina would
act as a strong base.
4. Litharge in the fluxes was reduced in
order to reduce the probability that
the slags become too basic.
5. The silver inquart weights were increased
over ten times in order to improve the
collection of palladium in the lead
buttons during the fusion process.*
6. The samples were reduced to 2916.7 0.5
mg (1/10 assay ton) in order to increase
the controllability of the fusions.
7. About 79 grams of flux were used per
one-tenth assay ton sample (2.9167 grams).
8. Twenty-gram clay assay crucibles were used
for each assay instead of 30-gram assay
crucibles.

A number of assayers, including Fulton,
circa 1911, preferred to use high quantities of
silver inquart in the crucible charges in order
to eliminate palladium from the slags and hold it
throughout the cupellation process. Fulton recom-
mended a ratio of ten to one, ten parts of silver
to one part of palladium expected in the assay
(Fulton and Sharwood, 1929).


46
Assay Set Number Six
The assay set was prepared following the pro-
cedures given in the experimental methods section.
In order to counteract the strongly basic
behavior of the flux and samples at fusion tempera-
tures, a flux was calculated that would maintain a
bisilicate slag. This yielded a flux which con-
tained 45 grams of litharge, 24 grams of soda ash,
7.5 grams of borax glass, 1.5 grams of silica, and
1.2 grams of flour. This was labeled the #4-0x flux
(see Flux Table, Appendix D).
The fusion temperature was lowered to 1030C
and the fusion was allowed 30 minutes time. The
cupellation furnace was started at 1000C and stopped
at 850C. The cupellation temperatures were set
according to estimates of the best performances from
the previous five assay sets.
The average palladium yield from this assay
set was 0.4988 wt. percent and the standard deviation
was 0.0393 percent. Four of the ten values from
the assays were rejected on the basis of the t-tests
at the 99 percent confidence level (see Table 10, and
Statistics, Appendix B).
The examination of the slags near the inter-
faces of the slags with the lead buttons revealed some
palladium.


Table 10.Assay set number six
Sample No. Sample Wt. (mg) Dor6 Wt. (mg) Ag Inquart W t.(mg J Pb Correction* W L.(mg) Pd in Sample Wt.(mg) pd %
51 2916.7 50.4 32 5.0 17.9 .6137*
52 N 47.3 32 4.0 14.9 .5109
53 M 47.8 37 4.0 10.4 .3566*
54 N 47.4 34 4.0 13.0 .4457
55 M 52.8 36 5.0 16.3 .5589
56 M 62.7 34 10.9 17.8 .6103*
57 H 52.2 37 5.0 14.7 .5040
58 H 48.9 37 4.0 11.5 .3943*
59 H 49.0 35 4.0 13.6 .4663
60 M 57.3 42 5.0 14.8 .5074
Statistics:
Highest % I'd in Sample: 0.6137 Lowest % Pd in Sample: 0.3566
Mean % Pd in Sample: 0.4852 Average % Pd in Sample: 0.4988
Standard Deviation: 0.0393
IValue nos. 51, 53, 56, 58 rejected at the 99% confidence level based on t-tests.
Conditions:
Date of Run: 4-83 Sample Description: 1) Pd in Alumina, Petroleum Refining
Catalyst
2) Dehydrated at 900C to 600C
1 hour, cooled, weighed out.
flux: 4 Ox. Furnace Temperature: 1030C Fusion Time: 30 minutes
Cupellation Temperatures: Start 1000C, Stop 850C
Slag and Crucible Analysis: pd in slag near interface of lead button and slag.
Remarks Concerning Assay: Used high %Ag Inquart, Pb in each sample assayed
by A-A and washing. Each sample nearly boiled
over. Used 20g Assay Crucibles.


48
The fusion charges boiled but did not boil
over.
Assay Set Number Seven
Since the previous assays boiled and there
was some palladium discovered in the slags, a com-
ponent of the flux was changed in order to lower the
temperature at which the charge would become liquid
during the fusion. Potassium carbonate was
exchanged for sodium carbonate (soda ash). This
allowed the fusion to be run at lower temperature:
the boiling activity would then be less vigorous
than in previous assay sets. The flux with potassium
carbonate was labeled the #5-0x flux (see Flux Table,
Appendix D).
The fusion temperature was set to 975C and
the time was increased to 35 minutes for this assay
set. The additional 5 minutes of time was to allow
the palladium sufficient time to fall into the lead
buttons at the bottoms of the crucibles.
The average palladium yield from this
assay set was 0.5262 wt. percent and the standard
deviation of the samples was 0.0199 wt. percent.
Only two of the samples were rejected from this
assay set using the t-tests at the 99 percent


49
confidence level (see Table 11 and Statistics,
Appendix B).
There was no palladium found in the slags of
these assays.
Assay Set Number Eight
The same procedures were followed in this
assay set as in Assay Set Number 7 except that the
starting cupellation temperature was lowered to 950C
in order to avoid spitting from the sample during
the early part of the cupellation process, as was
witnessed in the cupellations of the previous assay.
The average percent palladium recovered from
this assay set was 0.5412 wt. percent and the standard
deviation was 0.0130 wt. percent. Three values were
rejected from this set based on t-tests at the 99 per-
cent confidence level (see Table 12 and Statistics,
Appendix B).
Assay Set Number Nine
The same procedure was followed in this assay
as in the previous assay set. There was a power loss
at the furnace after 25 minutes of fusion time had
elapsed.
The average palladium returned from this
assay set was 0.5304 wt. percent and the standard


