Artigo sem citações.pdf

View Article Online / Journal Homepage / Table of Contents for this issue

C

Nanoscale

Dynamic Article Links <

Cite this: Nanoscale, 2012, 4, 4301

REVIEW

www.rsc.org/nanoscale

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

Luminescence nanothermometry† Daniel Jaque*a and Fiorenzo Vetrone*b Received 30th March 2012, Accepted 14th May 2012 DOI: 10.1039/c2nr30764b The current status of luminescence nanothermometry is reviewed in detail. Based on the main parameters of luminescence including intensity, bandwidth, bandshape, polarization, spectral shift and lifetime, we initially describe and compare the different classes of luminescence nanothermometry. Subsequently, the various luminescent materials used in each case are discussed and the mechanisms at the root of the luminescence thermal sensitivity are described. The most important results obtained in each case are summarized and the advantages and disadvantages of these approaches are discussed.

A. Introduction A.1

a

Fluorescence Imaging Group, Departamento de F
ısica de Materiales C-04, Insitituto Nicol
as Cabrera, Facultad de Ciencias, Universidad Aut
onoma de Madrid, Madrid 28049, Spain. E-mail: [email protected] b
Institut National de la Recherche Scientifique-Energie, Mat
eriaux et T
el
ecommunications, Universit
e du Quebec, Varennes, QC, J3X 1S2, Canada. E-mail: [email protected]; Fax: +1-450-929-8102; Tel: +1514-228-6847 † This work was supported by the Universidad Aut
onoma de Madrid and Comunidad Aut
onoma de Madrid (Project S2009/MAT-1756), by the Spanish Ministerio de Educacion y Ciencia (MAT2010-16161) and by Caja Madrid Foundation.

Daniel Jaque obtained his Ph.D. in Physics at Universidad Aut
onoma de Madrid (Madrid, Spain) working on Multifunctional Solid State Lasers. After working for several years on the optical and conducting properties of nanostructured superconducting and metallic thin films, he co-founded the Fluorescence Imaging Group at the Universidad Aut
onoma de Madrid. His current research Daniel Jaque activity is focused on the development of fluorescent nanoparticles capable of thermal sensing at the nanoscale. He has been Distinguished Invited Professor at Heriot Watt University (UK) and Swinburne University of Technology (Australia). He has published over 225 publications and attended as invited speaker to 16 international conferences.

This journal is ª The Royal Society of Chemistry 2012

Why nanothermometry?

Nanothermometry aims to extract knowledge of the local temperature of a given system with sub-micrometric spatial resolution. Such knowledge is required for the complete understanding of micrometric and nanostructured systems whose dynamics and performance are strongly determined by temperature. The recent development of nanotechnology and nanomedicine brought about the appearance of a large number of such systems. It is difficult (if not even impossible) to enumerate all the systems (fields) that presently exploit nanothermometry. Nevertheless, we can initially identify three areas that can clearly

Fiorenzo Vetrone received his Ph.D. in Chemistry at Concordia University (Montreal, Canada) followed by postdoctoral fellowships funded by the Natural Science and Engineering Research Council (NSERC) of Canada and the Royal Society (UK). Fiorenzo Vetrone is currently an Assistant Professor of Nanobiotechnology at Universit
e du Qu
ebec, Institut National de la Recherche
Fiorenzo Vetrone Scientifique-Energie, Mat
eriaux et T
el
ecommunications (INRSEMT) in Varennes, Canada where his research activities are focused around multiphoton excited luminescent nanoparticles for use in the development of multifunctional nanoplatforms for diagnostic and therapeutics of various diseases. His scientific contributions, including over 50 publications and 20 invited presentations, have been recognized by various national and international awards. Nanoscale, 2012, 4, 4301–4326 | 4301

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

View Article Online

make use of nanothermometry and can benefit significantly by the further development of nanothermometry towards higher sensitivities and resolutions. The three areas covered in great detail in this review are micro/nano-electronics, integrated photonics and biomedicine. In micro/nano-electronics, the reduced dimensions of electrical conduction channels lead to relevant resistances in such a way that the Joule heating effect becomes non-negligible. Therefore, any small variation in the local resistance of the conduction channels (caused by either a local modification of the micro/ nano-conductor geometry or by the existence of scattering defects within it) could cause the appearance of a ‘‘hot-spot’’, i.e. a well-localized temperature increment. These ‘‘hot spots’’ can be the catalyst for performance deterioration or ultimately, an irreversible failure.1,2 The knowledge of the exact location of such thermal singularities is crucial to avoid catastrophic damage, which can be averted by either design modifications or by the controlled incorporation of micro-cooling devices.3 Unfortunately, the magnitude and location of these temperature singularities are difficult to be predicted a priori, since they depend not only on the device design but also on the quality of the integrated circuits. Thus in micro/nano-electronics, nanothermometry could become a pivotal tool for the detection of ‘‘hot-spots’’ in integrated circuits during real operating conditions. In integrated photonic devices, temperature is also a critical parameter where the performance of such systems depends on the optical properties of the constituting materials. For example, light control and manipulation in integrated photonic devices are achieved by the sub-micrometric controlled variation of the local refractive index, which is a physical property that is strongly temperature dependent.4 In fact, temperature not only affects the refractive index but also the optical gain of the integrated laser sources, spectral properties of the band-gap structures and finally, the optical conversion efficiency of nonlinear converters such as optical super-lattices.5 In this context, any variation in the local temperature within an integrated photonic device could strongly affect its performance or in the worst case, could even lead to irreversible damage.6 Again, using nanothermometry to glean knowledge of temperature singularities induced during device operation is paramount for the development of reliable, efficient and robust integrated photonic devices, making nanothermometry an indispensable tool. This can be extended to the case of new generation optofluidic devices in which light is used to manipulate micron sized objects (including living specimens), which are propagating within micro-fluidics through the use of optical forces.7 In this case, residual absorption of fluids at laser radiation wavelengths can cause heat loading and can ultimately lead to malfunctions of the device resulting from the onset of convection currents.8 Another area that could benefit greatly from the high-resolution and high-sensitivity of nanothermometry is biomedicine. It is well-known that in any biosystem, temperature is known to play a crucial role in determining its dynamics and properties.9,10 For example, temperature is one of the critical parameters determining cell division rates and hence, determining the rate of tissue growth.11 Moreover, it also drastically affects the mechanical, optical and structural properties of fundamental biomolecules such as proteins where they can undergo a denaturation process when their temperature deviates by even a few 4302 | Nanoscale, 2012, 4, 4301–4326

degrees above 37
C.12 Thus, to understand the dynamics of biosystems, the simultaneous monitoring of their temperature is crucial to elucidate the origin of the observed behavior. Aside from this fundamental interest, thermal sensing of biosystems is also vital for the early detection and treatment of many diseases. It is commonly accepted that one of the first signatures of any given illness (such as inflammation, cancer or cardiac problems) is the appearance of thermal singularities.13 In the particular case of cancer, it has been postulated that the thermal singularity associated with incipient tumors becomes detectable when they reach a size consisting of thousands of cancer cells, i.e. when the tumor size is well below 1 mm. Exploiting this phenomenon for cancer detection would constitute a major leap forward when compared with traditional imaging techniques such as tomography, for which the minimum detectable tumor size is on the order of several millimeters.14 This is especially important since the early detection of tumors would allow for significantly earlier medical intervention. High-resolution thermal sensing is also required in cancer therapeutic processes such as hyperthermia, which is based on externally inducing an increment in the tumor’s temperature up to cytotoxic levels (43–45
C).15 This temperature increase should be performed in a controlled manner such that the thermally induced damage of surrounding tissues would be minimized. This can only be accomplished if the temperature of the tumor can be continuously monitored during the hyperthermia treatment such that it could be immediately terminated when cytotoxic levels are reached, thereby avoiding excessive (and unnecessary) heating. This thermal sensing platform is quickly becoming a reality with the recent development of nanothermometers, which due to their reduced size can be incorporated into tumors and cancer cells and subsequently provides the requisite thermal information. Finally, it should be noted that a universal thermal sensor does not exist since different applications require different properties. For example, the requirements that must be satisfied by nanothermometers used in biological applications are very different from those necessary for the thermal imaging of opto-fluidic devices. In bio-medicine, the nanothermometers should be nontoxic, water soluble and very stable under light irradiation so that toxic components are not delivered to the cells. In addition, the final temperature resolution achievable should be on the order of 0.2
C since temperature variations as low as 1
C are already relevant in biological dynamics. In the case of thermal imaging of opto-fluidic devices the typical temperature changes caused by laser radiation are on the order of several degrees so that the final temperature resolution of the nanothermometers is not required to be below 1
C. In opto-fluidics the toxicity is not a critical issue but, due to the high laser intensities achieved within microchannels, the physical and chemical stability with respect to laser irradiation becomes an essential aspect. A.2

Nanothermometry techniques: a brief summary

Motivated by the numerous fundamental and practical applications, many research groups have presently focused their efforts on the development of diverse techniques capable of achieving thermal sensing with sub-micrometric resolution.16 This has led to the development of several prospective avenues for nanothermometry and consequently, to the elaboration of several This journal is ª The Royal Society of Chemistry 2012

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

View Article Online

nanothermometry techniques. These techniques have been traditionally classified into three main groups: electrical, mechanical, and optical. The working principles of electrical techniques are the same as those governing traditional thermocouples and thermistors, in which the temperature reading is achieved from variation of the resistance, voltage, conductivity or electrical capacity of a given conducting probe.17 These traditional principles have also been applied in nanothermometry due to the rapid development of micro/nano-fabrication techniques, which have driven the dimensions of the electrical thermal sensor down to the submicrometric scale.18 By using the same principles of atomic force microscopes, these sub-micrometric thermal sensors were scanned (in pseudo-contact) over the surface to be thermally imaged (known as scanning thermal microscopy).19 The spatial resolution reached using this technique can, in principle, be as good as that of atomic force microscopes. Indeed, micro-thermocouples based on Pt–Cr thin films have been demonstrated to be capable of thermal sensing with sub-100 nm resolution over areas of hundreds of mm2.18 The scanning area could be increased up to a few mm2 by utilizing two-dimensional arrays of micro-thermocouples although in this case serious difficulties in the microfabrication procedure could be problematic.20 Despite these promising results, there are several potential drawbacks in using electrical nanothermometry. First, it is a surface technique and thus can be only used to obtain two-dimensional thermal images. In addition, the technique requires a physical contact between the thermal probe and the system under study, which can lead to complex heat transfer fluxes ultimately resulting in a modification of the system temperature due to the interactive measurement procedure. Finally, the advanced technology required for the fabrication of the thermal probes increases the costs of the required experimental set-up. Mechanical nanothermometry is similar to micro-thermocouple based scanning thermal microscopy but replaces the micro-thermocouple with a bi-material cantilever. Due to the different mechanical properties of the two materials that make up the cantilever, any temperature change in its surroundings results in cantilever deflection. Based on this simple principle, thermal sensitivities reaching the mK range have been demonstrated while maintaining the spatial resolutions typical of atomic force microscopes.21 As in the case of electrical nanothermometry, mechanical nanothermometry can be only used to obtain thermal images of surfaces and does not offer the possibility of reconstructing three-dimensional thermal images. In addition, due to the requirement of physical contact between the thermal probe and the specimen, the thermal measurement can also be affected by artificially induced heat fluxes.22 Optical nanothermometry is based on the analysis of temperature-induced changes in the inherent passive optical properties of materials. Optical nanothermometry includes a great variety of techniques. Some of them are based on temperature-induced changes in the optical path length of a transparent system due to the temperature dependence of both refractive index and thickness.17,23 Interferometric techniques exploit these thermally induced changes to reconstruct two-dimensional images of transparent materials with spatial resolutions close to the micron and with thermal sensitivities that in the best of the cases can surpass the mK limit.23 The thermal dependence of the refractive This journal is ª The Royal Society of Chemistry 2012

index consequently makes the reflectance of any interface sensitive to temperature variations. This is exploited by reflectance based thermometry techniques that relate temperature to the polarization and intensity of light reflected from a given surface. In combination with high numerical aperture optics and adequate choice of illumination wavelengths, real time thermal images of micro-resistors and micro-coolers with sub-micrometric spatial resolutions and with thermal sensitivities of few tens of mK have already been obtained.24 One of the limitations of this technique is that it requires the existence of an optical interface (surface) and consequently, can be only used for twodimensional thermal imaging. Traditionally, scanning Raman thermal microscopy has been also cataloged as an optical method. It is based on the analysis of the properties of Raman (vibration) modes that are affected by temperature variations through thermally induced structural changes. Raman thermal microscopy, therefore, requires the scanning of an excitation laser beam over the system to be thermally imaged. A subsequent spectral analysis provides the Raman images of the system in terms of the intensity, position and width variations of the Raman modes, which are in turn translated into temperature units. This procedure has been successfully applied to acquire thermal images of micro-heating devices and high power integrated laser devices achieving thermal and spatial resolutions no better than 10
C and 1 mm, respectively.17,25,26 Raman thermal imaging appears to be a universal technique since any material shows a Raman spectrum under laser excitation. In practice, however, this technique is only applicable to those materials showing large Raman efficiencies and that present high thermal sensitivity of the Raman modes restricting it to a few number of systems. Finally, the Pyrometric and Infrared (IR) Thermometry techniques should be also mentioned. Both are based on the measurement of the electromagnetic radiation emitted from a given surface at a specific spectral range (typically long wavelengths). The temperature of the surface can then be obtained by analyzing its emissivity within the Planck’s blackbody theory. Again, this is a universal technique that has already been applied to obtain thermal images of living specimens, optically pumped laser sources, integrated microelectronics and fluids. However, it too has several drawbacks such as the impossibility of obtaining three-dimensional images and the requirement of using long wavelengths (>1 mm).27–29 This last point necessitates that expensive infrared detectors are required and obviously limits the final spatial resolution that can be achieved (ranging from 5 to 30 mm). A.3

Luminescence nanothermometry

Luminescence is the emission of light from a given substance, occurring from electronically excited states that have been populated by an external excitation source (optical radiation, in the case of photoluminescence). The properties of the emitted photons depend on the properties of the electronic states involved in photon emission.30 These, in turn, depend on the local temperature and thus luminescence nanothermometry exploits the relationship between temperature and luminescence properties to achieve thermal sensing from the spatial and spectral analysis of the light generated from the object to be Nanoscale, 2012, 4, 4301–4326 | 4303

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

View Article Online

thermally imaged. The grouping of luminescence nanothermometry into different classes is based on the particular parameter of luminescence which is analyzed and from which the thermal reading is ultimately extracted. Fig. 1 schematically depicts the six parameters that define the luminescence emission of a given material: intensity, band-shape, spectral position, polarization, lifetime and bandwidth. Fig. 1 also qualitatively demonstrates how the luminescence emission spectrum is modified when each of these parameters is varied. Thus, based on these variations it is possible to define the following luminescence nanothermometry sub-classes: Intensity Luminescence Nanothermometry (ILNth). In this case, thermal sensing is achieved through the analysis of the luminescence intensity. When temperature changes, there is an overall change in the number of emitted photons per second such that the emission spectrum becomes less (or more) intense. Temperature induced changes in the luminescence intensity are generally caused by the thermal activation of luminescence quenching mechanisms and/or increases in the non-radiative decay probabilities. Band-Shape Luminescence Nanothermometry (BSLNth). The term ‘‘band-shape’’ refers to the relative intensity between the different spectral lines that make up the luminescence spectrum. Thermally induced variations in the band-shape usually take place when the electronic states from which emission is generated are very close in energy such that they are thermally coupled. It can be also present in mixed systems, i.e. systems containing more than one class of emitting centers.

