Journal of Minerals & Materials Characterization & Engineering, Vol. 5, No.1, pp 47-62, 2006
jmmce.org
Printed in the USA. All rights reserved
47
Recovery of Metals from Aluminum Dross and Saltcake
J.Y. Hwang*, X. Huang, and Z. Xu
Institute of Materials Processing, Michigan Technological University
Houghton, MI 49931, USA
*Corresponding author’s e-mail address: jhwang@mtu.edu
ABSTRACT
Various aluminum-smelting by-products from three production sources were received
and characterized. The waste materials were tested for compound identification and
environmental acceptance. A coarse metallic aluminum recovery test using an Eddy
Current separator (ECS) was performed using two different Circuit configurations. White
dross performed equally well with either Circuit, while black dross processing shows
significant difference on the separation results. It was found that ECS technology was
effective for particle sizes down to 6-10 mesh.
Keywords: Eddy current, recycling, aluminum slag
INTRODUCTION
Aluminum is a critical material in the U.S. construction, packaging, and
transportation industries. The aluminum industry produces approximately one million
tons of waste by-products from domestic aluminum smelting annually. The most
significant by-products are called salt cake and dross, and are generated in the smelting
process.
Over the last two decades, aluminum recycling has grown rapidly in terms of both
size and importance to the U.S. economy. Between 1950 and 1974, recycled aluminum
constituted only about 5% of the total domestic aluminum market [1]. Since then, both
the fraction of recycled materials and the total domestic aluminum market have grown
substantially. In January 1997, for example, total aluminum shipments to domestic
markets were 1,591 million lbs., an increase of 12.5% over January 1996 levels. Of this
total, 639 million lbs., or about 40%, was recovered from new and old metallic scrap. In
most applications, recycled aluminum materials perform as well as primary material, and
provide significant savings in both production costs and energy usage.
At present, most aluminum-bearing scrap is recycled through a smelting process.
Although the details of the smelting process differ between various installations, most
involve melting the scrap in the presence of chloride-based slag, generally using either a
reverberatory or rotary furnace. This slag is typically a eutectic or near-eutectic mixture
of sodium and potassium chlorides containing low levels of fluorides (cryolite) or other
additives. It serves two primary functions. First, since the material is molten and fairly
48 J. Y. Hwang, X. Huang, and Z. Xu Vol.5, No.1
fluid at typical aluminum smelting temperatures, the slag coats the metallic aluminum
being melted and minimizes oxidation losses during processing. Second, the presence of
the fluorides and other additives assists in breaking down prior surface oxide layers on
the aluminum charge and promotes improved separation between the aluminum and the
residual nonmetallics in the charge.
The aluminum-bearing scrap for recycle may be either reclaimed metallic aluminum
products (e.g. castings or used beverage containers) or metal-bearing aluminum oxide
drosses skimmed from primary aluminum melting furnaces. Drosses obtained from
primary melting operations (so-called “white drosses“) consist primarily of aluminum
oxide (with some oxides of other alloying elements such as magnesium and silicon) and
may contain from 15 to 70% recoverable metallic aluminum. Drosses from secondary
smelting operations (so-called “black drosses“) typically contain a mixture of
aluminum/alloy oxides and slag, and frequently show recoverable aluminum contents
ranging from 12 to 18%. Commercial smelting of both white and black drosses is often
done in a rotary salt furnace. The nonmetallic byproduct residue, which results from such
dross smelting operations is frequently termed “salt cake“ and contains 3 to 5% residual
metallic aluminum. It is normally disposed of in a landfill [2].
In response to increasing environmental pressures, the domestic primary aluminum
smelting industry has initiated a number of efforts to both minimize dross/salt cake
generation and to reprocess/recycle the byproduct wastes generated. The most common
approach has been to upgrade the metallic aluminum content of dross wastes prior to re-
melting. This is typically done by mechanically pulverizing the dross down to a coarse
powder and then screening it with a 10 to 20-mesh screen. The oversize fraction (the
concentrate) typically contains 60% to 70% metallic aluminum and is remelted. The
undersize fraction is then landfilled. This approach substantially reduces the amount of
dross smelted, lowering both the amount of energy and the amount of salt cake generated
during the smelting operation. Metal-rich salt cake residues can also be processed in the
same manner, further reducing the amount of material landfilled. Typical costs for this
concentrating operation generally range from about $5 to $60 per ton of material
processed [3].
