See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/343294387

PARTICLE SIZE AND COMPOSITION OF DROSS PARTICLES FROM GALVANIZE

AND GALVANNEAL OPERATIONS

Conference Paper · September 2002

CITATIONS READS

5 92

3 authors, including:

Frank Ajersch Frank E Goodwin

Polytechnique Montréal International Zinc Association
113 PUBLICATIONS   1,408 CITATIONS    543 PUBLICATIONS   1,515 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Metal Matrix Composites View project

Silver View project

All content following this page was uploaded by Frank E Goodwin on 29 July 2020.

The user has requested enhancement of the downloaded file.

 P A R T I C L E SIZE A N D C O M P O S I T I O N OF D R O S S P A R T I C L E S
FROM GALVANIZE AND GALVANNEAL OPERATIONS

F. Ajersch*, L. Tr6panier* and F.E. Goodwin**

*l~cole Polytechnique, D6partement de g6nie chimique

C.P. 6079, succ. Centre-Ville, Montr6al, Qc, H3C 3A7, Canada
Tel." 514-340-4711 (4533)
email" ffan k.aj ersch@poly,mtl.ca

** International Lead Zinc Research Organisation

2525 Meridian Parkway, P.O. Box 12036
Research Triangle Park, NC 27709-2036, USA
Tel.: 919-361-4647
email: fgoodwin@ilzro, org

Key Words: Galvanizing bath, dross particles, size distribution, composition, phase transformation

INTRODUCTION

The control of dross particles in the bath of continuous galvanizing lines can be considered to be of
major concern for the manufacture of quality coated steel products. Bath cleanliness depends to a large degree
on the bath chemistry, the temperature fluctuation as well as on the specific operating variables such as ingot
addition, dross skimming procedures, line speed and various steel characteristics. Many operations run the same
lines for both galvanized and galvannealed products where the transition from one type of operation to the other
takes considerable time, especially for large size baths. The physical and chemical characteristics of the dross
particles generated in the zinc solution can vary considerably.

This study was undertaken under the sponsorship of the International Lead Zinc Organization Inc.
(ILZRO) in order to quantify the size distribution and chemical composition of dross particles from different
industrial operations. Seven companies representing 14 different galvanizing lines had agreed to sample each of
their zinc pots according to a procedure recommended by the project investigators. Since samples were taken
from both galvanize and galvanneal operations and during transition periods, information on the history of
dross formation and growth as well as their dissolution could also be obtained. The objective of the study was
to measure the size, shape and chemical compositions of the dross particles of the samples from the different
baths in order to be able to interpret the variables that contribute to their formation, growth and movement in
the bath. Ideally, it was intended that all operations retrieve samples using the same standard prescribed
procedure. This was not possible resulting in samples of variable size from water quenched granules to two kilo
blocks. Consequently, some influence of the rate of solidification on the particle morphology was observed.
Particle sizes and shapes were measured using an image analysis system and quantitative energy dispersive X-
ray analysis was carried out to determine the chemical compositions.

44 th MWSP Conference Proceedings, Vol. XL, 2002 771

 Samples from galvanneal operations containing both Fe2AI5 and FeZnlo particles are indicative of a non-
equilibrium bath in a stage of transition from galvanize to galvanneal. The Fe2A15 particles in these samples
were observed to be in the process of dissolution as indicated by a non faceted surface and by a much smaller
size than the stable FeZn~o particles. It was generally observed that the number and size of particles increased
with a decrease in bath temperature and an increase in aluminum content of the bath. Zinc baths sampled
following ingot addition seem to indicate an increase both the number and size of the particles with a tendency
to agglomerate. The chemical composition of the particles from samples taken at different locations in the same
bath were almost constant which is indicative of good mixing within the bath. The analyses compared very well
with the analyses particles from the literature, clearly identifying then as ternary intermetallics.

