Reports on Industrial Group Meetings 2001

• Spring Meeting, Univ of Reading 7-10th April 2001
  • SM1New and Future Possibilities in Powder Diffraction 10 ^th April
  • SM2Why Industry uses Crystallography (including Alun Bowen Indus trial Lecture) 9 ^th April
  • SM3Basic Powder Diffraction Workshop 8 ^th- 9 ^th April
• Pharmaceutical SIG 7th June 2001, GSK Harlow
• Autumn Meeting 1st November 2001, Pilkington, Lancashire

BCA Spring Meeting

7^th to 10^thApril 2001, Reading

SM1 - New and Future Possibilitie s in Powder Diffraction

The second ICG symposium of the Reading Spring Meeting took place on the morning of Tuesday April 10th in room 104 of the Palmer Building. A very substantial number of people managed to shake off the previous evening's revelries - that evening having seen the Conference Dinner and an extension until 1am at the bar - and attended the session, indeed filling the room to capacity for much of the time. Chairing duties were ably fulfilled by Steve Norval and Phil Holdway , deputising for Chris Frampton who was unable to attend for happy reasons.

The session's theme was designed to review recent developments and look forward to near-future developments in a number of key fields in powder diffraction. Three talks covered specific application areas, with two more technology-based presentations from diffractometer manufacturers to round off the morning.

The session was kicked off by Rob Delhez of Delft University, with a talk on Line Broadening Analysi s; Yesterday and Tomorrow. Rob described a subject in a "transition state" - his title did not contain the word "today". He described established methods of size and strain analysis, such as the Warren-Averbach approach, and how these have now been supplemented by methods for accounting for the concentrations of different kinds of lattice defects. However, so far there is little evidence for the practical application of these methods to real materials problems, although much theoretical work has bee n published. Rob made a plea for the rectification of this situation.

John Faber from ICDD (the International Centre for Diffraction Data) described some recent and ongoing developments in the Powder Diffraction File (PDF), which is very much a staple of many powder diffractionists' work. There are now links with structural database providers; the PDF is now growing rapidly, replete with calculated patterns from the Inorganic Crystal Structure Database and the Cambridge Structural Database. A new relational structure to the database will also accompany the launch of new, targeted subsets of the PDF, the first of which - metals and alloys - is already available. John described some examples of datamining using the new systems, including looking at phases of the oxides of lanthanides, and determination of atomic coordinates for beta-tantalum through examination of isostructural compounds.

Jeremy Cockcroft (Birkbeck College) gave a splendid overview of the broad subject of Riet veld refinement. His description of the subject's history pointed out the evolution of different aspects of the topic, such as the kinds of radiation sources used, the computing power applied, the peak shape parameters applied, etc. Moving on to the present, he gave examples of the kind of complex organic and organometallic compounds that are now being refined by the Rietveld method (albeit with restraints and constraints), sometimes having been determined ab initio from powder data. There are now d ynamic studies being carried out as a function of temperature and pressure, and there is now even application in structural molecular biology, with recent reports of polypeptide structure refinements and refinement of the T3R3 human insulin zinc complex. Into the future, Jeremy posed some tough questions - more structures will be solved and refined ab initio with better data from better sources, but how can we weed out the errors ? And exactly what information and/or data should we archive ?

Technical presentations from Philips and Bruker AXS wound up the session. Anna Widdowson from Philips described the capabilities of their new X'Celerator system, which uses a new detector technology called RTMS - "real time multiple strip". Frank Stowasser of Bruker AXS described the many arrangements of Goebel mirrors and beam expanders or compressors that can be used to optimise beam optics for a range of different X-ray diffraction applications.

So an entertaining and informative sess ion came to an end at lunchtime. Many thanks are due to all the speakers and to the organisers for putting together such a worthwhile morning.

Steve Maginn
Cambridge Crystallographic Data Centre (CCDC)

SM2 - Why Industry uses Crystallography

The Alun Bowen Industrial Lecture:
Crystallography in the Aerospace industry - Colin Small (Rolls Royce plc, Derby)

The Crystallographi c Section (Colin) at Rolls Royce provides engineering answers from interpreted crystallographic data - phase identification, texture analysis (for titanium), residual stress in metallic materials - and any service Colin is unable to provide himself, is obtained from ISIS, Daresbury, various Universities and commercial organizations. These answers are provided for engineering colleagues who, although well qualified themselves, are unfamiliar with aspects of crystallography such as texture analysis a nd pole figures.

