Robert Thomas has more than 30 years of experience in trace element analysis. He is the principal of his own freelance writing and scientific consulting company, Scientific Solutions, based in Gaithersburg, MD.

Robert Thomas

Recent regulations on heavy metal testing have required the pharmaceutical industry to monitor a suite of elemental impurities in pharmaceutical raw materials, drug products, and dietary supplements. These directives are described in the new 

) Chapters <232>, <2232>, and <233>, together with 

The International Conference for Harmonization

) Q3D, Step 4 guideline, which suggests the use of plasma spectrochemistry to measure 24 elemental permitted daily exposure (PDE) levels in the various drug delivery categories. This month's column, adapted from a chapter in Robert Thomas's new book, 

Measuring Elemental Impurities in Pharmaceuticals: A Practical Guide,

 offers guidance on selecting the optimum atomic spectroscopic technique for orally delivered drugs, based on the J-values described in the validation protocols outlined in 

	Comparing an instrumental technique based on its performance specifications with simple standards is important, but it bears little relevance to how that instrument is going to be used in a real-world situation. For example, instrument detection limits (IDL) are important to know, but how are they impacted by the matrix and sample preparation procedure? What are the real-world method detection limits (MDL), and is the technique also capable of quantifying the maximum concentration values expected for this analysis in the same sample run? And what kind of precision and accuracy can be expected if working close to the limit of quantitation (LOQ) for the overall methodology being used? Additionally, on the sample throughput side, how many samples are expected and at what frequency will they be coming into the laboratory? How much time can be spent on sample preparation and how quickly must they be analyzed? Sometimes the level of interferences from the matrix will have a major impact on the selection process or even the amount and volume of sample available for analysis (1,2).

	Understanding the demands of an application is therefore of critical importance when a technique is being evaluated, and particularly if there is a minimum amount of expertise or experience in house on how best to use it to solve a particular application problem. This could be the likely scenario in a pharmaceutical or dietary supplement production laboratory that is being asked to check the elemental impurities of incoming raw materials used in the manufacturing process. They will now have to follow the new 

) Chapters <232> (or Chapter <2232> for dietary supplements) and <233> (3), which recommend the use of a plasma-based spectroscopic technique to carry out the analysis-or any other atomic spectroscopy technique as long as the validation protocols are met. Note: The 

International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals

) has it's own directives, which are described in its Q3D Step 4 guidelines (4).

	The expertise of the operator should never be underestimated, because if inductively coupled plasma–mass spectrometry (ICP-MS) is being seriously considered, it generally requires an analyst with a higher skill level to develop rugged methodology free of interferences that can eventually be put in the hands of an inexperienced user to operate on a routine basis. Again, this is a real concern if the technique is being used by novice users who have limited expertise in running analytical instrumentation, which may be the case in the pharmaceutical or nutraceutical industries. So, let's take a closer look at the new 

 chapters to understand what atomic spectroscopy techniques might best meet the demands of this application.

	To get a better understanding of the suitability of the technique being used and whether its detection capability is appropriate for pharmaceutical materials, it's important to know the permitted daily exposure (PDE) limit for each target element. In particular, it's important to understand what the 

, as described in Chapter <233>, which is defined as the PDE concentration of the element of interest, appropriately diluted to the working range of the instrument, after the sample preparation procedure to get the sample into solution is completed. Let's use Pb as an example. The PDE limit for Pb in an oral medication defined in Chapter <232> is 5 µg/day.

	Based on a suggested dosage of 10 g of the drug product/day, that's equivalent to 0.5 µg/g Pb. If 1.0 g of sample is digested or dissolved and made up to 500 mL, that's a 500-fold dilution, which is equivalent to 1.0 µg/L. So, the J-value for Pb in this example is equal to 1.0 µg/L.

	The method then suggests using a calibration made up of two standards: standard 1 = 1.5J, standard 2 = 0.5J. So for Pb, that's equivalent to 1.5 µg/L for standard 1 and 0.5 µg/L for standard 2.

	The suitability of a technique is then determined by measuring the calibration drift and comparing results for standard 1 before and after the analysis of all the sample solutions under test. This calibration drift should be <20% for each target element. However, once the suitability of the technique has been determined, further validation protocols described in detail in Chapter <233>, must be carried out to show compliance to the regulatory agency if required.

	It should also be pointed out that no specific instrumental parameters are suggested in Chapter <233>, but only to analyze according to the manufacturer's suggested conditions and to calculate and report results based on the original sample size. However, it does say that appropriate measures must be taken to correct for interferences, such as matrix-induced wavelength overlaps in inductively coupled plasma–optical emission spectrometry (ICP-OES) and argon-based polyatomic interference with ICP-MS.

	Let's examine this approach by taking an example of measuring 24 elemental impurities in an oral drug according to Chapter <232> and calculating the J-values for each elemental impurity and comparing them with the limits of quantitation (LOQ) for each atomic spectroscopy technique (such as atomic absorption, axial-ICP-OES, and ICP-MS) to give us an assessment of their suitability. For this analytical scenario, we'll take the LOQ for the technique as 10x the IDL. These LOQs were calculated by taking the average of published IDLs from three instrument vendors' application material and multiplying them by 10 to get an approximation of LOQ. In practice, a method limit of quantitation is typically determined by processing the matrix blank through the entire sample preparation procedure and taking 10 replicate measurements. The method LOQ, sometimes referred to as the method detection limit, is then calculated as 3–7x standard deviations of these 10 measurements, depending on the percent confidence level required.

	To make this comparison valid, the sample weight was adjusted for each technique, based on the detection limit and analytical working range. For atomic absorption and ICP-OES, we used a sample dilution of 2 g/100 mL, whereas for ICP-MS we used 0.2 g/100 mL. Atomic absorption and ICP-OES could definitely use larger sample weights, but for high-throughput routine analysis, we are probably at the optimum dilution for ICP-MS. (Note: For this assessment, it was felt that axial ICP-OES was the better choice over the radial configuration, because of its superior detection capability.)

	Tables I–III show the comparison of atomic absorption (flame and furnace), axial ICP-OES, and ICP-MS, respectively, for 24 elemental impurities in an oral drug with a maximum dosage of 10 g/day, according to Chapter <232>. The important data to consider is in the final column, labeled "Factor Improvement," which is the J-value, divided by the LOQ. Generally speaking, the higher this number the more suitable the technique.

	The "Class" column in these tables indicates the level of toxicity of the elements, which have been determined based on chronic exposure data and likelihood of occurrence in the drug product. (It is generally recognized that Class 1 and 2A elements are the most important to monitor.) Also it should be noted that the arsenic and mercury PDEs are based on the inorganic forms of the element.

	It should be emphasized again that LOQ in these examples is just a guideline as to the real-world detection capability of the technique for this method. However, it does offer a very good approximation as to whether the technique is suitable based on the factor improvement number compared to the J-values for each elemental impurity. Clearly, if this improvement number is close to or less than 1, as it is with the majority of elements by flame atomic absorption (FAA), the technique is just not going to be suitable, particularly for the "big four" heavy metals, which are the most critical. On the other hand, electrothermal atomization (ETA) would be suitable for the majority of the impurities (including the heavy metals), except for a few of the catalyst-based elements, which are not ideally suited to the technique. However, the ETA technique is very time-consuming and labor intensive, so it probably wouldn't be a practical solution in a high-throughput pharmaceutical production laboratory. The comparison between FAA and ETA is exemplified in Table I.

	Table II shows that axial ICP-OES offers some possibilities for monitoring oral drugs because the vast majority of the improvement factors are higher than 1. These numbers could be further improved, especially for the heavy metals, by using a much higher sample weight in the sample preparation procedure without compromising the method. (Note: Because most commercial ICP-OES instrumentation offer both axial and radial capability [dual-view], it was felt that the axial performance was most appropriate for this comparison.)

