In the June installment of "Focus on Quality" (1), we looked at the classification of software from the new GAMP guide version 5 (2). This was the classification of software into four, or if you preferred my version, five different categories. Now I would like to spend time in this column looking at the life cycles that are applicable to software in categories 3, 4, and 5.  

The reason is that the life cycle associated with each category of software is the main way of determining the amount of validation work you will need to undertake. But the key to this approach is an honest and accurate appraisal of the software category. If this is mistakenly or deliberately underestimated — that is, if a category 4 application is classified as category 3 software — then a laboratory has a compliance hole that will cost more in time, effort, and reputation to sort out later than doing the work correctly in the first place. However, before we begin, let's recap what these three categories of software are in case you did not read the thrilling June installment or simply forgot what I wrote.  

Software Categories 3, 4, and 5: A Quick Reprise

GAMP version 5 (2) has defined these three software categories as follows: 									

Off-the-shelf products that cannot be changed to match the business processes but this category also can include configurable software products but where only the default configuration is used.  			

Configured software products provide standard interfaces and functions that enable configuration of the application to meet user-specific business processes. However, configuration using a vendor-supplied scripting language should be handled as custom components (category 5). 									

These applications are developed to meet the specific needs of a regulated company. This definition implicitly includes spreadsheet macros written using visual basic for applications (VBA) and LIMS language customizations. It also will include macros written for some spectroscopy software as short cuts for performing a series of tasks. Note that this is the highest risk software, as there is the greatest likelihood of functional omissions, bugs, and errors in the software, and therefore, the life cycle model used needs to have sufficient controls to ensure that it is properly specified, designed, built, and tested before release.  			

GAMP 5 notes that these categories are not silos of software but a continuum: There might be elements of a higher or lower category depending how the software is used and/or configured or customized. 

Why Is a System Life Cycle Model Important?

A software application or a computerized system does not suddenly materialize out of thin air. Each one needs to be planned and implemented. Therefore, the use of a system life cycle is important as it provides a plan to use as a basis for the implementation or building of a computerized system.  

Plan: A plan is a blue print or road map for carrying out a task. It will outline the stages or phases that need to be gone through so that the system will be built correctly and function as required. 

Basis: Plans can be changed to fit the system that is being implemented or developed and you should accept that there is no one-size-fits-all approach. A specific system can have more, the same, or fewer needs than the standard life cycle model will provide and therefore, adaptation of the plan might be needed to accommodate them. 

Therefore, the key requirement of any life cycle model used is that it should be meaningful and applicable to the system that you are building or implementing. If it is not, then you have problems.  

Most of the life cycles used for validation of computer systems in the pharmaceutical and healthcare industries are based upon a V model. For the more observant of readers, this is because the phases form a V. The basic principle of a V model is shown in Figure 1. 

There are four basic principles of V models that you will need to understand as we go through the discussion of the variations of a theme in the rest of this column. 									

The life cycle starts on the left-hand side with the specification phases, which will define what the system will do and how it will be built. Documentation of the various specifications means that you have to write down what you want, which most people do not do until after the event.

From the specification phase, there should be enough information to build the system, which occurs at the base of the V. This is where the software is programmed and informally tested by the developer or programmer.

The testing and verification of the built system occurs on the right-hand side of the V. Testing is exercising the software against a specification and verification is confirmation that a requirement is met outside of the software by the writing of a procedure or the installation of a component.

Typically, there is a symmetry in the V as the testing and verification of the system is against the specification on the same level, so there is more detailed testing at the lower levels of the V (white box testing focused on the inputs and outputs of the modules of software) and higher level testing at the top of the V (black box testing looking at the overall functions of the system).

The model implies a stately progression from specification, build, and test/verify. What is never shown or described in the textbooks and guidance documents but happens in real life is the scenic trip through the life cycle. The user requirements specification is rushed or in many cases is not written at all and then the wrong system is selected or built and the rest of the life cycle is a scrabble to recover a degree of credibility of the system and demonstrate to the boss that it sort of works. As an alternative trip through the life cycle, after placing custom software in front of the users, the feedback "This is great but we don't work this way" can mean back to the drawing board. Spend time on the specification or you will pay much more in the future.  

Also, the V model in Figure 1 does not describe who does which portion of the life cycle phases. For some systems (for example, a spreadsheet), a spectroscopist in a laboratory could undertake the whole work themselves (of course under the watchful gaze of those helpful people in quality assurance who will be riding shotgun on the work). Alternatively, if you buy commercial software, part of the V model will be your responsibility and the rest will be performed by the vendor and together, the two parts make the whole. Having understood the basics of the life cycle, at this point, enter stage left, the original GAMP V model. 

In the Beginning: The Original GAMP Life Cycle Model

In the beginning was the GAMP life cycle V model. It is important to realize that since its inception in the early 1990s, the GAMP guide has always had a life cycle model, shown in Figure 2. However, the problem, from the laboratory perspective, is that the original aim of the GAMP guide was to control suppliers of manufacturing production equipment to the pharmaceutical industry. Thus, the V model used in the first four versions of the GAMP guide from 1994 to early 2008 was intended primarily for manufacturing equipment (3). For its original intention, this V model works well for manufacturing equipment and systems that consist mainly of equipment with some computer control elements that are supplied by engineering companies. Many of these systems have a basic equipment configuration that is then customized to fit an individual facility or manufacturing line.  

The left-hand side of the life cycle deals with the specification of the system: the users write the user requirements specification and then the selected vendor, in conjunction with the users, writes the functional specification and the design specification. The system is built at the bottom of the V, typically at the vendor's site, and tests called factory acceptance tests (FAT) can be conducted on the system to see if it works correctly, and the installation qualification (IQ) could be preceded by site acceptance tests (SAT). 

However, there are many problems when this is adapted for computerized systems, the primary one being it does not fit! So in many regulated organizations, the adoption of a life cycle model for manufacturing equipment caused problems when applied to computerized systems and laboratory systems. Let us look at the problems: 									

The assumption with this model is that the system is custom and not off the shelf, in contrast to most computerized systems used in the laboratory that are commercially available and are used as is or configured.

The left-hand side of the V dealing with the specification does not include any configuration of the category 4 software to match the business process.

It implies that all stages of the model need to be carried out regardless of the value of each to the overall validation project, or indeed, the nature of the software to be validated — simple or complex.

There is no explicit consideration of writing and testing the software modules of the system, integrating them, and then testing the whole application with any attached instrumentation. 

The right-hand side of the model when applied to computerized systems does not reflect real life: you cannot install commercial software against the design specification as the vendor will have written this document, not you, and will probably not release this to you. Even if they did, you might not understand  as it is written for programmers.

