That is a very reasonable question. I never came across any paper that precisely characterize Pt catalyst to identify the higher oxidation states. Sometimes I wonder whether it is excess precursor left after the synthesis.  

Well, from electrochemical data you can clearly see that ORR and HOR can ONLY occur on a Pt surface. At a high potential, where Pt-oxides are formed, there is no activity. Furthermore, Pt-oxides cannot exist at low potentials. In other words, Pt oxide, and ORR/HOR are not occurring in the same potential region.  See the attached link for more info.

dx.doi.org/10.1016/j.jpowsour.2012.06.019

Article Establishing the potential dependent equilibrium oxide cover...

Gustav,

The question was about the Pt-oxides present in the catalyst before any electrochemical test, not about the oxides that forms during electrochemical cycling.

Mazdak, I think if there is any oxide formed during the synthesis that would be much more stable compared to those forms during electrochemical cycles. In that case pre-formed oxides are not supposed to be active for ORR. But that is an assumption only, reality could be different.

Hi Gustav and Sourov,

Thanks for your answers and comments.

Gustav, thanks also for the paper you mentioned. I checked it and there were useful information there, further enlightening your viewpoint and the answer you gave.

I think your answer has implicitly a part of the reality, especially where saying “Pt oxide, and ORR/HOR are not occurring in the same potential region”. But before going further on, let me say few words about some points. Disregarding the type of the catalyst and its chemical state, there is most often an activation polarization that is controlled by the reaction sequence at the metal-electrolyte interface. This activation polarization is not exclusively determined by the quality of catalyst surface (other parameters like: activation energy of the redox event or the electron transfer contribution, chemical reactions that precede electron transfer, reaction byproducts and their effect on both surface and the engaged analyst and so on). So, the fact you pointed out about the less activity of Pt at high potentials (low overpotentials), is a phenomenological description of the activation polarization (in a wide sense) and is not exclusively caused by the Pt oxidation, though it is a part of that.

Now, let’s get back to the question and imagine that we are at high enough overpotentials (to avoid the abovementioned issue) guaranteeing a good electrocatalyric current. How can we ensure that 100% of this current is coming from 100% metallic Pt? Do we have any “in-situ spectroscopic evidence” for that?

Let’s look at this question from another angle and here is where Sourov’s comment arrives.

Imagine that you prepare some Pt nanoparticles supported on a carbon material. Then, you check it with XPS and you figure out that there are 40% of Pt oxides and 60% of Pt 0. Now you put your catalyst-support system on a GCE and immerse it in 0.1 M HClO4 and do the ORR. Of course you’ll read a current. But how do you know if the current is coming only from the 60% of Pt 0 or also that 40% of Pt oxides have “reduced” and are producing current. Moreover, even if that 40% of Pt oxides have not been reduced by any reason, how can we confirm that their contribution in the measured current is zero?

In the final reading, based on what you told and wrote in your paper, since only at high enough overpotentials “the stability of the oxygen bond decreases and the ORR may proceed”, regardless of the origin of the oxygen (you have mentioned indifferently whether water dissociation or oxygen molecules, and I would add, other sources such preexisting surface oxides from the synthesis procedure), one can expect that if such high enough overpotential is applied to your partially oxidized catalyst, the stability of the oxygen bonds will decrease and the ORR may proceed. This will in turn mean that one should consider all the 100% Pt content (both metallic and and oxidized) in his calculations as responsible for the electrocatalytic activity and all other electrochemical parameters he is trying to figure out.

Hello! I'd like to share some ideas on this point.

1) the data of  XPS provide with the composition not of the electrochemically active surface, but the surface of the catalyst till the depth of about 10 nm. Thus, the results of  XPS are not useful when determining the amount of Pt atoms responsible for the proceeding of an electrochemical reaction. 

2) the main method used to determine the real electrochemically active surface is the cyclic voltammetry: the maximums of hydrogen ad-atoms adsorption/desorption in the cathodic potential region (0-0.4 V vs she); maximums of oxygen ad-atoms adsorption/desorption in the anodic potential region (0.6-1.5 V vs she); maximums of copper ad-atoms adsorption/desorption; maximums of carbon monooxide ad-atoms adsorption/desorption and so on. Even this method cannot guarantee that all the Pt atoms on the electrochemically available surface will function as catalytic centers.

Hi Mazdak,

Let us look at a related system, Pt-Co nano particles. Particles of say 50/50 can easily be produce. As soon as this particles comes in contact with an acid and the potential cycled in the window for ORR, the Co will leach out from the utter part of the particle. The inner particle can however still contain Co. It is easy to see that the Co has been leached out from the surface in a normal CV. Further, it is also possible to see that these Co can migrate from the center out to the edge after a while, based on electrochemistry.

How does this affect ORR. ORR is highly surface sensitive.  So if ORR is performed with a unleached samples, you will see that the activity is not great, but as the Co is leached out the activity improves, and reaches close to pure Pt.

Turning to Pt and the Pt oxide. It is pretty clear that the Pt-oxide is not stable in the potential window of ORR, as previously discussed. So, if platinum oxide particles are created, in the ORR window, utter oxide will be reduced to Pt. The particle can however still contain oxide but they do not take part in the reaction directly. The oxide can however affect the activity due to the pulling of electrons, similar to alloying.

