Let’s chat GPC by chatGPT – Putting ChatGPT to the test and a first (?) article about ALD by AI

I’m sure you have heard the buzz in the field of artificial intelligence (AI)…

The promise of AI is to augment human abilities in areas such as decision-making, problem-solving, and other tasks. AI systems aim to simulate human intelligence and learn from data to improve their performance over time.

ChatGPT is an AI language model developed by OpenAI, it is capable of generating text in a conversational style, based on its training data, in response to prompts provided by users.

Well, the text in italic is generated by ChatGPT itself in response to my prompt: What is the promise of AI and what is ChatGPT?

But okay, that is an easy one! It is always easy to talk about yourself and what is keeping you busy. That even holds for a chatbot. But now let’s put ChatGPT to the test and see what it can tell us about atomic layer deposition (ALD).

To do so, I have done two things with ChatGPT .

First of all, I have asked ChatGPT several questions about ALD. You’ll find the responses it generated below. I have also added some comments at the end. By the way, note that others did ask ChatGPT some questions about ALD too, see for example Twitter here, here and here.

Next, in a separate chat, I asked ChatGPT two write a two-page article about ALD. I asked it to include a title and an abstract. This article is given below too. Note that this might be the first article ever about ALD written by AI! Or it might not be. As you can judge by yourself, AI can do amazing stuff!

(Just before going online, I found this blog by Jonas Sundqvist on BALDEngineering.com)

10 Questions about ALD and ChatGPT’s responses

What is ALD? What is the difference between ALD and CVD?

ALD stands for Atomic Layer Deposition. It is a thin film deposition technique used in the semiconductor industry to grow uniform and conformal films on substrates.

CVD stands for Chemical Vapor Deposition. It is a similar thin film deposition technique, but it involves the reaction of a gas phase precursor with a substrate surface to form a solid film.

The main difference between ALD and CVD is that ALD is a self-limiting process that produces ultra-thin films with atomic-level precision and control, whereas CVD is a non-self-limiting process that produces thicker films.

What are the key merits of ALD?

The key merits of ALD are:

Precision: ALD allows precise control over film thickness and composition, leading to uniform and conformal coatings down to the atomic scale.
Repeatability: The self-limiting nature of ALD provides high reproducibility and batch-to-batch consistency.
Film quality: ALD films exhibit excellent uniformity, high surface area, and low defect density.Selectivity: ALD can selectively deposit films on specific materials and patterns, making it ideal for complex and multi-layered structures.Compatibility: ALD is compatible with a wide range of materials and substrates, including flexible and 3D surfaces.
Scalability: ALD is a scalable process that can be applied to large areas, making it suitable for high-volume manufacturing.

Who invented ALD and when was that?

ALD was invented by Tuomo Suntola, a Finnish materials scientist, in the 1970s. Suntola was working at the Helsinki University of Technology (now Aalto University) when he developed the concept of ALD and published his findings in the late 1970s. The first commercial application of ALD was for the production of DRAM memory chips in the 1980s. Since then, ALD has become an important tool in the semiconductor industry for the growth of high-quality thin films and is widely used in the manufacture of various devices, such as microelectronics, batteries, and optoelectronics.

Which materials can be prepared by ALD? What are the most popular materials?

A wide range of materials can be prepared using ALD, including metals, oxides, nitrides, sulfides, and hydrides. The most popular materials that are prepared by ALD are:

Al₂O₃ (aluminum oxide): commonly used as a high-k dielectric for MOS gate insulation.

TiO₂ (titanium dioxide): used for optical applications, such as antireflection coatings and UV protection.

ZnO (zinc oxide): used as a transparent conductor in electronic devices.

HfO₂ (hafnium oxide): commonly used as a high-k dielectric material for DRAM memory devices.

Ta₂O₅ (tantalum pentoxide): used as a high-k dielectric material in advanced CMOS devices.

SiO₂ (silicon dioxide): used as a gate oxide and insulator in microelectronics.

These materials are popular in ALD because they can be deposited with excellent control over thickness and stoichiometry, leading to high-quality and uniform films.

How fast can you deposit a film by ALD? What does GPC mean and what are typical values of the GPC?

The deposition rate of an ALD film depends on many factors, such as the precursors used, the reaction conditions, and the substrate type. Generally, the deposition rate of an ALD film is on the order of a few angstroms per cycle, with a typical cycle time of a few seconds.

GPC stands for “Growth Per Cycle” and refers to the thickness of the film deposited in one cycle of the ALD process. The GPC is a measure of the deposition efficiency and is often used to compare the performance of different ALD systems or processes.

Typical values of the GPC for ALD films range from a few angstroms to a few nanometers per cycle, depending on the material and process conditions. For example, the GPC for Al₂O₃ films is typically in the range of 1-2 angstroms per cycle, while the GPC for ZnO films can be as high as 5-10 angstroms per cycle. The GPC can also be influenced by factors such as the choice of precursors, the reaction temperature, and the pressure, so it is often tailored for each application.

What are ALD precursors? What is the difference between thermal and plasma ALD?

ALD precursors are the starting materials that are used in the ALD process to form the desired film. They are typically volatile, low-molecular-weight compounds that can easily evaporate and react with the substrate surface. ALD precursors are chosen based on their reactivity, stability, and availability, as well as their compatibility with the desired film material.