Colin's talk began with an entertaining video presentation of aero engine testing. Engines such as the Trent 890, fitted to the Boeing 777, undergo rigorous ground tests to ensure that failure will never result in disintegration, rather containment of any components within the engine body. To the relief of the audience, especially those facing a return flight across the Atlantic, the engine passed all tests (crosswinds, icing, bombardment with hail, water, 2 .5lb chickens and 8lb Canada geese) with flying colours!

Thermal paints are used to show (by colour changes) the temperature to which certain parts of a jet engine rise during operation. No really systematic attempt had been made to establish a sound scientific basis for these colour changes. So, one of Colin's first jobs was to analyse the ingredients of the thermal paint, look at their high temperature chemistries, and then analyse their reaction products, in order to expl ain the colours produced at high temperatures.

One aircraft operating in the Middle East had suffered engine failure during take-off. Subsequent investigation revealed turbine blades coated in an orange-brown substance, which had blocked the blades' cooling holes. Phase analysis of the brown substance by XRD showed diopside (CaMg(SiO₃)₂), CaSO₄, quartz, Fe₂O₃, and CaCO₃. Analysis of dust from the tarmac accou nted for all these compounds except diopside. Diopside had been formed from some of the other compounds when they had struck the hot turbine blades. Reference to the diopside phase diagram suggested that the engine's operating temperature would cause the formation of solid diopside which had then blocked the turbine blades' cooling holes, causing overheating, resulting in engine failure. On the basis of this work, advice was given to raise the engine's operating temperature such that diopside would be present as a supercooled liquid phase rather than as a solid. As a result, the liquid diopside was blasted through the cooling holes and no further blockages occurred. A superb piece of all-round science!

The underlying theme of this year's Alun Bowen Lecture was really one of communication and education. Colin's response to engineers' dislike of complicated information about pole figures was to provide them initially with a simplified scheme consisting of a sketc h and 3 numbers. However, when a titanium wide chord fan blade (WCFB) had failed a bird test, and yet there appeared to be no changes in the chemistry and microstructure of the blade, the texture was checked. A 0002 pole figure revealed dramatic differences between the texture of the failed blade and that of a normal blade and it was apparent that on this occasion a simple explanation would not suffice. Some change in the manufacturing process must have occurred, and to convince the engineers, Colin supplemented a more detailed explanation with some 3-D pictures of the pole figures. Further investigation then revealed that in this particular forging process, it had taken 5 minutes longer than normal to take the ingot to the rolling mill and this time delay had caused the texture change.

The lecture highlighted just how important it is for a crystallographer (or any scientist for that matter) to be able to translate his/her analytical results into meaningful informatio n for the customer who may then act upon it. The examples quoted above illustrate the point admirably.

Pharmaceutical R&D: Protein Crystallography - David Brown (Pfizer, Sandwich, Kent)

In today's approach to drug design, teams of structural biologists are formed to develop screening methods and to look for compounds that form the basis for effective treatments for a disease. Compounds are made based on structural information from computer modelling of the biochemical targets. Thousands of compounds from existing compound libraries can be screened by high speed robots (high throughput screening, HTS), and those which show promising biological activity, called "leads", are then optimised by chemical modification. If these lead compounds pass early safety tests, they become development candidates and move forward into the exploratory development phase.

In terms of target molecules, the requirements for structu re-based drug design (SBDD) are: (a) a source of high yield protein, (b) high purity proteins, (c) crystallisable proteins, (d) high resolution diffraction patterns <2.5Å, and (e) stability to solvents such as DMSO, MeOH and EtOH. It is apparent that there is great value in data from protein crystallography for understanding the modes of protein binding in high throughput screening (HTS). There are currently some 14,000 structures stored in the protein data bank and use is made of virtual s creening in which 3-D structures of target protein and small molecules are screened for docking or selectivity.

Pharmaceutical R&D: How to make an impact with small molecule crystallography - Roy Copley (GlaxoSmithKline Pharmaceuticals, Harlow, Essex)

Single crystal studies are used at all stages of the Drug Discovery Process, for protein crystallography and for the drug itself. They are used for molecular conformation and absolute config uration studies (NMR is used almost exclusively for molecular connectivity) and are also used in conjunction with powder diffraction, solid state NMR, IR and Raman spectroscopy, DSC and TGA for studies of bulk properties and polymorphism.