Table 11.Assay set number seven
Sample No. Sample Wt. (mg) Dor 6 Wt. (mg) Ag Inguart Wt. (mg) Pb Correction W t. (mg) Pd i n Sample Wt.(mg) Pd %
61 2916.7 50.4 34.7 0 15.7 .5383
62 M 52.1 34.0 1.6 16.5 .5657
63 M 49.0 30.0 4.3 14.7 .5040
64 M 48.7 33.7 0 15.0 .5143
65 M 41.0 21.8 0 19.2 .6583*
66 M 56.7 34.2 3.8 18.7 .6411*
67 M 51.5 34.9 1.4 15.2 .5211
68 M 49.4 34.0 0.4 15.0 .5143
69 M 50.2 34.0 1 .2 15.0 .5143
70 M 61.2 34.0 11.5 15.7 .5383
Statistics: Highest % Pd in Sample: 0.6583 Lowest % Pd in Sample: 0.5040
Mean % Pd in Sample: 0.5812 Average % Pd in Sample: 0.5262
Value nos. 65 and 66 rejected at Standard Deviation: Die 99% confidence level based 0.0199 on t-tests.
Condi Lions:
bate of Run: 4-83 Sample Description:
Sample Lot No.: EL-990176
flux: 5 Ox. Furnace Temperature
1) Pd in Alumina, Petroleum Refining
Catalyst
2) Dehydrated at 900c to 600C
1 hour, cooled, weighed out.
975 C Fusion Time: 35 minutes
Cupellation Temperatures: Start 950C, Stop 825C
Slag and Crucible Analysis: No Pd
Remarks Concerning Assay: Used 20g Assay Crucible, Dor6 beads are gravimeLrically
determined Lhen decomposed for Pd, Pb, and Ag determination.
tn
O


Table 12.Assay set number eight
Sample No. Sample Wt. (mg) Dor 6 Wt. (mg) A9 Inquart Wt.(mg) Pb Correction Wt. (mg) Pd in Sample W t. (mg) cu 1
71 2916.7 53.1 34.9 8.0 17.4 .5966*
72 M 54.4 34.9 7.0 18.8 .6446 *
73 H 71.0 51.0 4.7 15.3 .5246
74 M 54.6 34.5 4.4 15.7 .5383
75 M 47.0 32.7 0 14.3 .4903 *
76 H 60.3 34.9 10.0 15.4 .5280
77 M 51 .7 34.9 8.0 16.0 .5486
78 N 57.2 34.0 7.5 15.7 .5383
79 H 51.4 34.9 5.0 16.0 .5486
80 H 50.0 33.6 0 16.4 .5623
Statistics: Highest 4 Pd in Sample: 0.6446 Lowest 4 Pd in Sample: 0.4903
Mean 4 Pd in Sample: 0.5675 Average 4 Pd in Sample: Standard Deflation: 0.5412 0.0130
value nos. 71, 72, and 75 rejected at the 994 confidence level based on t-tests.
Condi Lions:
Date of Run: 5-83 Sample Description: 1)
Sample Lot No. EL-990176
2)
l'lux: 5 Ox. Furnace Temperature: 975C
Cupellation Temperatures: Start 950C, Stop 825C
Slag and Crucible Analysis: No Pd
Remarks Concerning Assay: Used 20g Assay Crucible, Dor6 beads are gravimetrically
determined then decomposed for Pd, Pb, and Ag determination.
Pd in Alumina, Petroluem Refining
Catalyst
Dehydrated at 900C to 600C
1 hour, cooled, weighed out.
Fusion Time: 35 minutes


52
deviation was 0.0288 wt. percent. Three values were
rejected on the basis of the t-tests at the 99 per-
cent confidence level (see Table 13 and Statistics,
Appendix B).
Assay Set Number Ten
The procedure for the tenth assay set did not
vary from the previous assay except that the higher
silver inquart weights were used in order to once
again attempt to discover the relationship between
the weight of the silver inquarts and the percent
recovery of palladium. At no time during this final
experiment did the values of the silver inquarts
reach the values suggested by Fulton (Fulton and
Sharwood, 1929), since these inquarts would have
reached over 150 mg in weight and this weight of
silver in the dore bead normally brings about an
excess of spitting from the sample during the cupel-
lation process. Therefore, weights of silver inquarts
of approximately one-half the weights suggested by
Fulton were used. The high silver inquart weights
did not noticeably alter the quantities of palladium
recovered from this assay set (see Table 14). But
the uniformity and low values of the Pb corrections
should be noted.