"...This will in turn mean that one should consider all the 100% Pt content (both metallic and and oxidized) in his calculations as responsible for the electrocatalytic activity..." - As previously mentioned, the ORR is only taking part on the surface and not in the bulk, hence, only what is on the surface should be counted. The bulk properties can still affect, but this is minor.

Hi Lei Xing,

Thanks for the suggestion.  I checked the paper, but unfortunately I could not find relevant material to the question, said, role of oxides in electrocatalytic activity.

Hi Elena and Gustav,

Thanks for your contributions. Before going through the somehow long discussion below, I’d like to highlight some points that will be referred to with more details later on.

1- I feel the attribution of a “zero and one” model to electrocatalytic activity of oxidized and metallic Pt, respectively, is not prudential. I don’t believe that the oxide is as active as the metal, but zero activity is quite tricky.

2- Any thoughts about the contribution of oxides in the ORR in my discussions, relies on the prerequisite of availability of oxides at the time of occurrence of the ORR. Since the formation of oxides on the Pt surface takes place gradually and in a potential range (as the ORR plots suggest and not in a single point jump) while ORR is in the meanwhile continuing (and of course, slowing down uphill the potential), there are always some oxides nearby the metallic surface that can do something if they are good enough for that.

3- Even if what Gustav says about the reduction of the oxides and their contribution in the ORR is true, we already have a big issue with the literature highlighting the negative effects of oxidized states of synthesized Pt catalyst on the electrocatalytic activity (see these papers just as some examples:

a-Fabrication and Characterization of Well-Dispersed and Highly Stable PtRu Nanoparticles on Carbon Mesoporous Material for Applications in Direct Methanol Fuel Cell,

b-Synthesis and physical/electrochemical characterization of Pt/C nanocatalyst for polymer electrolyte fuel cells,

c-Rapid Microwave Synthesis of CO Tolerant Reduced Graphene Oxide-Supported Platinum Electrocatalysts for Oxidation of Methanol,

d-PtRh alloy nanoparticle electrocatalysts for oxygen reduction for use in direct methanol fuel cells). Although I am not confirming that part of literature (and from this point of view, I am in line with Gustav) and was actually casting some doubts on them by bringing up the possibility of participation of the oxides in ORR, I think indifference in the behavior of primarily metallic or oxidized Pt catalyst is a bit more than what I could have imagined.

Now, let me open a bit more the aim of the question for Elena. As you mentioned XPS provides the surface composition down to few nanometers, well deeper than the diameter of the catalyst nanoparticles, and thus talks about both surface and bulk of the particles. Of course it is not meant to tell us the electrochemically active surface not only because it cannot measure the surface area (also CV cannot measure the surface area, but using a well stablished conversion factor, we convert the current measured by CV to surface area), but also because it gives data from surface and depth of the particles. The question aim to understand if there are some oxidized points of the catalyst on its surface that can in the meantime be detected by XPS, should they be considered equally likely as the pristine metallic sites to show electrocatalytic activity. Gustav believes that although the XPS is not a proper tool to determine the total amount of the oxidized Pt that can participate in catalysis (since it gives both superficial and in-depth content of them), those oxides located on the surface can show the catalytic activity. And he believes that regardless of the chemical state of the Pt, as soon as it is exposed to potential region where its thermodynamically stable state is metallic, namely, at low enough potentials (high enough overpotentials), even if there had been some superficial oxides, they will get reduced and in the course of electrochemical testing (e.g., subsequent CV cycles), they will show electrocatalytic activity.

But I feel getting rid of oxides could be not that easy. Sometimes oxides formed in different ways are not easily removable. Especially if we view it in the wider context of oxygen functional groups. Being in the thermodynamically unstable potential region of oxides does not necessarily mean the absence of oxides. The same applies to some well-known issues of the catalysts like CO poisoning where although supposedly the poison should be unstable at high potentials in the presence of oxygen, it is absolutely not easy its complete removal and rather, it accumulates over time and further develops the poisoning (look this paper: Oxygen reduction reaction on Pt and Pt bimetallic surfaces: a selective review).

In other words, having a surface 100% free of oxides (pure metallic) could be as unlikely as a surface 100% covered with oxides. This, in addition to other things, means that these two are coexisting and the attribution of the electrocatalytic current to just one (namely, metallic one) could be incautious and audacious.

For instance, imagine that some Pt atoms are in Pt 2+ state in given platinum catalyst. This means that Pt 2+ can be further oxidized to Pt 4+ and for this to happen, some more oxygen should be adsorbed to the surface. This is similarly the case for ORR as well. Notwithstanding the difference in the final product of Pt oxidation and the oxygen reduction, the sequence of the events is quite similar, meaning that even an oxide could have contributed in the oxygen reduction. In other words, what will be the problem if in the course of CV, a fraction of Pt is just repetitively transformed between Pt 2+ and Pt 4+ state (Pt 2+ to Pt 4+ in the forward cycle and vice versa in the backward cycle). If such a thing can happen, it means that platinum oxide can contribute in the ORR.

