Metal oxide nanomaterials are being used for an increasing number of commercial applications, such as fillers, opacifiers, catalysts, semiconductors, cosmetics, microelectronics, and as drug delivery vehicles. The effects of these nanoparticles on the physiology of animals and in the environment are largely unknown and their potential associated health risks are currently a topic of hot debate. Information regarding the entry route of nanoparticles into exposed organisms and their subsequent localization within tissues and cells in the body are essential for understanding their biological impact. However, there is currently no imaging modality available that can simultaneously image these nanoparticles and the surrounding tissues without disturbing the biological structure.
Due to their large nonlinear optical susceptibilities, which are enhanced by two-photon electronic resonance, metal oxides are efficient sources of coherent anti-Stokes Raman Scattering (CARS). We show that CARS microscopy can provide localization of metal oxide nanoparticles within biological structures at the cellular level. Nanoparticles of 20–70 nm in size were imaged within the fish gill; a structure that is a primary site of pollutant uptake into fish from the aquatic environment.
© 2008 Optical Society of America
Nanotechnology has been described as potentially the next industrial revolution and a wide variety of engineered nanoparticles (ENPs) are of interest for medical and industrial applications. The same properties that give ENPs such enormous commercial potential -their large surface area and high degree of reactivity- also makes them potentially biologically harmful to human and ecosystem health. Assessing the potential health risks associated with exposure to nanoparticles is now recognised as being a major international research priority  and the Royal Society urgently recommended that research into the health, safety, and possible environmental impacts of nanomaterials be investigated before they become widespread in the human and natural environment. In response the UK Government and its regulatory agencies are now developing a risk assessment agenda and regulatory regimes for engineered nanoparticles (ENPs). Several other reviews support the call for research into the toxicity, epidemiology, biopersistence and bioaccumulation of ENPs, as well as an examination of their exposure pathways and the development of methods and instrumentation for monitoring their presence and concentrations in the environment . Recent research has demonstrated potentially significant and severe human health effects in lung and skin tissue[4–6]. This has heightened human health concerns for ENPs as currently some are being researched and developed for medical use, including vehicles for delivery drugs. Crucially, for the environment, there is still almost a complete lack of data on environmental fate of currently produced ENPs and very little information on their ecotoxicology.
Some of the earliest adopted synthetic nanoparticles for industrial use were a class of compounds know as metal oxides. Titanium dioxide has a high degree of market penetration in paints, self-cleaning glass, cosmetics, and sunscreens and shows promise for use in other diverse areas such as solar cells, medical technology, and the remediation of sites polluted with organic chemicals. Cerium dioxide is used as a fuel additive and catalyst for removing oxygen from tailpipe emissions from automobiles. Zinc oxide has widespread use as a catalyst and may also be useful in environmental remediation. All are likely to see an increase in environmental release in the near future.
Our freshwater and marine environments act as a sink for waste discharges and aquatic organisms receive some of the highest exposures levels for a wide range of pollutants. Some metal oxide ENPs will be discharged directly into the aquatic environment, others unintentionally (e.g. via the use of sunscreens), but as yet, there is almost a complete lack of data on their environmental concentrations. Studies in aquatic crustaceans, have shown that exposure to metal oxide nanoparticles via the water increased mortality and induced adverse behavioural and physiological changes [7, 8]. Exposure of fish to metal oxide nanoparticles via the water has been shown to induce several adverse effects, including decreases in ion channel activity in the gills and intestine, as well as oedema and thickening of the gill lamellae. However, the mechanisms behind these effects have yet to be established; and to do so requires knowledge of the fate of the metal oxide within the animal model. The first step in establishing how ENPs might induce harm is to establish their entry route and precise location in the animal system, which is not currently possible due to a lack of a suitable imaging modality with sufficient sensitivity and resolution to resolve the particles within an organism without perturbing the system. Furthermore, the ability to locate metal oxide nanoparticles within aquatic organisms could be of use to quantify their uptake from the aquatic environment.
Accurate localization of metal oxide nanoparticles within a biological structure places several constraints upon an imaging modality. Firstly, it must have sufficient spatial resolution to locate the particles at a cellular level. Secondly, the technique must be non-invasive; having sufficient depth penetration to image intact tissues and not rely upon contrast derived from invasive pre-processing. Processing, such as sectioning and staining, inherently perturb the tissue and alters the location of the particles. Thirdly, three-dimensional data is required to provide accurate localisation within a biological system. Finally, and most importantly, a modality is required that can derive contrast from both metal oxides and biological tissues; two materials of very different elemental composition. Current imaging modalities do not meet these criteria. Previous work in this field has relied upon transmission electron microscopy (TEM), which has sufficient resolution to visualize individual nanoparticles. However, it is limited to two-dimensions and requires highly invasive preparatory procedures which lead to the removal of cells and hence alter the position of the nanoparticles. Furthermore it does not give specific contrast of metal oxides.
