- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Wed Nov 12 01:16:58 1997
To check the hypothesis that the blue glow is from phosphorus in
detergent, the cell was emptied, leaving the old electrodes and
wetted sides in order to transfer some small amount of NaOH solution
and to have the benefit of conditioned electrodes. Then distilled
water was added. The voltage was turned up ovor the 50% mark (about
900 V) and no glow was seen. The cell was allowed to run for about
15 minutes, but no light was visible. This part of the experiment
should be done again using a longer time frame and more careful
checking for the glow, maybe using a video.
Then some "Electrosol with BAKING SODA Automatic Dishwasher
Detergent" was added. The detergent is rated as no more than 6.1
percent phosphorus. However, it contained white granules and blue
granules. I added one blue granule and 4 white granules to the
cell. Set voltage so as to get about 10 mA, which was at 50 percent
(about 900 V). Gradually over about 5 minute period, as granules
dissolved, current rose to about 20 mA. I then turned out the lights
and thought maybe I could see a blue green glow. (As usual, I was
cowering about 8 feet away, operating by remote control extension
cord.) I turned the voltage up to 100 % and current to about 40 mA
and the electrode could clearly be seen to have a fairly uniform non-
sparking blue-green glow. I got up close and could see a few stable
unmoving non-blinking spots.
I then put the voltmeter on the cell. I measured the following
voltages and variac percents:
Var.
% Volts
10 190
20 363
30 540
40 710
50 873
I stopped there because my meter is only rated at 750 V.
I then tried to discern the glow onset voltage by quickly wiggling
the knob at the visibility threshold level. That level was 30-32
percent, roughly 540-576 V. ABove that brightness seemed
proportional to added voltage.
After running at 50 percent for about a half hour water was warm to
the touch, over 90 F, which was a surprise. It shouldn't have been a
surprise. At 550 cc that would represent about 6105 J for the
estimated 11.1 C. For 30 Min that's about 3.4 W. I was running at
874 V at 19.5 mA, so that's 17 W. No surprise there, but it will be
interresting to do a better job of measuring the heat output. There
were no visible bubbles. I looked at the spots again at 50 percent
and 100 percent. At 50 percent the spots could be seen to be
twinkling a bit, much more at 100 percent. The spots seemed somewhat
orange. There were some dark patches on the electrode surface that
migh have been from dirt or varnish.
This was just a rough experiment to try to rule out detergent, but
was unsuccessful in that objective. It was very successful in
obtaining a consistant blue glow, higher operating voltages, and no
sparks.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Wed Nov 12 11:49:40 1997
I decided to try some foil electrodes in the detergent solution.
Made some foil electrodes similar in size to the heavy 1 mm thick
electrodes. Put into the solution and immmediately noticed a much
higher condutivity. Might be partially due to the fact I left the
old electroides inthe cell overnight, thus some salts may have
leached out of the counditioned coating of the old electrodes.
There was no blue glow and no appreciable gas generated. Set at 30 %
and about 40 mA and let run for 1/2 hour. Then turned up voltage and
could see blue glow. Much more glow on the smaller area electrode.
Then I decided to shorten the larger area electrode. It glowed
more. I then shortened the smaller electrode also. I let the cell
run in the shortened condition for another 1/2 hour. I turned up the
voltage and got much in the way of bubbles. Good glow but no sparks.
The water was hot so I figured that maybe the bubbles were steam.
Both electrodes had the glow on both front and back, which I thought
was unusual.
To check the steam hypothesis I lowered the small electrode from its
current depth of about 1 cm back to its original depth of about 3
cm. The bubble generation moved with the 1 cm patch at the tip of
the foil and stayed with it. This clearly meant that the bubble
production was a function of the conditioning the surface had had.
The steam hypothesis was thus proven false, as the patch had been
lowered into a lower cooler region, and the patch did not carry
enough specific heat to maintain any heat flow.
Another interesting feature of the bubble generation is that the
bubbles seemed to be ejected in small jets. The small bubbles went
sideways for about a cm before starting to rise to the surface. The
foil moved about in the water as if propelled by the jets.
