Yet it was not until the mid 1800's that methodical studies of fluorescence occurred. The German romanticist Goethe, aside from his literary efforts, was fascinated by the natural sciences and was likely the first person to note in 1810 that some minerals fluoresced. By 1833, the Englishman, Sir David Brewster, had directed a beam of sunlight through a chlorophyll solution and found that the beam was red in the solution yet became the expected green once on the other side. He termed this phenomenon internal dispersion. In 1845, when Sir John Herschel found that a container of dissolved quinine bisulfate glowed in sunlight, he called it epipolic or surface dispersion. These early experimenters expose the paradigm of the day, that fluorescent phenomena were just variations on known properties of light such as diffusion and dispersion.
As no artificial sources yet existed, and all these early experiments relied on sunlight, either directed through a prism onto the subject, or through a filter into a black box. These studies were limited to low intensity longwave ultraviolet light, since ozone in the atmosphere blocks radiation shorter than 3000 angstroms. (Incandescent lamps are no better.) An effective method of producing ultraviolet did not emerge until the early 20th century. (Bush 1966; Robbins 1983: 3-13)
Towards the end of the 19th Century, a new paradigm was developing in luminescence theory. In England, Sir George Stokes repeated Goethe's work, passing a fluorite sample through a spectrum created from sunlight. He noted that the sample glowed from the middle of the violet region well into the apparently dark space beyond (i.e. the ultraviolet band). Declaring this a new physical property, Stokes coined the term fluorescence after the mineral he had examined. Meanwhile, in France, three generations of the Becquerel family were conducting the first controlled and extensive studies of fluorescent phenomena. Antoine and his son, Edward, exposed fluorescent materials to different wavelengths, measured the spectra emitted at different temperatures, and described the first recognised activator, manganese, in calcite. In 1879, the Becquerels invented the phosphoroscope, to measure the duration of phosphorescence. Edward's son, Henri, experimented with recording uranium phosphorescence on photographic plates and made history by discovering the property of radioactivity. In 1886, another Frenchman, Verneuil, linked fluorescence to the action of trace impurities, such as copper, manganese, and silver. Luminescence, particularly phosphorescence, had become a hot topic, bringing attention from great minds like Crooks, Goldstein, and Rutherford. As a result some important discoveries were made. For example, radioactive elements other than uranium were detected, and this work led ultimately to the development of a theory of radioactivity by Rutherford. Further, what we know today as the fluorescent light bulb was developed as an offshoot from the study of the rare earth elements coupled with experiments with the first cathode ray tubes by Goldstein and Crooks. (Dake & De Ment 1941: 1-7; Pringsheim and Vogel 1943: 5-6; Robbins 1983: 1-5; Wedepohl 1971: 2)
1903 heralded the development of the iron arc lamp which was capable of producing abundant SW light. That year the British Museum of Natural History created the first public display of fluorescent minerals in London. The same year, George Kunz and Charles Baskerville produced the first catalogue of fluorescent minerals. They exposed over 13,000 specimens from the American Museum of Natural History collections to ultraviolet, x, and gamma rays. In 1912, Englehardt made a similar study of just ultraviolet fluorescence, including over 400 specimens from 147 mineral species. And, in the first practical application recorded, the mines at Franklin, New Jersey, were using the arc lamp to sort fluorescent willemite ores in the early twenties. In this same period, the fluorescence of certain fossils was noted (Miethe 1927a,b; Simpson 1926; Wagner 1928). But, with this exception, the more practical applications of the effect had not yet been appreciated. (Robbins 1983: 13)
In the 20's and 30's, new sources of ultraviolet light were developed. One, the argon bulb, was only able to produce low intensities. Another, known as the Nico lamp, was more effective, but very costly to manufacture. The first mercury vapour lamps were also being developed. Once Corning developed a filter capable of passing only ultraviolet, mercury vapour lamps became the most economical and effective sources. With a high demand from the food industry and hospitals for improved methods of sterilization, these lamps became even cheaper and readily available. However, despite all their refinements, even today these lamps are not very efficient. After filtering, the long wavelengths are reduced to as little as 11% of the original output, while the shortwave is only around 43%. In addition, both lamps and filters become more opaque with use, due to interaction of their materials with the light. (In some cases, the filters can be heated slowly to 430 degrees C and then slowly cooled to reverse this effect.) (Robbins 1983: 15-19, 20-30)
Modern research really began in the mid 1930's. E.M. Gunnell (1903) refuted an earlier supposition that fluorescence was related to the same phenomena that produce colour. T.W. Ward (1935) linked fluorescence to the presence of known elements, particularly manganese. He was the first to postulate the existence of active fluorescent agents. In the Soviet Union, S.I. Vavilov, V.L. Levshin, and others reported two important mechanisms: discrete centres at which glowed isolated ions and molecules, and crystallophosphors which glowed from internal ionization and photo conduction (Marfunin 1979: 144). In the mid 1940's, two U.S. Geological Survey researchers, Murata and Smith (1946), isolated the coactivation of lead and manganese in fluorescent halite. Researchers had yet to discern the actual nature of electron capture levels. It was not until the 1950's when laser research began to shed more light on this area.
Advances in theory, new analytical techniques, and the discovery of more examples of luminescence led to a dramatic increase in luminescence research from the fifties onward; this was especially so in mineralogy, but not in ultraviolet fluorescence per se (Marfunin 1979: 141). A major development was the ability to analyse luminescence spectra quantitatively, through which the precise identification of electron transitions became possible (Marfunin 1979: 142). Luminescence data aided in the study of a number of mineralogical properties, such as crystal formation and structure, the identification of impurities, and the analysis of valence, coordination, and symmetry (Marfunin 1979: 142-3; also see Yagoda 1946). Research shifted from ultraviolet irradiation, into the use of x-, gamma-, and cathode rays, as well as proton and neutron beams. It is also only in the last few decades that the first limited archaeological applications of UVFA occur, although these are rare. While geologists had appreciated the potential for lithic sourcing as early as the 1930's (e.g. Lutati 1930; Stanciu 1937), archaeological interest in this technique only began in the late 1960's (Renfrew and Peacey 1968), and is really only beginning to spread in the 1990's (see below).