The problem is that this model has been used in many organizations and in several regulatory documents. The PIC/S guide on computerized systems in GXP environments (4) uses this model indiscriminately. To misquote the eminent spectroscopist William Shakespeare, exit GAMP V model pursued by bear. However, these problems have changed with the release of GAMP version 5 (2), where the single model shown in Figure 1 has been replaced by several life cycle models, each applicable to a specific category of software rather then the one-size-fits-all approach. The models outlined here are just some of those that can be used. The V model is a relatively old concept in software engineering and other ways of working, such as the IEEE waterfall model or iterative models are just as pertinent, but space does not permit their discussion.  

In all of the models that follow, I will be focusing on the life cycle elements and phases that constitute them. However, it is important that you don't forget that you need to have control of a validation project. This can be achieved in a number of ways. 									

Validation policy: This is a document that will present how validation of different categories of software will be carried out. However, this document tends to be general in its nature and describes what needs to be undertaken for each type of software. It will be left to the project team to write the validation plan for a specific scenario.

Validation plan: This will be written for a system with a specific version of software and will describe the roles and responsibilities for the specific validation, the life cycle to be followed, and the documented evidence to be generated. This is the interpretation of the validation policy for a specific system and version of software and therefore, can justify changes in the approach predicated by the software, the system, and how the laboratory will be using it.

Validation SOP for a specific category of software such as a spreadsheet template. As such work tends to have the same stages, these can be defined in a procedure together with the documents to be written to save the effort of writing a validation plan for each template. 

Regardless of the approach, you will need to demonstrate a degree of control over the validation and these are the options for consideration. 

In the new approach advocated by GAMP 5, the simplest life cycle model is for nonconfigured product or category 3 software (2). Remember that all that is needed for this type of software is to install and configure the software and then test it against your user requirements. Therefore, the life cycle model can be compressed down into a three-phase model, as shown in Figure 2. However, what is not shown in this figure is the work that the software vendor has done in the specification of the product, coding, and testing the system that underpins the reduction in work that this model allows. This is risk management in practice as the software cannot be changed to match the business process and the only changes that are possible are setting up the run time configuration and documenting the parameters. We can rely upon the vendor's testing for the basis of our reduced testing. The focus of the testing and verification should be on the functions of the application that we use and that are documented in the URS.  

Now consider the philosophical question: Are you lazy? Don't answer, as your boss may be reading this column over your shoulder. Based upon work that we have undertaken in the past, the simplified category 3 model can be condensed to a single document called an integrated validation document. The background, approach, and detail of this document has been published recently (5). The aim of this document is to speed up further the validation of category 3 applications and systems by documenting and focusing testing on the "intended use" (6) functions of the software. Occasionally this also can require the documentation of the installation of the software if there is no IQ available from the vendor, so a section can be included in the integrated document. Space does not permit the detailed description of the approach so please read the publication to understand the approach and the use of a validation master plan (VMP) to support it (5). 

Articles

Focus on Quality

Bob McDowall looks at the different life cycle models that apply in the laboratory to GAMP software categories 3, 4, and 5.

Understanding and Interpreting the GAMP 5 Life Cycle Models for Software

By all accounts, the spectroscopy market and the materials analysis market in general have weathered the recession and downward economic turn of the last couple of years very well. To begin with, it has been well documented that the effects of these difficult economic times have been felt less severely in the spectroscopy market compared to other markets, as the demand for instrumentation and services has remained at respectable levels. 

However, recently, the concern for many has shifted to avoiding the "W" recovery forecast by many experts over the next 12 months. With budget deficits growing, there is a feeling that markets such as academia may begin to recede as pressure on public spending grows, which will have a contagious effect on other sectors of the economy as well. Combating this concern will be emerging, high-growth areas of spectroscopic research, and several of these are covered extensively in this month's issue. 

Terahertz technology is finding new applications each day, and as described in this month's "Laser and Optics Interface" column by Phil Taday, is now making a move to the process line, a development that could mean further growth for this emerging industry. Combine this application of cutting-edge techniques such as THz with innovative applications of more mature techniques such as EDXRF to electronics production, as described in this month's feature, and the field of spectroscopy is well positioned to avoid this dreaded "W" recovery pattern. 

We hope readers will find articles such as these useful not only in their daily research work but also in their attempts to grasp the bigger picture in the materials analysis market. For as has been asserted in this space many times, if a topic is being covered in these pages, it is a safe bet that it is having an impact on the market as a whole. 

From the Editor

By all accounts, the spectroscopy market and the materials analysis market in general have weathered the recession and downward economic turn of the last couple of years very well. To begin with, it has been well documented that the effects of these difficult economic times have been felt less severely in the spectroscopy market compared to other markets, as the demand for instrumentation and services has remained at respectable levels.

Avoiding the "W" Recovery

Retsch's GM 300 knife mill is designed for the sample preparation of foodstuffs. The mill has a 4.5-L capacity and a 1.1-kW motor. According to the company, the mill can be used with samples having high water, oil, or fat content, including frozen french fries, pizza, and gummy candies. The mill's grinding chamber and knife can be removed without tools, and all components that come in contact with the sample can be removed and autoclaved. The mill's speed can be adjusted for preliminary or fine size reduction throughout the 500–5000 min

  range. A digital display can store as many as 10 parameter combinations. 

InstaSpec X cameras from Oriel Instruments include a 1 × 256 pixel InGaAs array-based scientific camera and a 122 × 124 CCD camera. According to the company, both feature TE cooling and USB and ethernet connectivity. The cameras are designed to interface with the company's spectrographs and are controlled by the company's InstaSpec Pro software, which is designed to seamlessly integrate the output from both cameras into a single spectrum. 

Oriel Instruments, a Newport Corporation Brand,

BaySpec's RamSpec NIR long-wave Raman systems are designed to provide long-term reliability, compact size, and ultralow fluorescence interference for Raman analysis of biological samples such as tissue and skin. The instruments' Volume Phase Grating (VPG) serves as the spectral dispersion element and an InGaAs array detector serves as the detection element. The devices use a fiber-optic bundle or slit optics as an input, based on customer preferences. Wavelength ranges from 1064 to 1720 nm, and optimization for a 1064-nm excitation laser reportedly provides reduced fluorescence with biological samples. 

The MB3600-PH FT-NIR analyzer from ABB Analytical is designed for QA/QC, research and development, and at-line process analytical technologies applications in the life sciences and pharmaceutical industries. According to the company, the instrument is maintenance-free and includes software that enables operation in a 21 CFR Part 11-compliant mode. The instrument reportedly analyzes solid and liquid samples in a nondestructive manner and without specific sample preparation. 