Table 13.Assay set number nine
Sample No. Sample Wt. (mg) Dore Wt. (mg) Ag Inquart Wt.(mg) Correction Wt.(Mg) Pd in Sample Wt.(mg) Pd %
81 2916.7 59.6 34 3 12.5 12.8 .4389*
82 M 48.5 34.0 4.0 14.1 .4834
83 M 53.1 34.1 3.3 15.7 .5383
84 H 52.5 34.3 2.5 15.7 .5383
85 M 53.8 34.0 3.4 16.4 .5623
86 H 53.7 34.0 3.3 16.4 .5623
87 M 54.5 34.0 3.8 16.7 .5726*
88 H 52.8 35.9 1.7 15.2 .5211
89 M 60.9 34.9 11.2 14.8 .5074
90 M 74.1 62.1 0 12.0 .4114*
Statistics:
Highest % Pd in Sample: 0.5726 Lowest % Pd in Sample: 0.4114
Mean % Pd in Sample: 0.4920 Average Pd in Sample: 0.5304
Standard Deviation: 0.0288
Value nos 81, 87 , and 90 rejected at the 99% confidence level vased on t-tesi
Conditions:
Date of Run: 6-83 Sample Description: 1) Pd in Alumina, Petroleum Refining
Sample Lot No.: GL-990176 Catalyst
2) Dehydrated at 900C to 600C
1 hour, cooled, weighed out.
flux: 5 Ox. Furnace Temperature: 975C Fusion Time: 35 minutes
Cupellation Temperatures: Start 950C, Stop 825C minutes(81,82
r 89,90)
Slag and Crucible Analysis: No
Remarks Concerning Assay: Used 20g Assay Crucible, Dore beads are gravimetrically
determined then decomposed for Pd, Pb, and Ag determination.
Power loss at 25 minutes, sample Nos. 81, 82, 89, 90 affected.


Table 14.Assay set number ten
Sample Dor6 Ag Pb Pd in
Sample Wt. WL. Inquart Correction Sample Pd
No. (mg) (mg) W t. (mg) ML.(mg) Wt.(mg) %
91 2916.7 89.4 72.5 1.2 15.7 .5383
92 M 91.9 75.8 1.2 14.9 .5109
93 M 86.4 70.0 1.1 15.3 .5246
94 M 97.3 81.5 1.3 14.5 .4971
95 N 98.7 83.9 1.3 13.5 .4629*
96 M 80.5 62.5 1 1 16.9 .5794*
97 95.1 78.4 1.2 15.5 .5314
98 M 89.6 72.5 1.2 15.9 .5451
99 H 82.8 64.8 1 1 16.9 .5794*
100 M 87.4 70.0 1.2 16.2 .5554
Statistics :
Highest Pd in Sample: 0.5794 Lowest % Pd in Sample: 0.4629
Mean % Pd in Sample: 0.5212 Average % Pd in Sample: 0.5289
Standard Deviation: ( ). 0200
Value nos . 95, 96 , 99 rejected aL the 99% confidence level based on t-tests.
Conditions :
Date of Run: 6- 83 Sample Description: 1) Pd in Alumina , Petroleum Ref:
Sample I.ot No.: GL-990176 2) Catalyst, Dehydrated at 900C to 600C
1 hour, cooled, weighed out.
Flux: 5 Ox. Furnace Temperature: 975C Fusion Time: 35 minutes
Cupellation Temperatures: Start 950C, Stop 825C
Slag and Crucible Analysis: No Pd
Remarks Concerning Assay: Used 2og Assay Crucible, Dor£ beads are gravimetrically
determined then decomposed for Pd, Pb, and Aq determination.


55
The average palladium yield from this assay set
was 0.5289 wt. percent and the standard deviation
was 0.0200 wt. percent. Three sample values were
rejected on the basis of the t-tests at the 99 per-
cent confidence level (see Statistics, Appendix B).
Results of the Final Experiment
Assay Set Number 6 was the last assay set to
have vigorous boiling problems and the loss of samples
which accompanies such activity in the crucibles.
Assay Sets 7, 8, 9, and 10, which used I^CO-j instead
of Na2CC>2 as in previous assays, boiled but not
vigorously enough so that the charges reached the
tops of the crucibles. Thus, there was no apparent
loss of sample from these assays. The fusions were
more controlled with the potassium carbonate flux.
The cupellations for the last four sets of assays
showed no noticeable spitting. Therefore, the cupel-
lation temperatures were satisfactory.
The average yield of palladium was 0.525 wt.
percent for all five assay sets and the standard
deviation was 0.016 wt. percent. The relative
error of the combined assay sets was 3.1 percent
(see Table 15). The high relative error and the


Table 15.--Results of the final experiment
Set no. Sample wt.(mg) Dore bead wt. (mg) Ag inquart wt. (mg) Correction wt. (mg) Palladium wt. (mg) Mean percent of each set
6 17500.2 306.0 216.0 2.7 87.3 0.4988%
7 23333.6 412.5 269.3 20.4 122.8 0.5262%
8 20416.9 396.2 257.8 27.9 110.5 0.5412%
9 20416.9 375.3 241.2 25.8 108.3 0.5304%
10 20416.9 637.1 520.7 8.4 108.0 0.5289%
Aver. 20416.9 425.4 301.0 17.0 107.3 Std. Dev. 0.5318% = 0.0066
U1


57
low of palladium were due to the low values from
Assay Set Number 6. This set was rejected for
the summary on the basis of the t-tests (see
Statistics, Appendix B and Table 15)