Optical techniques offer non-invasive imaging and there are several label-free techniques currently available; such as high-resolution optical coherence tomography, two-photon fluorescence and harmonic generation microscopy, all of which have been shown to be suitable for 3-dimensional imaging with sufficient depth penetration in biological specimens. However, other than zinc oxide, which is known to exhibit two-photon fluorescence, these techniques do not produce sufficient contrast of metal oxide nanoparticles. In this study we show that Coherent Anti-Stokes Raman Scattering (CARS) microscopy derives sufficient contrast from titanium dioxide, cerium dioxide, and zinc oxide, to image low particles concentrations deep within a biological structure. We use the fish gill, a complex and highly vascular structure in which to demonstrate the effectiveness of the technique. Moreover, fish gills are a useful model to study the mechanisms of trans-epithelial transport of metal oxide nanoparticles.
CARS is the latest contrast mechanisms to be exploited for biological microscopy and has recently received a great deal of attention (for reviews see references [15–17]). CARS microscopy derives its contrast from intrinsic molecular vibrations in a sample. A pump beam, of frequency ωp and a Stokes beam, ωs, interact with the sample via a four-wave mixing process. When the beat frequency (ωp-ωs) is tuned to match a Raman active vibrational mode, molecules are coherently driven with the excitation fields resulting in the generation of a strong anti-Stokes signal. CARS microscopy is an excellent technique for three-dimensional non-invasive imaging of biological structures[18, 19]; the nonlinear generation of the CARS signal confines optical excitation to a focus where the photon flux is highest, thus providing intrinsic optical sectioning. Furthermore, the use of infrared excitation gives CARS an increased depth penetration over conventional optical microscopy, which removes the need for sample sectioning. Contrast is derived form intrinsic sample properties, removing the need to stain the sample.
However, CARS is not background-free; as with all a four-wave mixing processes, the signal intensity scales with the squared modulus of the third-order susceptibility at the anti-Stokes frequency. Third order nonlinear susceptibility, χ(3), can be expressed in the following general form;
Where Ω is the vibrational frequency of the Raman active vibrational mode; AR and AT are constants representing the Raman scattering and two-photon absorption cross-sections; ωT is the frequency of the electronic transition; and ΓR and ΓT are the widths of the Raman line and the two-photon electronic transition respectively. The first term in Eq. (1) represents the vibrationally resonant contribution; the second, a non-resonant electronic contribution, which is independent of the Raman shift; and the third, a two-photon electronic resonance enhanced nonresonant contribution. Nonlinear susceptibilities in biological samples are generally relatively small; however, adequate contrast is achieved by maximizing the resonant contribution of structures of interest by tuning ωp-ωs to match Raman active modes within structures of interest. For most biological imaging applications the non-resonant contribution limits vibrational contrast to lipids, which are abundant in highly Raman active C-H bonds and hence yield large CARS signals. For most applications in CARS microscopy investigators strive to minimize the non-resonant contributions and hence optimize image contrast. However, for this application we exploit the second and third terms of Eq. (1) to provide exceptional contrast of materials with intrinsically large non-resonant nonlinear susceptibilities. All three metal oxides investigated in this study are know to exhibit high non-resonant third-order susceptibilities; of the order of 10-12 e.s.u [21, 22]. Furthermore, all three are defined as wide bandgap semiconductors, with absorption wavelengths of 375, 390 and 400 nm for ZnO[23, 24], TiO2 and CeO2 respectively. This allows further enhancement of χ(3) by the third term in Eq. (1) when ωp is tuned on or near the two-photon electronic resonance of the bandgap transition [27, 28]. We show that although the size of individual nanoparticles is far too small to be resolved of CARS microscopy, the signal obtained is sufficient to provide location of nanoparticles deep within highly scattering biological tissues.
