8.3.2 Sputter etching: argon ion etching of gold

One commercial process for cutting inkjet printer nozzles uses sandblasting. Not surprisingly, the surface finish is rather poor and there are issues with particles contaminating the devices. However, it is a physical process very like this that we need if we are to achieve a vertical etch profile.

The key is directed bombardment by highly energetic particles. When processing on the microscale, these particles are not sand grains but ions accelerated towards the surface by an electric f
Author(s): The Open University

License information
Related content

Except for third party materials and/or otherwise stated (see terms and conditions) the content in OpenLearn is released for use under the terms of the Creative Commons Attribution-NonCommercial-Share

7.3.1 Electroplating

Electroplating is a relatively fast process, inexpensive and simple, although fairly messy and limited in applicability. The wafer is dipped into a solution with dissolved salts of the metal (e.g. CuSO4 + H2SO4) and is connected to a negative voltage. A positive metal electrode (anode), also in the solution, completes the circuit. Anywhere that current can flow into the wafer surface, metal will be deposited. Plating has several advantages: it will deposit met
Author(s): The Open University

License information
Related content

Except for third party materials and/or otherwise stated (see terms and conditions) the content in OpenLearn is released for use under the terms of the Creative Commons Attribution-NonCommercial-Share

7.3 Depositing metals and alloys

Metal layers are used extensively in device fabrication: to carry current for both power and signals, to apply the voltages that control transistors and generate forces for MEMS, as mirrors and optical coatings, and in magnetic devices for recording media. Different applications might require a continuous film, a long track, multiple thin layers or a plug filling a ‘via hole’ through to a buried layer. The electrical properties resulting from micro structure and composition must be contro
Author(s): The Open University

License information
Related content

Except for third party materials and/or otherwise stated (see terms and conditions) the content in OpenLearn is released for use under the terms of the Creative Commons Attribution-NonCommercial-Share

7.2.5 Stress

Perfect crystal structure can be achieved only by epitaxial growth, where deposits are formed atomic layer by atomic layer. The lattice planes of the deposited film merge seamlessly with those of the substrate on which it is deposited. Even when the crystal lattices of these two materials match in shape, however, they will never be a perfect match in size, owing to differences in atomic spacing between the deposit and the substrate; so, the deposited film will be stretched and distorted to ma
Author(s): The Open University

License information
Related content

Except for third party materials and/or otherwise stated (see terms and conditions) the content in OpenLearn is released for use under the terms of the Creative Commons Attribution-NonCommercial-Share

7.2.4 Microstructure

There are many practical requirements for a film at the microscopic level.

Firstly, it must be firmly bonded to the surface on which it's deposited. A poorly adhered film can peel or flake away under the stresses of later processing steps, or may lead to a reliability issue that will plague the device throughout its lifetime. Similarly, issues can arise with wetting if a deposited material prefers to form droplets on the surface rather than spreading out into a uniform film. A thin adhe
Author(s): The Open University

License information
Related content

Except for third party materials and/or otherwise stated (see terms and conditions) the content in OpenLearn is released for use under the terms of the Creative Commons Attribution-NonCommercial-Share

7.2.3 Chemical composition

As outlined in Table 2, some deposition techniques are best suited to conducting materials, whereas others come into their own only for chemical compounds. In either case, chemical composition may be an important consideration. Impurities can interfere with the conduction properties of the material (notably in t
Author(s): The Open University

License information
Related content

Except for third party materials and/or otherwise stated (see terms and conditions) the content in OpenLearn is released for use under the terms of the Creative Commons Attribution-NonCommercial-Share

7.2.2 Step coverage (conformality)

Not all layers require precise thickness control. Sometimes all that matters is that the film coats the entire surface, including vertical walls and (most difficult) the corners at the bottom of deep holes.

For example:

  • We may require an insulating layer of oxide between two conductors.

  • A protective titanium nitride barrier layer prevents aluminium from diffusing into silicon, and an underlying titanium adhesion layer ensures that
    Author(s): The Open University

    License information
    Related content

    Except for third party materials and/or otherwise stated (see terms and conditions) the content in OpenLearn is released for use under the terms of the Creative Commons Attribution-NonCommercial-Share

7.2.1 Thickness control and uniformity

Often, final device characteristics, such as the value of a capacitor, the threshold voltage of a transistor, the resistance per square in a thin-film resistor and the resonant frequency of an acoustic wave filter, depend strongly on the thickness of a deposited layer. Therefore it must be ensured that the layer thickness is the same at all points on the wafer, and on every wafer that comes off the production line. Specifications of ±1% uniformity and reproducibility are not uncommon, and so
Author(s): The Open University

License information
Related content

Except for third party materials and/or otherwise stated (see terms and conditions) the content in OpenLearn is released for use under the terms of the Creative Commons Attribution-NonCommercial-Share

6.5 Q-value

The rate at which the mass–spring system loses energy to its surroundings is referred to as the Q-value for the oscillator. The Q-value is defined as:

ΔE/E is the fractional energy loss per cycle of the oscillation. This can also be expressed in term
Author(s): The Open University

License information
Related content

Except for third party materials and/or otherwise stated (see terms and conditions) the content in OpenLearn is released for use under the terms of the Creative Commons Attribution-NonCommercial-Share

6.3.1 Damped harmonic oscillator

Starting once again with Newton's second law but including the additional damping force in the equation gives:

Author(s): The Open University

License information
Related content

Except for third party materials and/or otherwise stated (see terms and conditions) the content in OpenLearn is released for use under the terms of the Creative Commons Attribution-NonCommercial-Share

6.3 Damping

In the real world, most oscillations are subject to damping and so the amplitude of the oscillation dies away over time. For example, the bell mentioned earlier would not be very effective if it did not lose some of its energy as sound waves. The oscillating cantilever of the AFM will, like the simple mass-spring system, be subject to frictional forces from the air, the material of the cantilever itself, and the fixing point.