This paper (Kinetics and mechanism for the oxygen reduction reaction on polycrystalline cobalt–palladium electrocatalysts in acid media) shows in good way that there are possibilities for attachment of different number of oxygen atoms to the catalyst. Considering a catalyst atom with lower number of attached oxygens at a lower oxidation state and one with higher number of attached oxygens at a higher oxidation state, it can be inferred that oxides could also do the ORR.

Article Fabrication and Characterization of Well-Dispersed and Highl...

Article Synthesis and physical characterization of Pt/C nanocatalyst...

Article Rapid Microwave Synthesis of CO Tolerant Reduced Graphene Ox...

Article PtRh alloy nanoparticle electrocatalysts for oxygen reductio...

Article Oxygen Reduction Reaction on Pt and Pt Bimetallic Surfaces: ...

Article Kinetics and mechanism for the oxygen reduction reaction on ...

Hi Mazdak,

First and foremost I have to say that the word "platinum oxide" is very vaguely used in many studies.What is typically meant by the oxide is actually add atoms of either O and OH on the surface. This is what is being one of the main points of the paper I posted before. So, please be aware of this when reading papers

Again, returning to the paper that I posted above and the question weather or not Pt oxide can contribute to ORR. Let us assume that it is indeed active, why is it not possible to measure the reversible potential of oxygen?  Now, let us assume that is it not active. At a high potential the oxide is formed. If we now sweep the potential downwards having an oxide on the surface, how will the onset of the ORR wave be affected? From the mentioned paper, it is clear that the onset is shifted negatively, which clearly indicates the oxides inactivity. Further, this negative shift also shows that the oxide can exist in the ORR region, although only for a short while. After sweeping down to the RHE and then up again, the oxide is gone.

Now, could there be other oxides than Pt-oxide that are active for ORR and stable? Sure, if I remember right MnO2 is a good one.

The devil is in the experimental details. Most studies have the particle characterization before any electrochemical characterization. This is very problematic as the EC interface can very much affect the particles compositions. Something that is stable in air, is not necessarily stable in an electrolyte, and especially not after potential cycling. Like in the first study you referred to, there is no evidence that the oxides were also present after the EC measurement. The only correct way of doing these experiments are in-situ experiments, and they are tricky.

In the case of CO poising, the only reason for adding Ru to the Pt is that the Ru has an early adsorption of OH in comparison to Pt. If OH is present, the CO can be oxidized. Similarly, if you put in H2O2 in the solution, there wont be any CO poisoning of the Pt surface.

Hi everybody,

I'm sorry I cannot continue the discussion in a couple of the coming days since I will be out. As soon as I am back, I will join you again.

Thanks in the meanwhile Gustave for your helpful comments and patience.

Mazdak

Pt nanoparticles are oxidized by oxygen (the first works about this phenomenon can be related to articles on properties of EURO-Pt catalysts in J. Catal., 1978-1988). The smallest particles (1-1.5 nm) can be oxidized totally in air at ambient conditions, bigger - up to 2-3 external atomic layers.  

So, if a catalysts, after its storage in air,  was not reduced in situ in the XPS chamber, a high content of Pt2+ and Pt4+ detected by XPS testifies only  to a high dispersion of the supported metal. 

There are a lot of articles about particle size effects in dioxygen electroreduction or electrooxidation of CO, methanol...

Oxidation proceeds wery quickly and takes 2-5 min in the case of Pt or Pd. Electrocatalysts with 50-60% metal loading are pyrophoric because of this reason. 

Hi everybody,

Thanks for all contributions.

In this period, I checked all the comments and also consulted with some people in the field and I would like to conclude a couple of point that might be helpful for all.

An important point was the thermodynamic stability of Pt and its different chemical states. As Gustav told and as I could hear from more people, Pt oxides are not stable at low potentials, in the ORR region. Even if it happens so, it will be for a short while. Therefore, the central point wouldn’t be the likelihood of activity of the oxide in that potential region, but rather, the likelihood of existence of them at such potentials. Apart from this, even assuming a sort of activity for those oxides (like what I pointed out about Pt 2+ to Pt 4+ in the forward cycle and vice versa in the backward cycle, in previous post), this activity would be less than metallic Pt. Therefore, combining the two facts, namely, little probable amount of the oxide and its lower activity (if any), make the subject of exact determination of the contribution of the oxides in the total electrocatalytic current measured, an absolutely undesirable subject for investigation. Honestly, it is a fair decision.

The other point that I could conclude after some reading and also discussions with people, was that understanding the chemical states of the Pt catalyst right after the synthesis, could be more helpful from the viewpoint of interpretation of the synthesis mechanism rather than the activity. If we consider that a good amount of Pt oxides (yeah, let’s only say a good amount and not all) are reduced in the ORR region, those oxide fractions measured by tools like XPS, will be more helpful for enlightenment of the synthesis process rather than what comes up afterwards.

Indeed.

Mazdak, I agree. Good summery

May be it is possible to do quantum chemical or DFT calculations for adsorbed hydrogen on Pt clusters with differant amounts of oxidized Pt atoms?