Biomaterials - DOC by domainlawyer

VIEWS: 69 PAGES: 15

									1

Collaborative Research: Photonic Crystals in Biology: Butterfly Scales
1. Introduction
The beautiful iridescent colors found on the wings of butterflies have attracted the attention of brilliant minds over the past centuries starting with Newton, who understood that these colors must be due to the presence of “thin film structures” [1]. Since then, much progress has been made and it is now recognized that many of these brilliant colors are due to various kinds of microstructures present on butterfly wing scales [2-6]. In other words, the colors are produced by periodic structures of cuticle-air that mimic photonic crystals [7-11]. Many butterflies across all five major families (Nymphalidae, Lycaenidae, Riodinidae, Papilionidae, and Pieridae) display striking colors that result from the nano-morphology on the surface of their wing scales [4, 5, 12-15]. These structural colors can include deep blacks [16], reds, oranges and greens [17, 18], as well as the more common blues, violets, ultraviolets, and whites [19-23]. Typically these colors appear metallic due to the saturation or the purity of the colors produced [3]. The colors perceived by an observer results from the various types of interaction of the incident light with the nanostructure on the wings of the butterfly scales. This interaction may lead to the incident light being scattered in all directions, producing a color that is independent of the viewing angle [24], or it can have a strong directionality, producing colors that are only seen at certain angles [3, 6, 25-27]. Biologists and Physicists alike have been drawn to study the nanomorphology of these scales, by randomly sampling a few specimens across the 17,000-specimenrich butterfly superfamily (the Papilionoidea). This initial sampling has uncovered a relatively broad range of morphological structures that are involved in producing the structural colors. Because of the relatively sparse number of specimens examined to date, and because many of these specimens are not closely related, it is still unclear how readily these nano-morphologies can change to give rise to different colors, and what are the minimal number of morphological changes that can produce a change in a scale’s color. In this proposal we will describe and model the structural colors of a group of closely related species of butterflies with remarkably different colors. We will also apply artificial selection to a population of live butterflies over several generations to change the color of its wings from brown to blue in order to describe the causative changes in the scale’s nano-morphology. It is our hope that the resulting insight into the photonics of butterfly wing scales will stimulate a rapidly developing field of fundamental physical and biological research. It is also our hope that our results will inspire technical innovations in the field of applied material science to generate such structures on other substrates.

2. Specific aims
The aim of the present proposal is to combine descriptive and theoretical approaches coupled with experimental approaches to understand the structural colors of butterfly wing scales. We will take into consideration the evolutionary relationships among species, in order to properly evaluate the direction of morphological modification and the number of independent nano-structural solutions to similar “colors”, when more than one solution is found. The evolutionary framework will also help us predict which ancestral scale nanostructures most easily evolve into a diverse array of different colors. This information will facilitate the future design of synthetic nanostructures for photonic applications. By probing nature’s diversity, we hope to identify alternative solutions to the same “color” that may differ in ease of artificial synthesis, or archetypical solutions that can be easily modified to produce significantly different colors. 2.1. Study of the morphology and optical properties of butterfly wing scales

2
 We will perform a detailed analysis of the purple, violet, blue, white, yellow, orange, and brown scales on the wings of the members of a single butterfly genus, Bicyclus, and characterize their scale structural morphology using light microscopy, scanning electron microscopy (SEM), and transmission electron microscopy (TEM) (see Figs. 1 and 2).  We will describe their wing scales according to a series of parameters, e.g., scale color, UV reflectivity, ultra-structure, and iridescence.  A theoretical-computational analysis of the optics will be performed, using the obtained structural data. We will determine whether scale color reflectivity, ultra-structure, and other quantifiable variables of butterfly wing scales are co-varying with each other, i.e., whether a certain color always appear associated with a particular scale ultra-structure and vice-versa.  We will perform angle-dependent reflection and transmission microspectrophotometry of individual wing scales, which will yield the necessary data to test and fine-tune the theoretical models. We would like to point out that such measurements are nonexistent with Vukusic’s work being the single exception [24].