Fig. 1 Schematic representation of the possible effects caused by a temperature increment on the luminescence. Red lines correspond to higher temperatures.

4304 | Nanoscale, 2012, 4, 4301–4326

Spectral Luminescence Nanothermometry (SLNth). It is based on the analysis of the spectral positions of the emission lines, which are unequivocally determined by the energy separation between the two electronic levels involved in the emission. In turn, this depends on a large variety of temperature dependent parameters of the emitting material including refractive index and inter-atomic distances (density). Thus, in any emitting material the spectral positions of the luminescence lines are expected to be temperature dependent, and this is exploited by spectral luminescence nanothermometry to translate spectral shifts into temperature. Polarization Luminescence Nanothermometry (PLNth). In anisotropic media, the emitted radiation is generally non-isotropically polarized and consequently, the shape and intensity of emitted radiation are strongly dependent on its polarization. This allows for the definition of the ‘‘polarization anisotropy’’ parameter, which is the ratio between the luminescence intensities emitted at two orthogonal polarization states. As a result, polarization luminescence nanothermometry is based on the influence of temperature on this polarization anisotropy. Bandwidth Luminescence Nanothermometry (BLNth). The width of the various emission lines that make up any luminescence spectrum is determined by the properties of the material (such as the degree of disorder) and temperature. It is wellknown that as the temperature of a luminescent material is increased, a corresponding increase in the density of phonons occurs resulting from the spectral contribution of homogeneous line broadening. Generally, in the vicinity of room temperature, homogeneous line broadening leads to a linear relationship between bandwidth and temperature. The change in the bandwidth of the luminescence spectra is exploited in bandwidth luminescence nanothermometry to achieve a thermal reading. Lifetime Luminescence Nanothermometry (LLNth). Luminescence lifetime, sf, is defined as the time that the emitted luminescence intensity decays down to 1/e of its initial value after a pulsed excitation. This is an indication of the total decay probability of the emitted intensity (indeed this probability is defined as the inverse of the luminescence lifetime). Decay probabilities from electronic levels depend on a great variety of factors and many of them are related to temperature (such as phonon assisted energy transfer processes and multiphonon decays). This temperature dependence makes it possible to extract temperature readings from the determination of the luminescence lifetime. Thus, luminescence nanothermometry provides several options to achieve thermal sensing from the analysis of the emission spectrum generated by the system under study. It should be noted that the thermal sensitivity would vary from system to system and would obviously depend on the magnitude of the thermally induced spectral variations. Clearly, systems that show remarkable changes in luminescence properties for small temperature changes will provide the largest thermal sensitivities. Similarly, spatial resolution in luminescence nanothermometry is difficult to be estimated as it is determined by the spatial resolution at which the emission spectrum is recorded. High spatial resolution would require the active luminescence volume to be as small as possible, i.e. it requires acquisition of the luminescence emission spectrum with ‘‘high spectral resolution’’. For example, if the luminescent system to be thermally imaged This journal is ª The Royal Society of Chemistry 2012

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

View Article Online

possesses luminescent centers that are homogeneously distributed, achieving high spatial resolution would require the use of confocal luminescence microscopy (based on the combination of high numerical optics and spatial filters) or of near-field optical microscopes (based on the spatial scanning of low aperture optical fibers). On the other hand, luminescence nanothermometry of non-luminescent systems (such as fluids, electrical circuits or living cells) is also possible yet it requires the introduction of luminescent micro- and nanoparticles with a temperature dependent luminescence within the system to be imaged. Ultimately, the spatial resolution of the thermal measurement would be limited by the spatial dimension of the luminescent particle (provided it has been discretely incorporated into the system). Recent achievements in nanoparticle synthesis have made it possible to prepare highly efficient nanoparticles (sizes down to few nanometers), which after proper surface functionalization, can be dispersed in solutions allowing for their facile incorporation in different systems. Resulting from this great diversity in luminescence nanothermometry, numerous approaches and experimental demonstrations of sub-micrometric thermal images of a great variety of systems have appeared. In the next section we review the most relevant ones where we classify them in terms of the luminescence nanothermometry sub-classes. Moreover, the different thermal and spatial resolutions achieved by the different approaches are compared and discussed.

B. Luminescence thermal images B.1 Intensity luminescence nanothermometry As described in Section A.3, intensity luminescence nanothermometry is based on the thermal variation of the luminescence intensity to achieve a temperature reading. The basic elements of this technique are luminescence systems with a hypersensitive thermal response, i.e. whose luminescence intensity is strongly affected because of small temperature variations. Previous work reporting on thermal imaging and sensing based on luminescence intensity analysis can be grouped differently depending on the nature of the luminescence centers used as thermal probes. ILNth has been previously demonstrated in different systems including semiconductor nanocrystals (hereafter quantum dots, QDs), luminescent organic dyes, rare earth ions, transition metal ions, luminescent nanogels and luminescent polymers. In the following sections we will summarize the most relevant results obtained for each case. B.1.1 QD-based intensity luminescence nanothermometry. Semiconductor QD nanocrystals are perhaps one of the most ubiquitous optical nanoprobes due to their excellent photostability, large luminescence quantum yield and also because of their size-tunable absorption and emission wavelengths.31 Recombination of electron–hole pairs within the QD volume gives rise to a typical near Gaussian emission band whose peak wavelength depends on the QD material and the QD size.32 Most of the QDs used for imaging applications have their averaged emission wavelength in the visible range and as can be observed in Fig. 2(a), this emission band is strongly modified by temperature variations.31 One of the most remarkable changes is This journal is ª The Royal Society of Chemistry 2012

Fig. 2 (a) Luminescence of CdSe QDs as obtained at different temperatures. The inset shows the integrated luminescence intensity as a function of temperature. Dots are experimental data and the solid line is a linear fit. Data were extracted from ref. 33. (b) Scanning electron microscope image of a 200 nm sized silica sphere covered with CdSe/ZnS QDs. Figure 2(a) is reprinted with permission from G. W. Walker, V. C. Sundar, C. M. Rudzinski, A. W. Wun, M. G. Bawendi and D. G. Nocera, Appl. Phys. Lett., 2003, 83, 3555. Copyright 2003, American Institute of Physics. Figure 2(b) is reprinted with permission from L. Aiguoy, B. Samson, G. Julie, V. Mathet, N. Lequeux, C. Nı. Allen, H. Diaf and B. Dubertret, Rev. Sci. Instrum., 2006, 77, 063702. Copyright 2006, American Institute of Physics.

produced in the luminescence intensity where as the temperature increases the QD emission becomes weaker.33 This seems to be a universal phenomenon in QDs and is attributed to the activation of phonon assisted processes as well as to the presence of thermally assisted energy transfer processes from bulk to surface (non-radiative) states. A very interesting feature of thermally induced luminescence quenching in QDs is its linearity. As can be observed in the inset of Fig. 2(a), the emitted intensity of the QD decreases almost linearly in the vicinity of room temperature. Although the results included in Fig. 2 correspond to CdSe/ZnS QDs, similar linearities have been also observed in other systems such as CdTe QDs.34,35 Linearity is an outstanding feature for thermal sensing since it ensures a constant thermal sensitivity in the whole working temperature range of the thermal probe. Indeed a very simple calibration procedure is required to obtain thermal sensing from the analysis of the QD emission intensity. There are two different approaches to obtain luminescence images based on the analysis of QD luminescence. The first approach relies on the controlled scanning of a single QD over the system to be thermally imaged while recording its luminescence intensity. This approach was first demonstrated by Aigouy et al. who modified an AFM tip with a subwavelengthsized silica sphere covered with CdSe/ZnS QDs (see Fig. 2(b)).36 The continuous monitoring of QD luminescence and the Nanoscale, 2012, 4, 4301–4326 | 4305

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

View Article Online

subsequent translation into temperature units could generate the thermal image of the area under scanning. This technique has been previously used for optical transmission of sub-micrometric objects and its application in thermal sensing and imaging is imminent. The main disadvantage of this approach is that it is a pseudo-contact method so that it is not clear how the vicinity of the tip can modify the actual temperature distribution. In addition, this technique is restricted to surface sensing. The second approach involves a reduced technological complexity and involves the artificial incorporation of the QDs into the studied system. The spatial variation of the QD luminescence intensity (that is recorded by a conventional or by a scanning confocal microscope) would allow for thermal imaging to be achieved. A representative example of this approach is the pioneering work of Han et al. who used CdSe QDs to monitor the temperature of cancer cells incubated in a mixed solution of QDs and metallic nanoparticles.34 The authors designed a double beam experiment as that shown in Fig. 3(a). One near-infrared (NIR) laser beam (820 nm) was used as the heating source since it promoted the excitation of surface plasmons of the metallic nanoparticles. A second light source (halogen lamp) illuminated the whole incubation dish. The subsequent QD luminescence intensity was recorded by a CCD camera and was used to obtain thermal images of the cell population during NIR illumination. An example is given in Fig. 3(b) where it is evident that at the central region (where the NIR laser spot is focused) the QD intensity is decreased. This is a clear indication that the metallic nanoparticle mediated laser-induced local heating occurs (see the darker region indicated by a black

Fig. 3 (a) Schematic of the experimental set-up used for QD based ILNth of a cell culture during gold nanoparticle mediated heating by an infrared laser beam, (b) QD intensity luminescence image obtained during the heating process (20 s after starting). The black arrow indicates the position of the heating laser spot. Note that the intensity reduces radially with respect to the focus. (c) Micrographs of darkfield (left) and luminescent field (right) indicating that the cells are still attached to the dish after heat treatment. The luminescence image reveals cell death in the surroundings of the laser heating focus. Reprinted with permission from ref. 34. Copyright 2009 Springer.

4306 | Nanoscale, 2012, 4, 4301–4326

arrow in Fig. 3(b)). After a proper calibration, the authors concluded that the maximum temperature at the focal point was as large as 50
C. This temperature increment was high enough to produce cell ablation at the focal point as can be observed in Fig. 3(c). This was pioneering work since it demonstrated the first application of QD based luminescence intensity nanothermometry in a biosystem. The final temperature resolution achievable by using QDs is determined by the temperatureinduced luminescence quenching rate. In the case of CdSe QDs, this quenching rate is close to 1% per
C such that sub-degree thermal resolutions are easily achievable. An alternative approach that, because of its novelty, deserves to be mentioned in detail was proposed by Lee et al.37 They described a reversible nanothermometer built from two types of nanoparticles connected by a polymer that was acting as a molecular spring. Fig. 4(a) shows a schematic representation of the system proposed by Lee et al. The nanoscale superstructure consisted of an Au nanosphere covered by a poly(ethyleneglycol) (PEG) film with a thickness of few nanometers. The outer side of the PEG film was conjugated with 3.7 nm CdTe QDs and they were excited through plasmon–exciton interactions. The nanothermometer displayed the characteristic exciton luminescence of CdTe QDs (at approximately 550 nm) following optical excitation of the surface plasmon resonance of the Au nanoparticle (at around 633 nm). The efficiency of this plasmon–exciton energy transfer is strongly dependent on the distance between the Au surface and CdTe QDs, i.e. on the PEG film thickness. Since temperature variation from 20 to 50
C results in an expansion of the PEG molecule, the PEG thickness varies substantially within this temperature range, producing a concomitant change in the luminescence output (it becomes temperature dependent). Lee et al. demonstrated the ability of the superstructure depicted in Fig. 4(a) for ILNth by analyzing the luminescence intensity of the Au–PEG QD system when it was subjected to a time dependent temperature variation as schematically indicated in Fig. 4(b). As can be observed in Fig. 4(c), the CdTe QD luminescence intensity

Fig. 4 (a) Schematic representation of the superstructure proposed by Lee et al. for nanoscale thermometry that consists of a metallic nanoparticle covered with a PEG film and CdTe QDs. (b) Time evolution of the superstructure temperature and (c) corresponding time evolution of the CdTe luminescence intensity. Data were extracted with permission from ref. 37.

This journal is ª The Royal Society of Chemistry 2012

View Article Online

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

is shown as a function of time and the Au–PEG QD responds instantaneously to the temperature variations. From Fig. 4(b) and (c) it is possible to estimate a thermally induced luminescence quenching rate of 0.6% per
C for the Au–PEG QD system that is quite close to that of QDs under direct exciton excitation. B.1.2 Dye-based intensity luminescence nanothermometry. The term ‘‘dye’’ refers to organic compounds that show a strong luminescence (usually in the visible) when optically excited with short wavelength radiation (usually in the ultraviolet or blue). Light emission is generated by the electronic transitions taking place between vibrational energy levels of different electronic states. The emission and absorption bands of dyes are both strongly dependent on the particular composition of the compound. Furthermore, for a particular dye, the spectral properties of the luminescence band depend on many factors, such as the solvent, concentration, pH and temperature. As a general rule, the luminescence intensity generated by organic dyes decreases as the temperature increases. This can be explained in terms of the activation of multiphonon de-excitations as well as the thermal promotion of electrons up to vibrational levels of higher energies, from which radiative deexcitation probability is reduced.38 The magnitude of the thermal dependence varies from dye to dye where some dyes have a much stronger sensitivity to temperature. This is the case of Rhodamine B, which belongs to the xanthene group and possesses a luminescence band centered around 550 nm (see Fig. 5(a)). Rhodamine is probably the most popular dye as a result of its superior chemical stability as well as its relatively high luminescence efficiency. Fig. 5(b) shows the luminescence intensity generated from a UV excited Rhodamine B solution as a function of the solution temperature. As can be observed, temperature induces a remarkable luminescence quenching. Close to room temperature (25
C), the luminescence intensity reduces linearly with temperature at a rate close to 2% per
C, which is similar to that found for QDs. There are several reports in the

Fig. 5 (a) Room temperature luminescence spectrum of Rhodamine B. (b) Integrated emission of Rhodamine B as a function of temperature. Dots are experimental data and the solid line is a guide for the eyes.

