Cutting Aluminum with Oxy/Act...

By Stephan (***)

Date 06-07-2008 18:29 Edited 06-07-2008 22:26

Plasma-Brain,

you are quite near the mark...

I guess some others may explain it better, but without making it too tricky...

You can distinguish between:

1. Oxy-Fuel Cutting and

2. Fusion Cutting (e.g. Plasma Arc Cutting)

As by using the example unalloyed steel.

Unalloyed steel "combusts" when increasing its temperature upon its "ignition point" which lies at ~ 1100 °C when in contact with pure oxygen. The result is heat + slag. In contrast unalloyed steel melts at ~ 1500 °C.

However, not every metallic material is applicable for oxy-fuel cutting.

Oxy-fuel cutting applicable materials must meet the following 4 conditions:

a. The material must combust within the oxygen stream

b. The materials "ignition point" must lie below its melting point

c. The slag emerging from the combustion must be inviscid

d. The material to be cut should have a low thermal conductivity

Materials being applicable for oxy-fuel cutting are e.g.

- Unalloyed steel
- Low alloyed steel
- Cast steel
- Titanium

Now let's have a look upon some materials and their capability to be oxy-fuel cut. Always in mind, to meet this ability the material must meet the conditions as mentioned above.

- Aluminum - does not meet condition "b" (Oxide's melting point > Al melting point)
- Stainless steel - does not meet condition "c" (high viscid slag)
- Copper - does not meet condition "b" (Oxide's melting point > Cu melting point)
- Grey cast iron - does not meet condition "b" (Oxide's melting point > Iron melting point)

These materials however are capable to be fusion or plasma cut, respectively.

Best regards,
Stephan

By 803056

Date 06-08-2008 13:36

A good explanation.

Best regards - Al

By Stephan (***)

Date 06-09-2008 07:12 Edited 06-09-2008 10:26

Al,

such comments, made by you, mean honor for the post's author!! :-):-):-)

My best regards,
Stephan

Edit: Have added two more smileys, no... wait... 3 more :-)

By Plasma-Brain (**)

Date 06-08-2008 15:00

Indeed, very good explanation.
So basically you run into the same issue if you try and weld on a piece of aluminum without taking the oxide layer off first.
The oxides melting point is higher than the base metal so the base metal melts before the oxide has a chance too?
And with stainless steels the slag formed won't flow out of the cut area rapidly enough to allow for cutting?
Thank you Stephan
-Clif

By Stephan (***)

Date 06-09-2008 07:28

Clif,

thanks a lot for your kind reply!

The latter I basically would like to agree with.

The first named (aluminum oxide) whereas I would like to - at least partially - restrict.

Aluminum welding without aluminum oxide is a hard nut to crack.

On a first view one could mean it's a kind of contradiction, since aluminum oxide is actually an electrical isolator, but...

Since Al2O3 - even aluminum oxide - has a lower electron work function compared with pure aluminum (Al2O3 = 1.77 eV / Al = 3.95 eV) it is easier to "detach" electrons from the aluminum oxide surface layer.

This means, normally the aluminum oxide layer improves the welding process by up to a specific limit, which is again similar to a specific layer thickness.

I have just a minute for the forum but I hope to come back later on the day. I am certain to have some pictures anywhere showing how the aluminum oxide has a positive influence on the weld.

When I should find the time and the pictures, I'll post them later on.

However, basically you are nonetheless right when saying:

"The oxides melting point is higher than the base metal so the base metal melts before the oxide has a chance to."

Best regards,
Stephan

By AnaS Date 06-27-2008 08:13

Hi.

I am not a welder, but a physicist. Actually I'm a PhD student, and I do plasma physics research.

There is not much data on work functions of non-metallic compounds. Actually, this is the only place I could find the electron function for aluminum oxide. Could you please give me the reference for it? Was it an article? A book? I am currently writing a paper and this would really help me.

Best regards,

Ana

By Stephan (***)

Date 06-27-2008 11:56

Hello Ana!

Thanks for questioning this interesting item.

To be honest with you.

I can remember well the chat I had with my fellows (amongst them a chemist, a physicist and a mathematician - what a combination!) where we have discussed my question "How can an electron escape easier from a dielectric than from a metal to start the ionization of an arc plasma?" By the way, this discussion was based upon the question if a Laser can induce a plasma upon a material's surface - of course it can - but in how far even this plasma may act regulative and interactive to another plasma coming e.g. from an electrical welding arc (as we can see it with Laser-GMA-Hybrid Welding). In other words, does the second (arc) plasma work only properly when being centered towards the Laser induced plasma?