The amount of waste material also can be minimized by minimizing the amount of
salt used during the smelting operation. This so-called “dry“ smelting process uses about
half the amount of salt employed for the more traditional “wet“ operation, but requires
that the furnace be tilted at the end of the run to remove the residual salt cake material
[4]. In addition, smelting operations are using improved control of aluminum oxidation
after dross removal to maximize metallic aluminum recovery and minimize the amount of
aluminum oxide, which must be landfilled. The approaches employed include protective
coverings, forced cooling (typically in water-cooled steel drums), and storage under a
protective inert atmosphere [5]. Mechanical pressing of the hot dross to squeeze out (a
portion of) the metallic aluminum is sometimes also used as an alternative to salt
smelting.
While the smelting by-product is viewed by industry as a disposal problem, costing
Vol.5, No.1 Recovery of Metals from Aluminum Dross and Saltcake 49
producers millions of dollars in landfill costs and exposing them to severe environmental
liabilities, we view the by-products not as a waste stream, but as raw materials, which
need further processing to create value-added products to economically enhance the
bottom line of the aluminum industry. In this project, we are trying to develop a
technology to divert the aluminum smelting by-products into valuable feedstock materials
for the manufacturing of concrete products such as lightweight masonry, foamed
concrete, and mine backfill grouts.
Methods for the separation of aluminum from the waste stream include manual
separation as well as density separation methods such as hydraulic or
pneumatic classifiers [6]. These methods are either labor intensive or of limited
applicability since many plastics have the same density as aluminum. More recently, very
promising methods have been developed based on the idea of eddy current separation [7].
An alternating magnetic field induces eddy current in conducting bodies, which in turn
combine with the magnetic field to cause a Lorentz force which is capable of accelerating
conducting materials away from nonconducting products. The ratio of electrical
conductivity to the mass density, δ/ρ, is an indication of the separability of the various
materials. The δ/ ρ ratio of aluminum is 13.0 [8].
In the present paper, characterization of dross and salt cake by-product material will
be discussed. Also the efficiency of Eddy Current Separation (ECS) to recover metallic
aluminum will be evaluated.
EXPERIMENTAL
Toxicity Characteristics Leaching Procedure (TCLP)
An Inductively Coupled Plasma (ICP) Emission Spectrophotometer (Leman Labs
Inc., Lowell, MA) was used for TCLP quantification. TCLP [9] extraction procedure was
performed as follows: a portion of by-product material was treated with known volume of
extraction fluid composed of glacial acid and sodium hydroxide whose pH was kept at
4.88 to 4.98. The slurry was contained in a jar with a cover and mixed in a Burrell‘s
Wrist Action Shaker (Pittsburgh, PA) for 22 hours. The slurry was filtered through no. 1
Whatman filter paper (acid washed) and the liquid phase (defined as the TCLP extract)
was collected for the ICP analysis. The acidified extract (pH 2) was stored in the
refrigerator.
Gas Analysis of Aluminum Salt Cake
Evolution of gases during the salt cake process was determined with a Gas
Chromatographic method using thermal conductivity detector (GC/TCD). Known
amounts of saltcake sample (-100 mesh) were reacted with a known volume of deionized
water in a headspace analysis jar (250 mL). The headspace of the jar was collected after a
certain time (from 0.5 to 48 hours) and analyzed with a GC/TCD at the operating
conditions listed in Table 1.
50 J. Y. Hwang, X. Huang, and Z. Xu Vol.5, No.1
A toxic gas monitor (GC Industries, Inc.) for hydrogen sulfide (H
2
S) gas was
employed in addition to GC analysis. For the H
2
S analysis, the headspace of the jar was
released through tubing (valve attached) to the gas monitor. The slurry mixture in the jar
was mixed using a magnetic stirrer.
Table 1. GC Operating Conditions
Column Molecular Sieve 5A,
Haysep Q
Detector Thermal conductivity
detector
Column
temperature
50
o
C
Detector
temperature
60
o
C
Injector
temperature
60
o
C
Filament
temperature
150
o
C
Carrier gas Helium
Flow rate 20 mL/min
Characterization
In order to identify constituents of -100 mesh fines of the aluminum by-product
materials, Scanning Electron Microscopy (SEM) analysis has been performed on the
materials at a state of as pulverized, after water wash and after NaOH treatment. The
SEM was JEOL, JSM-820. Secondary electron image resolution was at 30 kV and 10
-12
to 10
-6
A of probe current with 10X to 300,000X magnifications.
X-ray diffraction (XRD) phase analysis was carried out to characterize the crystal
phases in aluminum wastes. Parameters used for the analysis were target: Cu; count time:
1.800 sec; scan range: 20-89.99
o
; scan rate: 1.00 deg/min.