This journal is ª The Royal Society of Chemistry 2012

literature dealing with thermal imaging from the analysis of the luminescence intensity variations of Rhodamine B. They can be classified into two types, depending on how the Rhodamine was introduced in the imaged system: liquid solution or thin film. In the first case an aqueous solution containing Rhodamine B was introduced into a microfluidic device that was optically excited at 525 nm. The spatial variation of the dye emission intensity was recorded by placing the microfluidic device in a conventional luminescence microscope. Ross et al. demonstrated that using this facile technique they were able to measure the temperature of the microfluidic device with sub-micrometric spatial resolution.39 In fact, the authors were able to measure two-dimensional temperature distributions resulting from Joule heating in a variety of electrokinetically pumped microfluidic systems. Fig. 6(a) shows the thermal image of a multi-branched microfluidic circuit that is schematically shown in Fig. 6(b) where a voltage difference of 1000 V was applied between the two fluid reservoirs (light gray circles). The area enclosed by the dashed line in Fig. 6(b) is the field of view of the thermal image where Ross et al. stated that the maximum temperature resolution achievable with their technique could surpass 0.1
C.39 These outstanding thermal sensitivities allowed the observation of complex phenomena, in which each branch of the fluid circuit carries a different fraction of electric current and, therefore, suffers from different thermal loading. In the second approach, a thin film containing the luminescent Rhodamine dye was deposited onto the surface to be thermally imaged. The physical contact between the surface and film creates a temperature distribution on the film that reproduces the

Fig. 6 (a) Color coded thermal image of a multi-branched microfluidic circuit. (b) Schematic showing a layout of the multi-branched circuit. Reproduced from ref. 39 with permission from The Royal Society of Chemistry.

Nanoscale, 2012, 4, 4301–4326 | 4307

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

View Article Online

surface of interest. Thus, surface thermal imaging was achieved from the visualization of the luminescence intensity of the thin film. This approach has been applied to obtain thermal images of microfluidic systems subjected to Joule heating as well as microand nanowires under real operation conditions.40,41 The latter case is of special relevance because of the low dimensions of the heating source. Fig. 7(a) shows the basic experimental set-up used by L€ ow et al. where a nickel nanowire is covered by a Rhodamine containing thin film that is optically excited in the 530–550 nm range with a mercury lamp.40 The film luminescence (containing thermal information through its intensity variations) was collected by a microscope objective and analyzed by a CCD detector. In the absence of applied voltage, the surface temperature is homogeneous and the luminescence intensity of the thin film is constant over the entire surface. Conversely, when an electrical current flows thorough the nanowire (Fig. 7(b)), Joule heating appears and the thin film surface temperature is no longer homogeneous, creating a reduction in the luminescence at the nanowire location (see the luminescence image included in Fig. 7(b)). From the analysis of the decrease in luminescence intensity, it is possible to determine the local temperature of the nanowire (as it is schematically illustrated in Fig. 7(c)). In this case the thermal image corresponds to that of an 80 mm-long, and 2 mm-wide nickel wire when a current of 20 mA is flowing. The thermal image clearly demonstrates that this technique allows for the measurement of local temperatures in excess of 90
C while keeping sub-micrometric spatial resolution. One of the clear drawbacks of using luminescent thin films is that they allow only

Fig. 7 (a) Depiction of a nanowire on a Rhodamine B thin film. (b) Spatial variation of the Rhodamine B emission intensity when a 20 mA current is applied along the nanowire (at the center of the image). (c) Thermal image of the nanowire as obtained from the luminescence of the image shown in (b). Reproduced from ref. 40 with permission.

4308 | Nanoscale, 2012, 4, 4301–4326

for the thermal imaging of surfaces, i.e. 2D thermal imaging. Moreover, it should be noted that the existence of a physical contact between the sensing film and the system under investigation could modify the actual surface temperature such that the thermal image could actually differ from that of the device in normal working conditions, i.e. in the absence of the sensing film. B.1.3 Rare earth-based intensity luminescence nanothermometry. Rare earth atoms lie within the sixth row of the periodic table, from cerium to ytterbium, inclusive. The electronic configuration of trivalent ions is considered as a singularity since it shows an unfilled 4f shell, partially screened from the environment by the electrons in the 5s2 and 5p6 shells.30 Optical transitions involving different levels within the 4f shell are normally forbidden. However, they may become partially allowed when the ions are located within a matrix, by virtue of the crystal field of the local environment. Due to the partial electronic screening from the environment, the luminescence lines of rare earth ions are typically very narrow and the corresponding lifetimes are relatively long, from a few ms to a few ms. Due to these properties, rare earth ions have been used in low threshold, high gain solid state lasers. The intensity of the luminescence lines of rare earth ions depends on several parameters among which, temperature is one of the most critical ones.30 There are a large number of mechanisms that could link the intensity of rare earth luminescence with temperature including activation of the multiphonon decay probability, activation of energy transfer between rare earth ions or quenching centers, population re-distribution due to Boltzmann statistics, appearance of phonon assisted Auger conversion processes and thermal enhancement of energy transfer processes between rare earth ions and the host levels or charge transfer states.30 Due to the simultaneous interaction of those different phenomena, it is very difficult to predict and understand how the luminescence intensity of rare earths is influenced by temperature. Indeed the thermal behavior of different lines could differ by a great amount to that observed for other lines. This is illustrated in Fig. 8 that includes the change in intensity observed for the various lines of the 5Dj / 7Fi transitions of Eu3+ doped in

Fig. 8 Schematic representation of the energy level diagram of trivalent europium (Eu3+) ions in La2O2S (left). The figure on the right shows the temperature dependence of the luminescence intensity associated with the different transitions indicated by the arrows on the energy level diagram. Data extracted from ref. 44 with permission from the American Institute of Physics.

This journal is ª The Royal Society of Chemistry 2012

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

View Article Online

La2O2S.42–44 This demonstrates that each line is affected in a different temperature range, so that this particular compound can be used to detect temperature changes in different ranges from 100 to 500 K. The inherent ability of Eu3+ ions for temperature sensing in the vicinity of room temperature has been previously exploited for the microscopic detection of thermogenesis in single living cells. The idea proposed by Suzuki et al. was to establish a physical contact between a HeLa cancer cell and a glass micropipette filled with an Eu3+ containing solution (see Fig. 9(a)).9 This compound suffers from a relevant temperature induced luminescence quenching in the biophysical range 20–50
C following optical excitation in the UV, as can be observed in Fig. 9(b). The intensity reduction rate was found to be close to 3% per
C, superior to that previously described for luminescent dyes, such as Rhodamine. The intensity of the visible emission of the Eu3+ compound is registered in real time so that the authors were able to follow the time evolution of the cell’s temperature. This system was used to detect, at the single cell level, the intracellular temperature changes caused by a modification in the intracellular free Ca2+ concentration. Fig. 9(c) shows the time variation of the Eu–TTA luminescence intensity. The arrow indicates the moment at which the Ca2+ concentration was externally varied

by adding a 1 mM Ca2+ solution and the arrowhead indicates the moment at which the initiation of positive thermogenesis was detected. The system proposed by Suzuki et al. was able to detect real-time intracellular temperature variations as small as 1
C in the biophysical range.9

Fig. 9 (a) Schematic illustration of the set-up used by Suzuki et al. for the measurement of the temperature of a single cell by using pipettes filled with a luminescent Eu3+ solution. (b) Temperature variation of the luminescence generated by the Eu3+ containing solution. (c) Time variation of the Eu3+ luminescence intensity and cell temperature variation. The arrow indicates the moment at which the Ca2+ concentration was externally varied. The arrowhead indicates the moment at which the initiation of positive thermogenesis was detected. Reprinted with permission from ref. 9. Copyright 2007 Elsevier.

Fig. 10 (a) Depiction of the luminescent nanogel proposed for ILNth by Gota et al. indicating the temperature induced shape change in the nanogel. (b) Luminescence intensity generated by the nanogel schematically shown in (a) as a function of temperature. (c) Phase contrast and luminescence images of living COS7 cells containing the nanogel of (a) at three different culture medium temperatures. (d) Time evolution of intracellular luminescence intensity and temperature after the addition of camptothecin to the culture medium at t ¼ 0. Reprinted with permission from ref. 45. Copyright 2009 American Chemical Society.

This journal is ª The Royal Society of Chemistry 2012

B.1.4 Luminescent nanogel-based intensity luminescence nanothermometry. The potential use of luminescent nanogels (system typically composed of several synthetic polymers or biopolymers which are chemically or physically cross-linked) as nanothermometers was initially proposed and demonstrated by C. Gota et al. in 2009.45 The base of their temperature sensitive nanogel was the thermoresponsive poly-NIPAM combined with the DBD-AA water sensitive fluorophore. The resulting compound was treated by an emulsion polymerization technique using a cross-linker (MBAM). The working principles of the nanogel based thermometer are schematically shown in Fig. 10(a). At low temperature (below 30
C) the nanogel swells by absorbing water into its interior and this inner water causes a luminescence quenching of the DBD-AA units. When the nanogel temperature is raised above 30
C the nanogel shrinks, causing the release of water molecules. As a consequence, the DBD-AA units are no longer quenched and the nanogel luminescence increases remarkably. This is illustrated in Fig. 10(b) in which the 515–550 nm luminescence intensity generated from the nanogel under 488 nm excitation is shown in the 20–40
C range. At first glance, the luminescent nanogel has the outstanding characteristic of a luminescence enhancement with temperature. This is at variance with the previously described cases (such as QDs and rare earth ions) allowing for the measurement of heating processes without a relevant loss of luminescence intensity and, hence, of thermal sensing accuracy. A particular disadvantage is that the nanogel intensity becomes highly temperature sensitive only in a reduced temperature range (27– 35
C). Although this limits its application to a very reduced temperature range, it allows for the investigation, for example, of

Nanoscale, 2012, 4, 4301–4326 | 4309

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

View Article Online

intracellular temperature variations. This possibility was demonstrated by Gota et al. and their results are summarized in Fig. 10(c) that shows the phase contrast and luminescence images (at three different culture medium temperatures) of living COS7 cells containing the nanogel.45 It is observed that the nanogel emission was strong enough to provide luminescence images only at medium and high temperatures (29 and 35
C, respectively), in accordance with the calibration curve shown in Fig. 10(b). Fig. 10(d) shows the time evolution of intracellular luminescence intensity and temperature after the addition of camptothecin to the culture medium at t ¼ 0, which is expected to cause intracellular heating due to the early response of apoptosis such as DNA cleavage. Gota et al. were able to detect this intracellular heating and found that the pattern of temperature variation was different among the cells, a behavior that they explained in terms of the cell-division-dependent effects of camptothecin. Moreover, they estimated the temperature resolution achievable using their nanogel-based nanothermometer and in the best of cases, they approached 0.3
C. Data included in Fig. 10 clearly reveal the excellent potential of nanogel-based luminescence nanothermometry for the future study of intracellular dynamics and single cell manipulation and therapy. B.1.5 Luminescent polymer-based intensity luminescence nanothermometry. A luminescent polymer is a large molecule (macromolecule) composed of repeating structural units (usually connected by covalent chemical bonds) that typically shows visible luminescence (500–600 nm) when optically excited by UV radiation. Although luminescent polymers show relatively low luminescence efficiencies (quantum yield values well below 0.4),46 they are becoming quite popular for luminescence imaging of fluids and biological systems due to their very good solubility in water. In polymers, luminescence is explained in terms of the existence of luminescent monomers within the macromolecule. The luminescence intensity of such monomers is affected by a great variety of parameters such as microenvironmental polarity and symmetry, strength as well as the number of chemical bonds in their surroundings.46–50 As a consequence, any change in the structural properties of the luminescent polymer would cause a large variation in the resulting luminescence intensity. Some of the most relevant examples of polymer luminescence nanothermometers are those based on N-isopropylacrylamide (NIPAM). As reported by Uchiyama et al., water dissolved NIPAM based luminescent polymers prepared under certain conditions undergo a temperature-induced phase transition at around 32
C.51 Therefore, the luminescence intensity generated by the luminescent monomers increases drastically, as can be observed in Fig. 11(a), in which we also included the polymer emission band obtained at different temperatures (inset). Data included in Fig. 11(a) correspond to the so-called poly(DBD-AEco-NIPMAM) compound for which the sharp increment on the luminescence intensity occurs as a consequence of the change in the microenvironmental polarity during the temperature-induced phase transition.52 The optical contrast of thermal images based on such polymers is expected to be great since the temperature-induced phase transition produces an intensity increment larger than one order of magnitude. Unfortunately, such an amazing intensity 4310 | Nanoscale, 2012, 4, 4301–4326

Fig. 11 (a) Temperature dependence of the luminescence intensity generated from a water solution of poly(DBD-AE-co-NIPMAM). The inset shows the luminescence spectra at different temperatures. (b) Temperature dependence of the luminescence intensity generated from luminescent polymers based on N-alkylacrylamide and fluorophore units. The inset shows the luminescence spectra at different temperatures. (c) Digital photos demonstrating the remarkable temperature increment in the luminescence intensity of an aqueous solution of the N-alkylacrylamide based polymers. Reprinted with permission from ref. 51 and 52. Copyright 2004 and 2003 American Chemical Society.

enhancement is produced only in a very small temperature range (21–24
C), which limits its application for thermal imaging to systems with larger temperature variations. However, Uchiyama et al. overcame this limitation the following year and reported on the temperature-induced luminescence intensity increment in polymers based on N-alkylacrylamide and fluorophore units.51 The UV-excited emission spectra are shown in the inset of Fig. 11(b) and similar to results from Fig. 11(a), the luminescent intensity increases with temperature by more than one order of magnitude. Although the magnitude of the luminescence intensity increment observed in NIPMAM and N-alkylacrylamide based polymers is very similar, the former shows that the intensity variation is produced in a larger temperature range (10– 40
C), thus it can be used for thermal sensing in a wider temperature range. Digital photos included in Fig. 11(c) show the remarkable temperature increment in the luminescence intensity of an aqueous solution of the N-alkylacrylamide based polymers. These pictures also reveal that even for temperatures close to 40
C, the polymer solution remains stable without any evidence of precipitation. Despite the excellent results and perspectives described in this section, it should be noted at this point that the use of luminescence intensity as the thermal indicator has several drawbacks. Perhaps the main shortcoming of the systems described above is that the luminescence intensity depends on temperature but also on the local concentration of emitting centers. Thus, when performing thermal imaging based on luminescence intensity, it is a likely possibility that the observed luminescence intensity This journal is ª The Royal Society of Chemistry 2012

View Article Online

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

variations could not only be caused by temperature variations but also by fluctuations in the local concentration of emitting centers. This, obviously, would cause a false thermal image. The thermal imaging methods described in the forthcoming section have been developed in an attempt to overcome this problem by extracting temperature information from a spectral magnitude that is independent of the local concentration of emitting centers. B.2 Band-shape luminescence nanothermometry Band-shape luminescence nanothermometry (BSLNth) is generally based on luminescent systems whose luminescence spectra consist of several lines/bands with a relative intensity that is strongly dependent on temperature. There are two main operation schemes for BSLNth. In the first approach, the luminescence lines/bands are generated by different emitting centers so that the temperature induced band-shape change arises from the different thermal quenching of each center or from the thermally induced changes in the energy transfer rates among these centers. In the second approach, the different lines/bands that contribute to the luminescence spectra are generated by a unique luminescent center. In this case, the modification in the luminescence band-shape is generally caused by a thermally induced population re-distribution among the different energy levels of the emitting center. Of note, in both cases the relative intensity of the luminescence lines/bands depends only on temperature but not on the local concentration of emitting centers. This represents a significant advancement over intensity luminescence nanothermometry since the temperature reading is not affected by fluctuations in the concentration of luminescent centers. There are a great variety of examples in the literature that report on luminescent systems capable of band-shape dependent thermal sensing based on either of these two schemes. These systems can be composed of rare earth ions, quantum dots and organic dyes. Below we summarize the most relevant results in each of these cases. B.2.1 QD-based band-shape luminescence nanothermometry. Conventional QDs show a simple luminescence spectrum consisting of a single broad luminescence band correlated to the lowest energy excitonic recombination. The spectral position and intensity of this luminescence band are temperature dependent and they are, in fact, widely used for both ILNth and SLNth. Nevertheless, the presence of this single band does not allow the use of simple QDs for BSLNth since, as explained previously, BSLNth requires the simultaneous presence of two or more luminescence lines/bands to extract a temperature reading from their relative intensities. Very recently, Vlaskin et al. demonstrated the potential use of QDs in BSLNth, in particular, CdSe QDs that were doped with Mn2+ ions.53 The presence of Mn2+ ions leads to the appearance of energy states within the QD band-gap as can be observed in Fig. 12(a). When the QD system is optically excited to its high-energy excitonic states, efficient energy transfer (kET) quenches excitonic emission and sensitizes the Mn2+ luminescence. As a consequence, the emission spectrum becomes more complicated and now consists of two luminescence bands: one corresponding to the exciton re-combination (characteristic broad emission band of QDs) and another broad band (at longer wavelengths, i.e. at lower energies) This journal is ª The Royal Society of Chemistry 2012