The XPIN Si-PIN detector from MOXTEK is intended for use in benchtop and handheld energy dispersive X-ray fluorescence instrumentation. The detector reportedly is optimized to provide maximized resolution, count rate, energy absorption, and peak-to-background for industrial and scientific applications. According to the company, the detector has a 625-μm-thick Si-PIN diode, an ultralow-noise JFET, a multilayer collimator, a preamplifier, and a two-stage thermoelectric cooler. The detector's 8-μm windows are coated to resist corrosion and harsh environments. The detector is available with a 6-mm

The CUV-DHC-XE-LED-DA cuvette holder with attenuator from Avantes is designed for use with the company's xenon and LED light sources. By attaching the cuvette holder directly to the light source, users reportedly eliminate the need for an extra fiber. According to the company, the holder improves the light throughput by a factor of 200 in the UV range and 60 in the NIR-vis range. The holder is intended for use with 10 × 10 mm cuvettes and is equipped with a 5-mm slot for filters. 

The XploRA INV compact analytical Ramam chemical imaging microscope from HORIBA Scientific reportedly combines the automatization features and small footprint of a standard confocal Raman microscope with the capabilities of an inverted microscope for biological applications such as cell research, cancer detection, pharmaceutical verification of intercellular activities, inclusion of microreactors, and incorporation of AFM units for tip-enhanced Raman spectroscopy. According to the company, the microscope's open structure permits the use of options and add-ons for inverted microscopes, such as micromanipulators, "optical tweezers," and specific enclosures for cell applications.  

CAMO Unscrambler software from Shimadzu Scientific Instruments is designed for multivariate analysis of FT-IR data. The software includes methods for PCA, multivariate curve resolution, PLS regression, three-way PLS regression, clustering (K-Means), SIMCA, and PLSDA classification. According to the company, multivariate quantitation and other statistical data analyses are required in nutraceuticals, pharmaceuticals, and food technology. 

Chemplex's 2100 XRF sample cups with integrated overflow reservoirs are designed with sizes to accommodate all X-ray instruments. According to the company, an internal overflow reservoir collects over-pours and prevents accidental spillage during handling or in the spectrometer. A vented snap-on cap is designed to eliminate pressure differentials within the sample cup and sample chamber to avert a convoluted thin-film sample support plane. The cups' sleeve component reportedly maintains the thin-film's planar surface during assembly for a leak-resistant seal. The cell body integrates an outer flange that serves as a mechanical stop for the sleeve to ensure the same excitation source-to-sample plane distances from one preparation to another. 

Agilent's model 1200 analytical supercritical fluid chromatography (SFC) system combines the company's model 1200 Rapid Resolution LC system with the Aurora SFC Fusion A5 system. The SFC system is designed to provide a 10-fold increase in sensitivity compared with existing SFC systems. The system reportedly provides pulseless carbon dioxide flow and a dynamic range of greater than 20,000 for measuring enantiomeric excess and quantifying impurities that are 0.05% of the main peak. 

Morpho Detection's Hardened Mobile Trace handheld trace detection system is designed to detect and identify chemical warfare agents and toxic chemicals that can be weaponized. The instrument reportedly is engineered to meet the U.S. Department of Defense's MIL-STD-810-G standard for temperature, dust, impact, and humidity. The device incorporates the company's ion-trap mobility detector. 

The Evolution Array UV-vis spectrophotometer from Thermo Fisher Scientific is designed with a photodiode array for fast acquisition of full-spectrum UV-vis data in QA/QC, pharmaceutical, life science, materials science, and educational laboratories. The spectrophotometer reportedly provides simultaneous detection of all UV-vis wavelengths in the 190–1100 nm region. According to the company, users can create standard curves, plot 3D graphical displays, and examine samples at any wavelength at any time. A mechanical shutter reportedly is the only moving part. 

Hamamatsu's model S11071 back-thinned CCD is designed to provide low etaloning, strong antiblooming, and fast readout speed for use in UV and Raman applications. The CCD reportedly has a maximum readout speed of 10 MHz, and an array readout reportedly can be performed in less than 1 ms. The CCD has a pixel size of 14 × 14 mm and is available in four configurations: 1024 × 16, 1024 × 64, 2048 × 16, and 2048 × 64 active pixels. 

Hard-coated bandpass interference filters from Edmund Optics are designed for use in biotech, biomedical, and quantitative chemical applications. Reportedly, the filters selectively transmit a narrow range of wavelengths while transmitting others. According to the company, the filters are fabricated with a plasma reactive sputtering process that produces single-substrate filters with deeper blocking, higher transmission, and steeper slopes than soft coated filters. The filters are available in UV, vis, and IR center wavelengths with 10–80 nm bandwidths. 

Departments

Products

			Manufacturers and suppliers of electronic and electrical products to countries within the European Union (EU) must comply with RoHS (1) and WEEE regulations (2). 			

				Typical products that currently fall under these regulations include											

Lighting equipment, including light bulbs

Under these directives, the use and disposal of six hazardous metals–materials used in the manufacture of these products is restricted. The six hazardous materials defined in the RoHS/WEEE regulations are lead (Pb), mercury (Hg), cadmium (Cd), hexavalent chromium (Cr

 ), polybrominated biphenyls (PBB), and polybrominated diphenyl ethers (PBDE). The maximum permitted concentrations for these materials are 1000 ppm (except for cadmium, which is limited to 100 ppm) by weight of homogeneous material. The term "homogeneous" means that the limits do not apply to the total weight of the final product, or even to an individual item in the product, but to any single component that could be mechanically separated out. So, for example, in a printed circuit board, this would refer to the plastic substrate material, the metallic connecting wire, the plastic sheath covering a cable, or any other material that is used in the manufacturing process of the board. 			

XRF spectroscopy is the most widely used analytical technique for carrying out the rapid trace element testing of electrical equipment to meet the RoHS/WEEE requirements. Small, portable, or handheld instruments using the principle of EDXRF instruments are used routinely for checking various locations on a component or piece of equipment to determine whether that item is in compliance. However, the inherent problem with this technique is that X-ray beam diameter is typically only a few hundred micrometers to a few millimeters. So if the component is extremely small, accurate quantitation could be problematic. In addition, if there is any elemental segregation in the circuit board material or electrical component, the composition could be different depending upon where the X-ray beam strikes the sample. These potential limitations could mean that an item can pass or fail the RoHS/WEEE levels based upon a measurement that is not truly representative of the entire component. 			

It is therefore extremely important to understand these sampling–heterogeneity issues by measuring the elemental composition of the component as it is received and comparing it with the same sample ground up into a homogeneous mixture. This information is critical in order to make a valid assessment as to whether use of the rapid "spot checking" XRF method of electronic components is accurate and truly representative of the trace element composition of that sample. Before we discuss the experimental details of the methodology, let us remind ourselves the principles of XRF, so we can get a better understanding of its strengths and weaknesses for this type of analysis. 