EXPERIMENTAL

Each of the participating companies received a protocol as to the time and interval of sampling. Since
the sampling techniques varied among the participants, the samples varied in size from water granulated shot to
two kilogram blocks. Four samples were taken over one hour period at 12 minute intervals for each of the
operation.

A total of 69 samples from the 7 different companies operating 14 lines were analyzed for both particle
morphology and chemical composition. Each line provided information as to the operation parameters (GA, GI,
temperature, pot size (tonnes) A1 and Fe content wherever available, location and interval between samples).
Not all the companies provided complete information as to bath composition and steel chemistry. The pertinent
information is summarized in Table 1.

Table 1" Bath and Sampling Conditions

Co. Bath Strip speed

No Line Type Size Temp. (°C) % A1 V (m*s-1)
(Tonnes)
1 GI 170 458-459 0.21 2.91
2 N GI 130 465-468 0.175-0.205 0.61
P GI 453-455 0.21
W GA 457-460 0.13 0.45
- 160 488-490 - 0.92
4 GI 170 503 0.17 eff 2.50
5 GI 250 470 0.18 eff 0.68
7 GI 170 455 0.28 eff 2.33
s GA 250 460-465 0.135 eff 1.20
_ _
469 0.135 0.98
6 A GI 60 460-462 0.20-0.23 2.66
B GI 60 460-462 0.20-0.23 2.66
7 I GI 170 458 0.18 2.50
A GA 170 455 0.135 2.00

The samples were sectioned and mounted in a 1 1/2 inch bakelite mould and polished down to 0.5 btm
with diamond paste using an automated polishing unit. Due to the limited difference in contrast between the
matrix and the FeZn~0 particles, the polished surfaces were chemically etched using Nitol (5-7 drops of citric
acid in 30 ml of ethanol) exposed for 5 to 10 seconds for optical microscopy and for 20 to 25 seconds for the
energy dispersive X-ray analysis.

772 44 th MWSP Conference Proceedings, Vol. XL, 2002

Image Analysis

The surface of the polished samples were scanned for a sufficient number particles of each of the
intermetallic FezA15 and FeZnl0 phases. An average of 10 photos for each type of particle was taken for each
sample, generating 980 digitalized photos of 45 samples containing only FezA15 particles and 24 samples with
both FezA15 and FeZn~0 particles. All photos were taken at the same magnification of 500 x.

An image analysis software (IMAGE TOOL UTHSCA) was used to analyze the data which digitalized
the image at a resolution of 375 pixels per line times 281 lines for a total of 105375 pixels per image.
Parameters such as surface area, perimeter, length of major and minor axes, equivalent diameter, roundness and
compaction were tabulated for each of the sets of particles measured. The average mean particle diameter was
taken as a measure of the particle size.

Analysis of the Chemical Composition

The same polished samples were also used to determine the chemical composition of the dross particles.
The analysis was carried out on a JEOL 840 scanning electron microscope equipped with a LINK energy
dispersive X-ray analyzer. At least two spectra per particle were obtained for each particle totaling 178
analyses. The operating parameters of the analyzer are presented in Table 2.

Table 2: Operating Parameters for SEM-EDS Analysis

Parameter Value
Accelerating voltage 15Kev
Current 1 x 10-9A
Focal distance 39 mm
Elements analyzed Zn, Fe, A1

Several analyses of the same particles were carried out to check the reproductibility as well as the
chemical homogeneity. All the spectral data were subjected to a correction for absorption and atomic number
effects for the quantification of the composition values. Analyses were carried out only for particles greater than
5 gm in diameter in order to minimize the error due to volume interaction. The calculation for the
approximation of the volume interaction for both types of particles was found to be only 0.496 gm 3, well within
the volume of a 5 gm diameter particle.

