Phosphors are usually white in appearance and exhibit luminescence when excited. Nowadays they have wide range of applications from CRT tubing, plasma displays. light bulbs and x-ray conversion screens. Alchemists were the first to synthesize luminescent materials, mainly by accident in their attempts to make gold [3, 5]. In 1603, Vincenzo Cascariolo created a material that glowed purple at night having been exposed to sunlight during the day. Later, La Galla in 1612 wrote the first publication on synthetic luminescent material. Another important publication in 1640 termed the word “phosphor” to mean any ‘microcrystalline solid luminescent material’. To distinguish it from the element phosphorous that was later discovered in 1669, long-lived luminescence became known as “phosphorescence” [5]. The synthesised phosphor was probably barium sulphide with a low efficiency. A more stable phosphor was synthesised in 1866 by Theodore Sidot by heating zinc-oxide in a stream of hydrogen sulphide. Soon it was known that these sulphides do not luminance in their pure state, but do when they contain small quantities of activators.

In the 18^th and 19^th centuries, phosphors were mainly used for detecting invisible particles (UV photons, cathode rays, x-rays and alpha particles) [5]. At this time, with many concurrent advances in other scientific fields such as vacuum science, ceramics, glass working, and electromagnetism, Karl Ferdinand Braun introduced the idea of the cathode ray tube in 1897 and won the Nobel Prize in Physics in 1909 for his contributions [7]. After the introduction of the fluorescent lamp by GEC in 1938, the demand for efficient lighting increased. The need for better CRTs and more efficient lighting accelerated research into properties of phosphors and luminescence.

During the 19^th century, Phillip Lenard and co workers synthesised phosphors by firing metallic/rare earth ions impurities (activators) that formed luminescent centres in the host [8]. P.W. Pohl and F. Sietz introduced the configurational coordinate model of luminescence centres, establishing the basis of modern-day luminescence physics [8]. Another figure that helped us understand luminescence is Alexander Jablonski whose work resulted in the Jablonski energy diagram, a tool that can be used to explain the kinetics and spectra of fluorescence, phosphorescence, and delayed fluorescence. Other key figures include Frank Condon and Fonger and Struck, who have helped us to understand certain temporal properties of luminescence.

Phosphors are thermographic if they exhibit emission-changing characteristics with temperature. The idea of emission analysis for sensing technology is not new. For the case of pressure measurements, the Stern-Volmer relationship between emission intensity and air pressure dates back to as early as 1919. The idea of using phosphors for temperature measurement was first cited in 1937 [9] during the development of the fluorescent lamp where a loss of luminescence was observed with increasing temperature. Although Neubert suggested the idea, it was not until the early 1950's when Urbach and Bradley, cited in Allison and Gillies [2], first utilised phosphors to obtain temperature distributions on a flat wedge. Czysz and Dixson carried out tests producing thermal maps of models in wind tunnels [10, 11]. By applying a phosphor at the tip of an optical fibre, Wickersheim and co-workers investigated many phosphors with a variety of applications and their work led to the commercialisation of fluorescence-based thermometry systems [12]. Grattan and associates also investigated a variety of fibre tip thermometry systems based on a variety of phosphors. Cates [13] and co-workers developed remote measurement systems by adhering a layer of phosphor onto a surface of interest rather than at the end of a fibre [14], offering the system greater flexibility and allowing remote measurements of moving surfaces.

The capture and analysis of fast pulses was mainly nuclear and particle physics territory and required very expensive and sophisticated instrumentation. Over the past two decades, advances in technology, electro-optics and electronics have opened up new techniques and possibilities. Affordable, new detection systems that allow time resolution measurements in the picosecond regime, and newer short pulsed laser systems with increasingly higher pulse energies are being used to devise more elaborate systems that enable greater accuracy and range, opening up newer application areas. Due to the sensitive emission profiles, spatial resolution and high specificity of luminescence, florescence spectroscopy is rapidly becoming an important tool in sensing technology, spanning across many scientific disciplines. It is now very common for biomedical and aerospace applications for detecting oxygen levels, pressure, temperature and even cancer cells.

There are many organisations, joint collaborations and universities that are advancing the field of phosphor thermometry. Examples include Oak Ridge National Laboratory, NASA, Southside Thermal Sciences, Rolls Royce, Pratt and Whitney, University of Lund, University of Manchester, Imperial College and Cranfield University; with key researchers that include: S. Allison, S. Alaruri, B. Noel, M. Cates, G. Gillies, D.L Beshears. A.L Heyes, Seefeldt, J. Feist, T. Bencic and A. Omrane.

This section introduces the fundamental physics of luminescence, later specialising into the luminescence in phosphors. It will attempt to explain various responses that change with temperature, giving phosphors their sensing properties. It will start off with the Jablonski diagram which explains luminescence in general, and later moves on to the configurational coordinate diagram and the charge transfer curve model that helps in the understanding of the sensing properties of thermographic phosphors.

Luminescent processes are governed by a few important events that occur on timescales orders of magnitude apart. In general, excitation causes the energy of luminescent molecules to jump to higher electronic states. However, the configuration does not permanently remain excited. Vibration relaxation, internal conversion, intersystem crossing and emissions soon follow, resulting in the excited state returning back to the ground or an intermediate state. This process can be neatly summarised with a Jablonski energy-level diagram (Figure 1).

For any particular molecule, several electronic states exist. There are a combination of different available orbits (singlet states – S₀, S₁, S₂, ) and spin orientations (triplet/intermediate states – T₁, T₂), represented by thick lines, that are further divided into a number of vibrational and rotational energy levels, represented by the thinner lines in Figure 1.

Excitation (e.g. S₀ to S₁, S₂) involves the absorption of sufficient energy to raise a molecule's electrons into electronic states of S₁ or S₂. This molecule does not remain excited continually. According to Bell et al. [15], the ground state (S_o) is the only stable state with all other states decaying back to this state. According to the conversation of energy principle, the amount of energy absorbed must be released. This happens via:

emissions of photons equal to the energy-level difference

These energy transfers are further detailed as follows, with typical timescales summarised in Table 2.

Phosphors can take a number of forms usually consisting of a host material/matrix doped with activator atoms. Many of the materials that fluoresce efficiently are those that originate from a deliberately added impurity [2]. The added activator atoms are usually rare earth (lanthanides) ions or transition metals, seen in Table 3. Other luminescence centres include antinides, heavy metals, electron-hole centres and ZnS-type semiconductors. Thermographic phosphors for high temperature application usually have rare-earth ion centres in ceramic hosts.

Example hosts include: -

Yttrium garnets e.g. Y₃(Al,Ga)5012 , YAG

The potential number of phosphor combinations is very large.

Lanthanide ions, found in the 6^th period of the periodic table, are characterised by an incomplete 4f shell that is shielded from the effects of the crystal lattice by outer filled shells. Therefore, when a rare earth is mixed into a host lattice in low concentrations it can be treated as a free ion [16]. An example described in Allison and Gillies [2] is that the host material Al₂O₃ is transparent and non-fluorescent until Cr³⁺ is added. Luminescent centres are said to be isolated if the dopant concentrations are a few percent [2]. Although this is the case, according to Heyes [16], the host lattice has a profound effect on the thermal response of the phosphor. The influence on the processes of absorption and emission can be explained with the aid of a configurational coordinate diagram (Figure 3). The environment of a luminescent centre is not static and the diagram shows the potential energy curves as a function of a configuration coordinate (deviation from the ion equilibrium distance). Although the model is very simplistic and the shapes of the curves are not parabolic in reality, it shares many features from the Jablonski diagram (Figure 1), and can illustrate several physical phenomena including Stokes Shift. In addition it can also illustrate:

Absorption and emission band widths

Like the Jablonski diagram (Figure 1) energy potentials and vibrational energy levels are represented by horizontal lines; similarly, absorption and emission transitions are indicated by vertical lines. After excitation, electrons occupying an upper vibrational level of an excited state (point B) will relax to the ground vibrational level of that state (C) losing energy via the release of phonons [16]. Following radiative emission, the electrons reaching a higher vibrational level of the ground state (D) will further lose energy (phonons) on their return to their ground state equilibrium (A). The difference in excitation and emission energy levels can be seen in the diagram illustrating Stokes Shift.