CHAPTER IV
RESULTS AND CONCLUSIONS
OF THIS RESEARCH
The Experimental Data
The mean palladium yield for the nine assay
sets was 0.5305 percent 0.0195 wt. percent? Assay
Set Number Two was rejected owing to the failure of
the fusion process, not based on statistical
analysis (see Table 16).
Based on statistical analysis of the data of
the nine assay sets, sets number 1, 3, 4, 5, and 6
were rejected (see Table 17). The palladium yield
was 0.5317 wt. percent and the standard deviation
was 0.0066 wt. percent for the remaining assay sets.
The relative error, based on the best known value of
the samples of 0.542 wt. percent, was 1.9 percent
(see Table 17). The rejection of assay sets 3, 4, 5,
and 6 was based on t-tests of the data at the 95 per-
cent confidence level. Assay Set 1 was rejected on
the basis of an "F" test which revealed that its
standard deviation was significantly different from
the standard deviations of assay sets 7, 8, 9, and 10.


Table 16-Tabulation of results from the ten assay sets
Assay set no. Text page no. Sample wt. (mg.) Do re wt.(mg) Ag wt. (mg) Pb Correction wt. (mg) Pd wt. (mg) Av.Pd percent
1 25 189322.0 1014.7 31.5 -0- 983.2 0.5151
2 28 87501.0 446.2 12.5 -0- 433.7 0.4956*
3 30 204169.0 1338.4 28.8 178.0 1131.6 0.5542
4 33 175006.0 1010.6 28.9 -0- 981.7 0.5609
5 36 233336.0 1652.8 152.3 294.0 1206.5 0.5170
6 46 17500.2 306.0 216.0 2.7 87.3 0.4988
7 48 23333.6 412.5 269.3 20.4 122.8 0.5262
8 49 20416.9 396.2 257.8 27.9 110.5 0.5412
9 49 20416.9 375.3 241.2 25.8 108.3 0.5304
10 52 20416.9 637.1 520.7 8.4 108.0 0.5289
Mean percent of the assay sets = 0.5303% Standard deviation = 0 .0195%
*Assay Set No. 2 was rejected due to failure of the fusion process (see Table 5 )
U1
VO


60
Table 17.Statistical tabulation of
results from the ten assay sets
Assay set no. Mean percent palladium Number of samples Standard deviation
1 0.5151 7 0.0392**
2 0.4956 3 0.0561***
3 0.5542 7 0.0293*
4 0.5609 6 0.0219*
5 0.5170 8 0.0090*
6 0.4988 6 0.0393*
7 0.5262 8 0.0199
8 0.5412 7 0.0130
9 0.5304 7 0.0288
10 0.5289 7 0.0200
Sets 7, 9 and 10 8, 0.5317 4 0.0066
*Sets 3, 4, 5, and 6 are rejected on the basis of
t-tests at the 95% confidence level; the mean
is significantly different than the known
value, 0.542 wt. percent.
**Set 1 is rejected on the basis of the F test; its
standard deviation is significantly different
than those of sets 7 and 8, 9 and 10.
***Set 2 was rejected since 7 of the fusions failed
and the remaining 3 fusions may have frothed
excessively, losing sample.


61
Comparisons of means and standard deviations
are computed at the 95 percent confidence level
(see Statistics, Appendix B).
Noteworthy Observations
The dord beads of Assay Sets 6 through 10
were dissolved in aqua regia and brought up to
volume in 100-ml volumetric flasks and analyzed
against standards on the AA spectrophotometer
for the contents of lead. The concentration of
palladium was also measured on Assay Sets 7
through 10. The results of these analyses
are given below in terns of mg average palladium
recovered per assay set. (15.80 mg. Pd should
be recovered per 1/10 assay ton since the best
known concentration was 0.542 percent Pd.)
Comparative graphs of fire assay results
and atomic absorption results are given in
Figure 8. The graph shows the average palladium
recovered in milligrams versus the assay sets
for each method of analysis. The fire assay
results are usually higher since the results are
apparent gravimetric values of Pd.


Avg. mg. Palladium Recovered
62
Figure 8- A comparative graph of fire
assay results versus Atomic
Absorption Spectrophotometry
Results
F.A. Fire assay
- Atomic absorption
AA.


63
Table 18. Average palladium per assay set recovered
Assay set no. Average Pd* fire assay Average Pd* AA spec.
7 16.07 15.00
8 16.10 15.75
9 14.98 14.50
10 15.5 3 15.55
*
Average palladium is given in milligrams.
Note: Outliers are omitted in this table.
Time and temperature relationships were
intimately related in the assay experiments. There
was a range of time and temperature values in the
last four assays within which the fusions were
more controllable than at any other time during
the experiments. This range was approximated by
multiplying the time for the fusion by the average
temperature. It was estimated that the range
factor, time x temperature, had limits between 20
and 60 minutes of time and between 875C and 1020C.
Beyond these limits the effects of the potassium
carbonate in the flux would render the fusions incom-
plete or uncontrollable (see Table 19). It was of
some value to note the relationship between the


64
Table 19. Time x temperature
factor in fusion
Set No. Average Fusion Temperature Time Factor*
1 1075C 40 minutes 43.000 X 103
2 1095C 45 minutes 49.275 X 103
3 1080C 35 minutes 37.800 X 103
4 1060C 40 minutes 42.400 X 103
5 1080C 40 minutes 43.200 X 103
6 1030C 30 minutes 30.900 X 103
7 975 C 35 minutes 34.125 X 103
8 97 5 C 35 minutes 34.125 X 10 3
9 975 C 30 minutes 29.250 X 103
10 975 C 35 minutes 34.125 X 103
*Dimension of the Factor is c x minutes.
Note: Fusion time of less than 20 minutes or
more than 60 minutes can
change the end product. So, these are
the lower and upper limits for this
fusion process. Fusion temperatures
less thanS75C or more than 1020 C
also change the end product.


