Journal of Undergraduate Research
Volume 4, Issue 1 - September 2002

Oxidation of Nickel (100) by Exposure to Atomic and Molecular Oxygen

David Winsberg

 

 

ABSTRACT

 
There are many reasons to study oxidation processes involving oxygen, one of which is to better understand the corrosion of spacecraft materials. We studied the adsorption of oxygen on the (100) surface of nickel by exposing a sample to molecular oxygen in an ultra-high vacuum environment. Before dosing the sample with oxygen, we cleaned our sample using a sputter and anneal process, and then used XPS to ensure cleanliness. After the dose, we then ran XPS scans to characterize our surface. We graphed coverage, which is related to the ratio of O 1s peak area to Ni 3p peak area, versus exposure. Our data shows a plateau at 2 to 10 Langmuir exposure and another plateau at 40+ Langmuir. We will compare the adsorption from exposure to molecular oxygen to exposure to high energy O-atoms. Although oxygen atoms are expected to oxidize Ni more quickly than molecular oxygen, it is unknown exactly how much faster or by what means. The chemistry and mechanisms involved could be different from those induced by molecular oxygen.

 

INTRODUCTION

 
Malaria Satellites that operate in low earth orbit (LEO) are exposed to high-energy O (oxygen) atoms, which can degrade the surface of the satellite. About 90% of the composition of the atmosphere at this altitude consists of atomic oxygen. The satellites orbit the earth at a speed of about 18,000 miles an hour, which causes the already reactive O-atoms to bombard the satellite with about 5 eV (electron volts) of kinetic energy. Polymers are especially susceptible to this form of corrosion, but metals also can be attacked in this manner. In order to design satellite materials with higher resistance to oxidation, the rates and mechanisms for surface oxidation by hyperthermal atomic oxygen need to be known. This paper focuses on the oxidation of Ni(100) due to molecular and hyperthermal atomic oxygen.

 

Molecular oxygen adsorbs on the clean surface by a concerted dissociative process whereby the intramolecular bond of O2 is broken and new bonds are formed between the oxygen atoms and the surface. The initial rate for dissociative adsorption is given by the expression R = F _ So, where F is the flux of O2 molecules striking the surface and So is called the initial sticking probability. Since the cleavage of O2 molecules requires energy, So is less than unity, which means that only a fraction of the O2 molecules adsorb on the surface. Based on experiments done by several different research groups [2,3], the adsorption of oxygen and the subsequent oxidation of Ni(100) by exposure to O2 occurs by the following mechanism. After about 25% of the surface is covered with oxygen atoms, "islands" of oxide start to form on the metal surface, and it is at the edges of these islands where most of the further oxidation is believed to take place. The reason is that after 25% coverage, it requires less energy for an O-atom to adsorb at the perimeter of an oxide island than elsewhere on the surface. Note that there is a significant energy barrier and surface concentration needed to start an oxide island. When these islands form, nickel atoms from the bulk diffuse to the surface region, allowing more oxygen to adsorb onto the surface. This means that the adsorption rate or sticking probability will initially be high, since there is not much oxygen on the surface, and it will decrease as the surface becomes covered with oxygen. It should then get larger after the islands have formed and begin to grow. Once an oxide layer is formed and oxide layers start to build up, the absorption rates should again decrease, since the growth of oxide is now limited by the diffusion rate of species through the oxide film [2]. However, if the incoming oxygen is not O2 but high-energy O-atoms, the adsorption and oxidation processes could be quite different [4].

 

PREVIOUS WORK

 
Background In order to gain a better understanding of the chemistry involved in adsorption and oxidation, scientists have often performed UHV (ultra-high-vacuum) experiments. A common experimental method is to expose the sample to a source of oxygen, and then use surface analysis techniques, such as X-ray photoelectron spectroscopy (XPS), to determine the chemical changes and how much oxygen has adsorbed at the surface. In order to interpret the experiments with atomic oxygen, we decided to first run experiments with molecular oxygen to obtain data for comparison, since part of our research is to determine the differences in oxidation by exposure to atomic versus molecular oxygen. Since other scientists have done similar experiments with O2 [2], we can use this to compare our data to those obtained by other scientists.

 

EXPERIMENTS

 
After "venting" (increasing the pressure to atmospheric level) our UHV chamber, we must "bake out" the chamber before running any experiments. Bake out involves heating the chamber to about 200oC for 24 to 36 hours under vacuum to desorb any contaminants from the chamber walls. After bake out and subsequent cooling, we would clean our sample by sputtering it with Ar ions, followed by annealing the sample. We would then use XPS to ensure that our sample was free from surface contaminants. If it was not, we would repeat the sputter and anneal processes until the surface was clean. After determining from XPS that our sample was clean, we would "flash" the sample to desorb any hydrogen that might have adsorbed from the background, since XPS does not directly detect hydrogen. This consists of briefly heating the sample to 1000K, and then cooling it back to 300K. Except for bake out, we would clean the sample according to the above method before running each experiment, including running XPS to ensure that the surface was clean, before running each experiment. We only ran bake out after venting the chamber.