The Frank-Condon principle states that electronic state transition times are much shorter than vibrational relaxation and therefore assumed to occur in static conditions. Based on this, excitation occurs to vibrationally excited levels of the excited electronic state. According to Royer [19], emissions occur from the lowest vibrational level of the excited state, because relaxation from excited vibrational states is much faster than emission.

According to Heyes [16], at temperatures above 0 K, electrons are distributed over different vibrational levels according to the Boltzmann's law.

(1)nexcited=nground(-ΔEkT)where ‘n’ is the electron population at a given state; ‘E’ is the energy difference between these states; ‘k’ is the Boltzmann constant and ‘T’ is the temperature.

If the temperature is high enough, electrons in the excited state can intersect the ground state curve (point E) allowing vibrational relaxation via phonon release to the ground state without any radiative emission. Ranson [20] describes this as the absorption of thermal energy (phonon) from point C, which excites the electrons to the intersection point E. Since non-radiative processes can now also take place, the observed luminescence intensity from a large quantity of excited ions will diminish, explaining the thermal quenching behaviour that is observed for most thermographic phosphors.

When the temperature is elevated, electrons are spread over a number of vibrational levels in the excited state. Since radiative transitions that can take place between any of the vibrational states in the excited and ground states, a broadening of the of the emission lines is expected [16].

Photo excitation alone can sometimes promote electrons into high vibrational levels at points beyond the intersection point (E) which results in a purely non-radiative emission, with no luminescence being observed. In some cases this may explain why higher energy photons (lower wavelengths) can actually dampen observed luminescence.

An add-on to the configurational coordinate model that explains quenching behaviour in different host materials is proposed by Fonger and Struck [21]. According to these authors, the outer crystal field, which is highly dependent on the chosen host, causes another energy potential (charge transfer state) that can be added on to the existing configuration coordinate diagram (Figure 3). Excited electrons can now return to the ground state via the charge transfer (CT) curve. Suppose an excited electron reaches an excited state of E3; it would normally return to the ground state by radiative emission. However, if the electrons are further excited by elevated temperatures (thermal activation), the electrons can intersect the crossover point of the CT curve, enabling the transfer of electrons to a lower energy level of E2 without any radiative emission. Likewise, electrons in E2 or E1 states can transfer its energy to the ground state in the same way. Different hosts will have the CT curve in slightly different places, thus explaining the different behaviour from various hosts.

When a continuous light source is used to excite the phosphor, electron populations are constantly being excited to higher states and returning back to their ground states. An equilibrium level is usually reached, indicated by a steady level of emission intensity. If the temperature is large enough, then probability of deactivation via a non-radiative process is more likely; this will be observed in a reduction in intensity. Various authors have investigated the effects of temperature on intensity for various phosphors and their emissions lines, and this has shown to be true for most cases. Figure 7 shows an example of intensity variations of some emission lines of La₂O₂S:Eu phosphor [2]. Most emission lines show a decrease in intensity with temperature. However, there are some emission lines where there is an increase in intensity over a certain temperature range. This may be due to increases in absorption at that wavelength. The will be explained further in section 5.5.

By calibrating the intensity response over a temperature range, temperature measurements can be made. A complete 2D acquisition can be achieved using CCD cameras with each pixel serving as a separate sensor. For a 1MP CCD, 1 million points can be monitored.

A common problem with intensity based techniques is that the observed intensity is also a function of other variables. If they are not taken into account, large errors can remain. Examples of such factors include: non-homogenous illumination, light source instabilities, phosphor coating thickness and densities, distance and detector viewing angle, surface curvature, reflections and shadings. These problems are documented especially in literature relating to pressure sensitive paints. Researchers have attempted to correct for these errors by using by reference imaging and by other mathematical means [25]. However, a better intensity approach that eliminates many of these issues is the intensity ratio approach.

The intensity ratio mode relies on taking a ratio of two emission lines. By doing this a number of errors can be eliminated. In pressure-sensitive-paint (PSP) literature, pressure sensitive paints were added with pressure insensitive reference dyes to make binary paints. The insensitive dye acts as an intensity monitor. Bell et al. [15] reports this technique to be the most successful approach for illumination compensation.

The same methodology can be applied to thermographic phosphors. Some phosphors exhibit a multiple emission response with some emission lines being insensitive/less sensitive to temperature. Ideally, the intensity of one of the emission lines should be independent of temperature. Figure 9 shows an ideal intensity variation of the two emission lines with temperature. Phosphors with these characteristics can act as binary paints, and a calibration between the ratio of emissions can be indicative of temperature. It is important that the reference can be excited with the same light and show emissions at different wavelengths so that they can easily be differentiated.

For low temperatures Chyu and Bizzak calibrated a 2D intensity measurement ratio for La₂O₂S:Eu to make surface heat transfer measurements for a hot jet impinging on a circular plate [26, 27]. The system reported a range of 292K-333K with an accuracy of 0.5K and repeatability of 0.15K. The cooling effectiveness was also determined from a row of cooling holes [28]. Until recently, dysprosium was the only known rare-earth activator to exhibit an intensity ratio response at high temperatures. Fiest and Heyes [29] showed similar response with samarium-doped phosphors. The main mechanism behind this phenomenon is thermailisation [30]. When two energy levels are closely separated by a difference of approximately 1,000 cm⁻¹, the upper level will not fluorescence at low temperatures due to high multi-photon relaxation that quenches the energy. As the temperature increases, the upper level becomes more populated and hence the fluorescence from this level gradually increases. Figure 10 illustrates the similarities between the energy diagram of free Dy and Sm ions. The diagram is only indicative of the physical principles, and in reality there will be host interactions that resulting in variations in the energy levels which could lead to energy splitting, line broadening and shifting [29]. Figure 11 illustrates the emission spectra of YAG:Dy and Y₂O₂S:Sm.

For YAG:Dy, the absorbed laser light excites the dysprosium into an excited state which relaxes to the ⁴F_9/2 level. This level undergoes fast thermal equilibrium and pumps a proportion of its population to the nearby ⁴I_15/2 level. As the temperature increases, there is a gradual build-up of the population to this level, and hence level of fluorescence. However, above a certain temperature, luminescence slowly begins to decrease due to the charge transfer state (CTS) transitions [32]. The ⁴F_9/2 level emission (496nm) almost stays constant with increasing temperature, and therefore can be used as an internal reference for calibration for level emission. This allows temperature determination as a relative, rather than a absolute measurement eliminating significant sources of error [30].

These two discrete energy states produce two distinct emission lines. According to Heyes [16], the electron population follows the Boltzmann's relation and is dependant on the temperature and the energy gap. The ratio of the two emission lines can be easily determined by monitoring the increase in fluorescence relative to the lower level.

