X-ray crystallography began in the Department of Chemistry in 1981 with the purchase of a Philips X-ray generator and a Stoe Weissenberg camera as a result of Professor Field's efforts.  The first structure to be solved was that of [Fe₂(m-Br)(m-P(OMe)₂NEtP(OMe)₂)]BF₄ using intensity data collected at the old CSIR in Pretoria on a 4-circle Philips X-ray diffractometer after carrying out the photography at UNP.  At that time, the only diffractometer in South Africa was the one administered by Jeff Gaffner in the NPRL (National Physics Research Laboratory).  Current students should note that data were submitted on cards with 1 instruction line per card and that relatively small least-squares refinements typically took 40 h of CPU time on the old mainframe.  Imagine the frustrations if, for some reason, the computer lost the information during the course of running such a long job!

The biggest problem in doing X-ray structures in the early '80s in South Africa was that low-temperature facilities did not exist.  Unstable samples (e.g., those with included volatile solvent molecules such as CH₂Cl₂) simply could not be examined.  Prof. Field recalls having to take a sample of relatively unstable crystals of a Ru=Ru bimetallic complex personally by air in liquid nitrogen to the NPRL in order to get a fast (low-resolution) structure of the complex.  Fortunately, as we all know, Oribi airport is 2.5 km from the Chemistry Department, so transport time could be minimized (at least in Natal).

Plotting of the crystal structures in the old days was done with the original version of ORTEP on the mainframe computer in the Science Building (yes, we did not have an abundance of workstations in Chemistry in those days).  However, the only available plotter was located 1 km away in the Faculty of Agriculture!  So, not only did 4 least-squares cycles of refinement take 40 h, but also the printout had to be retrieved from a plotter 1 km away!  (Sounds like the dark ages!)
 

In 1986, with the assistance of the NRF and the UN, an Enraf-Nonius CAD4 single crystal X-ray diffractometer was purchased for R600,000.  This was the 3^rd instrument of its kind in South Africa at the time (the others were at WITS and UCT).  The first operator of the CAD4 at UNP was Karen nee Edwards, who since obtained a PhD in protein crystallography at a cancer research institute attached to the University of London’s Hammersmith Hospital.  When she left to start her PhD work, Niyum Ramesar took her position for ca. 10 years.  Niyum left UNP in 1999 with an MSc in Inorganic Chemistry and is currently the scientific officer in charge of the neutron diffraction facility at Pelindaba outside Pretoria.

The CAD4 ran for 12 years at room temperature.  Unstable samples simply perished, or were done at WITS with their low-temperature facility.  In 1998, however, we obtained a Bruker LT3 low-temperature attachment and fitted it to the CAD4.  Several new people became heavily involved in low-temperature crystallography at UNP, namely Martin Watson, James Ryan, and Dr. Orde Munro.  Routine determination of low-temperature structures became possible and the facility was operated at ca. -80 °C or lower.  Users at the Durban and Pietermaritzburg campuses can certainly attest to the positive impact that low-temperature crystallography has made on their research efforts since 1998.

The most significant problem with the CAD4, now in its 18^th year of operation despite increasingly frequent electronic faults and other age-related problems, is that as a serial diffractometer, which measures each reflection individually in reciprocal space, data collections are slow.  The typical data collection time is 3-12 days, depending on the size of the unit cell.  We were therefore averaging 4 or 5 structures a month over the period 1998-2001 and always had a backlog of samples of between 5 and 20 crystal specimens!  Moreover, even with a 3 kW long fine focus tube, some small molecule crystal samples (notably metalloporphyrins) were close to impossible analyze as a result of intrinsically weak scattering power.  Many data sets could be solved but not refined to completion owing to the lack of high angle data.  Indeed, biological samples with large unit cells (e.g., proteins) cannot be done on the CAD4.
 

Efforts to garner funding for a sensitive and fast area detector (CCD) diffractometer began in about 1998.  In 2002, the University was able to put up 100% of the funding needed to purchase a new diffractometer.  The Chemistry Department owes uncountable thanks to the fund-raising efforts of Professors A. Bawa and S. Karim and their unwavering support for high-caliber research at the University.  The new X-ray diffractometer has undoubtedly breathed new life into research in all areas of chemistry at UNP, particularly since good structures can now be obtained even from relatively poor crystal specimens, such as natural products and bioinorganic compounds.

The situation in X-ray crystallography at the University of Natal is now very different.  The Xcalibur 2 CCD diffractometer from Oxford Diffraction is an instrument of outstanding design and craftsmanship.  Moreover, we now have one of the best low-temperature facilities in the country and routinely run ALL samples at -173 °C.  At these low temperatures, thermal motion is highly damped and all atoms can be located with higher accuracy.  The instrument is fitted with a 3 kW ceramic X-ray tube and fiber optic collimators that focus the high-intensity X-ray source on the crystal specimen.  The sensitive CCD detector and advanced electronics give high-quality data at even modest X-ray power.  Typical data collection times are now in the range 4-14 h, depending on the diffracting power of the crystal specimen under study.  Since the CCD detector can be located at any distance from 45-110 mm from the crystal sample, we can collect data on small proteins and nucleic acid specimens whose largest cell dimension is < 100 Å.

We now no longer have a problem with samples in a long queue decaying before the data can be collected!  With the modern software and PC's in our laboratory, full structure solution and refinement can take as little as 15 minutes for well-behaved samples.  Disordered samples still, however, require some serious work to solve.  Fortunately, the highly redundant data sets obtained from the CCD diffractometer usually have a high percentage (> 80%) of observed data for a reasonable crystal sample, which gives one a fair chance at fully resolving disordered structures.