Once it was discovered that certain valuable ore minerals fluoresced, prospectors began to use ultraviolet lights in the field. UVFA has also played an important role in the mining of valuable ores. Over ninety percent of the tungsten ore deposits in the U.S. were located by their fluorescence (Riecker 1959). Lead-zinc, lithium, manganese, petroleum, and uranium deposits have also been located in this way, and UVFA is frequently used for quick and inexpensive assaying on site and in the field (Hopkins 1990). (Waychunas 1988: 664)
Most present knowledge has been gained through theoretical and experimental research into artificial compounds and naturally fluorescent solutions. This has yielded information on which elements are involved, as well as the effects of changing ambient conditions and concentration, and the presence of other impurities. A general model of fluorescence, based on quantum theory, has been created, and the groundwork has been laid out to explain how the phenomenon is moderated by these other factors. The bulk of this research is being carried out on artificial compounds by chemists and biologists, but fluorescence spectroscopy is regularly applied to chemical analyses in pollution studies and forensics, as well as clinical, biological, medical, industrial, and agricultural studies. A slightly different application is the detection of nonvisible handwriting traces on fraudulent tax documents (Riordan 1991). Luminescence research in the earth sciences has mostly shifted to x-ray and cathode luminescence, and principally focusses on synthetic minerals and their structure, rather than on rocks. (Becker 1968: 1-18; cf. Dake and De Ment 1941: 64; Hurtubise 1986; McGown 1986; Verbeek, p.c. 1994)
Over the next 10-4 seconds, the molecule relaxes, losing its excess vibrational energy in small amounts (vibrational relaxation) until the lowest energy level of the present state is attained (Guilbault 1973). Since excited states are unstable, an excited molecule must lose still more energy. For this, several paths are possible (Becker 1968: 76-8; Dake and De Ment 1941:51-2; Hurtubise 1986; Wain 1965:12-15). The loss may involve a direct jump to the ground state. This route can release a large quantum of energy in the form of a photon, which we see as fluorescence. Since the jump may terminate at any of the ground state's vibrational levels, the resulting spectrum for an entire sample will typically exhibit several wavelengths, the intensity of each corresponding to the likelihood of a particular jump (Schulman 1985). While excitation depends on the overall energy of the bombarding light, the resulting emission spectrum is independent of its wavelength (McGown 1986). This final decay process takes roughly 10-9 to 10-7 seconds. Since some energy is typically lost due to vibrational relaxation prior to fluorescence, the energy of the emitted light is less than that of the absorbed light, and the wavelength is corresponding longer. In other words, while the substance was bombarded with ultraviolet light, the fluorescence will be either higher in teh ultraviolet spectrum, or more typically up into the visible spectrum, or even in the infrared. This phenomenon is known as Stoke's Law, and the gap in wavelength betwen the two is known as the Stoke's Shift.
In the simplest case, the transmission of energy occurs virtually unaffected by the rest of the crystal structure, active electrons being in orbitals that are not involved in ion-lattice interaction. Such independence is common among the rare earths, but the best example is uranium (in the form of uranyl). Other examples are transition metals, such as manganese, molybdenum, titanium, silver, and tungsten, as well as the elements boron, lithium, mercury, tellurium, and tin (Verbeek p.c. 1994). This insensitivity is substantiated by the small overall loss in energy, the emission energy approaches that of the excitation energy. Thus the Stoke's Shift is quite small and the resulting fluorescence will tend to occur in wavelengths approaching the ultraviolet. Other rare earth elements and transition metals are much less independent. Their active electrons are in structural bond sites, making them much more sensitive to crystal structure. Their behaviour will therefore vary according to the crystal structure in which they are found, the relationship reducing or enhancing fluorescence. (McGown 1986; Waychunas 1988: 641-5)
A return to the ground state without fluorescence is possible, known as nonradiative energy transfer, during which the absorbed energy is lost to adjoining molecules as heat (a form of quenching - see below) (McGown 1986). A third path involves intersystem crossing, during which the electron drops into an atypical state. Normally a molecule is in a singlet state, a term which relates to its electrons being paired with opposing spin directions. During intersystem crossing, an electron enters a slightly lower triplet state, in which paired electrons share the same spin direction, and then returns through photon emission to the ground state, resulting in phosphorescence. However, this is not a direct path. The electron must jump from the triplet state either to an adjacent, but lower energy singlet state, or all the way back to the singlet ground state. Both of these are termed "forbidden" paths, because they are statistically very unlikely. Phosphorescence therefore involves a relatively long delay, of the order of 10-4 to 10 seconds, which the brain is capable of distinguishing from fluorescence. Due to the extended length of time the molecule spends in this excited state, the chance of radiationless decay is much higher (Guilbault 1973). Very rarely, an electron may cross back up to a singlet state, and then back to the ground state. This is called delayed fluorescence. (Dake and De Ment 1941: 51; Marfunin 1979; Radley and Grant 1959: 6; Rendell 1987; Waychunas 1988: 649)
Fluorescence in rocks relates to the presence and purity of particular minerals and the elements they contain. A fluorescent mineral consists of inert materials and one or more activators. These activators may be major or minor components of a mineral, occupy defects in the crystal lattice, or be present along grain boundaries, and they need only be present in minute quantities, even below the ppm range (Waychunas 1988: 639, 645). Many elements have been shown to be activators, particularly the transition metals mentioned above, the lanthanides or rare earth metals, actinides such as uranium, isoelectrics like bismuth, calcium, germanium, indium, lead, mercury, O-2, and S-2, as well as some other elements such as boron, lithium, mercury, tellurium, and tin (Marfunin 1979; Radley Grant 1959; Verbeek, p.c. 1994). In addition, certain organic compounds such as hydrocarbons are also active. Substances containing oil bubbles will sometimes fluoresce (Dake and De Ment 1949), as will amber (Verbeek, p.c. 1994).