(2)R(T,ΔE)=IeIg=1ΔEkT

The intensity ratio technique using thermographic phosphors was first cited in Gross et al. [30] using YAG:Dy3 with a reported temperature range of 300K-1,500K and an accuracy of ±9 to ±50K. Kontis et al. [32], reported a similar system utilising two gated ICCD cameras. Temperature calibration was made between 295K-1,350K, with a reported accuracy and repeatability of ±2.5K and <0.3%. The system was used for thermal measurements on a ceramic plate exposed to an impinging jet flame [32], and surface heat transfer measurements in a supersonic combustor [33].

Heyes, Feist and Seedfeldt [34] investigated the intensity ratio for dysprosium using YAG and YSZ hosts. Temperature calibration was made between 300-900K, with data repeatability around ±0.6% [34]. The system was used for temperature measurement on ceramic and alloy plates that were heated by flame impingement. YSZ is used for making gas turbine thermal barrier coating; the tests demonstrated the capability of making ‘smart TBCs’ with instrumentation abilities. The same authors have also investigated Y₂O₂S:Sm phosphors using the intensity ratio mode between 300-1,100K and showed an uncertainty of ±1%; they also tested the lifetime decay response mode from 900-1,425K and showed an uncertainty of ±1% and 0.1% at higher temperatures [29].

The drawback of intensity ratio response is that two separate detections are required. The conventional way to achieve this is by using two cameras with appropriate optical filters to detected the intensity of the desired wavelength. Another way to achieve this is by using a filter wheel. Table 5 compares these techniques.

More recent approaches include the use of a cube beam splitter to ensure that the images are spatially identical. This approach was used by Kontis [32], with the schematic shown in Figure 10. The total intensity will split two ways and a reduction in the intensity would be expected.

Stereoscopes have also been used. A stereoscope has two apertures which allow two images to be independently filtered using a single camera. It provides similar advantages as the ‘filter wheel’ approach, with the additional advantage of having no moving parts. This approach was adopted by Heyes et al. [34] to image the dual emission ratio response of a YAG:Dy and YSZ:Dy. The system was later enhanced to also allow the simultaneous measurement of lifetime decay response. This system allows the cross checking of temperature using the two methods, and also extends the dynamic range of measurement [35]. Similar two mode response systems have been reported by Omrane and Hasegawa [36].

This method is based on the decay mechanism of phosphor emission. The method is a well established technique for studying emissions of fluorescent molecules, and is used in a number of disciplines. It eliminates many of the issues related with intensity based approaches. The approach is:

Insensitive to non-uniform excitation

Reponses are usually observed using fast responding detectors, such as PMTs. This method is extremely effective and current detectors can observe decay lifetimes as short as a few hundred picoseconds with single photon counting capability.

Excitation promotes a large number of electrons into an excited state. When excitation is ceased, electrons return to their ground equilibrium level. For simplicity, this is either a radiative or non radiative transition. The rate of the electron population returning to the ground state can be expressed mathematically as: (3)dNdT=λNwith the solution yielding to (4)Noe-λtwhere N(t) is the quantity of electrons at a given time, N₀ is the initial quantity of excited electrons at t=0, and λ is the decay constant, the rate at which electrons make this transition.

The mean lifetime of which an electron remains in the excited state can be easily calculated.

(5)τ=1λ

Since the two transition pathways (radiative and non-radiative) compete and mutually exclusive, the decay constant can be written as the sum of the two possible rates of transitions. The analysis excludes the effects of interaction between activators, impurities in the host that can lead to further processes and change the simple exponential decay signature.

(6)λ=kr+knr

The radiative rate (k_r) is an temperature independent term and can be considered as being a constant, whilst the non-radiative (k_nr) transition becomes highly temperature dependant after the quenching temperature. For a given temperature, the probability of an single electron taking a transition pathway can be calculated from basic probability theory, resulting in: (7)Probability of radiative emissions:Pr=krkr+knr (8)Probability of non-radiative emissions:Pnr=knrkr+knr

If the temperature is increased, the decay rate via non-radiative means (k_nr) also increases. This has the following consequences highlighted in Table 6.

In summary, the probability of radiative transition will decrease whilst the probability of non-radiative transition will increase. By assuming the electron population is proportional to the observed luminescent intensity. The lifetime decay relation can be represented as: (9)I(t)=Ioetτwhere I_o is the initial intensity at time t = 0, and τ is the decay constant.

Figure 13 illustrates typical lifetime characteristics with increasing temperature. The graph shows faster decays with temperature. The relation is only held after the quenching temperature. Researchers have also observed variation in intensity levels with temperature that is not shown in the figure. Figure 14 illustrates the decrease in lifetime decay with temperature for a range of phosphors. It also shows the quenching temperature for some of the phosphors.

Since the lifetime approach is independent of illumination energy, the problems associated with model deformation, movement, shading and uneven light distribution do not exist [25]. In terms of disadvantages, the lifetime method suffers a lack of signal strength [37] as excitation light, in pulsed form, is only available for fractions of the time. To compensate for this, high-powered laser pulses are commonly used. Increasing the pulse strength risks the destruction of the paint. This is true for pressure sensitive paints; however, phosphors have much higher damage tolerances.

The highest temperature recorded using phosphor thermometry was obtained by researchers at ORNL, who successfully calibrated YAG:Dy to 1,705°C using the lifetime decay approach [24].

An investigation by Rhys-Williams and Fuller [43] noted that there are rise times associated with the response of thermographic phosphors. Their research showed that it was dependant on activator concentrations. The phosphor under investigation was Y₂O₃:Eu at room temperature. Ranson later analysed risetime characteristics in the late nineties and realised that it could be used for detecting temperature [44].

Ranson et al. [45], notes that the crystal structure of Y₂O₃:Eu has two sites of symmetry producing energy levels shown in Figure 17. They note the previous work of Heber et al. [46] who gives evidence for three potential energy transfers (a,b and c) to level D₀. The energy transitions of paths ‘a’ and ‘b’ have been observed to be very fast compared to that of ‘c’ [47]. It is this transition that gives this phosphor the rise time characteristics.

The emission of 611 nm (path d) follows the lifetime decay relation shown previously in section 5.2.

(10)N(t)=Noe-tτdwhere N₀, in this case, is the total number of electrons at D_0. This is not fixed and depends on the transition paths ‘a’, ‘b’ and ‘c’. The fast transitions ‘a’ and ‘b’ can be modelled as being instantaneous; but the transition of ‘c’ is dependant on the decay of electrons from C_3i to D₀ which decay at (11)Nc3i(t)=Nce-tτd

Thus, the number of electrons accumulated from path ‘c’ as a function of time is: (12)Nc(t)=Nc-Nce-tτr=Nc(1-e-tτr)

The total number of electrons at D_o is then: (13)N0(t)=Nab+Nc(1-e-tτr)

Combining the equations yields the full characterisation of the decay: (14)N(t)=[Nab+Nc(1-e-tτr)]e-tτdwhere τ_d = lifetime decay, τ_r is the risetime, N_ab and N_c are the number of electrons by transitions a, b and c, respectively.

The investigations were carried out were carried out using Y₂O₃:Eu phosphor with approximately 3% Eu concentration. Previous investigations by Rhys-Williams and Fuller [43] noted that rise times ranged from 60 μs at 5% mole concentration to 320 μs at 0.27% mole concentration. Recent work by Allison et al. [48] underwent investigations at 0.5% Eu. The results shown in Figure 18 clearly demonstrate the effects of temperature on risetime, showing a noticeable clear decrease in risetime due to increasing temperatures. Another temperature related response, that is further discussed section 5.5, is also shown; there is an increase in luminescence strength due to increasing temperatures. According to Allison et al. [48], this is due to increased phosphor absorption at the excitation wavelength (337 nm nitrogen laser).