RESULTS

Particle Morphology

Particle sizes and shapes were determined from the 980 digitalized photographs of optical images taken
from the samples of the seven different companies. These images were analyzed in order to obtain the surface
area, perimeter, length of major and minor axes, the aspect ratio, roundness, equivalent diameter and
compaction. Table 3 summarizes the data for area, perimeter and equivalent diameter only. Particles are
identified as either eta (Fe2A15) or delta (FeZn~0)even though in all cases they are of a ternary composition. The
results are presented for the mean values of 5 representative particles from each sample. Photographs of
representative particles from each line were also taken.

44 th MWSP Conference Proceedings, Vol. XL, 2002 773

 Table 3" Average Size and Particle Morphology

Company/Samples Phase Are a (lam)2 Perimeter gm Equiv. dia. ~tm

Fe2A15 673.98 127.59 28.52
2N Fe2A15 431.49 104.25 22.87
2P Fe2Als 507.78 117.04 24.64
2W Fe2A15 422.62 89.57 22.59
2W FeZnl0 115.87 52.60 10.93
2W Fe2Znl0 2213.32 173.06 39.73
3.1 Fe2A15 565.28 115.84 25.49
3.2 FeZn~0 2398.94 251.27 48.35
4.4 Fe2A15 186.22 64.91 14.98
4.5 Fe2A15 294.97 89.75 18.22
4.7 Fe2A15 237.07 332.72 53.66
4.7a Fe2A15 467.83 135.01 24.06
4s Fe2A15 407.01 104.50 22.46
4sa FeZnl0 3548.46 358.18 65.28
Fe2A15 2326.39 233.09 47.91
6A Fe2A15 299.20 342.41 19.04
6B Fe2A15 3335.88 366.59 63.74
6B 1 Fe2A15 386.94 99.07 21.74
71 Fe2A15 3327.33 282.52 64.07
711 Fe2A15 773.06 123.47 30.25
7A Fe2A15 773.51 133.41 29.33
7A1 FeZn~0 4790.96 320.22 74.59

It should be noted that the photographed particles are selected from the larger particles (generally > 5
~tm) in each sample. Quantitative image analysis of all the particles was not carried out in this study as was
done by Cominco ~. The task of measuring all dross particles in a sample would not only be extremely arduous,
but also prone to errors in composition due to the interaction volume of particles smaller than 5 gm.

The equivalent mean particle diameter of Fe2A15 for company 1 with a bath composition of 0.21% A1 is
28.52 ~tm with a standard deviation of 4.79 lam. The shapes are variable with polygonal faces but the smaller
particles appear to be more rounded.

Company 2 supplied samples from three lines (N, P and W). The equivalent mean particle diameter of
line N is 22.87 ~tm with a standard deviation of 3.72 ~tm and generally of the same polygonal from as those of
company 1. Representative particles are shown in Figure 1a and 1b for baths operating in similar conditions.

Line W of company 2 shows both Fe2A15 and FeZn~0 particles for a bath operating at 0.13% A1. The
FezA15 particles measured 22.59 + 6.24 ~tm. Some of the FeZn~0 particles were observed to be much smaller, in
the 10 ~tm range, but numerous large particles greater than 60 gm were also observed. Overall standard
deviations therefore have little meaning unless a bimodal distribution is quantified. The number of Fe2A15
particles is much smaller as would be expected in a GA operation. In addition, we can observe the presence of
some FeZn~0 with inclusions of small FezA15 particles which would indicate the transition of a GI to a GA
operation.

774 44 th MWSP Conference Proceedings, Vol. XL, 2002

 .. ' I F ~' -- .... , , . Q

, . p - ... ,.
l

(a) (b)

Figure 1.a. Fe2A15 particles, Company 1, bath at 0.21%A1

Figure 1.b. FezA15 particles, Company 2, line N, bath at 0.175-0.205 %A1

The results from company 3 also indicate the presence of both Fe2A15 and FeZn~0 particles. This line is
therefore assumed to be in a GA operation or in a transition period. The FezA15 particles are very uniform in
size at 25.49 + 1.52 ~m whereas the FeZnl0 particles are almost twice that size at 48.35 + 21.97 ~tm but much
less numerous.

