Flower–like structures on „LiPON“ solid electrolyte

In document Sacrificial interlayers for all-solid-state batteries (Page 81-87)

Battery cycling

3.3 Experimental Issues

3.3.1 Flower–like structures on „LiPON“ solid electrolyte

The first phenomenon addressed in this chapter is the morphological change of „LiPON“

electrolyte thin–films. This topic has found hardly any attention in literature. Only the Indian Institute of Science has reported the observation of structures on the thin–film surface [158]. However, as these structures can have a huge influence on the properties of the interface between „LiPON“ and any adjacent phase, they will be discussed below.

During this work, the formation of surface anomalies was visible. Directly after removing the samples from the deposition chambers, the samples occurred to be opaque (Figure 13). These anomalies were not just single deposits, isolated on the substrate, but many deposits spread all over the substrate.

The morphology of the observed surface anomalies had a variable size and shape. Deriving from the most common shape these structures are referred to as “„LiPON“ flowers”. Their size was depending on the film thickness but was typically in the range of 2 µm – 10 µm in diameter. As Figure 14 shows, they could also reach diameters of more than 20 µm and were large enough to be observed with the bare eye.

Figure 13: „LiPON“ film after deposition for 2 h on silicon (100) substrate.


The shape and orientation of these structures varied depending on their position on the sample. Only in the center of the sample they were perfectly round. The closer they were located to the sample edge, the more they were shaped crescent–like. It could also be observed that these structures were always oriented toward the edge of the sample (as shown in Figure 15).

5 µm

Figure 14: Flower–like structures on „LiPON“ solid electrolyte.

Figure 15: Orientation of „LiPON“ flowers in dependence of the position on the substrate. SEM images.

20 µm

30 µm 30 µm

30 µm


The origin of these structures is currently unclear. The cross–sectional image suggests that their formation is initiated during the sputtering process (Figure 16). By creating a cross–

section of the sample it was possible to get a side–view on one of these structures. The first observation was that these structures were growing on top of the electrolyte layer but not instead of it. In Figure 16 the smooth electrolyte is still visible below the surface deposit. It is interesting to see that the thickness of the electrolyte below is not constant. Instead it is gradually decreasing from the edge to the center of the structure and these structures seem to grow in a certain angle to the surface. This finding suggests that, from a certain point during the deposition on, the formation of these structures begins on top of the electrolyte layer and they are not formed right from the beginning. There must be an initiation for the formation of these structure. Then they seem to spread out across the electrolyte leading to the covering of a linearly growing electrolyte layer. This assumption is supported by the fact that „LiPON“ layers with a thickness of around 100 nm do show hardly any surface anomalies (see chapter 4.1). The way the structures grow suggests that in line with nucleation theory a threshold needs to be overcome to initiate the growth of these „LiPON“ structures.

It could also be observed that sometimes anomalies were formed in areas where there were scratches from substrate handling. However, when attempts were made to create artificial nucleation sites (e.g. by creating scratches on the sample surface) no preferred growth of structures in these areas could be observed.

Figure 16: Cross–sectional SEM image of a surface anomaly on „LiPON“ electrolyte.

2 µm


In addition to the growth of these structures also their composition was examined.

Determining the stoichiometry of the surface via XPS was not possible. Although the deposits (“flowers”) were comparably large, they were still smaller than the lateral resolution of the method. The average stoichiometry of the film surface in an area with many structures compared to a surface area with less structures did not show any difference bigger than the error of the measurement. To further examine the composition of these structures, EDX was used as an additional method.

A sample was examined where there was a surface deposit and in one position a structure deposit to have broken off during handling (Figure 17). With this sample, the stoichiometry of the films could also be compared to „LiPON“ below these structures. The results are summarized in Table 3. The O/P ratio of the „LiPON“ film was around 3.8 whereas the N/P ratio was 1.1. This is in good agreement to the results that can be obtained by varying the nitrogen flow and with respect to the accuracy of the detecting method. However, the oxygen ratio is slightly higher than expected, probably due to surface impurities. The amount of carbon is around 4.5 %. Only traces of Si, probably due to holes in the film, are visible. The stoichiometry of one of the „LiPON flowers” differed strongly from the surface stoichiometry. The oxygen content was reduced by about one third and was only 41.4 at%.

Figure 17: EDX image of the „LIPON“ surface, a surface deposit and an area where a surface deposit had been removed.


The nitrogen content was reduced by almost 2/3 and the phosphorous content decreased to only 3.1 at%. The carbon content of the structures was much higher (45.6 at%). In the area where there was no deposit, the oxygen content was the highest with 73.2 at%. The nitrogen and phosphorous content were just as high as in the deposit but the carbon content was reduced to 15.1 at%.

Table 3: Comparison of the elemental contents of a „LiPON“ electrolyte surface, a deposit and an area where a deposit is missing (shown in Figure 17).

However, a second line scan on another sample was performed and led to contradicting results (Figure 18). In this case the nitrogen and carbon content remained more or less the same but the phosphorous content in the surface deposit was smaller than in the „LiPON“

thin–film, whereas the phosphorous content was higher. Although the composition of these surface deposits could not finally be determined, it can be concluded from the results above that they do not consist of „LiPON“ and therefore might possess different properties than the flat films.

Nimisha et al. suggest that these deposits are formed due to contact with humid air [158].

However, in the present study humidity as the origin of these structures can be excluded.

The samples were always kept under protective atmosphere. The PM II chamber had a very low base pressure (10–8 mbar) and the plasma led to a high temperature during the deposition. Therefore, the fraction of water in the sample should be very low, even though the deposition chamber needs to be vented to change one of the four targets. Even if there was still residual humidity in the chamber, it should not be enough to create surface deposits of more than 10 µm in diameter and thicknesses larger than the electrolyte films in the time–

frame of the deposition.

If these deposits were formed in contact with atmosphere, humidity could be a reasonable explanation. However, the presence of surface deposits could already be observed with the

At% Surface Deposit Deposit missing

O 61.1 41.4 73.2

N 17.7 6.8 6.4

P 16.2 3.1 2.3

C 4.5 45.6 15.1

Si 0.5 3.1 2.9


bare eye directly after removing the samples from the deposition chambers. For the above mentioned reasons it can be assumed that the humidity was too low and that the deposits must have a different origin.

The only clear dependence that could be found was the dependence of the size of these deposits on the deposition times. Films deposited for only a few minutes had an almost perfectly flat surface. Only a few deposits were visible and they were very small (diameter

< 100 nm). This finding suggests that „LiPON“ films should have very homogeneous properties as long as they are thin enough but from a certain thickness on, surface anomalies will occur that will influence the electrolyte properties and performance of cells. In the case of „LiPON“ as an interlayer, it must be ensured that these deposits do not form during deposition.

Figure 18: SEM image and EDX signals of a „LIPON“ film with surface deposit. The nitrogen (green) and carbon signal (red) show no difference between the films and the surface deposit. The phosphorous signal (violet) decreases on the deposit whereas the oxygen signal (blue) increases.

15 µm


In document Sacrificial interlayers for all-solid-state batteries (Page 81-87)