A particularly interesting example of the use of crystallography concerned Augmentin, an oral antibacterial containing sodium amoxycillin. It was important to see if the crystals were solvated with methyl acetate, to determine why methanol is needed for crystallisation, and to explain the powder pattern. The space group was P2₁2₁2₁ and the asymmetric unit contained sodium amoxycillin, methyl acetate and methanol. The sodium ions and solvent molecules lay in channels parallel to the a axis, and methyl acetate did not bind to the Na ion. However, the predicted powder pattern did not match that of the bulk material. Another crystal was selected, and in this structure there was no methanol, the meth yl acetate molecule was bound to the Na ion, but still the powder pattern did not match the predicted one. A moment of inspiration resulted in the production of a calculated powder pattern derived from a 2:1 mixture of the two crystal structures, which matched the experimental pattern.

X-ray diffraction in the Minerals Industry - Nick Elton (Exeter Advanced Technologies)

Analysis in the minerals industry is used in all stages from exploration through extraction, comminution, separation, blending, storage and transport. Chemical analysis is done by wet chemistry and XRF, and phase analysis is undertaken by XRD.

Typical problems arising with quantitative XRD analysis are: (a) complex mixtures, (b) variable phases (solid solution polymorphs), (c) poor crystallinity, (d) low concentration, (e) preferred orientation, and (f) speed of analysis. The cement industry is a prime example of where these problems occur and they are compounded by lack of standards and peak overlaps in the XRD patterns. The best approach would seem to be Rietveld or whole pattern analysis. Sophisticated approaches such as these are unheard of at the moment in the minerals industry! XRD is usually only used for semi-quantitative analysis in troubleshooting, but now some European plants are using on-line XRD for clinker analysis. For on-line work, the diffractometer is literally assembled at a suitable place in the process. Phase analysis is achieved at both ambient and non-ambient temperatures, and crystallite size distribution (especially for the oil exploration industry) can be measured. Another recent advance has been the prediction of 3 & 28 day compressive strength from XRD data using only 2 measurements from the XRD pattern.

In the aluminium industry too, bauxite processing suffers from lack of good standards, variable crystallinity and preferred orientation. As a consequence, the XRD/XRF mass ba lance approach is used. The conversion of goethite-diaspore to haematite is related to the amount of Al substitution, the solid solution obeying Vegard's Law. Preferred orientation is solved by spray drying methods of sample preparation. In the China Clay industry, quantification of quartz, feldspar, kaolinite and mica is achieved through Peter Salt's Quanticlay software.

XRD will continue to be a key analysis tool in terms of QA and QC, but a fast throughput is needed, beca use as is the case with all industries, cost reduction is essential.

David Rendle

SM3 - Basic Powder Diffraction Workshop

The Basic Powder Diffraction Workshop was made up of three half-day sessions and was aimed at people relatively new to the field of powder diffraction. Many graduates enter the field of powder diffraction with very little knowledge of crystal structure or how a diffractometer f unctions. This workshop aimed to correct those deficiencies. The first session provided a well-balanced introduction to the principles of crystallography relevant to an understanding of powder diffraction. In the second session the X-ray diffractometer and sample preparation were described and in the third session the problem of sample identification was covered. Each of the Workshops sessions was well attended.

Introduction to Crystallography

The speaker for this session was Jeremy Cockcroft (Birkbeck College) who covered the crystallographic principles underpinning powder diffraction. Topics discussed included symmetry groups and crystal systems, space group diagrams, factors affecting peak position, peak intensity and line profiles in a powder diffraction pattern and reflection multiplicity. The principles and practice of indexing reflections by assigning hkl values to each diffraction peak, various indexing methods, the choice of re flections for indexing and indexing problems were also covered. Other topics ranged from unit cell refinement (including the need to take into account zero errors and instrumental aberrations), space group determination and whole pattern fitting. Finally Jeremy discussed reflection conditions for the various crystal classes including conditions for glide planes and screw axes.

Powder Diffraction Instruments

Judith Shackleton (Manchester Materi al Science Centre) started the second session with an introduction to the production of monochromatic radiation, X-ray tube construction and choice and the correct tube settings to use. This was followed by a discussion on diffraction geometry, including the Bragg-Brentano geometry which is generally used in powder diffractometers. Other topics included the choice of filters and monochromators to remove unwanted radiation and the use of Soller slits and divergence slits and of primary and secondary optics in general. Judith then talked about the sources of errors in powder diffraction data. These included systematic errors such as instrumental design and errors caused by the instrument being incorrectly set up. Sample errors were also discussed and included sample preparation, sample height, surface roughness, particle statistics, transparency of sample, crystal size and texture.