Company 4 supplied the largest number of samples from four different lines, each operating at a
different aluminum content. Line 4 (0.17% eff. A1) produced very few particles of FezA15 which were small and
of uniform size at 14.98 + 1.01 jam and which tended to agglomerate. Particles from line 5 (0.18% eff. A1) were
somewhat larger and much more numerous at 18.22 + 3.94 mm. As opposed to samples from company 2, the
shapes are generally much less faceted and appear to be undergoing dissolution with little or no tendency to
agglomerate. Representative particles from these two lines of company 4 are shown in Figures 2a and 2b. Line
speeds are different and line 4 temperature is 503°C, considerably higher than conventional operations.

Line 7, operating at a very high effective A1 content of 0.28%, shows very large Fe2A15 particles of
53.66 + 5.45 gin. It appears that the large particles are most likely the result of the agglomeration of small
particles due to their relatively irregular shape. Line S from company 4 was sampled for a GA run at 0.135%
eff. A1. The FezA15 particles were much less numerous and smaller at 22.46 lam with rounded surfaces and large
pores suggesting a dissolution mechanism. Some of these particles were enclosed within the FeZn~0 particles
was also previously observed for company 2. The FeZn~0 particles however were large (65.28 jam) and
polygonal, of a relatively even distribution, but not very numerous which would indicate a stage at the
beginning of the GI to GA transition. Representative particles for each of the lines are shown in Figures 3a and
3b.

Company 5 also operated a GA line at 0.135% A1 but the samples produced only FeZn~0 particles of
average size of 47.91 lam with a large standard deviation of 28.32 gm. The large spectrum of particle sizes
would indicate a continuous nucleation and growth mechanism that occurs throughout the sampling period for a
line that has had a long GA run. Representative particles are illustrated in Figure 4a and 4b.

44 th MWSP Conference Proceedings, Vol. XL, 2002 775

 100~tm

(a) (b)

Figure 2.a. Fe2A15 particles, Company 4, line 4, bath at 0.17 % eff .A1.
Figure 2.b. Fe2A15 particles, Company 4, line 5, bath at 0.18 % eff.A1.

(a) (b)

Figure 3. Fe2A15 (dark grey) and FeZnl0 (light grey) particles, Company 4, line S, bath at 0.135 % eff.A1
showing separate (a) and included (b) particles showing transition stage from galvanize to galvaneal.

(a) (b)
Figure 4. FeZn~0 particles from company 5 showing a heterogeneous nucleation on Fe2A15 particles (a) and
spontaneous nucleation in the matrix (b).

776 44 th MWSP Conference Proceedings, Vol. XL, 2002

 Particles in samples from company 6 for lines "A" and "B", both for GI operations (0.20 to 0.23% A1)
indicated a wide range of sizes. Line "a" samples showed Fe2A15 particles of an average size of 19.04 gm which
are generally polygonal but were not found to be in an agglomerated form. Line "B" produced large Fe2A15
particles ranging from 21.74 to 63.76 gm in diameter where the external form appears to be diffuse rather than
faceted, indicative of dissolution. The difference in the particle morphology between the two lines could be
attributed to the sampling location. Line "A" was sampled at about half depth of the bath whereas line "b" was
sampled near the bath surface. The surface samples contained agglomerated particles resulting in larger sizes
and also exhibited large porosities.

The GI line of company 7 operating at 0.18% A1 showed large well defined particles with an average
size of 64.01 gm. The particles seem to be in a state of dissolution since the external surfaces are much more
rounded and with surface porosities. The GA line operating at 0.135% A1 however shows much smaller Fe2A15
particles averaging 29.33 jam and larger FeZn~0, particles of 74.59 gm. The distribution is more or less
homogeneous. Some Fe2A15 particles are located within the larger FeZn~0 particles or may appear attached to
the surface. This type of distribution is characteristics of the transition from GI to GA operations and was also
observed for company 2 with comparable particles as shown in Figures 3a and 3b..