The principles of XRF spectrometry are well-documented in the literature (3). A sample is irradiated with a beam of high-energy X-rays to excite the ground state atoms of the elements being measured. As the excited electrons fall back into a lower energy orbital, fluorescence occurs and secondary X-rays are emitted that are characteristic of those elements present in the sample. The individual X-rays then are separated and measured via a system of crystals, optics, and detectors. Elemental concentrations in unknown samples are quantified by comparing the X-ray intensities against known calibration standards of a similar matrix to the sample. 			

XRF is not considered a truly ultratrace technique like atomic spectroscopic techniques, but typically is used to quantitate elemental concentrations from parts-per-million levels up to percentage levels of elements in the periodic table from Na (atomic mass 23) up to U (atomic mass 238). It generally is not used for elements below Na, because fluorescence is significantly less in this region, due to the shallow depth and angle of penetration of the emitted X-rays. Apart from the elemental coverage, the analytical domain of XRF is very similar to that of arc or spark emission optical techniques. However, the major benefit of XRF over arc/spark emission is that it can analyze both conducting and nonconducting materials as well as inorganic and organic matrices with minimal sample preparation (4). 

It is important to emphasize that XRF is available in two separate configurations: energy-dispersive (ED) and wavelength-dispersive (WD) XRF. The EDXRF approach measures the intensity of the photon energy of the individual X-rays generated, whereas in WDXRF spectrometry, the polychromatic beam emerging from a sample surface is dispersed into its monochromatic wavelengths with an analyzing crystal. A specific wavelength, corresponding to the analyte of interest, is then calculated from knowledge of the crystal and the diffraction angle. It generally is recognized that WDXRF offers higher performance than EDXRF, because of its more sophisticated optical configuration. This translates into lower detection capability (particularly for the lighter elements), higher precision, wider dynamic range, and superior spectral resolution. This makes WDXRF the preferred technique for research-type work and applications that require the highest performance. However, the major benefit of EDXRF is that it is less expensive, and is physically more compact. This makes it the technique of choice for field applications in which portability is important and for handheld XRF devices that carry out the high-throughput screening of an industrial production line like metallurgical samples or for checking toxic elements in children's toys or electrical equipment (5). 			

It is also important to emphasize that accurate quantitation in XRF is based upon the availability of good calibration standards of a similar matrix to the unknown samples. When X-rays are directed onto a solid or liquid sample, they are absorbed by the matrix elements before they eventually reach and excite the analyte atoms. If the concentration of the matrix changes for any reason, the absorption of the X-rays  also will change. This will result in different excitation characteristics, which will ultimately impact the intensity of the emitted secondary X-rays. There are ways to minimize these matrix effects, but with EDXRF in particular, it is very important to match the calibration standards with the unknown samples in order to obtain good accuracy. However, for applications in which high accuracy is not required, such as in the checking of electrical equipment, where known calibration standards are not available readily, XRF is ideally suited for the high-throughput pass–fail screening of unknown samples using semiquantitative routines. 

The experiment was split into two sections. The first set of samples was taken from a printed circuit board that was known to have failed the RoHS/WEEE regulations, while the second part of the experiment involved taking some of the circuit board electrical connectors and components. For each experiment, two sets of samples were analyzed. The first set of samples were the individual parts of the circuit boards and electrical components with minimal sample preparation, while the second set of samples were the same parts ground up into a fine powder with a cryogenic (using liquid nitrogen) laboratory grinding mill (6770 Freezer/Mill, SPEX SamplePrep, Metuchen, New Jersey) (6). For both experiments, each set of samples were analyzed for bromine, cadmium, chromium, lead, and mercury using the semiquantitative capability of an EDXRF system (S2 Ranger, Bruker AXS, Madison, Wisconsin) (7). Calibration curves were created using polyethylene standards spiked with known concentrations of RoHS/WEEE elements. It should be noted that the instrument used in this study is a small benchtop EDXRF system that is based upon similar technology to the commercially available handheld devices,  the major difference being that calibration standards were used for this study, whereas handheld devices are typically standardless. Before we describe the sampling procedures used in this study, let us briefly discuss the principles of cryogenic grinding. 			

Cryogenic grinders are basically laboratory mills that chill samples in liquid nitrogen and pulverize them into a fine powder with some kind of impacting device. The freezer mill technology used in this study incorporates an insulated tub into which liquid nitrogen is poured. The grinding mechanism is a magnetic coil assembly suspended in the nitrogen bath.  

Each sample is placed in a closed grinding vial along with a metal impactor and the vial is then inserted into the coil assembly and lowered into the liquid nitrogen. When the sample is chilled thoroughly, which is typically 10–15 min, grinding begins. The magnetic coil shuttles the impactor rapidly back and forth, pulverizing the sample against the end plugs of the vial, as shown in Figure 1. When the grinding cycle is complete, the vial is removed from the freezer mill, emptied, and cleaned. 

It's well accepted that chilling materials with liquid nitrogen to temperatures approaching -200 °C, has two important benefits for sample preparation. First, it makes flexible samples extremely brittle so they can be pulverized by impact milling relatively quickly, and second, it preserves structural and compositional aspects usually damaged or lost during room temperature pulverizing. Virtually any kind of solid sample can be pulverized with this kind of technology, including bones, rocks, ceramics, polymers, plastics, metals, asbestos, teeth, and plant materials. It is simply a matter of choosing the optimum vial and impactor, based upon the type of sample and the suite of analytes being determined. A typical turnaround time is in the order of 45–60 min/sample. 			

 Sampling Procedure for the Analysis of Noncompliant Circuit Boards, Electrical Connectors, and Components by EDXRF

Three circular, 40-mm diameter samples were cut from different locations on a circuit board known to be noncompliant with RoHS/WEEE regulations, as shown in Figure 2. Both the front and back sides of the circuit board were analyzed by EDXRF. 

Next, the same circuit board samples were cut into small pieces and all copper items were removed. Each sample was placed into a chromium-free vial and submerged in liquid nitrogen in the freezer mill for 20 min. Samples were then ground for three 2-min cycles at 10 Hz, using a 15-min precooling segment prior to grinding and 2-min cooling segments between cycles. Following this treatment, unground material was then removed from the powdered sample and submerged again in liquid nitrogen for 20 min. This material was then returned to the freezer for additional grinding. 			

Ground samples were pressed into pellets measuring 40 mm in diameter by layering the sample material onto a base of boric acid and pressing at 30 tons with a laboratory press (3630 X-Press - SPEX SamplePrep). The final pelletized samples are shown in Figure 3. These pellets were then analyzed by EDXRF, as described earlier. 