	However, it can be seen in Table III that ICP-MS shows significant improvement factors for all impurities, which are not offered by any other technique. Even for the four heavy metals, there appears to be ample improvement to monitor them with good accuracy and precision. The added benefit of using ICP-MS is that it would also be suitable for the other methods of pharmaceutical delivery, such as parenteral or inhalation, where the PDE levels are typically one or two orders of magnitude lower. Additionally, if arsenic or mercury levels were found to be higher than the PDE levels, it would be relatively straightforward to couple ICP-MS with high performance liquid chromatography (HPLC) to monitor the speciated forms of these elements if required.

	It is important to understand that there are many factors to consider when selecting a trace element technique that is most suited to the demands of your application. Sometimes, one technique stands out as being the clear choice, whereas other times it is not quite so obvious. And, as is true with many applications, more than one technique is often suitable. However, the current methodology described in the new 

 Chapters <232> and <233> presents unique challenges, not only from a perspective of performance capability, but also because of the validation protocols that have to be met to show suitability of the technique to the analytical procedure being used. From a practical standpoint, there is no question that to meet all the PDE limits in all pharmaceutical delivery methods, particularly for parenteral and inhalation drugs where the PDEs are significantly lower, ICP-MS is probably the most appropriate technique. However, for oral delivery products, especially liquid medications or those that can be easily brought into solution with a suitable aqueous or organic solvent, axial ICP-OES could offer a more cost-effective approach. In addition, ICP-OES can use larger sample weights and lower dilutions, which will improve its detection capability. In addition, if the sample workload requirements are not so demanding, ETA could provide a solution for monitoring a small number of samples for the critical Class 1 and 2A elements. However, I think it's fair to say that ICP-MS has shown it has the detection limits and throughput capability to be the optimum technique of choice for carrying out the measurement of elemental impurities in a wide and diverse range of pharmaceutical materials.

	Note: Elemental contaminants in dietary supplements were not evaluated for this article, but a full explanations of the J-value calculations compared to the limit of quantitation for the four elemental PDEs specified in Chapter <2232> are given in a chapter of the new book cited in the abstract (5).

https://www.labmate-online.com/article/mass-spectrometry-and-spectroscopy/41/scientific-solutions/money-to-burn-do-you-know-what-is-costs-to-run-your-atomic-spectroscopy-instrumentation/2030

	(3) Elemental Impurities in Pharmaceuticals: Updates: USP Website: 

http://www.usp.org/chemical-medicines/elemental-impurities-updates

	(4) International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use, 

http://www.ich.org/products/guidelines/quality/article/quality-guidelines.html

Measuring Elemental Impurities in Pharmaceuticals: A Practical Guide

 (CRC Press, Boca Raton, Florida, 2018), ISBN 13:978-1-138-19796-1.

 is principal of Scientific Solutions, a consulting company that serves the application and writing needs of the trace element user community. He has worked in the field of atomic and mass spectroscopy for more than 40 years and has written over 80 technical publications including a 15-part tutorial series on ICP-MS. He recently completed his fourth textbook entitled 

. He has an advanced degree in analytical chemistry from the University of Wales, UK, and is also a Fellow of the Royal Society of Chemistry (FRSC) and a Chartered Chemist (CChem). He has served on the ACS Committee on Analytical Reagents for the past 17 years.

Articles

How to select the optimum atomic spectroscopic technique for orally delivered drugs, based on the J-values described in the validation protocols outlined in USP Chapter 

Choosing the Right Atomic Spectroscopic Technique for Measuring Elemental Impurities in Pharmaceuticals: A J-Value Perspective

This installment describes the development of two novel X-ray diffraction (XRD) techniques that enable the rapid analysis of samples using handheld instrumentation for remote applications. Both techniques can be applied to unprepared samples in the field, which is a highly favorable characteristic in many applications since the time required for laboratory-based sample preparation is avoided. The first development is based on back-reflection energy-dispersive XRD, which offers the benefit of being insensitive to sample morphology and therefore opens the possibility of analyzing some sample matrices with no sample preparation at all. The second area of research focuses on using the technology to enhance the XRD signal from a specific mineral or phase within a mixture specifically for metallurgical applications. 

	This column installment describes the development of two novel X-ray diffraction (XRD) techniques that show exciting potential in handheld and portable applications. Although the ways in which these two techniques work are distinct, they are linked through their applicability in compact instrument formats with similar application areas in mining and metallurgy. The X-ray technologies required for implementation of both techniques are low cost and are essentially the same as those used in existing handheld X-ray fluorescence (XRF) instrumentation, representing a relatively low-risk path to the development of a commercial product.
	
	


	
	The first of these methods is based on the application of energy-dispersive XRD (EDXRD) in a back-reflection geometry (1,2). This combination yields a property that is essentially unique in the field of XRD--almost complete insensitivity to sample morphology. An idealized configuration for this technique is illustrated in Figure 1, in which the X-ray beam from a tube source passes through an annular detector and is incident on a sample.

	This configuration maximizes the detected diffraction signal from the sample while conforming to the angular constraints of the back-reflection method. The insensitivity to sample morphology arises because, for an X-ray photon diffracted exactly through 180° for example, the distance of the interaction point on the sample relative to the instrument becomes immaterial. It can readily be shown that, in practice, constraining the scattering angle to 2θ > ~160° retains this valuable property (1).

	A lack of sensitivity to the shape of the sample opens up the opportunity to analyze samples with no preparation at all, and this is the primary advantage of the back-reflection technique. In addition to the convenience of avoiding the crushing and grinding of samples, and the equipment needed to accomplish this procedure, a considerable time-saving is achieved. In its most basic form, back-reflection EDXRD can be implemented using the same X-ray technologies found in handheld XRF instrumentation, a mature technology with global annual sales of several thousand units into a wide range of end-user applications (3). Furthermore, EDXRD tends to be much faster than the angle-dispersive XRD methods used in the majority of laboratory-based instruments, and the back-reflection geometry is ideal for handheld-based applications. All of these factors come together to favor the commercialization of a handheld device, which is a relatively minor evolution of an existing, highly successful instrument (4).

	The primary application envisaged for a handheld XRD instrument of this type is in mining and related areas, with iron ore mining likely to be the single largest commercial application. A handheld XRD device based on back-reflection EDXRD will not be able to compete with laboratory diffractometers in terms of data quality and accuracy, but nevertheless there are major opportunities to add value with rapid XRD measurements in the field. For example, identification of the dominant iron oxide species in an iron ore sample, such as hematite (Fe

), will allow improved XRF accuracy of the total iron content of the sample since the form of the oxide will no longer need to be assumed and the consequent potential errors can be avoided. In addition, the improved accuracy will feed directly through to the XRF quantification of all other elements in the sample. A working prototype instrument has recently been developed by modification of a handheld XRF device. An image of the prototype is shown in Figure 2 and it features a measurement head that has been revised to achieve the required back-reflection geometry.

	Figure 3 shows the recorded spectrum for a rock sample known to be predominantly hematite, together with a model fit of the diffraction spectrum of hematite.

	
	The model separately includes minor amounts of the elements that show fluorescence peaks (labeled on the plot) to achieve a good overall fit. It is important to note that the Rh X-ray tube source emits strong Rh-L lines approximately in the range 2.5-3.1 keV, which scatter from the sample as well as part of the measurement head, and give a signal that dominates the spectrum in this energy range. By recording a spectrum with no sample, the background has been subtracted but the peak intensities in this range are not reliable and have been excluded from the fitting process. The strong off-scale peak in Figure 3 is because of the coincidence of a Rh-Lβ line with a hematite diffraction peak. Overall, the fit to the data is extremely good and demonstrates the potential of the back-reflection technique in handheld XRD analysis of samples of this type. For the results shown here, the measurement head was evacuated, but this would not necessarily be required for a purpose-designed instrument.

	Recent laboratory testing with the prototype instrument shows that, in many cases, secondary minerals in iron ore samples can be identified and approximately quantified. For active mining operations, such measurements will enable rapid delineation of ore boundaries and real-time assays at the mine face. The assessment of ore grades will allow more efficient management of blasting, excavation, and haulage operations. The efficiency of the on-site enrichment process can be monitored via the analysis of raw materials, concentrates, and tailings. All of these advantages serve to improve the overall mine productivity while reducing costs relative to sending samples off-site for laboratory analysis. It is anticipated that the same advantages would be realized for other metallic ore mining operations.