65
relative errors of the assay sets and the time x
temperature factor in estimating the outcome of the
assays (see Figure 9). Eighty percent of the assay
sets within the dashed boundary (see Figure 9) were
easily controllable and provided data with relative
errors between 0.1 percent and 3 percent.
Conclusions
Greater than 98 percent recovery of palladium
(based on the best known value of 0.542 percent pal-
ladium in the alumina matrix) was achieved in the
final experiment. The accuracy for the determination
method used in the last four assay sets was 1.9
percent. Two and one-half percent relative error
was the best achieved by standard wet chemical methods
for the determination of palladium in the alumina
matrix (Hoffner, 1982).
The potassium carbonate flux used in the
last four assay sets helped to achieve controllability
of the fusions so that palladium was no longer apparent
in the slags from the samples,
These experiments were aimed at deriving
pure palladium and silver dore beads without the
presence of lead. This was not achieved. Some


66
30 35 40 45 50
Temperature x Time Factors x 10^
Temperature x time function x 103. Numbers
in parentheses are assay set numbers.
Figure 9.


67
lead-palladium complexing was recognized in the dore
beads which necessitated the use of AA spectrophoto-
metric analyses of the dole beads for the determin-
ation of palladium, silver, and lead. The latter
analyses were simple since only the metallic dore
beads had to be digested.
The fire assaying procedures used in the last
four assay sets gave nearly the same economy as any
classical fire assaying method with the added bene-
fits of controllability and accuracy, even with the
alumina matrix. The average cost per sample was $7.50,
including the AA analyses. There was no other method
for the determination of palladium in an alumina
matrix which had the same cost effectiveness at the
time of this research.
The following procedures are recommended in
summary:
1. One-tenth assay ton (2.9167 grams) of
palladium sample (in alumina or similar
matrix) with 0.030 grams of silver
inquart and 79 grams of #5 Ox flux (see
Appendix D, Fluxes) are fused at a furnace
temperature of 975C for 35 minutes.
2. The lead button derived from this fusion
is separated from the slag and cupelled


68
at an initial temperature of 950C and a
finishing temperature of 825C in order
to obtain the dore bead (see Appendix G).
3. The dord bead is weighed out after cool-
ing and the silver inquart is subtracted
to give the apparent weight of palladium
in the sample.
4. The dord head is digested in dilute, hot
aqua regia and brought up to volume in a
100-ml. volumetric flask. The pH is
maintained at 1 or less.
5. An aliquot solution is diluted in order to
bring the concentration of the palladium
within the linear working range of the AA
spectrophotometer.
6. The concentration of the palladium is
measured on the A.A. spectrophotometer
and this value, yg/ml, is converted to ppm.
The palladium content in the original
sample can now be determined.


BIBLIOGRAPHY
Analytical methods for atomic absorption spectro-
photometry, 1973: Norwalk, Connecticut,
Perkin Elmer Corp.
F. E. Beamish, 1966, The analytical chemistry of
the noble metals: New York, Pergamon Press,
pp. 109-233.
Beazley, Phillip M., 1982-1983, interviewed by Michael
P. Thomas regarding the possibilities of applying
fire assaying in the determination of palladium in
refining catalysts and the problems with palladium
determination via wet chemistry: Littleton,
Colorado, Marathon Denver Research Center, May
16, 1982; October 20, 1982; November 16, 1982;
January 15, 1983; June 14, 1983.
Bugbee, Edward E. 1940, A textbook of fire assaying,
3rd ed.: Golden, Colorado, Colorado School of
Mines Press (1981), pp. 4-19, 120-131, 260-280,
300-305.
Considine, Douglas M., ed., 1976, Van Nostrand's
scientific encyclopedia, 5th ed.: New York,
Van Nostrand Reinhold Co.
Crook, Welton J., Applications for the Vreeland
Direct Reading Spectroscope and Spectrograph:
Redwood City, California, Spectrex Corp.
Currie, L. A., Filliben, J. J., and DeVoe, J. R.
(April 1972), Statistical and mathematical methods
in analytical chemistry, Analytical Chemistry,
v. 44, no. 5, pp. 497R-507R.
Davis, John C., 1973, Statistics and data analysis
in geology: New York, John Wiley and Sons Co.,
Inc.