2. Materials and methods
2.1 Light source
CARS microscopy has strict excitation source requirements. Firstly, tunable dual-wavelength excitation is required to provide the Stokes and pump fields, with the difference in frequency tunable over a range of vibrational modes found in biological samples. Secondly, pulsed excitation is required to provide high Stokes and pump fields whilst keeping the average intensity low enough to avoid sample damage. Previous investigators have shown optimal pulse widths to be of the order of several picoseconds. Efficient CARS generation requires perfect temporal synchronization between the Stokes and pump pulse trains. Thirdly, near infrared excitation is preferable to minimize sample damage and maximize depth penetration. For these reasons we opted for an Optical Parametric Oscillator (Levante Emerald, APE Berlin) pumped with a frequency doubled Nd:Vandium picosecond oscillator (High-Q Laser Production GmbH). The pump laser generates a 6 ps, 76 MHz pulse train of 532 nm laser light with adjustable output power up to 10 W. The Optical Parametric Oscillator (OPO) uses non-critically phased matched optical parametric generation to produce signal and idler beams which exit the laser cavity collinearly with a perfect temporal overlap. The OPO provides continuous tuning over a wide range of wavelengths; from 670 nm to 980 nm for the signal, which is used as the pump, and between 1130 nm and 1450 nm for the idler output, used for the Stokes beam. The maximum combined output power of the signal and idler is approximately 2 W.
2.2 CARS microscope
Imaging was performed using a modified commercial inverted microscope and confocal laser scanner (IX71 and FV300, Olympus UK). A schematic of the optical setup is shown in Fig. 1. The FV300 confocal unit is idea for conversion to multiphoton microscopy and has consequently been used by many previous investigators [29, 30]. It has a simple beam path making it easy to align and its accessible design makes it easy to modify without disturbing other components. To maximize the NIR throughput the standard galvanometer scanning mirrors were replaced with silver galvanometric mirrors and the tube lens was replaced with a MgF2 coated lens. A 60X, 1.2 NA water immersion objective (UPlanS Apo, Olympus UK) was used to focus the laser excitation into the sample. The nonlinear dependence of CARS confines the optical excitation to a focus where the photon flux is highest, bypassing the need for confocal detection. The scanning confocal dichroic was replaced by a silver mirror with high reflectivity throughout the visible and NIR (21010, Chroma Technologies). Due to the directional nature of the CARS generation, simultaneous forwards- and epi- detection is desirable. The redundant internal detectors were replaced by external, non-descanned, detection, which has the advantage of increased efficiency for deep tissue imaging. The forward-CARS signal was collected by an air condenser (NA=0.55) and directed onto a red-sensitive photomultiplier tube (R3896, Hamamatsu) via a mirror and collimating lenses. The epi-CARS signal was collected using the objective lens and separated from the pump and Stokes beams by a long-wave pass dichroic mirror (z850rdc-xr, Chroma Technologies) and directed onto a second R3896 photomultiplier tube at the rear microscope port. The anti-Stoke signal was isolated at each photodetector by a single band-pass filter cantered at 750 nm (HQ750/210, Chroma Technologies). Three-dimensional data was acquired by taking stacks of 2-dimensional images in the x-y plane each separated by an increment in the z-direction, which was achieved by alteration of the objective focus.
Uncoated zinc oxide (>99 %, 50–70nm), titanium dioxide (99.9%, 25–70nm), and cerium oxide (>99%, 20–70nm) nanoparticles were purchased from Sigma-Aldrich (Poole, UK). Size distributions were verified by viewing under a transmission electron microscope (Jeol 100S, Jeol UK). Nanoparticles were diluted to 250µg/l with milliQ water and 10µl were dropped on to copper 200 hexagonal mesh grids and examined at 80 kV. As a simple model system in which to investigate epi- vs forward-CARS detection, the metal oxide nanoparticles were suspended in agarose. Low temperature agarose gel and histological stains were obtained from Sigma-Aldrich (Poole, UK) and used without further purification.