Some activators are known to interact in what is called sensitization luminescence. In one variety, cascade fluorescence, the energy of one activator is immediately absorbed by a second activator, which fluoresces in turn. Energy may be transferred through emission and subsequent reabsorption, or in the form of vibrational energy transferred at ion bond sites. Pb2+ and Ce3+ are common sensitizers. (Marfunin 1979: 160-5; Waychunas 1988: 645)
Despite the presence of activators, many materials will still not fluoresce. The reason lies in compounds which block fluorescence, known variously as quenchers, inhibitors, and even poisons. Quenching refers to any process through which fluorescence is reduced or blocked.
Quenching processes may be static or dynamic. In static quenching, also known as complexational quenching, the normally fluorescent molecule has formed a "nonluminescent complex" (Schulman 1985) with an adjoining inhibitor molecule. Any energy that is absorbed during ultraviolet bombardment is lost through intersystem crossing while the molecule is still in its ground state; thus, excitation, and fluorescence, never occur. Elements such as mercury and bismuth are typical static quenchers (Schulman 1985). In dynamic quenching, also called collisional or diffusional quenching, the energy is lost after the normally fluorescent molecule has become excited, but again before fluorescence can occur. Through excitation energy transfer, the absorbed energy is passed to another molecule, which may then itself fluoresce. The degree of quenching of the original activator depends on the concentration of the quencher (McGown 1986: 73). Three variants exist. In resonance energy transfer (also known as the dipole mechanism) (Schulman 1985), the "donor" and "acceptor" molecules are separated by 10-6 cm. The donor acts as an electrical dipole with an associated field, and any acceptor molecules within range may absorb the energy, becoming excited in turn. In exchange energy transfer (Schulman 1985), the donor and acceptor are adjacent and their electrons can interchange. When donor electrons become excited, they join the acceptor in return for unexcited electrons. The absorbed energy also leaves the donor, at which point it returns to its ground state. Exchange energy transfer depends on the degree of contact between the two molecules and is thus temperature dependent. A more generalised form of exchange energy transfer is vibrational energy transfer, in which energy is lost as heat during physical interaction between neighbouring molecules. This mechanism is thermally dependent, and, at higher temperatures, is referred to as thermal quenching. As the temperature rises, so does the rate of molecular collisions. Whenever one of the excited molecules collides with one of its neighbours, energy is passed on. The reduction in intensity during thermal quenching has been estimated at one percent for every Celsius degree the temperature increases (Guilbault 1973). After a certain threshold, much above room temperature, the energy drain is generally sufficient to retain all the potential fluorescence energy within the crystal lattice and no fluorescence occurs. (Hurtubise 1986: 75-8; McGown 1986; Schulman 1985; Waychunas 1988: 647)
Oxygen is such a strong inhibitor that patination greatly reduces fluorescence even in its early stages (Radley and Grant 1959:269). Fe3+ is usually referred to as a poison and is the most common inhibitor. Iron can quench fluorescence in concentrations as low as 0.1% (Waychunas 1988: 647), and thus iron-rich rocks, such as jasper, will rarely fluoresce (Verbeek, p.c.). Co2+, Fe2+, and Ni2+ are weaker quenchers, and only take effect when their concentration reaches several percent (Waychunas 1988: 647). Some halogens and organic compounds can also act as inhibitors (Pringsheim and Vogel 1943: 70).
Typically, when the concentration of activators or sensitizers reaches a high enough level, self-quenching occurs. Self-quenching is common among the rare earth elements when they exceed one percent, and also with some transition metals, such as chromate (Cr3+) (Waychunas 1988: 647).
Although it is unlikely in archaeological applications, if the intensity of the activation energy is sufficient, an effect called photodecomposition may develop. Photodecomposition occurs when the energy that is absorbed is exceeds the dissociation energy of the molecule's chemical bonds. (Rendell 1987: 22)
With all this potential for variability, mineral specimens and rocks from one location may fluoresce, while those from another may not. Additionally, the presence of other impurities can alter the colour, reduce the intensity, or even block the effect entirely.
Fluorescent response is governed by the wavelength(s) of the applied light. Typically, shortwave (2300 to 3100 angstroms) and long wavelength (3100 to 3800 angstroms) light sources are used. Mineral response often depends on this wavelength; the colour may differ or the mineral may not react under one or both. However, typically shortwave light evokes the strongest and most frequent response. (Bush 1966; Radley and Grant 1959)
Ambient conditions and prior exposure to certain rays and chemicals can have an effect on fluorescence. For example, previous exposure to heat may alter colour and intensity (for an archaeological comment on this, see Hofman et al. 1991), but this effect is not well established. Bush (1966:238) notes that some minerals fluoresce only at extremely low (liquid oxygen) temperatures, while others show a marked increase in intensity (cryoluminescence) (see also McDougall 1952). Rendell (1987) explains that at reduced temperatures there is little vibrational motion, hence there is a corresponding reduction in the chance of energy loss due to molecular collision. A number of authors (McDougall 1952; McGown 1986; Pringsheim and Vogel 1943) note that an increase in ambient temperature generally reduces fluorescence intensity, and, when high enough (e.g 100 to 400 degrees Celsius (Pringsheim and Vogel 1943: 69), prevents it. By comparison, elevated temperatures significantly reduce the duration of phosphorescence by shortening the reaction time, while reduced temperatures ultimately extend the delay indefinately (Pringsheim and Vogel 1943: 12-13). Pressure increases will initially boost fluorescent intensity, up to a threshold, after which it will decline (Dake and De Ment 1941:60-1). Prior exposure to beta, gamma, and x-rays. Cathode and x-ray bombardment, for example, can make substances fluoresce that normally would not (Pringsheim and Vogel 1943: 69). Exposure to certain acids and alkalis may also activate inert substances (Radley and Grant 1959:364).
