OR WAIT 15 SECS
Volume 28, Issue 11
Evaluation guidelines and important factors for laboratory scientists to consider when choosing an inductively coupled plasma–mass spectrometry instrument
This article presents a set of evaluation guidelines — developed from the author's 20 years of experience demonstrating inductively coupled plasma–optical emission spectrometry (ICP-OES) and ICP–mass spectrometry (MS) instrumentation and running customer samples — to help laboratory scientists make the right decision when selecting an ICP-MS instrument. It focuses on the evaluation process and the most important factors to consider rather than on how to compare the analytical performance of the instruments being evaluated.
So you have read the chapter in my book (1) on comparing inductively coupled plasma–mass spectrometry (ICP-MS) with other atomic spectroscopic techniques and convinced your boss that it is the ideal technique for your laboratory. However, understanding how the instrument works doesn't necessarily give you an insight into how to compare different designs, hardware components, and the difficult-to-define intangible factors such as ease of use, reliability, and technical support; these factors are of critical importance when you have to make a decision regarding which instrument to purchase or what vendor to go with. A number of excellent commercial instruments are available on the market that all look very similar and have very similar specifications, but how do you know which instrument is the best for your application needs or which company will take care of you the best? This article presents some guidelines on how best to evaluate the capability of the instrumentation and to assess the expertise and commitment of the ICP-MS vendor.
It is very important before you begin the selection process to decide what your evaluation objectives are. This is particularly important if you are part of an evaluation committee. It is fine to have more than one objective, but it is essential that all the members of the group begin the evaluation process with the objectives clearly defined. For example, is detection limit performance an important objective for your application, or is it more important to have an instrument that is easy to use? If the instrument is being used on a routine basis, good reliability may be very critical. On the other hand, if the instrument is being used to generate revenue, perhaps sample throughput and cost of analysis are of greater importance. Every laboratory's scenario is unique, so it is important to prioritize before you begin the evaluation process. So, in addition to looking at performance, instrument features, and hardware components, you should make the comparison with your evaluation objectives in mind, based on your particular application needs.
Although price can be considered a valid evaluation guideline, especially if you only have limited funds to purchase an instrument, it will not be discussed in this article. However, you will probably own the instrument for 10 years or more. For that reason, you should purchase the instrument that is best suited for your needs and not the instrument with the lowest price tag. If you are interested in the financial side of the evaluation process, it is presented in greater detail in my book (1). With that in mind, let's take a look at the three most common selection criteria used in the evaluation process.
They typically include
Let's examine these in greater detail.
Analytical performance can mean different things to different people. The major reason that the trace-element community was attracted to ICP-MS almost 30 years ago was its extremely low multielement detection limits. Other multielement techniques, such as ICP-optical emission spectrometry (OES), offered very high throughput but just could not get down to ultratrace levels. Although electrothermal atomization (ETA) offered much better detection capability than ICP-OES, it did not offer the sample throughput capability that many applications demanded. In addition, ETA was predominantly a single-element technique and so was impractical for carrying out rapid multielement analyses. These limitations quickly led to the commercialization and acceptance of ICP-MS as a tool for rapid ultratrace-element analysis. However, there are certain areas where ICP-MS is known to have weaknesses. For example, dissolved solids for most sample matrices must be kept below 0.2%; higher levels can lead to serious drift problems and poor precision.
Polyatomic and isobaric interferences, even in simple acid matrices, can produce unexpected spectral overlaps that will have a negative impact on your data. High-resolution instrumentation and collision–reaction cell and interface technology are helping to alleviate these spectral problems, but they also have their limitations. Depending on the types of samples being analyzed, matrix components can dramatically suppress analyte sensitivity and affect accuracy. These potential problems can all be reduced to a certain extent, but different instruments approach and compensate for these problem areas in different ways. With a novice, it is often ignorance or a basic lack of understanding of how a particular instrument works that makes the selection process more complicated than it really should be. So, any information that can help you prepare for the evaluation will put you in a much stronger position.
