font face=arial size=-1 color=black>Yahoo! is not affiliated with the authors of this page or responsible for its content.
CORRECTING INHERENT IMBALANCE AND CONSEQUENT FAILURE OF VRLA CELLS BY THE USE OF CATALYSTS 1. Introduction
VRLA cells have been very successful in replacing flooded cells
in long-life telecom and UPS applications. Much of this success
is due to assumptions that VRLA cells will provide 20 years of
service without the addition of water. However, despite the fact
that these cells are often built to superior quality standards, their
actual performance in service has been problematic, with
persistent reports of early failures.
Evidence from our own laboratory testing, taken over a 2 year
period, tended to confirm the field reports. The tests indicated
two possible failure modes:
-
First, the cells were emitting hydrogen at rates that
were generally too high to sustain a 20 year life without
loss of capacity due to dryout.
-
Second, and of more concern, the negative plates of
some of these cells appeared to be discharging under
long term, steady-state float with the possible result of
negative plate capacity failure.
The purpose of the present work was to find out what was
happening inside the VRLA cell that would explain these
apparent problems and, if possible, to find practical solutions.
2. Summary of long-term float test.
This test has been described in detail elsewhere [1]. Three pairs
of AGM cells of three different designs from two separate
manufacturers were selected for the test. They were all long-life
cells designed for 20 years of maintenance free service. Each
pair was floated, relatively undisturbed, for a period of two years
at an average voltage of 2.27 volts per cell at a steady
temperature of 80o F (27o C). All cells were equipped with gas
collection apparatus and reference electrodes.
We had assumed initially that the cells would fail from dryout and
our original intent was simply to calculate the rate of dryout from
the hydrogen emissions. Based on the amount of water
available in the typical VRLA cell, the cells had a maximum
permissible gas emission rate of 20 ml per day per 100 Ah of
capacity. That is, if any cell emitted more than this amount of
gas, it would dry out before its 20 year life goal.
The results of the test were pessimistic. The gas emission rates
on 5 out of the 6 cells were well above the 20 ml per day target.
In one case, it was twice the target value, indicating an expected
dryout life of only 10 years.
Gas analysis showed that the emissions were largely hydrogen.
This was expected because, in accordance with conventional
understanding of VRLA cell behavior, positive grid corrosion
removes oxygen from the oxygen cycle and an equivalent
amount of hydrogen will be generated at the negative plate and
emitted form the cell. What was unexpected, however, was that
pairs of identical cells would have gassing rates varying by a
factor of 2:1. It seemed very unlikely that the positive corrosion
rates would vary by this amount. We had not realized at that time
that our cells were "unbalanced" - a concept we will describe
later.
Reference electrode readings indicated that the negative plate
potentials were generally slightly below open circuit values; but
this seems typical of these kinds of cells so we were not
concerned. We did become concerned, however, when some
cells showed a serious and steady decline in conductance over
the test period [1]. This strongly suggested that negative plate
discharge was taking place under steady-state float conditions.
Since this reaction can evolve hydrogen, we suspected that the
excess hydrogen emission was the result of this reaction.
Calculations showed that if negative discharge and dryout are
taking place simultaneously, then, for a given amount of
hydrogen emitted, negative discharge will be the first to limit
capacity.
Our concern appeared justified when, at the end of the two year
test period, a capacity test was carried out on two cells in our
test series. Both cells failed to meet the minimum 80% capacity
target; one cell yielded 75% capacity while its twin yielded only
60%. The results are shown in Figure 1. Both cells failed by
negative limitation as measured by their reference electrodes.
This was unmistakable proof the negative plates on these cells
were discharging on steady-state float. The question now was:
what was the cause?
The test itself had done nothing unusual to the cells that would
explain such a continuous loss of capacity - no discharges, no
long open circuit periods, no high temperature excursions -
indeed, no disturbances of any kind. The positive grids of these
cells were made from a non-antimonial alloy so capacity
reduction cannot be blamed on antimony transfer. (Note: There
was some tin in the positive alloy but normally this is not
considered a negative impurity). A prime suspect, namely
leakage of air into the cells during the test, can also be excluded
because these cells were emitting gas into collection apparatus
during the entire test period and were pressurized to 1.5 psi (100
millibars) at all times. Further, since the conductance readings
had shown continuous rather than sudden decline, the
mechanism of discharge also had to be a continuous process
rather than a sudden event.