The difference between thermal ALD and plasma ALD is the method used to activate the precursors for reaction. In thermal ALD, the precursors are activated by heating the substrate to a specific temperature, whereas in plasma ALD, a plasma is used to activate the precursors.

Plasma ALD has several advantages over thermal ALD, including faster reaction kinetics, improved precursor activation and utilization, and the ability to deposit films at lower temperatures. However, plasma ALD can also introduce unwanted impurities into the films, so careful control of the plasma conditions is required. Thermal ALD is a more established and well-understood process, but it is typically slower and requires higher temperatures. Both thermal ALD and plasma ALD have their strengths and weaknesses, and the choice between the two methods depends on the specific application and requirements.

What are typical temperatures for ALD? What is the ALD temperature window?

The typical temperatures for ALD depend on the specific precursors used and the desired film material. In general, ALD reactions are carried out at temperatures in the range of 150°C to 500°C, with most processes performed at temperatures between 200°C and 400°C. The exact temperature required for a specific ALD process can vary widely depending on the precursors used and the reaction kinetics.

The ALD temperature window is the range of temperatures over which an ALD process can be carried out without undergoing unwanted side reactions or degradation of the precursors. The ALD temperature window is determined by several factors, including the thermal stability of the precursors, the reaction kinetics, and the surface chemistry of the substrate. The ALD temperature window is an important consideration when choosing precursors and optimizing the ALD process, as it defines the range of temperatures over which the desired film can be deposited with high quality and uniformity.

For example, the ALD temperature window for the deposition of Al₂O₃ using trimethylaluminum (TMA) and water as precursors is typically in the range of 250°C to 300°C. The temperature must be carefully controlled within this range to ensure the desired film quality and avoid unwanted side reactions.

What are the applications of ALD? For which applications is ALD implemented in high-volume manufacturing?

Atomic layer deposition (ALD) is a thin-film deposition technology used in various applications:

Semiconductor devices: ALD is used in the fabrication of transistors, memory devices, and photovoltaic cells.
Coatings: ALD is used for deposition of protective and functional coatings for applications such as biomedical implants, optical devices, and energy storage.
Catalyst support: ALD is used to deposit uniform and conformal metal oxide layers for catalytic applications in fuel cells and electrocatalytic water splitting.

In high-volume manufacturing, ALD is commonly used in the following applications:

Semiconductor devices: ALD is used for high-k gate dielectric and metal contacts in integrated circuits and memory devices.
Display devices: ALD is used for deposition of thin-film encapsulation layers and polarizer films in LCD displays.
Energy storage: ALD is used for deposition of battery materials, such as cathode coatings and separator coatings.

Which companies are selling ALD equipment? Which companies are using ALD in their applications?

ALD equipment companies:

ASM International
Cambridge NanoTech
SUGINO Machine
AJA International
Aixtron
Beneq
TEL Tokyo Electron

Companies using ALD in their applications:

Intel
Samsung
TSMC
GlobalFoundries
SK Hynix
Micron Technology
LG Display
Sony
Panasonic
BASF.

What are the latest developments in ALD? Where is the field going?

The latest developments in ALD include:

Scalable ALD processes: There is ongoing work on developing scalable ALD processes to meet the needs of high-volume manufacturing in various industries.
Alternative precursors: Researchers are exploring alternative precursors for ALD that can provide better control over film properties, improve selectivity, and reduce the environmental impact of ALD processes.
New materials: Researchers are developing ALD processes for new and emerging materials, such as 2D materials, perovskite materials, and transition metal carbides, for a range of applications.
Integration with other technologies: There is ongoing work on integrating ALD with other thin-film deposition technologies, such as plasma-enhanced ALD and molecular beam epitaxy, to provide new capabilities and improve film properties.

The field of ALD is moving towards more scalable and cost-effective processes, improved material and process control, and the development of new and emerging materials. There is also increasing interest in the integration of ALD with other technologies, such as machine learning and data analytics, to improve process control and increase efficiency.

As I said, AI can do amazing stuff. I’m very impressed with the text ChatGTP comes up with. And it does it fast! To generate the answer for the first question took the longest (perhaps 1 minute), as well as the question about applications. But in this case, I also had to restart the chat.

Obviously, there are quite some mistakes in the text too. Quite a few things are clearly wrong, so be very careful! Just to name a few: ALD was not used for the production of DRAM memory chips in the 1980s, values of the GPC do not range up to a few nanometers per cycle, and ALD is not (yet!) used for deposition of battery materials in high-volume manufacturing. Re-iterating what many others have said before: ChatGPT can be very helpful but only for those topics in which you are already knowledgeable! In other cases, you might not be able to spot ChatGPT’s mistakes.

Furthermore, I also noticed that the argumentation is sometimes awkward. If it would be the text of one of my students, I would have the student make corrections. And references are missing too. That is quite an issue.

But anyway, I think this is just the start of something very big. I’m convinced that things will move rapid now and that we will be using AI extensively in our writing very soon. For me it has already started.

See also the two-page article about ALD that ChatGTP has generated. I think that proves my point!