NCER
Assistance Agreement Final Project Executive Summary
Date
of Final Report: August
10, 2008
EPA
Agreement Number: X-83254101-1
Center:
Center
for Environmental and Energy Research (CEER)
Project
Title: Emissions Reduction of Commercial
Glassmaking Using Selective Batching
Investigator(s): William M. Carty
Institution(s)
of PI(s): Alfred
University
Research
Category: Congressionally
Mandated Center
Project
Period: September 1, 2006 – May 28, 2008
Description
and Objective of Project:
Selective batching involves separating a glass
batch into two or more, chemically unique mixtures. When the two mixtures are added together, they yield the
final batch chemistry. Selective
batching technologies offer a substantial opportunity to close the gap between theoretical
and practical glass melting energy consumption.
Selective batching controls reaction paths to
avoid a eutectic liquid forming between sodium carbonate and calcium carbonate. In conventionally batched melts, this
leads to high temperature melt segregation. Batch free time reductions (up to 80%) have been achieved
using selective batching. Batch
free time is defined as the amount of time a glass batch takes to complete the
dissolution of all batch ingredients.
While
increased efficiency has been observed, the fining behavior of selectively
batched melts has yet to be investigated.
Fining refers to the removal of bubbles present in the glass as
undissolved gas, and implies a homogenous glass. It is proposed that selective batching can improve on fining
in two ways (1) by yielding a developing melt whose viscosity is more
consistent and higher during the melting process, and (2) by tailoring the size
of the voids between granules in a selective batch to achieve fast fining by
enhanced StokesÕ fining.
Summary
of Findings:
The vertical bubble populations of selectively
batched melts were compared to the vertical bubble populations of
conventionally batched melts.
ÒConventionalÓ refers to the use of a powdered batch. Bubble position and diameter
measurements were taken on 24 crucibles encompassing five trials. The glass composition studied was a
generic float glass (Table I). Three
melts at 1350¼C revealed distinguishable differences in bubble diameter, bubble
concentration, and gas volume percent results between conventional and
selective trials (Figures 1-3).
Table I. Glass Composition Studied.
|
Oxide |
Mol % |
Weight % |
|
SiO2 |
71.9 |
73.0 |
|
Na2O |
13.2 |
13.0 |
|
CaO |
10.5 |
10.2 |
|
MgO |
4.3 |
3.0 |
|
K2O |
0.13 |
0.2 |
|
Al2O3 |
0.06 |
0.1 |

Figure 1. Mean bubble diameter of the
distribution as a function of melting time. The conventional trial shows the largest mean bubble
diameters throughout the time scale shown, while the Selectivet
trial has larger mean bubbles than the Selectivew trial.

Figure 2. Bubble
concentration (N) per cubic centimeter of glass. The Conventional trial at 1350¼C has the fewest number of
bubbles throughout the time scale measured.

Figure 3. The (undissolved) gas content for each
melting trial is displayed and compared against one another. The lines between symbols are a guide
for the eye.
The selectively batched trials contained a
higher concentration of bubbles than conventionally batched melts. The mean bubble diameter was smaller in
selectively batched melts than conventionally batched melts. This means more, smaller bubbles.
Vertical bubble position versus bubble diameter
were plotted as an alternative, static approach for evaluating fining
behavior. The plots revealed that
conventionally batched melts contain significant (up to one third of the melt
depth) regions with few to no bubbles while selectively batched melts all
exhibited consistent vertical bubble distributions.
In response to these findings, the viscosity of the
melts was calculated using the Vogel-Fulcher-Tammann Equation. The calculation was made twice for each
sample, using two sets of data.
Hot stage microscopy and dilatometry measurements were used to arrive at
one set of VFT parameters, while chemical analysis data were used to arrive at
another (Table II).
Table II. Calculated Viscosity Results for
Samples Melted at 1350¼C for 5 hours
|
Sample |
Method |
|
|
HSM/Dilatometer (Pa*s) |
ICP (Pa*s) |
|
|
Conventional |
103.65 |
101.28 |
|
Selective, truncated GSD |
102.52 |
101.37 |
|
Selective, wide GSD |
102.84 |
101.33 |
The
abnormally high viscosity result for the conventional trial measured using HSM
and dilatometry can be explained by sample preparation differences. HSM and chemical analysis samples are
powdered glass and are therefore more homogenized when compared to the bulk
glass. The dilatometer samples are
not homogenized and represent the true homogeneity of the bulk glass. The Selectively batched melts return
consistent results because selective batching results in a more consistent
developing viscosity profile.
Selectively
batched melts contain a consistent, vertical bubble population as a result of
two granules with two starting viscosities that are much higher than the
viscosity of the eutectic liquid formed in a segregated melt (Figure 4). The low viscosity liquid in a
selectively batched melt (granule #1) is on the same order of magnitude as the
viscosity of the final melt. The starting
viscosities in a selectively batched melt are also closer together than the
viscosities of the two regions in a segregated melt.


Figure 4.
Proposed dynamic viscosity model of conventionally and selectively
batched melts at 1350¼C. Each melt
is assumed to have two starting viscosity reference points. G2Õs viscosity is undefined at the
melting temperature of 1350¼C.
Conclusions:
Selectively batched melts result in a more
homogenous melt earlier on in the melting process; the Òpre-homogenizationÓ of
selectively batched melts can help reduce furnace residence times, leading to
energy savings. In contemporary,
industrial glass melting operations, homogenization is achieved by increasing
the temperature of the furnace and stirring the melt. Selective batching could reduce or eliminate the need to
increase the temperature of the furnace and/or stir the melt.
Understanding high temperature melt segregation
in conventional batches is central to understanding bubble populations in both
types of melts. By tailoring the
granules size/distribution and viscosity of the granules used in a selectively
batched melt, bubble populations can be altered. Selective batching can close the gap between practical and
theoretical melting energy requirements in industrial glass melting operations
by achieving a more homogenized glass earlier in the melting process, and at
lower temperatures.
Publications/Presentations:
D.
Fiordimalva, The Fining Behavior of a Selectively Batched, Generic Float
Glass Composition,
M. S. Thesis, Alfred University, 2008.
Supplemental
Keywords: selective
batching, batch segregation, batch free time, hot stage microscope, HSM,
Vogel-Fulcher-Tammann equation, VFT model, high temperature melt segregation,
fining, green house gas emissions, CO2
Relevant
Web Sites: http://ceer.alfred.edu