Standards for the 21st century; establishing trust in measurement

Instrument manufacturing

How has instrument manufacturing changed in the period under question? Essentially, it can be divided up in to three discrete phases, which not surprisingly get shorter, the nearer one gets chronologically to the present day, as the pace of change accelerates. Also at each phase boundary, there is a noticeable change in the physical characteristics of the instruments, usually associated with a reduction in size.

Phase 1: 1940–1975

Instruments manufactured for their optical quality, with minimal electronics. Customers initially relied on the quality of manufacture, to give “peace of mind”, but very soon like-minded individuals established the Photoelectric Spectrometry Group (1948) in the UK to act as a forum where mutual problems could be discussed. Manufacturers relied on the skills of their workforce to precision build the instruments. To date many of the instruments produced in this era still match, or even exceed, the quality of the optics being used in modern instruments. At the end of the era, welcome to the age of the microprocessor, and the age of regulatory control, see below.

Phase 2: 1975–1995

Incorporation of one (or two) microprocessor-based electronic controls, and use of internal control materials, e.g. neutral density glasses. At the start, these filters were simply measured, and then subsequent values measured against these stored values for comparison. By the end of the phase, regulatory pressures require these filters to have traceability (established by an unbroken chain of comparisons) to NMI certified materials, with acceptance criteria built on expanded uncertainty budgets for a given system. Acceleration in the processing power of microprocessor-based systems brought about the birth of the IBM 5150 PC in 1981, and evolution to the next logical step.

Phase 3: 1995 to current

There has been a shift of control and data processing to external computers. Spectrometers are now reduced simply to an “optical bench” with necessary support electronics to provide an electronic data output. Availability of fibre optic technology, semiconductor lasers etc. produced primarily for the telecommunications industry allows the development and production of even smaller semiconductor array-based devices, and sample presentation accessories. Couple this with the ability of systems to process large amounts of data, and you have spectroscopy being taken routinely into areas of the spectrum, previously either unexplored, e.g. NIR, or techniques considered confined to the research lab, e.g. FT-IR, Raman etc.

Dimensionally, the trend has obviously been to progress to smaller and smaller units. However, one could ask is this progress or is there a point at which “small is no longer beautiful”? To quote a certain fictional engineer “... I (we) can’t disobey the laws of physics”.

For applications where high performance is required, you still need good energy throughput, i.e. don’t expect the performance of a unit with an A5 footprint (or smaller) to match that of instrument sized at A0. In the modern environment, the risk analysis, specifically “... is the spectrometer capable of adequately performing the required measurement?” has replaced “... will it physically fit in the laboratory?”. As we see in the next section, this ability to take the spectrometer to the sample, and not the other way around again re-defines the Quality environment required.

The user base

As the instrumentation has changed over the years, so too has the spectroscopic user base. This may be described as a transition from specialist users with detailed knowledge of the inner workings of the instrument towards an application driven “Black box” mentality. As mentioned above, many modern instruments are application software control computers with spectroscopic peripherals. This transition from specialist to workhorse instrumentation is not all bad news of course! The ever increasing spectroscopic application metrology is a major asset to society assuming that we can rely upon the data quality produced.

In our view, one of the major issues of today is that the majority of the user base are too trusting of the numbers generated without appreciating the necessity for Good Spectroscopic Practice and the underlying assumptions and limitations of optical measurements. The problem is exacerbated by our institutions of higher education being either unable or unwilling to teach practical aspects of analytical spectroscopy. From a data integrity point of view the lack of a detailed knowledge of relevant quantum theory is unlikely to have the same impact as a fingerprint on the face of an cuvette or use of a solvent with insufficient transmittance. At the same time we are experiencing a spectroscopic diaspora whereby:

• spectroscopic measurements and testing is moving out of the laboratory into the warehouse, manufacturing at point of use and into the field;
• relatively untrained operators, especially from a spectroscopic view point, may have skill base issues in routine operation.

There remains a need to be assured of holistic operation rather than just modular performance using appropriate reference materials to confirm “fitness for purpose” and/or operability within defined limits.

One source of help, from a user instrumentation operation perspective, is the Instrumental Criteria Sub Committee of the Analytical Methods Committee under the auspices of the Royal Society of Chemistry3 who have worked for many years producing reports and technical briefs on analytical instrumentation which are free for download from their website.

International Standards and their regulatory impact4

As stated above, 1975 can be shown to be a key date, from which many of the international standards and the associated control bodies came into existence.

From 1975 to 1981

• 1975–ISO-REMCO formed
• 1976–USA FDA introduces GLP
• 1977–ILAC formed
• 1978–ISO/IEC Guide 25 issued (to become ISO/IEC 17025 in 1999)
• 1981–OECD GLP Guidelines produced

There then followed a period of relative quiet, until 1987 when ISO issued the 9000 series of standards, and the emphasis on quality was expanded to a much larger audience. In the early ’90s, following a major review with respect to the understanding of traceability in analytical chemistry, CITAC was formed, and the national accreditation bodies implementing accreditation to ISO Guide 25 organised into regional co-operations; namely APLAC in the Pacific Rim, and EA in Europe. The use of accreditation to ISO Guide 25 continued to expand slowly and gain acceptance as more and more laboratories used this route to assist the acceptance of the data being generated, but it was the conversion of ISO guide 25 to ISO standard 17025 in 1999 that really accelerated the process.

Throughout this period, another major standards organisation, ASTM, reflected this increased emphasis on Quality by publishing a range of guidance standards, shown in Table 1. In many areas these standards provide essential guidance in areas not otherwise covered. UV-visible instrument manufacturers use for example E387, as the “de facto” standard for the estimation of instrumental stray light performance. It is interesting to observe the changes brought about by the five-year review cycle, on these standards. In January 2002, ASTM reflected the global nature of its standards work, by changing its name to ASTM International.

Table 1. ASTM international standards for spectroscopy.

Abbreviations

APLAC: Asia Pacific Laboratory Accreditation Cooperation; CITAC: Cooperation on International Traceability in Analytical Chemistry; EA: European Cooperation for Accreditation; FDA: US Food & Drug Administration; GLP: Good Laboratory Practice; ILAC: International Laboratory Accreditation Cooperation; OECD: Organisation for Economic Cooperation & Development.

Summary

It is a basic tenet for all analytical scientists that, before the commencement of an analytical procedure, they must ensure the suitability and proper operation of any instrument or system that is part of the measurement process ... in order to demonstrate “fitness for purpose”.5 This challenge is becoming increasingly difficult as spectroscopic applications increase.

This article has hopefully provided an overview and introduction to the topic. In future columns we intend to cover specific application areas in more detail.

References

1. NBS Special publication 260–102, Holmium oxide solution wavelength standard from 240 to 640 nm—SRM2034 (1986)
2. J.C. Travis et al., J. Phys. Chem. Ref. Data 34(1), 41–56 (2005). https://doi.org/10.1063/1.1835331
3. http://www.rsc.org/Membership/Networking/InterestGroups/Analytical/AMC/AMCReports.asp
4. For more detail, see P.J. Jenks, “REACH....... into the future? Possible consequences of the EU REACH Directive on the analytical laboratory”, Spectrosc. Europe 20(4), 22 (2008). https://www.spectroscopyeurope.com/quality/reach-future-possible-consequences-eu-reach-directive-analytical-laboratory
5. Using analytical instruments and systems in a regulated environment; http://www.rsc.org/images/brief28_tcm18–99611.pdf