It will be necessary to view the responses of as many specimens from each resource as possible. Samples from different exposures, as well as those representing the full range of appearance and condition will be necessary in order to assess the full range of ultraviolet fluorescence which might occur, for it has long been recognised that fluorescent response can be extremely variable. Differences in response may relate to only slight variations in the concentration of trace element constituents. Apparently inert substances may be revealled as mildly fluorescent under stronger illumination or once the eyes are allowed to fully adjust to conditions. (Pringsheim and Vogel 1943:68)
A variety of ultraviolet light sources are commercially available, although even lamps of the same type do not necessarily produce the same output. Not only may their filaments emit somewhat different wavelengths, to block the visible light which is a side effect in the production of ultraviolet, manufacturers employ a variety of filters which also have dissimilar properties (Radley and Grant 1959:59). The spectra emitted by different types of lamp are even more different (Hurtubise 1986). The most popular mercury vapour lamps can have high intensities, but emit narrow bands of light, known as line emissions. Lower powered tungsten filament lamps have proportionally lower intensities. Xenon arc lamps have both a high intensity and a relatively even output spectrum. As mentioned previously, different wavelengths of excitation light may all produce some degree of fluorescence, but the intensity of fluorescence may vary significantly (Guilbault 1973). And, since any fluorescence is the sum of emissions from all activators that are present, producing a combined spectral emission, changes in any of the component emission intenisties will alter the overall colour. Thus, it is advisable to use the same source throughout a study, and to run comparative tests of different sources before drawing intensively on another's results.
Shortwave and longwave lights will also produce different results in some minerals. As a generalization, where shortwave light tends to produce purple-blue colours, longwave will produce orange-red colours. In other words, since the shortwave light has a shorter wavelength than the longwave, the resulting fluorescence will also tend to have a shorter wavelength (Stoke's Law).
In some specimens, bright colours may obscure other, paler, hues. However, for archaeological applications, this is not a problem, as we are primarily concerned with the overall effect. It is important to remember that human eyes are not all receptive to the same range of light wavelengths. Some eyes are not receptive to certain wavelengths (colourblindness) and, with age, they can lose sensitivity to the lower end of the visible spectrum (Dake and De Ment 1941:71,80). There are a number of known colour perception disorders. In dim light, the Purkinje Effect sees blues fade, reds vanish, but greens become enhanced. With high intensity, greens, green-yellows, and red-yellows become yellowish, while blue-greens become blue. The brain further adjusts so that colours become more complementary. And, in the Abney Effect, it is recommended that the observer spends at least several minutes in acclimatising to the darkness before viewing samples (Radley and Grant 1959:66). I have found that intermittent use of a small penlight allows notetaking in the darkroom without the necessity of destroying night vision by turning on the lights. Use of a red light such as in a photographic darkroom is also feasible.
The work area must also be considered. Around the specimen, non-reflecting surfaces, ideally such as black velvet, are recommended to limit interference. There should be a shield between the specimen and the observer, such as a piece of glass. (Regular eyeglasses work, and there are also inexpensive plastic safety glasses available that easily block most ultraviolet.) This serves two purposes. First, the human eye itself fluoresces under ultraviolet, with the result that reflected source light can create a false colour impression. Secondly, exposure to shortwave ultraviolet irritates and will eventually damage the eye, as occurs with snow-blindness. These wavelengths are blocked by glass and some clear plastics. There are also filters which block all or most ultraviolet, but they are not easily obtained. (Dake and De Ment 1941:71; cf. Radley and Grant 1959:59)
Certain details should be recorded during the analysis. For example, fluorescent intensity and even colour may vary depending on the proximity of the ultraviolet source. Similarly, the distance of the sample from the observer (or camera) will affect the apparent intensity. For consistency, these distances should be kept constant, and recorded along with the type of ultraviolet source. Since irradiation, including thermal alteration, can influence UVFA, it is critical to know the past history of all material analysed.
Since the brain is highly subjective with regard to colour, various devices have been developed over the years to measure the intensity and colour of fluorescence, such as fluorimeters, colorimeters and spectrophotometers. Other simpler devices may also be used, such as handheld photographic lightmeters (cf. Filippelli and Delaney 1992; Shockey 1993). Some of these devices operate by comparing the sample to artificial light. Others use photo-cells and actually measure the wavelength of the emitted light. (Dake and De Ment 1941; Radley and Grant 1959).
At present, there is no recognised colour standard for fluorescent light. The Fluorescent Mineral Society is in the process of developing such a standard, but the project is proceeding slowly. (FMS p.c. June? 1996). Several colour systems do exist, mostly in the world of arts, such as the Pantone and Admark colour systems. The main and critical failing of all such standards is their design for use under white light. It is awkward, and inaccurate to categorise fluorescent colours seen in what must be a dark viewing area with a colour chart which must properly be viewed under white light. A number of colour systems have been used by archaeologists, including the Pantone and Admark systems, as well as quite subjective systems such as the five generalised colour groups recommended by Wain: blue, orange-yellow-gold, green, red-pink, and white-cream (1965: 37). Robbins (1983) provides an list of 25 colours, but these tend to be even more subjective, such as "butter yellow" and "deep sky blue". Earl Verbeek, ex of the U.S.G.S., recommends that in the absence of an ideal standard, a generalised system be used, in which colours are described as clearly as possible, and confusing or exotic terms, such as hazel or turquoise are avoided (p.c. November 1993). He describes three attributes, namely brightness (the fluorescent intensity), saturation (colour richness or how much white is mixed with the basic color, ranked pale, medium, deep), and hue (the basic shade, such as red, orange, blue, green, etc.) (Verbeek p.c. November 1993). An example would be weak medium blue-green. I have that this system is the most useful, although intensity is very difficult to measure subjectively. I have found it very useful to compile actual examples of each colour hue category, which I have glued to a small card. This tool is very valuable in giving consistent colour descriptions of samples.
It is possible to photograph samples on colour film. This not only aids in publication and presentation of results, but also provides a means of recording results. Because regular film is extremely sensitive to ultraviolet light, it is critical that light reflected from the sample or transmitted directly from the source does not reach the film. Use of a filter which only transmits visible light (4000 to 7000 A) is recommended (Dake and De Ment (1941:81). While a variety of different filters exist (such as the Corning Noviol C, Ilford Q, and Kodak Wratten 2A, 2B, and K2 filters), these do not necessarily exclude all ultraviolet. Further, some of these filters may exclude visible wavelengths which will add enormously to the required exposue time (Stoll 1954). Films such as Ektachrome 160 Tungsten are available which are insensitive to wavelengths below 4000 angstroms (ultraviolet). Even with this film, infrared contamination might still occur, but these wavelengths should be detectable as heat or with an infrared detector. The lens and lamp must also be situated carefully to minimise exposure to reflected visible light that typically leaks past the lamp's filters (Bowen 1947; Stoll 1954). This is best accomplished by placing the lamp(s) at 45 degrees to the sight line from camera to subject. Robbins (1983: ix) had success with Kodachrome 64 film, and exposures between 7 and 50 seconds at f/4.5. Pitt and Pitt (1993) used Kodak Plus-X Pan film and a yellow filter to photograph fossils under longwave light. They shot at f/8 for two minutes. Another researcher (Tankersley p.c. 1994) has had excellent luck using Ektachrome 200 film, a Blue 40 filter, and exposures of circa two seconds at f/16. This illustrates that some experimentation will initially be necessary because different UVFA wavelengths and intensities may require different exposures; red colours require a longer exposure than do blue.
