This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
Site-specific SEM dopant mapping could be useful tool to semiconductor industry. Previous work has shown that using Focused Ion Beam milling on the sample for trench sizes. This project looks at optimising trench geometry for reducing artefacts in SEM dopant mapping.
The position and number of dopant atoms in semiconductor determines device functionally.  Therefore it is vital to measure dopant atom distributions and it is clearly observed from the recent developments in the technology that as semiconductor dimensions shrink traditional measurement techniques will be reaching their limits and an alternative method is urgently needed in order to keep further miniaturisation on track.  Dopant contrast in SEM has been considered as alternative but so far it is only easily observed on cleaved surfaces. This project concentrates on site-specific preparation of specimens for SEM dopant mapping using Focused Ion Beams (FIB). This will provide the basic understanding to enable 3d dimensional SEM dopant mapping.
Figure 1: Contrast seen in two trenches of different lengths (Work done by Dr Mark Jepson)
We can observe from figure 1 that dopant contrast of two samples which are different lengths are taken into consideration by et.al Dr Mark Jepson for getting contrasts.
2.1 Scanning Electron Microscope (SEM)
The scanning electron microscope (SEM) is a type of electron microscope that images the sample surface by scanning it with a beam of electrons in a raster scan pattern. The electrons interact with the atoms that make up the sample producing signals. In this project we use the secondary electrons. In the most common or standard detection mode, secondary electron imaging or SEI, the SEM can produce very high-resolution images of a sample surface, revealing details about less than 1 to 5 nm in size. The low energy of secondary electrons i.e. <50ev makes them sensitive to change in electric potentials at the specimen which can be used to dopant mapping.
A Field emission gun is used to produce an electron beam that is smaller in diameter, more coherent and with up to three orders of magnitude greater current density or brightness than can be achieved with conventional thermionic emitters such as tungsten filaments. The result in SEM is significantly improved and spatial resolution, and greatly increased emitter life and reliability compared with thermionic devices
As mentioned above Secondary electrons are also the main means of viewing images in the scanning electron microscope (SEM). In this project the range of secondary electrons depends on the energy. The distance measured is on the order of a few nanometers in metals and tens of nanometers in insulators. This small distance allows such fine resolution to be achieved in the SEM. P Type Bridge is normally visible brighter than n type but at a very low KV. Therefore in order to get that imaging in this project we have used a low voltage scanning electron microscope (LV SEM) FEG.
2.2 Dopant Mapping
The dopant atoms that are either of n-type or p-type are the main properties for increasing mobility's in the semiconductors. In the recent years there is a vast improvement in this technology by which tiny devices at microscale and nanoscale are been possible. So, for the improved scale at nano there is a need of dopant atoms, which has a vital role to play in functionality and development. In attaining this challenge the dopant mapping of the atoms are been developed.
2.3 Focused Ion Beam (FIB)
Focused ion beam also known, as FIB is used in semiconductor industries for analysis, deposition and ablation of material. There is a vast increase of technology in the recent years in which many developments occurred in the semiconductor industry. Introduction of FIB technology affected the whole process as FIB can modify and create microstructures. The ability of imaging while removing or depositing materials also makes FIB important tool for semiconductor industry. The FIB became a powerful tool for nanostructures for fabrication and manipulation. FIB instrument is similar to that of a Scanning electron microscope (SEM) but the beam used is Gallium ion beam in FIB instead of the electron beam used in SEM. The basic functions of FIB used in this project is namely imaging and sputtering with an ion beam also described as milling. The size and shape of the beam intensity of sample helps in determining resolution.
From the basic interaction it is seen that milling is continuous process, which is done when ions interact with the surface of the substrate. The milling rate is proportional to the beam currents and high currents means high beam proportion and thus high milling of the substrate. Milling allows sample to open up towards the three-dimensional imaging of the structure. This project includes trench size. Due to this the more knowledge could be obtained and the semiconductors can be developed and known in future in the three dimensional ranges. The aim of the project is the optimization of the Focused Ion Beam preparation to dopant mapping in SEM. Only project pictures were focused on influence of gallium ion beam energy. This project has in artefacts trench geometry.
As FIB is frequently used for milling the other great advantage of this is deposition of material on to the substrate. Most commonly the FIB is equipped with a deposition gun and when the deposition mode is selected the deposition gun interacts with surface and deposits the required material what we have chosen on to the substrate. Most commonly the FIB has Carbon (C) deposition, Tungsten (W) deposition, Platinum (Pt) deposition and Silicon oxide (SiOx) deposition.
The energy used of gallium ion beam in this project was only 30 kV and 10 kV respectively. As to protect the area that should not be eroded localised deposition can be carried out only in low kV so that only these energy levels can be used. As it is known that the FIB is functioned as dual beam, which has both columns. But it is not applicable in our project because as we look in low KV SEM, we use separate SEM for this purpose and there is no use of the SEM column, which is extended with Focused ion beam system.