Metallic Aluminum Assay
Two grams of sample were mixed with 50 ml of 3N NaOH in a beaker and left under
the fume hood overnight. The slurry was filtered through #4 Whatman filter paper. The
filtrate was diluted to 250.0 mL with deionized water and analyzed with ICP.
Vol.5, No.1 Recovery of Metals from Aluminum Dross and Saltcake 51
Flowsheet of Eddy Current Separation Process
Circuit 1 flowsheet
White dross and black dross samples processed using Circuit 1 (Figure 1), were
crushed to -6" and then screened at 2”. The +1/2" fraction and -1/2" fraction were
separately fed to Eddy Current Separators with machine parameters adjusted for
optimized separation.
Figure 1. Circuit 1 flowsheet.
Circuit 2 flowsheet
In Circuit 2 (Figure 2), white dross and black dross samples were crushed to -12"
and then screened at 3/8", which shifted more material to the coarse fraction for further
liberation prior to ECS processing.
Since the top size for crushing was -12, it proved easier to handpick the large
aluminum pieces from the 12" x 3/8" fraction than running a mechanical scalping
operation. After handpicking, the rest of particles were shredded. The dust and spills were
combined with the coarse ECS reject since it was visually observed that the size of the
material was below the size of the coarse ECS reject stream and it would serve no
purpose to try and reprocess the material on a separate ECS system. The -3/8" fraction
was separately fed to an Eddy Current Separator.
52 J. Y. Hwang, X. Huang, and Z. Xu Vol.5, No.1
Figure 2. Circuit 2 flowsheet.
RESULTS AND DISCUSSION
Environmental Assessment and Characterization
Table 2 shows the type of aluminum smelting by-product wastes that were received
from three project team companies. The samples are here identified by letter designation,
i.e. Supplier A, Supplier B, etc., along with the type of material (saltcake, dross, dust,
etc.).
TCLP testing on as-received samples
Toxicity Characteristic Leaching Procedure (TCLP) tests were performed on the as-
received samples to identify environmental acceptability. The results are displayed in
Table 2 along with the concentration limits for acceptability. None of the as-received
samples pose environmental concerns for the heavy metals tested under TCLP guidelines.
Gas Generation from aluminum salt cake
The 20-drum salt cake sample received from Supplier A had a distinct odor of
hydrogen sulfide (H
2
S). A gas analysis study was performed on the material in both a dry
and wet state. The results are shown in Table 3. None of the gases generated when the
Vol.5, No.1 Recovery of Metals from Aluminum Dross and Saltcake 53
salt cake was wet were found to be hazardous.
Table 2. TCLP Results on As-Received Aluminum By-Product Wastes
Sample Concentration (ppm)
Se As Ba Cd Cr Pb
Supplier A
Salt Cake 0.083 0.124 0.526 0.164 0.262 0.303
Supplier B
Black Dross 0.154 0.111 0.545 0.211 0.095 0.109
Gray Baghouse Dust 0.039 0.054 0.030 0.009 0.003 0.065
Black Baghouse Dust 0.035 0.039 0.035 0.099 0.041 0.038
Supplier C
Baghouse Dust 0.027 0.030 0.039 0.010 0.016 0.057
Coarse Black Dross 0.033 0.043 0.008 0.009 0.023 0.054
Fine Black Dross 0.031 0.055 0.006 0.009 0.023 0.068
White Dross 0.033 0.049 0.007 0.009 0.029 0.061
ACA Oxide 0.165 0.251 0.664 0.135 0.272 0.038
Recovered Al Fines 0.173 0.279 0.749 0.134 0.300 0.402
TCLP Limits 1.00 5.00 100.00 1.00 5.00 5.00
No H
2
S gas was detected using a GC/TCD and a gas monitor. The concentration of
evolved hydrogen sulfide was not high enough to be detected from both wet and dried
salt cake, where the limit of detection for GC/TCD is 2.5 % and the H
2
S monitor can
determine the concentration of H2S in the atmosphere in the range of 0 to 100 ppm with
+/- 3ppm accuracy. Hydrogen sulfide gas could be recognized even when it was dry by
the characteristic odor. Notice that H
2
S gas odor is perceptible in air in a dilution of 0.002
mg/L (2 ppb).
Vigorous H
2
gas evolution was observed when the salt cake is exposed to water.
Four percent (v/v) of 2 was detected by the GC method, not exceeding the regulation
value, 4.5 % of minimum flammability. During the GC analysis, no methane and
ammonia gases were observed from the salt cake by Molecular Sieve 5A, and only trace
amounts of methane could be seen by Haysep Q column.