It should be emphasized that these evaluation guidelines are based on my personal experience and should be used in conjunction with other material in the open literature that has presented broad guidelines to compare figures of merit for commercial instrumentation (2–4). In addition, you should talk with colleagues in the same industry or application segment as yourself. If they have gone through a lengthy evaluation process, they can give you valuable pointers or even suggest an instrument that is better suited to your needs. Finally, I strongly suggested that you narrow the actual evaluation to two or three commercial products. By carrying out some pre-evaluation research, you will gain a better understanding of what ICP-MS technology or instrument to focus on. For example, if funds are limited and you are purchasing an ICP-MS system for the very first time to perform high-throughput environmental testing, it is probably more cost-effective to focus on quadrupole technology. On the other hand, if you are investing in a second system to enhance the capabilities of your quadrupole instrument, it might be worth taking a look at magnetic sector or triple-quadrupole collision–reaction cell instruments. Or, if fast multielement transient peak analysis is your major reason for investing in ICP-MS, time-of-flight (TOF) or Mattauch-Herzog simultaneous sector systems should be given serious consideration. Alternatively, if your instrument is going to be used to characterize nanoparticles using the single-particle ICP-MS technique, the speed of data acquisition and the response time of the ICP-MS quadrupole and detector electronics must be fast enough to capture the time-resolved nanoparticles pulses, which typically last only a few milliseconds or less.
So depending on your application requirements, several major performance criteria should be evaluated:
It is beyond the scope of this article to go into the evaluation of all these performance criteria, because it would require more detail than can be conveyed in a magazine article. There are many ways to compare performance, particularly when it comes to the demands of different application areas. However, it will be easier to compare the capabilities of each instrument being evaluated by using exactly the same set of evaluation samples and calibration standards. The samples can include unknowns, your own quality control (QC) standards, and certified reference materials, but it is important that you supply these samples already prepared; you should take all the necessary precautions in preparing the samples for an ultratrace-element technique like ICP-MS (1). In this way, you are assessing the capability of the instrument and not the expertise of the vendor to carry out dissolution and preparation of the sample or making up the calibration standards and blanks. It's certainly valid to evaluate the expertise of the vendor's analytical chemist, but try to separate that evaluation from the performance of the instrument.
It's also important to emphasize that in today's ICP-MS marketplace, all commercially available instruments are of the highest quality. Despite what each vendor might tell you, they will all do an excellent job for you. There will clearly be subtle differences in the capability of each instrument, based on your application requirements. More-detailed information is available from my textbook and the short course that I teach at the Pittsburgh Conference every year (1,5).
But even after you have run a set of evaluation samples, it might not be possible to declare a clear winner. So the final decision will often come down to factors such as ease of use, reliability, and technical support from the vendor. The rest of this article will therefore focus on these intangible factors, which are sometimes extremely difficult to compare.
In most applications, analytical performance is a very important consideration when deciding what instrument to purchase. However, the vast majority of instruments being used today are being operated by technician-level chemists. They usually have had some experience in the use of trace-element techniques such as atomic absorption (AA) spectroscopy or ICP-OES, but in no way could they be considered experts in ICP-MS. Therefore, a system's usability aspects might be competing with its analytical performance as the most important selection criteria, particularly if the application does not demand the ultimate in detection capability. Although usability is in the eye of the user, there are some general considerations that need to be addressed. They include but are not limited to the following aspects:
First, you need to determine the skill level of the operator who is going to run the instrument. If the operator is a PhD-level scientist, then maybe ease of use is not such an important factor. However, if the instrument is going to be used in a high-workload environment and possibly be operated around the clock, there is a strong possibility that the operators will be less qualified. Therefore, you should be looking at how easy the software is to use and how similar it is to other trace-element techniques that are used in your laboratory. Ease of use will definitely have an impact on the time it takes to fully train a person to use the instrument. Another issue to consider is whether the person who runs the instrument on a routine basis is the same person who will be developing the methods. Correct method development is critical because it impacts the quality of your data; therefore it usually requires more expertise than just running routine methods.
This article will not discuss software features or operating systems, because they are complicated to evaluate and decisions tend to be based more on a personal preference or comfort level than on the actual functionality of ICP-MS software features. Instrument software is also a moving target, because it is continually being modified and updated. However, there are differences in the way software feels. For example, if you have come from an MS background, you are probably comfortable with more-sophisticated software. Alternatively, if you come from a trace-element background and have used AA or ICP-OES, you are probably used to more routine software that is relatively straightforward to use. You will find that different vendors have come to ICP-MS from a variety of different analytical chemistry backgrounds, which is often reflected in the way they design their software. Depending on the way the instrument will be used, an appropriate amount of time should be spent looking at software features that are specific to your application needs. For example, if you are working in a high-throughput environmental laboratory, you might be interested in turnkey methods that are used to run a particular United States Environmental Protection Agency (EPA) methodology, such as Method 200.8 for the determination of elements in water and wastes. In addition, maybe you should also be looking very closely at all the features of the automated QC software, or if you do not have the time to export your data to an external spreadsheet to create reports, you might be more interested in software with comprehensive reporting capabilities.