According to Gross et al. [30] temperature can cause the crystal lattice containing the rare-earth to vibrate creating a changing crystal field that produces a broadening of emission linewidths. Frequency shift of the spectral lines can also occur due to thermal expansion of the crystal lattice [30]. Both lines shift variation and broadening can be calibrated to reveal temperature. However, these effects are usually small. The variation in the line shift at 1,000K is only 3 nm, making the sensitivity very small and difficult to detect [16]. Kusama et al. [49] utilised this approach for measuring temperature using Y₂O₂S:Eu phosphor and the following graph shows variations at ^-15ºC and 72ºC. Kusama et al. [49] suggested quadratic shift according to the equation: (15)E=A+BT2where ‘E’ is the expected energy, ‘A’ is the ground energy at 0K, ‘B’ is a constant, and ‘T’ is the temperature in Kelvin.

Various studies have shown a variation in the excitation and absorption band of some phosphors due to changes in temperature. When a nitrogen laser (337nm) or third harmonic Nd:YAG laser (355 nm) is used to excite a Y₂O₃:Eu phosphor, there is a gradual increase in the emission intensity with increasing temperature. According to Allison and Gillies [2], this is because the absorption, at these wavelengths, is weak at room temperatures, and slowly increases with temperature. The absorption spectra for Y₂O₃:Eu at room temperature is shown in Figure 20, illustrating the weak absorption lines at 337 and 355 nm. Figure 21 illustrates the shift in absorption band due to increases in temperature. If a linear trend is assumed, it seems reasonable to use these wavelengths for higher temperature detection.

It has been shown that the activator concentration affects the temporal decay profile and the intensity of the emission. Y₂O₃:Eu concentrations less than 5% leads to strongest lines of shortest wavelength [16]. Greater concentrations lead to dispersion with no sharp lines being observable. With increasing concentrations, the energy gap between lines is reduced so electrons reach lower levels from neighbouring ions by non-radiative means. Allison and Gillies [2] notes that higher activator concentrations may alter the fluorescent decay so that it follows a multi-exponential rather than a simple exponential profile, making measurements more difficult to characterise and prone to errors.

As previously discussed, the risetime of the phosphor's response is also affected by the activator concentration. Reducing the dopant concentration increases the rise time for Y₂O₃:Eu phosphor [43]. Not much information is available to see whether this is universally true for other phosphors.

It most applications, it can be assumed that thin coatings of the phosphor exhibit the same temperature as the surface of interest. However, in some applications, where temperatures are changing at fast rates, knowledge of the phosphors thermal response is required to properly unfold the temperature [24]. In YAG phosphors, increasing the dopant concentration reduces the thermal conductivity. Kontis [32] notes that most 1% dopant YAG phosphors has a thermal conductivity of 4 Wm^-1K^-1, which is reduced to 2 Wm^-1K^-1 when the concentration is increased to 3%.

High excitation energies can lead to luminescence saturation. This is where the luminescent intensity does not change with increasing energy from the source. In fact, above a threshold, there have been reported cases where luminescent intensity actually decreases with faster decay profiles. There are a number of explanations for this. The laser beam can induce an increase in temperature [50]; in this case thermodynamic consideration must be given for these beam related effects. According to Allison and Gillies [2], this is probably due to the increased probability of two ions being excited in close proximities. This increases the chances of the energy being transferred from one ion to the other, with only one photon being emitted instead of two.

Pressure sensitive paints respond to both thermal changes and changes in the level of oxygen. Thermographic phosphors were originally thought to be independent of oxygen changes. Recent investigations are challenging this assumption. These investigations are important if phosphors are to be utilised in areas where the partial oxygen level is likely to change e.g. consumption of oxygen in combustion chambers.

Feist et al. [51] investigated the oxygen quenching of Y₂O₃:Eu and YAG:Dy. Volumetric percentage of oxygen was changed from 21% to 5% by flooding the furnace with nitrogen. No absolute changes were noted, but the readings resulted in increased uncertainties in temperature measurement. For Y₂O₃:Eu, the uncertainties, due to changes in oxygen, were an order of magnitude greater than uncertainties at fixed concentrations, providing a convincing case for oxygen quenching. However, for YAG:Dy, the uncertainty was the same order of magnitude, and so the results for this case may be considered as being inconclusive.

A more recent investigation by Brubach et al. [52] showed the effects of various gas compositions on three different phosphors. The results show that variations in oxygen, nitrogen, helium, carbon dioxide, water vapour and methane concentrations do not influence the decay time of Mg₄FGeO₆:Mn and La₂O₂S:Eu phosphors. (Figure 22 a,b). These phosphors are only influenced by thermal quenching and are suitable for environments where changing gas environments are expected. Y₂O₃:Eu (Figure 22c) however showed high sensitivity to oxygen.

Apart from pressure causing an increase in partial oxygen levels, there is also evidence that application of pressure/strain can affect luminescent properties of thermographic phosphors. This phenomenon is not very well understood but becomes very relevant when extreme pressures are concerned. The application of pressure can be viewed as the imposition of compressive strain that can result in changes in both chemical bonds and atomic level orbital configurations. The decay time of Gd₂O₂S:Tb decreased by an order of magnitude with application of 2 GPa, while the decay time of La₂O₂S:Eu increased by an order of magnitude with application of 3.5 GPa [2].

Although some phosphors may not exhibit oxygen sensitivity, for example La₂O₂S:Eu [52], they may possess pressure sensitivity; it is important that both parameters are treated independently. However, in most flow conditions, it is unlikely that these sorts of pressures will be reached (1GPa = 10,000 Bar). Although, Brubach et al. [52] investigations showed no change in lifetime for La₂O₂S:Eu up to a pressure of 10 Bar (1 MPa), the results presented in

Figure 23, illustrates the decrease in decay lifetime at higher pressures (0 – 50 MPa). In very harsh flows, such as those experienced in gas turbine engines, the maximum pressure is around 50 bar (5 MPa), and the effects of this phenomena may become relevant.

Y₂O₃:Eu phosphor showed sensitivity to oxygen quenching and showed irreversible changes after the absolute pressure was increased to 6 Bar [52]. According to these findings, Y₂O₃:Eu, which has been a very popular choice of phosphor for turbine engine thermometry, is unsuitable for environments where the pressure and oxygen level is expected to change.

Impurities in the phosphor can affect luminescence. In a simple case, excitation energy acts directly on the activator, as shown in Figure 24, which consequently produces radiative emissions with some energy being lost by other non-radiative means. Impurities in the host material can change atomic electronic environment experienced by the activators. Transition metal impurities, even at low concentrations (1 ppm), can decrease luminescence due to them extracting energy that would otherwise be used to produce radiative emissions. A representation is shown in Figure 25. Since there is a change in the probability of non-radiative and radiative energy transfers, the decay rate of luminescence is also expected to be altered.

It is possible for the energy transfer to act in the other way. UV radiation on impurities can further excite the activator by energy transfer. These added impurities are termed sensitizers if their presence increases luminescence. In some cases, the activator only produces radiative emissions with a sensitizer is present. (Figure 26– case A). The host lattice can itself act as a sensitizer, for example YVO₄:Eu³⁺. In other cases, both the activator and the sensitiser can be directly excited, (case B), and the sensitizer can also be luminescent (case C). The sensitisers could be additional activators, which further complicates the analysis. Some experiments have revealed that the addition of small amounts of other activators, such as Dy and Tb or Pr, to Y₂O₃:Eu decreased the lifetime decay by a factor of 3, with little change in the quantum efficiency [2].