 We will determine whether different colors present on the same individual are more likely to
have similar scale ultra-structure, than similar colors found across individuals of the same species, or across different species. Figure 2 (left). Evolution of transverse band colors on the dorsal forewing of females of the genus Bicyclus. Estimation of ancestral states was carried out using Figure 1 (above). The diversity parsimony of scale coloration on the dorsal implemented in surface of the forewing of several MacClade 4.0. Bicyclus species. The white pupils Character states were of the eyespots of B. anynana (top assigned to terminal left), as well as the white, purple, taxa based on all the and blue transversal bands are transverse band color structural colors that reflect UV variants found in Bicyclus. light.

2.2. Experimental approaches to modifying particular colors on the wing  One of the most exciting things we intend to pursue is to target a change in color of the scales using artificial selection in replicate lines, and observe which morphological components of the ultra-structure of the scale change as a result. We will use the “lab-rat” butterfly, Bicyclus anynana, to experimentally modify wing scale coloration from brown, found in B. anynana, to blue or purple, found in other Bicyclus species (Fig. 1).

3. Background and significance of project relative to long-term goals of the investigators
3.1. Monteiro Lab: The long-term goals of the Monteiro Lab are to explore the molecular and developmental mechanisms of morphological evolution using butterflies as model organisms.

3
This lab is the only lab that has established transgenic technology for butterflies [28] and is currently developing functional genetic tools in order to test the role of several candidate genes in differentiating specific colors on butterfly wings [29]. This lab has extensive experience with artificial selection experiments on pigment-based color patterns [30-32] but would like to continue exploring structural-based colors as well [33]. The traits previously selected were the rings of colored scales, present in the margin of the wings, which make up an eyespot pattern. These experiments demonstrated that populations of this species of butterfly carry substantial amounts of genetic variation for eyespot size, ring color composition, and eyespot shape. In all cases the mean value of the trait under selection changed to values well beyond those present in the individuals of the original population, and resembling the phenotypes of different species of the genus. It has never been assessed in Bicyclus, or in any other butterfly or moth, whether artificial selection can be employed to change the nano-structure of the wing scales, i.e., whether there is sufficient genetic variation underlying morphology of scale nano-structures in individuals of a population, and whether the alleles of the genes that produce such variation can be recombined in new ways so as to produce new structural colors on the wing. Heritability experiments performed on the “brightness” of the structural scales of another butterfly, Colias eurytheme, showed that this trait has an heritability of around 50-60% (R. Rutowski, personal communication), which is comparable to the estimates obtained in Bicyclus for eyespot traits. It is very likely, therefore, that selection on the structural component of the brown scales in B. anynana will also yield a strong response to selection. The genus Bicyclus contains around 80 species showing dramatic changes in the color patterns on their wings [34, 35]. Structural colors in Bicyclus butterflies are present in scales at the centers of the eyespot patterns [36], in brown-purple colored rings around the eyespots, and in transversal white, blue, and purple bands of scales in the dorsal surface of the wings (Fig. 1). All of these Figure 3. a) SEM photograph scales of different colors are of the wing area where the large Figure 4. SEM photographs of the posterior eyespot on the dorsal highly reflective in the UV [37]. ultrastructure of a) white, b) black, surface of B. anynana is located To date, only the scales of a and c) gold scales that make up the (red rings indicate the rings of the eyespot pattern on the single Bicyclus species, B. approximate areas of the dorsal wings of B. anynana. d) anynana, have been studied at differently colored scales), b) brown background scales on the the level of their ultra-structure close-up of the white central same surface. Darker colors, with scales, c) black disc scales, d) using low-resolution scanning the exception of the white scales, gold ring scales, and e) brown have more cross-ridges (vertical electron microscopy (SEM) background scales. lines) per unit area. (Figs. 3 and 4) [33]. The experiments outlined here, where we will attempt to alter scale color in Bicyclus anynana via alteration of the nano-structures at the surface of wing scales will be the first experiments of its kind. As in the selection experiments for eyespots traits, however, we expect to attain gradual changes in the structural morphology of the B. anynana brown background scales that that will mimic the structural colors of other Bicyclus species. Ultimately, and using the selection experiments described in this project, it will be possible to begin targeting the genes involved in producing the specific scale nano-structures.