CORRECTING INHERENT IMBALANCE AND CONSEQUENT FAILURE OF VRLA
CELLS BY THE USE OF CATALYSTS
William E. M. Jones, Philadelphia Scientific, Lansdale , Pennsylvania, USA
Dr. D. 0 Feder, Electrochemical Energy Storage Systems, Madison, New Jersey, USA
Laboratory tests indicate that high quality AGM cells, designed for 20 year standby service, can fail from negative plate
discharge even under steady state float conditions. This paper shows that such discharge is not only possible but inevitable
if these cells are "out of balance". For a cell to be in balance the hydrogen equivalent of the positive grid corrosion must not
be less than the hydrogen emitted by the negative plate. Imbalance results in excess or unrecombined oxygen being generated
in the cell which can discharge the negative plate. A catalyst placed in the cell can correct imbalance thereby preventing
negative discharge, reducing water loss and increasing the useful life of the cell, especially in hot operating environments. Two possible explanations for the negative plate capacity loss
remained, involving two different kinds of discharge:
-
Self discharge due to impurities in the negative active
material.
-
Discharge due to oxygen produced inside the cell.
Figure 1: Five hour rate capacity tests showing negative plate
limitation after 2 years on float at 2.27 vpc. Cell 6 gave 75% and
Cell 7 gave 60% capacity.
3. Basics revisited
The design of VRLA cells is different from that of flooded cells in
one crucial respect: gas management. In a flooded cell, the
gasses produced during charging simply leave the cell and can
generally be ignored. (Note: The gas evolution in a flooded cell
must be compensated for by periodic water addition, but it is not
an aging factor for the cell.)
In a VRLA cell, on the other hand, the success or failure of the
cell is largely defined by how much hydrogen and oxygen it
produces and at what rate these gases leave the cell. Yet gas
measurements have received little attention in the literature. The
following is a review of some basic VRLA topics from the
perspective of gas management.
3.1 Oxygen cycle. In the oxygen cycle, everything begins at
the positive plate where water is broken down into three
components:
-
oxygen gas which travels in the gas space
-
hydrogen ions which travel in the liquid
-
electrons which travel in the electrical circuit
All three components converge on the negative plate where they
recombine into water; the water then travels back toward the
positive plate to complete the cycle. The oxygen gas
depolarizes the negative plate to approximately its open circuit
value. The hydrogen evolution rate is thereby reduced to a value
close to the self discharge rate which results in the cell having a
much reduced water consumption.
3.2 Hydrogen oxidation or recombination. For present
purposes, we agree with the conventional wisdom that VRLA
cells do not normally recombine hydrogen gas (only hydrogen
ions). If hydrogen gas is generated, it must leave the cell. (Note:
For the record, we have seen examples of cells that appear to
be recombining hydrogen but they are the exception, not the
rule).
3.3 Negative plate discharge. The negative active material
plays a much larger role in the design balance of VRLA cells
than it does in flooded cells. Its inherent tendency to produce
hydrogen at all normal voltages, thereby consuming water, is in
conflict with the limited water loss allowed in a VRLA cell.
Oxygen depolarization does not eliminate hydrogen evolution of
negative plates but reduces it to its minimum level possible,
namely, its open circuit value. This limiting value has been
clearly stated by Berndt [2] and is so important to the
understanding of the VRLA cell that we will repeat it here as a
basic law:
The open circuit self-discharge of the negative plate is a simple,
independent chemical reaction. Yet, in practice, it indirectly
determines the service life of many, if not all, long-life VRLA
cells. The argument is as follows:
-
In all the long-life AGM cells we tested, the hydrogen
emission rates on float (controlled by the negative
plates) were much higher than the positive grid
corrosion rates.
-
Therefore, if only two modes of failure were possible,
these cells would fail by dryout long before the positive
grids corroded. (Note: This assumes, of course, that
they