	Another very promising application area is the assessment of the quality of dolomitic limestones. Limestone consists mainly of calcite (CaCO

) and is commonly added to the blast furnace in the production of iron to combine with and remove the impurities in the iron ore. The blast furnace slag that results is used as a base material for the construction of roads and highways. However, the presence of magnesia (MgO) is a critical component in the slag because it lowers the viscosity and improves the efficiency in the removal of impurities from the iron. It also reduces degradation of the refractory lining of the blast furnace and thereby increases its useful lifetime, a significant financial benefit in the production of iron and steel. Dolomite (CaMg[CO

) in the limestone provides a source of magnesia when fired and significantly increases the value of limestone for this application. Unfortunately, Mg is a light element and as a result is difficult to quantify or even detect using handheld XRF. Conversely, XRD methods can quite easily distinguish calcite and dolomite as a consequence of their differing crystal structures.

	Test results with the handheld XRD prototype for this analysis are shown in Figure 4. It shows three unprepared rock samples corresponding to the samples from St. Louis, Hope Valley, and Mindelheimer Klettersteig (5). Based on this earlier analysis, it was known that the Hope Valley and Mindelheimer Klettersteig rock specimens are dominated by calcite and dolomite, respectively, while the St. Louis sample consists of mainly calcite with minor amounts of dolomite. The signal intensities in Figure 4 are significantly lower than in Figure 3 because the tube excitation voltage is considerably reduced, 4 kV compared to 7 kV for the iron ore application. The part of the spectrum affected by scattering of Rh-L lines is also proportionately greater, but nevertheless the results are consistent with the known compositions. In particular, a dolomite diffraction peak at ~3.5 keV is clearly visible in the St. Louis data but absent in the Hope Valley spectrum, illustrating the detection of dolomite in the former. These measurements could be significantly improved through the use of an anode material in the X-ray tube that has no emission lines in the 2-4 keV range.


	The back-reflection EDXRD technique is also likely to find application in the analysis of high value objects because of its completely nondestructive characteristic. The detailed scientific analysis of objects with cultural heritage significance is a growing field. Examples include archaeological artefacts such as pottery (including pigments and glazes), jewelry, any objects made from stone or rock, and art pieces such as paintings and sculptures. Studies of this type are usually carried out either to answer questions relating to origin or manufacture, giving insight into the material history of the objects, or to understand the stability and deterioration of materials to ensure proper conservation and develop new conservation methods. Instrument portability is a great advantage in this application area, enabling the instrument to be taken to the target object, and a handheld XRD device would certainly find application in museums and galleries.

	As previously mentioned, a handheld instrument based on the back-reflection EDXRD technique cannot compete with the performance of laboratory diffractometers. However, this drawback is a consequence of the achievable resolution of energy-resolving detectors which, in this context, is quite limited even for state-of-the-art detectors such as silicon-drift devices. Our studies have recently demonstrated the implementation of the back-reflection technique in a high-resolution configuration at the Diamond Light Source synchrotron at Harwell in the United Kingdom (6). The benefit of using a synchrotron is that it generates intense X-ray beams by accelerating a beam of electrons around a circular vacuum tube at approximately the speed of light. As the electrons circle the tube, they emit energy with wavelengths in the range of 0.1-100 nm. However, to produce X-rays with enough intensity to generate high analyte signals with low background noise, the circumference of the vacuum tubes are in the order of several hundred meters. These are such large, specialized facilities that only about 40 synchrotron light sources exist throughout the world. Naturally, this analysis method is restricted to objects that can be taken to the synchrotron. Although acquisition times would be considerably longer, a laboratory version of the high-resolution configuration is certainly possible and constitutes a direction of future research for our laboratory.

Mineral or Phase-Specific XRD Technology


	The second novel XRD technique under development by our laboratory enables the enhancement of the diffraction signal of a specific mineral or phase within a mixture. Mineral or phase-specific XRD is suited to applications in which the user is interested in one key mineral or phase, such as an important impurity in the sample. The presence or absence (within the limit of detection) of a specific phase can be determined with this technique.

	Phase specificity in the new XRD technique relies on finding coincidences in the ratios of crystal d-spacings and elemental characteristic X-ray energies. Such coincidences can be exploited so that the two crystal planes diffract through the same scattering angle at two different X-ray energies, as illustrated in Figure 5. An energy-resolving detector placed at the appropriate scattering angle will detect a significantly enhanced signal at these energies if the target mineral or phase is present in the sample.

	
	Mathematically, the basis of the technique can be explained as follows. The well-known Bragg law for diffraction of X-rays from crystal planes can be expressed in the energy domain as follows:
	
	

 is the crystal d-spacing, and θ is the scattering angle. For the coincidence-based XRD method to work, the following equation must be satisfied:
	
	


	where δ must be small (the coincidence criterion), 

 are characteristic energies emitted simultaneously by the X-ray tube, and 

 are d-spacings of the target mineral. Rearranging gives
	
	


	showing that the ratio of characteristic energies must closely match the inverse ratio of d-spacings.

	For any given target mineral or phase the anode material must be carefully chosen to meet the coincidence criterion. As an example, quartz can be selectively detected by exploiting the following coincidence: quartz d-spacings of 2.2366 and 2.1277 Å (ratio 1.0512) and Pd (palladium) characteristic emission lines Lα

 at 2.990 keV (ratio 1.0536), for which 2θ = 154.6° and δ = 0.3°. The specific detection of quartz in an unprepared mudstone (which is a type of fine grained sedimentary rock whose original constituents were clays or muds), known to contain quartz, is illustrated in Figure 6.

					Figure 6: (a) A mounted mudstone rock sample known to contain quartz. (b) The combined EDXRD and EDXRF spectra for the on- (red) and off-coincidence (black) measurements, illustrating the signal enhancement because of quartz near 3 keV. The more prominent XRF peaks are labeled.

	In general, for any given target phase several candidate characteristic X-ray energy and d-spacing ratios can be found. Consequently, one of the parameters that can be chosen is the scattering angle. A wide range of angles can, in principle, be considered, but there are several advantages to choosing a high scattering angle, for example 2θ > 150°. A key advantage is reduced sensitivity to sample morphology and relaxed constraints on precise positioning of the sample. For example, moving the sample 2 mm further away from the instrument will have a negligible effect on the measurement and sample morphology on a similar scale is also tolerable. This benefit works in much the same way as in back-reflection EDXRD, though the mineral-specific XRD technique does not have the extreme insensitivity engendered by the former. Fortuitously, it also turns out that a high scattering angle improves specificity and signal intensity, and reduces sensitivity to poor powder averaging. Just as for handheld XRD based on back-reflection EDXRD, a high scattering angle additionally favors a compact instrument design.
	
	


	An exciting potential application of this technology is the quantification of retained austenite in steels using a handheld or production line instrument. Austenite is a high-temperature allotrope of iron or iron alloy that allows higher levels of carbon to be dissolved in the iron at elevated temperature. As the steel cools, the crystal structure changes and other allotropes (phases) form, including ferrite and martensite. The amount of austenite retained in the microstructure is an important factor in determining some of the strength characteristics of the steel. Depending on the intended use of any given steel, the retained austenite may be detrimental to the mechanical properties (7). For example, the tool and die industry requires minimum amounts of retained austenite and higher levels of ferrite and martensite because maximum hardness and resistance to wear is usually crucial in these applications and austenite reduces the attainable hardness. Conversely, some retained austenite is beneficial in the bearing and gear industry because it suppresses crack propagation. During the manufacture of certain steels the amount of retained austenite must be tightly constrained through optimization of the heat treatment of the steel.

	Key to this control is accurate measurement of the retained austenite in the steel as it comes off the production line. Although austenite in steels has traditionally been carried out by metallographic equipment, more recently laboratory-based XRD instrumentation has been considered the most accurate method especially at low concentrations below ~15%. However, the main disadvantages of using XRD for this application is that it is relatively expensive, and also the samples have to be brought back to the laboratory for analysis. By enhancing the diffraction signal specifically from the austenite phase in a steel sample, phase-specific XRD can achieve relatively high sensitivity to low concentrations of this critical phase. It is easy to envisage a handheld “retained austenite sorter” that could be used to quickly accept or reject finished parts based on a threshold amount of retained austenite, analogous to the use of handheld XRF devices in scrap metal sorting. A handheld instrument would cost a fraction of the price of laboratory diffractometers and could prove very desirable for this type of application.