Chemical Composition of Particles

Typical spectra generated by dispersive energy analysis for Fe2A15 and FeZnl0 particles were divided
into two groups; that is, for companies operating GI lines and for those operating GA lines. Mean values are
tabulated for the entire series of samples consisting of 178 analysis summarized in Tables 4 and 5.

The results clearly show that the compositions of the particles for the GA lines are much more
consistent than the particles form the GI lines. It should be noted that the operating conditions for the GA lines
are kept within much closer limits of A1 content than for the GI lines. Variations in temperature from 455°C to
490°C did not seem to have a great influence on the composition of the particles.

Table 4" Summary of Compositions of Particles for GA Operations

Co Type Temperature of bath Bath composition Particle composition

(°C) % A1 % A1 %Fe % Zn
2-W Fe2AI5 457-460 0,13 45.50 31.22 23.28
2-W FeZn~o 457-460 0,13 2.98 4.49 92.54
Fe2A15 488-490 45.41 30.81 23.79
FeZn~o 488-490 2.10 5.28 92.62
4-S Fe2A15 460-465 0.135 eff 46.59 31.44 21.97
4-S FeZnlo 460-465 0.135 eft 2.71 5.12 92.17
FeZn7 469 0.135 2.21 5.30 92.5
7-A Fe2A15 455 0.135 47.28 30.92 21.81
7-A FeZn~o 455 0.135 2.70 5.05 92.25
Mean particle composition Fe2A15 (40 spectra) 46.20 31.10 22.71
Standard deviation 0.90 0.29 0.97
Mean particle composition FeZn~0 (46 spectra) 2.54 5.05 92.42
Standard deviation 0.37 0.33 0.20

44 th MWSP Conference Proceedings, Vol. XL, 2002 777

 Table 5" Summary of Compositions of Particles for GI Operations

Co Type Temperature of bath Bath composition Particle composition

(°c) % A1 % A1 %Fe % Zn
Fe2A15 456-459 0,21 39.46 36.94 23.60
2-n Fe2AI5 465-468 0.175-0.205 28.20 39.47 32.34
2-p Fe2A15 453-455 40.39 38.13 21.49
4-4 Fe2A15 503 0.17 eff 34.49 36.69 28.82
4-5 Fe2A15 470 0.18 eff 25.91 38.04 36.09
4-7 Fe2AI5 455 0.28 eff 37.98 38.34 23.75
6-A Fe2AI5 460-462 0.20-0.23 40.13 37.24 22.33
6-B Fe2AI5 460-462 0.20-0.23 41.14 37.27 21.60
7-I Fe2A15 458 0.18 46.86 30.96 22.19
Mean particle composition Fe2Als (92 spectra) 37.21 37.01 25.80
Standard deviation 6.62 2.42 5.34

Results of Table 5 for the GI lines do show a variation in composition which may be due to the
operating parameters but which are difficult to quantify. As opposed to the results of the GA lines, the variation
in composition can be attributed to bath composition and bath temperature.

DISCUSSION

The type of particles observed in the samples are basically in accordance with what would be expected
from the phase diagram. Lines operating in the galvanizing composition range above 0.16% A1 showed only
Fe2A15 particles, whereas the GA lines between 0.13 to 0.14% A1 (near the knee point) generally showed the
two types of particles, and where Fe2Als appear to be dissolving. Company 5 did not show any FezA15 particles
for their GA operation, indicative of a stable operation.

The size of Fe2Als particles for the GI lines varies to a large extent from 15 to 65 btm with an overall
average of 34.08 btm. For the GA lines, the FezAls particles were smaller and the FeZnl0 particles were larger.
As would be expected, the transition from GI to GA operation takes a long time and the particles suspended in
the bath do not reach equilibrium immediately. For these cases, it is evident that the FezAls particles are in the
process of dissolution, whereas the FeZnl0 particles are in a nucleation and growth stage. Since the iron
solubility increases in the transition from GI to GA operations, the bath becomes supersaturated in iron which
becomes the driving force for the precipitation and growth of FeZn~0 particles.