4
One of the most well established techniques in Quantitative Genetics to find genes responsible for specific morphologies, quantitative trait loci (QTL) mapping, involves producing divergent populations of organisms for the trait under study (e.g. large versus small, blue scales versus brown scales, etc.), using artificial selection over several generations [38]. Once the populations have evolved non-overlapping morphologies, a few individuals are crossed and back-crossed between the two populations, and their offspring are measured with respect to the trait of interest. An association study is then performed between specific genotypic markers. 3.2. Srinivasarao lab: One of the long-term goals of the Srinivasarao lab is to gain a better understanding of how light interacts with matter sculptured at the nanometer scale to generate various optical phenomena. The lab also wants to replicate some of these nano-morphologies and has already had some success toward these goals (see below). The work detailed in the current grant will broaden our understanding of the multiple ways that nano-morphologies can produce color.

4. Aims and results from previous support
4.1 Monteiro has been supported by NSF since 2003. The aims of the first grant, IBN–

0316283 “The role of developmental genes in controlling butterfly eyespot patterns” (08/01/03-12/31/05, $240,00) were to develop transgenic tools for butterflies. The results from
this work resulted in a highly publicized publication [28] that was featured in BBC News and several other newspapers and websites around the World. The second grant IOB-0516705 (10/01/05-09/30/08), currently ongoing,, aims to apply these transgenic tools to test the role of several candidate genes in the development of the wing color patterns. These grants have supported a total of eleven publications [28, 29, 36, 39-46] , one other [36] highly publicized and featured in a piece in Science magazine and in the German newspaper, Der Spiegel, among other media; trained one PhD student who is a woman minority, a member of a Native American tribe, the Choctaw Nation of Oklahoma; six additional MS and MA students (3 women); and 19 undergraduates (10 women). Also, since 2003, A Monteiro has been invited to speak at 26 venues, including a Gordon Conference on Molecular Evolution, and has co-organized 3 international meetings and/or symposia. 4.2 Srinivasarao currently has 5 active NSF grants, DMR-0312792 (11/15/03-10/31/07) and DMI-0423619 (07/01/04-06/30/07), DMR-0637233 (08/15/06-07/31/08), CMS-0600600 (07/01/06-06/30/09), DMR-0603026 (07/01/06-06/30/09) and has completed CAREER Award, DMR-0096240. Currently this has resulted in 22 papers published, 8 submitted, 4 in preparation, and 55 invited talks presented worldwide. We summarize here only part of our results from the awards due to the space constraints.
4.2.1 Anchoring Transitions of Nematic Fluids at a Polymer Interface: Control by Sidechain Branching. We reported a

temperature-driven anchoring transition in a polymer/nematic fluid composite [47], where the transition temperature is far from the bulk nematic-isotropic transition temperature. In order to probe the subtle effects of the side chain structure of the polymer on control of the anchoring, a series of polyacrylates with different side

Figure 5. Left: Polarized microscopy (POM) images under crossed polarizers, showing only planar anchoring in a wide range of temperature. Right: POM images under crossed polarizers, showing homeotropic-toplanar anchoring transition.

5
chain structures have been synthesized for the. A polymer dispersed liquid crystal (PDLC) film made from TL205/1-methyl-heptyl acrylate shows only planar anchoring in the temperature range of -14o C to TNI (nematic-isotropic transition temperature), while the films made from all other methylheptyl acrylates are homeotropic at low temperature and show the homeotropic-toplanar anchoring transition at temperatures from 70C to 78C, which are close to that of TL205/poly(n-heptyl acrylate) system, as shown in Fig. 5. We postulated that this dramatic difference in the anchoring is due to a tilted conformation of the 1-methylheptyl side chain. By using two different acrylates in the photopolymerization, it is possible to tune the anchoring transition temperature at the copolymer interface to be in a wide range of temperatures between those of the individual homopolymers [48]. When these two acrylates have different anchoring tendencies and different polymerization rates, photopolymerization through a photomask produces a spatially varying anchoring that leads to a spatially varying refractive index, which then behaves as a diffraction grating [49]. We also demonstrated that stable Bloch walls [50] can be successfully trapped and may be analyzed to provide a very simple method to determine the anchoring energies. Such walls are not stable in nematic fluids but are in our system and we have studied the 3-dimensional structure by using a laser scanning confocal microscope in the polarization mode. We also observed that the LC anchoring is governed by the control of polymer structures in phase-separated LC-polymer composite structures [51].
4.2.2 Three-Dimensionally Ordered Array of Air Bubbles in a Polymer Film. We