Phase Identification

The speaker for this session was Joh n Faber (ICDD) who began with a brief history of the development of powder diffraction data bases in qualitative analysis up to the present day and the introduction of the more versatile Relational Data Base (RDB). The ICDD powder diffraction file (PDF) currently contains about 130,000 entries. John demonstrated the Boolean search system and the Hanawalt and Fink search methods for qualitative analysis and followed this with a discussion of fully automated search methods, which are based on these. T he latest developments in automated search methods now include programs, which have total pattern matching. Finally John demonstrated the Relational Data Base (RDB) version of the PDF which gives a more flexible search environment and allows easier updating of information.

Jo Jutson

Pharmaceutical Special Interest Group Meeting 7th June 2001

Morning Session:

The meeting was opened with the prese ntation "Automated Solutions to High Throughput Crystal Screening" by Paul Higginson of Pfizer. Paul's presentation highlighted the cost that could occur in the pharmaceutical industry should the wrong solid form be taken into development. He highlighted the benefits of automated screening, not just to the researcher who is relieved of the monotony of repeated experiments, but also to the company who can reduce the risk of an unwanted form appearing late in development. Paul demonstrated the use of automated screening, and the instrumentation involved, with a successful case study. He concluded by emphasizing that a high quantity of data needs high quality data management to produce usable knowledge.

"Prediction, Morphology and Mechanical Properties of Paracetamol" was presented by Sally Price of UCL. Sally's presentation demonstrated how difficult polymorph prediction can be, but also highlighted some of the ways that may make polymorph prediction more feasible in the future. She hig hlighted how the use of global minima in lattice energy would always be necessary, but pointed out the high number of minima that normally occur within a feasible range. She then went on to explain how you could eliminate unlikely structures based on their morphologies. For instance, crystal structures whose elastic constants mean they would be too compressible to be realistic are unlikely to form. In addition, crystal structures whose slowest growing face is too slow are also unlikely to form. Sall y concluded by suggesting that involving nucleation kinetics may be the way forward in this area.

Our hostess Clare Anderton gave the final talk of the morning entitled "On-Line Monitoring of Solid-State Form during Crystallisations by Raman Spectroscopy". Clare went on to explain what Raman Spectroscopy is and how it works. She then highlighted some of the advantages over methods such as XRPD, especially for online monitoring. These included: high throughput, real time investigation, mult iplexing with more than one probe, portability of the instrument and the sensitivity of Raman to solids forming in solution. Clare then took us through two case studies to show just how effective the method can be. The first case study showed how the crystalline form could be monitored from the mid-point to the end of a crystallisation procedure. The second case study showed how the form of a product varied during the manufacturing procedure, previously unseen by off-line measurements. Both case stu dies allowing quantification of forms to as little as 5-10% in situ.

Brett Cooper, Merck Sharpe & Dohme

Afternoon Session:

The first presentation of the afternoon was by Dr. Arjen van Langevelde of Crystallics on "High-Throughput XRPD in Polymorph Discovery". Crystallics is a company which specialises in the investigation of solid form pharmaceutical materials, in particular polymorph identification, characterisation and production.

Polymorph identification is achieved through an automated high-throughput crystallisation screen. Here concentration, solvent composition, temperature, cooling gradients and ripening time can be individually controlled to produce a wide range of crystallisation condition thus maximising the chances of discovering polymorphs. More than 1000 different crystallisation experiments can be undertaken at one time, yet each crystallisation takes place in 20 microlitre wells, minimising the amount of material required. The resultant solids are screened for crystallinity by an automated X-Ray Powder Diffraction (XRPD) system, while proprietary software has been developed for the interpretation of the XRPD patterns to allow for fingerprint phase analysis through peak search and cross correlation. Once identified, Crystallics can characterise these polymorphs, off-line, by thermal analysis (DSC, TGA) and by crystal structure determination, typically using Single Crystal X-ray Diffraction. If ne cessary, a 'MultiMax' reactor can be used to optimise crystal growth conditions to produce suitable quality single crystals. If only small, poorly diffracting crystals are available the crystal structure can be determined from high-resolution XRPD data. Ab initio structure prediction is also offered as a means of structure determination. Finally, once polymorphs have been identified and characterised, the crystallisation conditions to optimise the production a specific polymorph on scale up c an be precisely determined using the 'MultiMax' reactor in batches up to 50 ml. This system is able to automatically determine the meta-stable zone through solution turbidity measurements. (Further information can be found at http://www.crystallics.nl/)