Well, my unaware uderstanding that time - you see, I am a welder (proud of it :-)) and no physicist - was, that there should exist a specific likelihood an electron may escape as well from a dielectric, even based upon Heisenberg's indeterminacy principle. You certainly may imagine, it was a great discussion, but the bottom line at that time was, that most likely the reduced binding energy level of the outermost electrons might be the reason for the lowered work function of the oxide layer. In this case even Al2O3.

I can remind as well, as I have asked my colleagues (I know the question might occur silly and bothering but however, it's even me) if when we would imagine to have an oxide layer of a "very large" thickness, the layer's electron work function value might change across its thickness, i.e. having a lower work function again at the surface vs. the center. Hmm... no one could answer this question. What the fellas told me whereas was that the physical properties of metal oxides are quite intricate to deal with. And by reading what you have so kindly mentioned, I guess they were right by saying so.

But we've had enough of words.

To reply your question regarding the reference.

The reference I have used for making my predication was:

HAAS, B.
"Schutzgasschweißen von Aluminiumwerkstoffen" (in German)
Jahrbuch Schweißtechnik '94
pp. 114-123
Deutscher Verlag für Schweißtechnik, DVS-Verlag,
Düsseldorf, 1994

The author has used as references again as follows:

- Merkblatt DVS 0933 (09.91) "MIG-Schweißen von Aluminium, Werkstoffe, Schweißparameter". Deutscher Verlag für Schweißtechnik, DVS Verlag, Düsseldorf (in German) and

- Merkblatt DVS 0913 (06.86) "Metall-Inertgasschweißen von Aluminium" Deutscher Verlag für Schweißtechnik, DVS Verlag, Düsseldorf (in German)

As I have stated when discussing Cliff's topic, HAAS has worked with even the data as I have taken it from his paper. And that was:

Al2O3 = 1.77 eV and
Al = 3.95 eV.

I have read once a very interesting PhD-Thesis (in German) on the influence of chemical active shielding gas constituents in Argon for GMA-Welding of aluminum and aluminum alloy base materials. It was repeated there that metal oxides may have lower electron work functions compared with the metals they are emerging on, and he has repeated this as well for aluminum. But unfortunetly he made no further predications on the references his statements were based upon.

Sorry for not being able to say more, but nonetheless I hope it helps.

Best regards,
Stephan

P.S. I guess it is a great honor for all of us that you have found the way to this the world's best welding forum. Saying this both as a welder + as an admirer of your profession!

By AnaS Date 07-01-2008 12:46 Edited 07-01-2008 12:53

Hi, Stephan!

Thanks for such an elaborate reply. It is indeed helpful, right now I'm trying to get the book in your reference.

I am currently employed by the Eindhoven University of Technology in The Netherlands, working in group called EPG (Elementary Processes in Gas Discharges). The work I do is related to the very transition from gas to plasma.

You probably know that when you take some volume of some gas, stick two electrodes in it, ground one and put voltage on the other one, the gas gets ionized and a conductive channel will form between the electrodes, thus turning the gas in a conductor. In high pressures (around atmospheric pressure), the channel will be thin. At low pressures (a few milibar), the channel will in fact be a big, diffuse thing, that we cannot call a channel any more because the bigger part of the gas is then ionized.

After this starting phase, in high pressures, an arc develops, similar to the welding arc, I assume.

But, I'm interested in the starting phase. It is hard to do measurements because everything happens in nanoseconds, and it is quite unpredictable. This is why it is still not fully understood. And I'm looking at the interaction between plasma in its starting phase and dielectric surfaces. The thing is that people have noticed that if you have a piece of dielectric material close to the electrodes, it will influence the starting phase.

As I'm currently writing a paper on the subject, I can't tell you about my measurements. Anyway, I have already talked about it enough, I don't know if you're even interested in it.

If I use the reference in my paper and if my boss allows it, you'll be acknowledged. It is so nice to know that someone else but physicists actually likes what you do!

Thanks again.

By rlitman (***)

Date 07-01-2008 15:47

Well, I for one find this quite interesting, even if its far from the original topic.

As I see it, up until the initiation of the plasma, you're describing a capacitor, so, clearly, the introduction of a dielectric in proximity to the electrodes, will alter the permittivity in the voltage field, which will lower the voltage across the surfaces for a given charge density. I would think having a higher charge density, would also lower the work function.

