The Ar-Ar mass spectrometer, imaginatively named MS1, was built in the Sheffield University Physics Department in 1966 (Fig. 1). It was used in the early development of Ar-Ar dating (Turner, 1968, 1969) and in 1969 produced the first (at that time only) Ar-Ar ages of Apollo 11 lunar rocks (Turner, 1970a,b).
MS1 – Some personal reflections by Prof. Grenville Turner
My time between leaving UC Berkeley in 1964 and obtaining a first ion beam in the MS1 on 25th November 1966 was extremely frustrating and I had to content myself writing up the Berkeley observations (Merrihue and Turner, 1966), developing theoretical models on the effects of diffusive loss of 40Ar, and repeating measurements on Craig Merrihue’s Bruderheim sample, using a borrowed mass spectrometer in Cambridge (Turner et al., 1966).
Like the workman’s hammer (as new but for two replacement heads and three handles) the MS1 has changed considerably in the intervening 47 years, reflecting new technologies which came, and in many cases went. The only original parts remaining are the magnet (purchased for £495 from AEI Ltd. in Manchester), the spectrometer frame and the main vacuum housing. The latter was machined in the Physics Department workshop from a single ingot of stainless steel (260mm x 95mm x 95mm) with a view to minimising the system volume. Bored into the steel are seven interconnecting valves, a 0.2ml metering volume, a getter housing, and what was originally the source housing. Attached precariously on top, by means of a metal to glass seal, is the activated charcoal cold finger. The cold finger with its original sample of ‘Berkeley charcoal’ has been a constant source of wonder having resisted breakage by generations of students and post docs for nearly half a century. The spectrometer was calibrated using a small hand operated McCleod gauge to trap off a measured volume of air for admission to the metering system. According to my notebook, the cost of MS1, excluding items made in house and all workshop costs, was £1,947, paid for by the University of Sheffield.
In the early days MS1 was operated using a fixed EHT (J&P Engineering, £350) and a variable magnet current (APT valve based power supply, £106) swept over the argon mass region. Ion currents, measured on an IDL vibrating reed Electrometer (£275), were recorded on an industrial chart recorder. The imminent rise of computers was becoming evident in that the charts were read using a DMac (not that Mac) digitizer table which produced punched paper tape output of peak tops, zeros and time. In Berkeley, charts were read by hand using a ruler and pencil until I was given the task by John Reynolds to design and build a digitizing chart reader. Reynolds was fond of telling a story of mass spectrometry in Chicago in the 1950s, where a student had supposedly been fired for measuring charts using a 6H pencil instead of the required 12H! In 1969 computations were done in batches on the University’s mainframe IBM computer which involved a wait of several hours for the return of results on reams of 15” x 11” line printer output. As often as not developing programs involved delving through masses of printout to locate trivial errors before resubmitting and a further wait.
In 1970 the MS1 showed its quality when it was abandoned for a year while I went for a Sabbatical in Jerry Wasserburg’s lab in CalTech (this was prompted in part by lack of support for my research from the Science Research Council in the UK). Before leaving I went round the machine, closed off all the valves, turned off all the electrics, including the ion pump supply, before ceremonially unplugging the mains connection. One year later I returned to Sheffield, inserted the mains plug and switched on the ion pump supply. After a little twitch from the needle the ion pump pressure read 10-8mm, more twitches as I opened each UHV valve in turn. No leaks, I was back in business to analyse 15415, the ‘Genesis rock’ (Turner, 1972).
Support from NERC in the 1970s allowed the lunar work to flourish on MS1 (see appended references). NERC and SRC shared responsibility for supporting lunar science in the UK. The division between the two was very simple. NERC was responsible for science below the lunar surface (geology), hence my dating work; SRC was responsible for science above the surface (astronomy). Colin Pillinger’s work on carbon isotopes mainly detected solar wind, ergo SRC’s responsibility! My support came partly in the form of new equipment but, more importantly, people, post doc Peter Cadogan and technician Dave Blagburn. Dave’s arrival in 1975 heralded many changes and improvements to MS1.
Reading charts on the DMac ended in 1972, with spectra being collected by a Dynamco data logger on ¾” magnetic tape (Fig. 1). The logger also recorded magnetic field measured by a Rawson Lush rotating coil gauss meter. The tapes were taken by hand to the mainframe along with my Fortran analysis program on IBM cards. The program located peaks and zeros and carried out regressions on specified isotope ratios. Dave’s arrival coincided with the acquisition of our first computer a DEC (Digital Equipment Corporation) PDP11, the ‘must have’ minicomputer of the 1970s. Dave started work the same week that the PDP11 arrived and, never having seen or used a computer, was naturally given the job of interfacing it to the data logger.