Table 3. Gas Analysis Results with Regulations
Gas Flammable
Limits
OSHA*
Limits
Experimental Results
H
2
S 4.3 to 45 % 20 ppm ND**
H
2
4.5 to 75 % Not toxic 4.2 %
NH
3
15 to 28 % 50 ppm ND**
CH
4
5.3 to 15 % Not toxic ND**
* Occupation Safety & Health Administration; ** Not detected
Characterization Results
Tables 4 and 5 provide a summation of the characterization results for aluminum
smelting by-product wastes. From ICP analysis (Table 4), it has been found that these
54 J. Y. Hwang, X. Huang, and Z. Xu Vol.5, No.1
materials contain three major components: water soluble salts (NaCl, KCl), metallic
aluminum, and oxides (Al
2
O
3
, MgAl
2
O
4
). The minor species detected by SEM and X-ray
diffraction were Si, SiO
2
, Ca
3
Al
2
O
6
, Ca
2
Al
2
SiO
7
, CaAl
2
Si
2
O
8
•H
2
O (Table 5).
Table 4. Characterization of Aluminum By-Product Wastes
Sample
Soluble Salts
(wt %)
Metallic
Aluminum (wt %)
Residual Oxides
(wt %)
Supplier A
Salt Cake 65.00 2.06 32.94
Supplier B
Black Dross 39.80 22.90 37.30
Gray Baghouse
Dust
19.10 1.40 79.50
Black Baghouse
Dust
14.90 2.20 82.90
Supplier C
Baghouse Dust 23.00 18.60 58.4
Coarse Black Dross 43.00 7.13 49.87
Fine Black Dross 34.00 5.83 60.17
White Dross 12.00 43.38 44.62
ACA Oxides 3.80 14.19 82.01
Recovered
Aluminum Fines
14.00 10.47 75.53
Table 5. Chemical and XRD analysis of various aluminum smelting by-products.
Sample Soluble salts Metallic
Aluminum (wt %)
Phases
Black Dross 43% 25% NaCl,
18% KCl
7.1 NaCl, KCl, Al, Al
2
O
3
, SiO
2
,
Si, MgAl
2
O
4
, Ca
3
Al
2
O
6
Baghouse
Dust
23% 16% NaCl,
7% KCl
18.6 NaCl, KCl, Al, Al
2
O
3
, SiO
2
,
Si, MgAl
2
O
4
, Ca3Al
2
O
6
,
Ca
2
Al
2
SiO
7
White Dross 2.7% 1.5% NaCl,
1,2% KCl
43.4 NaCl, KCl, Al, Al
2
O
3
,
MgAl
2
O
4
, Ca
3
Al
2
O
6
,
Ca
2
Al
2
SiO
7
SPF Fines 34% 20% NaCl,
14% KCl
5.8 NaCl, KCl, Al, Al
2
O
3
, SiO
2
,
Si, MgAl
2
O
4
, Ca
3
Al
2
O
6
,
Ca
2
Al
2
SiO
7
ACA Oxide 0.8% 0.35%
NaCl, 0.46% KCl
10.1 Al, Al
2
O
3
, SiO
2
, Si,
MgAl
2
O
4
, CaAl
2
Si
2
O
8
.H
2
O
RAF 2.5% 1.53%
NaCl, 0.98% KCl
13.9 Al, Al
2
O
3
, SiO
2
, Si,
MgAl
2
O
4
, CaAl
2
Si
2
O
8
.H
2
O
Vol.5, No.1 Recovery of Metals from Aluminum Dross and Saltcake 55
Eddy Current Separation
White dross and black dross samples were studied with an Eddy Current Separator to
evaluate its capability for recovery of aluminum. Circuit 1, which will be discussed in
further detail below, was first tested. A major finding from Circuit 1 was that the coarse
fraction aluminum recovery was relatively low. This finding initiated performing a
second test where the coarse fraction was further liberated by comminuting prior to
receiving ECS processing; this work was performed using Circuit 2. A detailed
discussion of this work follows.
Circuit 1 ECS processing performance
White Dross testing on Circuit 1
White dross with a feed composition of 41.05% metallic aluminum (Al
o
) was
processed first. Combining product streams the process generated a yield of 40.0 wt%.
Each stream (product and reject) had a screen analysis performed on it. This was
followed by analytical determination of Al
o
content for each individual size fraction. The
complete by-size analysis is provided in Table 6, including Al
o
recovery within size
fractions and Al
o
recoveries by size fraction, for both the coarse and fine materials.