You have to remember that when we talk about Aluminum welding here, its typically A/C, so the Al2O3 only comes into play in the half phase when it is the cathode.
In DC welding, the reduced work function of oxides has been put to use for a long time, as additives to the tungsten used in GTAW.
Cerium oxide is the most extreme example, and it makes arc starting the easiest, but numerous other oxides have been used as well.
I'm curious as to whether the radioactivity of thorium oxide, also helps with the plasma initiation.

By Stephan (***)

Date 07-02-2008 15:44 Edited 07-02-2008 15:51

Hi Ana,

Wooow!

Thanks a lot for both the honoring words and your kind descriptions on what you are performing (could make me quite envious :-))

This is truly an extraordinary^2 field of research you are working on and surely much too much to find out for only one lifetime.

You know I truly love "swimming in the great ocean" of arc physics (even though I am just a devotionally astonishing layman) and the honest and hard try to learn and learn and learn and absorb everything I can get thereof. :-)

However, as I have read what rlitman has stated I have remembered a paper coming from TANAKA et al and been held with the Annual International Institute of Welding (IIW) Assembly 2006 in Quebec Canada. It has been presented within the IIW Study Group 212 (Physics of Welding) and I had the outstanding honor to have participated in the meeting that time.

The paper is called "Predictions of Current Attachment at Thermionic Cathode for Gas Tungsten Arc at Atmospheric Pressure" and has got the IIW Document Number : "IIW Doc. 212-1091-06".

It deals explicitely with the effects rare earth metal oxides (La2O3, Ce2O3) as well as the actinid oxide ThO2 do have with the reduction of the work function of tungsten used in Gas Shielded Tungsten Arc (DC) Welding. It is immense interesting and I just can recommend to get this paper eventually as well. TANAKA has additionally listed some very good references finally in his paper, perhaps being likewise helpful for you in your research efforts. Perhaps you might know it already but I'd like to state one particular reference hereinafter:

V.S. FOMENKO: Emission Properties of Materials, Kiev, Naukova Duma, (1970)

I mean to know, by having lots of instances proving this, that the Russians have carried out loads of loads of fundamental research (cohering with their cold war activities) in material science. Perhaps you may find what you're looking for...

Another suggestion from my side, if you allow, I have the great honor to attend the IIW Annual Assembly 2008 to hold in Graz Austria next week(!).

I am attending as well the SG 212 meetings and I will meet Prof. Tanaka personally then. Besides him many others of the greatest living arc physicists (e.g. John J. Lowke(!)) are participating as well over there. I am sure to find the opportunity to talk to him a/o them. Since you have awoke my greatest interest in where the values of work functions do actually emerge from, please let me ask those fantastic people on where they have their values from or at least if they could recommend some sources of knowledge and information to this interesting field.

Furthermore I can just recommend to have a look into the IIW database (in particular SG 212). There is a tremendous amount of information most likely being interesting for you and your research activities.

By the way, I have of course already asked my fellows (PhD plasma physicists... I love them :-)) within our company if they were able to explain where the values of metal oxide work functions do come from. Huuh, they have answered that there is extremely less information on this available (again approving your statement) and the values being available reach - partially - back to the 1930's(!). I am wondering how these values have been measured at that time and how great their accuracy and thus their reliability truly is.

Finally and last but not least, please let me attach an interesting link (probably you'll know it already) showing the path to the PhD Thesis of WENDELSTORF: http://www.wendelstorf.de/doc/2000/jw3282.00/index.html

I am sure - as far as you may not know it already - you will like it...

Well, if you may come back to this world's best welding forum it would be surely a pleasure for all of us...

I promise to ask Prof. Tanaka and will try to discuss the interesting issue of work functions of metal oxides with him a/o the others as I get the opportunity. Subsequently I promise to drop a short note of the outcome here in the forum.

Best regards and thanks again for your kind reply,
Stephan

By AnaS Date 07-07-2008 09:15

The change in the electric field due to the permittivity of the dielectric is indeed important for the initial formation of streamers (thin, ionized channels) in gas and the direction of their propagation, but later some other effects kick in, that have much bigger effect than the influence of the dielectric on the electric field. I will be able to discuss this further as soon as the paper is published, but not sooner.

In welding, if I understand this correctly, you charge an electrode in a welding machine and use the substrate (material you want to weld) as ground, which makes a capacitive plasma. So, when you use AC power on your charged electrode, don't you have trouble on polarity crossings? Your arc doesn't extinguish? And why do you use AC power? DC with positive charge on the electrode would produce nice, thin and powerful arcs. Wouldn't it?