I turned off the lights and turned up the power to 100 % to see if
there was a noticeable difference between the 1 cm^2 end patch and
the rest of the electrode with respect to glow. There was. The rest
of the electrode had a uniform glow, while the 1 cm^2 patch was
mottled and clearly had much more in the way of glowing active spots
than the upper portion of the electrode (where there were almost no
bubbles being generated.) The bubbles quickly clouded up a wide area
of the upper 1-2 cm of the surface of the water so that the upper
portion of the electrode could not be seen. There was clearly a fast
current of water moving outward from the surface of the electrode,
and it moved about. The other electrode reactedsimilarly, but not
nearly as much. It is assumed that the larger area reduced the
effects of conditioning on that electrode.
This was primarily meant to be an experiment to get acquainted with
foil electrodes, and to check for the blue glow in non-conditioned
aluminum electrodes. The result is that electrode conditioning time
is clearly needed for even the detergent electrolyte cell, and the
time required and/or the results obtained appear to be a function of
the *current density* at the electrode surface during the conditioning.
Next step using foil electrodes is to try foil with NaOH electrolyte
and to attempt to get the underwater spark effect with foil electrodes.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Nov 12 20:26:24 1997
The following results are so tentative I almost didn't document
them. However, this info may help prepare for a following up.
Looking at my to do list for a quick experiment that might be done in
the time before picking up the kids form school I decided on a quick
test of the effect of magnets on glow/spots.
I started with the cell repoted in the "Report of first foil
experiment". It had cooled for a few hours. When I turned it on
there was notably more current used. At 50 percent it used 100mA.
The cells seem to always be different when you turn them back on,
esp. if you leave the electrodes in them while they are off.
I started up the cell at 50% voltage and it was glowing nicely in a
few minutes. Turned it off and put the magnet near the front
electrode of the cell, which faces inward so you only normally see
the back of the electrode, not the front part that normally glows.
The glow around the fringe of the electrode was clearly visible as an
outline. There was no apparent affect from a 35 MGO magnet roughly
1"x1"x2", comprised of 4 smaller magnets. I had to get up close to
see the glowing face, but it didn't look any different. Not much
happening, even when I turned up the voltage to 100%. I decided to
move the cell forward so the magnet could go in the rear of the cell
and affect the more visible glow on the front of the rear electrode.
The magnet was harder to manipulate when in the back, but I wasn't
planning on doing much of that while the power was on.
Then I got a surprise. I don't know if it was due to the magnet or
the big surge of power used in the prior step. However the back side
(as well as the front side) of the front electrode now had a very
good glow. The rear electrode appeared unaffected. I turned up the
voltage to 100% (about 1900 V) again. Same, just more bright.
Removed magnet from back. When I power up again both electrodes were
bright and lit on both front and back. I turned up the power and
heard that familiar sound of the underwater sparks. Went up close and
then verified the electrodes were changing over into spark mode, just
at a higher voltage than with the NaOH solution. Don't know for sure
what got them there, magnets, or the 100 % surge voltage, or
something else, but the electolyte was still weak detergent plus
small amount of contaminates from prior runs, and everything was the
same as prior when only the glow would show.
I have a feeling this experiment may be difficult to reproduce,
unless a strong magnetic field does have some effect on electrode
conditioning. If true, then something really wierd is going on.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Nov 13 04:26:52 1997
Using the solution from the experiment "First magnet effects?", and
two new electrodes, I ran the voltage up to 100 % and got about 142
mA current, but that rose to over 150 mA in less than half a minute.
There was no glow.
I set to 50 % voltage and 65 mA at 2204. Put magnet near front
electrode. There ws no glow. Did note a water current moving
*sideways* from the rear electrode. It appeared to be caused by a
process at the edge of the foil. There was some small bubble, or
turbidity, formation, esp. near the rear electrode. The rear
electrode was about 1 cm^2, the front about 1 cm wide by 2 cm deep.
By 2212 both electrodes glowed, but the back of the front electrode,
even though twice as big as the back electrode, was glowing
brighter. Current had dropped to about 55 mA and was still dropping.
