We're Moving to
The content you are trying to access has moved to new site. Explore, discover (and in certain contries, buy) our products and services on our new website. Click to visit now or you will be automatically redirected.
to view non-migrated solutions
On an almost daily basis, industrial and consumer products manufacturers and news outlets highlight the many struggles happening within the microelectronics and semiconductor industry. New quandaries are shaping the industry to answer important questions like: How will the semiconductor industry continue to innovate and properly serve automotive, consumer electronics and medical device industries considering the current chip shortage?
Electrification of the drivetrain, microcontrollers or micro-components for intelligence functions like for safety features and for analog functions such as cabin temperature, speed, and sound – these are just a couple of the categories of vehicle semiconductors.
Innovation and production are directly affected by the chip shortage, and there are also logistical issues which have disrupted the supply chain during the Covid-19 crisis. However, we are here to discuss the ways in which the industry is employing quality analysis tools that are proving critical to semiconductor R&D, manufacturing, and fabrication (FAB). Here, we answer questions about the impurity management analyses that will enable the industry to provide quality and throughput of product.
Below are three Q&A sessions, by topic, transcribed from the PerkinElmer Semiconductor Virtual Symposium, featuring:
To go deeper than this Q&A, please watch the full on-demand virtual symposium sessions now.
Q&A with Aniket – Materials Characterization
Moderator:Why are hyphenation techniques important when analyzing semiconductor materials for quality?
Aniket: The quality and purity of raw materials directly affects semiconductor performance and lifetime use of the devices they serve. The mounting pressures of supplying the industry with chips are leading to increased adoption of analysis solutions that increase manufacturing productivity and the accuracy of detecting contaminants in chemicals, materials, gases, and the FAB environment.
Hyphenated solutions are an industry game-changer for semiconductors as they “complete the picture” of raw materials and parts characterization – by analyzing both the compositional make-up of the material, as well as the gases evolved from its decomposition.
When employing hyphenation techniques, typically a Thermogravimetric Analyzer (TGA) system is coupled with one of more analytical techniques, like FT-IR (Fourier Transform Infrared), MS (Mass Spectrometry), and/or GC/MS (Gas Chromatography/Mass Spectrometry), to provide the semiconductor industry with the most complete and advanced platforms for characterizing materials.
Hyphenated techniques contribute analyses with more power, higher levels of accuracy, and faster analyses by acquiring more information from a single run.
Moderator:How much material is needed for analysis with hyphenation techniques?
Aniket: The answer to this question has 2 parts. When discussing techniques that utilize systems like Thermal instruments, the PerkinElmer TGA 8000 can work with low sample mass, such as one milligram of material, due to its superior sensitivity and resolution. However, with 10 or 20 milligrams of material, an advantage is that there will be more gases evolved upon decomposition. Since nitrogen (or helium) is the carrier gas in a TGA analysis, the evolved gases from material is diluted with the carrier gas. So, the sample size is inversely proportional to the dilution that occurs.
We must make a compromise between signal quality and sample mass, especially in utilizing an infrared (IR) technique, where the carrier gas can dilute evolved gases. In contrast, when utilizing Mass Spectrometry (MS) or GC Mass Spectrometry (GC/MS), the effect of dilution in carrier gas is minimal, because the sensitivity is very high. So optimally, we want 1 to 10 milligram of sample size to start.
There are strategies and tricks that we can do in the hyphenation setup to minimize the dilution of gases. For example, we can utilize an enhanced siphon tube that is only a few milliliters from the sample surface, so that it draws in more and more of the evolved gases, so carrier gas dilution is reduced, resulting in better signal.
Moderator:Measurement of coefficient of linear thermal expansion (CTE) is the measure of a material to expand or contract with temperature change, and it is critical to semiconductors, not just in packaging but also in fabrication. Do materials characterization techniques allow for the measurement of non-self-supporting samples, such as films?