As to radioactivity, it does indeed help plasma initiation. When you apply voltage on your electrodes, plasma does not begin to form immediately. It has a delay, which consists of two effects. On of them is formative lag time, and the other one is statistical lag time. The latter is essentially the delay caused by the fact that you need an initial electron in your gas, which will start the electron avalanche and start the ignition process. There is always an associated probability that the electron will start an avalanche, which is less than 100%, which means that you need a number of initial electrons in your gas. This is where the radioactivity comes into play. It ensures that more free electrons are present in your gas prior to the ignition process, thus lowering your lag time, and increasing probability of a successful ignition.

Stephan, thanks for recommending the Fomenko book, I found an American edition and it contains the data I was looking for. However, the number is slightly different here (4.7 eV for Al2O3 compared to 4.25 eV for Al). This really shows the spread of values which a work function of the dielectric can have. I personally think there are several resons for that. One is clearly that Al has a strong affinity to oxygen and forms several oxides. So, when one says "aluminum oxide", you can't know which aluminum oxide is it, before you see the chemical formula. The second one is that you will rarely find a dielectric (or any compound, for that matter), of perfect purity. The impurities can influence the work function, depending for example on their electronegativity. So, you get a spread of values.

Thanks, I have an idea of the Al2O3 work function, which is really helpful.

Thanks for the PhD thesis as well. i couldn't find the paper, but I know some people investigating arc attachment. A plasma physics group from Bochum University in Germany lead by prof. Awakowicz published a few articles on the subject. For example, "The boundary layers of ac-arcs at HID-electrodes: phase resolved electrical measurements and optical observations", Journal-of-Physics-D-Applied-Physics. 21 Jan. 2007; 40(2): 415-31. I should mention, though, that they examined lamp electrodes, not electrodes in welding machines. Anyway, this guy has published some other articles you would maybe find interesting.

I will try to get access to the IIW database, sounds like it could be very useful.

So, how is the weather like in Graz? :) Enjoying the conference? have you met Prof. Tanaka?

Best regards,

Ana

By Stephan (***)

Date 07-14-2008 19:49

Hi Ana,

thanks for replying!

To answer your last questions first. :-)

I have enjoyed the IIW Assembly very much! It were five great days in Graz and I have met again some of the very greatest ones in the welding world over there! By the way, the weather was fine as well was Graz. A 100% recommendation for a visit - even a pure university's city, many students and an appealing old downtown, amazingly pulsating... All in all it was a true blessing again to having had the honor to participate this event.

To come back to our particular subject. Yes! I have met Prof. Manabu Tanaka as I have met Prof. Yoshinori Hirata as well, who is the Chairman of the IIW Study Group 212 (Physics of Welding).

In the course of the SG 212 Meeting, TANAKA has presented his paper on "Visualizations of 2D Temperature Field of Molten Metal in Arc Welding Process" (IIW Doc. 212-122-08). Hmmm, most excellent! Even TANAKA!

HIRATA has presented his paper "Magnetic Control of Arc Plasma and its Modelling" (IIW Doc. 212-1128-08). As well and as always an enjoyment to listen to...

One of the - at least for me, but surely not only for me - most impressive papers however came from John LOWKE (et TANAKA) dealing with "Electrode Heat Transfer in MIG Welding" (IIW Doc. 212-1119-08). To be honest with you. For me Prof. Lowke is one of the most outstanding arc physicists in the world and it is a true blessing for everyone who has the honor to listen devotionally to his wise words. He has presented a new and self developed theory on even the subjects we are discussing about here presently in the AWS forum. He has tried to find a reasonable explanation of the cathode sheath behavior in GMAW by checking the existing theories dealing with this subject (field emission, thermionic emission) and has found out that no existing approach can precisely explain the voltage drop with a GMA cathode. He has suggested an approach dealing with "Metastables" as he calls the particles which might arrange and govern the observable GMA regimes. Extremely interesting and presented by John Lowke himself(!). An unforgettable experience...

But to come to the point - forgive me that I am raving... :-).

Well, I have asked Prof. Tanaka as Prof. Hirata on their experiences in finding values for metal oxide work functions or the accuracy of even the known existing values. Both answered that they are getting the values they are using for their work from even the FORMENKO book, you have already found, as you said. These are the information they are using for their calculations and they have never tried to doubt on these values or have tried to question even those. I felt reminded on myself, since I have learned once the values for Al2O3 and pure Aluminum and since then I am using them for any discussion on this subject. You were the one who has "stung" me for questioning the values for the first time in my life, to be honest.