The coarse fraction (+ 2”) averaged 74.07% aluminum recovery. The recoveries
within size fractions ranged from 66.70% to 78.73%, not extremely high, but fairly
consistent. The assay for the product fraction averaged 78.85% Al
o
with a range of 71.20-
88.76% Al
o
. A clue that liberation may be a problem can be seen in the product assay
where the +1" material showed an assay of 71.20% Al
o
and the quality climbed to
88.76% for the +3/4" fraction and remained high for the rest of the fraction. Also, the ½"
x 2”size had the highest by size recovery at 22.33% of an overall recovery of 74.07% for
the coarse fraction, even though the calculated head assay of that material was not the
highest. Again, these results identify potential liberation problems.
The fine fraction (2” x 0) showed that product assay performance dropped off
sharply for the -10 mesh fraction. The recovery within the size fraction steadily declined
at particle sizes below 6 mesh. Since the +3 mesh fraction shows a 94.12% Al
o
recovery
at 55.63% Al
o
grade, which is higher than the 65.23% Al
o
recovery with 35.09% Al
o
grade at 3 x 6 mesh, the decline is believed to be related to the equipment capability
rather than the liberation.
Overall the Circuit 1 white dross test produced product quality of 69.02% Alº from
feed of 41.05% Alº, with a total Alº recovery of 67.25%.
Black Dross testing on Circuit 1
Black dross with a feed composition of 42.44% Alº was processed on Circuit 1.
Combining product streams the process generated a yield of 35.42 wt%. From Table 7,
the coarse fraction (+²") performance was consistent with regards to product grade at
various particle sizes, maintaining above 50% Al
o
. The Al
o
recovery within each size
fraction declined steadily from a high of 77.59% for +1" present in the coarse fraction.
Actually, this is somewhat surprising considering the calculated feed assay for each size
fraction was fairly homogeneous with an average of 48.61% Al
o
and a range of 40.81 to
56 J. Y. Hwang, X. Huang, and Z. Xu Vol.5, No.1
58.60% Al
o
. In direct comparison to white dross, where the average feed assay for the
coarse fraction was 52.55% Al
o
, it is fairly close to that of the black dross. However, the
white dross Al
o
recovery was fairly tight within size fractions, averaging 74.07% with a
range of 66.70-78.73%.
In the fine fraction (2” x 0), the product assay drops off below 10 mesh. Al
o
recovery
within each size fraction drops off below 3 mesh. What was surprising was that the +3
mesh size fraction had a 71.08% Al
o
recovery, better than most of the sizes in the coarse
fraction, following the pattern established in white dross testing.
Overall, the Circuit 1 black dross test produced product quality of 67.73% Al
o
from
feed of 42.44% Al
o
, with a total Al
o
recovery of 56.53%.
The performance of quality was quite similar between black dross and white dross,
but it was obtained differently, by processing more coarse material in the coarse fraction,
68.7 wt% for black dross compared to 57 wt% for white dross.
Again, with black dross, based on +3 mesh performance in the fine fraction and
somewhat irregular behavior in coarse fraction performance, the question of further
liberation needs to be addressed.
Table 6. White Dross Circuit 1 Test Results
Size
Fraction
Wt.%
Feed
ECS
Product
Product
Al
ο
Rejects Wt% Rejects
Al
ο
Calculate
Assay
Al
ο
Recovery
Al
ο
Recovery
+1/2 inch fraction
+1" 26.42 13.23 71.20 13.19 29.90 50.58 70.49 17.92
+3/4" 22.76 11.63 88.76 11.12 25.03 57.61 78.73 19.64
+1/2" 27.07 14.95 78.49 12.12 26.21 55.08 78.70 22.33
-1/2" 23.75 9.56 77.92 14.19 26.21 47.02 66.70 14.17
+1/2 inch
feed
100.00 49.37 78.85 50.62 26.91 52.55 74.07 74.07
-1/2 inch fraction (mesh)
3 mesh 15.77 13.12 55.63 2.65 17.21 49.17 94.12 28.28
6 22.05 11.84 35.09 10.21 21.69 28.89 65.23 16.10
10 25.91 1.79 54.14 24.12 17.40 19.94 18.76 3.75
14 10.07 0.84 22.02 9.23 16.81 17.24 10.65 0.72
20 7.37 0.00 0.00 7.37 20.10 20.10 0.00 0.00
28 4.51 0.00 0.00 4.51 21.09 21.09 0.00 0.00
35 2.72 0.00 0.00 2.72 21.20 21.20 0.00 0.00
48 3.88 0.00 0.00 3.88 22.60 22.60 0.00 0.00
-48 7.72 0.00 0.00 7.72 11.65 11.65 0.00 0.00
-1/2 inch
feed
100.00 27.59 45.70 72.41 18.24 25.81 48.84 48.84
Vol.5, No.1 Recovery of Metals from Aluminum Dross and Saltcake 57
Table 7. Black Dross Circuit 1 Test Results
Size
Fraction
Wt.%
of
Feed
ECS
Product
Wt%
Product
Assay
Al
ο
Rejects
Wt%
Rejects
Assay
Al
ο
Calc.