Fig. 12 (a) Schematic representation of the energy levels and bands of a QD + Mn2+ system. (b) Luminescence emission spectra at different temperatures as obtained from colloidal Zn0.99Mn0.01Se/ZnCdSe nanocrystals under UV excitation. (c) Temperature dependence of the intensity ratio between the excitonic related emission and overall total emission of Zn0.99Mn0.01Se/ZnCdSe nanocrystals under UV excitation. (d) Digital picture of the colloidal suspension of Zn0.99Mn0.01Se/ZnCdSe nanocrystals under UV excitation, as obtained at 210 and 400 K. Reprinted with permission from ref. 53. Copyright 2010 American Chemical Society.

corresponding to optical-de-excitations from the inter-gap energy levels of Mn2+.54 The relative intensities between these two bands simultaneously depend on the luminescence efficiency of the excitonic and inter-gap states as well as on the relative population of these two states. In turn, this depends on several parameters including temperature, energy separation, DE, between the Mn2+ excited state and the lowest energy excitonic level, as well as on the energy transfer rate (also temperature dependent).53–55 Consequently, the intensity ratio of these two emission bands is expected to be temperature dependent and in fact, this is exactly what has been observed. Fig. 12(b) shows the Nanoscale, 2012, 4, 4301–4326 | 4311

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

View Article Online

normalized emission spectra of ZnMnSe/ZnCdSe nanocrystals as obtained for different temperatures. It is clear that the temperature induces a remarkable change in the intensities of the Mn2+ and excitonic bands (centered at around 570 and 600 nm, respectively). Fig. 12(c) shows the relative contribution of the excitonic emission to the total (integrated) luminescence intensity and as can be observed the relative contribution of the excitonic emission increases with temperature with an almost linear dependence near room temperature. Thus, the temperature reading can be achieved through the simple measurement of the relative contribution of the excitonic emission with respect to the overall emission or by the ratio between the excitonic and Mn2+ emissions. Indeed, the temperature induced change in this ratio is strong enough to produce a clear change in the color of the colloidal solution as can be observed from the pictures included in Fig. 12(d). The temperature range in which thermal sensing is achieved can be tailored by a rational design of the energy level scheme during QD synthesis. In particular, Vlaskin et al. demonstrated that the temperature at which the change in the intensity ratio occurs depends on the energy separation between the first excitonic state and the inter-gap Mn2+ state (DE in Fig. 12(a)).53 This energy separation depends on the QD size in such a way that large QDs lead to a reduced energy separation between these states so that the temperature sensing can occur at lower temperatures. In particular, the thermal sensing rate reduces from 300 K down to 150 K when the QD size is reduced from 4.7 to 3.5 nm.53 In the pioneering work of Vlaskin et al. it was also demonstrated that the QD + Mn2+ luminescence system can provide thermal reading with a resolution as good as 0.2
54 C. Nevertheless its application in real thermal imaging has not yet been demonstrated although great advances towards the incorporation of this system into biological structures have been made through the synthesis of QD + Mn2+ in water, which has been recently reported.55 B.2.2 Rare earth-based band-shape luminescence nanothermometry. There are numerous examples in the literature of rare earth (RE) based BSLNth (RE-BSLNth). The fundamental mechanism of RE-BSLNth depends on whether the luminescence lines under analysis are generated by a single type of rare earth or by a combination of different rare earth ions. Based on this criterion, it is possible to classify the RE-BSLNth into two groups: single-center RE-BSLNth and multi-center RE-BSLNth. The basic working principles, materials and applications of each class are briefly described below. B.2.2.1 Single-center RE-BSLNth. Rare earth ions (also known as lanthanide ions), with a few exceptions, show a rich energy level scheme.30 Due to the partial screening of the f electrons, the energy level scheme of a particular rare earth ion remains practically unchanged from host to host. The emission probability for each energy level depends on a great variety of factors but one of the most critical parameters is the energy separation between the RE energy levels. If this energy separation is small (i.e. comparable to the thermal energy kBT where kB is the Boltzmann constant and T is the temperature), then it is not possible to populate a single level since, due to the Boltzmann statistics, the population will be re-distributed among energy levels with similar energy. Let’s suppose that a RE energy level 4312 | Nanoscale, 2012, 4, 4301–4326

(labeled as 1) is populated as a consequence of optical excitation from the ground state. Due to the proximity of a second level of higher energy (labeled as 2), the population initially excited into level 1 is thermally re-distributed among levels 1 and 2. In a steady state condition, the population of the high-energy state, N2, will be given by: N2 ¼ N1exp(DE/kBT)

(1)

where N1 is the population of level 1 and DE is the energy separation between states 1 and 2. Since both states are populated, both contribute to the overall luminescence spectrum with the presence of two luminescence lines at different energies. The intensity of the luminescence line corresponding to the de-excitations from state 1 down to the ground state, I1, is given by: I1 ¼ 41N1

(2)

where 41 is a constant whose value depends on a great variety of geometrical factors as well as on intrinsic properties of the emitting level (such as branching ratios and luminescence quantum efficiency). Similarly, the intensity of the luminescence line corresponding to the de-excitations from state 2 is given by: I2 ¼ 42N2

(3)

Thus, the ratio between both intensities is given by: I2/I1 ¼ (42/41)(N2/N1) ¼ (42/41)exp(DE/kBT)

(4)

such that from the experimental determination of the intensity ratio, a temperature reading can be reached. Of course, appreciable thermal sensitivity would be achieved only if the energy separation between emitting levels is small, so that small temperature changes can induce large population re-distributions. As a consequence, not all the rare earth ions can be used for BSFNt but only those possessing radiative states with an energy separation of the order of few hundreds of wavenumbers, i.e. with an energy separation comparable to the room temperature thermal energy (the so-called ‘‘thermally coupled states’’). Fig. 13 includes some of the energy level diagrams of the rare earth ions that have been previously used for BSLNth: neodymium (Nd3+), thulium (Tm3+), europium (Eu3+), erbium (Er3+) and dysprosium (Dy3+).56–63 In each case the luminescence lines used for BSLNth are indicated. The bottom graph of Fig. 13 shows the temperature variation of the normalized luminescence intensity ratios determined experimentally for a series of Er3+, Dy3+, Eu3+ and Nd3+ doped systems.57–59 As can be observed, a remarkable change in the luminescence ratio is observed in all the cases. The temperature-induced variation of the luminescence ratio has been found to be very similar in the cases of Er3+, Eu3+ and Dy3+. The luminescence ratio has been found to be much more temperature dependent for Eu3+-doped La2O2S system. In this particular case, the luminescence intensity ratio between the emission lines generated from the 5D0 and 5D1 levels varies by almost one order of magnitude when the temperature is only increased few tens of degrees.56,59 This surprising result has been tentatively explained in the past by involving a third state (charge transfer state corresponding to the La2O2S host) in the thermal coupling between the two emitting states. Therefore, in this case This journal is ª The Royal Society of Chemistry 2012

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

View Article Online

studies, a thermally sensitive luminescent nanoparticle was mechanically attached to an AFM tip (see Fig. 14(a)). Based on the same principles of conventional AFM, the luminescent nanoparticle was scanned in pseudo-contact mode over the surface to be thermally imaged. The visible emission generated from the nanoparticle after 975 nm multiphoton excitation was simultaneously collected and analyzed by a fixed microscope objective, as is schematically shown in Fig. 14(a). The spatial variation of the intensity ratio between the 520 and 540 nm luminescence bands provides the thermal image of the surface after proper translation into temperature units. The reduced size of the luminescence probe, in combination with the accurate control of both position and height of the AFM tip, allowed the achievement of high spatial resolutions. This is illustrated in Fig. 14(b) in which the thermal image of an integrated circuit, where hot spots have been artificially introduced, is shown.73 This approach has been also applied for the thermal imaging of fluids, however, it still has the same limitations as any other thermal scanning microscopy technique.75 It is restricted to bidimensional thermal imaging of surfaces and since it is a pseudocontact method, can alter the original temperature distribution. Since these first works demonstrating the ability of Er3+ ions as ratiometric temperature sensors, they have been widely used in a great variety of applications ranging from the determination of local temperatures in the surroundings of optically excited metallic nanoparticles, to the thermal imaging of microfluidics in the presence of external heating sources.60 Very recently, Er3+ ions have emerged as versatile biocompatible luminescent nanothermometers. This possibility has been demonstrated by using Er3+ and Yb3+ co-doped NaYF4 (NaYF4:Er3+, Yb3+)

Fig. 13 Energy level diagrams of the trivalent rare earth ions used for BSLNth: neodymium, dysprosium, erbium and europium. The transitions used in BSLNth are also indicated. Bottom graph: temperature variation of the intensity ratio between the thermally coupled luminescence transitions in different systems. Data were extracted with permissions from ref. 57–59. Dots are experimental data and dashed lines are guides for the eyes.

the hyperthermal sensitivity is caused not by a Boltzmann redistribution but by a strong thermal dependence of the energy transfer rate between the charge transfer states and the Eu3+ 4f emitting levels.56,59 Among the different rare earth ions capable of single-center BSLNth, erbium is probably the most used one. Er3+ ions typically show a very intense green emission that consists of two luminescence bands centered at 520 and 540 nm, whose relative intensity is strongly temperature dependent.57,61,62,64–67 Er3+ ions have been widely used as BSLNth since the two bands, which provide the thermal reading, lie within the green spectral range where highly efficient detection is possible. Moreover, these two bands can be easily populated through multiphoton absorption of NIR radiation with the assistance of sensitizer ions such as ytterbium (allowing for spatial resolution enhancement).68 Due to these two facts, it is possible to find numerous examples in the literature of thermal sensing and imaging using Er3+ ions.60–70 One of the most attractive approaches is the use of Er3+-doped nanoparticles in scanning thermometry.36,60,71–74 In pioneer This journal is ª The Royal Society of Chemistry 2012

Fig. 14 (a) Experimental set-up used for scanning thermal microscopy based on an erbium doped luminescent particle. (b) Electron microscopy image of an integrated conducting wire in which a series of hot spots were artificially created in the form of wire width reductions (left). The corresponding thermal image, as obtained by BSLNth, is shown at the right. Reproduced from ref. 73 with permission.

Nanoscale, 2012, 4, 4301–4326 | 4313

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

View Article Online

biocompatible nanoparticles.65 NaYF4:Er3+, Yb3+ nanoparticles are widely used because of their outstanding luminescence brightness, their low toxicity and their good chemical and physical stability.76–79 The main goal of the Yb3+ ions is to provide efficient excitation of the Er3+ ions under NIR optical excitation (980 nm) via a two-photon absorption process (such that autoluminescence is avoided while achieving high spatial resolutions). The two-photon excited emission of NaYF4:Er3+, Yb3+ nanoparticles (resembling the typical Er3+ emission) obtained at two different temperatures is shown in Fig. 15(a). The presence of two thermally coupled levels is evidenced by the temperature induced change in the ratio between the luminescence bands centered at 525 and 545 nm.80 The temperature variation of the luminescence intensity ratio follows an almost linear relationship up to 40
C allowing for the straightforward determination of temperature.65 Once the NaYF4:Er3+, Yb3+ nanothermometers were calibrated they were introduced to HeLa cancer cells by a simple incubation procedure. Once incorporated within the HeLa cancer cell, they were used to monitor cell apoptosis caused by an external heating source (in

Fig. 15 (a) Luminescence emission spectra generated from Er3+/Yb3+ codoped NaYF4 nanoparticles under 920 nm excitation as obtained at two different temperatures. (b) Intracellular temperature (as obtained from the spectral analysis of the intracellular Er3+ luminescence) as a function of the applied voltage to a resistance attached to the micro-chamber containing the cell. The optical transmission images of the cell at different applied voltages are also shown. Reprinted with permission from ref. 65. Copyright 2010 American Chemical Society.

4314 | Nanoscale, 2012, 4, 4301–4326

this case an ohmic resistance attached to the micro-chamber containing the HeLa cancer cells). A 920 nm laser beam was focused inside the cell so that the Er3+ luminescence was induced through a two-photon absorption process involving the excited states of Yb3+ ions and was collected by the same microscope objective that was used for focusing the excitation laser beam inside the cell. A subsequent spectral analysis of the intracellular luminescence non-invasively and accurately determined the intracellular temperature. This simple approach has been used to determine the intracellular temperature of a single cell as a function of the voltage applied to the ohmic resistance in physical contact with the micro-chamber containing the cells. Results are summarized in Fig. 15(b). The intracellular temperature was found to follow a quadratic behavior with respect to the applied voltage, in accordance with a Joule heating process. Fig. 15(b) also shows the transmission optical images of the investigated cells at different temperatures where at room temperature (25
C), the cancer cells show their typical irregular shape (green dotted outline). A subsequent temperature increase to 35
C does not produce any relevant changes to its morphology. However, increasing the temperature to 45
C, a small membrane fragment was observed, which is indicative of cell death. Results included in Fig. 15 demonstrated the potential use of Er3+ doped nanoparticles for the investigation of intracellular thermal dynamics. B.2.2.2 Multi-center RE-BSLNth. Multi-center RE-BSLNth is based on the incorporation of a luminescent compound containing two rare earth ions (or two different species of the same rare earth ion), within the system under thermal investigation. The luminescence intensities of each luminescent center follow very different thermal behaviors, in such a way that the luminescence intensity ratio between their emissions would be strongly temperature dependent. Among the most relevant examples of RE-BSLNth is the work of Brites et al., who reported on the synthesis of luminescent molecular thermometers consisting of terbium (Tb3+) and Eu3+ co-doped g-Fe2O3 nanoparticles.70,81 The luminescent system was specially designed in such a way that it shows an excited triplet state with energy slightly above that of the Tb3+ 5D4 emitting state. The small energy difference between the Tb3+ and host states caused the occurrence of a Tb3+ to host energy transfer process, which is strongly temperature dependent as it is a phonon-assisted process.70,81 Fig. 16(a) shows the temperature dependence of the luminescence generated by the Eu3+, Tb3+ co-doped g-Fe2O3 nanoparticles after UV optical excitation. It should be noted that in this system coordination compounds of Tb3+ and Eu3+ ions are embedded and, thus, the triplet state refers to the organic part of the complex. The two luminescence lines used for temperature sensing are indicated and properly labeled in terms of the emitting ion. As can be observed, the luminescence line at around 547 nm (generated by Tb3+ ions) suffers from a relevant quenching near room temperature due to the thermal activation of energy transfer processes from the 5D4 (Tb3+) state to the host triplet state. On the other hand, the luminescence line at around 610 nm (corresponding to Eu3+ ions) remains almost constant over the entire temperature range due to the large energy difference between the host band and the 5D0 excited state of the Eu3+ ions. This journal is ª The Royal Society of Chemistry 2012

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

View Article Online

Fig. 16 (a) Temperature dependence of the luminescence generated by the Eu3+, Tb3+ co-doped g-Fe2O3 nanoparticles after UV optical excitation. (b) Temperature dependence of the ratio between the Eu3+ and Tb3+ luminescence intensities. Reproduced from ref. 81 with permission.