Stopped for a minute at 2217 to put a mirror behind cell, and to
remove magnet. Both electrodes had grown brighter. When I turned
power back on at 50% voltage,
I didn't see much difference between the two electrodes, except both
were showing active spots around the periphery of the face andon th
edges.
By 2222 the "sparking" sound could clearly be heard. Current was
down to 40 mA. The bubbling was much subdued, but there were still
fluid currents visible about the rear electrode. Currents were about
like this viewed from the top:
electrode
============ -------->
/\ /
/ \
/ \
X | X |
V ^
Could be that edge effect is creating a vortex at X above that makes
it look like bubbles are jetting forward from electrode faces. Kind
of like high voltage needles make a wind away from the point tip.
Very strange that this could happen in water, though, as it is a
conductor, carrying no net charge.
At 2228 both electrodes were beginning to exhibit the moving spark
look, though the sites were clearly fixed. Most sparking activity
towards the edges.
By 2232 sparks were clearly visible in the light, though primarily
around the edges. Current was down to 29.4 mA.
At 2237 the current was down to 23.8 mA. Light from the discharge
was mostly from spots. Good flashing of the spots.
Happened to notice that foil electrodes from the first run, laying
there drying, looked corroded, dark grey, and warped.
At 2244 current was 20.6 mA. Meter turned off automatically due to
time out. Turned off for power for about 10 sec. to reset meter.
When I turned power back on on current was up to 34 mA.
Current began dropping. I ran voltage up to 100% for about 5 secnds,
and back to 50 %. Current only dropped to about 50 mA at 50%.
Another case fo surge conditioning of the electrode. AT 100 %
sparking and sound was intense, especially with the back electrode.
I think I saw a small fire like discharge at the surface near the
back electrode.
At 2250 both electrodes were very active and noisy and sparkly.
Current was 35 mA. Turned off transformer and noted that switch was
in up position, indicating that the last time I turned it on all
voltages were increased by 140/120, or about 16 percent. Means new
voltage record for me in these runs of about 2200 V. No wonder it
was so intense. Was too pre-occupied to observe current reading.
Well, the effects of a magnetic field on electrode conditioning are
still very inconclusive and even leaning toward null. The original
objective of the test is not clearly established, but it is clearly
established that (even) using weak detergent the spark effect can be
achieved at over 2200 V. This is starting to get up toward Claytor's
gas cell voltage range.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Wed Nov 19 00:07:18 1997
I got my oscilloscope attached to the experiment and have a 10 ohm 25
W resistor bridged for a current probe, giving 100 mA/V. I have a 1
ohm 25 W resistor, but it was easier to measure the resistance of the
10 ohm to 1%.
I tried using the scope with a cell containing the NaOH solution used
prior, and new foil electrodes.
The voltage waveform was a nice and clean sign wave, but the spikes
did show up on it when the sparking started. It looked kind of like
a rounded version of the following:
----
/ \
/ \
------ \
/ \
/..................\..........................
\ /
\ /
------- /
\ /
\ /
----
At the end of the experiment the scope measured .581 V rms, and 1.34
V pp on the current probe, giving 58.1 mA rms and 134 mA p-p or 67 mA
peak current. The DMM showed 53.9 mA at the time. The spikes were
very small. I measured the small spikes at only 15 mV, or
0.015V*100mA/V = 1.5 mA each, peak. I did not use a bypass capacitor.
Also got a 1 foot long closed end glass thermistor sleve built to
protect the thermistor probe from high voltage. Positioned
thermistor in the sleve at the middle depth of the cell. Used my Cole
Parmer digital thermometer. Noted that sparks turned on at about 30
C for this experiment.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Nov 20 00:36:05 1997
Test of Na2SiO3 at 0.1 g/l
I did a 90 minute test of 0.1 g/l Na2SiO3 electrolyte with new Al
foil electrodes. Used a 1000 pF 3000V bypass cap. Started out at
10% variac, 470 V (p-p) (about 127 V rms). Lots of bubbles
throughout run, even when sparks very active. Got excellent uniform
glow on both sides of electrodes, each about 2 cm^2, in only 9
minutes. Used strong magnet at distance of 1.5" to one electrode and
about 3" to the other. Don't know if magnet made any difference, but
electrode nearest magnet did end up with more active spots. Starting
temp was 24.5 C, volume was 500 cc.