2.4 Fish handling and nanoparticle exposures
Rainbow trout, Onchrhynchus mykiss, a species widely used in aquatic (eco)toxicology, approximately 200g in weight and 25cm in length were obtained from Houghton Springs Fish Farm (Dorset, UK) and maintained in 500 L tanks supplied via a flow-through system with deionised tap water. Environmental conditions were simulated with a water temperature maintained between 9 and 11°C and 12/12 hour light/dark cycle. Fish were fed a maintenance ration of food (Emerald Fingerling 30, Skretting, UK) 1% body weight and starved for 3 days prior to the experiments. For dosing the fish tanks, stock suspensions of the nanoparticles were prepared by suspending 2.5g/l of each compound in milliQ water. The suspensions were each agitated vigorously for 30 seconds and then sonicated for 1 hour to break up large particle aggregates. In addition, stock suspensions were sonicated for 30 minutes before dosing tanks and fish. Suspensions of nanoparticles (TiO2, CeO2, or ZnO) were sampled with a 1ml pipette and injected into agar for imaging. Fish were exposed to 5000µg/L titanium dioxide for 24h for an acute dose or 14 days for short-term dose. The pH throughout the experiments maintained between 7.3 and 7.5, with total conductivity ranging from 183 to 201 µS cm-1. The cation content of the water in the experimental aquaria were: Na+=8.27mg/l, K+=2.07 mg/l, Mg2+=4.38 mg/l, and Ca2+=24.50 mg/l.
Fish were terminated by euthanasia (MS222) and brain destruction, according to Home Office Animals’ License procedures, and gill filaments were dissected and blotted in cold trout ringer’s solution (pH 7.4), fixed in trout ringer’s with 3% glutaraldehyde, and imaged within 2 hours. At all times extreme care was taken to avoid cross contamination between normal and dosed fish/tissues.
2.5 Image processing
Processing of the 3-dimensional data sets was performed using OsiriX (OsiriX, version 1.7.1, 2005, open-source software). To compare the CARS signal intensity of the metal oxides at different excitation wavelengths signal was normalised against the nonresonant signal a region of the image containing only a bare coverslip. The nonresonant signal is independent of molecular orientation and vibration frequency, but varies with the excitation intensity and detection efficiency at different wavelengths. To rule out the dependence of the CARS intensity on the setup, we normalized the CARS signal from the sample with the nonresonant CARS signal from the coverslip.
3. Results and discussion
3.1 Imaging normal gill structures
In order to interpret the CARS images it is first necessary to briefly describe the normal structure and function of the fish gill. On each side of the fish there are four gill arches; each bearing a double row of elongated, laterally projecting structures which are referred to as gill filaments. On the upper and lower surface of each filament, projecting at right angles to its axis, are rows of closely packed, leaf-like structures called secondary lamellae. It is in these lamellae that gaseous exchanges takes place.
The lamella (L), shown in Fig. 2, are composed of two parallel sheets of epithelia separated by a narrow space through which blood circulates [32, 33]. A TEM of the gill lamellae is shown in Fig. 2(a). Blood flows from the central venous space (CVS) via capillaries (C) into the secondary lamellae space. The epithelium (E) forms the water/blood barrier and consists predominantly of pavement cells (PVC), which are squamous in appearance and cover 95% of the lamellae surface area. Mitochondrion-rich chloride cells (CC) are the next most abundant epithelial component and are responsible for ion-transport . Separation between the epithelial sheets is maintained by post-like pillar cells (PC), which extend between the two epithelial layers. These cells are analogous to posts in mammalian alveoli and restrain the lamellae under the internal force of blood pressure. Blood flow, and therefore gaseous exchange, takes place in the blood channels (BC) which are bounded by the pillar cells.
To identify the structures in the CARS images a comparison with a histologically stained section was performed. Figures 2(b) and 2(c) compare two longitudinal sections from the same region through the lamellae. Figure 2(c) shows a 60X wide field image of a Schiff, carbohydrate specific, stained slice taken approximately 10 microns deeper than an unstained slice imaged with CARS, shown in Fig. 2(b). The predominant components of the lamellae have been labeled. The Stokes and pump wavelength were 1255 nm and 924 nm respectively, providing contrast of CH rich structures. The CARS image shows good position agreement with Fig. 2(c). Areas with large CARS signal correspond to structures showing high stain uptake in Fig. 2(c).
Previous investigators  have shown that the coherent nature of CARS signal produces a far-field radiation pattern that depends on the size and shape of scatterers. As a consequence forward- (F-CARS) and epi-detected (E-CARS) images can provide complimentary information about a sample. The fish gill is rich with scatters (both cellar and extracellar) of differing shape and size. To investigate the directional effects of our sample and hence fully exploit the structural information we compared E- and F-CARS images of an un-dosed gill.
Figures 3(a) and 3(b) compare E-CARS and F-CARS images of the gill lamellae. The images are orientated such that the secondary lamellae lie in the x-y imaging plane; orthogonal to those shown in Fig. 2(d). The most significant difference between the two images is the background noise in the forward image. This arises from non-resonant electronic contribution to the CARS signal, which in the forward direction is generated in the surrounding buck media, such as a solvent. The epi-signal has been shown to reject the non-resonant background due to destructive interference of the backward CARS signal in scatterers larger than the excitation wavelength. Epi-detection allows detection of small features and can also arise from an interface between a sizable scatterer and its surrounding medium. On the other hand, forward detection is required for imaging objects with an axial length comparable to or larger than the excitation wavelength.