 

After ensuring that our sample was clean, we exposed the sample to our source of oxygen. To achieve a certain dosage with molecular oxygen, we went according to the following procedure. We started with a low background pressure, usually around 3.5 _ 10-10 torr, and then opened a valve that allowed a low pressure of oxygen into the chamber. Once the O2 partial pressure reached a desired value, (low dosage, 10-8 torr; high dosage, 10-6 torr) we began timing our exposure. During the dose, we adjusted the valve to maintain the desired pressure. After the amount of time had passed for the correct dosage, we closed the valve and let the chamber pump out. The number of O2 molecules that strike the surface is proportional to the partial pressure of O2 multiplied by the time of the exposure. The units of exposure are given in Langmuir (L) where 1L = 10-6 Torr _ sec. An exposure of 1L corresponds to approximately one atomic layer of O2 molecules striking the surface. All of our experiments were run at a surface temperature of 300K.

 

After exposing our sample to a specific amount of oxygen, we ran XPS scans to determine how much oxygen had adsorbed on the surface. The scans included a broad scan between 100 and 1500 eV, and narrow (more precise) scans to get detailed information about the peaks O 1s, Ni 3p, Ni 2p, and C 1s. XPS data on O 1s was taken to obtain the peak area (gives surface coverage) and peak shape (gives chemical state). Data for Ni 2p was taken to obtain peak shape, Ni 3p for a peak area basis, and C 1s to ensure that there was no carbon contamination. Peak areas were obtained using the Shirley smoothing method. The ratio of the O 1s peak area to the Ni 3p peak area multiplied by 3.238 provides a reasonable estimate of the surface coverage of oxygen in monolayers, where one monolayer corresponds to one atom of oxygen adsorbed for each surface atom of nickel. This factor of 3.238 involves several parameters, which include sensitivity, mean free path of an electron, detection probability, and photoemission cross section. We plotted coverage in monolayers resulting from a certain exposure. We also analyzed the O 1s peak shape to determine chemical state characteristics.

 

EXPERIMENTAL CHALLENGES

 
After we had taken enough data on exposure to molecular oxygen to generate a curve, we found considerable scatter in our data. We determined that the questionable data points were those that were taken when we did not "flash" the sample to desorb hydrogen before dosage. We identified this problem by looking closely at the O 1s peak shapes for the unflashed exposures, and noticed a shoulder at high binding energies due to -OH groups. Since XPS does not directly detect hydrogen, we might think that our sample is clean even though some hydrogen may be present. We solved this problem by flashing the sample before each dosage. By following this procedure, we were able to consistently reproduce the data.

 

RESULTS

 
Below is a graph of the oxygen coverage on Ni(100) as a function of the O2 exposure.

 

 

oxygen coverage on Ni(100) as a function of the O2 exposure

Figure 1. Oxygen coverage on Ni(100) as a function of the O2 exposure

 

 

As may be seen from the graph, the curve flattens at approximately an exposure of 3L to 15L and again at 40L, indicating that the rate of adsorption decreases abruptly at these conditions. This suggests that chemisorption occurs before 5L, the rate of which decreases as the surface becomes covered by oxygen. Islands of oxide start to form at around 10L, causing the rate to increase again as the islands grow, but then decrease when islands start to coalesce at around 40L. At this point, the oxygen uptake rate decreases again, since it is now limited mostly by diffusion of species through the oxide film.

 

Below are two graphs for the O 1s peak. One was taken at 8L exposure and the other at 9L, but the 8L peak has a shoulder and is shifted to higher binding energy, which indicates the presence of -OH on the surface. This is due to hydrogen on the surface before the dose, since the sample was not flashed in this case.

 

O 1s peak

Figure 2. O 1s peak

 

In the coming weeks, we will obtain the uptake curve for exposure of Ni(100) to atomic oxygen. Since atomic oxygen is very reactive, it should oxidize Ni(100) at a much higher rate than O2. Oxidation with gaseous O-atoms may also proceed by different mechanisms than oxidation with O2. For example, oxide islands may not form at low coverages, but rather the high reactivity of the O-atoms could produce unique, high density chemisorbed phases and chemical states of oxygen which serve as precursors to complete oxidation [4].

 

CONCLUSIONS

 
After running our experiments, we have obtained data for oxygen coverage on Ni(100) versus exposure of molecular oxygen for comparison with data that will be taken using atomic oxygen, as well as experiments done by others. Our experiments show that the chemisorption controlled regime occurs in the exposure range of 0 to 5L, the island growth controlled region occurs from about 5L to 40L, and diffusion controls the oxygen uptake rate after about 40L. Also, since we have performed system characterization experiments, we now have a better understanding of our system. We will use an oxygen atom source to determine the coverage as a function of exposure to atomic oxygen, and will compare it to exposure of molecular oxygen.

 

 

ACKNOWLEDGEMENTS

 

In this paper, I would like to acknowledge my mentor, Dr. Jason F. Weaver as well as the graduate students Alex Gerrard, Paulo Herrera, and Jau Jiun Chen.

 

 

REFERENCES

  1. Hoflund G B and Weaver J F (1994) Measurement Science and Technology 5 201

  2. Holloway P H and Hudson J B (1974) Surface Science 43 123

  3. Wang, W D, Wu N J, and Thiel P A (1990) Journal of Chemical Physics 92 2025

  4. Slezak J A, Zion B D, and Sibener S J (1999) Surface Science 442 L983

  5. Gibson K D, Viste M, Sanches E C, and Sibener S J (1999) Journal of Chemical Physics 110 2757

 

--top--

 

Back to the Journal of Undergraduate Research