For the mass-spring system the damping force Fd
Author(s): The Open University

License information
Related content

Except for third party materials and/or otherwise stated (see terms and conditions) the content in OpenLearn is released for use under the terms of the Creative Commons Attribution-NonCommercial-Share

6.1 Why is resonance important?

This section aims to take you through some general ideas about vibrations, which will help you understand the principles behind the resonant behaviour of the AFM probe tip. Vibrations and oscillations crop up in many contexts. They can be modelled mathematically and form a general topic in mathematics about vibrations and oscillations in which the appropriate balances between forces and accelerations are formulated into differential equations.

Students of physics and chemistry also get
Author(s): The Open University

License information
Related content

Except for third party materials and/or otherwise stated (see terms and conditions) the content in OpenLearn is released for use under the terms of the Creative Commons Attribution-NonCommercial-Share

5.1.2 Dipole-dipole forces

In the case of dipole-dipole interactions, the molecules that bond together have a fixed asymmetry in their charge distributions (as is the case in Figure 22); if their orientations are favourable the two will bond together. All molecules produce London forces. The dipole-dipole interactions are in addition to t
Author(s): The Open University

License information
Related content

Except for third party materials and/or otherwise stated (see terms and conditions) the content in OpenLearn is released for use under the terms of the Creative Commons Attribution-NonCommercial-Share

5.1.1 London forces

The distribution of charge in one atom or molecule occurs naturally as the electrons move around the nucleus. If a second atom or molecule is introduced, the charge distribution from the first will induce a complementary charge distribution in the second. Looking at Figure 22 you can see that the negative bias o
Author(s): The Open University

License information
Related content

Except for third party materials and/or otherwise stated (see terms and conditions) the content in OpenLearn is released for use under the terms of the Creative Commons Attribution-NonCommercial-Share

4.2 The piezoelectric effect at the atomic scale

It has been mentioned above that by changing the state of polarisation of a piezoelectric material we can generate movement, and vice versa. Let's examine a little more deeply what is meant by ‘state of polarisation’ and how we can maximise its effect to get the best out of electrically controlled micro-actuators.

In order to electrically polarise a material we need, by definition, to cause a separation of charges within the material. The more we can do this the greater the d
Author(s): The Open University

License information
Related content

Except for third party materials and/or otherwise stated (see terms and conditions) the content in OpenLearn is released for use under the terms of the Creative Commons Attribution-NonCommercial-Share

4.1 The piezoelectric effect

The phenomenon of piezoelectricity was first predicted and demonstrated in the late nineteenth century using naturally occurring materials. It has a vast number of applications, ranging from spark ignitors to inkjet printers. It is also utilised in timing circuits, where an oscillating electric field is used to make a quartz crystal resonate at its natural frequency. In MEMS, the effect is used to generate small-scale movements in a range of devices known as micro-actuators.

The effect
Author(s): The Open University

License information
Related content

Except for third party materials and/or otherwise stated (see terms and conditions) the content in OpenLearn is released for use under the terms of the Creative Commons Attribution-NonCommercial-Share

3.8 Review

In the first three sections, we have looked at devices whose usefulness is dependent on their form. In the case of the Pirani sensor, it was the dimensions of the microbridge that affected its sensitivity; in the AFM probe, its ability to resolve features on a surface is determined mostly by the form of the last few nanometres of its very tip. With devices the emphasis is not so much on the form of the structure as on how to make it move in the right way and, just as importantly, how to detec
Author(s): The Open University

License information
Related content

Except for third party materials and/or otherwise stated (see terms and conditions) the content in OpenLearn is released for use under the terms of the Creative Commons Attribution-NonCommercial-Share

3.7.4 The carbon-nanotube tip

A way of escaping the issues affecting process compatibility that arise from the use of techniques such as oxidation sharpening is simply to assemble the probe from separate parts – and this has been successfully done using carbon nanotubes. Single-walled carbon nanotubes can have diameters as small as 0.4 nm, but more typically they are of the order of 1 to 2 nm. This represents a great improvement on the radii of curvature achieved with oxidation sharpening. One might have thought that it
Author(s): The Open University

License information
Related content

Except for third party materials and/or otherwise stated (see terms and conditions) the content in OpenLearn is released for use under the terms of the Creative Commons Attribution-NonCommercial-Share

3.7.3 Oxidation sharpening

The all-important radius of curvature of the tip can be made smaller – both in the silicon nitride and pure silicon tips – by the trick of oxidation sharpening. Figure 15 shows the principle. The original tip is first heated to around 1000 °C in an oxidising atmosphere, such that the outermost micrometre
Author(s): The Open University

License information
Related content

Except for third party materials and/or otherwise stated (see terms and conditions) the content in OpenLearn is released for use under the terms of the Creative Commons Attribution-NonCommercial-Share

3.7.2 The hybrid probe

Where the benefit to be gained from combining the narrow-angled silicon tip with the super-light, any-shape silicon nitride cantilever outweighs the expense and difficulty of the more complex process sequence, AFM probes can be made with silicon tips on silicon nitride cantilevers. Figure 14 shows one such pro
Author(s): The Open University

License information
Related content

Except for third party materials and/or otherwise stated (see terms and conditions) the content in OpenLearn is released for use under the terms of the Creative Commons Attribution-NonCommercial-Share