The structure of secondary lamella is clearly visible in both the epi- and forward-CARS images. The pillar cells (PC) and epithelium (PVC) produce contrast due their lipid content. The striking difference between the two images is the red blood cells (RBC). In the forward-CARS image RBCs show very high contrast against the lamellae structure, and can be seen in the spacing between the pillar cells. However, in Fig. 3(b) the RBC are barely visible. This difference is explained by the spatial dependence of CARS generation. The RBCs are homogeneous objects far greater in size than the excitation wavelength, and therefore do not generate epi-CARS. Previous investigator  have show that χ(3) for hemoglobin exhibits two-photon resonance around the pump wavelength employed in this study, which account for the uniform high F-CARS signal from the centre of the cells.
The pillar cells are visible in both figures (a) and (b), however, the epi-CARS image shows more intercellular detail. In Fig. 3(a) periodically distributed bright ‘dots’ are visible around the periphery of the cells, which do not appear in Fig. 3(b). These ‘dots’ have previously been studied using confocal immunofluorescence  and correspond to strands of collagenous extracellular matrix proteins that provide tension which prevent ballooning of the lamellae.
Combined epi- and forward-CARS provides an excellent visualistion of the microvasculature structure. Figure 3(c) shows the epi- (green) and forward-CARS (blue) data from Figs 3(a) and 3(b) combined in a single image. Figure 3(d) shows the same region under increase magnification. This visualisation clearly shows the RBCs occupying the spaces between the pillar cells. The structural arrangement of the pillar cells is responsible for orienting the RBC to face that their largest surface area towards the direction of water flow, therefore the RBCs appear as elongated rods in the x-y image plane rather than bi-concaved discs.
3.2 CARS detection of metal oxide nanoparticles
As discussed earlier, epi- and forward-CARS images contain different, and often complementary, information depending on the size, shape and distribution of scatters and the properties of their surrounding medium. We used nanoparticles suspended in agraose as a simple model system in which to investigate epi- vs forward-CARS detection of the metal oxide nanoparticles.
Representative E- and F-CARS images of the three types of metal oxide particles embedded in agarose are shown in Fig. 4. Due to their large χ(3) and two-photon resonance, all three types of metal oxide generate a significantly large CARS signals in both epi- and forward-detection. The most obvious difference between the forward- and epi-images is the large signal from the agarose in the forwards-detected images, which is not detected in the epi direction due to destructive interference of the backwards CARS field from bulk media. All three forward images show at least an order of magnitude greater peak intensity than their corresponding epi-detected images. However, the absence of signal from the agarose in the epi images greatly enhances the contrast of the nanoparticles by a factor of 10.
An interesting characteristic of the forwards images is the appearance of a dark ‘halo’ around the particles of lower intensity than the surrounding agar. Previous investigators [36, 37, 40] have ascribed this ‘dip’ in intensity to destructive interference between CARS field from the scatterer and it’s surrounding media or the mismatch in refractive index between the scatterer and the media, which distorts the foci of the excitation fields and reduces the forward-detected signal.
Due to particle aggregation, the gels contain particles of various sizes. Another obvious difference between the E- and F-CARS images is that only the larger particles appear in the forwards images. The smaller particles present in the epi-images relate to black-holes, or are not present, at corresponding locations in the forwards-images. Scanning in the z-direction verified that the absence of these particles was not due to particles being out of the imaging plane. This absence of smaller particles in the F-CARS images can be explained by the same phenomenon responsible for the ‘dip’ in signal at the particle-agar interface, or purely due to the lower signal from the smaller particles falling below the background level of the agarose.
It is not possible to determine the size of the smallest resolvable aggregate due to the point-spread-function of the imaging system exceeding the size of an individual nanoparticle. This is responsible for uneven brightness of particles appearing to be of similar size; the effect is exaggerated by the nonlinear concentration dependence of the CARS signal.
3.3 Three-dimensional imaging of nanoparticles
To accurately locate the metal oxides nanoparticles within the complex structure of the lamellae, three-dimensional data is required. Z-stacks were acquired by taking a series of 2-dimensional images (x-y plane) each separated by an increment of 0.25 µm in the z-direction.