A few spots occurred 38 minutes into the run. at 34.97 C. There were
more spots on the electrode near the magnet. The spots quickly went
through the foil (in about 7 minutes) and became little rings of
sparks. The rings did not grow very fast. This solution was one of
the least corrosive of the Na electrolytes. At 65 minutes into the
run I stopped to remove the oscilloscope voltage probe as the voltage
vas 1.32K (p-p) (474rms), and at 28 %. I realize now that I could
easily go twice that without removing the voltage probe. At 82
minutes the temperature was 48.76 C, and I pushed the voltage to 38%,
about 660 V rms. The sparking was intense and the noise was readily
audible over the din of the hood fan. At 90 min into the run there
were a lot of bubbles and the temp was 60.69. Current averaged 111 mA
for the final 8 minutes, and temp increased 11.93 C. That's 5965
cal, or 51.9 W for the 8 minute period. At 660 V rms and .111 A rms,
that about 73 W power in. This is one of the better energy ratios
I've observed. However, the energy estimate has errors of various
kinds including poor insulation and no stirring. Still, comparing
apples to apples, this was the best heat producing electrolyte so
far, and the best for producing the initial static blue glow. It was
one of the best bubble producers second to alum. Attempts to measure
radiation with a geiger counter met with no success, but there still
is a small possibility of a small count increase, as the geiger meter
did go into higher ranges when the sparks were on than not.
Changes I need to make or consider:
(1) fully submerged electrodes to avoid risk of igniting,
and heat lost to surface boiloff
(2) add a stirrer or go to boiloff calorimtry
(3) insulate cell
(4) try bigger area of electrode, thicker electrode
(5) try more dilute concentration
(6) make rod and cap configuration to do recombiation
(7) eventually interface to computer to get true rms I*V.
(8) if boiloff cell not used, then make electrode depth deeper
in order to absorb steam bubbles
(Note - I used a TDS220 (100 MHz with 1 GHz sample rate) with 2
standard P6112 10X (300V) probes and one P5200 2500 V(DC + peak AC)
probe. I have a spare 100 mHz 10X, 1X probe available for use with
the external trigger port.)
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Nov 26 00:00:34 1997
We noted before that with aluminum electrodes, with either low
conductivity or high conductivity electrolytes, that the glow, and
eventually sparking if sufficient voltage is used, does not occur
unless and until a coating of some kind is conditioned onto the
electrode surface, and that coating has a significant capacitance, or
at least phase shifting capability. Starting out with two fresh
electrodes the x-y plot is straight line. As a phase shift occurs,
and the line breaks into two parts, forming a diagonal eye shape, the
electrodes begin to glow. The phase shift seems to reach maximum
about the time the sparking begins.
To test the idea that a film forms on the electrodes, I removed one
electrode and replaced it with a fesh one, after the sparking was
going good, and a clean disticnc eye was formed on the x-y plot of
current (y) as function of voltage (x). This particular test was
with a weak electrolyte (using a weak combination of Li2SO4, Na2SiO3,
and detergent) so the active voltage was about 480 V rms, current
about 12 mA rms. The scope probe measured voltage on (was attached
to) the active electrode (the old one, not the one replaced with a
new one). The x-y plot immediately changed upon replacing an
electrode. The positive voltage segment looked as before - with half
an "eye", showing hysteresis on the positive part of the cycle. The
negative portion, however, was a straight line - except for a brief
part of the phase where the current leads the voltage into the
positive range. The split into two lines occurs below and before the
voltage crosses zero. This indicates that an oriented film is being
made, in that it acts like a diode. It conducts from the electrode to
the electrolyte when negative, but not as well when positive.
Looking at the normal XT and YT plots, the current dips negative much
further than to the positive, providing a net current flow, with the
active (old) electrode acting as a net cathode. When the active
electrode is max negative then negative current is max negative and
much larger than when the active sparking electrode is positive. Many
more and larger bubbles were created in this mode, especialy from the
new electrodes.
Evetually, as the new electrode conditions, the x-y plot returns the
symmetric eye configuration.