	Another possible application of the mineral-specific XRD technique is the detection and quantification of crystalline silica (quartz) in mine dust, which has been collected on respiration filters of mine workers. Crystalline silica is responsible for silicosis (8), so there are strict protocols in place for the measurement and control of exposure of mine workers to airborne crystalline silica. Currently there are over 2000 service laboratories worldwide that are carrying out this type of analysis. This application is an excellent example of the required detection of a specific mineral within a potentially complex mixture. As in the retained austenite application, XRD methods are considered to be the most accurate and also the most widely applicable to all types of mine dust, compared to other competing methods such as infrared spectroscopy.

	The specific detection and quantification of quartz also has potential in other market segments. For example, quartz is a common impurity in iron ores and, because it is a very hard mineral, its presence influences decisions on the ore processing methods. Rapid quantification of the quartz in these ores, perhaps alongside an analysis of the iron mineral content by a back-reflection XRD instrument, would allow rapid decision-making at the mine site on blending different grades before initial processing of the iron ore.


	This column installment reports significant progress in handheld XRD technology, with the recent development of two enabling techniques. The first technique has been developed as a handheld mineral analyzer, which is intended for the rapid mineral identification and quantification of ore-based mining samples in the field through a combination of X-ray diffraction and X-ray fluorescence. This technique will be capable of directly analyzing the mineralogical components of a sample within a few minutes with no sample preparation--traditionally beyond the capabilities of conventional XRD equipment.

	The second development is based on a phase-specific XRD technique, which enhances the diffraction signal of a specific mineral or phase within a mixture and is suited to applications in which the user is especially interested in one specific component or impurity in the bulk sample. Both techniques offer the capability of analyzing samples in the field, without having to bring them back to the laboratory for analysis.


	Graeme Hansford would like to thank Alexander Seyfarth and Terry Smith of Bruker Elemental for their support of these projects and the provision of equipment, the UK’s Science and Technology Facilities Council for funding support.

 102-106 (2013).
	(3) A. Seyfarth and B. Kaiser, “History of Handheld XRF,” FIRST Newsletter, Issue 23 (2013), Bruker Elemental.
	(4) University of Leicester Press Release: 

http://www2.le.ac.uk/offices/press/press-releases/2014/january/innovative-handheld-mineral-analyser-2013-2018the-first-of-its-kind2019

.
	(5) G.M. Hansford, S.M.R. Turner, D. Staab, and D. Vernon,

 1708-1715 (2014).
	(6) G.M. Hansford, S. Turner, N. Karim, and J. Hall, “Completely Non-Destructive High-Resolution XRD Analysis of Cultural Heritage Objects,” presented at the 64th Denver X-ray Conference, Westminster, Colorado, 2015.
	(7) D.H. Herring, “A Discussion of Retained Austenite,” IndustrialHeating.com, March 2015. Available at 

http://www.heat-treat-doctor.com/documents/RetainedAustenite.pdf

 has more than 40 years of experience in trace-element analysis. He is the principal of his own freelance writing and scientific consulting company, Scientific Solutions, based in Gaithersburg, Maryland.

 received his PhD in high-resolution infrared spectroscopy at the University of Cambridge in the UK and worked for 10 years developing balloon- and aircraft-born instruments for atmospheric measurements. In 2006, he moved to the Space Research Centre at the University of Leicester’s Department of Physics and Astronomy, to develop X-ray instrumentation for space missions. Since 2010, his work has focused on novel XRD techniques, instrumentation, and applications. Please direct correspondence to: 

This installment describes the development of two novel X-ray diffraction (XRD) techniques that enable the rapid analysis of samples using handheld instrumentation for remote applications. Both techniques can be applied to unprepared samples in the field, which is a highly favorable characteristic in many applications since the time required for laboratory-based sample preparation is avoided.

Handheld X-ray Diffraction for Remote, Field-Based Applications

This installment evaluates the application requirements for the determination of 15 elemental impurities in pharmaceutical materials as described in the new 

) Chapters <232> and <233> and offers suggestions about which atomic spectroscopic technique might be the most suitable to use.

	Since the introduction of the first commercially available atomic absorption (AA) spectrophotometer in the early 1960s, there has been an increasing demand for better, faster, easier-to-use, and more flexible trace-element instrumentation. A conservative estimate shows that in 2015, the market for atomic spectroscopy instruments will be close to $1 billion in annual revenue - and that doesn't include aftermarket sales and service costs, which could increase this number by an additional $400–500 million. As a result of this growth, we have seen a rapid emergence of more-sophisticated equipment and easier-to-use software. Moreover, with an increase in the number of manufacturers of atomic spectroscopy instrumentation and its sampling accessories, together with the availability of newer techniques like microwave-induced plasma optical emission spectroscopy (MIP-OES) and laser-induced breakdown spectroscopy (LIBS), the choice of which technique to use is often unclear. It also becomes even more complicated when budgetary restrictions allow only one analytical technique to be purchased to solve a particular application problem.

	To select the best technique for a particular analytical problem, it is important to understand exactly what the problem is and how it is going to be solved. For example, if the requirement is to monitor copper at percentage levels in a copper plating bath and it is only going to be done once per shift, flame atomic absorption would adequately fill this role. Alternatively, when selecting an instrument to determine 15 elemental impurities in pharmaceutical products according to 

) Chapters <232> and <233>, this application is probably better suited for a multielement technique such as inductively coupled plasma–optical emission spectrometry (ICP-OES) or perhaps inductively coupled plasma–mass spectrometry (ICP-MS), depending on the materials being analyzed.

	So, when choosing a technique it is important to understand not only the application problem, but also the strengths and weaknesses of the technology being applied to solve the problem. However, there are many overlapping areas between the major atomic spectroscopy techniques, so it is highly likely that for some applications, more than one technique would be suitable. For that reason, it is important to go through a carefully thought out evaluation process when selecting a piece of equipment.

	The main intent of this installment is to look at the application requirements for the determination of 15 elemental impurities in pharmaceutical materials as described in the new 

 Chapters <232> and <233> (1,2). The goal is to offer some insight as to which might be the most suitable atomic spectroscopic approach, particularly for new users who do not have a strong background or training in the real-world application of these techniques. We will not discuss either the MIP-OES or LIBS techniques in this evaluation. They are both very useful techniques for many sample types and have been described at length in the open literature (3). However, for the purpose of space and clarity, we will be focusing on the four most commonly used atomic spectroscopy techniques: flame atomic absorption (FAA), electrothermal atomization (ETA), ICP-OES, and ICP-MS.

	To set the stage, let's first take a brief look at their fundamental principles. Simplified schematics of each of these techniques are shown in Figure 1.


	Figure 1: Simplified schematic of AA, ICP-OES, and ICP-MS.

	This is predominantly a single-element technique for the analysis of liquid samples that uses a flame to generate ground-state atoms. The sample is aspirated into the flame via a nebulizer and a spray chamber. The ground-state atoms of the sample absorb light of a particular wavelength, either from an element-specific, hollow cathode lamp or a continuum source lamp. The amount of light absorbed is measured by a monochromator (optical system) and detected by a photomultiplier tube or solid-state detector, which converts the photons into an electrical signal. As in all atomic spectroscopy techniques, this signal is used to determine the concentration of that element in the sample, by comparing it to calibration or reference standards.

	FAA typically uses a liquid sample flow of about 2–5 mL/min and is capable of handling in excess of 10% total dissolved solids, although for optimum performance it is best to keep the solids down below 2%. For the majority of elements, its detection capability is 1–100 ppb with an analytical range up to 10–1000 ppm, depending on the absorption wavelength used. However, it is not really suitable for the determination of the halogens, nonmetals like carbon, sulfur, and phosphorus, and has very poor detection limits for the refractory, rare earth, and transuranic elements. Sample throughput for 15 elements per sample is typically 10 samples per hour.