The data presented in Fig. 5 is a typical example of an isolated nanoparticle or small nanoparticle cluster in a gill exposed to titanium oxide nanoparticles for a period of two weeks. The figure shows a 2-dimensional representation of 3-dimensional image stack of a fish gill that has been exposed to titanium dioxide nanoparticles for a period of two weeks. A small cluster of extremely bright/saturated pixels, corresponding to a diameter of approximately 1 µm in diameter appear to be in the capillary [labelled C in Fig. 2(c)]. It is not possible from examination of x-y plan alone to conclude whether the particles are located inside the capillary, and hence verify that metal oxides can indeed cross the epithelial membrane. However, examination of the reconstructed x-z and y-z planes verifies that the nanoparticles are indeed inside the capillary.
As well as finding small particles within fish gills exposed to metal oxides, large particle aggregates ranging from 10s to 100s µm in size where also observed. These larger aggregates are better visualized using a 3-dimensional projection. Figure 6 shows a 3-dimensional projection of a 150 µm z-stack of a representative example of a gill containing a large particle aggregate following a one week exposure to titanium dioxide nanoparticles. The aggregated particles can be seen to occupy the region between the secondary lamellae in the vicinity of the capillary. Unlike those in Fig. 5, the particle is trapped within the mucus layer coating the lamellae. Smaller particles can be seen at the periphery of the particle aggregate, which may be correspond to the smaller particles that subsequently cross the epithelial membrane after an extended exposure, such as that shown in Fig. 5.
We have shown that CARS microscopy provides excellent label-free contrast of, potentially harmful, metal oxide nanoparticles deep within a biological structure. Although individual particles can not be identified, they can be located at the cellular scale, which is necessary to provide information on the entry route and final location of nanoparticle in a biological system. To this end, CARS microscopy is an excellent tool for application in ecotoxicological assessments of the potential for metal oxides entering the environment to cause biological harm.
We found that tuning ωp−ωs to match the lipid CH stretch gave excellent contrast of the metal oxides against the surrounding biological structure. To achieve this vibrational resonance, a pump wavelength of 924 nm was employed. Although the pump wavelength is off resonance with the two-photon excitation of the semi-conductor bandgap the signal obtained from all three metal oxides was found to produce sufficient signal for their detection, and in most cases, was found to saturate the photodetector. Tuning ωp to exactly match the bandgap resonance would increase the CARS efficiency of the nanoparticles and allow low excitation powers to be employed. However, resonance excitation of the metal oxide at powers required to produce contrast from the tissue are likely to induced photo damage of the metal oxide and thus complicating their detection. Furthermore, using a longer wavelength excitation allows deeper imaging within the tissue.
We have shown that E-CARS provides superior detection of the nanoparticles over F-CARS; giving better contrast against a bulk solvent and allowing detection of smaller particle aggregates. However, the F-CARS images contain complimentary information regarding the biological structures in the tissue, which, when combined with the E-CARS images aid the localization of the nanoparticles within the gill structure.
Although we have only shown results for the detection of TiO2 nanoparticles in the fish gill, the results from the agarose phantom study indicate that similar contrast would be obtained from the ZnO and CeO particles.
For this investigation we used the fish gill as a model system, as this is likely to be a primary target tissue and route of uptake for metal oxide nanoparticles into this organism from the aquatic environment. However, the results presented in this study can be extended to other organisms and tissues, such as lung alveoli in humans, where the potential harmful effects of ENPs is also of concern.
We wish to acknowledge the help of the Xie group at Harvard University for generously sharing information regarding CARS microscopy during the 4th Annual CARS workshop in 2007. Special thanks go to Brian Saar for his advice on setting up the CARS microscope.
BDJ was supported by the Natural Environmental Research Council (NE/D004942/1). Thanks also go to Tessa Scown for help with animal husbandry and fish handling during the experiment.