reported the formation of a three dimensionally ordered array of air bubbles of monodisperse pore size in a polymer film through a templating mechanism based on thermo-capillary convection [52] (See Fig. 6). We have also demonstrated that both linear polymers [52] and highly conjugated semiconducting polymers with stiff backbones [53, 54] can also generate such films (Figs. 7 and 8), supporting our model [52, 55]. This finding is in contrast to the arguments in the literature that, in order for such ordered structure formation, one needs Figure 6. Schematic illustration of a mechanism to have polymers with a given architecture proposed for the breath figure process, by which 3dimensionally ordered array of holes are formed in (like star polymers) or copolymers that have polymer films. (g) polystyrene blocks. The dimensions of these bubbles can be controlled simply by changing the velocity of the airflow across the surface. Depending on the type of the solvent used, either single layer or multilayer array of hole may be generated. When these three dimensionally ordered macroporous materials have pore dimensions comparable to the wavelength of visible light, they are of interest as photonic band gaps and optical stop-bands. We have also shown that one can fabricate picoliter beakers in a facile manner using the method of self29.7  29.7  3.7 m Figure 8. Left: Microporous film made of assembly described azide substituted poly(para-phenyleneFigure 7. Macroporous polymer ethynylene). The 3D image size is 15.2 above. Organic
3

film showing the interconnected pores with high degree of order and uniformity in size.

x15.2x 4.7 (m). Right: Same bubble arrays after heating to 300oC. The 3D image size is 10.8x10.8x2.2 (m).

6
semiconducting polymers have been used to form these picoliter beakers (Fig. 8), which will find use in analytical chemistry and analytical biochemistry [54].

5. Research Methodology
5.1. Biological material: Butterflies A. Monteiro is already in possession of specimens for each of 54 species of Bicyclus butterflies and will share wing samples with M. Srinivasarao. A small travel budget is included to allow additional specimens to be collected for the destructive SEM and TEM work. For the artificial selection work, we will use a large lab population of B. anynana butterflies being permanently reared in the Monteiro lab. 5.2. Measurements of scale morphology and correlations of scale color and scale morphology From the SEM and TEMs performed for individual scales from each species we will score a variety of parameters such as density of ridges on the surface of the scale, density of cross ridges, density of trabeculae, thickness of the cuticle at several positions, and number of cuticle layers. In order to determine whether there is a correlation between a scale’s color and its nano-morphology, we will perform a correlation analysis using transformed data sets (phylogenetic independent contrasts), where we remove the effect of shared ancestry from biasing the value of the correlation [45, 56, 57]. For these analysis we will use the single morphological parameters independently or a compound measure of a scale’s morphology (using principle components analysis), and correlate it with a scale’s color and/or reflective light spectral distribution (using the X,Y,Z color system described below). It is likely that only some aspects of a scale’s morphology will be involved in producing the observed color. In order to dissect colordependent and color-independent morphologies apart, and avoid biasing our color-morphology correlation with points from many closely related species, which are not strictly independent, we will transform our data to phylogenetic independent contrasts [56], which takes the species evolutionary relationships into consideration and produces sets of independent data points for further analysis. We have started collecting transmission electron microscopy (TEM) data for the scales of B. anynana and two other species, B. mandanes and B.sambulos (Fig. 9) that show considerable variation in ultra-structure. In particular the purple scales of B. medontias (Fig. 9B) show a considerable thinker cuticle than the brown colored scales from other species. The brown scales of B. anynana (Fig. 9A) and B. Figure 9. TEM micromedontias (Fig. 9C) also show considerable structural differences. The graphs showing cross brown color is likely produced through the presence of pigments sections through A) deposited inside the scale. The variations in ultra-structure found for Bicyclus anynana brown brown scales may reflect past instances of selection (in ancestral scales; B) B. medontias species) for particular structural colors that are no longer visible, or purple scales; C) B. current instances of selection for structural colors in flanking regions of medontias brown scales; the wing, that produce correlated changes in other parts of the wing. D) B. sambulos light Untangling what characteristics of a scale’s ultra-structure are violet scales. actually functionally producing its color, and which characteristics are due to historical and phylogenetic constraints will be an essential component of this project. By using a group of closely related species, such as members of the Bicyclus genus, that show considerable variation in scale coloration and morphology, and which have a well described evolutionary history, will enable us to dissect these two aspects apart. We will continue to collect