The energy transfer from the sensitizer to the activator is termed Resonance Energy Transfer (RET). RET is also possible when the emission spectra of the sensitizer (donor) overlaps the absorption spectra of the activator (acceptor). The transfer is manifested by the quenching of the donor and the increased absorption from the activator that consequently results in increased emissions. These complex mechanisms can be used to explain risetimes and complex multi-exponential decay profiles.

There have been a number of studies to suggest that lifetime decay and intensity changes with phosphors particle size. Investigations into nano-crystalline and coarse grain particles of Y₂O₃:Eu phosphors reveal that the excited state parabola on the configuration coordinate diagram may be affected. Konrad et al. [54], explains there is an increasing slope of the excited parabola with reducing particle size, shown in Figure 27.

This results in the intersection point between the excited and ground state being increased. Consequently, the quenching temperature is expected to be higher and the lifetime decays are expected to last longer. Work by Christensen et al. [55], has shown an increase in lifetime ranging between 436-598 µs, due to reductions in particle size from 0.42 to 0.11 µm. As different preparation and surface bonding techniques produce particle sizes, it seem reasonable to assume that the decay lifetime is not absolute, and therefore it is important that calibration is deployed for that type of preparation or bonding technique.

This process involves mixing powdered phosphors with chemical bonding agents to create a paint that can be either brushed or air-sprayed on to a surface. The nature of the binder will depend on the surface and the operating temperature range. Epoxy binders have a temperature limit that is reached at a few hundred degrees. Apart from survivability at higher temperatures, chemical binders must allow transmission characteristics that enable the phosphor to be excited and emissions to be detected. Chemically bonded phosphors usually require curing by raising the temperature to 700°C and slowly bringing it back down to room temperature. In the past few years, a variety of commercially available binders have been investigated [56-58]. Propriety binders manufactured by thermal paints experts at Rolls-Royce Plc have been successfully tested up to 1,100°C [16]. Some of the higher surviving commercially available binders include ZYP-ZAP and Coltronics-Resbond, which have shown survivability and fluorescence detection up to 1,600°C [59]. Table 7 compares some of these binders. Goedeke et al. [59] notes that although ZYP-ZAP has stronger survivability, the observed fluorescence is higher in Resbond at 1,500°C.

Problems with chemical binders include the possibility of changing the phosphor's atomic configuration, and hence luminescence and thermographic properties. Ideally, chemical binders should suspend the phosphor without changing atomic properties.

Problems associated at high temperatures include differences in thermal expansion that causes the paint and substrate to expand at different rates. According to Allison et al. [56], one of the most challenging surface for bonding is high strength nickel alloy due to high differences in the thermal expansion coefficients. At high temperatures, this causes the paint to flake off. To increase the thermal conductivity of the paint, to reduce thermal shock and increase survivability of the paint, tests were conducted with the addition of MgO₂ in the binder.

Another problem associated with chemical binders is the effects due to thermal exposure. Figure 28 demonstrates the reduction in emission intensity of Y₂O₃:Eu phosphor with Resbond 793 binder at 1,400°C after 4 hours of thermal exposure [59]. Similar results were reported by Ranson et al. [60] using a different chemical binder. The results showed a reduction in intensity to approximately 10% of their initial value following thermal exposure at 1,200°C for two hours. The reason for this could be simply due to the paint layer flaking off. Other cases report the paint transitioning to a yellow/brown colour, and is still unclear whether it is problems with the optical transmission of UV, or the optical passing of emissions, or a combination of the two that is responsible for the reduction in intensity. Thermal exposure for long periods may force chemical reactions within the phosphor changing its characteristics.

Chemical binders allow the use of spray painting. The advantage of this is that large areas of various sizes and shapes can easily be covered. However, maintaining a uniform surface and controlling the thickness and roughness can be difficult. Tests indicate a variation in intensity across different test pieces [61].

The disadvantage of having binder paints is that the minimum coating size that can be produced is around 10 µm (typical is around 30-60 µm). This is relatively large compared to the vapour deposition and plasma spraying techniques. Greater thicknesses provide greater thermal gradients between the phosphor coating and substrate, constituting to a greater error in measurement.

In this process, a coating is applied by vaporising the phosphor and allowing it to condense on the surface of interest. There are a variety of ways this can be achieved including electron beam (EB-PVD), pulsed laser disposition (PLD), chemical vapour deposition (CVD) and RF-frequency sputtering. No chemical binders are required and therefore there is no interference or problems concerning optical transmittability of UV and emission wavelengths. The resulting coatings are very robust and long-lived with fluorescent intensity being constant throughout its life. They can be made very thin compared to chemical binder paints, and can be finely controlled to have a uniform surface finish. However, the equipment required to produce these coatings can be very expensive and the coatings areas are usually very limited.

During vapour deposition, dopant atoms can be situated in a variety of positions and rotations within the hosts crystal structure and therefore experience a variety of crystal field effects, leading to weaker and wider spectral emissions. Post annealing is required to realign the ions to restore crystalline quality and increase luminescent intensity. Allison et al. [2] notes that the high temperatures generated during vapour deposition can irreversibly break down some phosphors, such as oxy-sulphides, and post annealing will offer no benefit.

Ranson et al. [60] investigated thin coatings (0.1 µm – 3 µm) produced by RF sputtering and thick coatings produced by binder paints. They found that emission intensity was related to coating thickness, up to a certain level. It was shown that 0.7 µm post annealed coatings provided the same level of intensity as thick coatings (approx 10 µm) produced by chemical binders. These tests were undertaken for Y₂O₃:Eu phosphors, and whether this is true for other phosphors is yet to be investigated.

While thick coating produced by binder paints show declining intensity with thermal exposure, Ranson et al. [60] showed that thin coatings produced by vapour deposition show no decline in intensity level. Figure 30 compares shows the variation in luminescent intensity as a function of time hours of constant thermal exposure at 1,200ºC. The results show that the binder paints showed a very quick decline in intensity, whilst thin coatings produced by vapour deposition showed no change in intensity.

Flame spraying is portable and applicable to objects of diverse geometries, but has a lower impact velocity than plasma spray [1]. Plasma spraying involves injecting powdered phosphors into an electrical discharge to liquefy them. The technique allows robust, well-adhered coatings made up from a multiplicity of solidified droplets [16]. It is usually used to produce thicker films than vapour deposition methods and can produce coatings exceeding a millimetre. Like vapour deposition, annealing is also necessary to restore crystalline quality and improve luminescent efficiency.

For point measurements, there are a number of detectors that can be used, with the main ones being PMTs, silicon and avalanche photodiodes, and newly developed Si-Photomultipliers. This section compares these, and the main findings and typical characteristics are highlighted in Table 8.

Continuous lasers and other light sources are suitable for intensity measurements. Pulsing can be introduced to enable lifetime mode analysis. One way is to use a mechanical shuttering mechanism. There are limits on how fast these can operate. Previous mechanisms operated in the sub milliseconds regime, and were considered too slow and unsuitable to detect lifetimes shorter than this. Newer optical choppers/mechanical shutters can provide sub-microsecond [75] pulses, and now there are a range of optical shutters (LCDs) that can provide responses faster than this. The problem with continuous lasers is that the energy per pulse is relatively low. For example, to obtain the same power as you would from a typical 266 nm 10 ns 10 mJ Q switched Nd:YAG laser pulse, you will need a continuous laser operating continually with 1 MW of power at that wavelength. Typical high energy laser systems only operate with average powers of 5 KW. This lack of power from continuous sources have lead to researchers resorting to high peak power pulsed laser systems, even when they are utilising the intensity mode [29, 33, 76].