An interesting aside relates to the DEC salesman who later made a name for himself as the science fiction writer, James P Hogan. Along with details of the PDP11 he brought along his first manuscript entitled, Inherit the Stars, which he asked if I would mind commenting on. It’s a fascinating tale of a 50,000 year old body found on the lunar surface. My only criticism was its terse reference to the ‘limited amount of information from earlier (Apollo) lunar missions’. The upshot of this was a lunchtime stroll round the local park where I gave James a one hour lecture on the scientific results of Apollo and he made copious notes in his notebook. As a result, when ‘Inherit the Stars’ was published by Del Ray in 1977 it contained lots of authentic Apollo science! Chapter eleven begins with a commendable summary of lunar chronology. Topics such as KREEP and crustal asymmetry (‘in the Awaiting Explanation drawer’) appear naturally in discussions between the scientists. The age of the body is even determined by cosmic ray exposure dating.
Initially the PDP11 collected data directly from the data logger and carried out the computations previously done on the mainframe. Eventually the logger was dispensed with and the PDP11 took on the now familiar role of controlling the spectrometer in peak switching mode and collecting DVM readings directly. The PDP11 computer had several long forgotten ‘old world’ features. On the front panel a row of switches was used to input a short program in binary to boot the computer. The 32KB core memory was an amazing object, consisting of a bird’s nest of read and write wires passing through individual small ferrite loops, one loop per bit of memory. Fortunately this type of memory is non-volatile so that programs read in by the inefficient and temperamental cassette tapes would remain stored indefinitely. The cassette reader was soon replaced by the new dual 8” floppy disks (128kB per side). I was envious of the ionospheric physics group next door who had large removable hard disks which held a whopping 20MB!
Programs were typed in on a DecWriter, a bizarre device printing on 15” wide line printer paper one row of dots at a time and producing a sound like a gear box with broken teeth. Peak fitting and graphics were performed on a Tektronix 4010. Before the advent of commercial spreadsheets I wrote a general purpose program, Plot.bas, which was used for many years to carry out a range of calculations on tables of isotope data (ratios, regressions, ages, cumulative release, etc.), plotting the results either on the Tektronix screen or a very fancy Hewlett Packard flatbed plotter. This was a time when personal computers were being developed at a furious rate and beginning to compete with the more expensive minicomputers, especially in situations requiring limited computing power. Many of them appeared and disappeared almost overnight; who has now heard of the Exidy computer? In 1984 the Macintosh introduced the graphical user interface and WYSYG (what you see is what you get) spelling the end of the daisy wheel printer with just a single font and the flatbed plotter, soon to be universally replaced by the inkjet or the laser printer.
Dave Blagburn’s arrival led to a stream of major developments on the MS1 and its peripherals. The original version could be described as MS1-lite, a very minimalist machine, just enough to get some results from the first lunar samples. Two new low blank furnaces were constructed, complete with proper temperature controllers to replace the old Variacs. The extraction system was expanded to accommodate the new furnaces and provide a laser ablation port with a Nd-glass laser ablation system for spot analyses (Fig. 2). The oil diffusion pump and cold trap were replaced with a turbo pump and the original 8 l/sec Ferranti ion pump upgraded. A new calibration manifold with capacitance baratron replaced the glass system and the Mcleod gauge.
In 1985 I decided to replace the Nier ion source with the new more efficient source developed by Heiri Baur at ETH in Zurich. This had several knock on effects. First of all the source housing, 95mm x 38mm diameter, in the stainless block was far too small. The solution was to make a new housing, reverse the connections to the magnet, and put the collector in what had been the source housing. This involved coming up with a new compact design for the collector with Faraday, electron multiplier, and adjustable collector slit (Fig 3). Based on ion optics calculations, the new source also demanded a new flight tube with better transmission. To round things off a more secure magnet cradle was constructed, supported on linear bearings to what could safely be described as a military specification.
For several years a small Micromass 6” radius mass spectrometer for the analysis of terrestrial samples was connected to the opposite end of the extraction manifold (Fig. 4). By the time the group moved across the Pennines from Sheffield to Manchester, the MS1 was well into middle age, a work horse capable of doing what was asked of it and technical developments in the 1980s were directed to a new instrument, the MAP (Mass Analyser Products) spectrometer, purchased for terrestrial noble gas analysis.