Fig. 16(b) shows the temperature dependence of the ratio between the Eu3+ and Tb3+ luminescence intensities. It is clear that this luminescence ratio guarantees a precise and absolute temperature measurement with an accuracy, estimated by Brites et al., to be close to 0.5
C.81 The relevance of the results summarized in Fig. 16 is even more notable given the flexibility in the nanoparticle synthesis allowing, for example, for thin film production. This, in turn, would allow for the opportunity of real time temperature sensing during magnetic hyperthermia possible by taking advantage of the superparamagnetism character of the temperature sensitive nanoparticles. Despite these good perspectives, the potential application of Eu3+, Tb3+ co-doped gFe2O3 based luminescent nanothermometers for biological applications has not yet been demonstrated. At this point it should be noted that temperature sensing based on Eu3+ and Tb3+ co-doped systems has been recently proposed by Cui et al. who also reported on the remarkable temperature induced changes in the Eu3+ to Tb3+ luminescence intensities in double doped metal– organic luminescent compounds.82 A different approach was proposed by Ishiwada et al. where the phosphor thermometer consisted of a Tb3+/Tm3+ co-doped Y2O3 system.83 BSLNth was achieved by recording the ratio between the Tm3+ emission in the blue (at around 466 nm) and the dominant green emission of Tb3+ ions (at around 540 nm). The intensity ratio between these two emissions is strongly This journal is ª The Royal Society of Chemistry 2012

temperature dependent, varying almost a 100% in the 300– 1100
C range as can be observed in Fig. 17. In this case, the temperature induced ratio variation is explained on the basis of the different thermal quenching of the individual ions: the blue emission of Tm3+ does not show any appreciable thermal quenching in the 300–1100
C, whereas the 540 nm Tb3+ emission is fully quenched at 1100
C. Despite the large ratio contrast shown by the Tb3+/Tm3+ co-doped Y2O3 system, it should be noted that it is achieved over a large temperature range (over almost 800
C) so that the net thermal sensitivity is low. In addition the luminescence ratio is linear only in the 500–900
C range and is nearly flat close to room temperature. Therefore, the system recently proposed by Ishiwada et al. would find practical application in thermal imaging of high temperature systems (such as combustion cells) rather than in biosystems and microelectronic devices. As a final example of multi-center RE-BSLNth we will briefly describe the original idea of Seaver and Peele84 They realized that when the Eu3+–EDTA complex is dissolved in water, two different species appear: Eu3+–(EDTA)(H2O)2 and Eu3+– (EDTA)(H2O)3. The presence of these two species leads to the existence of two slightly different crystal fields (local environments) for Eu3+ ions, i.e. two different Eu3+ centers. This causes the appearance of two 7F0 / 5D0 emission lines close in energy (separated by less than 1 nm). The ratio between these luminescence lines was found to be strongly temperature dependent as a consequence of the different thermal quenching rates of both Eu3+ centers as well as because of the temperature induced chemical equilibrium between them. The luminescence ratio varied at a rate of 8% per
C, allowing for thermal sensing with a thermal resolution of 1
C. Seaver and Peele applied this novel thermal sensing approach to determine the temperature of acoustically levitated water drops as small as 200 mm, being capable of determining single drop temperature rises of 7
C.84 B.2.3 Dye-based band-shape luminescence nanothermometry. As described in previous sections, luminescent dyes (such as Rhodamine B) have been widely used for ILNth. The sole use of a luminescent dye for temperature sensing by ILNth can result in problems due to local fluctuations in both excitation light intensity and dye concentration. This is especially true in dynamical systems such as micro-fluidics and living cells. Both fluctuations make the calibration of ILNth very difficult if not

Fig. 17 Images of the temperature-sensitive visible luminescence of Y2O3:Tb3+, Tm3+ phosphors. Rows (a), (b) and (c) correspond to different Tb and Tm concentrations as described in ref. 83. Data reproduced from ref. 83 with permission from Optical Society of America.

Nanoscale, 2012, 4, 4301–4326 | 4315

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

View Article Online

even impossible. These problems could be avoided by using a second luminescent dye with temperature independent luminescence intensity, such as Rhodamine 110, that would act as a luminescence reference. This was the approach used by Ebert et al. who were able to determine the laser-induced thermal loading of an optofluidic device that consists (see Fig. 18(a)) of a micro-channel (capillary) side-pumped by two optical fibers both coupled to a 1064 nm laser.8 An aqueous solution containing both Rhodamine B and Rhodamine 110 was injected into the capillary and the laser induced temperature increment, created by the non-vanishing absorption of water at 1064 nm, was determined from the confocal luminescence image obtained from the ratio between the luminescence intensities generated by the two dye species (found to decrease at a rate of 1.3% per
C). Fig. 18(b) shows a typical thermal image of the optofluidic device as obtained when a 1064 nm laser with a total power of 2 W was incident onto the microfluidic device. The thermal images obtained by the dual dye system were compared to theoretical predictions and were in excellent agreement, revealing the feasibility and applicability of this approach.8,85 B.3 Spectral luminescence nanothermometry As previously discussed, spectral luminescence nanothermometry (SLNth) provides thermal reading from the

temperature induced spectral shift of luminescence lines. At variance with ILNth and BLNth, SLNth is not based on the analysis of absolute or relative intensities but on the determination of the spectral position of the luminescence lines. Consequently, temperature reading is not affected by luminescence intensity fluctuations caused by variations in the local concentration of emitting centers. Obviously, high-resolution SLNth implies the use of luminescence systems with large temperature induced spectral shifts in such a way that small temperature variations would cause remarkable shifts in the luminescence lines. Although the presence of temperature induced spectral shifts has been observed in many luminescent systems, only in the case of semiconductor QDs has it been successfully used for highresolution SLNth. Fig. 2(a) demonstrates that temperature increments cause an overall intensity reduction of the QD luminescence as well as a remarkable spectral shift. This is a general phenomenon in QD luminescence as it has been observed and reported in many systems.86–90 The physical origin of the temperature induced spectral shift of QDs is far from being simple. It has been stated that it occurs as a result of the combination of different phenomena. The thermal spectral coefficient of QDs (dl/dT, where l denotes the spectral position of the luminescence line) can be written as:86–88 dl/dTf (dEg0/dT) + (dEconf/dT) + (dJe–ph/dT)

Fig. 18 Top. Temperature dependence of the ratio between luminescence intensities generated by Rhodamine B and Rhodamine 110 in an aqueous solution containing both dyes. The optical transmission image in the middle corresponds to an opto-fluidic device consisting of a microchannel side pumped by two optical fibers both coupled to a 1064 laser. The graph at the bottom is the thermal image of the micro-fluidic when the total 1064 nm power was set to 2 W. Data reproduced from ref. 83 with permission from Optical Society of America.

4316 | Nanoscale, 2012, 4, 4301–4326

(5)

where the first term (dEg0/dT) corresponds to the temperature coefficient of the bandgap energy of the bulk material (Eg0), a term that only depends on the QD constituent elements. The second term (dEconf/dT) accounts for the temperature coefficient of the quantum confined energy (Econf). According to previous works, the magnitude of this second term depends on both the thermal expansion coefficient of the QDs and on the radius of the QD (dEconf/dT increases as the QD size decreases).88 Finally, the third term (dJe–ph/dT) accounts for the temperature-induced change in the electron–phonon coupling energy (Je–ph), caused by the temperature-dependence of both the phonon energy and density of states.87,88 Due to the presence of these different terms, the a priori estimation of dl/dT of a given QD system is difficult to undertake, since it depends on the intrinsic material properties as well as on the geometrical properties (size) of the QDs. As a matter of fact, it is possible to find some systems (such as PbSe QDs) in which the sign of the dl/dT coefficient changes with the QD size.88 This fact is illustrated in Fig. 19(a) and (b) in which we have shown the temperature variation of the spectral shifts of the 3.9 and 6.9 nm diameter PbSe QDs, respectively. Note that for small PbSe QDs the dl/dT coefficient is positive whereas for larger PbSe QDs it is negative. More recent studies have shown that QD size can be used to optimize the SLNth sensitivity. In the particular case of CdTe QDs, Maestro et al. demonstrated that the spectral thermal coefficient grows monotonously as the QD size is reduced, as can be observed in Fig. 19(c).90 Note that for the particular case of CdTe, ultra-small QDs lead to dl/dT coefficients close to 0.8 nm per
C. For such large coefficients, thermal sensitivities as good as 0.2
C are, in principle, expected. The first demonstration (to the best of our knowledge) of QD based SLNth was provided by Li et al. who used the luminescence of CdSe QDs to determine the temperature of an electrical This journal is ª The Royal Society of Chemistry 2012

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

View Article Online

Fig. 19 Temperature dependence of the luminescence wavelength of (a) 3.9 nm and (b) 6.9 nm PbSe QDs, respectively. (c) Temperature coefficient of luminescence emission wavelength as obtained from CdTe QDs of different sizes. In all the cases solid dots are experimental data and solid lines are guides for the eyes. Reprinted with permission from ref. 88 and 89. Copyright 2010 American Chemical Society.

micro-heater on an aluminum microwire and their experimental approach to achieve non-contact thermal sensing is illustrated in Fig. 20(a).91 Individual CdSe QDs were placed on top of the aluminum micro-heater and were optically excited by a tightly focused green (532 nm) laser beam. The CdSe QD luminescence was collected by the same objective used for focusing the 532 nm radiation and then spectrally analyzed by a CCD camera attached to a high-resolution monochromator. Any local change in the micro-heater temperature was then detected by the appearance of a red shift in the CdSe QD luminescence shifting at a rate of 0.1 nm per
C.91 By scanning the focusing/collecting microscope objective authors were able to measure the temperature profiles along the micro-heater with sub-micrometric resolution. Representative results are shown in Fig. 20(b), where the temperature profiles obtained for different applied voltages are shown. It is clear that the experimentally obtained profiles account well for the predictions made based on Joule’s law, i.e. the temperature increment rises with the applied voltage. From the temperature profiles the temperature uncertainty can be estimated to be close to 1
C. The results described above demonstrated the ability of QDs to be used as highly sensitive optical probes for SLNth. This fact, in combination with the possibility of incorporating QDs in biological systems (such as cells and tissues) makes them unique This journal is ª The Royal Society of Chemistry 2012

Fig. 20 (a) Experimental set-up used by Li et al. in ref. 91 to measure the thermal images of a micro-heater using CdSe QDs. (b) Temperature profiles obtained along the micro-heater for different applied voltages. Reprinted with permission from ref. 91. Copyright 2007 American Chemical Society.

probes for high-resolution SLNth of biological systems.92,93 It should be noted that for the purpose of thermal imaging of biological systems, QDs offer an additional advantage: they can be optically excited within the biological window through a multiphoton excitation process.94 This last possibility simultaneously allows for the achievement of an improved spatial resolution as well as large penetration depths into tissues.95 During the last few years several groups have experimentally demonstrated QD based SLNth of biological systems. The first work reporting on the use of QDs for SLNth in a biological system was published by Maestro et al.89 In this case, CdSe QDs were incorporated into HeLa cancer cells by a simple incubation procedure. These cancer cells were subsequently subjected to an external heating process using an air-assisted micro-heater as is schematically depicted in Fig. 21(a). At the same time, 800 nm femtosecond laser pulses were focused into the cell in order to excite the CdSe QD luminescence via a multiphoton excited process. The intracellular QD luminescence was spectrally analyzed allowing the authors to obtain, in real time, the time evolution of the intracellular temperature during the heating procedure. This is presented in Fig. 21(b), which shows the intracellular QD emission spectra as obtained at two different heating times. As can be observed, the QD intracellular emission is partially quenched but concomitantly undergoes a relevant red shift. From the analysis of this spectral red shift and based on the Nanoscale, 2012, 4, 4301–4326 | 4317

View Article Online

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

analysis was used to monitor the local temperature response inside single living cells upon external chemical stimuli. In particular, they managed to map the intracellular heat generation in living cells following Ca2+ stress and cold shock. Their results demonstrated the presence of an inhomogeneous intracellular temperature and are currently boosting the use of QDs as high-resolution bio-nanothermometers.