Dr. Alastair Florence of the University of Strathclyde continued the session with his presentation entitled 'DASHtastic Adventures with the Bruker-D8'. Alastair related his recent success in using the DASH program to solve molecular crystal structures from Powder X-ray diffraction data collected on the Bruker-D8 diffractometer. The diffractometer in his laboratory is equipped with a primary monochromator to provide Cu K_A1 radiation and a Position Sensitive Detector (PSD) which enables the accurate location of reflection positions along with good angular resolution and superior counting statistics. This specification produces data with a quality that maximises the chances of solving the structure. It is vital that the effect s of preferred orientation on intensities are removed from the measured patterns and this is achieved by collecting data using a rotating capillary geometry. To reduce absorption effects borosilicate capillaries are used, while data is improved still further by using a 1mm slit which reduces the background at no cost to the measured intensities. Data analysis with DASH begins with unit cell and space group determination. The peak positions of the first 20 reflections are located using DASH and passed to an input file for a third party indexing program such as DICVOL-91. After indexing, a careful process of unit cell and space group verification takes places by visual comparison of predicted 'tick marks' vs observed peak positions. The next stage is to perform a Pawley fit of the raw data which seeks to model the pattern background, zero point, intensities, peak widths and also refine the unit cell parameters. Typically data is collected to 60° 2 Theta but the high angle data is often omitted from the fit due to a high degree of peak overlap and poor signal-to-noise in this region. After data processing, the business of solving the structure can take place and this was described using Chlorpropamide as an example. Initially a trial structure was generated by placing the asymmetric unit at random within the unit cell. Comparison of the simulated XRD pattern with the experimental one clearly demonstrated that the structure was incorrect. A Simulated Ann ealing algorithm was then used to move the molecules about the cell and vary torsion angles within the molecules to generate more trial structures and minimise the difference between simulated and experimental patterns. 300,000 structures were tested to find the one that gave an excellent fit across the whole pattern at which point the structure was considered to be solved. The process took just 5 minutes!

The third presentation of the afternoon was by Detlef Beckers of Phillips Analytical and was entitled "Temperature and Humidity Controlled X-Ray Diffraction Analysis on 4-Epi-Oxytetracycline". The ability to monitor structural changes of a material as a function of temperature and relative humidity is of great importance to pharmaceutical industry. An unexpected phase transition at high temperature and high humidity during storage or transport can turn a powerful medicine into a useless powder. Phillips Analytical have developed a stage for the X'Pert Pro Diffraction System where sa mple chamber temperature and humidity can be independently controlled. Humidity levels of 5% to 95% can be achieved between room temperature and 50 ° C. An example of the use of the system was described by reference to measurements on 4-epi-oxytetracycline (OTC). OTC is a basic compound for the production of various types of antibiotics, however it is unstable under ambient conditions and can transform to 4-epi-OTC. The level of the 4-epi- OTC impurity must be controlled to be less than 1% during production. HPLC is used to monitor the 4-epi-OTC level but the system must first be calibrated against standards of accurately known 4-epi-OTC content. In order to reproducibly generate these standard samples, the environmental conditions that govern the formation of 4-epi-OTC must be understood. Using the X'Pert Pro system, three crystalline phases were found to be associated with the formation of 4-epi-OTC and by recording PXRD patterns as a function of temperature(5 ° C steps) and RH (5 % steps) it was possible to construct a phase diagram relating all three of the phases. Using this information the conditions required to generate samples of precisely known 4-epi-OTC content could be determined.