In possibly the earliest employment of UVFA in lithic research, Colin Renfrew and J.S. Peacey (1968) tried to distinguish several Aegean marble sources based partly on their fluorescent colour, but met with only limited success. UVFA has further been employed in lithic use-wear analysis, for detection of organic residues (Broderick 1979; Ocampo p.c. October 1993). Kinnunen et al. (1985) include UVFA responses in their overview of Finnish lithic resources.
For the North American continent, application has been limited to the Plains and Southwest. Larry Banks (1990: 8) mentions using UVFA, apparently in the late 1970's, to unsuccessfully differentiate similar Wyoming cherts. Banks (1990: 8), in his opus on lithic materials of the south central states, briefly mentions "the potentially extraordinary implications" of UVFA (abbreviated UVFL). Tim Church (1990) recorded UVFA data on an extensive collection of northern Plains materials housed at the State Archaeological Research Center in Rapid City, South Dakota. In Texas, Michael Collins compared the fluorescence of Edwards Plateau chert lookalikes (Collins and Headrick (in press) mentioned in William et al. 1992: 72). Following Collins' lead, Banks and Roger Bowers made another attempt on Edwards Plateau chert lookalikes, this time with some success. Hofman, Todd, and Collins (1991, esp. 297) used UVFA to source artifacts from the Folsom and Lindenmeier palaeoindian Sites, finding it "reliable, expedient, and inexpensive". They further list the UVFA responses for a series of other rocks. Mathew Hillsman (1992) studied central Texas and eastern New Mexico chert look-a-likes, finding UVFA better than visible characteristics for distinguishing between the rocks, so long as the objects were not small flakes. With a spectrometer, he was able to clearly distinguish some of the look-a-likes. Mark Luther (p.c. September 1992; Bluemle 1993: 5), of the North Dakota Geological Survey, has been begun to assess the range of response of many North Dakota materials. Church has a second extensive study underway at Fort Bliss, in Texas. Don Shockey (1993) sourced artifacts from the Woodard-Benefield Site, in Oklahoma, by comparison to reference samples of Boone and Keokuk cherts. And finally, new research planned by Mariah Associates will attempt to distinguish between different chert types, while assessing some aspects of intra-formational variation and the complicating effects of thermal alteration. The above studies exemplify the use of UVFA as a means of distinguishing between different lithic materials to allow sourcing of artifacts.
A number of researchers have addressed procedural concerns. Of these, the classification of fluorescent colour is typically foremost. It is generally agreed that a standardised system is needed, but to date a number of schemes have been used. Hofman, Todd, and Collins (1991) relied on the generalised colour groups suggested by Wain (1965: 37), namely orange-yellow-gold, green, blue, red-pink, and white-cream. Luther recommends the Pantone Color Sytem which is a standard in the printing industry (p.c. September 1992). Robert Christensen tested systems, finally recommending the Pantone Color System (Williams et al. 1992: 79). Church (p.c. January 1994) has used generalised, subjective colour categories, while Shockey (1993) used a system designed by AdMark. Hillsman (1992) used a subjective classification scheme based on combinations of green, yellow, orange, brown, red, blue, white, grey, and purple.
Another concern addressed has been the quantification of fluorescence colour and intensity. Shockey (1993) measured fluorescent intensity using a Soligar photographic lightmeter, with mixed success. Hillsman (1992) used a spectrometer to measure the visible wavelength response to longwave ultraviolet. The previously mentioned glass study measured fluorescent responses from the ultraviolet to infrared wavelengths with a Perkin-Elmer 330 spectrophotometer (Green and Hart 1987).
A variety of ultraviolet sources have been used, but the favourites, by far, are those made by Raytech. Raytech lamps have been used by the Oklahoma Archaeological Survey, William et al. (1992) and Robert Christensen (William et al. 1992: 79). Shockey (1993) used a Raytech PLS60 or a PP-FLS lamp, while Church (p.c. January 1994) has used a Versalume (PP-FLS) lamp. Hillsman (1992) used a Raytech LS-88CB lamp. In this study, a Raytech hand-held Versalume (PP-FLS 10-020) and a larger, stationary Model 218 were used.
Finally, certain archaeologists note complicating factors. Bob Elston, of Intermountain Research, notes that surface samples tend to have a reduced response, possibly due to exposure to sunlight, and that heating in a reducing environment produces an orange fluorescence (p.c. January 1994). Britt Bousman (p.c. November 1993), of the Texas Archaeological Research Laboratory in Austin, notes that thermally altered samples of Edwards Plateau chert had reduced fluorescence. Tim Church (p.c. January 1994), at Fort Bliss, notes varying colour depending on the light source used. He also recommends differentiating matrix fluorescence from that of any patina or cortex, and presumably inclusions. Hofman et al. (1991) note that thermal alteration, rock texture and even mineral grain orientation may influence the fluorescent response. Hofman (p.c. February 1994) notes that development of a "good control collection" will be necessary to prevent error due to phenomenon such as thermal alteration, cortex, and weathering, as well as the normal internal variation in each rock.