References and links
2. R. Society, Nanoscience and Nanotechnologies: Opportunities and Uncertainties, 2004). http://www.nanotec.org.uk/finalReport.htm
3. G. Oberdörster, E. Oberdörster, and J. Oberdörster, “Nanotoxicology, An emerging discipline evolving from studies of ultrafine particles, environmental health perspective,” Environ. Health Perspect. 113, 823–839 (2005). [CrossRef] [PubMed]
5. A. A. Shvedova, V. Castranova, E. R. Kisin, D. Schwegler-Berry, A. R. Murray, V. Z. Gandelsman, A. Maynard, and P. Baron, “Exposure to carbon nanotube material: Assessment of nanotube cytotoxicity using human keratinocyte cells,” J. Toxicol. Environ. Health: Part A 66, 1909–1926 (2003). [CrossRef]
6. D. B. Warheit, T. R. Webb, C. M. Sayes, V. L. Colvin, and K. L. Reed, “Pulmonary instillation studies with nanoscale TiO2 rods and dots in rats: Toxicity is not dependent upon particle size and surface area,” Toxicol. Sci. 91, 227–236 (2006). [CrossRef] [PubMed]
7. S. B. Lovern and R. Klaper, “Daphnia magna mortality when exposed to titanium dioxide and fullerene (C60) nanoparticles,” Envir. Toxicol. Chem. 25, 1132–1137 (2006). [CrossRef]
8. S. B. Lovern, J. R. Strickler, and R. Klaper, “Behavioural and physiological changes in Daphnia magna when exposed to nanoparticle suspension (titanium dioxide, nano-C60, and C60HxC70Hx).” Environ. Sci. Technol. 41, 4465–4470 (2007). [CrossRef] [PubMed]
9. G. Federici, B. J. Shaw, and R. D. Handy, “Toxicity of titanium dioxide nanoparticles to rainbow trout (Oncorhynchus mykiss): Gill injury, oxidative stress, and other physiological effects,” Aquatic Toxicol. 84, 415–430 (2007). [CrossRef]
10. R. Leitgeb, W. Drexler, A. Unterhuber, B. Hermann, T. Bajraszewski, T. Le, A. Stingl, and A. Fercher, “Ultrahigh resolution Fourier domain optical coherence tomography,” Opt. Express 12, 2156–2165 (2004). [CrossRef] [PubMed]
12. P. J. Campagnola and L. M. Loew, “Second-harmonic imaging microscopy for visualizing biomolecular arrays in cells, tissues and organisms,” Nat. Biotechnol. 21, 1356–1360 (2003). [CrossRef] [PubMed]
13. F. Stracke, B. Weiss, C. M. Lehr, K. Koenig, U. F. Schaefer, and M. Schneider, “Multiphoton microscopy for the investigation of dermal penetration of nanoparticle-borne drugs,” J. Invest. Dermatol. 126, 2224–2233 (2006). [CrossRef] [PubMed]
14. V. Bagalkot, L. Zhang, E. Levy-Nissenbaum, S. Jon, P. W. Kantoff, R. Langer, and O. C. Farokhzad, “Quantum dot - Aptamer conjugates for synchronous cancer imaging, therapy, and sensing of drug delivery based on Bi-fluorescence resonance energy transfer,” Nano Lett. 7, 3065–3070 (2007). [CrossRef] [PubMed]
15. J. X. Cheng and X. S. Xie, “Coherent anti-Stokes Raman Scattering Microscopy: Instrumentation, Theory, and Applications,” J. Phys. Chem. B 108, 827–840 (2004). [CrossRef]
16. L. G. Rodriguez, S. J. Lockett, and G. R. Holtom, “Coherent anti-stokes Raman Scattering Microscopy: A Biological Review,” Cytometry Part A 69A, 779–791 (2006). [CrossRef]
17. E. O. Potma, X. L. Nan, E. Conor, and X. S. Xie, “Cars microscopy: Coming of age,” Abstracts of Papers of the American Chemical Society 228, U291–U291 (2004).
18. C. L. Evans, E. O. Potma, M. Puoris’haag, D. Cote, C. P. Lin, and X. S. Xie, “Chemical imaging of tissue in vivo with video-rate coherent anti-Stokes Raman scattering microscopy,” Proceedings of the National Academy of Sciences of the United States of America 102, 16807–16812 (2005). [CrossRef] [PubMed]
19. C. L. Evans, X. Xu, S. Kesari, X. S. Xie, S. T. C. Wong, and G. S. Young, “Chemically-selective imaging of brain structures with CARS microscopy,” Opt. Express 15, 12076–12087 (2007). [CrossRef] [PubMed]