Assay of
Size
Fractions
Al
ο
Recovery
w/in Size
Fraction
Al
ο
Recovery
by Size
Fraction
+1/2 inch fraction
+1" 62.58 31.53 75.77 31.05 22.22 49.20 77.59 49.15
+3/4" 18.60 8.99 56.38 9.61 26.24 40.81 66.78 10.43
+1/2" 12.00 4.19 51.59 7.81 62.36 58.60 30.74 4.45
-1/2" 6.82 1.14 52.67 5.68 45.70 46.87 18.79 1.24
+1/2”
feed
100.00 45.85 69.18 54.15 31.19 48.61 65.26 65.26
-1/2 inch fraction (mesh)
3 11.55 5.84 53.18 5.71 22.13 37.83 71.08 10.74
6 25.37 5.14 59.61 20.23 29.57 35.66 33.87 10.59
10 29.49 1.54 55.88 27.95 29.36 30.74 9.49 2.97
14 8.77 0.06 39.90 8.71 23.12 23.23 1.17 0.08
20 8.36 0.00 0.00 8.36 22.27 22.27 0.00 0.00
28 5.93 0.00 0.00 5.93 18.26 18.26 0.00 0.00
35 3.19 0.00 0.00 3.19 18.79 18.79 0.00 0.00
48 2.14 0.00 0.00 2.14 18.11 18.11 0.00 0.00
-48 5.20 0.00 0.00 5.20 9.20 9.20 0.00 0.00
-1/2”
feed
100.00 12.58 56.07 87.42 25.02 28.93 24.38 24.38
Circuit 2 ECS processing performance
White Dross testing on Circuit 2
White dross from a different batch with a feed composition of 60.27% Al
o
was
processed using Circuit 2. Combining all product streams, the process generated a yield
of 75.87 wt%. A breakdown of the by size performance is provided in Table 8.
As seen in Table 8, the product assays for the coarse fraction were in the 45-55% Al
o
range down to 10 mesh, but the Al
o
recovery within size fractions was outstanding, being
in the 90% level down to 10x14 mesh. The Al
o
recovery within the 14x20 mesh size was
78.28%, comparable to the best coarse fraction recoveries from Circuit 1. By including
the additional liberation step, the coarse fraction material obtained higher recoveries at
approximately the same quality. In addition the ECS was able to effectively perform on
slightly finer material than experienced under Circuit 1 conditions.
The fine fraction (3/8" x 0) showed that product assay by size maintained quality
down to the 20x28 mesh size, however, the recoveries within each size fraction were
extremely good for the +3 mesh size (94.13%) and 3x6 mesh size (83.06), but dropped
off significantly at 10x14 mesh size. In general, reducing to a 3/8" size changed the
weight distribution to the fine ECS unit, but pretty much maintained performance
compared to the ² inch split of Circuit 1. Overall Circuit 2 produced better product quality
(73.49% Al
o
) with a remarkably high aluminum recovery of 92.51%. This may be
58 J. Y. Hwang, X. Huang, and Z. Xu Vol.5, No.1
somewhat misleading because the feed assay of the Circuit 2 white dross was 60.27%
Al
o
, up considerably from the material tested on Circuit 1 (41.05% Al
o
). For true
comparison the performance has to be normalized with regards to feed quality, which will
be addressed later.
Black Dross testing on Circuit 2
Black dross with a feed composition of 15.44% Al
o
was processed using Circuit 2.
Again there is a disparity in feed quality between black dross material in Circuit 1 and
Circuit 2 and this will have to be normalized for fairness of comparison. A breakdown of
the by-size performance is provided in Table 9.
Because of the low feed grade quality, considerably lower product qualities were
experienced than expected. For the coarse fraction the quality drops off below 10 mesh,
and the recovery within size fractions also experienced significant deterioration below 10
mesh. For the fine size fraction (3/8" x 0), product quality was fairly consistent down to
20 mesh, but Al
o
recoveries within size fractions were small below 10 mesh. Overall the
Circuit 2 black dross test produced a product quality of 30.16% Al
o
from a feed of
15.44% Al
o
, with a total Al
o
recovery of 56.81%.
