Frequently Asked Questions
Arconic Spectrochemical Reference Materials has been making and using CRMs for many years and have answered many of the questions you may have concerning the production, certification and use of our CRMs. Below you will find the answers to many of the most frequently asked questions about Arconic CRMs.
Q
How does Arconic select what CRMs and compositions to offer as catalog items?
A
While Arconic offers one of the widest ranges of CRMs in the industry, it is impossible to produce CRMs for every existing alloy and alloy variation. We attempt to choose the alloys that are most widely used within the aluminum industry and then supplement them with range CRMs in an attempt to cover compositional variations. Our "SS" CRMs are generally near the nominal registered Aluminum Association composition for that alloy, with some modifications for elements not covered in the registration. For those who need highly specific CRMs, we offer Arconic Specialty CRMs.
Q
Can I use a CRM that doesn’t correspond exactly to my alloy?
A
The most desirable situation is to have CRMs that match your particular composition as closely as possible. Recognizing that this may not be possible, CRMs that are close in composition or a series of CRMs that bracket the composition of interest can be used and can provide a high degree of analytical accuracy. Contact us if you have any questions regarding suitability of a CRM got your application.
Q
What are the differences between "SS", "WA" and "KA" CRMs? How are they best used?
A
The "SS" CRM is the nominal alloy composition and is generally close to the midpoint in Aluminum Association registered composition. The "WA" and "KA" CRMs are range CRMs and provide a means to vary composition within the same alloy type. They are also useful in analyzing related alloy compositions. The "Wx" and "Kx" designations are for wrought and casting alloys respectively.
Q
How are your CRMs produced and certified?
A
Arconic CRMs are produced from a continuously cast ingot with length and diameter optimized to assure compositional homogeneity from edge to center and along the length. Certification follows a set of statistically intensive procedures that check for both homogeneity and concentration. Compositional accuracy is assured through the use of NIST reference CRMs (when available), Arconic master composition CRMs and most importantly the use of at least two independent analytical techniques.
Q
Are Arconic CRMs traceable to NIST?
A
Yes. NIST Standard Reference Materials® (SRM) are used as part of the certification process whenever possible. In many cases there are no NIST SRMs that correspond to the alloys we make. In those cases, the reference analytical methods, such as ICP-AES, will always use NIST traceable solution CRMs. Traceability is further established through the use of balances that are calibrated and verified using NIST traceable weights. Note: Arconic provides material analysis for NIST SRMs as one part of NIST’s certification process.
Q
What do the brackets around certain compositions mean?
A
The brackets, e.g. (0.52), mean that this is an uncertified composition and is given for reference only. A reference value is established with at least one analytical technique, however, analysis using a second or third independent verification was not performed. In the current Arconic catalog, certified values are provided for almost every element in every CRM. The exceptions to this are the Arconic "SQ" RMs and some trace level elements.
Q
What does a certified value of 0.0000% with an estimated uncertainty of 0.0002% mean on a certificate of analysis?
A
When a certifed value of 0.0000% or 0.00000% is assigned to an element, we are certifying that the amount of that given element rounds to zero in the last given decimal place. An estimated uncertainty of 0.0002% on a certified value of 0.0000% signifies that we are 95% confident that the amount of that given element falls within 0.0002% of 0.0000% - we understand that it is not physcally possible to have a negative amount of an element present in an alloy; however, it is possible to measure a negative value via an established calibration curve. An estimated uncertainty of 0.0002% on 0.0000% gives the user a reasonable range for type correction or standardization when using the CRM and helps the user determine if the correction is adequate for their given application. For example, a reading of -0.0001% is much more reasonable when correcting to 0.0000% than a reading of -0.02% which may signify a bias or incorrect correction for the specific element of interest.
Q
What are your "SQ" RMs for? Why isn’t the composition certified?
A
In most spark optical emission systems the analytical response curves are established only once and will remain the same throughout the life of the instrument. This differs from many other quantitative analytical systems which are standardized prior to each use. Even though the shape of the analytical curves is fixed, no OES system is perfectly stable and the response curves will "drift" over time. In order to monitor and correct this drift, Arconic provides “SQ” RMs for drift correction. They are not certified because they are used simply to correct instrument drift based on intensities. SQ RMs do not correspond to any particular alloy and therefore are generally not suited as CRMs for compositional determination. They are used at a frequency that depends on the drift behavior of the Spark-AES system. For current generation instruments, this is usually performed at some frequency between once per day (24 hours) and once per week.
Q
What does the uncertainty mean? Why do some values come with no uncertainty?
A
The uncertainty reflects the errors associated with the certification of the composition for a particular element. Statistically, it contains the errors due to inhomogenity, instrumental precision and bias between analytical techniques used in the certification processes. In most cases the predominate error is the between technique bias. The given limits of uncertainty represent this combined uncertainty and seek to estimate, with a 95% confidence level, a range in which the true value may be expected to lie. In some cases no uncertainty is given because of limited data or the lack of a second independent measure. For those interested in more detail see the ISO document "Guide to the Expression of Uncertainty in Measurement" and NIST Technical Note 1297 "Guidelines for Evaluating and Expressing the Uncertainty of NIST Measurement Results."