B.4

Polarization luminescence nanothermometry

Polarization luminescence nanothermometry (PLNth) is based on the temperature dependence of the polarization state of the luminescence generated by luminescent molecules under Brownian dynamics. In general, when a luminescent molecule is illuminated by a linearly polarized laser beam it re-emits partially polarized luminescence due to the random orientation of the molecules. The polarization anisotropy factor of the luminescence, r, is defined as:97,98 r ¼ (Ik  It)/(Ik + 2It)

(6)

where Ik and It are the intensities of the luminescence polarized parallel and perpendicular to the incident polarization. In the absence of any molecular motion, the polarization anisotropy reaches its maximum value of 0.4. On the other hand, in the presence of molecular rotation (induced by, for example, its Brownian dynamics) the polarization anisotropy decreases due to an average-effect caused by the molecular rotations taking place during the luminescence emission (during the lifetime temporal window). When a luminescent molecule with a hydrodynamic molecular volume of V is incorporated in a fluid with a h viscosity, its polarization anisotropy can be written as:97 r ¼ 0.4(1 + (Vh/kBTsf))1

Fig. 21 (a) Schematic drawing of the experimental set-up used by Maestro et al. in ref. 89 for intracellular luminescence thermal sensing based on CdSe QDs. (b) Intracellular luminescence spectra generated by CdSe QDs as obtained at two different heating times (0 and 3 min). (c) Intracellular temperature determined by SLNth using CdSe QDs. Reprinted with permission from ref. 89. Copyright 2010 American Chemical Society.

temperature spectral coefficient of CdSe QDs (close to 0.15 nm per
C in this case), the authors were able to determine the cell temperature during the different stages of the heating procedure (see Fig. 21(c)). A similar approach as that used by Maestro et al. was adopted by Yang et al.96 CdSe QDs were also incorporated into HeLa cancer cells by an incubation procedure, although in this case the CdSe QD intracellular temperature obtained from the spectral 4318 | Nanoscale, 2012, 4, 4301–4326

(7)

where T is the molecule temperature and sf its luminescence lifetime. Thus, the polarization of the luminescence is unequivocally related to the molecule temperature, so that a thermal reading can be obtained from the analysis of the polarization dependence of the luminescence. One of the molecules used in the past for PLNth is fluorescein, whose calibration curve obtained in aqueous solution (temperature dependence of its polarization anisotropy) is shown in Fig. 22(a). It should be noted that in the biological range (20– 50
C) the polarization anisotropy is modified by more than a 100% making it a highly sensitive temperature sensor. Moreover, the temperature is determined from the analysis of an optical parameter that is independent of the local concentration of luminescence molecules so that neither normalization nor reference procedures are necessary. Fluorescein based PLNth has been recently applied to determine the heating potential of diverse metallic nanostructures acting as nanometer-size heat sources such as gold nanorods (GNRs).99 The procedure used for PLNth is schematically illustrated in Fig. 22(b). The GNRs were dispersed onto a fluorescein-containing film and placed in a confocal luminescence microscope. The fluorescein molecules were excited by a 473 nm beam and by an NIR laser beam tuned to the surface plasmon resonance of GNRs. The parallel and orthogonal components of the fluorescein luminescence were This journal is ª The Royal Society of Chemistry 2012

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

View Article Online

Fig. 22 (a) Luminescence polarization anisotropy as a function of temperature for fluorescein dissolved in water. Dots are experimental data and the solid line is a guide for the eyes. (b) Schematic of the experimental set-up and data analysis procedure adopted by Baffou et al. to obtain a thermal image of optically excited GNRs by PLNth. Data reproduced from ref. 99 with permission from Optical Society of America.

separated by a polarizing cube and sent to two independent detectors. The blue and infrared beams were scanned across the sample plane for simultaneous local heating and temperature measurement. From the confocal luminescence images obtained in terms of the luminescence intensities at both polarizations, the polarization anisotropy image was obtained. Based on the calibration curve of Fig. 22(a), this polarization anisotropy image was translated into temperature units as it is shown in Fig. 22(b). Based on this operation scheme, Baffou et al. were able to achieve temperature resolutions as good as 0.1
C while keeping spatial resolutions below 0.5 mm.99

B.5 Bandwidth luminescence nanothermometry In general, when a luminescent center is heated, homogeneous line broadening causes spectral broadening of the emission lines as a consequence of the local environment fluctuations due to the thermal vibration of the luminescent center and also that of its neighboring atoms/molecules.30 The magnitude of the temperature induced luminescence line broadening is in general small and thus, can be only observed in systems showing inherent narrow emission lines and in which homogeneous line broadening This journal is ª The Royal Society of Chemistry 2012

dominates over the inhomogeneous one. This is the case of rare earth ions incorporated in some crystalline hosts. One of the most pertinent examples is the case of Nd3+ ions in the well-known YAG host (Nd:YAG).100 Nd:YAG shows two hypersensitive luminescence lines at around 940 nm that correspond to the low energy emissions within the 4F3/2 / 4I9/2 transition (see Fig. 23(a)) and the bandwidth of these two transitions is strongly temperature dependent as can be observed in Fig. 23(a). In the proximity of room temperature, the luminescence linewidth increases almost linearly with temperature in a wide temperature range (100
C) with a rate close to 0.04 cm1 per
C.101 This linear dependence allows for temperature reading from the determination of the Nd3+ luminescence linewidth with a constant sensitivity in the whole temperature range. Bandwidth luminescence nanothermometry (BLNth) based on the hypersensitive luminescence lines of Nd:YAG has been recently demonstrated by Benayas et al. who managed to obtain thermal images of a Nd:YAG ceramic microchip laser component in the presence of a tightly focused pump beam at 808 nm (where the Nd:YAG absorption is as large as tens of cm1).101 For this purpose, the authors proposed the experimental set-up schematically shown in Fig. 23(b). In this setup, the Nd:YAG microchip device that is pumped by a tightly focused 808 nm laser beam, was placed in a modified confocal microscope. The luminescence generated by the Nd3+ ions was collected by a scanning microscope and, after passing several filters and pinholes, was analyzed by a high-resolution spectrometer. This set-up allowed the authors to obtain the two dimensional spatial distribution of the Nd3+ bandwidth in the surroundings of the 808 nm tightly focused beam. This luminescence image can be easily translated into temperature units by using the calibration curve presented in Fig. 23(a). From the experimental set-up schematically shown in Fig. 23(b), the authors were able to get thermal images (with a sub-micrometric resolution) of a microchip laser element in the presence of high intensity pump radiations. A typical example of the thermal images obtained by BLNth is shown in Fig. 23(c). As expected, the maximum temperature increment is produced at the 808 nm focus position (center of the thermal image) being on the order of tens of degrees for a total pump power of 1 W. The large thermal conductivity of the YAG host produces a relevant spreading of the pump-induced thermal loading as was observed in the thermal image. Note that the thermal image is obtained from the confocal luminescence image that in turn can be expanded to three dimensions since the typical confocal axial resolutions (few microns) are well below the microchip element thickness (hundreds of microns). By acquiring the thermal images at different pump powers, the authors found that the laser induced thermal gradients at the focus position (determining the laser performance and stability of the microchip device) follow a non-linear behavior at variance to the linear one expected from theory. This information can be critical in understanding the real world laser performance of microchip elements. B.6

Lifetime luminescence nanothermometry

The different methods reviewed up to this point require the acquisition of the luminescence spectrum of the thermally sensitive optical nanoprobe. In general, the requirement of high spatial resolution leads to low signal levels because of the use of Nanoscale, 2012, 4, 4301–4326 | 4319

View Article Online

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

noise ratio in the luminescence spectrum so that an accurate temperature reading could be achieved. This, in turn, requires long illumination times leading to the possibility of laser-induced local heating of the system under investigation due to the activation of multiphonon decay processes. In order to avoid any externally induced heating of the system to be thermally imaged, acquisition times should be reduced. This can be accomplished by extracting the thermal information not from the luminescence spectrum but from the luminescence decay curve. In this case, the signal can be acquired at a time interval on the order of the luminescence lifetime (typically in the range of nanoseconds to milliseconds). Furthermore, for the acquisition of luminescence decay curves the use of high-resolution spectrometers is not necessary since spectral analysis is no longer required. Moreover, the time evolution of luminescence is usually recorded by using high gain fast detectors (such as avalanche photodiodes or photomultipliers) such that high signal-to-noise ratios are usually achieved. Of course, it goes without saying that for lifetime luminescence nanothermometry (LLNth), systems showing decay curves that are strongly dependent on temperature are obviously required. In general, temperature is one of the most critical parameters determining the luminescence decay time (lifetime). This is because temperature increments lead to the activation of phonon-assisted processes and of multiphonon decays. Both effects cause an increase in the net de-excitation probability of the emitting level and, hence, a decrease in its luminescence lifetime. Up to now the systems that have been proposed, based on the hyperthermal sensitivity of their luminescence lifetime, can be classified into four different groups depending on whether they were developed from rare earth ions, luminescent dyes, quantum dots, polymers or membranes. The main results achieved in each case are summarized below.

Fig. 23 (a) Temperature variation of the luminescence bandwidth of the Nd:YAG emission line at approximately 940 nm. The inset shows the luminescence spectrum of Nd:YAG at circa 940 nm showing the two hypersensitive lines of Nd:YAG. (b) Experimental set-up used for thermal imaging of tightly pumped Nd:YAG ceramic microchip devices. (c) Thermal image of an 808 nm tightly pumped Nd:YAG ceramic based microchip. The laser spot is at the center of the thermal image, where temperature is maximum. Reprinted with permission from ref. 101. Copyright 2011 Springer.

high numerical optics, high-resolution spectrometers and of luminescence particles with reduced dimensions (with a relatively low absorption efficiency).80 In these conditions, large acquisition times have to be employed to ensure an adequate signal-to4320 | Nanoscale, 2012, 4, 4301–4326

B.6.1 Rare earth based lifetime luminescence nanothermometry. Several works have been published proposing rare earth doped crystals and complexes as good candidates for LLNth. One of the most interesting cases is that of Ce3+ doped YAG nanophosphors. Allison et al. reported that when the size of Ce3+:YAG micro-crystals is reduced to the sub-micrometric range (down to the nanoscale) the temperature dependence of the Ce3+ luminescence lifetime changes dramatically.102 Allison et al. found that while in micrometric Ce3+:YAG crystals the luminescence lifetime was almost temperature independent near room temperature, the lifetime of Ce3+:YAG nanocrystals becomes strongly temperature dependent, following an almost linear decay near room temperature (see Fig. 24). The ability of any given luminescent system for LFTNth is usually evaluated from its ‘‘normalized lifetime thermal coefficient’’, as, which is defined as as ¼ dsnor(T)/dT where snor(T) is the luminescence decay time at temperature T normalized to the room temperature value (i.e. snor(T) ¼ sf(T)/sf(25
C)). For nanosized Ce3+:YAG crystals, as is close to 0.01 per
C, in the 10–70
C temperature range. Additionally, the luminescence lifetime of Ce3+:YAG nanocrystals is in the tens of nanoseconds range and can be considered as a short lifetime value when compared with other rare earth ions. This is a very favorable property for LLNth as it increases the measuring rate. In addition to Ce3+, Eu3+ and Tb3+ have been recently proposed as good luminescent probes for LLNth in the 0–70
C temperature range. Yu et al. demonstrated that Eu3+ and This journal is ª The Royal Society of Chemistry 2012

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

View Article Online

Fig. 24 Temperature induced luminescence lifetime variation in different rare earth based systems. Dots are experimental data and solid lines are guides for the eyes. Data were extracted with permissions from ref. 102–104.

Tb3+ complexes, with large luminescence quantum yields, showed a drastic shortening of the luminescence lifetimes when the temperature was increased from 0 to 70
C (see Fig. 24).103 The normalized lifetime thermal coefficients found for Eu3+ and Tb3+ complexes were determined to be close to 0.02 and 0.012 per
C, respectively. According to Jiangbo et al., these values of as ensure a temperature resolution of approximately 0.1
C (this being constant over the entire temperature range due to the linear relation between lifetime and temperature).103 At this point, the work of Cai et al. should be mentioned as they report on the thermal sensitivity of the luminescence lifetime of Er3+ ions incorporated in microspheres composed of a heavy-metal fluoride glass (ZBLALiP).104 The authors determined that the lifetime decreases monotonously (but not linearly) with temperature in the range of 170 to 50
C. When compared with the other rare earth based LLNth systems, Er3+ has the disadvantage of a non-linear relation between lifetime and temperature and of an almost temperature independent lifetime in the vicinity of room temperature. Nevertheless, the interest of the work proposed by Cai et al. resides in the fact that the luminescence thermal sensor is incorporated in a glass microsphere that can be potentially controlled by optical tweezers.104 For example, this possibility could lead to the achievement of three-dimensional thermal images by the three-dimensional scanning of the thermal sensor. Despite their vast potential, to the best of our knowledge, there are no reported thermal images obtained by rare earth based LLNth. B.6.2 Dye-based lifetime luminescence nanothermometry. The temperature-induced reduction of the luminescence lifetime of luminescent organic dyes is a well-known phenomenon and is especially noticeable in the case of the ubiquitous Rhodamine B dye. Fig. 25(a) shows the Rhodamine B luminescence lifetime as a function of temperature in the 10–70
C range. The clear temperature–lifetime relation, obeying the Arrhenius equation, has been tentatively explained in the past in terms of the intramolecular activated non-radiative decay from the first excited This journal is ª The Royal Society of Chemistry 2012

singlet of the Rhodamine B molecule.105–107 A luminescence lifetime reduction close to 75% is observed. This yields an averaged normalized lifetime thermal coefficient of approximately 0.0125 per
C, although the nonlinear relation between lifetime and temperature leads to larger lifetime based sensitivities at lower temperatures. In the biological temperature range (25– 45 mine B is estimated to be as y 0.03 per
C, allowing for thermal reading with resolutions better than 1
C. Rhodamine B has been extensively used for LLNth specifically of fluids in the presence of external heating sources. The additional possibility of multiphoton excitation of the characteristic red luminescence of Rhodamine B has opened the possibility of true 3D thermal imaging of fluids with sub-micrometric spatial resolution.108,109 This has been, indeed, already demonstrated by Benninger et al. who were able to obtain a 3D thermal image of a 130  40  100 mm3 micro-channel device specifically designed for polymerase chain reaction (PCR) and is shown in Fig. 25(b).110 In this case, one of the micro-channel walls (upper one) was held at a constant temperature of 73
C by using an integrated heating source. It is clear from this thermal image that the intra-channel temperature in a PCR device is far from being constant. Indeed, variation as large as 5
C has been measured, this being a non-negligible temperature variation. As a matter of fact, such intra-channel

Fig. 25 (a) Temperature dependence of the Rhodamine B luminescence lifetime. Dots are experimental data and the solid line is a guide for the eyes. (b) Thermal image obtained by Rhodamine B based LLNth of a 130  40  100 mm3 micro-channel device specially designed for polymerase chain reaction. Reprinted with permission from ref. 110. Copyright 2006 American Chemical Society.