On a different plane, UVFA has potential value for lithic technologists and dating purposes. Hofman et al. (1991: 306) note the capacity for UVFA to highlight subtle or invisible differences in the degree of patination due to "time gaps between flake removals" on artifacts. This quality allows insight into technological realms such as curation and recycling of older artifacts. Mariah Associates (Mike Quigg p.c. November 1993), of Austin, Texas, have been using UVFA to discern areas of retouch, and attempting to sort out problems related to mixing of zones on sites. William et al. (1992) tested the potential of UVFA for the relative dating of Knife River flint artifacts. They announcing a correlation between the age of artifacts and the intensity of fluorescence, but unfortunately did not determine the mechanism. They used the Pantone Color System. A planned study by Jack Jackson (p.c. November 1993) of Fort Hood, Texas, will make use of UVFA during an evaluation of patina thickness for absolute dating. Another planned by Mariah Associates (1993) will examine the applicability of UVFA toward relative dating of patinated artifacts, and address the effects of thermal alteration on fluorescence.
UVFA analysis is beginning to be used in general archaeological research, where, for example, Bluemle (1993: 5) notes fluorescence colour as part of a general description of Knife River Flint. Several CRM companies, such as Intermountain Research (R. Elston p.c. January 1994) of Silver City, Nevada, PIII (D. Zeanah p.c. November 1993) of Salt Lake City, Utah, and Mariah Associates, of Austin, Texas, plus the Oklahoma Archeological Survey (R. Drass p.c. November 1993), Texas Archaeological Research Laboratory (B. Bousman p.c. November 1993), and archaeologists at Witchita State University (D. Hughes p.c. November 1993) routinely use it to sort look-a-likes (see above). Typically, no colour system is used for this work.
The materials examined were all North American lithic resources picked arbitrarily from the John D. Holland Lithic Laboratory collection at the Buffalo Museum of Science. These samples were chosen to include a wide range of different rock types, as well as material from across North America. Included in the sample set were several series of macroscopic similar "look-a-likes". The rocks and their UVFA reactions discussed below are not intended to be a preliminary list to indicated the range of responses one might expect form different types of rocks. As a number of pioneers of this technique have noted, intensive study of each lithic resource will be necessary before generalisations can be safely made.
It has been noted in the literature (Radley and Grant 1954) that the study of UVFA in rocks is a complex and highly subjective past-time. To compensate for the lack of a proper colour standard, and to acquire more quantitative data, some have advocated the use of spectrometers (cf. Hillsman 1992). While it is true that they do have the potential to achieve an objective accounting of the fluorescent response, the average archaeologist does not have direct access to such equipment. Therefore, this study suggests the alternative approach of developing a comprehensive comparative collection, something that can be done by all archaeologists on the appropriate local, regional, or broader scale.
Both naturally patinated and fresh surfaces were examined, when possible, for each specimen. These surfaces must truly be fresh, as coatings may be present on surfaces that appear to be newly exposed. Samples were studied separately under short and longwave light. Ultraviolet-blocking, plastic safety glasses were worn at all times. A large number of samples was first examined in order to establish the range of colour and intensity. Sample colour response was recorded based on the system recommended by Verbeek (see above).
Earl Verbeek, with the U.S. Geological Survey, recommends that in the absence of an ideal standard system, a generalised system should be used, in which colours are described as clearly as possible, and confusing or exotic terms, such as hazel or turquoise are avoided (p.c. November 1993). He describes three attributes, namely brightness (the fluorescent intensity), saturation (colour richness or how much white is mixed with the basic color, ranked pale, medium, deep), and hue (the basic shade, such as red, oramge, blue, green, etc.) (Verbeek p.c. November 1993). An example would be weak-medium-blue-green. This approach has shown itself to be the most practical, although intensity proved to be so difficult to quantify subjectively that it was dropped as a category. (However, once more intensive studies of individual resources are conducted, it is recommedned that range of intensity be recorded.) I have found it very useful to compile actual examples of each colour hue category, which I have glued to a small card. This tool is very valuable in giving consistent colour descriptions of samples. The use of such a control collection has subsequently been recommedend by Jack Hofman (p.c., February 1994).
Apparently inert samples were coded as "non-fluorescent" (nf), with the background hue, typically "purple-black" in the case of dark-coloured rocks, or "purple-white" for light-coloured rocks. A weak purple response is likely reflected violet light that has leaked past the filter on the ultraviolet source lamp, however, reflected light may be concealing weak fluorescence. To check for this problem, put a pane of glass between the sample and light. If the sample colour changes or dims, the response observed is at least partly fluorescence (glass only affects shortwave). A second technique involves looking for phosphorescence. If the sample glows after the lamp is shut off, it is indeed fluorescent. (Verbeek p.c. November 1993)
Once the range of colours is assessed, I will experiment with taking colour photos of each active specimen.
Included in the study were several groups of look-a-likes (Table 2). Only with some of the groups in Table 2 .... was UVFA successful at differentiation.
UVFA colour response varieties such as orange-brown, yellow-brown, green, yellow-green, brown, and dirty-white were most commonly recorded. Intensity ranged from pale to bright, but these terms were seldom used as they are highly subjective, and most samples had some degree of patination present which masked the more brilliant response. Non-fluorescent samples appeared to range from purple-black to a dirty-purple-grey, correlating with how dark the original sample colour was in visible light (i.e. translucency): light-coloured specimens typically were a purple-grey when not fluorescing another colour, while dark specimens were purple-brown to purple-black under ultraviolet light.
In addition, samples were normally heterogenous in colour. Often mottling and banding were present, either as varying shades of one dominant colour, or as regions of different colours. At times this pattern mirrored that seen in normal light, but different patterns were not unusual. As with the description of rocks in visible light, fluorescence description proved rather difficult at times due to this heterogeneity. Samples were described as mottled or banded (concentric or linear), and inclusions were noted. Additionally, mineral inclusions, such as dendrites, veins, vugs, and fossils, often fluoresce, sometimes quite brilliantly (cf. Rolfe 1965). These phenomena may prove useful in resource characterisation; however, caution is advised as there is no guarantee that an artifact from a particular resource will bear all the attributes of that resource.
The colour response of rocks may correlate with their form of diagenesis. Hillsman (1992) notes that the impurity content of rocks, including that of the fluorescent activators and inhibitors, is highly dependent on the mode of deposition. Secondary replacement cherts, for example, tend to be quite impure. The nature of their impurities will naturally control the presence and nature of any fluorescence. And, for my rocks.....?