Comparisons of the ECS
The efficiency of separation for two ECS processing circuits was compared, by
including normalization because of variations of feed quality. To provide a basis for
comparison, we know that if we produced 100% Al
o
at 100% recovery a perfect 100%
efficiency of separation would be obtained. Secondly, to compensate for feed quality
variations between tests, the ratio of Assay of Product over Assay of Feed ratio can be
used for normalizing. Combining these aspects, an efficiency index value can be obtained
from the following equation.
[[(% Al
o
in Product)/ (% Al
o
in Feed)] x wt% recovery]/100 = Efficiency Index Value
Based on the equation the higher the index value, the more efficient the separation.
Table 10 displays the efficiency of separation index values as determined for Circuit 1
and Circuit 2 white and black dross tests.
Vol.5, No.1 Recovery of Metals from Aluminum Dross and Saltcake 59
Table 8. White Dross Circuit 2 Test Results
Size
Fraction
Wt.% of
Feed
ECS
Product
Wt%
Product
Assay
Al
ο
Rejects
Wt%
Rejects
Assay
Al
ο
Calc.
Assay of
Size
Fractions
Al
ο
Recovery
w/in Size
Fraction
Al
ο
Recovery
by Size
Fraction
12 x 3/8 inch fraction (mesh)
3 30.47 30.47 55.77 0.00 0.00 55.77 100.00 40.90
6 20.44 20.30 49.68 0.14 29.50 49.54 99.59 24.28
10 16.38 15.86 45.95 0.52 24.75 45.28 98.26 17.54
14 5.44 4.85 38.12 0.59 22.18 36.39 93.39 4.45
20 4.52 3.19 32.92 1.33 21.91 29.68 78.28 2.53
28 4.25 1.98 27.15 2.28 20.12 23.39 53.96 1.29
35 1.40 0.41 24.27 0.99 19.33 20.78 34.21 0.24
48 6.24 0.96 15.53 5.28 16.60 16.44 14.54 0.36
-48 10.85 0.78 13.95 10.07 12.57 12.67 7.92 0.26
12 x 3/8"
feed
100.00 78.81 48.43 21.20 15.97 41.54 91.85 91.85
-3/8" fraction (mesh)
3 4.77 3.82 45.02 0.95 11.29 38.30 94.13 6.26
6 13.82 9.29 49.14 4.53 20.56 39.77 83.06 16.62
10 17.07 7.14 53.08 9.93 25.57 37.08 59.88 13.80
14 7.20 1.71 50.37 5.49 28.24 33.50 35.71 3.14
20 8.11 1.04 51.43 7.07 27.97 30.98 21.29 1.95
28 9.80 0.62 43.62 9.18 24.27 25.49 10.82 0.98
35 3.33 0.13 25.29 3.20 24.73 24.75 3.99 0.12
48 12.33 0.28 20.12 12.05 19.32 19.34 2.36 0.21
-48 23.57 0.39 11.90 23.18 13.55 13.52 1.46 0.17
-3/8" feed 100.00 24.42 48.64 75.58 20.63 27.47 43.23 43.23
Products Wt% of feed
Al
ο
assay Al
ο
Recovery
SK12 58.08 81.15
SK4 4.60 48.64
SK14 13.18 48.43
Subtotal 75.87 73.49 92.51
Rejects
SK5 14.24 20.63
SK15 3.55 15.96
Baghouse 6.35 15.96
Subtotal 24.13 18.71 7.49
TOTAL 100.00 60.27 100.00
60 J. Y. Hwang, X. Huang, and Z. Xu Vol.5, No.1
Table 9. Black Dross Circuit 2 œ By size Performance Analysis
Size
Fraction
Wt.% of
Feed
ECS
Product
Wt%
Product
Assay
Al
ο
Rejects
Wt%
Rejects
Assay Al
ο
Calc.