Q
How do I use the certified value and uncertainty to judge how well my instrument is performing?
A
Since there are always errors associated with both the certification process and the actual analyses done on your instrument, it is unlikely that you will consistently reproduce the certified value exactly. Certainly, the uncertainty interval of your analysis should overlap to some degree with the given uncertainty but you must make a judgment on whether that is sufficient criteria for acceptance. For instance, an analysis of Si may yield 8.65% against a certified value 8.55% with an uncertainty of 0.09%. If a control chart showed that this was a random value and that on average (day-to-day) your analyses were consistent with the certified value you would judge this to be acceptable. On the other hand if you consistently showed an average of 8.65% you may be seeing a bias in your analyses and may want to take action to correct the response.
Q
What is a matrix effect and how do I minimize or avoid this effect?
A
There is some confusion when people speak of matrix effects. In most cases in Spark-AES we are simply speaking of the fact that different alloys can produce different responses for the same elemental concentration. This effect has less to do with instrumentation and more to do with the micro and macro grain structure and the formation of second phases associated with different alloys. These are what give alloys properties such as corrosion resistance, strength, fatigue resistance, etc. The best way to avoid the undesirable analytical effects of matrix is to use CRMs whose compositions are similar to that of the alloy being analyzed.
Q
I have a piece of rolled or forged aluminum. Can I spark it directly and get the right composition?
A
Maybe - this brings us back to matrix effects. Processes such as rolling, forging, heat treating can significantly affect the grain structure and the formation of second phases. Even large scale castings can show significant macro segregation. Sparking directly on a fabricated piece of aluminum may subject the analyses to errors due to these metallurgical effects. That’s why we always recommend chilled cast samples taken directly from the molten metal. These samples will have structures most similar to CRMs. Alloy identification can be done on fabricated pieces but we would recommend that certified composition be established using chilled cast samples. Follow the procedures outlined in ASTM B985 to obtain a representative sample from any product.
Q
When I analyze your CRM I do not get the same number you have on the certification sheet. Am I wrong or is the certification incorrect?
A
This depends on the magnitude of the difference- see the discussion above on uncertainty. When a customer comes to us with this problem, we most often find that there’s either a problem with instrumentation, calibration curves or a matrix effect that has not been considered. Occasionally, we are wrong. We certify thousands of CRMs and tens of thousands of elements each year and rarely have a problem because of the numerous verification steps we have in place. If you have a problem with a CRM, we would be happy to help you resolve the issue whether it is within your laboratory or with our CRM.
Q
What’s the difference between a Certified Reference Material (CRM), a Reference Material (RM) and a Standard Reference Material® (SRM) ?
A
ISO-REMCO defines a Certified Reference Material as a Reference material characterized by a metrologically valid procedure for one or more specified properties, accompanied by a certificate that provides the value of the specified property, its associated uncertainty, and a statement of metrological traceability. A Reference Material is defined as a material, sufficiently homogeneous and stable with respect to one or more specified properties, which has been established to be fit for its intended use in a measurement process. An SRM is a Reference Material produced by NIST. Arconic’s reference materials are CRM’s except for the “SQ’s” which are RMs. We are accredited by ANAB to ISO Guide 34 as a Reference Material Producer.
Q
Why do you have individual element CRMs?
A
They provide a convenient means of establishing response (calibration) curves over a fairly wide range of concentrations. Remember if you choose to use them for day-to-day analytical work, you need to determine that there are no matrix effects since individual element CRMs are most often not in a particular alloy matrix. There are some exceptions.
Q
Why are some elements, such as sodium, not generally found in your alloy CRMs but only as individual element CRMs?
A
Some elements are simply difficult to alloy into aluminum and can cause undesirable effects in the CRMs such as segregation through an ingot or formation of inclusions. Sodium for instance has a tendency to burn off during casting and interfere with silicon grain refining or modification when using strontium or phosphorus. Boron on the other hand, readily forms borides with elements such as titanium. This might not be an issue in general production but can cause significant problems in CRMs where high homogeneity and consistency are mandatory. Individual element CRMs allow us to tailor our casting methodology specifically to avoid these problems.
Q
My Spark-AES supplier says I have universal calibration curves. What does this mean? Do I still need CRMs?
A
Your supplier generated calibration curves for each element using a large number of different alloy CRMs. This gives a generic calibration curve that provides reasonably accurate analyses but will generally not give the same accuracy as those generated for a specific alloy. CRMs for your particular alloy are therefore essential. The supplier’s software will allow you to perform a “type standardization” that helps tie the generic calibration curve to a specific alloy. If done correctly, this will greatly increase the accuracy of the analyses.
Q
I am using your CRMs and have good precision and accuracy but my customer says I am still out of spec. What should I do?
A
First assure yourself that you have taken into consideration all possible sources of error. Is the lab analyzing good quality samples being taken and prepared according to industry standards such as ASTM E716? Are you analyzing chilled cast samples while your customer is looking at a fabricated piece - see above? Are you using the right CRM? Have you verified your analyses using an independent method? If you feel comfortable with these answers, one of the best ways to handle such a problem is to independently verify accuracy between laboratories by analyzing the same Arconic CRM.