For the second part of the investigation, the connectors and components were removed from the circuit boards, sorted by type, placed into 40-mm sample cups covered with Mylar film, and analyzed by EDXRF. The samples were then ground in the freezer mill using the same protocol described previously, and analyzed not as a pressed pellet, but as loose powders. The individual connectors and components sorted by type are shown on the left, while one of the ground samples (sample C in Table II) is shown on the right in Figure 4.  

Quantitative results for the three sections of the circuit board (front and back) using both sampling procedures are shown in Table I. The elements cadmium and mercury were not detected in any of the samples before or after grinding and, therefore, are not shown in the table. Bromine levels exceeded RoHS/WEEE limits for all samples. However, by switching to nonbrominated flame retardants in the manufacture of circuit boards, this obstacle could be overcome. 

Table I: Quantitative results (ppm) for Cd, Cr, and Pb by EDXRF in three sections of the printed circuit board using the two sampling procedures described 			

The results for chromium and lead are more interesting. While chromium was not detected in any of the samples prior to grinding, it was found in samples 1 and 2 after they were ground and pressed into pellets. In particular, sample 2 was found to contain 1166 ppm of chromium, which exceeds the RoHS/WEEE maximum limit of 1000 ppm. This was not contamination from the sampling container, because the vial was made from a chromium-free hardened steel. Lead levels were measured within the acceptable range in the cut disc samples, but the results increased significantly in samples 1 and 2 after grinding and as a result, both samples then failed to meet RoHS/WEEE regulation levels. These higher levels of chromium and lead in the pelletized powders appear to be real and indicate that the EDXRF is generating spurious results because of the elemental heterogeneity of the circuit board. 

 Analysis of the Electrical Connectors and Components

Quantitative results for the various electric connectors and components (samples A-E) before and after grinding are shown in Table II. As in the first experiment, mercury was not detected in any of the samples. Since sample E was a steel alloy, containing over 95% iron and small amounts alloying elements, the presence of chromium was expected. Interestingly, cadmium was detected in sample E before grinding, but not after. This could indicate that the XRF analysis of the loose metal connectors occurred at a point where the instrument was able to detect a cadmium signal. However, after grinding, the cadmium level in the mixed, homogenized powder was, in all probability, too low to be detected.  

Table 2 Table II: XRF results (ppm in the solid) for the various electric connectors and components before and after grinding

Bromine concentrations increased dramatically for samples B–E after grinding, and as a result were all above the RoHS/WEEE limits. Lead levels increased in all cases, but significantly for samples B–E. Sample D just met RoHS/ WEEE limits for lead before grinding, but failed by a wide margin after grinding. Again, this could indicate that EDXRF might not be the optimum technique for the measurement of these elements in loose circuit board components and connectors where the elements are distributed throughout the sample in a segregated manner. 

This is just a preliminary study, but it clearly indicates that the method of sample preparation and the measurement location of the circuit board have a significant impact on the XRF results. Concentrations of metals were generally much higher for the ground-up, homogeneous samples compared to the individual parts of the circuit boards and connectors, because of heterogeneity sampling issues. This means that spot-checking various locations on a circuit board with a portable or handheld EDXRF device could produce misleading results, based upon the elemental segregation of the sample. On the other hand, processing a section of the board or component to ensure a homogeneous sample is likely to yield a result that is more representative of the overall composition of the circuit board. More work needs to be done to confirm these findings, but based upon the data generated in this investigation, extreme care should be taken when testing all kinds of electrical equipment for RoHS/WEEE regulated elements using rapid XRF screening methods. And to be absolutely certain that the reported results are truly indicative of the components being tested, serious consideration should be given to the alternative sampling procedures described in this study. 

 John E. Martin, Lea L. Anderson Smith, and Gloria Adjei-Bekoe 

are with Spex Sample Prep, Metuchen, New Jersey, 

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

(1) Directive 2002/95/EC of the European Parliament and the Council of the European Union on  Restriction Of the use of Hazardous Substances (RoHS) in electrical and electronic equipment: - 

http://www.rohs.gov.uk/Docs/Links/RoHS%20directive.pdf

(2) Directive 2002/95/EC of the European Parliament and the Council of the European Union on Waste Electrical and Electronic Equipment (WEEE): - 

http://www.berr.gov.uk/files/file29931.pdf

Article

This study evaluates the merits of different sampling techniques for the determination of a group of hazardous elements in printed circuit boards as defined by the European Union, RoHS, and WEEE.

Comparison of Different Sample Preparation Procedures for the Determination of RoHS/WEEE Regulated Elements in Printed Circuit Boards and Electrical Components by EDXRF

In the first installment of this series (1), I introduced the five fundamental symmetry elements that we consider for physical objects. In the second installment (2), I discussed why these symmetry elements form mathematical groups. Here, we will extend the group concept in terms that can be used more mathematically — ultimately, a form that can be used in spectroscopy. 

In the last column, I made the point that a symmetry operation can be represented by matrix multiplication. This is a point worth making again, so here is a quick review. Consider the point (3, 4) in two-dimensional space, as shown in Figure 1a. If a reflection plane of symmetry σ were to cut the plane of the page coincident with the 

 axis, that point would be reflected to the point (-3, 4). This is equivalent to performing the following matrix multiplication: 

				 (You may want to review the procedures for performing matrix multiplication.) The 2 x 2 matrix in the equation above is a matrix representation of σ, the reflection symmetry element. 

Figure 1b shows that this concept is not limited to points, but to shapes as well: if every point on the shape in Figure 1b, represented by (

), were operated on by the 2 x 2 matrix element representing σ, it would transfer every point to another point on the shape. That is, the expression 

holds for each and every point on the shape. Therefore, the shape has a reflection symmetry element. Furthermore, if this equation were not true for every point on the shape, the shape would not have the symmetry element σ. 

As was mentioned in the last installment, matrix multiplication is extendable to three dimensions and appropriate to all five symmetry elements.  

Spectroscopy deals with atoms and molecules. How do we construct matrix multipliers for molecules that have symmetry? Consider the 

 axis of symmetry for the hydrogen molecule H

 as shown in Figure 2a. Performing the symmetry operation switches the two hydrogen atoms, which have "a" and "b" labels to better see the switch. This suggests that the matrix multiplication that represents this symmetry operation is simply 

 symmetry operation is represented by the 2 x 2 matrix on the left. However, this is overly simplistic. That's because when we are talking about molecular properties, like spectroscopic properties, many (most?) of them are vector properties (3). This means that direction matters, not just position. The better way to represent the H

 molecule is presented in Figure 2b, with each atom having a set of unit vectors representing the positive 

 axis should be pictured as going behind the plane of the page.) On the left side, the unit vectors on the original atoms are all pointing in the positive directions. However, after operating on the molecule with the 

 symmetry element as shown on the right side, not only have the two atoms exchanged positions, but some of the unit vectors have switched to the negative directions (as shown by the minus signs on the appropriate axis labels). 