3. Materials & Methods
Experiments were carried on plain silicon, simple p-n junction, stair case structure, and resolution structures. An undoped plain silicon sample is taken for determining the correct geometrical dimensions in order to get the perfect dimensions for the other two samples. We can see from figure 2 the schematic structure of staircase structure, which we used for our project. There is also figure 3 showing the schematic representation of PN junction structure and also Figure 4 showing the schematic representation of resolution structure.
Figure 2: Schematic Diagram of Staircase Structure Used in our project.
Figure 3: Schematic representation of PN Junction structure.
Figure 4: Schematic Diagram of resolution Structure.
3.1 FIB Preparation
An undoped plain silicon sample is taken and it is milled on FIB. Sample is placed on stage and it is closed. Then the vacuum is pumped in and the sample is ready for imaging. Then we should arrange the sample for milling by setting the focus and after a stage tilt of 520 towards the ion beam the milling is started on the sample. The interaction of ions with the sample damages the area where it is scanned for milling. To avoid more damage a carbon strip is deposited near the area of milling in order to secure the region of trench for obtaining good results. Different trench geometry sizes are milled on the sample and it is a step-by-step procedure.
Angle of Milling
Nature of Milling
Regular cross Section
Cleaning Cross Section
Cleaning Cross Section
Cleaning Cross Section
Table 1: Step By Step procedure for milling of trench in FIB.
The first step is depositing carbon strip at 30kv voltage and 0.30ÂµA current. Then a regular cross section of desired trench dimension is made on the sample at 30kv voltage and 7nA current. The next step is cleaning cross section, which is conducted at 30kv voltage and 3.0nA current and also at 30KV voltage 1.0nA current. Then finally for a low voltage of 10kv and 0.41ÂµA current and this process is repeated for other different dimensions. We can clearly observe the procedure of milling at different voltages and current with their nature of milling in table 1. The trench thus obtained from the milling process can be observed in figure 5 which is taken from FIB.
Figure 5: Image of FIB milled sample after completing milling procedure.
3.2 SEM Imaging and processing
SEM is used for getting the image of the milled trench for results. The sample is now placed in SEM. The sample is tilted at 520 stage tilt in SEM also because it is milled at 520 stage tilt in FIB for getting best results. This is carried in UHR mode of scanning at 1kv accelerating voltage and spot size of 3.0 at working distance of 5.2. The contrast and brightness are kept at 28.2 and 47.6 respectively. The images are scanned at two magnifications at 10000x and 5000x mag respectively. The correct values are given in table 2 given below.
Mode of Cam
Table 2: SEM values Taken at imaging
The contrast is determined by the equation Cpn = (Ip - In /Ip + In)200 where Ip and In are intensity of p and n areas. The data is processed in image J software which is used for processing the image for getting the results. The data collected from image J is calculated using the Microsoft excel sheet by evaluating the intensity profile and depth. The image J software used is shown in figure 6.
Figure 6: Picture of Image J tool.
We can observe the image in figure 7 which shows the image of the trench taken from SEM imaging after 520 tilt which is used for determining the results.
Figure 7: Image after 520 tilt in SEM
In FIB the ion beam is operated at 520 tilt for milling process and for getting the results the contrast must be taken at an angle of 520 in SEM for that reason the sample is tilted for the same angle in the SEM. But as there is an angle tilt so we had calculated the exact value of the tilt correction which is 1.27 for cos 520 using trigonometry and according to that results are calculated. The different trench sizes that are taken for milling on the sample are considering x as length, y as width and z as depth of the trench as represented in figure 7 and the trenches are represented by T1, T2 and T3 for the first section of results and the value are given in table 3 on plain silicon sample.
a. b. c.
Figure 8: a. 2d View of the trench representing the length as x, width as y and depth as z. b. Top view of the trench. c. Side view of the trench.
Trench dimension in Âµm
X Y Z
25 40 5
25 20 5
25 5 5
Table 3: Trench dimensions used in our project keeping the x value constant.
Figure 9: a. Bigger trench b. Smaller trench.
3.3 SEM Energy filtering
For obtaining better dopant contrast from the images obtained energy filtering is processed. In this the voltage used is 16KV. Due to energy filtering there is increase in contrast of the image and results obtained is far better from normal imaging.