Well, to keep the long story short. There was another, extremely likeable, fellow, Michael Schnick, a great expert in numerical simulation coming from the University of Dresden and who has held a presentation on his paper: "Numerical Investigations of the Influence of Design Parameter, Gas Composition and Electric Current in Plasma Arc Welding" (IIW Doc. 212-1127-08). He has participated with our discussion and he has given me the following very interesting information. First of all, he said, the discussion on the validity of work function values is a very reasonable one, since nobody knows exactly how the values have been taken - just as we have mentioned and assumed now for several times. He said that he has visited the Max Planck Institute Department for Molecular Physics in Berlin. Here they should have a very particular apparatus making it possible to deposit metal oxide layers of 1(!) Angstrom upon substrates to subsequently measure their work functions.

To exactly determine the correct work function - e.g. Al2O3 - which should actually normally exist in this stoichiometric composition - even just a layer of that minimal thickness is necessary, the fellow said.

The "normal" layers of metal oxides whereas but in particular Aluminum Oxide (let us remain with Al2O3 to not unnecessarily complicate the issue) showing thicknesses of 50... 60 Angstrom(!). And now it comes...

When we are considering the tunnel effect as the main effect for that a valence electron may escape from the oxides surface structure, then we have to consider as well the electron's wave length in relation to the width of the potential well. Are both values - please correct me when I am wrong (you know I am nothing but a layman!!) - comparable similar, the tunneling is caused.

Now were my question, how great is the likelihood for an electron to escape when it has to tunnel a layer's thickness of 50... 60 Angstrom? Wouldn't the resonant frequency of all the atoms a/o molecules hinder an interior electron to tunnel throughout the oxide layer thus the likelihood of an electron to escape drops down to ~ 0 ?

Hmmm, questions over questions - as usual. However, as the colleague from the University of Dresden said, there are presently investigations being performed to approve the well-known values for metal oxide work functions or to determine even new values with the highest technological accuracy achievable at present, respectively.

And now coming finally to a work function value for Al2O3 I could find out with the great help of my colleagues. The value comes from the book:

"Gaseous Conductors - Theory and Engineering Applications"; James Dillon Cobine, 1941.

Cobine should state within the mentioned book another reference:

A.L. Hughes and L.A. DuBridge, "Photoelectric Phenomena", McGraw-Hill Book Company, Inc., N.Y. 1932

and

S. Dushman, Rev. Mod. Phys., 2, 381 ,1930.

for the Al2O3 work function value:

3.77 eV.

As my fellas from the numerical simulation department say, by comparing other metal oxide work functions the value appears to be quite sufficient as to be accepted for calculations.

As you can see, there is not that much left from the very first beginnings of this thread where we have used 1.77 eV for the work function of Al2O3. However, perhaps the people who have prepared the papers (data sheets) have failed by writing 1.77 eV instead of 3.77 eV. Assuming that the pure Aluminum might have a work function of a slightly > 3.77 (e.g. 3.95eV) the oxide would nonetheless provide the first electrons under the influence of an exterior force (field or thermionic). At least as long Prof. Lowke were wrong by stating that the field forces with welding are quite too low to dissolve an electron from the cathode (workpiece) and thus the existing theories might prove. :-)

To be honest, I truly trust Prof. Lowke and his calculations very much, thus I doubt currently on the existing theories. So to speak... I am appearing to be a little confused, since the more I try to learn the less I mean to know.

Perhaps you, Ana, can point all of us welders in the right direction when you have prepared your paper, which I guess will be extraordinary good!

Please let me make one short sentence to your wonderful predication:

" In welding, if I understand this correctly, you charge an electrode in a welding machine and use the substrate (material you want to weld) as ground, which makes a capacitive plasma. So, when you use AC power on your charged electrode, don't you have trouble on polarity crossings? Your arc doesn't extinguish? And why do you use AC power? DC with positive charge on the electrode would produce nice, thin and powerful arcs. Wouldn't it?"

Yes, you are absolutely right in your understanding. And yes also, we would having troubles when we were using AC for - in particular - Gas Shielded Tungsten Arc (GTAW) Aluminum Welding (destroying the oxide layer) without having power supplies who would support us in the periods the arc does extinguish in polarity crossings. GTAW DC arcs (electrode as cathode) are feasible here when using Helium as a shielding gas. This however runs smooth and quick and yields a high fusion depth at all.