70
Faye, G. Y., and Inman, W. R., 1961, The determination
of platinum and palladium in copper-nickel matte
and ore concentrates, Analytical Chemistry, v. 33,
no. 2, pp. 278-283.
Fraser, J. G., and Beamish, F. E., 1954, Fire assay
for palladium, Analytical Chemistry, v. 26, no. 9.
pp. 1474-1477.
Fraser, J. G., and Beamish, F. E., 1954, Separation
of palladium from lead and the colorimetric
determination of palladium with potassium iodide,
Analytical Chemistry, v. 26, no. 3, pp. 495-498.
Fulton, C. H., and Sharwood, W. J., 1929, A manual of
fire assaying, 3rd ed.: New York: McGraw-Hill
Co.
Gerlach, Dr. W., and Schweitzer, Dr. Eugen, 1929,
Die chemische emissionspectralanalyse: Munich:
Physical Institute of the University of Munich.
Hach Chemical Company, 1977, water analysis handbook,
expanded ed., Ames, Iowa.
Haffty, Joseph, and Riley, L. B., 1968, Determination
of palladium, platinum, and rhodium in geologic
materials by fire-assay and emission spectrography,
Talanta, v. 15, pp. 111-117.
Havinga, J., 1973, Less-common metals: Eindhoven,
Netherlands, Philips Research Laboratories.
Hoffman, E. L., Naldrett, A. J., Van Loon, J. C.,
Hancock, R. G. V., and Manson, A., 1978, The
determination of all the platinum group elements
and gold in rocks and ore by neutron activation
analysis after preconcentration by nickel sulfide
fire assay technique on large sample sizes,
Analytica Chimica Acta 102, pp. 157-166.
Hoffner, Sonia, E., 1982-1983, interviewed by Michael
P. Thomas regarding the problems with palladium
determination in refining catalysts: Littleton
Colo.: Marathon Denver Research Center, October
20, 1982; November 16, 1982; January 15, 1983;
June 14, 1983.


71
Krauskopf, Konrad B., 1967, Introduction to geo-
chemistry: New York, McGraw-Hill.
Natrella, Mary Gibbons, 1963, Experimental statistics;
NBS Handbook 91, reprinted October 1966: U.S.
Government Printing Office, Washington, D.C.,
pp. 2-1 to 2-15, 3-1 to 3-40.
Page, Norman J., Riley, L. B., and Haffty, Joseph,
1969, Platinum, paladium and rhodium analyses
of ultramafic and mafic rocks from the Stillwater
Complex, Montana, U.S. Geological Survey Circular
624, pp. 1-12.
Plummer, M. E. V. and Beamish, F. E., 1959, Determi-
nation of platinum and palladium in ores and
concentrates, Analytical Chemistry, v. 31, no. 12,
pp. 1141-1143.
Plummer, M. E. V. Lewis, C. L., and Beamish, F. E.,
1959, Fire assay for platinum and palladium in
ores and concentrates, v. 31, no. 12, pp. 254-258.
Scanlon, James, 1979, interviewed by Michael P.
Thomas regarding fluxes for fire assaying: Golden,
Colorado, Colorado School of Mines Research Insti-
tute.
Scott, Wilfred W., ed., 1925, Standard methods of
chemical analysis, 4th ed., vol. 1: New York,
D. Van Nostrand Co.
Seath, John, and Beamish, F. E., 1940, Assay for
platinum metals in ore concentrates, Industrial
and Engineering chemistry, v. 12, no. 3, pp. 169-
171.
Seath, John, and Beamish, F. E., 1953, Spectrographic
examination of assay beads for platinum, palladium,
and gold, Analytical Chemistry, v. 25, no. 6,
pp. 980-981.
Shirley, Russell L., 1983, interviewed by Michael P.
Thomas regarding fire assaying of palladium in
neutral, acid ores including alumina: Wheat-
ridge, Colorado, Auro-Tech Assaying Co.,
January 18, 1983; March 20, 1983.


72
Smith, 0. C., 1953, Identification and qualitative
chemical analysis of minerals, 2nd ed.: New
York: Van Nostrand co.
Smith, S. L., 1979, Fire assaying for gold and
silver: Tucson, Arizona, Jacobs Assay Office.
Weast, Robert C., ed., 1973, Handbook of physics
and chemistry, 54th ed.: Cleveland, Ohio,
Chemical Rubber Co.


APPENDICES


APPENDIX A
RESULTS TABULATION


Table A-l.Tabulation of Results from the Ten Assay Sets
Assay set Text page no. Sample wt. (mg) Dore wt.(mg) Ag wt.(mg) Pb Correction wt. (mg) Pd wt. (mg) Av. Pd percent
1 25 189322.0 1014.7 31.5 -0- 983.2 0.5151
2 28 87501.0 446.2 12.5 -0- 433.7 0.4956*
3 30 204169.0 1338.4 28.8 178.0 1131.6 0.5542
4 33 175006.0 1010.6 28.9 -0- 981.7 0.5609
5 36 233336.0 1652.8 152.3 294.0 1206.5 0.5170
6 46 17500.2 306.0 216.0 2.7 87.3 0.4988
7 48 23333.6 412.5 269. 3 20.4 122.8 0.5262
8 49 20416.9 396.2 257.8 27.9 110.5 0.5412
9 49 20416.9 375.3 241.2 25.8 108.3 0.5304
10 52 20416.9 637.1 520.7 8.4 108.0 0.5289
Mean percent of the assay sets = 0 .5303% Standard deviation = 0.0195%
k Assay Set No . 2 was rejected due to failure of the fusion process (see Table
5)