Aniket: A thermomechanical analyzer (TMA) like the TMA 4000, is not just for bulk measurement, but also for directional analysis. It is possible to measure non-self-supporting films on the TMA by utilizing a clamp setup that will allow you to position the sample in a tension mode and that is how we can measure the CTE. Any change in the sample dimensions due to heating or cooling is captured by means of a floating TMA probe setup.
Moderator:QA/QC in manufacturing and failure analysis, what are some of the challenges that need to be overcome to ensure final product quality?
Aniket: Problems can be introduced at any stage of the manufacturing process, requiring testing at multiple stages of the wafer manufacturing process, starting with the raw materials. Multiple analytical techniques can be employed to determine elemental and organic composition as well as physical testing of the materials.
Also, the presence, content and uniformity of the coatings that are applied to wafers need to be determined. And finally, residual impurities can have a negative impact on the performance of the wafer and/or the final manufactured device. So, control and detection of impurities is essential throughout the manufacturing process.
Another important facet of QA/QC is ensuring consistent quality across the entire wafer, since each wafer will produce many individual components once processed. Emerging technologies like accessories for Fourier transform infrared (FT-IR), specifically, the PerkinElmer Spectrum 3 MappIR system, enables automated measurement of a silicon wafer over the complete size range of wafers, ranging from 2” (50mm) up to 12” (300mm).
Spectrum 3 MappIR enables automated measurements at multiple points on the wafer, using either standard preset patterns or user-customizable patterns to determination critical wafer parameters across the entire wafer rather than just at one point in the center of the wafer. The software controls the measurement position, data collection and data analysis according to the method and calculation required.
Q&A with Dr. Ewa Pruszkowski – Inorganic Analyses
Moderator:How are ICP-MS analyses helping the semiconductor industry?
Pruszkowski: Elemental analysis by inductively coupled plasma mass spectrometry (ICP-MS) is a well-known technique to provide the detection capability needed by the semiconductor industry. Additional attributes, such as robustness, reliability, and speed, make it a must-have solution for silicon wafer manufacturing facilities.
In the semiconductor industry it’s especially important to analyze the incoming raw materials and chemicals to ensure purity. This really helps semiconductor chip manufacturers to avoid chip failures and obtain higher product yield. Also, analysis of particles and deposits on the wafers or in the native silicon oxide layer using VPD (vapor phase decomposition) with the ICP-MS as a detector is an important technique to prevent problems with final semiconductor products.
Moderator:What kind of semiconductor goals is a cold plasma technique utilized for?
Pruszkowski: During semiconductor production acids and acid mixtures are utilized for many important processes, such as cleaning silicon wafer surfaces. In the cold plasma technique, also known as cool plasma technique, energy of the plasma is lowered, so ionization of argon and argon species is not efficient. This leads to less spectral interferences and more accurate data. So, we effectively have less spectral interferences, leading to better data.
Only elements with a low ionization potential can be effectively run in cold plasma conditions. This technique is mainly used for clean chemicals without high total dissolved solids (TDS). Hot and cold plasma conditions can be run in one method with conditions changing automatically.
Moderator:In terms of utilizing a cleanroom, we know we are going to get better detection limits in a cleanroom, but what cleanroom class do you need to achieve sub parts per trillion (PPT) background equivalent concentrations (BECs)?
Pruszkowski: A lot of our experiments and publications were done in the cleanroom 10,000 which is ISO 7. This is not the best cleanroom standard in the world. I would recommend performing such analyses in a cleanroom, at least class 1000 or higher for the best results.
Alternatively, you could use sample preparation systems which are enclosed or have neutral gas coming in to separate them from the atmosphere. For example, ESI prepFAST S (for semiconductors) system not only analyzes samples but also prepares standards.
Moderator:How do you distinguish between metal ions and metal nanoparticles in high purity chemicals for semiconductors? What are typical detection limits in terms of size for metal nanoparticles?
Pruszkowski: We can distinguish between metal ions and nanoparticles by looking at the real-time signal. The dissolved metals will have constant background, however, with particles you will see spikes. We can measure dissolved and its nanoparticles at the same time.