Nanoscale, 2012, 4, 4301–4326 | 4321

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

View Article Online

temperature variations could cause important variations in the PCR reaction efficiency.111 Finally, the work recently published by Bennet et al. should be mentioned due to its original approach for thermal imaging and sensing of a microfluidic device.112 In this case, the authors did not fill the microchannel with a Rhodamine B solution, rather they encapsulated the temperature-sensitive Rhodamine B within a micro-droplet, which could be held and manipulated in the microfluidic flow by using optical tweezers. This novel approach retains the capability of LLNth to deliver quantitative mapping of microfluidic temperatures without the disadvantageous need to introduce a luminescent dye that pervades the entire microfluidic system. As a disadvantage, it should be mentioned that in this case the spatial resolution of the thermal imaging technique is limited by the size of the micro-droplet containing the Rhodamine B. For instance, the size of the micro-droplets used by Bennet et al. were close to 10 mm in diameter,112 so that the final spatial resolution of their technique was far from that achieved by a direct incorporation of Rhodamine B into the micro-channel, given by the spatial resolution of the microscopy system (typically close to 1 mm). Although Rhodamine B is the most popular dye for lifetime thermal sensing, it has some serious drawbacks such as its low solubility in water. This makes it unsatisfactory for measuring aqueous microfluidics temperatures because of problems such as undesired deposition on surfaces. In this respect it should be mentioned that Mendels et al. have used Kiton Red, a watersoluble derivative of Rhodamine B, that shows the same lifetime– temperature response as the parent fluorophore, for LLNth of aqueous microfluidic systems.113 The ability of this ‘‘alternative’’ lifetime based luminescent thermal probe was investigated by Mendels et al. by comparing the thermal Kiton Red LLNth based thermal images of a microfluidic T-mixer when two fluids of very different temperatures (25 and 60
C) were injected, as it is schematically drawn in Fig. 26. On the left hand side of Fig. 26, a series of LLNth thermal images are shown as obtained at different locations of the T-mixer device. These images reveal how the local temperature of the microfluidic becomes homogenous as the two fluids are mixed along the common branch. The accuracy of their experimental measurements was checked by comparing these images with theoretical simulations (right image in Fig. 26) showing an excellent agreement.113 B.6.3 Polymer-based lifetime luminescence nanothermometry. The use of luminescent polymers for LLNth is not as widely known as luminescent dyes despite the recently published results where they have emerged as highly sensitive lifetime thermal probes. As it has been previously described in Section B.1.6, there are luminescent polymers (such as the so-called poly(DBD-AEco-NIPAM)) that, in an aqueous solution, undergo a thermally induced phase transition that is accompanied by a drastic decrease in the microenvironmental polarity in the vicinity of the main polymer chain.52 Since the luminescence properties of the polymer are affected by the solvent polarity, the phase transition causes a drastic change in the luminescence properties of the polymer. This fact has been used for thermal sensing based on the luminescence intensity variations during phase transition. Graham et al. demonstrated that the phase transition does not only cause a drastic change in the luminescence efficiency of the 4322 | Nanoscale, 2012, 4, 4301–4326

Fig. 26 Thermal images (left column) of a microfluidic T-mixer at different positions. Two fluids at different temperatures (60 and 25
C) are injected into the microfluidic device. The thermal images were obtained from the spatial variation of the luminescence lifetime of the Kiton Red dye. The figure at the right is a schematic representation of the device and includes the theoretical temperature distribution. Reprinted with permission from ref. 113. Copyright 2008 Springer.

polymer but also in its luminescence lifetime.114 This fact is shown in Fig. 27(a) that includes the temperature dependence of the luminescence lifetime of the poly(DBD-AE-co-NIPAM) polymer in the vicinity of the phase transition temperature (close to 32.5
C). As can be observed the lifetime based thermal sensitivity achieved is large. In fact, the normalized lifetime thermal coefficient is as large as 0.06 per
C, i.e. one order of magnitude larger than those typically achieved with rare earth ions or luminescent dyes. These large values of as are only achieved during the phase transition so that the temperature operation range of the polymer based LLNth is very limited. Despite this narrow operation range, luminescent polymers have been already used in thermal imaging of PCR microfluidic devices. A typical example is given in Fig. 27(b), showing the thermal image of a microfluidic chamber positioned between a heating source (at the left-hand of the chamber) and a cold sink (at the righthand of the chamber). The LLNth image clearly shows the temperature gradient across the chamber. The temperature This journal is ª The Royal Society of Chemistry 2012

View Article Online

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

cytoplasm was found to be strongly dependent on the particular cell cycle. In addition, authors also found that the majority of cells showed a well-localized temperature singularity (indicated by the arrowheads in Fig. 28(b)) that was tentatively associated with a centrosome-specific thermogenesis.116,117 Results obtained by Okabe et al. demonstrated that polymeric LLNth could be used to identify the relationships between the temperature and organelle functions. B.6.4 Quantum dot-based lifetime luminescence nanothermometry. As described in Section B.3, the energy level diagram of QDs is temperature dependent due to a combination of different factors, including temperature-induced changes in QD size, band-gap, electron–phonon interaction and confinement energy. This temperature induced variation in the QD energy level scheme is known to cause a luminescence shift that has been already used for thermal sensing and imaging based on SLNth. As it has been also described, this temperature induced

Fig. 27 (a) Temperature dependence of the luminescence lifetime of the poly(DBD-AE-co-NIPAM) polymer in the vicinity of the phase transition temperature (close to 32.5
C). (b) Thermal image of a microfluidic chamber positioned between a heating source (left-hand of the chamber) and a cold sink (right-hand of the chamber). Data were obtained from the analysis of the spatial variation of the luminescence lifetime of the poly(DBD-AE-co-NIPAM) polymer that is introduced into the microchamber in an aqueous solution. Reproduced from ref. 114 with permission from The Royal Society of Chemistry.

resolution of the thermal image included in Fig. 27(b) has been estimated by the authors to be less than 0.1
C.114 The achievement of these outstanding thermal sensitivities within the physiologically relevant temperature range makes luminescent polymers, such as poly(DBD-AE-co-NIPAM), valuable lifetime based thermal probes for biomedical applications, in which temperature control is critical. Very recently, Okabe et al. have taken advantage of the outstanding properties of luminescent polymers for LLNth for the achievement of high-resolution thermal images of single cells with spatial and temperature resolutions as good as 200 nm and 0.18
C, respectively.115 Okabe et al. incubated COS7 cells in a solution containing a NNPAM based luminescent polymer with a phase transition at around 35
C. During phase transition the luminescence lifetime of the polymer was found to increase linearly with temperature with a normalized lifetime thermal coefficient as large as 0.065 per
C.115 Once the luminescent polymer was incorporated into the cells Okabe et al. obtained the corresponding thermal image from the analysis of the spatial variation of the luminescence lifetime of the polymer. Representative data are included in Fig. 28. Okabe et al. found experimental evidence of the existence of a highly inhomogeneous intracellular temperature distribution. Indeed, as can be observed in Fig. 28(b), the nucleus of the COS7 cells showed significantly higher temperatures than the cytoplasm. The observed temperature gap between the nucleus and the This journal is ª The Royal Society of Chemistry 2012

Fig. 28 (a) Confocal luminescence image of living COS7 cells incubated with a luminescent polymeric thermometer. (b) Thermal image of the same cells as those shown in (a) as obtained by LLNth. Arrowheads indicate the location of the well localized hot-spots observed in the majority of cells examined by Okabe et al. in ref. 115. Reprinted with permission from ref. 115. Copyright 2012 Nature Publishing Group.

Nanoscale, 2012, 4, 4301–4326 | 4323

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

View Article Online

spectral shift is accompanied by a relevant luminescence quenching. The observed temperature induced luminescence quenching is attributed to several causes. On the one hand, the luminescence efficiency of the excitonic level decreases as a consequence of the thermal activation of multiphonon assisted electron–hole recombination. On the other hand, temperature enhances the energy transfer probability from bulk (excitons) to non-radiative surface (trap) states (since thermally induced QD expansion reduces the energy separation between excitonic states) and favors energy migration among them until a nonradiative state is reached.118–121 Both effects lead to an increment in the non-radiative de-excitation probability and, thus, in the total de-excitation probability. Therefore, it is expected that temperature increments in the surroundings of ambient temperature would also cause a decrease in the luminescence lifetime. This is indeed what has been recently demonstrated by HaroGonz
alez et al. who examined in detail how the CdTe QD luminescence lifetime varies within the biophysical thermal range (25–50
C).122 The authors found that the luminescence lifetime decreases almost linearly for all the QD sizes, however, the smaller the CdTe QD, the larger the temperature induced lifetime reduction. Fig. 29(a) shows, as an example, the luminescence decay curves of 1.2 nm CdTe QDs obtained at both 30 and 50
C and reveal a drastic temperature induced luminescence decay

Fig. 29 (a) The luminescence decay curves of 1.2 nm CdTe QDs obtained at 30 and 50
C. (b) Luminescence decay time of 1.2 nm CdTe QDs as a function of temperature. Dots are experimental data and the solid line is a guide for the eyes.

4324 | Nanoscale, 2012, 4, 4301–4326

time reduction. This is even more noticeable in Fig. 29(b) that shows the luminescence decay time of 1.2 nm CdTe QDs as a function of temperature. As can be observed, luminescence lifetime decreases linearly with temperature. From Fig. 29(b), Haro-Gonz
alez et al. estimated, for CdTe QDs, a luminescence lifetime thermal coefficient as large as 0.017 per
C, this being one of the largest as values ever reported (superior than that reported for Rhodamine B).122 This fact makes CdTe QDs very promising luminescent probes for high (spatial and thermal) resolution luminescence lifetime thermal imaging. Haro-Gonz
alez et al. already demonstrated the ability of CdTe QDs for lifetime thermal sensing in micro-fluidics and the obvious next step is their application for thermal imaging of biosystems (such as single cells and tissues).

C. Conclusions In summary, we have presented a detailed review of the diverse methods proposed to date for the achievement of high-resolution thermal sensing from the analysis of luminescence. It has been shown that many luminescent systems (including polymers, organic dyes, rare earth doped crystals, rare earth doped nanocrystals, semiconductor nanocrystals, rare earth doped complexes, phosphorous and quantum dots) can be used as basic light-emitting materials for nanothermometry. It is not possible to highlight one material over the rest since the most suitable one would depend on the actual system to be thermally imaged. This review describes in detail how these luminescence systems have been previously used for high-resolution thermal imaging of a great variety of systems including living cells, microfluidics, electronic nano-devices and solid state lasers. The continuous development of new microscopy techniques coupled with novel and cutting edge synthesis processes (allowing for the rational design of novel luminescent materials) will ensure the speedy development of luminescence nanothermometry, although working principles will be probably the same as those summarized in this review. A plethora of examples have been described and explained in detail. The achievement of temperature resolutions well below the
C limit has fostered the use of luminescent nanothermometers in biological applications. This is by far the most challenging of all possible applications described in this review, due to the complex nature of the biological milieu. To overcome some of the hurdles associated with working in a biological environment, employing luminescent nanothermometers optically excited in the near-infrared (NIR) region is particularly promising. In combination with multiphoton microscopy, this would allow for high-resolution, three-dimensional thermal imaging of living specimens. Although a great deal of effort has been invested in the development of NIR excited luminescent nanothermometers, their real-world application in three-dimensional thermal imaging of bio-systems is far from being a reality. For the most part, this is because these nanothermometers typically emit in the visible range, where tissue transparency is low. In our opinion the forthcoming advances in the field of luminescence nanothermometry will be propelled by the development of luminescent nanothermometers, which operate within the biological window (700–900 nm). That is, both their excitation and emission wavelengths lie within this optimal window. This journal is ª The Royal Society of Chemistry 2012

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

View Article Online

Furthermore, we are also firmly convinced that the most groundbreaking results in the field of luminescent nanothermometry will be obtained by using ratiometric thermal sensors, such as those described throughout this review. The main reasons, of course, are that exploiting the luminescence ratios of particular emission bands to obtain the thermal reading is not affected by local intensity fluctuations (i.e. concentration of emitting centers) but moreover, would require simple experimental instrumentation. Finally, we also believe that luminescent nanothermometry will be widely used in the future not only for disease detection but also for monitoring and control of laser induced thermal treatments (i.e. hyperthermia). In this sense, the development of multi-functional nanoparticles capable of simultaneous laser induced heating and luminescent nanothermometry will yield significant advancements in disease diagnoses and therapeutics in the near future.

Notes and references 1 B. A. Shakouri, Proc. IEEE, 2006, 94, 1613–1638. 2 J. I. M. Colvin, Microelectron. Reliab., 1998, 38, 1705–1714. 3 A. Bar-Cohen and P. Wang, Microgravity Sci. Technol., 2009, 21, 351–359. 4 D. H. Jundt, Opt. Lett., 1997, 22, 1553–1555. 5 J. J. Romero, C. Arag
o, J. A. Gonzalo, D. Jaque and J. Garc
ıa Sol
e, J. Appl. Phys., 2003, 93, 3111. 6 G. Nenna, G. Flaminio, T. Fasolino, C. Minarini, R. Miscioscia, D. Palumbo and M. Pellegrino, Macromol. Symp., 2007, 247, 326– 332. 7 A. Ashkin, J. M. Dziedzic and T. Yamane, Nature, 1987, 330, 771. 8 S. Ebert, K. Travis, B. Lincoln and J. Guck, Opt. Express, 2007, 15, 15493–15499. 9 M. Suzuki, V. Tseeb, K. Oyama and S. Ishiwata, Biophys. J., 2007, 92, L46–L48. 10 O. Zohar, M. Ikeda, H. Shinagawa, H. Inoue, H. Nakamura, D. Elbaum, D. L. Alkon and T. Yoshioka, Biophys. J., 1998, 74, 82–89. 11 F. Tardieu, M. Reymond, P. Hamard, C. Granier and B. Muller, J. Exp. Bot., 2000, 51, 1505–1514. 12 G. N. Somero, Annu. Rev. Physiol., 1995, 57, 43–68. 13 M. Stark, Cancer, 1974, 33, 1664–1670. 14 J. S. Michaelson, S. Satija, D. Kopans, R. Moore, M. Silverstein, A. Comegno, K. Hughes, A. Taghian, S. Powell and B. Smith, Cancer, 2003, 98, 2114–2124. 15 T. B. Huff, L. Tong, Y. Zhao, M. N. Hansen, J.-X. Cheng and A. Wei, Nanomedicine, 2007, 2, 125–132. 16 J. Lee and N. A. Kotov, Nano Today, 2007, 2, 48–51. 17 P. R. N. Childs, J. R. Greenwood and C. A. Long, Rev. Sci. Instrum., 2000, 71, 2959–2978. 18 L. Shi, O. Kwon, A. C. Miner and A. Majumdar, Design, 2001, 10, 370–378. 19 A. Majumdar, Annu. Rev. Mater. Sci., 1999, 29, 505–585. 20 J. Park and M. Taya, Electron. Lett., 2004, 40, 4–5. 21 O. Nakabeppu, M. Chandrachood, Y. Wu, J. Lai and A. Majumdar, Appl. Phys. Lett., 1995, 66, 694–696. 22 K. Kim, J. Chung, G. Hwang, O. Kwon and J. S. Lee, ACS Nano, 2011, 5, 8700–8709. 23 G. Baffou, P. Bon, J. Savatier, J. Polleux, M. Zhu, M. Merlin, H. Rigneult and S. Monneret, ACS Nano, 2012, 6, 2452–2458. 24 J. Christofferson and A. Shakouri, Rev. Sci. Instrum., 2005, 76, 024903. 25 T. Beechem, S. Graham, S. P. Kearney, L. M. Phinney and J. R. Serrano, Rev. Sci. Instrum., 2007, 78, 061301. 26 M. Kuball, J. M. Hayes, M. J. Uren, T. Martin, J. C. H. Birbeck, R. S. Balmer and B. T. Hughes, IEEE Electron Device Lett., 2002, 23, 7–12. 27 S. Ch
enais, S. Forget, F. Druon, F. Balembois and P. Georges, Appl. Phys. B: Lasers Opt., 2004, 79, 221–224.