One one ocasion I pushed the voltage up to 2500 V p-p and noted that
the eye configuration changed, adding two more loops at the ends.
Beyond about +-1000 V, the current increased much more with
increasing voltage, but lagged on the return.
Using a strong electrolyte vs a weak one seems to only lower the
threshold voltages where the glow and then sparking occurs. However,
when I made too weak of an electrolyte solution of Li2SO4 and Na2SO3,
it didn't seem to want to start glowing, etc, until a few granules of
detergent were added. Then you could see the eye quickly form.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Nov 26 11:59:00 1997
A test of Zr electrodes was done using the weak Li2SO4, Na2SiO3, and
detergent solution of the prior "Polarity of hysteresis" experiment.
Conductivity was higher than Al but less than Mg, starting at 100.0
mA rms at 450 V rms (30.34 C) and dropping, despite with temperature
increase, to 43.7 mA rms at 476 V rms (45.54 C) in 27 minutes. A very
slight current phase lead was noted from the beginning which
continually expanded to reach 30.5 deg. at the conclusion. There
were many more bubbles than typical for Al electrodes at first, but
they diminished as the current dropped due to conditioning. The
experiment was stopped when an active sparking spot could be seen in
daylight.
Zr appears not as good as Al, but works fine. Careful calorimetry
should be done on it though, as in the initial 3 minutes it put out a
measured 41.45 W heat vs 42.9 W supplied, despite the fact the
electrolyte was not in a dewar, and was producing lots of bubbles.
The COP dropped rapidly with increasing temperature though due to
thermal losses and maybe changes in cell conditions. If it was ou
initially it would be difficult to make practical use due to the fact
the condition dissappears quickly.
I suspect that maximum heat is put out by both AL and Zr in the
intial time increments due to the fact that all the "excess heat" is
due to oxidation. The diode tests done earlier confirm the probably
essential nature of oxygen to the process of sparking and glowing,
and possibly conditioning. The use of foil electrodes makes it clear
that there is a sizeable consumption of Al happening. I suspect that
conditioning is a process of deepening the oxide layer. One problem
with this is the fact that conditioning does not seem to take place
if the electrolyte is weak enough, despite sufficient amp seconds and
watt seconds.
The fact that Al is being consumed in the process is probably enough
to turn off most investigators for pursuing further. However this
genre has so much in common with other supposedly ou things it feels
worth pursuing. In particular the "sparks", which really seem to me
to be brief arcs, suppressed by fast oxide formation at the active
site, do seem to be a mechanism for creating cavitation and thus
sonoluminescence.
It would be useful to focus a microscope, with a photomultiplier or
very sensitive photo receptor, on an active spot and measure
brightness compared to current, or compared to dI/dt during the spark
moment, which can be obtained using a small ferrite core transformer
with the primary inserted into the electrode current loop. I think
there could be a bright flash at the tail end of the spark bubble
collapse. This could help explain the destruction of the electrode
by the sparking process.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Nov 30 22:26:50 1997 Notes
(1) Aluminum burns hot. (The probable source of excess heat observed
is likely from oxidation of Al.) Its heat of combustion is 399,000
cal/g mole, or 7200,00 BTU/lb mole. With atomic weight of 26.98154
the energy is 14,800 cal/g. A 709.5 cm^2 sheet of Al foil used for
electrode material weighs 2.91 g. That is 4.10x10^-2 g/cm^2 or about
607 cal/cm^2 consumed. It was not unusual in an hour for about 1/2 of
the foil to be consumed on each of two electrodes, each about 2 cm^2
in area, so the heat generated would be about 1200 Cal. - enough Al
consumption to account for 2% ou reading? (Why the heck is a
"hydrogen economy such a big deal - we should got to the "Al
economy". Besides great availability and compactness, it burns
without air pollution!) Some solvent produced a white grainey powder
at the bottom of the cell which needs to be explored further. Might
be blown off the surface by cavitation shock.
(2) The Al2O3 layer conducts electrically, so the insulator is
something else, or it is an Al2O3 layer that is thicker due to
conditioning. To test I simply measured the resistance of the Al
electrodes. There was no noticible resistance though the current has
to go through the Al2O3 layer twice to get through the electrode.