	This is also mainly a single-element technique, although multielement instrumentation is now available. It works on the same principle as FAA, except that the flame is replaced by a small heated tungsten filament or graphite tube. The other major difference is that in ETA, a very small sample (typically, 50 μL) is injected onto the filament or into the tube, and not aspirated via a nebulizer and a spray chamber. Because the ground-state atoms are concentrated in a smaller area, more absorption takes place. The result is that ETA offers detection capability at the 0.01–1 ppb level, with an analytical range up to 10–100 ppb.

	The elemental coverage limitations of the technique are similar to the FAA technique. However, because a heated graphite tube is used for atomization in most commercial instruments, it cannot determine the refractory, rare earth, and transuranic elements because they tend to form stable carbides that cannot be readily atomized. One of the added benefits is that ETA can also analyze slurries and some solids because no nebulization process is involved in introducing the sample. This technique is not ideally suited for multielement analysis because it takes 3–4 min to determine one element per sample. As a result, sample throughput for 15 elements is in the order of one sample per hour.

	Hydride generation (HG) is a very useful analytical technique to determine the hydride forming metals, such as As, Bi, Sb, Se, and Te, usually by AA (although detection by either ICP-OES or ICP-MS can also be used). In this technique, the analytes in the sample matrix are first reacted with a very strong reducing agent, such as sodium borohydride, to release their volatile hydrides, which are then swept into a heated quartz tube for atomization. The tube is heated either by a flame or a small oven that creates the ground state atoms of the element of interest and then measures by atomic absorption. When used with ICP-OES or ICP-MS, the volatile hydrides are passed directly in the plasma for excitation or ionization.

	By choosing the optimum chemistry, mercury can also be reduced in solution in this way to generate elemental mercury. This is known as the cold vapor (CV) technique. HGAA and CVAA can improve the detection for these elements compared to FAA by up to three orders of magnitude, achieving detection capability of 0.005–0.1 ppb with an analytical range up to 5–100 ppb, depending on the element of interest. It should also be pointed out that dedicated mercury analyzers (some using gold amalgamation techniques), coupled with atomic absorption or atomic fluorescence are capable of better detection limits. Because of the on-line chemistry involved, these techniques are very time-consuming and are normally used in conjunction with FAA, so they will most likely impact the overall sample throughput.

	Radially viewed ICP-OES is a multielement technique that uses a traditional radial (side-view) inductively coupled plasma to excite ground-state atoms to the point where they emit wavelength-specific photons of light that are characteristic of a particular element. The number of photons produced at an element-specific wavelength is measured by high resolution optics and a photon-sensitive device such as a photomultiplier tube or a solid state detector. This emission signal is directly related to the concentration of that element in the sample. The analytical temperature of an ICP is about 6000–7000 K, compared to that of a flame or a graphite furnace, which is typically 2500–3500 K.

	For the majority of elements, a radial ICP instrument can achieve detection capability of 0.1–100 ppb with an analytical range up to 10–1000 ppm, depending on the emission wavelength used. The technique can determine a similar number of elements as FAA can, but has the advantage of offering the capability of nonmetals like sulfur and phosphorus, together much better performance for the refractory, rare earth, and transuranic elements. The sample requirement for ICP-OES is approximately 1 mL/min, although lower flow nebulizers are available. Instruments are capable of aspirating samples containing up to 10% total dissolved solids; for optimum performance, the level is usually kept below 2%. ICP-OES is a rapid multielement technique, so sample throughput for 15 elements per sample is in the order of 20 samples an hour.

	The principle of axial ICP-OES is exactly the same as that of radial ICP-OES, except that in axial ICP-OES, the plasma is viewed horizontally (end-on). The benefit is that more photons are seen by the detector, and for this reason detection limits can be as much as an order of magnitude lower, depending on the design of the instrument. The disadvantage is that the working range is also reduced by an order of magnitude. As a result, for the majority of elements, an axial ICP instrument can achieve detection capability of 0.01–10 ppb with an analytical range up to 1–100 ppm, depending on the emission wavelength used. The other disadvantage of viewing axially is that more severe matrix suppression is observed, which means that the total dissolved solids content of the sample needs to be kept much lower. Sample flow requirements are the same as for radial ICP-OES.

	Most commercial ICP-OES instruments have both radial and axial capability built in. However, because of the different hardware configurations available, some instruments work by carrying out the analysis either using the radial or the axial view (sequentially). Others can use both the radial and axial view at exactly the same time (simultaneously), which for some applications may be advantageous. Strategic use of both radial and axial view, together with the optimum wavelength selection, can extend the analytical range by 2–3 orders of magnitude. Sample throughput is approximately 20 samples per hour, the same as that of radial ICP-OES.

	The fundamental difference between ICP-OES and ICP-MS is that in ICP-MS, the plasma is not used to generate photons, but to generate positively charged ions. The ions produced are transported and separated by their atomic mass-to-charge ratio using a mass-filtering device such as a quadrupole or a magnetic sector. The generation of such large numbers of positively charged ions allows ICP-MS to achieve detection limits approximately three orders of magnitude lower than ICP-OES, even for the refractory, rare earth, and transuranic elements. As a result, for the majority of elements, an ICP-MS instrument can achieve detection capability of 0.0001–1 ppb with an analytical range up to 0.1–100 ppm, using pulse counting measurement, but can be extended even further up to 100–100,000 ppm by using analog counting techniques. However, it should be emphasized that if such large analyte concentrations are being measured, expectations should be realistic about also carrying out ultratrace determinations of the same element in the same sample run.

	The sample requirement for ICP-MS is approximately 1 mL/min, although lower flow nebulizers are available. ICP-MS systems are capable of aspirating samples containing up to 5% total dissolved solids for short periods with the use of specialized sampling accessories. However, because the sample is being aspirated into the mass spectrometer, for optimum performance, matrix components should ideally be kept below 0.2%. This is particularly relevant for laboratories that experience a high sample workload. Sample throughput will be approximately 20 samples per hour for the determination of 15 elements in a sample

	There is no question that the multielement techniques like ICP-OES and ICP-MS are better suited for the determination of the 15 elements defined in 

 Chapters <232>, especially in laboratories that are expected to be carrying out the analysis in a high-throughput, routine environment. If the best detection limits are required, ICP-MS offers the best choice, followed by graphite furnace AA (ETA). Axial ICP-OES offers very good detection limits for most elements, but generally not as good as ETA. Radial ICP-OES and FAA show approximately the same detection limits performance, except for the refractory, rare earth, and the transuranic elements, for which performance is much better with ICP-OES. For mercury and those elements that form volatile hydrides, such as As, Bi, Sb, Se, and Te, the cold vapor or hydride generation techniques offer exceptional detection limits. Figures 2 and 3 as well as Table I show an overview of the detection capability, working analytical range, and sample throughput of the major atomic spectroscopy approaches - three of the metrics most commonly used to select a suitable technique. This information should be considered an approximation and used only for guidance purposes.


	Figure 2: Typical atomic spectroscopy detection limit ranges.

	It is also worth mentioning that the detection limits of quadrupole-based ICP-MS, when used in conjunction with collision-reaction cell or multiple mass separation-selection devices, is now capable of sub-parts-per-trillion (ppt) detection limits, even for elements such as Fe, K, Ca, Se, As, Cr, Mg, V, and Mn, which traditionally suffer from plasma- and solvent-based polyatomic interference.


	Figure 3: Typical atomic spectroscopy analytical working ranges.

	This is not meant to be a detailed description of the fundamental principles and comparisons of each technique, but a basic understanding as to how they differ from each other. For a more detailed comparison of the strengths and weaknesses of the major atomic spectroscopy techniques, please see reference 5. Now, let's turn our attention to selecting the best technique based on the demands of the application.