Although nanoparticles could be analyzed using traditional single quadrupole ICP-MS instruments, the NexION® 5000 multi-quadrupole ICP-MS, allows you to perform nanoparticle analyses more efficiently because spectral interferences are practically removed. This enables better detection limits for dissolved metals as well as nanoparticles.
Regarding the typical detection limit in terms of size, it depends on the element, but the NexION 5000 ICP-MS can detect particles down to 10 or 15 nanometers.
Moderator:What is the difference between MS/MS (also known as on-mass) and Mass Shift modes in ICP-MS?
Pruszkowski: There are two important modes that ICP-MS can operate in, MS/MS and Mass Shift. In MS/MS mode, Q1 and Q3 are set to the same mass, and through reactions in the cell, interferences are removed from that mass.
However, in Mass Shift mode, Q1 is set to the analyte mass but Q3 is set to the mass where the cluster of ions of analyte which are created with the cell reside. For example, in an analysis of titanium with ammonia, the ammonia reacts very quickly with titanium, creating several clusters. The most prominent clusters at masses 114 and 131 can both be used for analytical work. The advantage of this is that the background at masses where the cluster exists is free from interferences, so we can obtain very good detection limits and BECs.
Many elements react with gases such as NH3, O2 and Ch4 creating clusters. This allows us to utilize a multi-quadrupole instrument while taking advantage of Mass Shift mode.
Q&A with Miles Snow – Gas Chromatography
Moderator:Why is GC/MS so important to the semiconductor industry?
Snow: GC/MS can do so much more than the historically popular GC with a Flame Ionization Detector (FID). FID is utilized to analyze for atmospheric molecular contaminations; however, the compound identification is based upon retention time. Often compound interference or a totally separated peak can both occur and with FID by itself, it’s very difficult to identify exactly what that component is.
A mass spectrometry (MS) detector offers several advantages as a technique to identify contaminants in semiconductor raw materials and all the way through manufacturing and FAB. GC/MS not only obtains the retention time, but also obtains a spectra of the various compounds that may come out as your targets. You can use that data to verify the identification of your targets and it’s also extraordinarily useful for unknown components.
If unknown components are found, you can get the associated spectra of those, and then match them against a library for tentative identification and then run an actual sample for verification. GC/MS gives you that extra layer of confidence that you’ve identified exactly what the contamination is.
Moderator:Is it possible to add a liquid autosampler on the GC/MS system with an online Thermal Desorption (TD) System?
Snow: Yes, you can hyphenate a GC/MS by adding a liquid autosampler with an online TD, and we can look at two scenarios for this. You could inject the sample into a different column with a different detector. Or you could inject the sample into a different injector and have the sample path merge into the same column as the online thermal desorption system. In this case, it would go to that same detector or detectors, and it would then produce your results.
Moderator:Can hydrogen or nitrogen be utilized as carrier gases in GC/MS techniques?
Snow: The simple answer is yes and there are a couple of different strategies. Some operators feel, if they are running helium; it's extraordinarily expensive and might want to switch to another gas. One thought is to switch 100% over to nitrogen or switch over to hydrogen, but these gases have some disadvantages. Hydrogen, as we know, is flammable, so you must take some extra precautions, but it can be done.
Nitrogen is considered because it is very convenient, and inexpensive, but it degrades chromatographic performance. Neither one is ideal, so another strategy is using online, thermal desorption, you are trying to collect and concentrate the analytes. In this case, the purging can be done with nitrogen and then introduce it on the column - and use a different gas, such as helium which would be relatively low consumption of the expensive gas. In this case, we get effective helium like separations, but the amount used is small.
In some examples, we’ve set up equipment like this and a tank of helium can last six months, nine months and the usage stays down in the area of 30, 40 or 50 mils a minute. In this case, one tank every 9 months is rather good. And for a similar scenario, you could and switch to hydrogen; it is up to the user how they want to conserve by switching gases to optimize their operational costs.