7
scale morphology data at larger resolution that the one displayed in Fig. 9 (using SEM and TEM) and color data for all Bicyclus species 5.3. Artificial selection on a scale’s color Using a thin fiber optics cable connected to a spectrophotometer, we will obtain spectra from a constant area in the wings of live virgin male and female specimens of Bicyclus anynana (see Fig. 10 as an example), and determine the level of color spectral variability for the colony population. After each measurement the butterfly will be marked with a unique number and kept in a separate cage with others of its sex, to prevent matings from taking place. Once all the measurements are processed, we will select a subgroup of butterflies, displaying the largest deviations from the mean color spectral values for the population, to mate with each other and provide offspring for the next generation. We will measure the color spectral value of all offspring from the next generation and compare it with the mean spectral value from the parental population. From these measurements, we will determine the heritability for scale coloration for our lab population. This measure will indicate the proportion of scale color variation in the population that is due to genetic variation and will indicate the Figure 10. Top: Spectrotrait’s potential photometer measurements of the “evolvability”, or brown scales of ten individuals of degree to which it B. anynana showing variation in will respond to wavelength at which reflection isS artificial selection. at the highest intensity (red and For a successful and white dots). Individuals with the largest wavelength measurements timely alteration of a are marked with red dots. Bottom: scale’s coloration The individuals identified in the using artificial plot above also show variation in selection, we will absolute levels of intensity of light need the heritability reflection. The light blue for this trait to be individual most closely significantly approximates the color spectrum different from zero of the purple scales of B. and relatively high mandanes (yellow curve) and (around 50%) [58]. would be marked as the most extreme individual in our selection As already experiment. mentioned above, we predict that heritabilities of this order of magnitude are likely to be present for structural colors in Bicyclus. From the measurements already performed in different individuals of the population (Fig. 10) along with repeatability measurements performed in each individual (not shown), it is clear that there is substantial amounts of color variation between individuals relative to the variation within an individual. This indicates that selection for color change is likely to succeed. We will rear large populations in each generation and will perform two replicate selection lines, each containing between 300-400 individuals/per generation. The strength of selection applied in each generation, i.e., the difference in the mean trait value for the individuals selected as parents for the next generation relative to the population mean will depend on the initial variance for the trait. We will select the most extreme 30 males and 30 females as parents in each generation. These numbers have been shown to minimize inbreeding depression and maximize selection pressure across several generations of selection for this species. Artificial selection experiments are done routinely in this species and the Monteiro lab has extensive experience in this area [30-32].