Instead of a 2 x 2 matrix to represent the two hydrogen atoms, we really need a 6 x 6 matrix to represent not just the exchange of the atoms, but the effects on the unit vectors. The appropriate 6 x 6 vector is given in Figure 3. We concede that the matrix multiplier for 

 is now more complicated than the 2 x 2 matrix first proposed. However, we also note two things: most of the matrix elements are zero, and the nonzero matrix elements show a pattern. This pattern can be described as a 3 x 3 "submatrix" that has the same form wherever it appears: 

Note that the only nonzero elements are on the main diagonal of this square matrix. Although we have used the word previously, it will be defined formally here: the entire 6 x 6 matrix in Figure 3 that represents the 

 symmetry operation is called a representation of the 

 symmetry operation. It is not the simplest representation (as we'll demonstrate shortly), so it is also called a reducible representation. 

As we have pointed out, the 3 x 3 matrix earlier is actually a simpler, more concise way of representing the entire 6 x 6 matrix in Figure 3. As such, the 3 x 3 matrix is also a representation of the 

 symmetry operation: we say that we have reduced the 6 x 6 representation to a simpler 3 x 3 representation. However, as we will see, it too is a reducible representation. The general 3 x 3 representations for each type of symmetry operation are what are given in Figure 4. Figure 4 reprints the general three-dimensional forms for each type of symmetry element (and includes information omitted its previous appearance [2]). Any given molecule's complete representation is given by a 3

 is the number of atoms in the molecule. Each of the 3 x 3 matrices in Figure 4 will appear up to 

 matrix, with all of the other matrix elements being zero. 

 molecule is potentially problematic for us, because it contains a C

 axis coincident with the molecular axis, and infinities can be troubling to deal with. As another simple example, let us consider the H

O molecule, shown in Figure 5 with its unit axes labeled. Because we now have a three-atom molecule, our complete representation for a symmetry operation is going to be a 9 x 9 matrix. We have already established (1) that the H

 symmetry, a point group that contains four symmetry elements: 

,σ, and σ'. The bottom part of Figure 5 shows how the unit vectors change with each of the indicated symmetry operations. Notice that the identity element 

 does nothing to the molecule and vectors, so there no change at all (but remember that 

 is a symmetry element that is required by the mathematical definition of a group! [2]). The 

 symmetry operation switches the two hydrogen atoms (again, labeled by "a" and "b" so we can keep track of them) and reverses the 

 unit axes on all atoms — even though the oxygen atom itself does not move, some of its axes do. The σ reflection plane is in the plane of the molecule, so while the atoms do not move, the 

 axes on all atoms are reflected to their negative directions. The σ' reflection plane bisects the H-O-H bond angle, so the hydrogen atoms are reflected onto each other and the 

 unit axes of all atoms are reflected to their negative directions. 

The matrix multiplication that leads to the vector changes in Figure 5 are given in Figure 6. Notice that, while each symmetry operation can be represented by a 9 x 9 representation, there is a simpler, different 3 x 3 matrix representation for each symmetry operation that is repeated three times in each complete representation. So while the 9 x 9 matrix is certainly complete information, it's nine times more numerical data than we need. We really only need the 3 x 3 matrices to understand how the unit axes are affected by the symmetry operations in the 

Here's the fun part: the same 3 x 3 matrix representations will be found for any object that has 

 point group symmetry. Consider diethyl ether, for example: C

O. Its complete matrix representation would be 3

 = 3·15 x 3·15 = a 45 x 45 matrix = 2025 matrix elements! What you would find in that 45 x 45 matrix, however, is fifteen 3 x 3 blocks that are the same 3 x 3 matrices we found for the H

This same concept applies to any group of symmetry operations in any point group. Each symmetry operation will be defined by a 3 x 3 matrix that appears repeatedly in the complete 3

 representation. That's what Figure 4 is giving us; all we need to do is apply the correct angle for C

 axes, and put the -1 in the correct position for the different forms of reflection planes (see reference 1 for a discussion of the various types of reflection planes), and you've got the 3 x 3 matrix you need. 

However, it can get even simpler than this. Even the 3 x 3 matrices are not the simplest way of representing the characteristics of the symmetry operations in a point group. What we are looking for is the simplest possible representation, the so-called irreducible representation. That is our next goal. 

Another Way: Introduction of Character Tables

Looking at the atomic unit vectors is one way to understand how to simplify the representations of the point groups and their symmetry operations. Another way to construct a representation is to use numbers whose values will mimic the multiplication table of the symmetry elements in the point group. Again, to refresh you, Figure 7 reproduces the multiplication table for the 

 point group that we constructed in the last installment of this series (2). Any set of numbers that reproduces the patterns in Figure 7 is an acceptable representation; ultimately, we want to find the simplest set of numbers.  

 = 1, σ = 1, and σ' = 1? We get the multiplication table shown in Figure 8. 

While trivial (everything is, after all, just one), the table in Figure 8 does mimic the overall structure of the multiplication table in Figure 7 — that is, we can break down the table above into four quadrants, and the structure of each quadrant is mimicked by the pattern of the numbers. So, the number set {1, 1, 1, 1} is a perfectly acceptable (though, again, trivial) representation of the symmetry elements {

This is not the only set of numbers whose multiplication will mimic the structure of the table in Figure 7. If we set 

 = 1, σ = -1, and ?'= -1, we get the multiplication table shown in Figure 9. This table, too, mimics the structure of Figure 7 (although there is some ambiguity, because the same number represents two different symmetry elements), so the set of numbers {1, 1, -1, -1} serve as a representation of the symmetry operations in the 

Two other sets of numbers can be easily constructed: {1, -1, -1, 1} and {1, -1, -1, 1} are two sets of numbers which, when taking the places of the symmetry operations in Figure 7, give results that mimic the distribution of products in the table. I will leave it to the reader to construct the multiplication table and verify that this is so. Let us try a set that does not work. If we consider the set {1, 0, 1, 0} to represent the four symmetry elements, we get the multiplication table shown in Figure 10, which, upon inspection, does not have the same general structure as the table in Figure 7. What this demonstrates is that not any random set of numerical values can be used in a representation. 