Fluorescence was visibly inhibited on some samples. Weathering was the most obvious correlate, apparently reducing the fluorescent intensity of some samples or even blocking it entirely. Also, regions that had been coated with clear nail polish for labelling purposes were inert.
At the same time, a variety of fluorescent contaminants were noted, including lichens, lint, and surface washes. Lichens are known to fluoresce, indeed, longwave ultraviolet is typically used by biochemists to detect the complex organic acids found in lichens. Their UVFA can range from a bright white to a bluish or greenish white, or a bright orange (Hale 1974: 121). Small patches of lichen, circa 0.5 to 3 mm in diameter, were found on a number of samples and glowed bright orange under uv. Lint often glows bright white due to the presence in many fabrics of commercial brighteners (added in the manufacturing process and in detergents and colour-safe bleach). Lastly, some surfaces of samples, which appeared to be very fresh under daylight, were actually coated with what showed as a reddish or orange wash under ultraviolet light. This coating is possibly caliche. Caliche is a mixture of mostly calcite, plus organic materials and clay minerals, which is a quite common feature in arid climes usually on the undersides of rocks. Caliche is the evaporites left behind when ground water drawn to the surface by capillary action evaporates (Verbeek, personal communication 1994). According to Leudtke (1994 personal communication), calcium deposits on the undersides of artifacts are also a not uncommon phenomenon in New England shell middens. The observed red-orange response may relate to the presence of manganese in the calcite, possibly sensitized by lead (Waychunas 1988: 655). Washing Garden City (Edwards variety) chert with 30% hydrochloric acid failed to remove it. Also Verbeek (pers comm) warns that exposure to even weak acid solutions may alter some minerals, so more severe washing may not be desirable. Overall, these contaminants are easily detected, and providing the analyst is wary and sample preparation is adequate, erroneous results should be avoidable.
Certain generalisations can now be made.
"In the early days the results were so encouraging that (UVFA) was hailed as a rapid, accurate and reproducible method, and for many purposes, indispensable to the analyst. Maturer consideration showed that the accuracy is limited, and the reproducibility is dependent on strict standardisation of working conditions.... The conclusion to be drawn... is, that if applied with discretion and under standard conditions, fluorescence analysis is a most valuable aid to the scientific worker, especially in routine work or sorting tests...."
This quotation stresses important points. Ultraviolet fluorescence is a potentially very valuable tool for the lithic analyst. However, it is crucial that it be used in a careful and controlled manner. While it is tempting to shine a ultraviolet lamp on a tray of artifacts to sort out certain types of material, the full range of behaviour for each lithic resource type must be assessed and the potential for alteration in response due to heating, weathering, and surface contaminants, and differences in light sources must be appreciated before a source can be legitimately ascribed.
This study has shown that a wide range of response can be expected for prehistorically utilised lithic resources, but the responses listed are not, however, meant to be the final values for these resources. Researchers will have to investigate resources further on their own. Yet the results of this study do provide some insight into what may be expected from such research.
Banks, Larry D.
1990 From Mountain Peaks to Alligator Stomachs:
A Review of Lithic Sources in the Trans-Mississippi South,
the Southern Plains, and Adjacent Southwest. Memoir #4.
Norman, OK: Oklahoma Anthropological Society.
Becker, Ralph
1968 Theory and Interpretation of Fluorescence
and Phosphorescence. New York: Wiley Interscience.
Becquerel, Edward
1848 Ann chim phys 22: 244
Bluemle, John
1993 State rocks, minerals, and fossils: a
state rock for North Dakota? North Dakota Geological
Survey Newsletter 20(1): 4-6.
Bowen, E.J.
1947 Fluorimetric analysis: fluorimeter design.
The Analyst 72(858): 377-379.
Broderick, Michael
1979 Ascending paper chromatographic technique
in archaeology. In Brian Hayden (ed.) Lithic Use-Wear
Analysis. New York: Academic Press.
Bush, Edward A., Jr.
1966 Mineral fluorescence and ultraviolet
radiation: a review. The Compass 43(4): 237-41.
Church, Tim
1990 An investigation into prehistoric lithic
procurement in the Bearlodge Mountains, Wyoming.
Unpublished M.A. Thesis. University of Montana.
Collins, Michael B. and P. Headrick
(supposed to be in press as of 1995???)
Fluorescent Properties of "Edwards" and Related Cherts
of Central Texas. Ms. submitted to Journal of Field
Archaeology.
Crooks, W.
1881 Proceedings of the Royal Society 32:206
1883 Proceedings of the Royal Society of London
35: 262.
Dake, H.C. and J. De Ment
1941 Fluorescent Light and its Applications. New
York: Chemical Publishing Company.
Englehardt, E.
1912 "Luminescenzerscheinungen im ultravioletten
Lichte" Unpublished Dissertation. University of Jena.
Eyman, Charles E.
1965 Ultraviolet fluorescence as a means of
skeletal identification. American Antiquity 31(1): 109-112.
Filippelli, Gabriel M. and Margaret L. Delaney
1992 Quantifying cathodoluminescent intensity
with an on-line camera and exposure meter. Journal of
Sedimentary Petrology 62(4): 724-5
Goethe, Johann W. von
1810
Goldstein, E.
1876 Wren Ber 80 (II): 151
Guilbault, George (ed.)
1973 Practical Fluorescence: Theory, Method,
and Techniques. New York: M. Dekker.
Gunnell, E. Mitchell
1933 The photo-luminescence of Illinois
fluorite and certain zinc minerals and associated species
from the Joplin, Missouri, district. America Mineralogist
18: 68-73
Hale, Mason E., Jr.
1974 The Biology of Lichens. New York: American
Elsevier Publishing Company.
Hillsman, Mathew J.
1992 Evaluation of Visible and Ultraviolet-
Excited Attributes of Some Texas and Macroscopically
Similar New Mexico Cherts. Unpublished M.A. Thesis.
Eastern New Mexico University.
Hofman, J.L., L.C. Todd, and M.B. Collins
1991 Identification of central Texas Edwards
chert at the Folsom and Lindenmeier Sites. Plains
Anthropologist 36(137): 297-308.
Hopkins, Delmont N.