8
5.4 Measurement of the optical properties of Individual wing scales There have been a number of reports where the “color” of large areas of the butterfly of interest has been studied. Such measurements are subject to a number of inconsistencies due to the fact that the precise alignment of the scales for different measurements may be difficult and that multiple scale types, with different optical qualities, are measured together, among other things. In order to avoid such complications, one must perform the necessary optical measurements on single individual wing scales. Such studies have not been carried out on any species of butterflies, with Vukusic’s study as a single exception in the literature [24]. Therefore in order to better understand the color generating properties of a collection of wing scales, we propose to study the optical properties of individual wing scales. We will make measurements of absolute reflectivity and transmission of various butterfly wing scales as well as angle dependent reflectivity measurements. To a large extent, optical modeling of the wing scales is hindered by not having accurate values of the refractive index of the materials that the wing scales are made of. Hence we will make measurements of refractive index of wings scales of all the butterfly wings we intend to study. To this end, we will build a dedicated instrument that will measure the angle dependent reflectivity of the wing scales. The optical arrangement of the instrument will enable us to position an individual wing scale at the center of a motorized rotational stage that coincides with the path of the light source with a detection system mounted on a rotating arm so as to allow us to make measurements of the absolute intensity both in reflection as a function of angle as well as the transmitted intensity (intensity at zero angle). The detection arm will allow for a scan from 0 to 2 around a horizontal plane about the centrally mounted wing scale. 4 An example of reflectance spectrum of a butterfly wing at a 0 fixed angle is shown in Fig. 11. 2 While data on individual wing scales can be collected 0 over a large enough angles, the above instrument will also 0 allow us to obtain data in the transmission mode, thus 4 5 6 7 8 allowing us to interrogate the scattering as a function of 0 0 0 0 Wavelength (nm) 0 angle as well, from zero degrees to about 90 degrees or so. 0 0 0 0 0 Such data provides us with information on the structure Figure 11. Reflectance spectrum of factor of the object doing the scattering. We will perform a wing scale of a Morpho butterfly such measurements on both the cover and ground scales of From ref. 3. all the butterfly species we will investigate. Measurements such as the angle dependent scattering and angle dependent reflection will be done as a function of wavelength of the incident beam as well with a variety of wavelengths available in Srinivasarao’s lab with the use of an Ar ion laser (456, 476, 488, 514nm), a Kr ion laser (568nm) and several He-Ne laser (541nm and 633nm) sources. We will use a Na lamp and a xenon arc lamp to obtain several other wavelengths covering the entire visible region of the electromagnetic spectrum. We will also use a micro-spectrophotometer to make measurements of the reflectivity of the wing scales without having to remove the scales from the wings of the butterfly. The microspectrophotometer is basically a microscope with a spectrometer head with a spectral resolution of 1nm. The illumination spot size can be adjusted to be in the range of 5-25 m, thus ensuring that data is collected from a single wing scale. Spot sizes smaller than 5 m can be achieved by using objectives that have higher numerical apertures. We will add a tilt stage on the rotational stage of the microscope, thus enabling us to make angle dependent measurements of reflectivity of the chosen wing scales. 5.5. Modeling the Optical Properties of an Individual Wing Scale
% Reflectance

9
5.5.1.Specification of color of the wings The modeling of the colors produced by the wings of

butterflies will take three different forms. One is to use the language of color science to convert the set of complex data of the reflectivities to a simple point on the CIE (Commission International de l’Eclairage (International Commission on Illumination) xy coordinate system, and the second is to use the structural details obtained from various techniques (TEM, SEM, and optical) to model the optical properties of the wing scale – in other words to be able to predict the color based on the well known optics [59]. As a third avenue for modeling the optical properties, we will use the language of photonic crystals to compute the band gaps, fully recognizing that the band gaps may only be pseudogaps due to the weak refractive index contrast of these systems, since the index contrast arises primarily from the difference in indices between the cuticle and air. The mathematical description of 2 color by chromaticity diagram relies on three parameters: an object, an illuminant, and a detector. An object 1.5 can be described by its Spectral Reflectance or Spectral Transmission 1 Curve, which maps the reflectance or transmission of light that interacts with the object at various wavelengths. An 0.5 illuminant is described by its Spectral Power Distribution, which denotes the relative power of the illuminant 0 through its gamut of wavelengths. 350 400 450 500 550 600 650 700 750 Again, the range of wavelengths Wavelength (nm) between 400 and 700 nm are of Figure 12. Color matching functions of 1931 CIE standard particular interest for the purposes of 2o observer. The filled circles, open circles, and filled color science. Finally, the sensation of squares represent the color sensitivity for red ( x ), green color requires a detector, such as the ( y ), and blue ( z ), respectively. From ref. 3. human eye. The CIE decided on a standard set of Color Matching Functions in 1931 so that color could be mathematically described [60, 61]. This type of function can be generated by using a split field with one half illuminated by a single wavelength of light. The test for the observers is to match the other half of the field to the standard by adjusting the relative intensities of red, green, and blue light. In 1931, a large group of observers with normal color vision were asked to match every wavelength of visible light at a 2 viewing angle (so as to exclude any rod vision participation), and the CIE 1931 2 Standard Observer was born [60, 61]. The Observer’s perception of red, green, and blue is denoted by x , y , and z , respectively. These Color Matching Functions are displayed graphically in Fig. 12 [62]. By summing the reflectivity curves, R(), power distribution, P(), and color matching functions for a given object and light source over Figure 13. CIE 1931 2 Observer chromaticity the visible spectrum of light, three so-called
Spectral tristimulus values