My best regards and please keep us updated with your fascinating profession!

Stephan

By rlitman (***)

Date 07-14-2008 20:39

I wonder how different lamp electrodes are from GTAW welding. I know that in GTAW, the shielding gas pressure is always at one atmosphere, and pure gases like hydrogen or neon are not used, but there certainly are a lot of similarities.

As for A/C vs D/C.
D/C certainly makes a "nicer" GTAW arc, but the electrode is actually most often negative, because the positive side receives more heating.
D/C is the preferred waveform for most GTAW welding.
Yes, for sinusoidal A/C the arc extinguishes at each crossing. That is why a high voltage, high frequency, signal is often overlayed, to reestablish the arc at each crossing. My welder is powered by an inverter, and produces a square enough A/C waveform that does not extinguish at each crossing.
The electrode positive cycles tend to break up the oxides on the surface of aluminum (and other highly reative metals such as magnesium, etc.).
So, A/C is used, to minimize the heating of the electrode, and to have some "cleaning" action on the workpiece.
One interesting effect of using A/C, is that the frequency controls the "tightness" of the arc. Higher frequencies have narrower cones than lower frequencies.
Varying the frequency from 60hz, up to 180hz, has an effect that is similar to, but more powerful, than adjusting the taper angle of the electrode tip.

By Stephan (***)

Date 06-09-2008 10:41

Clif,

using quickly the lunch break for a short addition on your interesting topic.

Have found the pictures and would like to attach them hereinafter.

- On Aluminum_Oxide_1.jpeg, you can see the difference between an "Al2O3 afflicted" and an Al2O3 free* sheet metal (AA 5754 base material) welded by using AC GTAW.

- On Aluminum_Oxide_2.jpeg, you can see the current and voltage patterns of an "Al2O3 afflicted" and an Al2O3 free* base material (same grade as above) by using the same process as above.

- On Aluminum_Oxide_3.jpeg, you can see the current and voltage patterns of a scraped and a brushed base material (same grade as above) by using the same process as above.

As you can see, the aluminum oxide has an improving effect - although it appears actually rather as a contradiction (electrical isolator) - on the welding process.

"Without" or better, with too less aluminum oxide you can expect:

- Rather wider weld with an irregular delineation
- Dark precipitates
- Irregular and in worst case no "cleaning" area besides the seam, and last but not least

- Arc Instabilities(!)

Best regards,
Stephan

* Which is of course to be seen relatively

Attachment: Aluminum_Oxide_1.jpg (74k)

Attachment: Aluminum_Oxide_2.jpg (0B)

Attachment: Aluminum_Oxide_3.jpg (0B)

By Plasma-Brain (**)

Date 06-09-2008 20:59

Stephan,
All I have to say is Wow, Thank you for such an in depth reply!
So aluminum can be "too clean" to properly weld? That's a new one to me but it makes sense.
Would that same mechanism be why adding Co2 or O2 to a steel puddle improves its characteristics? I remember reading about that in one of the Journals but the details behind it went over my head. Ill dig through my collection and see if I can find it again.

Would that also make it so that it's easier to cut aluminum before you clean it with a plasma torch, or is the voltage of the plasma arc too high to make much of a difference?

Slightly overwhelmed but very appreciative,
-Clif

By Stephan (***)

Date 06-10-2008 07:27

Clif,

my true pleasure, thanks!

In terms of discussing CO2 a/o O2 and their influences on the weld, I am certain there's a whole bunch of superb information to find here within the forum's archives but as well as on the famous Ed Craig website. But from a very fundamental standpoint active components in mixed shielding gases are - amongst others - to major reduce the surface tension of both droplet and weld pool to improving hereby the droplet detachment as the wetting behavior of the weld pool. But I am sure you'll find all the details by "digging" through your collection. :-)

In terms of Plasma Arc Cutting (the cutting experts may correct me when I am wrong), the Plasma stream has to be seen as a high energy containing medium having relatively high temperatures (~ 30.000 °C). At least to my best knowledge, there is no metallic oxide known having higher evaporation temperatures than 30.000 °C. Thus it is as you already have assumed. The Plasma stream might rather not recognize any considerable differences between the original base material (Al - temperature of vaporization T ~ 2800 K) and its oxide (Al2O3 - temperature of vaporization T ~ 3600 K). Both will be vaporized relatively fast...

Best regards and thanks again,
Stephan