APPENDIX B
STATISTICAL TESTS ON EXPERIMENTS
F TESTS, T-TESTS


Table B-l.F Tests* on Standard Deviations**
Assay set -* ** 1 1 2 3 4 5 6 7 8 9 10
1 2.048 1.790 3.204 18.970 1. 005 3.880 9.092 1. 853 3.842
2 2.048 3.666 6.562 38.854 2.037 7.947 18.623 3.794 7.864
3 1.7 90 3.666 1.790 10.599 1.799 2.168 5.080 1.035 2.146
4 3.204 6.562 1.790 16.810 1.134 3.438 8.054 1.642 3. 404
5 18.970 34.954 10.599 16.810 19.068 4.889 2.086 10.240 4.938
6 1.005 2.037 1.799 1.134 19.068 3.900 4.139 1.862 3.861
7 3. 840 7.947 2.168 3. 458 4.889 3.900 2.343 2.094 1.010
8 9.092 18.623 5.080 8.057 2.086 9.139 2.343 4.908 2.364
9 1. 853 3.794 1.035 1.642 10.240 1.862 2.094 4.908 2.074
10 3.842 7.968 2.146 3.404 4.938 3.861 1.010 2.367 2.074
*Computed at the 95% confidence level.
**Comparative variability of performance of s.


Table B-2.T'Values*
Assay
set 1 1 2 3 4 5 6 7 8 9 10
1 0.0195 0.0391 0.0758 0.0019 0.0163 0.0111 0.0261 0.0153 0.0138
.0988 .0399 .0456 .0358 .0472 .0371 .0360 .0397 .0371
2 0.0195 0.0586 0.0653 0.0214 0.0032 0.0306 0.0456 0.0348 0.0233
. 0988 . 1089 .0992 . 1400 . 1003 . 1426 .1410 . 1084 . 1431
3 0.0391 0.0586 0.0067 0.0372 0.0554 0.0280 0.0130 0.0238 0.0253
. 0399 . 1089 .0412 . 0273 . 0429 .0286 .0274 .0333 . 0292
4 0.0458 0.0653 0.0067 0.0439 0.0621 0.0347 0.0197 0.0305 0.0320
.0456 .0992 .0412 .0377 .0480 .0383 .0388 .0709 . 0389
5 0.0019 0.0214 0.0372 0.0439 0.0182 0.0092 0.0242 0.0134 0.0119
.0358 . 1400 .0273 .0377 .0400 .0170 .0128 .0268 .0186
6 0.0163 0.0032 0.0854 0.0621 0.0182 0.0274 0.0424 0.0316 0.0301
.0472 . 1003 .0429 .0480 .0400 . 0404 .0411 .0427 .0409
7 0.0111 0.0306 0.0280 0.0347 0.0092 0.0271 0.0150 0.0042 0.0027
.0371 .1426 .0286 .0383 .0170 . 0404 .0184 .0282 . 0220
8 0.0261 0.0456 0.9130 0.0197 0.0242 0.0424 0.0150 0.0108 0.0123
.0360 . 1410 .0274 .0388 . 0128 .0411 .0181 . 0270 .0196
9 0.0153 0.0348 0.0238 0.0305 0.0134 0.0316 0.0042 0.0108 0.0015
.0397 . 1087 . 0333 . 0409 . 0268 .0927 . 0242 .0270 .0289
10 0.0138 0.0333 0.0253 0.0310 0.0119 0.0301 0.0027 0.0123 0.0015
. 0371 .1431 .0292 .0389 .0186 . 0409 .0220 .0196 .0289
lXA - XBI > u Legend: |A X| * lXA " xbI difference in means
u u = u for comparison
Computed at the 95% confidence level.


APPENDIX C
A COMPARISON OF FIRE ASSAY RESULTS WITH
A. A. SPECTROPHOTOMETRY RESULTS
FROM ASSAY SET NOS. 7, 8, 9, 10


80
Table C-l.A Comparative Study of the Dore Bead Fire Assay Results versus A-A Spectrophotometer Results
Sample No. Fire Assay Results Dord Wt. (mg) A-A Spectrophotometer Results Dore Wt. (mg)
61 15.7 15 .5
62 16.5 16 .0
63 14.7 14 .0
64 15.0 14 .5
65 19.2 13 .5
66 18.7 19 .0
67 15.2 14 .5
68 15.0 14 .0
69 15.0 14 .5
70 15.7 14 .5
71 17.4 17 .5
72 18.8 17 .5
73 15.3 14 .5
74 15.7 16 .0
75 14.3 13 .5
76 15.4 14 .0
77 16.0 15 .0
78 15.7 17 .0
79 16.0 16 .0
80 16.4 16 .5
81 12.8 12 .0
82 14.1 14 .0
83 15.7 16 .0
84 15.7 16 .0
85 16.4 16 .5