There are other sets of numbers that can be used to represent the symmetry operation multiplication table, but the four sets of positive and negative ones presented so far are the simplest ones. As such, they are referred to as the irreducible representations of the point group. For any point group, the irreducible representations can be listed in a table, like the one shown in Figure 11. Each individual numerical value is called a character, so the table in Figure 11 is called the character table for that point group. Each irreducible representation has a letter label to identify it, here either an 

. These labels are called the Mulliken labels, after early 20th-century physicist Robert Mulliken. These are usually the symbols used by spectroscopists; another system of labels, based upon the capital Greek letter gamma Γ are called Bethe symbols and are used by a small fraction of scientists and engineers (although Γ is sometimes used to represent any general reducible representation).  

 point group contains only 1s and -1s, character tables for other point groups can be other numerical values like 0, positive or negative 2 or 3, the imaginary root 

 =√-1 ), an imaginary exponential, a sine function, or a 2 x 2 matrix. Figure 12 shows the complete character table for 

, which describes the symmetry of the ammonia molecule, NH

, and has some non-unit characters. Note that both axes of rotation and all three planes of reflection have the same sets of characters. When this is the case, we say that these symmetry elements involved belong in the same class. Only symmetry elements of the same type will ever be in the same class; that being said, 

 are always in classes by themselves because any shape that has 

symmetry element. Rather than listing all six symmetry elements of 

 explicitly, we can combine the columns of symmetry elements in the same class and write a slightly more compact character table, also shown in Figure 12. Note that the third and fourth columns now represent more than one symmetry element, as indicated by the numeral before the symmetry element in the first row. In this case, 

 has only one element in its class, the C

 axes have two, and the reflection planes have three. It comes out as a simple fact that the number of possible irreducible representations in a point group equals the number of classes of symmetry elements. Also, although it is indeed trivial, every character table has as its first irreducible representation the one that has all +1s as characters. This is referred to as the totally symmetric irreducible representation. 

Character tables are tabulated in many standard references in group theory, symmetry, physical chemistry, inorganic chemistry, spectroscopy, and other fields of science and mathematics. They summarize how the symmetry operations of a point group behave mathematically, and are the simplest way to do so. In our next installment, we will see how we can use the character tables to help us understand the behavior and properties of things that have symmetry. 

 is a professor of chemistry at Cleveland State University in Ohio. Many of his "Baseline" columns have been reprinted in book form by SPIE Press as The Basics of 

, available through the SPIE Web Bookstore at 

. His book Field Guide to Spectroscopy was published in May 2006 and is available from SPIE Press. He can be reached at 

The Baseline

In the third part of this series, David Ball starts getting into the mathematical aspects of group theory, aspects that ultimately become useful in spectroscopy.

Group Theory and Symmetry, Part III: Representations and Character Tables

Raman spectroscopy in general, and specifically, handheld and portable Raman, has developed very rapidly over the past decade due to technological developments that have unleashed the advantages of the technique. Portable/handheld Raman spectroscopy has gone from an insignificant market to the largest portable molecular spectroscopy technique in just a few years. 

   Raman spectroscopy, like other molecular spectroscopy techniques, has the advantages of being nondestructive and providing instantaneous analysis. Its biggest advantages over other techniques include being able to analyze samples through glass and plastic. However, the optical technology to build even an effective and reliable laboratory-based Raman instrument largely did not exist until the mid-1980s, while modern holographic notch filters were not developed until around 1990. The combination of these improvements with the miniaturization and ruggedization of lasers and electronics has led to the proliferation of handheld and portable Raman instruments in the past several years. 

   The market for handheld and portable Raman instruments was less than $10 million in 2006. Fast-forward to the end of 2009, and the market has grown to more than $50 million, with no less than four new vendors entering the market space in the previous two years. While military, security, and first responders continue to be the largest markets for portable and handheld Raman, raw material inspection and final product QA/QC in the pharmaceuticals industry is a rapidly growing application, as is the technique's use in forensics and counterfeit analysis. The market should continue to see strong double-digit growth for the near future. 

   The foregoing data were based upon SDi's market analysis and perspectives report entitled Global Assessment Report, 10.5th Edition: The Laboratory Life Science and Analytical Instrument Industry, August 2009. For more information, contact Stuart Press, Vice President – Strategic Analysis, Strategic Directions International, Inc., 6242 Westchester Parkway, Suite 100, Los Angeles, CA 90045, (310) 641-4982, fax: (310) 641-8851, 

Market Profile

Raman spectroscopy in general, and specifically, handheld and portable Raman, has developed very rapidly over the past decade due to technological developments that have unleashed the advantages of the technique. Portable/handheld Raman spectroscopy has gone from an insignificant market to the largest portable molecular spectroscopy technique in just a few years.

Market Profile: Portable and Handheld Raman

The terahertz region of the electromagnetic spectrum lies between 0.06 and 3 THz, which corresponds to 2–100 cm

 . It is a region that has been largely ignored by the pharmaceutical industry until recently due to the difficulties of undertaking measurements. These difficulties include the lack of conveniently sized sources and the use of cryogenically cooled detectors. But, recent advances in semiconductors and the availability of robust industrial femtosecond pulsed lasers have resulted in a rapid exploration of terahertz region by the pharmaceutical industry. The new semiconductor sources produce extremely short bursts of terahertz radiation, which are emitted from a point source. These bursts have a broad spectral range that is applicable to spectroscopy. The terahertz region is ideally suited to the investigation of the solid-state properties of pharmaceutical materials. Recent studies have included polymorphism (1,2), phase transitions (3,4), and hydrate forms (5,6). One of the major challenges has been assigning features observed in spectra obtained in the terahertz region; assigning of vibrational absorption bands has been difficult due to the complex intermolecular interactions in organic crystalline materials. But a number of groups have begun to attempt to interpret the spectral features observed in the terahertz region (7,8). The ability to generate terahertz radiation from a point source coupled with the fact that many pharmaceutical excipients are semitransparent in this region makes has resulted in the development of imaging systems capable of nondestructive, three-dimensional investigations of solid dosage forms. Terahertz radiation generated at a source can penetrate up to several millimeters into a sample, and whenever there is change in refractive index due to either a change in chemistry or physical (that is, layers) characteristics of the tablet, a reflection results. This reflection is a ballistic reflection that occurs from the surface of the interfaces within a pharmaceutical product and because of the finite speed of light, occurs at different times. We are able to make use of the time-of-flight capabilities of the technique to discriminate between pulses coming from different boundaries in the sample and randomly scattered photons (9,10). By measuring the differences between pulses arriving at different times, maps of the coating thicknesses of dosage form can be obtained. This can be done not only for the top surface but multiple surfaces within the products. Ho and colleagues (11) used the technique to investigate the scale-up failure between a laboratory (4-kg) scale batch and a pilot (20-kg) batch. The researchers reported significant differences in the dissolution profiles of the two batches, which they assigned to differences in the density of coatings. They also were able to develop a robust partial least squares (PLS) model to predict mean dissolution time from the raw terahertz waveform. More recently, Malaterre and colleagues (12) reported the correlation between the coating thickness on push–pull osmotic pump tablet; the thicknesses were obtained via terahertz pulsed imaging and the mean dissolution time. The authors also report the observation of internal within the coating, which correlated to an interruption to the coating process. 