In some highly regulated industries, the operating and reporting software needs to be compliant with a set of regulated standards and guidelines. For example, the pharmaceutical and food manufacturing industries are affected by federal regulations set down by the Food and Drug Administration (FDA) in 21 CFR Part 11, which specifies detailed requirements that computerized systems need to fulfill to allow electronic signatures and records in lieu of handwritten signatures on paper records. There are other industries that are highly regulated, so it is very important that any ICP-MS system used in a regulated environment has all the necessary software to be compliant.
As mentioned previously, optimized method development is absolutely critical in ICP-MS. Factors such as torch alignment, plasma and nebulizer gas flows, ion lens voltages, and collision–reaction cell gas conditions are all variables than can potentially impact the sensitivity, detection limits, interference reduction capabilities, and precision of your analysis. You should set aside enough time in your evaluation to carefully consider these factors and understand how different instruments compare with regard to ease of method development.
This is most definitely the case with collision–reaction cells and interfaces, especially with a new method. It can take a great deal of time and effort to select the best gases, gas flows, and cell parameters to maximize the reduction of interferences for certain analytes in a new sample matrix. Using a single gas in a simple collision cell can be beneficial to setting up a method, but can it determine your full suite of elements at the detection limits you require in all your sample matrices? And does the use of a collision cell degrade the detection capability of the analytes that don't actually need it? On the other hand, using reaction chemistry typically offers the best interference reduction capabilities and lowest detection limits, but it can be quite time-consuming, particularly if more than one reaction gas is required. Reaction mechanisms are very well understood nowadays because vendors have been investigating the optimum reaction gases for different sample matrices for almost 15 years, since collision–reaction cells were developed. As a result, a large amount of application data is available in the public domain.
ICP–MS systems are complex pieces of equipment that, if not maintained correctly, have the potential to fail when you least expect them to. For that reason, a major aspect of instrument usability is how often routine maintenance has to be performed, especially if complex sample matrices are being analyzed. You must not lose sight of the fact that your samples are being aspirated into the sample introduction system and the resulting ions generated in the plasma are steered into the mass analyzer via the interface and ion optics. In other words, the sample, in one form or another, is in contact with many components inside the instrument. Modern instruments require much less routine maintenance than older-generation equipment, but it is still essential to find out which components need to be changed and at what frequency to keep the instrument in good working order. You should be asking the vendor which components need to be replaced or inspected on a regular basis and what type of maintenance should be done on daily, weekly, monthly, or yearly intervals. In other words, determine the tasks required to keep the instrument in good working order. I also encourage you to talk to real-world users of the equipment to make sure you get their perspective of these maintenance issues. This step is particularly important if you are investing in a brand new instrument that hasn't been on the market long. There won't be a large number of users in the field, so it's important you talk to them to get their real-world perspective of the routine maintenance issues.
Alternative sample introduction techniques are becoming more necessary because ICP-MS is being utilized to analyze more complex sample types. Therefore, it is important to know if a particular sampling accessory is made by the ICP-MS instrument company or by a third-party vendor. Obviously, if it has been made by the same company, compatibility should not be an issue. However, sampling accessories made by a third party may work much better with some instruments than with others. It might be that the physical connection of coupling the accessory to the ICP-MS torch has been better thought out, or that the software "talks" to one system better than another. A good example of this is in the coupling of liquid or gas chromatographic techniques with ICP-MS for the separation and detection of different elemental species. There are many excellent commercially available chromatography systems that are capable of performing low-level speciation determinations, but the chromatography system must connect relatively seamlessly to the ICP-MS system. In addition, not all chromatography data handling software is the same, so make sure you fully evaluate its capabilities, flexibility, and ease of use, because it will have a direct impact on the time required to develop and run a speciation method on a routine basis.
You should also understand how much routine maintenance a sampling accessory needs. The benefits of a rugged ICP-MS system that requires very little maintenance is negated if the sampling accessory needs to be cleaned every 5–10 samples. (See reference 1 for more details about the suitability of sampling accessories such as laser ablation for the analysis of solid materials, high performance liquid chromatography for trace-element speciation studies, or field-flow fractionation for the characterization of nanoparticles.) And if a sampling accessory is required, software–hardware compatibility should be one of your evaluation factors.