	Comparing an instrumental technique based on its performance specifications with simple standards is important, but it bears little relevance to how that instrument is going to be used in a real-world situation. For example, instrument detection limits (IDLs) are important to know, but how are they impacted by the matrix and sample preparation procedure? What are the real-world method detection limits (MDLs) and is the technique capable of quantifying the maximum concentration values expected for this analysis? Also, what kind of precision and accuracy can be expected if you're working close to the limit of quantitation (LOQ) for the overall methodology being used? Additionally, on the sample throughput side, how many samples are expected and at what frequency will they be coming into the laboratory? How much time can be spent on sample preparation and how quickly must they be analyzed? Sometimes the level of interferences from the matrix will have a major impact on the selection process or even the amount or volume of sample available for analysis.

	Understanding the demands of an application is therefore of critical importance when a technique is being purchased, and particularly if there is a minimum amount of expertise or experience in house on how best to use it to solve a particular application problem. This could be the likely scenario in a pharmaceutical production laboratory that is being asked to check the elemental impurities of incoming raw materials used to manufacture a drug compound. They will now have to follow the new 

 Chapters <232> and <233>, which recommends the use of a plasma-based spectroscopic technique to carry out the analysis . . . or any other atomic spectroscopy technique as long as the validation protocols are met.

	The expertise of the operator should never be underestimated. If ICP-MS is being seriously considered, it generally requires an analyst with a higher skill level to develop rugged methodology free of interferences that can eventually be put in the hands of an inexperienced user to operate on a routine basis. This again is a real concern if the technique is being used by novice users who have limited expertise in running analytical instrumentation, which may be the case in the pharmaceutical industry. So, let's take a closer look at the new USP chapters to understand what atomic spectroscopy techniques might best meet the demands of this application.

New USP General Chapters on Elemental Impurities 

	Although the risk factors for heavy metal impurities in pharmaceutical materials have changed dramatically, standard methods for their testing and control have not changed much for more than 100 years, and as a result, most heavy metals limits had little basis in toxicology. For that reason, one of the most significant standards introduced by the USP in the past decade has been new methodology for determining elemental impurities and contaminants in drug products and dietary supplements. The new methods have been summarized in the 

United States Pharmacopeia 37–National Formulary 32

 General Chapters <232> and <233> for pharmaceutical products and Chapter <2232> for dietary supplements (4). These three methods will be replacing Chapter <231> (6), which is a 100-year old colorimetric test for heavy metals based on precipitation of the metal sulfide in a sample and comparing it to a lead standard.

	These new methods have been going through a review and approval process for a number of years, but by a combination of pushback from the pharmaceutical industry and discussions with other global pharmacopeias, toxicity levels and implementation timelines have been changed a number of times. However, a recent announcement by the USP (7) has indicated that these new chapters will finally be implemented on January 1, 2018, to coincide with the full approval with the 

International Conference on Harmonization

 Step 4 guidelines, which is expected to be December 16, 2017, for existing pharmaceutical products. (Note: the International Conference on Harmonization [ICH] of Technical Requirements for Registration of Pharmaceuticals for Human Use is an international body that brings together the regulatory authorities and pharmaceutical industry in Europe, Japan, and the United States to discuss scientific and technical aspects of drug registration) (8).

 Chapters <232> and <233> to get a better understanding of the analytical requirements for this application.

	Chapter <232> specifies limits for the amount of elemental impurities in drug products, drug substances, active ingredients, and excipients. These impurities may be present naturally, derived from the production catalysts, or introduced inadvertently through the manufacturing process, or they could be environmental contaminants in the pharmaceutical raw materials. When elemental impurities are known to have the potential to be present, compliance to the specified levels is a requirement. Additionally, because of the ubiquitous nature of arsenic, cadmium, lead, and mercury, those four elements at a very minimum must be monitored. The elemental impurity levels in the drug products, unless otherwise specified in an individual drug product monograph, must show compliance with the limits specified and must be made available to the regulatory agency upon request.

	A total of 15 elemental impurities (Cd, Pb, As, Hg, In, Os, Pd, Pt, Rh, Ru, Cr, Mo, Ni, V, and Cu) are specified together with their toxicity limits, defined as maximum permitted daily exposure (PDE) levels in micrograms per day for the four major drug delivery categories. The PDE limits, which represent updated levels proposed by USP in an attempt to align itself with ICH guidelines, are shown in Table II. Note: ICH guidelines define impurity levels for an additional nine elements. Of the 15 common elements, the majority of PDE levels are similar in both USP and ICH methods, but some are slightly different, which is the basis of ongoing discussions between the two organizations. For specific details about elemental impurities and their limits, please refer to the cited ICH reference (8). It's also worth pointing out that USP and ICH are going through the alignment process, and until this is finalized, it is difficult to know what the final USP PDE limits will be. For that reason, please refer to the USP elemental impurities web pages (7) for updates.

	Chapter <232> also addresses speciation, although it does not specify an analytical procedure. Each of the elemental impurities has the potential to be present in differing oxidation states or species. However, arsenic and mercury are of particular concern because of the differing toxicities of their inorganic and organic forms. The arsenic limits are based on the inorganic form, which is the most toxic. Arsenic can be measured using a total-arsenic procedure under the assumption that all arsenic contained in the material under test is in the inorganic form. When the limit is exceeded using a total-arsenic procedure, it should be demonstrated using a suitable procedure to separate the species and to ensure that the inorganic form meets the specification.

	The mercury limits are based on the inorganic form because methyl mercury, the most toxic form, is rarely an issue for pharmaceuticals. Therefore, the limit was established assuming that if mercury was present in the drug compound it would exist as the most common inorganic form. However, if there is a known potential for the material to contain methyl mercury (such as drugs, supplements, or compounds derived from fish or kelp), an appropriate speciation procedure is required.

	Chapter <233> deals with the analytical procedure, including the sample preparation procedure, instrumental method, and validation protocols for measuring the elemental impurities using one of two plasma-based spectrochemical techniques - ICP-OES, ICP-MS, or alternatively any other trace-element technique such as FAA or graphite furnace AA, as long as it meets the data quality objectives of the method defined in the validation protocol section. In addition, before any technique is used, the overall analytical procedure must be confirmed to be appropriate for the instrument being used and the samples being analyzed by meeting the Alternative Procedure Validation protocol (9). The chapter also recommends reading 

 General Chapter <730> on plasma spectrochemistry for further guidance (10).

	Meeting the validation protocol is critical when selecting the best technique for this application because all aspects of the analytical procedures, including the instrumental technique and dissolution process, must be validated and shown to be acceptable, in accordance with the validation protocol described in Chapter <233>. The requirements for the validation of a procedure for elemental impurities is then determined by following a set of quality protocols, which cover a variety of performance tests, including

	Meeting the performance requirements defined in these tests must be demonstrated experimentally using an appropriate system suitability procedure and reference material. The suitability of the method must be determined by conducting studies with the material under test supplemented or spiked with known concentrations of each target element of interest at the appropriate acceptance limit concentration. It should also be emphasized that the materials under test must be spiked before any sample preparation steps are performed.

	To get a better understanding of the suitability of the technique being used and whether its detection capability is appropriate for the analytical task, it's important to know the PDE limit for each target element, and in particular, what the 

 calls the J-value. In Chapter <233>, the 

 is defined as the PDE concentration of the element of interest, appropriately diluted to the working range of the instrument, after completion of the sample preparation procedure to get the sample into solution.

	To understand the J-value, let's use Pb as an example. The PDE limit for Pb in an oral medication defined in Chapter <232> is 5 μg/day. Based on a suggested dosage of 10 g of the drug product per day, that is equivalent to 0.5 μg/g Pb. If 2 g of sample is digested or dissolved and made up to 100 mL, that's a 50-fold dilution, which is equivalent to 10 μg/L. So the J-value for Pb in this example is equal to 10 μg/L. (Note: Further sample dilutions can be made if required, based on the technique being used.)

	The method then suggests using a calibration made up of two standards: standard 1 = 2.0 J, standard 2 = 0.5 J. So for Pb, this is equivalent to 20 μg/L for standard 1 and 5 μg/L for standard 2.

	The suitability of a technique is then determined by measuring the calibration drift and comparing results for standard 1 before and after the analysis of all the sample solutions under test. This calibration drift should be <20% for each target element. However, after the suitability of the technique has been determined, further validation protocols described previously must be carried out to show compliance to the regulatory agency if required.