Fibre lasers are increasing becoming more popular due to increased reliable up-time, beam quality, reduced running costs and servicing operation. In principal, fibre lasers are similar to DPSS lasers. The generic design includes laser diodes for pumping; a scheme for coupling the pump energy into the gain medium; a fibre based resonator configuration with brag gratings instead of mirrors; and a method for getting rid of excess heat. In a fibre laser, the laser is created directly inside a fibre. Therefore, there is no need for optical setup that requires the beam to delivered to a target via a series of steering mirrors. The use of fibre optics opens up areas of application that may have restricted optical access. Fibre lasers eliminate the need to for fibre optic coupling from a conventional laser. Fianium Ltd has recently created the worlds' first commercial high-powered 266 and 355 nm UV fibre laser [77]. The system is capable of producing 1W average power at 100 MHz, with a pulse width of 10 ps. This yields to 0.01 uJ/pulse and a peak power of 1 KW, which is relatively low.

Allison et al. [78] reports that high energy UV LEDs could be used to excite phosphors. They can be used in continuous or pulsed mode. In the pulsed mode, even though they have relatively low powers, they can be operated with pulse widths to increase the total amount of pulse energy. They have fall times of a few ns, and have emissions spectra much broader than that of a laser. This may be absorbed better by the phosphor, and may produce intensities that are of similar magnitude. Newly developed high powered UV LEDs can produce 450 mW of continuous 380 nm UV light [79]. In pulsed mode at a pulse width of 1 µs yields to energies of around 0.45 µJ/pulse. This is relatively low and would therefore be unsuitable for high temperature measurements where blackbody radiation becomes significant. It may be suitable in applications where relatively low temperatures are concerned.

LEDs have been very successful and have replaced lasers in many applications including pressure-sensitive-paints and fluorescence detection for biological purposes [80]. Since the pulse widths can be finely controlled, the decay lifetime can be determined using the frequency domain approach highlighted in Section 5.2. A sinusoidal wave using blue LEDs was used to excite the phosphor and determine lifetime in experiments conduced by Allison et al. [42].

Kontis et al. [32] investigated a turbulent flame that was impinged onto a 100 µm YAG:Dy phosphor coated onto an alumina ceramic disc of thickness 0.05 m and diameter of 0.035 m. The intensity ratio method was utilised to map the temperature every 2.5 s for a total of 142.5 s for heating, and 150 s for the cooling of the plate. The system used was previously shown in Figure 12. The results show that the maximum temperature (1250K) was attained after 135 s of initial flame impingement. The temporal variations in the temperature profiles around the centre of the disc is an indication of the strong and localised effects of hot jets, and the transient heat over the entire disc surface. Kontis et al. [32] notes that such systems can be used to evaluate the local heat transfer coefficients by using an appropriate theoretical model of the thermal response to heat flux of the configuration under consideration.

Similar transient heating experiments using the intensity ratio method was also undertaken by Heyes et al. [34]. A ceramic plate was brush-painted with YAG:Dy forming a thickness of 150-200 µm. A Nimonic alloy was coated with YSZ:Dy using plasma spray, producing a thickness of 100 µm. A Bunsen burner was ignited after 20 s, and was used to heat the rear surface the plates for 60 s. A K-type thermocouple was mounted on the front surface for comparative analysis. A Nd:YAG laser, with a pulse energy of 50 mJ, was used to excite the phosphors. Ten pulses were averaged for each measurement with a sampling period of 1 s.

Results, shown in Figure 42 and Figure 43, for the ceramic plate heating indicate that the front surface of the plate continued to heat up after the heat was removed. This is to be expected from ceramic material that has low thermal conductivity. The maximum temperature recorded was at 390K at 207 s. Results for the Nimonic plate indicate that the maximum temperature was observed at 840K after 82 s. Thermocouple measurements were in good agreement below 400K (5K difference). At higher temperatures there were larger differences, highlighting the difficulty of accurately measuring temperature of hot surfaces using conventional methods. Heyes et al. [34] notes that thermocouples are expected to under-predict the temperature due to changes in the local temperature caused by thermocouples acting as a heat sink, imperfect contact, heat loss by thermocouple convention and radiation, with additional errors caused by the thermocouple junction being above the surface.

Phosphor thermometry to measure the temperatures of surfaces within operating turbine engines dates back at least 20 years. Noel et al. [81] proposed such system in 1986. In a study by Saner et al. [82], optical temperature diagnostics was performed on the afterburner of an Volvo RM12, the engine used in the Gripen fighterjet. The intensity ratio mode of YAG:Dy was used to map the temperature inside an afterburner. HPC binder paint was used to attach the phosphor on to the surface of interest. A 355nm Nd:YAG laser was used to excite approx 100 mm². The phosphorescence signal was sampled using a stereoscope and an ICCD, allowing the simultaneous imaging of two spectrally filtered images. The ICCD was gated at 100 µm to suppress background radiation. The images of 455 nm and 493 nm were digital divided to obtain a ratio at each pixel that was processed through a calibration curve to reveal temperature. Temperature maps were taken at different engine running conditions. Figure 44 shows thermal maps at full afterburner load.

The measurements corresponded well to previous measurements made with thermocouples. This experiment demonstrates phosphor thermometry could be used in very harsh environments. In addition to increasing blackbody radiation at high temperatures, another problems that was encountered was that of flame emissions that overlapped emissions from the YAG:Dy phosphor. The flame emissions peaked at approx 415 nm. Although, background images were subtracted, it would have been more beneficial to use phosphors with emission wavelengths that did not interfere with the environment. In more recent experiments involving afterburner temperatures, the lifetime approach was used to determine temperature using an 266 nm Nd:YAG laser, a PMT and Mg₃FGeO₄:Mn phosphor [83].

Surface temperature measurements were made in a laboratory combustion rig by researchers at Imperial College, London [22]. The system was used to assess whether thermographic phosphors could be used for surface temperature measurements in gas turbines, and to evaluate film cooling. Film cooling involves the ejection of air over a surface, and is generally used to cool combustor liners, turbine blades and nozzle guide vanes in gas turbines. It provides thermal protection from high gas temperatures that usually exceeds the materials melting point, and would otherwise cause failure.

The combustor rig had a quartz window for optical access. Cooling air was directed over the window by a row of holes to enable it to survive the temperature, while keeping it free from wetting and carbon build up. The system was first tested using 20 µm Y₂O₃:Eu paint, prepared by Roll-Royce Plc, over an underlying TBC of thickness 250 µm. The phosphor was also applied within the depression of the cooling hole exits. The optical setup was arranged so the lifetime decay response could be measured using a PMT (Figure 45). A camera lens was used to focus the detection on a single point, and a traverse was used to successively scan an area of 8mm² with 128 point measurements.

This system evolved and was later modified to simultaneously measure both the intensity ratio and the decay response of YAG:Dy phosphor, allowing temperature to be determined by two independent methods [84]. A dichroic mirror was used to split the light equally to two PMTs that had spectrally different filters capture the relevant wavelengths of 455 nm and 494 nm. A schematic is shown in Figure 48.

The results (Figure 47) indicate a temperature distribution ranging from 300–600°C, with the uncertainty reported to be better than 2% [22]. The effects of film cooling holes can be clearly seen, demonstrating phosphor thermometry's capability in measuring such phenomena.