This journal is ª The Royal Society of Chemistry 2012

28 K. S. Schwartzkopf-Genswein and J. M. Stookey, Can. J. Anim. Sci., 1997, 77, 577–583. 29 S. Huth and O. Breitenstein, J. Appl. Phys., 2000, 88, 4000–4003. 30 G. F. Imbusch and B. Henderson, Optical Spectroscopy of Inorganic Solids, Oxford Science Publications, London, 2006. 31 X. Michalet, F. F. Pinaud, L. A. Bentolila, J. M. Tsay, S. Doose, J. J. Li, G. Sundaresan, A. M. Wu, S. S. Gambhir and S. Weiss, Science, 2012, 538, 538–544. 32 O. I. Mi
ci
c, H. M. Cheong, H. Fu, A. Zunger, J. R. Sprague, A. Mascarenhas and A. J. Nozik, J. Phys. Chem. B, 1997, 101, 4904–4912. 33 G. W. Walker, V. C. Sundar, C. M. Rudzinski, A. W. Wun, M. G. Bawendi and D. G. Nocera, Appl. Phys. Lett., 2003, 83, 3555–3557. 34 B. Han, W. L. Hanson, K. Bensalah, A. Tuncel, J. M. Stern and J. A. Cadeddu, Ann. Biomed. Eng., 2009, 37, 1230–1239. 35 S. Wang, S. Westcott and W. Chen, J. Phys. Chem. B, 2002, 106, 11203–11209. 36 L. Aigouy, B. Samson, G. Julie, V. Mathet, N. Lequeux, C. Nı. Allen, H. Diaf and B. Dubertret, Rev. Sci. Instrum., 2006, 77, 063702. 37 J. Lee, A. O. Govorov and N. A. Kotov, Angew. Chem., Int. Ed., 2005, 44, 7439–7442. 38 J. R. Lakowicz, I. Gryczynski, V. Bogdanov and J. Kusba, J. Phys. Chem., 1994, 98, 334–342. 39 D. Ross, M. Gaitan and L. E. Locascio, Anal. Chem., 2001, 73, 4117–4123. 40 P. L€ ow, B. Kim, N. Takama and C. Bergaud, Small, 2008, 4, 908– 914. 41 R. Samy, T. Glawdel and C. L. Ren, Anal. Chem., 2008, 80, 369–375. 42 S. W. Allison, Rev. Sci. Instrum., 1997, 68, 2615–2650. 43 S. V. Yap, R. M. Ranson, W. M. Cranton, D. C. Koutsogeorgis and G. B. Hix, J. Lumin., 2009, 129, 416–422. 44 H. Fonger and C. W. Struck, J. Chem. Phys., 1970, 52, 6364–6366. 45 C. Gota, K. Okabe, T. Funatsu, Y. Harada and S. Uchiyama, J. Am. Chem. Soc., 2009, 131, 2766–2767. 46 A. Rassamesard, Y.-F. Huang, H.-Y. Lee, T.-S. Lim, M. C. Li, J. D. White, J. H. Hodak, T. Osotchan, K. Y. Peng, S. A. Chen, J.-H. Hsu, O. M. Hayashi and W. Fann, J. Phys. Chem. C, 2009, 113, 18681–18688. 47 S. Uchiyama, Y. Matsumura, A. P. de Silva and K. Iwai, Anal. Chem., 2004, 76, 1793–1798. 48 I. F. Pierola, D. Radic and L. Gargallo, Chem. Mater., 2006, 18, 2081–2085. 49 C. David, M. Piens and G. Geukens, Eur. Polym. J., 1972, 8, 1019– 1023. 50 G. Kwak, S. Fukao, M. Fujiki and T. Sakaguchi, Chem. Mater., 2006, 18, 2081–2085. 51 S. Uchiyama, N. Kawai, A. P. de Silva and K. Iwai, J. Am. Chem. Soc., 2004, 126, 3032–3033. 52 S. Uchiyama, Y. Matsumura, A. P. de Silva and K. Iwai, Anal. Chem., 2003, 75, 5926–5935. 53 V. A. Vlaskin, N. Janssen, J. van Rijssel, R. Beaulac and D. R. Gamelin, Nano Lett., 2010, 10, 3670–3674. 54 E. J. McLaurin, V. A. Vlaskin and D. R. Gamelin, J. Am. Chem. Soc., 2011, 133, 14978–14980. 55 C.-H. Hsia, A. Wuttig and H. Yang, ACS Nano, 2011, 5, 9511–9522. 56 D. J. Bizzak, Int. J. Heat Mass Transfer, 1995, 38, 267–274. 57 P. Haro-Gonz
alez, I. R. Mart
ın, L. L. Mart
ın, S. F. Le
on-Luis, C. P
erez-Rodr
ıguez and V. Lav
ın, Opt. Mater., 2011, 33, 742–745. 58 D. J. Bizzak and M. K. Chyu, Rev. Sci. Instrum., 1994, 65, 102–107. 59 A. L. Heyes, S. Seefeldt and J. P. Feist, Opt. Laser Technol., 2006, 38, 257–265. 60 M. T. Carlson, A. Khan and H. H. Richardson, Nano Lett., 2011, 11, 1061–1069. 61 Z. P. Cai and H. Y. Xu, Sens. Actuators, A, 2003, 108, 187–192. 62 B. S. Cao, Y. Y. He, Z. Q. Feng, Y. S. Li and B. Dong, Sens. Actuators, B, 2011, 159, 8–11. 63 D. Li, Y. Wang, X. Zhang, K. Yang, L. Liu and Y. Song, Opt. Commun., 2012, 285, 1925–1928. 64 J. Petit, B. Viana and P. Goldner, Opt. Express, 2011, 19, 1138–1146. 65 F. Vetrone, R. Naccache, A. Zamarr
on, A. Juarranz de la Fuente, F. Sanz-Rodr
ıguez, L. Martinez Maestro, E. Mart
ın Rodriguez, D. Jaque, J. Garc
ıa Sol
e and J. A. Capobianco, ACS Nano, 2010, 4, 3254–3258.

Nanoscale, 2012, 4, 4301–4326 | 4325

Published on 16 May 2012. Downloaded by Universidade Federal de Alagoas on 15/09/2015 15:05:40.

View Article Online

66 M. Quintanilla, E. Cantelar, F. Cuss
o, M. Villegas and A. C. Caballero, Appl. Phys. Express, 2011, 4, 022601. 67 M. A. R. C. Alencar, G. S. Maciel, C. B. de Ara
ujo and A. Patra, Appl. Phys. Lett., 2004, 84, 4753–4755. 68 E. Cantelar and F. Cusso, Appl. Phys. B: Lasers Opt., 1999, 69, 29– 33. 69 M. T. Carlson, A. J. Green and H. H. Richardson, Nano Lett., 2012, 12, 1534–1537. 70 C. D. S. Brites, P. P. Lima, N. J. O. Silva, A. Mill
an, V. S. Amaral, F. Palacio and L. D. Carlos, New J. Chem., 2011, 35, 1177– 1183. 71 L. Aigouy, B. Samson, E. Sa, L. Peter and C. Bergaud, Thermal Nanosystems and Nanomaterials, Springer Berlin Heidelberg, Berlin, Heidelberg, 2009, vol. 118. 72 E. Sa€ıdi, B. Samson, L. Aigouy, S. Volz, P. L€ ow, C. Bergaud and M. Mortier, Nanotechnology, 2009, 20, 115703. 73 E. Sa€ıdi, N. Babinet, L. Lalouat, J. Lesueur, L. Aigouy, S. Volz, J. Lab
eguerie-Eg
ea and M. Mortier, Small, 2011, 7, 259–264. 74 B. Samson, L. Aigouy, P. Low, C. Bergaud, B. J. Kim and M. Mortier, Appl. Phys. Lett., 2008, 92, 023101. 75 L. Aigouy, L. Lalouat, M. Mortier, P. L€ ow and C. Bergaud, Rev. Sci. Instrum., 2011, 82, 36106. 76 F. Vetrone, R. Naccache, A. Juarranz de la Fuente, F. SanzRodr
ıguez, A. Blazquez-Castro, E. M. Rodriguez, D. Jaque, J. Garc
ıa Sol
e and J. A. Capobianco, Nanoscale, 2010, 2, 495–498. 77 L. M. Maestro, E. M. Rodriguez, F. Vetrone, R. Naccache, H. L. Ramirez, D. Jaque, J. A. Capobianco and J. Garc
ıa Sol
e, Opt. Express, 2010, 18, 23544–23553. 78 D. K. Chatterjee, A. J. Rufaihah and Y. Zhang, Biomaterials, 2008, 29, 937–943. 79 M. Wang, C.-C. Mi, W.-X. Wang, C.-H. Liu, Y.-F. Wu, Z.-R. Xu, C.-B. Mao and S.-K. Xu, ACS Nano, 2009, 3, 1580–1586. 80 D. Jaque, L. M. Maestro, E. Escudero, E. M. Rodr
ıguez, J. A. Capobianco, F. Vetrone, A. Juarranz de la Fuente, F. SanzRodr
ıguez, M. C. Iglesias-de la Cruz, C. Jacinto, U. Rocha and J. Garc
ıa Sol
e, J. Lumin., 2011, DOI: 10.1016/j.jlumin.2011.12.022, in press. 81 C. D. S. Brites, P. P. Lima, N. J. O. Silva, A. Mill
an, V. S. Amaral, F. Palacio and L. D. Carlos, Adv. Mater., 2010, 22, 4499–4504. 82 Y. Cui, H. Xu, Y. Yue, Z. Guo, J. Yu, Z. Chen, J. Gao, Y. Yang, G. Qian and B. Chen, J. Am. Chem. Soc., 2012, 134, 3979–3982. 83 N. Ishiwada, S. Fujioka, T. Ueda and T. Yokomori, Opt. Lett., 2011, 36, 760–762. 84 M. Seaver and J. R. Peele, Appl. Opt., 1990, 29, 4956–4961. 85 E. J. G. Peterman, F. Gittes and C. F. Schmidt, Biophys. J., 2003, 84, 1308–1316. 86 A. Narayanaswamy, L. F. Feiner, A. Meijerink and P. J. van der Zaag, ACSNano, 2009, 3, 2539–2546. 87 A. Olkhovets, R.-C. Hsu, A. Lipovskii and F. Wise, Phys. Rev. Lett., 1998, 81, 3539–3542. 88 Q. Dai, Y. Zhang, Y. Wang, M. Z. Hu, B. Zou, Y. Wang and W. W. Yu, Langmuir, 2010, 26, 11435–11440. 89 L. Martinez Maestro, E. Mart
ın Rodr
ıguez, F. Sanz Rodr
ıguez, M. C. Iglesias-de la Cruz, A. Juarranz, R. Naccache, F. Vetrone, D. Jaque, J. A. Capobianco and J. Garc
ıa Sol
e, Nano Lett., 2010, 10, 5109–5115. 90 L. M. Maestro, C. Jacinto, U. R. Silva, F. Vetrone, J. A. Capobianco, D. Jaque and J. Garc
ıa Sol
e, Small, 2011, 7, 1774–1778. 91 S. Li, K. Zhang, J.-M. Yang, L. Lin and H. Yang, Nano Lett., 2007, 7, 3102–3105.

4326 | Nanoscale, 2012, 4, 4301–4326

92 R. S. Yang, L. W. Chang, J.-P. Wu, M.-H. Tsai, H.-J. Wang, Y.-C. Kuo, T.-K. Yeh, C. S. Yang and P. Lin, Environ. Health Perspect., 2007, 115, 1339–1343. 93 W. J. Parak, T. Pellegrino and C. Plank, Nanotechnology, 2005, 16, R9–R25. 94 D. R. Larson, W. R. Zipfel, R. M. Williams, S. W. Clark, M. P. Bruchez, F. W. Wise and W. W. Webb, Science, 2003, 300, 1434–1436. 95 L. M. Maestro, J. E. Ram
ırez-Hern
andez, N. Bogdan, J. A. Capobianco, F. Vetrone, J. Garc
ıa Sol
e and D. Jaque, Nanoscale, 2012, 4, 298–302. 96 J.-M. Yang, H. Yang and L. Lin, ACS Nano, 2011, 5, 5067–5071. 97 A. Kawski, Crit. Rev. Anal. Chem., 1993, 23, 459–529. 98 B. Valeur, Molecular Fluorescence: Principles and Applications, Wiley-VCH, 2002. 99 G. Baffou, M. P. Kreuzer, F. Kulzer and R. Quidant, Opt. Express, 2009, 17, 3291–3298. 100 A. Kaminskii, Laser Photonics Rev., 2007, 1, 93–177. 101 A. Benayas, E. Escuder and D. Jaque, Appl. Phys. B: Lasers Opt., 2012, 107, 697–701. 102 S. W. Allison, G. T. Gillies, A. J. Rondinone and M. R. Cates, Nanotechnology, 2003, 14, 859–863. 103 J. Yu, L. Sun, H. Peng and M. I. J. Stich, J. Mater. Chem., 2010, 20, 6975–6981. 104 Z. P. Cai, L. Xiao, H. Y. Xu and M. Mortier, J. Lumin., 2009, 129, 1994–1996. 105 D.-A. Mendels, E. M. Graham, S. W. Magennis, A. C. Jones and F. Mendels, Microfluid. Nanofluid., 2008, 5, 603–617. 106 T. Karstens and K. Kobs, J. Phys. Chem., 1980, 84, 1871–1872. 107 J. J. Shah, M. Gaitan and J. Geist, Anal. Chem., 2009, 81, 8260–8263. 108 A. Nag and D. Goswami, J. Photochem. Photobiol.,A, 2009, 206, 188–197. 109 P. C. Jha, Y. Wsang and H. Agren, ChemPhysChem, 2008, 9, 111– 116. 110 R. K. P. Benninger, Y. Koc¸, O. Hofmann, J. Requejo-Isidro, M. A. A. Neil, P. M. W. French and A. J. deMello, Anal. Chem., 2006, 78, 2272–2278. 111 D. Y. Wu, L. Ugozzoli, B. K. Pal, J. Qian and R. B. Wallace, DNA Cell Biol., 1991, 10, 233–238. 112 M. A. Bennet, P. R. Richardson, J. Arlt, A. McCarthy, G. S. Buller and A. C. Jones, Lab Chip, 2011, 11, 3821–3828. 113 D.-A. Mendels, E. M. Graham, S. W. Magennis, A. C. Jones and F. Mendels, Microfluid. Nanofluid., 2008, 5, 603–617. 114 E. M. Graham, K. Iwai, S. Uchiyama, A. P. de Silva, S. W. Magennis and A. C. Jones, Lab Chip, 2010, 10, 1267–1273. 115 K. Okabe, N. Inada, C. Gota, Y. Harada, T. Funatsu and S. Uchiyama, Nat. Commun., 2012, 3, 705. 116 S. Doxsey, Nat. Rev. Mol. Cell Biol., 2001, 2, 688–698. 117 J. S. Andersen, C. J. Wilkinson, T. Mayor, P. Mortensen, E. A. Nigg and M. Mann, Nature, 2003, 426, 570–574. 118 A. M. Kapitonov, A. P. Stupak, S. V. Gaponenko, E. P. Petrov, A. L. Rogach and A. Eychm€ uller, J. Phys. Chem. B, 1999, 103, 10109–10113. 119 M. Leistikow, J. Johansen, A. Kettelarij, P. Lodahl and W. Vos, Phys. Rev. B: Condens. Matter Mater. Phys., 2009, 79, 1–9. 120 G. Schlegel, J. Bohnenberger, I. Potapova and A. Mews, Phys. Rev. Lett., 2002, 88, 137401. 121 Y. Ebenstein, T. Mokari and U. Banin, J. Phys. Chem. B, 2004, 108, 93–99. 122 P. Haro-Gonz
alez, L. Martinez Maestro, I. R. Martin, J. Garc
ıa Sol
e and D. Jaque, Small, DOI: 10.1002/sml.201102736, submitted.

This journal is ª The Royal Society of Chemistry 2012