However, sampling-accessory compatibility is not the same factor as how the sampling accessory impacts the performance of the instrument. For example, when a dry sample aerosol generated by a laser ablation system is introduced into an instrument, how do the ablated particles affect the torch, interface cones, and ion optic system? A liquid sample aerosol might not cause any problems, but dry particles ablated from a solid surface could possibly be deposited on components in the interface region and have a negative impact on the instrument's sensitivity or signal stability.
Installation of the instrument and where it is going to be located does not seem to be an obvious evaluation consideration at first, but it could be important, particularly if space is limited. For example, is the instrument freestanding or bench-mounted? Maybe you have a bench available but no floor space, or vice versa. It could be that the instrument requires a temperature-controlled room to ensure good stability and mass calibration. If this is the case, have you budgeted for this kind of expense? If the instrument is being used for ultratrace detection levels, does it need to go into a class 1, 10, or 100 clean room? If it does, what is the size of the room and do the roughing pumps need to be placed in another room? In other words, it is important to fully understand the installation requirements for each instrument being evaluated and where it will be located (for more details, see reference 1).
Technical and application support is a very important consideration, especially if you have no previous experience with ICP-MS. You want to know that you are not going to be left on your own after you make the purchase. Therefore, it is important to know not only the level of expertise of the specialist who is supporting you, but also whether they are local to you or located in the manufacturer's corporate headquarters. In other words, can the vendor give technical help whenever you need it? Another important aspect related to application support is the availability of application literature. Is a wide selection of material available to help you develop your methods, either in the form of Web-based application reports or references in the open literature? Also, find out if there are active users or Internet-based discussion groups, because they will be an invaluable source of technical and application help. One excellent source of help in this area can be found on the PlasmaChem Listserver, a plasma spectrochemistry discussion group out of Syracuse University (6).
Find out what kind of training course comes with the purchase of the instrument and how often it is run. Most instruments come with a 2–3 day training course for one person, but most vendors will be flexible regarding the number of people who can attend. Some manufacturers also offer application training, where they teach you how to optimize methods for major application areas such as environmental, clinical, and semiconductor analysis. Talk to other users about the quality of the training they received when they purchased their instruments and also ask them what they thought of the operator's manuals. You will often find that this is a good indication of the importance a vendor places on customer training.
To a certain degree, instrument reliability is impacted by routine maintenance issues and the types of samples being analyzed, but it is generally considered more of a reflection of the design of an instrument. Most manufacturers will guarantee a minimum uptime percentage for their instrument, but this number (which is typically ~95%) is almost meaningless unless you really understand how it is calculated. Even when you know how it is calculated, it is still difficult to make the comparison. But at least you should understand the implications if the vendor fails to deliver the uptime they promise. Good instrument reliability is taken for granted nowadays, but it has not always been the case. When ICP-MS was first commercialized in 1983, the instruments were a little unpredictable, to say the least, and were quite prone to frequent downtime. However, as the technique became more mature, both the quality of instrument components and overall reliability improved. You should therefore be aware of the instrument components that are more problematic than others. This is particularly true when the design of an instrument is new or a model has had a major redesign. You will therefore find that in an instrument's design life cycle, instruments that are "more mature" in age will have a known track record, particularly when being used in a high-sample-throughput environment. For this reason, it is very important that you talk to real users in your application field to get their input.
So when it comes to instrument reliability, it is important to understand whether it is related to the samples being analyzed, the inexperience of the operator, a failed or unreliable component, or an inherent weakness in the design of the instrument. For example, how does the instrument handle highly corrosive chemicals such as concentrated mineral acids? Some sample introduction systems and interfaces will be more rugged than others and will require less maintenance. On the other hand, if an operator is not aware of the dissolved solids limitation of the instrument, he or she might attempt to aspirate a sample, which will slowly block the interface cones, causing signal drift and, in the long term, possible instrument failure. Or, it could be something as unfortunate as the failure of a major component such as the radio frequency (RF) generator power amplifier tube, detector, or turbomolecular pump. Although this scenario is extremely unlikely, all these components have a finite lifetime and as the instrument approaches retirement, the likelihood of a component failure will increase.