It was observed that the number and size of the particles seem to decrease with increasing bath
temperature although the results of this study do not demonstrate a close correlation. Also, the number and size
seems to increase with the bath A1 content. This is consistent with particle growth from theoretical
considerations based on Ostwald ripening, and the same phenomenon was observed by Gagn62 who showed a
parabolic increase in particle size with decreasing bath temperature. He also found that the size increased
linearly with time after an increase of A1 content of the bath.

It is however difficult to quantify the particle size variation from the set of samples analysed in this
study. Sample sizes and location of sampling varied considerably for the 7 companies surveyed. Also, since the
solidification time increases with sample size, it would be expected that particles from the large samples would
be larger. The effect of sampling location and sample size is difficult to quantify due to the various sampling
techniques. It is quite clear however from company 6, that samples taken at mid-depth result in much smaller

778 44 th MWSP Conference Proceedings, Vol. XL, 2002

particles than from samples taken at the surface. The Fe2A15 particles are very slow to float and tend to
agglomerate into large irregular shapes at the surface of the bath.

The small variations in chemical analysis of the particles in the GA lines is proof of both the
reproducibility and reliability of the experimental method. Also, the narrow range of compositions is confirmed
from thermodynamic considerations. Analyses of particles from the same bath gave results with very little
difference, indicating that the compositions in the same bath are almost constant, and that sampling location
does not have any significant influence.

The results also clearly show that the composition of Fe2A15 particles form GI lines are different than
those from the GA lines which are generally higher in A1 and lower in Zn and Fe content. This variation is
attributed to the operating characteristics of each bath since the compositions in a specific bath are uniform but
are different for different baths. It would be expected that the FezA15 particles richer in A1 content would
originate from a bath of higher A1 content but this was not observed to be the case. If all baths had reported
effective A1 contents, a better correlation might have been possible. The composition of the FeZn~0 particles
were very consistent in all samples analyzed.

Analysis of particles of both Fe2A15 and FeZnl0 from this study are compared to the values from other
studies in Table 6 covering a 30 year period and which have used various different analytical techniques.
Considerable variation in composition can be observed in A1 (33 to 40%) in Fe (31 to 44%) and in Zn (12 to
26%). Variations in FeZnl0 are much smaller ranging from 2 to 4% A1, 5 to 9% Fe and 87 to 93% Zn.

Table 6: Comparison of Composition of FezAls and FeZn7 with Data from the Literature

Study Fe2A15 FeZn~0

% A1 %Fe % Zn % A1 %Fe % Zn
This study GA 46,20 31,10 22,71 2,54 5,05 92,42
This study GI 37,21 37,01 25,80
Tang 3 (1996) 40,50 39,30 20,20
Kato and Koumura 4 (1995) 40-43 38-40 13-17 2-4 6-9 87-91
Toussait and Segers 5 (1995) 40,11 38,20 21,66
Dauphin and Bretez 6 (1993) 33,16 42,13 24,71
Chen and Sharp7 (1990) 43,10 36,00 20,90 9,00 2,80 88,20
Urednicek and Kirkaldy 8 (1973) 42,90 43,70 13,40 8,20 3,60 88,20
Harvey and Mercer 9 (1973) 43,00 39,00 18,00
Ghuman and Goldstein lO (1971) 42-44 34-39 20-22

For the GA operations of this study the values for A1 content are higher and those of Fe are lower, whereas for
the GI lines the values are lower in Fe, higher in Zn and also lower in A1 than the values reported in Table 6.
The values for FeZn~0 particles in this study were slightly lower in A1 and Fe and higher in Zn. It should be
noted that the average compositions in this study represent a vary large number of both types of particles,
whereas the number of particles analysed in the literature results in generally much smaller.