Chris Weston of Bruker AXS concluded the session with his presentation on "X-Ray Rapid Screening System for Combinatorial Chemistry". Combinatorial chemistry refers to techniques to fabricate, test a nd store the resulting data for a material library containing tens, hundreds or even thousands different materials or compounds. Combinatorial investigations require rapid screening techniques to test and evaluate variations of composition, structure and property within a material library. X-ray diffraction is one of the most suitable screening techniques for solids since it is fast and non-destructive and abundant information can be revealed from the diffraction pattern. A two-dimensional X-ray dif fraction system designed for rapid screening, D8 Discover GADDS for Combinatorial Screening, has been developed for this purpose. The system consists of a theta -theta vertical goniometer on which are housed X-ray tube and optics, a two-dimensional X-ray detector, and an XYZ sample stage plus laser/video for sample alignment and monitoring. This geometry allows the combinatorial library of samples to be mounted with ease. The two-dimensional multiwire detector can collect a large area of a diffracti on pattern with high speed, high sensitivity, low noise and derives intensities by integrating around the powder rings thus limiting the effects of preferred orientation. The laser/video system ensures that each sample is aligned accurately on the instrument center. Different areas of a sample can be probed by the system since the X-ray beam can be collimated from 1000 to 50 µm. Once all the data has been collected the GADDS software can be used to perform the user specified screen. This could be phase identification (qualitative and quantitative) or measurements such as degree of crystallinity, particle size, texture or stress. To achieve this a wide selection of screening parameters can be extracted from the patterns such as integrated intensity, maximum intensity, peak width, peak 2 theta position, crystallinity and various stress components. The screening results can be displayed in colour coded map, 3D surface plot, or pass/fail map with user defined criteria.

�Neil Feeder, Pfizer

BCA Autumn Meeting Pilkington

This year's meeting was held at Pilkington European Technical Centre, Lancashire. Thirty-four people came. Jack Brettle, Head of Science Support, welcomed us to Pilkington and opened the meeting. He briefly described the scope of the company with manufacturing plants in 25 different countries. Pilkington's major achievement was the development of the process to make float glass in the 1950s . The business is divided into 2 main lines, building and automotive products.

Most of us had an opportunity to view many products in the Exhibition area which is fascinating. It describes how different types of glass are made. Glass can be seen in a wide variety of products from complex curved windows for helicopters and cars, bullet proof glass, thermal control, optical systems and a variety of ascetically pleasing products. We tested it mechanically by using the glass stairs.

O ur host and local organiser Mark Farnworth described the sort of X-ray work carried out at Pilkington.� Powder diffraction is used to check the quality of raw materials, identify the phases in dusts and corrosion products and do troubleshooting on final products. Many of Pilkington's special products are thin films deposited on glass to improve thermal, optical or mechanical properties. The XRD data can be made surface sensitive by coming in at a glancing angle of 1 or 2 degrees which is useful. Fina lly these thin films make ideal samples for X-ray reflectometry (or reflectivity). By calculating the data from models and comparing these data with experimental data one can obtain information about the film thickness, density and roughness. Mark showed data from a Round Robin and the agreement was impressive.

The polymorphic form of a drug influences its stability, solubility and hence bioavailability, thus powder diffraction is an important tool for characterising pharmaceuticals as Anne Kavanagh (Astra Zeneca) described.� Current regulations mean that there is a requirement to check samples at all stages in their lifecycle from discovery, though manufacture to trouble shooting. Anne gave a few examples from the literature of polymorphs with different physical properties and emphasised the importance of using complementary techniques such as DSC, TGA and hot stage optical microscopy. High temperature XRPD provides additional information.� Mirrors and a position sensitive detector s peed up data collection and are beneficial for hot stage work. Astra Zeneca also have a new system with a point source and a 2-dimensional position sensitive detector. This is ideal for very small samples and could be used as a many small sample stage or for spatially resolved work.

Matthias Abraham University of Oxford described how nanocrystalline nickel can be made by electrodeposition. This material is about 5 times harder than conventional large grained nickel and thus has many potent ial industrial applications. However during wear local heating can cause grain growth which reduces the hardness. Thus grain growth and hardness were studied after heat treatments. Line broadening showed that just a few grains grow and the rest remain roughly the same size and the hardness remains high for quite a long time.

Keith Rogers' talk (Cranfield University) covered 3 distinct topics. First he talked about his work on CdTe-CdS solar cells. After describing these systems, Keith showe d how diffraction depth profiling had improved the understanding of the structural changes than can occur in the anneal to make a CdTe-CdS heterojunction. It appears that strain is the driving force for recrystallisation. Next he talked about some of his colleagues' work on bone. The challenge is to produce synthetic materials that mimic human bone and can be used to repair breaks and make implants. The effects of ageing and disease are also of great interest. Human bone is poorly crystalline and th us not well defined.� We were shown some data from the nose of a whale which was well crystalline and oriented and thus more information could be obtained from it. With thermal treatment new crystalline phases may be observed in bone. Keith showed that with increasing temperature some peaks got sharper and the modulus increased. Continuing the medical theme, the final topic was SAXS on breast tissue with and without cancer. The results show that the collagen in cancerous tissue is less ordered than that in healthy tissue. One hospital is setting up SAXS equipment to investigate the possibility of using this as a diagnostic test on biopsy samples. Current tests usually involve staining and optical microscopy.