Instrument reliability is very difficult to assess at the evaluation stage, so you have to look very carefully at the kind of service support offered by the vendor. For example, how close is a qualified support engineer to you? What is the maximum amount of time you will have to wait to get a support engineer at your laboratory or at least to call you back to discuss a problem? Ask the vendor if they have the capability for remote diagnostics, where a service engineer can remotely run the instrument or check the status of a component by "talking to" your system computer via the Internet. Even if this approach does not fix the problem, at least the service engineer can come to your laboratory with a very good indication of what the problem could be.
You should know up front what a service visit is going to cost you, irrespective of what component has failed. Also find out which routine maintenance jobs you can do and which ones require an experienced service engineer. If a maintenance task does require a service engineer, how long will it take them to complete the service? Most vendors charge an hourly rate for a service engineer but when travel time or an overnight stay is required, fully understand what you are paying for. Some companies might even charge for mileage between the service engineer's base and your laboratory. If you work in a commercial laboratory and cannot afford the instrument to be down for any length of time, find out what it is going to cost for 24–7 service coverage.
You can take a chance and just pay for each service visit, or you might want to budget for an annual preventative maintenance contract, where the service engineer checks out all the important instrumental components and systems frequently to make sure they are all working correctly. This might not be as critical if the instrument is in an academic or research environment, but it is absolutely critical if it is being used in a commercial or contract laboratory to generate revenue. Also find out what is included in the service contract, because some cover the cost of consumables or replacement parts in addition to service visits, whereas others just cover the service visits. These annual preventative maintenance contracts typically make up about 10–15% of the cost of the instrument, but they are well worth it if you do not have the expertise in-house or if you just feel more comfortable having an insurance policy to cover instrument breakdowns.
Once again, talking to existing users will give you a very good perspective of the quality of the instrument and the service support offered by a manufacturer. There is no absolute guarantee that the instrument of choice is going to perform to your satisfaction 100% of the time, but if you work in a high-throughput, routine laboratory, make sure it will be down for the minimum amount of time. Fully understand what it is going to cost you to maximize the uptime of all the instruments being evaluated.
As mentioned earlier, it was not my intention to compare instrument designs and features in this article, but to give you some general guidelines as to what are the most important factors to take into consideration. Besides being a framework for your evaluation process, these guidelines should be used in conjunction with the other materials and references mentioned in this article.
However, if you want to find the best instrument for your application needs, be prepared to spend a few months evaluating the marketplace. Don't forget to prioritize your objectives and give each of them a weighting factor based on their degree of importance for the types of samples you analyze. Be careful to take the evaluation in the direction you want to go and not where the vendor wants to take it. In other words, it is important to compare apples with apples and not to be talked into comparing an apple with an orange that looks like an apple. However, be prepared that there might not be a clear-cut winner at the end of the evaluation. So talk to as many users in your field as you possibly can — not only ones given to you as references by the vendor, but ones chosen by yourself. This approach will give you a very good indication of the real-world capabilities of the instrument, which can often be overlooked at a demonstration.
Never forget that it is a very competitive marketplace, and that your business is extremely important to all of the ICP-MS vendors. You may be pulled in many different directions by your own colleagues, external people who are advising you, and the vendors who want to influence your choice. Good luck with your evaluation!
(1) R. Thomas, Practical Guide to ICP-MS: A Tutorial for Beginners, Third Edition (CRC Press, Boca Raton, Florida, 2013).
(2) K. Nottingham, Anal. Chem., 35–38A (January 1, 2004). http://pubs.acs.org/subscribe/journals/ancham-a/76/i01/toc/toc_i01.html.
(3) Royal Society of Chemistry, Report by the Analytical Methods Committee, The Analyst122, 393–408 (1997).
(4) Inductively Coupled Plasma Mass Spectrometry: An Introduction to ICP Spectrometries for Elemental Analysis, A. Montasser, Ed. (Wiley-VCH, Weinheim, Germany, 1998), pp. 16–28, Chapter 1.4.
(5) "How to Select an ICP Mass Spectrometer: The Most Important Analytical Consideration," Pittcon, 2014 Short Course. http://ca.pittcon.org/Technical+Program/tpabstra14.nsf/SCoursesByCat/FE8395AD9A9CEBF385257B94004ECC03?opendocument.
(6) PlasmaChem Listserver: A discussion group for plasma spectrochemists worldwide, Syracuse University, New York. http://www.lsoft.com/scripts/wl.exe?SL1=PLASMACHEM-L&H=LISTSERV.SYR.EDU.
Robert Thomas is a consultant and science writer specializing in trace element analysis. This installment was printed with permission and adapted from reference 1.