81
Table C-l (Continued)
Fire Assay Results versus A-A Spectrophotometer Results
A-A
Spectrophotometer
Sample No. Fire Assay Results Dord Wt. (mg) Results Dord Wt. (mg)
86 16.4 16.5
87 16.7 17.0
88 15.2 16.0
89 14.8 14.5
90 12.0 6.5
91 15.7 16.0
92 14.9 15.0
93 15.3 15.5
94 14.5 14.0
95 13.5 14.0
96 16.9 17.0
97 15.5 15.0
98 15.9 16.0
99 16.9 17.0
100 16.2 16.0


APPENDIX D
TABLE OF FLUXES USED FOR THE EXPERIMENTS


Table D-1.--Flux Table
Fluxes
Number 1 Flux for Oxides (1 ox.)
Litharge . FbO ...5895
Soda Ash .NagCO^.. . 3095
Borax Glass... ,Na2B47 . .11%
Flour * C6H105' . .195
Number 2 Flux for Oxides (2 ox.)
Litharge FbO . .6995
Soda Ash .Na2C0^.. .. 2595
Borax Glass Na2B1|07. 6%
Number 3 Flux for Oxides (3 ox.)
Litharge .FbO .. .7^.995
Soda Ash .NagCO^.. . .15.0%
Borax Glass... .Na2B^0?. 7.595
Silica .SiOg 1.195
Flour * C6K105 * 1.595
Number k Flux for Oxides ox.)
Litharge . PbO .. .57.095
Soda Ash .NagCO^. . .. .30.095
Borax Glass... . NagB^Or,. 9.595
Silica .SiOg 2.095
Flour c6H105 1.595
Number e, Flux for Oxides (5 ox.)
*Same as Number NagCO^ with Kg U- but replace CO, J *


APPENDIX E
APPARATUS AND MATERIALS


85
Apparatus
1. Thermolyne Model 1620-1 Furnace, 240 Vac.,
1200C maximum
2. Jelrus Model TFA Furnace, 115 Vac., 1100C
maximum
3. Ainsworth Type 10 Analytical Balance, precision
limit to 0.00005 gm
4. Troemmer Analytical Balance (Pulp Balance),
precision limit to 0.0001 gm
5. Vreeland Model 7 Emission Spectrometer, with
quantifying attachments and video camera and
recorder. Readability to less than 1 ppm for
Pd, Ag, and Pb
6. Perkin Elmer Model 103 Atomic Absorption
Spectrophotometer
7. Perkin Elmer Model 290 Atomic Absorption
Spectrophotometer
8. Bausch and Lomb Spectronic-20 Spectrophotometer
9. Bausch and Lomb Spectronic-710 Spectrophotometer
10. Corning Model 610 expanding pH meter
11. Hack Water Analysis Kit and DR/2 Spectrophotome-
ter
12. Thermolyne Model 1400 Furnaces (2), 115 Vac.,
1093C maximum
Materials
1. 30 gm Assay Crucibles, DFC
2. 20 gm Assay Crucibles, DFC
3. 15 gm Assay Crucibles,DFC


86
4. 1H inch sheep bone ash cupels
5. 26-gauge silver wire
6. Kimax 100 ml volumetric flasks
7. Kimax 250 ml Erlenmeyer flasks
8. Teflon beakers, 400 ml
9. Pyrex #7603 1 ml in 1/100 ml pipets
10. Repipet dilutor, 0 to 5 ml
11. Na2CC>3 anhydrous powder, VWR #JT3602-1
12. (Borax glass), VWR #AL61460-7
13. PbO, Litharge, Red Lead Oxide, Pueblo Brand
14. K2C03, VWR #JT3014-1
15. Si02, VWR #AL58550-3
16. Flour, dried wheat flour, Topco
17. 40-dram sample vials
18. 10-dram sample vials
19. HNO^ for trace metal analysis, SW #SC13769
20. HCl, SW #SC12701
21. HF, SW #SC12719
22. HC104 (Lead Free) SW #SC13990
23. Carbon rods, Spectrex Corp.
24. Ceramic hearths, Spectrex Corp.
25. 3000 ASA Polaroid Film, BS
Alumina, A^O^, Ultrapure, SW #SC10552-25KG
26.


APPENDIX F
CHEMICAL STANDARDS


88
Chemical Standards
1. Pd Std., PdClp in dilute HCl, 1000 ppm, Bancc,
Lot #M2-05
2. Pd Std., Pd on Carbon 5%, Eastman Kodak #15267,
Lot #A2A
3. Pb Std., Pb in dilute HNO.,, 1000 ppm, Banco,
Lot #B0-1 J
4. Ag Std., AgN03 in H20, SW #SC16380
5. Pt std., PtCl. in dilute HCl, 1000 ppm, Banco Lot
#Bo-2
6. PtC>2, power, 1 gram, SW #SC14118
7. Ni std., Ni(NO0)2-6H20, crystal SW #SC13744
8. Re std., KReO^, solid in dilute H2SO^, SW 1621
9. Rh std., (NH4)3RhCl, solid, SW. 1620
10. Cu std., Cu in 0.3M HN03, 1000 ppm, SW #16260


APPENDIX G
OUTLINE OF ESSENTIAL STEPS FOR
CLASSICAL FIRE ASSAYING