INDUSTRIAL GROUP AWARD

Chris Frampton, Chairman of the IG, presented Ian Langford with the Industrial Group Award for his work on profil e analysis. Ian started working on powder diffraction in the early days with Arthur Wilson at Cardiff. He has recently retired and now holds an honorary position at Birmingham. He has published extensively and in several of his papers data on zinc oxide were used as an example to describe a particular method of analysis. Thus it was appropriate that Ian chose to have a Beevers model of zinc oxide for his award and then gave a talk on this material.

I had never realised that zinc oxide is us ed so widely in industry. Uses include: paint, ceramics, catalysts, lubricants, paper, electronic devices, pharmaceuticals and even chemical smoke. Most of these uses relate directly to its crystal structure which has a hexagonal lattice and polar c axis. When pure it is white and impure it is red hence its use as a pigment. The crystallites are often submicron and anisotropic and thus can be characterised by line broadening. Ian mentioned the early work by Scherrer and Debye (1916, 1918) on line br oadening, his practical and widely used approach of defining peaks as partly Gaussain and partly Lorenztian and working with integral breadths and Scardi's approach for analysing size distributions. Looking to the future, Ian said he thought that newer methods of analysis would be based on building physical models and comparing simulated and experimental data. This approach is widely used in many other areas of X-ray scattering.

�

Most industries are currently in a state of change and Ilford is no exception. David Beveridge's talk started with work relating to silver halide photography and then moved onto the challenges of trying to identify dyes and pigments in new inkjet products. Many of these are poorly crystalline and can occur in several polymorphs. The crystal structure can strongly influence colour. Returning to the silver halides, their standard method of analysis is to remove the gelatine and collect relatively high angle data on the halide particles which have a c ore shell structure. Changes in the ratio of AgI to AgBr and line broadening can be followed.

Ian Slipper talked about work done at University of Greenwich for Sandberg Testing Laboratories. This involves running samples of Ground Granulated Blast furnace Slag (GGBS) to determine its glass content in accordance with BS6699. GGBS is a by-product of the steel industry and it is sold on as an additive to cement. It improves various properties of cement such as workability, setting times, perm eability and strength but these improvements require that the material is predominately amorphous. The test method discusses scan speeds, chart paper, separation of crystalline and amorphous peaks then cutting and weighing the paper to determine the amorphous content. Ian tried to update the procedure using Bruker's software EVA. His experience showed that some things which a person would do automatically require more effort to make a computer program do the job.� After some initial difficulties he was able to obtain comparable results by both approaches.

Many industrial systems are messy and cannot be analysed by classical line profile approaches. The lines are often overlapped, backgrounds non-flat and for dynamic data there are usually many data sets with poor counting statistics. Thus neither analytical profile fitting nor Rietveld are suitable. Steve Norval (ICI) described a new approach, LOSS, linear optimisation of a simulated set. This involves simulating data for a given pha se, broadening the peaks with a single, or a range of crystallite sizes, microstrain and/or stacking faults and then convoluting this with a parameterised instrumental function. The instrumental function can be checked by using the new NIST lanthanum hexaboride standard. Steve pointed out that although this standard is better than its predecessor it still has a bit of microstrain and it should be stored under nitrogen. Both experimental data and simulated data can be used as input to LOSS. Steve dem onstrated this pragmatic approach to analysing real time data with an experimental set for the catalyst support and simulated sets for Ni and NiO. Thus the appearance and disappearance of phases during thermal treatment could be followed.

This meeting was well attended (particularly considering that Lathom is not one of the easiest places for most people to reach). Several well-known faces who are now retired, or no longer still working in diffraction, came and everyone appeared to have en joyed themselves. Credit for this is due to Judith Shackleton, who put together a meeting which covered a wide range of interesting industrial topics, and to all the speakers. Thanks are also due to Mark Farnworth, the local organiser, for his work and to Pilkington for allowing us to use their excellent facilities.

�Mary Vickers, Cambridge University