Until now, much of this work has been achieved with laboratory-based instruments measuring tablets at high spatial resolution. In this column, we review the progress from the laboratory into the process environment and the  monitoring of coating applied to solid dosage forms in-line.  

The generation of pulsed terahertz radiation in system is based upon a photoconductive switch. In this technique, a semiconductor material is illuminated by an ultrafast laser with a pulse duration less than 90 X 10

  s and wavelength normally 800 nm. The laser pulse photo-excites electrons in the semiconductor. An electric field is applied by use of two metallic contact pads placed on either side of the illuminating laser beam, and the metallic contacts are fabricated on the front surface of the semiconductor material, as shown in Figure 1. When the electric field and the laser pulse are both present, the electrons are accelerated, resulting in the generation of an ultrashort terahertz pulse (pulse duration less than 200 fs). These pulses are inherently broadband, with the emitted power distributed over a range from 0.06 THz to 3 THz (or 2 cm

 ). The strength of the terahertz electric field (

 is the change in current density of the electrons in the semiconductor. The change in density mainly arises from the acceleration of electrons under the application of the external electric field.  

Detection of terahertz radiation can be performed in a similar photoconductive system as the generation technique and also occurs at room temperature. The commonly used scheme is to optically gate a photoconductive switch with a portion of femtosecond pulse. The pulse again generates electrons in the semiconductor. An electric field is provided by the terahertz pulse, which drives photogenerated electrons to generates a dc electric current in the semiconductor. The current is measured by a lock-in amplifier, and it is directly proportional to the terahertz electric field. By changing the optical path difference between terahertz waveform pulse and the femtosecond pulse at the detector, the entire terahertz time domain can be sampled. The photogated switch is opened only for about 200 fs and thus, a very good signal-to-noise ratio can be achieved. It should be noted that we are measuring the terahertz electric field and not intensity, as is common in most other optical techniques, and because we are measuring the electric field, we are allowed to have negatively as well as positively going signals. Fast Fourier transforming the waveform results in a maximum signal height of greater than 70 dB in the frequency domain. 

A Laboratory-Based Terahertz Pulsed Imaging System

A laboratory-based terahertz pulsed imaging system, the TPI Imaga 2000 tablet imaging system (TeraView, Cambridge, UK), consists of two units: a core unit and a scanning unit. The core unit consists of all the electronics, the laser, and most of the terahertz generating and detecting optics. The scanning unit consists of a six-axis robot and the terahertz focusing optics. The system currently samples a pharmaceutical product in 200-µm step sizes. The spatial resolution of the system is about 250 µm, while the axial resolution is about 38 µm. The spatial resolution is a function of the wavelength and the optics used in the system, while the axial resolution is determined by the duration of the terahertz pulse. Typically, a 7-mm tablet can be measured in about 20 min, but this depends upon the sample geometry.  

The Importance of Coating on Solid Dosage Forms 

To illustrate the importance of coating thickness on product performance, we have studied some commercially available pantoprazole tablets (pantoprazole sodium). These tablets are used as a short-term treatment of erosive esophagitis symptoms, which are associated with gastroesophageal reflux disease. There is known to be an interaction between the pantoprazole molecule (the active ingredient) and the acidic moieties often used in the enteric polymers. To reduce the chance of an interaction, pharmaceutical companies place a thick inert layer between the controlled released coating (the enteric coating) and the tablet core. A schematic diagram of the tablet coating is shown in Figure 2. The enteric coating protects the tablet from the acid environment in the stomach and allows the tablet to be absorbed in the alkaline environment in the gut. Therefore, any imperfection in the enteric coating could result in the drug being delivered to the wrong part of the body. A total of 12 pantoprazole tablets were studied using terahertz pulsed imaging, and the coating thickness of both the inert layer and the enteric layer. After measurement by the tablet imaging system, each tablet was sent for dissolution. The dissolution profile of each tablet is shown in Figure 3. From the terahertz data, we can obtain the coating thickness of the enteric layer, shown in Figure 4. We are able to calculate the mean dissolution time from the data in Figure 3 and correlate it to average coating from enteric coating thickness and the inert layer. A correlation plot is shown in Figure 5; the linear correlation factor between the combined layer and mean dissolution time is 

				 = 0.9147. The spread of the dissolution data (a spread of about 33%) highlights the requirement to control the coating process. 

While there is excellent correlation between the average tablet layer coating thickness and dissolution, this is only half the story; attention must be paid to both the variation in the tablet-to-tablet distribution across the batch and the coating thickness on individual tablets. To illustrate this variation, we have investigated the coating thickness of a 10-mm diameter biconvex placebo with an HPMC coating layer with samples taken at various times during a standard coating run. At each sample time we have measured 48 tablets off-line on the system. Initially, we validated the thickness measurements against the industry standard technique of percentage weight gain, demonstrating excellent correlation, as depicted in Figure 6. 

We then look in detail at the individual tablets in the following plots where, plotted as a histogram, we show the distribution of the average thickness of the coating for each tablet taken after 105, 120, and 135 min coating time.  

This clearly demonstrates that while the average coating thickness can be correlated directly with weight gain the distribution of average coating thickness for individual tablets is wide. For example, tablets coated for 135 min with an average coating thickness of (approximately) 64 µm show a distribution for individual tablets between 54 and 75 µm. Note that in all cases so far, we have investigated the average coating thickness for tablets. If we extend this to now look at the variation for the coating thickness for individual tablets (measured point-to-point), the distribution widens further.  

In Table I, we show data from three tablets chosen from the 135-min sample time.  At the extremes of the distribution, the minimum tablet coating thickness is seen to be as low as 40 µm with the maximum thickness some 2.5 times that at 104 µm. Such a large distribution has the potential to significantly impact the dissolution time for individual tablets within a manufacturing batch. Being able to measure not only the average coating thickness of a tablet batch during manufacture but also to establish the distribution of this in real time is clearly desirable, not only from the perspective of quality control but also as a development tool. 

 Table I: Data from three tablets chosen from the 135-min sample time			

Consequently, we have developed a system in which terahertz sensors can be placed directly within a tablet coater to take direct measurements of the increasing coating thickness on tablets during a coating run. These measurements are made on individual tablets and thus provide us with not only the coating thickness but also its distribution across tablets within the batch. 

Guest columnists Phil Taday and colleagues look at the emerging technique of terahertz spectroscopy and its move from the lab to the process world.

Spectroscopy-04-01-2010