20. Y. Shen, The Principles of Nonlinear Optics, (John Wiley and Sons, 1984).
21. M. C. Larciprete, D. Haertle, A. Belardini, M. Bertolotti, F. Sarto, and P. Gunter, “Characterization of second and third order optical nonlinearities of ZnO sputtered films,” Appl. Phys. B 82, 431–437 (2006). [CrossRef]
22. W. E. Torruellas, L. A. Weller-Brophy, R. Zanoni, G. I. Stegeman, Z. Osborne, and B. J. J. Zelinski, “Thirdharmonic generation measurement of nonlinearities in SiO2-TiO2 sol-gel films,” Appl. Phys. Lett. 58, 1128–1130 (1991). [CrossRef]
23. X. W. Sun and H. S. Kwok, “Optical properties of epitaxially grown zinc oxide films on sapphire by pulsed laser deposition,” J. Appl. Phys. 86, 408–411 (1999). [CrossRef]
24. L. Ja-Hon, C. Yin-Jen, L. Hung-Yu, and H. Wen-Feng, “Two-photon resonance assisted huge nonlinear refraction and absorption in ZnO thin films,” J. Appl. Phys. 97, 033526 (2005). [CrossRef]
25. S. P. Kowalczyk, F. R. McFeely, L. Ley, V. T. Gritsyna, and D. A. Shirley, “The electronic structure of SrTiO3 and some simple related oxides (MgO, Al2O3, SrO, TiO2),” Solid State Communications 23, 161–169 (1977). [CrossRef]
26. S. Tsunekawa, J. T. Wang, Y. Kawazoe, and A. Kasuya, “Blueshifts in the ultraviolet absorption spectra of cerium oxide nanocrystallites,” J. Appl. Phys. 94, 3654–3656 (2003). [CrossRef]
27. J. J. Burris and T. J. McIlrath, “Theoretical study relating the two-photon absorption cross section to the susceptibility controlling four-wave mixing,” J. Opt. Soc. Am. B 2, 1313 (1985). [CrossRef]
29. A. Majewska, G. Yiu, and R. Yuste, “A custom-made two-photon microscope and deconvolution system,” Pflugers Arch. Eur. J. Physiol. 441, 398–408 (2000). [CrossRef]
30. H. F. Wang, Y. Fu, P. Zickmund, R. Y. Shi, and J. X. Cheng, “Coherent anti-Stokes Raman Scattering Imaging of axonal myelin in live spinal tissues,” Biophys. J. 89, 581–591 (2005). [CrossRef] [PubMed]
31. OsiriX, “http://www.osirix-viewer.com/,”
32. G. M. Hughes and A. V. Grimston, “Fine Structure of Secondary Lamellae of Gills of Gadus Pollachius,” Quarterly Journal of Microscopical Science 106, 343–353 (1965).
33. G. M. Hughes and S. F. Perry, “Morphometric study of Trout Gills - Light-Microscopic Method suitable for evaluation of pollutant action,” J. Exp. Biol. 64, 447–460 (1976).
34. G. G. Goss, S. F. Perry, J. N. Fryer, and P. Laurent, “Gill morphology and acid-base regulation in freshwater fishes,” Comp. Biochem. Physiol. Part A Mol. Integr. Physiol. 119, 107–115 (1998). [CrossRef]
35. S. S. Sobin, H. M. Tremer, and Y. C. Fung, “Morphometric basis of sheet-flow concept of Pulmonary Alveolar Microcirculation in Cat,” Circ. Res. 26, 397–414 (1970). [PubMed]
36. J. X. Cheng, “Theoretical and experimental characterisation of Coherent anti-Stokes Raman Scattering (CARS) Microscopy,” J. Opt. Soc. Am. B 19, 1363–1375 (2002). [CrossRef]
37. A. Volkmer, J. X. Cheng, and X. S. Xie, “Vibrational imaging with high sensitivity via epidetected coherent anti-Stokes Raman scattering microscopy,” Phys. Rev. Lett. 8702, (2001).
38. G. O. Clay, A. C. Millard, C. B. Schaffer, J. Aus-Der-Au, P. S. Tsai, J. A. Squier, and D. Kleinfeld, “Spectroscopy of third-harmonic generation: evidence for resonances in model compounds and ligated hemoglobin,” J. Opt. Soc. Am. B 23, 932–950 (2006). [CrossRef]
39. H. Kudo, A. Kato, and S. Hirose, “Fluorescence visualization of branchial collagen columns embraced by pillar cells,” J. Histochem. Cytochem. 55, 57–62 (2007). [CrossRef]
40. D. Gachet, F. Billard, N. Sandeau, and H. Rigneault, “Coherent anti-Stokes Raman Scattering (CARS) Microscopy imaging atinterfaces: evidence of interference effects,” Opt. Express 15, 10408–10420 (2007). [CrossRef] [PubMed]