Assay of
Size
Fractions
Al
ο
Recovery
w/in Size
Fraction
Al
ο
Recovery
by Size
Fraction
12 x 3/8 inch fraction (mesh)
3 22.87 22.02 34.06 0.85 5.70 33.01 99.36 45.61
6 11.91 6.54 24.06 5.40 8.85 17.16 76.62 9.53
10 17.87 5.01 20.68 12.86 9.57 12.68 45.71 6.30
14 7.85 1.36 14.06 6.49 12.41 12.70 19.19 1.16
20 10.43 0.72 13.67 9.71 9.38 9.68 9.75 0.60
28 7.25 0.31 12.06 6.94 9.51 9.62 5.36 0.23
35 4.89 0.13 10.60 4.76 8.61 8.66 3.25 0.08
48 4.51 0.03 12.80 4.48 8.58 8.61 0.99 0.02
-48 12.42 0.05 7.10 12.37 8.61 8.60 0.33 0.02
12 x 3/8"
feed
100.00 36.14 28.92 63.86 9.38 16.44 63.56 63.56
-3/8" fraction (mesh)
3 6.69 3.33 26.99 3.36 11.10 19.01 70.67 7.90
6 15.82 5.02 23.96 10.80 9.24 13.91 54.65 10.57
10 16.46 2.24 30.15 14.22 9.83 12.60 32.58 5.93
14 8.33 0.37 32.37 7.96 10.48 11.45 12.55 1.05
20 8.85 0.16 30.88 8.69 9.55 9.94 5.62 0.43
28 10.21 0.21 10.34 10.00 9.25 9.27 2.29 0.19
35 8.56 0.00 0.00 8.56 9.63 9.63 0.00 0.00
48 7.04 0.00 0.00 7.04 10.27 10.27 0.00 0.00
65 5.27 0.00 0.00 5.27 10.00 10.00 0.00 0.00
100 4.40 0.00 0.00 4.40 9.20 9.20 0.00 0.00
-100 8.36 0.00 0.00 8.36 6.90 6.90 0.00 0.00
-3/8" feed 99.99 11.33 26.19 88.66 9.49 11.38 26.08 26.07
Products Wt% of feed Al
ο
assay Al
ο
Recovery
BD4 1.54 26.19
BD14 26.86 28.92
BD12 0.68 88.32
Subtotal 29.08 30.16 56.81
Rejects
BD5 12.05 9.49
BD15 47.46 9.38
Spills 11.41 9.38
Subtotal 70.92 9.40 43.19
TOTAL 100.00 15.44 100.00
Vol.5, No.1 Recovery of Metals from Aluminum Dross and Saltcake 61
Table 10. Efficiency Index Values from
ECS Testing
Circuit 1 Circuit 2
White Dross 1.131 1.128
Black Dross 0.90 1.11
CONCLUSIONS
Based on the Eddy current separation tests, the following conclusions can be
made from the test work:
White dross performed equally well with either circuit.
Black dross processing was more efficient with Circuit 2, realizing the effect of
liberation on the coarser fraction.
Both circuits were more efficient using white dross than black dross or white
dross is easier to upgrade.
ECS technology is effective processing down to 6-10 mesh size material.
ACKNOWLEGEMENT
Funding for this research was provided by the U.S. Department of Energy.
REFERENCE
[1] P. A. Plunkert, “Aluminum Recycling in the United States–A Historical Perspective“,
J. H. L. van Linden, D. L. Stewart, Jr, and Y Sahai (eds.), Second International
Symposium–Recycling of Metals and Engineered Materials, Warrendale PA, TMS, 1990.
p. 61.
[2] G. J. Kulik and J. C. Daley, “Aluminum Dross Processing in the 90‘s“, J. H. L. van
Linden, D. L. Stewart, Jr, and Y Sahai (eds.), Second International Symposium–
Recycling of Metals and Engineered Materials, Warrendale PA, TMS, 1990. p. 427.
[3] D. J. Roth, “Recovery of Aluminum from Rotary Furnace Salt Cake by Low Impact
Rotary Tumbling“, W. Hale (ed), Light Metals 1996, Warrendale PA, TMS, 1996. p.
1251.
[4] E. H. Geus, V. Lussi and H. Spoel, “Process for the Recovery of Aluminum From
Dross Without Use of Salt Flux“, (in) P. B. Queneau and R. D. Peterson, ed., Third
International Symposium–Recycling of Metals and Engineered Materials, Warrendale
PA, TMS, 1995, p. 813.
[5] H. Spoel, “Safety Aspects of the Design and Operation of Rotary Aluminum Dross
Coolers“, W. Hale (ed.), Light Metals 1996, TMS, 1996, p. 1241.
[6] A.F. Barton, Resource Recovery and Recycling. New York: Wiley, 1979.
[7] G.H. Nijhof, “Aluminum separation out of household waste using eddy current
technique and reuse of metal fraction,“ Recourses, Conservation Recycling, no. 10, p
161, 1994.
62 J. Y. Hwang, X. Huang, and Z. Xu Vol.5, No.1
[8] Exhibitors at ENVITEC, Bulk Solids Handling, 18, 145, 1998.
[9] Code of Federal Regulations: 40 CFR Appendix II to Part 261 Method 1311 TCLP
(7-1-92 Edition)