diagram.

10
Tristimulus Values X, Y, and Z are obtained [62]: X = k  R( ) P() x () d Y = k  R() P() y () d Z = k  R() P() z () d (1) (2) (3)

Here k is a constant determined by normalizing the relationships such that Y = 100 for a perfect white reflector which reflects 100% of all incident light. After normalization, CIE chromaticity values x, y, and z are defined as : x = X / (X + Y + Z) (4) y = Y / (X + Y + Z) (5) z = Z / (X + Y + Z) (6) Since the sum of the three chromaticity coordinates is obviously equal to unity, only two of the coordinates are needed to describe a color. By convention, x and y are used, and these two values decide an object’s color point on the CIE 1931 2 Observer chromaticity diagram (Fig. 13). These two chromaticity coordinates aptly describe two of the three components of color, namely hue and saturation. Once the reflectivities of the desired wing scales are measured, it is relatively simple to convert the reflectivities to their respective tristimulus values (X, Y, Z), from which the chromaticity coordinates can be calculated. We will make reflectivity measurements of all the wing scales including those of individual wing scales and map the reflectivity on to the chromaticity diagram. We fully recognize that the color maps we will obtain are based on the color perception of human observers. Hence, we will make every attempt to understand the “color perception” of butterflies by looking at the visual photo-pigments of the butterfly species involved and attempt to construct a diagram similar to the x, y chromaticity diagram shown in Fig 13. The data necessary to do this will come from the literature where the spectral characteristics of the photo-pigments have been garnered using “eyeshine spectroscopy”. Once the tristimulus (X, Y, Z) values are known one can convert these back to RGB values for the purpose of imaging and displaying the colors.
5.5.2.Modeling of Interference and diffraction

Interference colors are observed in thin films with a periodic variation in refractive index. Such colors are produced by light waves interfering after reflections from the two or more surfaces of these films, like soap films seen in sunlight [63], butterfly wings [2-6, 12], or beetles [3, 64]. Here we are studying colors of butterfly wing scales that are a few microns thick, we can use the theory of interference developed for coherent sources. The light source under consideration might be incoherent in comparison to a laser light source, however, the coherence length of these sources are on the order of a few microns. One sees interference colors when the film thickness is on the order of the wavelength of visible light. Therefore, in order to manipulate colors, one changes the film thickness or the viewing angle. The theoretical foundation for the interference phenomena in thin film structures is provided by the Fresnel equations [59, 63]. Constructive interference occurs when Fresnel reflection coefficient is positive for a light beam reflected from the liquid-air interface and is negative for reflection from the air-liquid interface. A negative reflection coefficient can simply be thought of as a beam undergoing a 180o phase shift between the incident and reflected light waves. Therefore the 180 o phase shift suffered at the air-liquid interface together with the 180 o phase shift suffered in a round-trip through the quarter-wave plate (the thickness of the film is  /4, where  is the wavelength of the light beam) leads to perfect constructive interference for all the reflected waves. The optical path difference experienced by the two rays that interfere constructively is simply equal to the extra distance the light beams have to travel in the medium.