In Manchester the major changes to MS1 (Fig. 5) have been to the gas extraction system and the extension of the Ar-Ar techniques to Kr and Xe and halogen geochemistry of crustal fluids. Analysis of fluids typically involves crushing using simple modified vacuum valves, followed by stepped heating of the residue. The Nd-glass laser used in Sheffield was replaced in 1993 by two new laser systems: a Nd:YAG laser with a wavelength similar to the Nd:glass laser (1064 vs. 1060 nm) but with the advantage of continuous output that enables stepped heating. The other laser was also Nd:YAG but frequency quadrupled to generate 266nm output so that translucent minerals, which did not absorb efficiently in the near infra-red, could be analysed. The latter was usefully employed to extract noble gases from fluid inclusions in quartz (Kendrick et al., 2001). Both these laser system were manufactured by Spectron lasers (a former UK company) and remain in regular use (Fig.5). The extraction line and mass spectrometer ion pump were fitted with pneumatic valves with the long-term goal of automation, still to be realised. Halogen analysis requires the facility to cryogenically separate the heavy noble gases (Ar, Kr and Xe) and was achieved by modifying the charcoal finger on the inlet manifold to a variable temperature design.
The 32 year-old APT valve based magnet power supply was replaced in 1999 with a solid state Kepco high voltage power supply. In the late 1980s John Saxton re-wrote the SPEC control software, which I had originally written in 1978 in RT11-based FORTRAN77, allowing a PC to replace the ageing DEC computer. In 1999 the discrete dynode electron multiplier used in the MS1 source became obsolete, and was replaced with a continuous dynode channeltron. Now, like its originator, long in the tooth, the MS1 may need to seek retirement in the foreseeable future. For the time being though it plays a valuable role in the combined analysis of noble gases and halogens in fluid inclusions and in some Ar-Ar work. It also plays a role in introducing a new intake of students and postdocs to a very basic mass spectrometer. A list of some early publications follows:
Grenville Turner (Feb. 2014)
Early publications based mainly on MS1.
Merrihue, C and Turner, G (1966) Potassium-argon dating by activation with fast neutrons, J Geophys Res, 71, 2852-2857.
Turner, G, Miller, J A and Grasty, R L (1966) The thermal history of the Bruderheim meteorites, Earth and Planet Sci Lett, 1, 155-157.
Turner, G (1968) The distribution of potassium and argon in chondrites, in Origin and Distribution of the Elements, ed L H Ahrens (Pergamon Press), 387-398.
Turner, G (1969) Thermal histories of meteorites by the 39Ar-40Ar method, in Meteorite Research, ed P M Millman (D Riedel), 407-417.
Turner, G (1970a) 40Ar-39Ar dating of lunar rock samples, Science, 167, 466-468. Turner, G (1970b) 40Ar-39Ar dating of lunar rock samples, Geochim Cosmochim Acta suppl 1, 2, 1665-1684.
Turner, G (1970c) 40Ar-39Ar age determination of lunar rock 12013, Earth Planet Sci Lett, 9, 177-180. Planet Sci Lett, 10, 227-234.
Turner, G (1971b) Argon 40 - argon 39 dating: the optimization of irradiation parameters, Earth.
Turner, G (1971a) 40Ar-39Ar ages from the lunar maria, Earth Planet Sci Lett, 11, 169-191.
Turner, G (1972) 40Ar-39Ar age and cosmic ray irradiation history of the Apollo 15 anorthosite, 15415, Earth Planet Sci Lett, 14, 169-175.
Turner, G, Huneke, J C, Podosek, F A and Wasserburg, G J (1972) Ar40-Ar39 systematics in rocks and separated minerals from Apollo 14, Geochim Cosmochim Acta suppl 3, 2, 1589-1612.
Turner, G, Cadogan, P H and Yonge, C J (1973a) Apollo 17 age determinations, Nature, 242, 513-515. Turner, G, Cadogan, P H and Yonge, C J (1973b) Argon selenochronology, Geochim Cosmochim Acta suppl 4, 2, 1889-1914.
Turner, G and Cadogan, P H (1974a) Possible effects of 39Ar recoil in 40Ar-39Ar dating, Geochim Cosmochim Acta suppl 5, 2, 1601-1615.
Turner, G and Cadogan, P H (1974b) The history of lunar basin formation inferred from 40Ar-39Ar dating, in Lunar Science VI, 826-828 (Lunar Science Institute, Houston).
Turner, G and Cadogan, P H (1975) The history of lunar bombardment inferred from 40Ar-39Ar dating of highland rocks, Geochim Cosmochim Acta Suppl 6, 2, 1509-1538.
Cadogan, P H and Turner, G (1976) The chronology of the Apollo 17 station 6 boulder, Geochim Cosmochim Acta Suppl 7, 2, 2267-2285.