Results and Discussion
Figure 10: Graph showing Depth vs Intensity profile graph for first three trenches
From figure 10 we can clearly understand that the contrast of the trench T1 and T2 are actually identical in spite of their differences in length, T3 has similar length as T2 but due to smaller width it results in a drop in the intensity curve which is due to artefact and are not the correct value. Such small trench width must therefore be avoided. In order to find the correct geometry there must be another graph with width size in between T2 and T3 for knowing the best possible dimension. As we had kept the value of x constant and calculated the results and due to the shrink of the dimensions of the devices we can't access a large area to mill which takes a long time and in industrial purpose it is not possible as millions of chips and devices have to be made in a very stipulated time so going for bigger trench size is worthless and so we decided to go for the smaller one and which is accurate and from the results obtained the value of y must be small such that the milling time would be minimal. But as there is a problem in depth the calculation of depth is also important for getting the critical value for the trench and so the trenches are made such as considering the same as x=length, y=width, and z=depth the value are given in table 4. Calculating the data obtained from table 4 we have the following results seen in figure 11.
Trench dimension in Âµm
X Y Z
15 20 5
30 20 5
35 20 5
10 20 5
Table 4: Trench dimensions used in our project keeping the y value constant
Figure 11: Graph showing Depth vs Intensity profile graph for results in table 4.
The redeposition of the sputtered material near the walls of the trench is more but the other two trenches are having a good response but then these are not the accurate geometrical dimension because of which there must be one accurate trench size which can be applied to the later samples and also which is reliable for the industrial purpose. For that purpose the SEM image of the smallest trench size is taken and the distance to the trench wall in x direction is calculated in order to find out the critical length and desired length and the following graph in figure 12 show us the results when observed in figure 13 and it is observed that the position nearer to the edge of the trench is not having the accuracy and as we go into the middle the graph becomes better and remains identical for all the other values which is observed in figure 13. The dark part in figure 13 represents that from that region the results are identical and remain equal.
Figure 12: Results showing the distance to the trench wall in x direction when taken from figure 13.
Figure 13: Trench Geometry of x=10, y=20, z=5 used for knowing the distance to the trench of wall in x direction
The optimum trench size was achieved by taking considrable steps of calculating the whole size of the image and also by vitue of the graphs we considered to take the smallest and accurate size which in our case is given by 10Âµm. For this particular trench size we concluded that the dimentions are as given in table 5. This means the smallest possible trench is used for doing the experments.
Figure 14: Image of Staircase structure at 10000x mag
Figure 15: Image of Staircase structure at 20000x mag.
Figure 16: Image of Staircase structure after energy filtering at 20000x mag
Figure 17: Image of resolution structure with diagonal stripes on the image.
Figure 18: Image of pn junction with diagonal stripes observed on the image.
Trench dimension in Âµm
X Y Z
10 20 5
10 20 5
10 20 5
Table 5: Optimum trech size used for different samples.
Pn contrast Flat n in n doped
Figure 19: Graph of Staircase structure result at 10000x mag
Figure 20: Graph of Staricase structure result at 20000x mag
Figure 21: Graph of Energy Filtered Staircase structure result.
We have considered the optimum trench size and cleaved the staircase structure, pn junction and also the resolution structure which you can have the images from figure 13-18. In the staircase structure we have observed that in the unfiltered imaging i.e in graph of figure 19 and figure 20 we can observe that it has poor resolution and low intensity profile. But from the previous results obtained it has the n type profile having much better curvature which is observed here. But for other results due to the charging effect of the sample there cannot be much seen as the results are affected by charging which gave us diagonal stripes on the imaging when imaged by SEM.
In staircase lower doping levels are low but higher doping levels is not expected profile possible because of oxide layer build up between FIB and SEM. The flat n doped region is observed in figure 19 which is the major observation which is observed. This is one new observation seen until this project.
In the case of pn junction and resolution structure also there was charging of sample and also the polishing step in not in much quality due to which the results could not be obtained. From figure 17 and figure 18 you can clearly observe the diagonal stripes. In pn junction and resolution structure due to charging it is typical for charging. No image analysis.
The energy filtered imaging in the case of staircase structure is observed without any charging and we can see from figure 21 that the result obtained shows that it has a low contrast but the profile is much better enhanced in this.
The effect of oxidation and carbon contamination is also one more effect that we observed here in this process. Due to the oxidation for much too longer it form a direct contact with the contrast such that the contrast would become very low and also we cannot see much profile from the resulting contamination. And also the carbon contamination is also observed in these results because of which there is change in the total profile.
From our results obtained one min thing we can observe is that our aim of the project which is trench geometry optimization to avoid artefacts could be achieved succesfully. Trench size has to exceed the ares by 3 Âµm in parralle for optimum trench. And also the oxidation effect we came to know the problems involved in the intensity profile. This was shown on the staircase structure. However also slope due to trench eliminated high dopant concentratioins are effected by oxidation and which redused the contrast. In future these oxidation effets and charging effects are to be rectified for getting the slope right.