(3) The fact plain H2O alone doesn't create the insulating layer
suggests Al2O3 is not the insulating layer. Some electrolyte appears
necessary. Alum didn't work. Baking soda didn't work.
(4) The insulating layer acts as a diode.
(5) The two opposing layers conditioned on opposing electrodes by AC
make the cell act as a capacitor.
(6) The insulating layer in some cases works to over 1000 V peak.
(7) There is a phase shift that builds as the electrode is
conditioned. This pahse shift is due to the induced capacitance due
to the opposing diode effect. The phase shift is critical to measure
when determining power input because the true power input is not as
large as otherwise measured unless a true integral of I*V is determined.
(8) Successful electrolytes tested on Al electrodes contained either
Li or Na salts and were alkaline.
(9) Na salts can condition the Zr electrode. No Li test yet made.
(10) Al conditioned with N2SiO3 has a very good insulator, and few
active spots, and they tend to be in the center of the electrode -
indicating the edges are quickly conditioned. The active spots
turned in to holes of active spots eating away the outer rim of the
hole. Very interesting possibilities.
(11) Zr appeared ou during initial conditioning.
(12) Upon mesuring the resistance of the films on various Al and Zr
electrodes it was noted that the Zr electrodes had a very strong
insulator (infinite) not easily poked through with the probes as the
Al insulator was.
(13) Upon looking at the Al electrodes under a microscope, it
appeared the active spots were shiney crystals on the rim of, or on
raised areas in, holes in the surface insulating layer about 20 to 60
microns wide, and about 5 microns deep. The crystals were from 1 to
20 microns across and often in close groups of 3 or 4. The normal
unconditioned foil was shiney and had 1 micron grooves with 1 micron
elevated ridges. There were roughly 1 micron bright spots atop the
ridges in the unconditioned foil. The conditioned foil looked white
to the eye but dark and cratered under the microscope, with no ridges
visible. The diameter of the visible area under the microscope was
1.1 mm.
(14) A strong magnetic field (but under .5 T) appeared to cause
active spots to form on the back side of the electrodes. This will
have to be examined more carefully to see if (a) the effect is real
and (b) if the spots simply bored through the foil from the other
side (faster.)
(15) During the conditioning process the eventually successful
electrodes glow a blue or blue green.
(16) Electrodes made of solder did not glow or form underwater sparks
or a non-conductive coating but produced a lot of bubbles and heat.
The heat probably due to the high conductivity of the electrodes, but
should be looked at again carefully.
(17) When a fullwave bridge was inserted in glowing circuit the glow
went with the anode. This indicates the glow is probably related to
oxidation of the Al. More carefull experimentation on this needs to
be done with respect to sparks.
(18) As the sparks begin the glow subsides. A sign that the
conditioning is complete and only a few holes remain in the surface
for conductivity. This is an indication that bubble formation and
closeoff probably limits the conduction time for an active spot. This
jives with the fact that there are bright crystals (probably
conductive) visible on the surface. Need to test conductivity of
bright spots with fine needle. Would be useful to do SIMS on the
spots to see if they are Al or an alloy or what.
(19) Alum, K2Al2(SO4)4*34(H2O), molecular weight of 808.57, (the new
element K replaces Li or Na) did not work well as an electrolyte.
Need to try lower concentrations. Too conductive. Should get some K
in the electolyte somehow. KNO3 not so hot in some tests but nothing
conclusive there either. KOH with LiOH is the obvious thing to try.
(20) Some candidates for Li based "conditioning" of Al electrodes are
spodumene (LiAlSi2O6) or petalite (LiAlSi4O10). I think spodumene is
the likely candidate for formation from a Na2SiO3 electrolyte, which
seemed to create the most effective insulator.
(21) Once conditioning is complete and the cell is sparking, ramming
the voltage up to over 1000 V rms creates more spots which remain
when the voltage is dropped back to the more normal 400-500 V rms
range. It is a form of superconditioning.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Best regards,
Horace Heffner
http://www.mtaonline.net/~hheffner/