	It should also be pointed out that no specific instrumental parameters are suggested in this section, but only to perform the analysis according to the manufacturer's suggested conditions and to calculate and report results based on the original sample size. However, it does say that appropriate measures must be taken to correct for interferences, such as matrix-induced wavelength overlaps in ICP-OES and argon-based polyatomic interference with ICP-MS.

	So with this information, let's take as an example the oral drug delivery method and calculate the J-values for each elemental impurity and compare them with the limits of quantitation (LOQ) for each technique to give us an assessment of their suitability. For this analytical scenario, we'll take the LOQ for the method as 10× the IDL. These LOQs were calculated by taking the average of published IDLs from three instrument vendors' application material and multiplying them by 10 to get an approximation of the LOQ. In practice, a method LOQ is typically determined by processing the matrix blank through the entire sample preparation procedure and taking 10 replicate measurements. The method LOQ, sometimes referred to as the method detection limit (MDL), is then calculated as 3× the standard deviation of these 10 measurements.

	To make this comparison valid, the sample weight was adjusted for each technique, based on the detection limit and analytical working range. For AA and ICP-OES, we used a sample dilution of 2 g/100 mL, whereas for ICP-MS we used 0.2 g/100 mL. AA and ICP-OES could definitely use larger sample weights, but for high-throughput routine analysis, we are probably at the optimum dilution for ICP-MS. Tables III, IV, and V show the comparison of AA (FAA and ETA), ICP-OES, and ICP-MS, respectively. The important data to consider are in the final column, labeled "Factor Improvement," which is the J-value divided by the LOQ. Generally speaking, the higher this number, the more suitable the technique is.

	It should be emphasized again that LOQ in these examples is just a guideline as to the real-world detection capability of the technique for this method. However, it does offer a very good approximation as to whether the technique is suitable based on the factor improvement number compared to the J-values for each elemental impurity. Clearly, if this improvement number is close to or less than one, as it is with the majority of elements by FAA, the technique is just not going to be suitable, particularly for the "big four" heavy metals, which are the most critical. On the other hand, ETA would be suitable for the majority of the impurities (including the heavy metals), except for a few of the catalyst-based elements, which are not ideally suited to the technique. However, the ETA technique is very time-consuming and labor intensive, so it probably wouldn't be a practical solution in a high-throughput pharmaceutical production laboratory.

	Table IV shows that axial ICP-OES offers some possibilities for monitoring oral drugs because the vast majority of the improvement factors are higher than one. These numbers could be further improved, especially for the heavy metals, by using a much higher sample weight in the sample preparation procedure without compromising the method. (Note: Because most commercial ICP-OES instrumentation offers both axial and radial capability, it was felt that the axial performance was the most appropriate for this comparison.)

	However, it can be seen in Table V that ICP-MS shows significant improvement factors for all impurities, which are not offered by any other technique. Even for the four heavy metals, there appears to be ample improvement to monitor them with good accuracy and precision. The added benefit of using ICP-MS is that it would also be suitable for the other methods of pharmaceutical delivery, such as intravenous feeding, injection, or inhalation, where the PDE levels are typically an order of magnitude lower. Additionally, if arsenic or mercury levels were found to be higher than the PDE levels, it would be relatively straightforward to couple ICP-MS with high performance liquid chromatography (HPLC) to monitor the speciated forms of these elements.

	There are many factors to consider when selecting a trace-element technique that is the most suited to the demands of your application. Sometimes, one technique stands out as being the clear choice, whereas other times it is not quite so obvious. And, as is true with many applications, more than one technique is often suitable. However, the current methodology described in the new 

 Chapters <232>, <233>, and <2232> presents unique challenges, not only from a perspective of performance capability, but also because of the validation protocols that have to be met to show suitability of the technique for the analytical procedure being used. From a practical perspective, there is no question that to meet all of the PDE limits in all pharmaceutical delivery methods, ICP-MS is probably the most appropriate technique. However, for oral delivery medications, especially those that can be easily brought into solution with a suitable aqueous or organic solvent, axial ICP-OES could offer a less costly approach. Also, if the sample workload requirements are not so demanding, ETA could provide a solution.

	It will be interesting to eventually see how USP limits and methodology will be aligned with ICH directives, but the recent announcement delaying implementation should not discourage global pharmaceutical and nutraceutical manufacturers to have the appropriate analytical capability in place as soon as possible to show compliance to regulators that their products are free from elemental impurities and contaminants. This is reinforced by the statement in the frequently asked questions that went along with the most recent USP announcement (7): "Given that General Chapters <232> and <2232> provide significant improvements over existing approaches in the control of elemental impurities, USP encourages users to implement the new methods as soon as reasonably possible." So it's clear, the USP takeaway message is that even though full implementation has been delayed to coincide with the completion of the 

 approval process, the pharmaceutical industry should continue to move forward with the adoption of the new methodology.

	(1) General Chapter <232> "Elemental Impurities in Pharmaceutical Materials – Limits," in second supplement to 

2014 (United States Pharmacopeial Convention, Rockville, Maryland).

	(2) General Chapter <233> "Elemental Impurities in Pharmaceutical Materials – Procedures," in second supplement to 

(United States Pharmacopeial Convention, Rockville, Maryland).

This installment evaluates the application requirements for the determination of 15 elemental impurities in pharmaceutical materials as described in the new United States Pharmacopeia (USP) Chapters  and  and offers suggestions about which atomic spectroscopic technique might be the most suitable to use.

Determining Elemental Impurities in Pharmaceutical Materials: How to Choose the Right Technique

This study focuses on United States Environmental Protection Agency (US EPA) Method 524.3 for volatile organic compounds (VOCs) in water using gas chromatography–mass spectrometry (GC–MS). The GC sample introduction and separation technique specified by the EPA is purge and trap and the detector is MS. This article discusses optimizing MS and chromatographic separation parameters when using alternative carrier gases. In particular, it discusses the results of using nitrogen and hydrogen as alternate carrier gases to helium for EPA Method 524.3 for VOC analysis and specifically the impact of MS and chromatography parameters on the optimization process. Data are presented showing that nitrogen could be a more suitable carrier gas than hydrogen when compared to helium.

	The United States Environmental Protection Agency (US EPA) has established several methods to ensure safe and contaminant-free drinking water. One of these methods is described in EPA Method 524.3 (1), the determination of 75 volatile organic compounds (VOCs) using purge-and-trap sample concentration and gas chromatography (GC) coupled with mass spectrometry (MS). In this method, the volatile components are concentrated onto a trap, which is then heated to elute the compounds onto a GC column, typically using helium as a carrier gas, where they are separated and finally detected by MS. However, the concerns of a helium shortage and the significant increase in market prices have resulted in chromatographers investigating the use of other carrier gases to see if acceptable performance can be achieved. This article describes a suite of MS and chromatographic separation optimization experiments that were performed to better understand whether hydrogen and nitrogen can be used as alternative carrier gases to helium and, in particular, how they compare from a performance perspective when monitoring VOCs according to EPA Method 524.3.

What Are Some of the Challenges to Using Hydrogen and Nitrogen?

	Even though hydrogen has excellent chromatographic separation efficiency with the widest range of useful linear velocities when compared to other carrier gases, it is extremely reactive and poses a safety risk, and as a result may not be the most suitable choice. In addition, the pumping efficiency of hydrogen is less than that of helium. Pumping efficiency is an important consideration because a good vacuum is critical to minimize collisions between target analytes and the carrier gas during electron ionization (EI) spectral acquisition. Additional collisions may result in atypical EI spectrum for targets and degraded sensitivity, mass resolution, and linear dynamic range. Furthermore, collisions in the source between the analytes and hydrogen atoms present may result in protonation. For these reasons, there are concerns about meeting the method tuning criterion using hydrogen as the carrier gas.

	Nitrogen also has its challenges. The heavier nitrogen molecule may scatter the analyte ions to a greater degree in the ionization source and in the analyzer, and may degrade the detection limits for the same analyte masses compared to helium. Finally, nitrogen's narrower range of useful linear velocities than helium or hydrogen can be easily overcome by careful flow rate selection.