Both efficiency (fuel economy) and performance (thrust) can be improved without increasing the size of the engine if higher turbine inlet temperatures are achieved [85]. However, there are consequences; a few degrees of operation at over-temperature can result in drastic reductions in blade life. Therefore, the operating temperature is prescribed by the balance between the benefits of thermal efficiency at higher temperatures and material stability and life.

At the moment, Khalid and Kontis at the University of Manchester are researching into methods of successfully measuring surface temperatures on both rotating and static components of development aeroengines using phosphor thermometry. This will eventually help predict heat transfer distributions, verify the effects of design changes, cooling effectiveness, and aid designers optimise aeroengine designs to enable higher temperature operation.

Lifetime imaging for thermal measurements using thermographic phosphors has been intensively used by a team of researchers at Lund University. They claim to be the first to obtain such 2D measurements [40]. A high speed framing camera (Imacon-Hadland), containing an eight faced prism is used to split light to eight independent intensified CCD cameras, where the exposure time and time separation between images can all be precisely configured.

Figure 48 shows a schematic of the framing camera. Using these images, an exponential curve can be fitted for each pixel, and evaluated against a calibration curve (Figure 49). This produces a complete 2D thermal map that was determined using lifetime decay analysis.

The group have obtained 2D surface measurements on low density fibre boards covered with Mg₄FGeO₆:Mn phosphor. Burning alcohol was used to heat the sample because it generates less soot that could interfere with measurements.

The radiation energy was 24 mJ/pulse which was expanded to an area of 100 cm² yielding a fluence of 0.25 mJ/cm². The results obtained showed a standard deviation of ±5K (less than 1%) at temperatures between 680-780K [40]. Recent experiments involved measurement of combustible and non-combustible surfaces using both the intensity ratio and lifetime imaging modes [86]. Figure 50 shows thermal temperature maps during a flame spread experiment at various times.

The surface temperature is a key parameter for modelling the decomposition of solid materials. It has a strong influence on the heat flow into and out of the material, and also determines the ignition temperature. Omrane et al. [87] used phosphor thermometry for surface temperature measurements in pyrolysis studies. Pyrolysis is the chemical decomposition of organic materials by heating in the absence of oxygen or any other reagents. According to the authors, accurate measurement using conventional methods prove to be problematic. Thermocouples and thermistors require direct contact, and may suffer heat losses and induce catalytic effects. Pyromerty can be difficult because emissivity of the surface may not be known and may be changing during the investigation.

Phosphor particles were deposited on the material under investigation and were placed inside a high temperature reactor that was pre-stabilised at a temperature of 733K [40]. A 266 nm Nd:YAG laser was used to excite the sample with subsequent emissions collected by a PMT, ICCD and a spectrograph, for both temporal and spectral analysis. A schematic is shown in Figure 51. The results show that the surface temperature increased to reach the temperature of the reactor.

In later work, pyrolysis on construction materials were studied. These included low-density fibre boards (LFB), medium fibre boards (MDF), particle board (PB) and polymethylmetharcrylate (PMMA) [87]. Cubic samples (5 mm³) were doped into Mg₄FGeO₆:Mn phosphor that was dispersed in ethanol or toluene. The samples were dried at 95°C and inserted through an air cooled fall tube onto to a holder. This was connected to a balance to monitor weight of during the investigation. A molecular beam mass spectrometer was used to monitor gas composition and ensure oxygen levels stayed below 0.1%.

The rapid pyrolysis of construction materials was successfully monitored using phosphor thermometry. The results covered a temperature range of 300-600°C. This covers pyrolysis initiation and completion for most construction materials during typical fire spread situations. The results are another demonstration of phosphor thermometry for detailed temperature measurements during these complex combustion-related situations where common techniques fail.

Temperature measurement of internal component surfaces enhances understanding of the processes inside the combustion engine. Armfield et al. [88], cited in Allison et al. [53], reports the use of thermographic phosphors for temperature measurements of an intake value and a piston.

For the intake value experiments, LaO₂S:Eu phosphor was coated on the stem-side of the value. A nitrogen laser was delivered through a 1mm optical fibre accessed through the head of the car. The same fibre was used to direct the luminescent emissions to a PMT. The lifetime decay response was used to determine temperature. A schematic is shown in Figure 53. For piston measurements, a quartz window was used to gain direct optical access as shown in Figure 54.

The results shown in Figure 55 indicate show that in both cases, there is a rapid increase in temperature and is followed by a slower steady increase.

Similar temperature measurements of the intake and exhaust values of an optically accessible laboratory engine was conducted by Omrane [86]. Measurement would increase the understanding of fuel evaporation process when it is in contact with high temperature surfaces. In the investigation, the engine was operated at 1,200 RPM using gasoline and iso-octane fuel.

A thermographic phosphor was bonded onto the values using commercial binder. A fourth harmonic Nd:YAG (266 nm) was used to excite the phosphor through a quartz window. The emission signals were collected through a quartz fibre and digitised using a PMT and a fast oscilloscope (Figure 56). The lifetime decay response was used to determine temperature.

The success of single point temperature measurements was later extended to provide 2D thermal maps of engine walls, values and piston. A direct injection stratified charge(DISC) engine was used to provide such thermal maps [89]. An ICCD camera with a stereoscope and two interference filters (632 nm/657 nm) was used to image emissions at those wavelengths [86]. The intensity ratio approach was used to provide temperature data using pre-calibrated data.

Both results, shown in Figure 57 and Figure 58 illustrate the exhaust values start to heat up earlier than the intake value, and also reaches a higher temperature. This is due to new fresh air entering the intake value allowing it to be cooled. In addition, the exhaust value also experiences burned gases from the combustion chamber that heat up the value [76].

The thermographic phosphor method has recently been used to measure gas temperatures. Hasegawa et al. [36] investigated the intensity ratio of YAG:Dy for such measurements with the ultimate aim of measuring the gas temperatures in an operational engine.

A static calibration, followed by a steady flow validation, was conducted before actual engine tests. In the calibration procedure, an imaging stereoscope and two band pass filters were used in front of an ICCD. The lifetime of the phosphor was also detected using a PMT to optimise the gate times for the ICCD. In gas flow validation experiments, phosphor-seeded air was passed through a 10 mm diameter exit at a velocity of 80 cm/s. A temperature-controlled heater was used to heat the flow to 573K 15 mm below the tube exit. A schematic is shown in Figure 59.

Figure 60 compares three different methods of measurement. The thermocouple readings were lower, as expected, since its intrusiveness perturbs the flow field with ambient room temperature, decreasing its temperature. According to Hasegawa et al. [36], the fluctuations in the intensity ratio method are mainly due to heterogeneity of gas seeding, causing a temperature error of around 5%.

During engine tests, a four cylinder diesel engine running at 1,200 RPM was used. A 355 nm Nd:YAG laser was used at 80 mJ/pulse and a homebuilt seeding device was used to obtain homogenous seeding. According to the authors, the intrusion of the phosphor caused an average drop in temperature by 2%. It was shown that phosphor thermometry could be used to measure temperatures of un-burnt gas flows. In turbulent combustion conditions inside the engine, phosphor temperature data agreed well within 5% error with the thermodynamic calculated data. However, chemi-luminescence effects caused measurements to be restricted after 10º after top dead centre. Further details on this can be found in the reference.

Laser induced phosphoresce from thermographic phosphors was used to measure temperature of single falling droplets [90]. The droplets were excited using 266 nm Nd:YAG. The resulting emissions were evaluated spectrally, using a stereograph and Mg₃FGeO₄:Mn phosphor; and temporally, using a PMT and La₂O₂S:Eu phosphor. Phosphor (1%, by weight) was added to the liquid under investigation. A Nd:YAG laser was triggered when a droplet crossed a He-Ne beam. The setup is shown in Figure 61. The droplets were approx 3-4 mm in diameter. The temperature of the liquid was controlled using a thermocouple and heating wires around the container.