CONCLUSIONS

1) Galvanizing (GI) lines with greater than 0.16% A1 generated Fe2A15 particles, whereas galvannealing
(GA) lines generally showed the presence of both FezA15 and FeZnl0 particles. GA lines operating
exclusively below 0.135% A1 generated only FeZnl0 particles.

44 th MWSP Conference Proceedings, Vol. XL, 2002 779

2) Particle sizes of Fe2Als for the GI lines varied from 15 to 64 gm in equivalent diameter but were
generally smaller than 30 jam for the GA lines. The surfaces of the Fe2A15 particles from the GA lines
were much less defined (non-faceted) and showed signs of dissolution, whereas the FeZn~0 particles in
the same bath were generally faceted and well formed, indicative of particle growth.

3) The chemical composition of both the Fe2Als and FeZnl0 particles in GA lines was much more stable
than the Fe2A15 particles from GI lines. This is attributed to the narrow range of both composition for the
GA lines as compared to the GI lines which varied from 0.17 to 0.28% A1. Compositions of the particles
varied slightly with bath chemistry and temperature and are generally consistent with values found in the
literature.

4) The chemistry of the dross particles corresponds to the composition that would be expected from
thermodynamic considerations where Fe2Als particles are present below the knee point (> 0.14% A1) and
FeZnl0 particles are predominant for baths with A1 content below this value. Both types of particles can
be found in GA lines which are in transition where the residual Fe2A5 particles are not completely
dissolved.

ACKNOWLEDGEMENTS

The authors wish to acknowledge all the companies who have participated in this study by supplying
samples form their operations.

REFERENCES

1) N.Y. Tang, G.R. Adams and P.S. Kolisnyk, "Galvannealing Bath Samples", Research Report, ILZRO
Project ZCO-3-2, (1998).
2) M. Gagn6, "Intermetallic Formation in Continuous Galvanizing Baths", Research Report, ILZRO
Project ZM-385, (1992).
3) N.Y. Tang and G.R. Adams, "Analysis of Steady State Galvanizing Bath Sampling Results", Research
Report, ILZRO Project ZCO-3-2, (1996).
4) C. Kato, N. Koumura, H. Mochigaki and N. Morito, "Dross Formation and Flow Phenomena in Molten
Zinc Bath ", Galvatech 1995, pp. 801-806, (1995).
5) P. Toussaint, P. Vernin, L. Segers, R. Winand and M. Dubois, "Experimental Study and Mathematical
Modeling of Zinc Ingot Melting Behavior in Continuous Hot-Dip Galvanizing Process", Galvatech
1995, pp. 663-666. (1995).
6) A. Gast-Bray, J.Y. Dauphin and M. Gretez, "Aluminum and Dross Repartition in Continuous Hot-Dip
Galvanizing Process", Proc. 85 th Meeting of the Galvanizers Association, Baltimore, pp. 1-25, (1993).
7) Z.W. Chen, R.M. Sharp and J.T. Gregory, "Lattice Parameters and Interplaner Spacings of Fe2A15 and
Fe2Als-Znx", Materials Forum, Vol. 14, pp. 130-136, (1990).
8) M. Urednicek and J.S. Kirkaldy, "An Investigation of the Phase Constitution form Zinc-Aluminum at
450°C '', Zeitschrift fur Metallkunde, Vol. 64 (6), pp. 419-427, (1973).
9) G.J. Harvey and P.D. Mercer, "Aluminum-Rich Alloy Layers Formed During the Hot-Dip Galvanizing
of Low Carbon Steel", Met. Trans., Vil. 4, pp. 619-621, (1973).
10) A.R.P. Ghuman and J.I. Goldstein, "Reaction Mechanism for the Coatings Formed During the Hot-
Dipping of Iron in 0 to 10 Pct A1-Zn Baths at 450°C to 700°C '', Met. Trans., Vol. 2, pp. 2903-2914,
(1971).

780 44 th MWSP Conference Proceedings, Vol. XL, 2002

View publication stats