	Another concern with using nitrogen is that it is more easily ionized than helium and therefore changes the space charge conditions in the ion source. Ionization of nitrogen within the source can also reduce the number of electrons available to ionize target compounds. In addition, nitrogen can cause scattering of molecular target ions in the analyzer. Tuning MS conditions can help to alleviate some of these problems, but unfortunately it may also defocus the electron beam, which can lead to less efficient analyte ionization. All of these limitations can cause reduced sensitivity when switching from helium to nitrogen as a carrier gas. Table I shows a comparison of the three gases, with regard to a number of performance metrics.

	The goal of the study was to meet all EPA Method 524.3 validation protocols using hydrogen and nitrogen as carrier gases by overcoming the traditional limitations described previously. The drinking water method was chosen because it has the most aggressive detection limits of any of the EPA methods for VOCs. It was believed that low detection capability was the most critical challenge to overcome, and if Method 524.3 requirements could be met, performance metrics for other EPA methods could also be achieved.

	Using a 20 ppb Method 524.3 standard, GC–MS conditions were established that provided adequate target component resolution and comparable peak efficiency for nitrogen and hydrogen compared to the reference helium data. The purge-and-trap multimatrix sample preparation system (Atomx, Teledyne-Tekmar) conditions are listed in Table II; the GC–MS (Clarus SQ 8 GC–MS, PerkinElmer) conditions are shown in Table III.

	A narrow-bore column with a low flow rate was used, which reduced the amount of nitrogen and hydrogen molecules in the source. This approach resulted in improved vacuum efficiency, increased sensitivity, and reduced protonation, thus overcoming the mass spectral challenges of these carrier gases.

	To prove this concept an example is provided using nitrogen. Two columns were investigated. The first was a 20 m × 0.18 mm, 1-μm 

 Restek VMS column, coupled to a narrow-bore piece of deactivated fused silica. The VMS column offered excellent separation of the compounds of interest, and the deactivated fused silica offered the increased backpressure necessary to operate the inlet at the low flow rates being investigated. Since connectors can often bring ease of use issues, the method was also developed on a 40 m × 0.18 mm, 1-μm 

Restek VMS column. The 40-m length provided enough back pressure that the added fused silica was unnecessary. Figure 1 shows the total ion chromatograms (TIC) of the EPA Method 524.3 20 ppb test mix for each of the three carrier gases.


	Figure 1: TICs of the Method 524.3 20 ppb test mix for (a) helium, (b) hydrogen, and (c) nitrogen.

	To investigate the effect of reducing the amount of nitrogen in the source, the target compound intensities were compared using nitrogen column flow rates of 0.3, 0.4, 0.6, and 0.8 mL/min. Ultimately, 0.3 mL/min was selected as the nitrogen flow rate because it provided efficient peak widths and target compound separation at a vacuum of 1.8 × 10

	The EPA tuning criterion for bromofluorobenzene was met for each carrier gas before calibration and was also continuously verified during the experiments. The results are displayed in Table IV.

	EPA Method 524.3 allows the use of a first order, 1/

 weighted calibration, which was used in our analysis. Calibrations ranged from 0.2 μg/L to 40 μg/L for all three carrier gas studies. Seven concentration levels were used for the helium and hydrogen studies, and five concentration levels were used for the nitrogen study.

	The following additional experiments were carried out on all three carrier gases to ensure Method 524.3 initial demonstration of capability (IDC) was met:

		Recovery and precision of the midpoint standard was performed using the 20 ppb standard.

		The minimum reporting limits (MRL) of 0.2 ppb and 0.5 ppb were investigated by injecting seven replicates of each concentration to calculate the upper and lower limits for prediction interval of results (UPIR and LPIR, respectively).

		All concentration levels were processed as quality control (QC) checks using the respective calibration.

		A blank was analyzed after the high-level standard and several times during the experiments to ensure system cleanliness and determine if any carryover was occurring.

	Using the optimum linear velocity for nitrogen, the chromatographic performance is very similar to helium on both columns tested; therefore, chromatographically, nitrogen works for this method, as demonstrated by the TICs in Figure 1. Additionally, the tuning criteria were met using all three carrier gases as demonstrated in Table IV. Therefore, under these GC–MS conditions, hydrogen protonation is not a concern for bromofluorobenzene tuning.

	An increase in nitrogen flow rate resulted in a decrease in target response and an increase in the background of N

 42). This result suggests that the increase of nitrogen molecules in the source decreased the response of the target analyte compounds, which could be attributed to a combination of increased electron consumption by nitrogen ionization, resulting in reduced target analyte ionization, and increased collisions of nitrogen molecules and ions with the analytes.

	Further study is required to determine the cause of this decrease in response and will be the focus of future work.

	Calibrations were qualified as required under EPA Method 524.3 (section 10.1.10). Table V contains the correlation coefficients and the reporting limit results. The range of correlation coefficients was 0.9990–1.0000 across the components. In addition, Table V contains the signal-to-noise ratio (S/N) achieved at the reporting limit. Please note a shorter list of targets is tabulated for space requirements; however, the full list of 75 compounds is available in the on-line version of this article.


	Table V: Correlation coefficients, reporting limit results, and signal-to-noise ratios (S/N)

	The method precision and accuracy criteria in Method 524.3 for the midpoint standard are ±20% and 80–120%, respectively. The performance criteria for all three carriers at the midpoint were exceeded. The precision results on selected targets are presented in Table VI. The results for both UPIR and LPIR criteria, calculated using the MDL, which are a part of the EPA Method 524.3 IDC requirements (section 9.2), are also displayed in Table VI. The method limits for the UPIR and LPIR are from 50–150%.


	Table VI: Precision performance criteria for the midpoint standard and LPIR and UPIR requirements

	The MRL criteria for the 0.2 ppb concentration were met for all 75 target analytes using helium as a carrier gas. For nitrogen, 72 of the target analytes met these criteria for the 0.2 ppb concentration, whereas the three remaining targets met the criteria for a concentration of 0.5 ppb. For hydrogen, 67 of the target analytes met these criteria for the 0.2 ppb concentration, whereas the remaining eight targets met the criteria for the concentration of 0.5 ppb. Even though these exceptions met the ±50% criteria at the 0.2 ppb standard concentration, they did not meet the LPIR and UPIR criteria, so for those three targets in nitrogen and eight targets in hydrogen, an MRL of 0.5 ppb was chosen. Since the 0.2 ppb calibration point was linear, this point was maintained on the calibration.

	The results of this study show that by using optimized GC–MS conditions, hydrogen and nitrogen are appropriate to use as alternate carrier gases to helium for EPA Method 524.3 analysis. With these optimized conditions the sensitivity limitations of these two gases have been overcome. Since nitrogen is more inert, less expensive, and doesn't have the safety concerns of hydrogen, it is recommended as the superior option for helium carrier gas replacement.

	The authors acknowledge and thank Dr. Adam Patkin, PhD, Principal Applications Scientist at PerkinElmer Instruments for his expertise and understanding of mass spectrometry, and Brett Boyer, Senior Service Specialist at PerkinElmer Instruments for his expertise and help with this project.

 is with PerkinElmer Instruments in Shelton, Connecticut. 

Tom Hartlein, Jacob Rebholz, and Roger Bardsley

 are with Teledyne Tekmar in Mason, Ohio. 

 is with Scientific Solutions in Gaithersburg, Maryland. Direct correspondence to: 

	(1) B. Prakash, A.D. Zaffiro, M. Zimmerman, D.J. Munch, and B.V. Pepich, 

Measurement of Purgeable Organic Compounds in Water by Capillary Column Gas Chromatography/Mass Spectrometry

 (U.S. Environmental Protection Agency, Washington, D.C., EPA Method 524.3, 2009), 

http://www.epa.gov/ogwdw/methods/pdfs/methods/met524-3.pdf

This study focuses on United States Environmental Protection Agency (US EPA) Method 524.3 for volatile organic compounds (VOCs) in water using gas chromatography–mass spectrometry (GC–MS).

Author | Robert Thomas

Articles