The results for both spectral and temporal methods are presented in Figure 62. The phosphor determined temperatures correspond well with the thermocouple readings from inside the tank. The thermocouple readings do not take into consideration the cooling and heat exchange to the surrounding air after it has left the nozzle.

To compensate for this, a model from Kinciad and Longley [91] was used. The results from the phosphors agreed very well to this with deviations less than 1%. Although the temperature of the droplets is of relatively low temperature, it does demonstrate the principle and may be used to measure much higher temperatures.

The same methodology was extended to form 2D thermal measurements of droplets using fast framing ICCD cameras and decay lifetime imaging. The temperature at each pixel was evaluated using calibration procedure of lifetime against temperature. The technique was first applied to free falling water-based droplets, then to a suspended droplet in an ultrasonic levitator [92], and later to sprays [93]. Figure 63 illustrates some example of results that were obtained. Brubach et al. [94] also performed 2D spray and droplet thermometry using the intensity ratio approach using a Mg₄GeO_5.5F:Me thermographic phosphor. Further details can be found in the references.

Kontis [33] investigated surface transfer measurements inside a supersonic combustor. The experiment was conducted on a blown-down design supersonic wind tunnel. The run time was approximately 8 s with 4 s of steady state flow conditions at nominal values. The schematic of the test section can be seen in Figure 64 with further details explained in the reference. Hydrogen fuel was injected through a slit located along the back surface of the step. Two alumina ceramic pieces were secured using a high temperature adhesive both upstream and downstream of the step. These were coated with a thin layer (<50 µm) of YAG:Dy. The total thickness of the phosphor + ceramic was approx 0.01 m. Frequency tripled Nd:YAG laser (355 nm) was used to excite the phosphor after 1 s, allowing time for hydrogen fuel combustion. Emission profiles were then recorded every second using the intensity ratio method.

Due to the short run time, the surface under consideration did not have sufficient time to attain an equilibrium temperature. Transient techniques could be employed to measure heat flux. A one-dimensional unsteady heat transfer conduction was employed for both the phosphor layer and alumina-zirconia ceramic substrate.

The experimental results were then compared with computational simulations. Taking uncertainty factors into consideration, detailed in the reference, the overall uncertainty for heat flux determination was in the region of ±5%, which compares well to conventional transient techniques, such as thin-film or thermocouple gauges.

Advances in image processing and optical sciences have made the luminescent coating technique practical for aerodynamic wind tunnel testing. The fundamentals of aerodynamic testing with organic luminescent coatings (PSPs and TSPs) are well documented [15, 95]. Hubner et al. [96] reports the use of such paints to obtain heat transfer measurements at Mach 11.1 with the temperature ranging from 270K to a maximum of 400K. However, at higher temperatures, organic TSP became completely quenched. At higher temperatures, the use of thermographic phosphors would be more appropriate.

NASA Langley Research Centre has been using the relative-intensity two-colour phosphor thermography system for at least 15 years. It has become a standard technique to measure temperature and flux. An UV illuminated phosphor coated model is exposed to the wind tunnel flow. Subsequent emissions from two wavelengths are observed and converted to surface temperature maps using pre-calibrated data. With temperature maps acquired at different times during the run, global heat transfer images can be computed, assuming an one-dimensional semi-infinite heat conduction model. Scaled hypersonic models of X-33 [97], X-34 [98], X-38 [99, 100] and X-43 [101] have all been tested in wind tunnels using phosphor thermometry at Mach 6 and 10 in air, and some have been tested at Mach 20 in Helium [97]. Further details can be found in the references.

According to NASA, the primary advantage of thermographic phosphors is the global resolution of the quantitative heat transfer data which can be used to identify heating footprints of complex, 3D flow phenomena, including transitional fronts, turbulent wedges and boundary layer vortices, that are extremely difficult to resolve using discrete measurement techniques. In addition, the technique does not need corrections that are required for infrared thermometry.

According to Hovarth et al. [97], measurement accuracy via phosphor thermometry is believed to better than around 8%, and the overall experimental uncertainty in heating data due to all factors is estimated to be around 15%, including uncertainties in the in the thermo-physical properties of the ceramic model. This is somewhat similar to that of thin films. Repeatability was found to be generally better, around 4% [100]. Merski [102] notes that the total uncertainties associated with the phosphor technique is shown to be approximately 7–10% in NASA's 31-Inch Mach 10 Tunnel and 8–10% in the 20-Inch Mach 6 Tunnel. A comparison with thin film measurements showed phosphor data to be within 7% and agreed better than CFD predictions. Figure 66 illustrates an example of the match between experimental phosphor data and predicted CFD data.

The phosphor technique provides a wealth of information critical to the design of thermal protection systems for applications involving engine design, re-entry vehicles, missiles and supersonic and hypersonic transport. Further research and work with phosphor thermometry at hypersonic speeds is planned at the University of Manchester in the near future.

Thermal Barrier Coatings (TBCs) provide thermal protection. They are usually found in very hot regions of a gas turbine engine. TBCs consist of a thin bond coat and an insulating layer, usually YSZ. They have a thickness in the order of 250 µm. TBC can be modified to behave like thermographic phosphors[103], exhibiting temperature dependable properties. The advantage of this is that no additional phosphor layers are required for temperature measurement. This concept was first proposed by Choy, Feist and Heyes [104].

YSZ:Eu and YSZ:Dy have been investigated both in powdered form and various forms of vapour depositions. Powdered YSZ:Eu response was observed with a dynamic range of 50-800ºC with a repeatability of ±0.1% using the lifetime method. YSZ:Dy was investigated using the intensity ratio method and was calibrated through a temperature range of 300-900K. This showed a repeatability of data around ±0.6% [34].

Using this methodology, it is possible to measure the temperature of the TBC. However, at high temperatures, there is a huge temperature gradient across the TBC, with temperatures being 200°C higher than that of the actual substrate. Figure 67 illustrates typical temperature variation across a TBC that is attached to a turbine blade experiencing high temperature flows. The actual substrate temperature is important for designers because it is this that determines material failures.

Thermographic TBC may be created from multi-laminar construction layers as illustrated in Figure 68. With this arrangement, discreet points in the coating could be measured, including the temperature of the bond coat. The actual doped YSZ could be sandwiched in between the undoped YSZ allowing temperature measurement at various distances through the TBC. This information could be used to determine the health of thermal barrier coatings. Southside Thermal Sciences, UK are further developing this concept.

The protective galvanneal process involves the dipping and heating of steel into molten zinc until the iron and zinc atoms form an alloy on the surface. The metals surface temperature may vary in the furnace, causing product quality and non-uniformity problems. Researchers at the ORNL developed a measurement system, based on the lifetime decay mechanism of phosphor thermometry to strictly control surface temperature, enabling the production of uniform, high-quality galvanneal steel [106]. The system, shown in Figure 69, includes a computerized phosphor-deposition device that is used to dust phosphor powder on the steel sheet. A portable nitrogen laser (377 nm, 0.3 mJ/pulse, 30Hz) is used to excite the phosphor, and subsequent emissions are detected and analysed in real time to evaluate and control the surface temperature by adjusting furnace settings. In 1998, the system was successfully installed and tested at Bethlehem Steel's facility in Burns Harbor, Indiana.