1990 Determination of uranium in geologic
materials by UV fluorescence. U.S.G.S. Open File Report
90-668: 111-3.
Hurtubise, Robert J.
1986 Trace analysis by luminescence
spectroscopy. In G.D. Christian and J.B. Cullis (eds.)
Trace Analysis: Spectroscopic Methods for Molecules.
New York: J. Wiley.
Kempe, D.R.C. and Anthony P. Harvey
1983 The Petrology of Archaeological
Artefacts. Oxford: Clarendon Press.
Kinnunen, Kari, R. Tynni, K. Hokkanen, and J.-P.
Taavitsainen
1985 Flint Raw Materials of Prehistoric
Finland: Rock Types, Surface Textures and Microfossils.
Geological Survey of Finland Bulletin 334.
Kunz, George F. and Charles Baskerville
1903 The action of radium, roentgen rays and
the ultra-violet light on minerals and gems. Science
n.s. 18: 669-783.
Lutati, F.V.
1930 Index Chimica 5 (15): 1222
Mariah Associates
1993 Proposal for UV Fluorescence Study.
Unpublished manuscript in possession of author.
Marfunin, A.S.
1979 Spectroscopy, Luminescence and
Radiation Centers in Minerals. Translated by V.V.
Schiffer. Berlin: Springer-Verlag.
McDougall, D.J.
1952 Some factors influencing fluorescence
in minerals. American Mineralogist 37(5-6): 427-437.
McGown, Linda
1986 Molecular fluorescnce spectroscopy. In
Ruth E. Whan (ed.) Metals Handbook 9th Edition Volume
10. Metals Park, OH: American Society for Metals.
McKern, T.W.
1958 The use of short wave ultra-violet rays
for the segregation of commingled skeletal remains.
Environmental Protection Research Division Technical
Report EP-98. Natick: U.S. Army Quartermaster Research
Engineering Command.
Miethe, A.
1927a Uber die Photographie von Fossilien bei
ihrem eigenen Fluoreszenzlicht. Photogr. Korrespondenz
63: 69-70.
1927b Fossilien-Photographie; Koralle. Magazin
alle Freunde von Natur Technik 3: 145-7.
Morse, H.W.
1905 Astrophysics Journal 21: 85, 410
Murata, K.J. and R.L. Smith
1946 Manganese and lead as coactivators of
red fluorescence in halite. American Mineralogist 31:
527-538.
Pitt, William D. and Lois J. Pitt
1993 Ultra-violet light as a useful tool for
identifying fossil mollusks, with examples from the
Gatun Formation, Panama. Tulane Studies in Geology and
Paleontology 26(1): 1-15.
Pringsheim, Peter and Marcel Vogel
1943 Luminescence of Liquids and Solids and
its Practical Applications. New York: Interscience.
Radley, J.A. and J. Grant
1959 Fluorescence Analysis in Ultra-violet
Light. London: Chapman and Hall.
Rendell, David
1987 Fluorescence and Phosphorescence
Spectroscopy. Chichester: John Wiley.
Renfrew, Colin, and J.S. Peacey
1968 Aegean marble: a petrological study.
Annual of the British School in Athens 63: 56-66.
Riecker, Robert E.
1959 Geologic applications of fluorescence.
The Compass 32(2): 85-93.
Riordan W.M.
1991 Detection of nonvisible writings by
infrared luminescence and ultraviolet fluorescence.
Journal of Forensic Sciences 36(2): 466-9.
Robbins, Manuel A.
1983 The Collector's Book of Fluorescent
Minerals. New York: Van Nostrand Reinhold.
Rock-Color Chart Committee
1984 Rock-Color Chart. Boulder, CO: Geological
Society of America.
Rolfe, W.D. Ian
1965 Uses of ultraviolet rays. In Bernhard
Kummel and David Raup (eds.) Handbook of Paleontological
Techniques. San Francisco: W.H. Freeman.
Schulman, Stephen G.
1985 Luminescence spectroscopy: an overview.
In S.G. Schulman (ed.) Molecular Luminescence
Spectroscopy. New York: J. Wiley.
Shockey, Don
1993 Ultraviolet tests and the Woodard-
Benefield Site. Oklahoma Anthropological Society
Newsletter 41(9): 5-7.
Simpson, G.G.
1926 Are Dromatherium and Microoconodon
mammals? Science 63: 548-9.
Stanciu, Victor
1937 Die Bestimmung des Fundortes der
Gesteine und Mineralien durch UV-Strahlen. Cluj
University, Mus. Geol.-Miner., Revista, Vol. 6: 321-7.
Stokes, C.G.
1852 Phil. Trans. 143: 463.
Tankersley, Ken and John Meinhart
1982 Physical and structural properties of
ceramic materials utilized by a Fort Ancient group.
Midcontinental Journal of Archaeology 7(2): 225-243.
Verneuil, A.
1886 Comptes rendu Academie de Science 103: 600
Wagner, Ernst
1928 Zur Fluorgraphie von Fossilien; Ein
neues und einwandfreies Glasfilter. Palaont. Zeitschrift
10.
Wain, Harry C.
1965 The Story of Fluorescence. Middlefield,
CT: Raytech Equipment Co.
Ward, T.W.
1935 Mineralogist 3:19
Waychunas, Glenn A.
1988 Luminescence, x-ray emission and new
spectroscopies. In F. C. Hawthorne (ed.) Spectroscopic
Methods in Mineralogy and Geology. Reviews in Mineralogy
Volume 18. Mineralogical Society of America.
William, J.D., E.R. Hajic, S.A. Ahler, and J.L. Overturf
1992 "Prehistoric Site Investigations:
Ultraviolet Fluorescence Analysis." In Stanley A. Ahler
(ed.) Phase II Cultural Resource Investigations Asscoiated
With Proposed Dam Safety Modifications at Lake Ilo
National Wildlife Refuge, Dunn County, North Dakota.
Grand Forks, ND: Department of Anthropology, University
of North Dakota.
Yagoda, Herman
1946 Luminescent phenomena as aides in the
localization of minerals in polished sections. Economic
Geology 41: 813-19.
res. ph: 716-873-8489 dept fax: 716-645-3808 hjarvis@acsu.buffalo.edu