11
When a light beam is incident (at an angle 1) on a thin film of liquid of thickness d, the path difference turns out to be 2n2dcos2, where n2 is the refractive index of the film and 2 is the angle of refraction. In addition to this optical path difference,there will be an additional path difference of ½ due to the additional phase difference  that occurs at the air-film interface whenever an incident light beam is reflected by a medium of higher refractive index than the initial medium. Thus the effective path difference between the two rays is 2n2dcos2 + ½ . Consequently if 2n2dcos2 + ½ = m , where m is an integer, the two rays will interfere constructively and give an intensity maximum. On the other hand, if 2n2dcos2 + ½ = (m + ½) , one has destructive interference resulting in zero intensity. Since by implication we have assumed the amplitudes, A, of the two beams to be equal, the resulting amplitude of the reflected wave will be given by, Ar, where Ar = A + Ae i with the phase difference, , given by  = (2/ ) (2n2dcos2 + ½ ). The total reflected intensity is Ir = Ar Ar*, and is equal to 4A2cos2(½). This equation can be 1 rewritten in terms of the reflectivity, R, to have the following form: 4IiRsin2((2 / )n1dcos2), where Ii is the incident intensity. One can of 0.8 course rather easily eliminate the angle of refraction from the above formula to show the dependence of intensity on the incident angle. 0.6 600 An attempt to quantify the above discussion 800 y 400 leads us to the task of deriving expressions for the reflected and transmitted light intensity. 0.4 Fresnel equations predict the amplitude of the 750 reflected light from thin film structures and can 440 300 be written as [65]:
n cos 1  n2 cos  2  r s   1 ; n1 cos  1  n2 cos  2 
  2n1 cos  1 t s   250  n1 cos  1  n2 cos  2 
0 0.2
500

where n1, n2 are the refractive indices of the two 0 0.2 0.4 0.6 0.8 media in which light propagates, rs and ts are the x amplitudes of reflection and transmission for SFigure 14. Chromaticity diagram calculated polarization of the incoming light beam. Here S- for the interference colors. The numbers are polarization implies the polarization is the path differences used to calculate the color perpendicular to the plane of incidence, where seen in reflection. From ref. 60. the plane of incidence is defined as the plane which contains both the incident and reflected light beams. Similar expressions can be derived for P-polarization. In order to obtain the angular dependence, one can easily eliminate the angle of refraction and write the optical path difference in terms of the incident angle. The maximal reflectance can then be easily shown to be given by:

 max  

2   4n2 d  n1 2 1 2 sin  1    n2 2n  1 

where n is an average refractive index This expression clearly shows that the reflectance is shifted to shorter wavelengths with increasing angle of incidence, consistent with all the interference colors, including those due to butterfly wings. In our preliminary study, we calculated the chromaticity diagram for various interference colors, and Fig. 14 is an example of such a color seen at a fixed angle [66].

12
We will use the structural data obtained from the SEM micrographs to compute the Fresnel coefficients and the  max for a variety of butterfly wings and compare them to the experimentally measured values to get an understanding of the color production. We will also convert the reflectance data into chromaticity coordinates so that a realistic representation of the color properties of the wings can be made as discussed in section 5.5.1.
5.5.3.Fourier modeling of the wing scale structures As already pointed out, the colors on the

wings of butterflies arises from a complex combination of interactions of light with the structures on the wing scales, coupled with absorption of light that is not reflected to the observer by the pigments that exist on the wings. Whenever the colors are structural and angle independent, such colors have often been attributed to incoherent scattering mechanisms like Rayleigh scattering. However, when the structures are periodic in nature, it is worthwhile to consider that the colors produced might be due to coherent scattering. Coherent scattering occurs when the elements doing the scattering of light are not random, in other words, they are either highly periodic or quasi-periodic. Consequently, when structures are periodic one can use models that describe coherent scattering for the production of color. Unlike in incoherent scattering, the phases of the two scattered beams are the same in coherent scattering. Since coherent scattering occurs only when there is spatial periodicity in the refractive index, we will analyze all the wing scale structures and attempt to classify the nature of the periodic structures found on the wing-scales. By analyzing the real-space images obtained by various microscopy techniques, we will characterize the extent of order on the individual wing-scales as well as over large enough areas that pertain to real viewing conditions that produce the color that one sees on the wings of butterflies. We will also obtain a Fourier transform of the real-space images of the wings and wing-scales, which will provide us the q-space image. This will of course be compared with the scattering measurements from individual wing scales as was discussed in the section on optical measurements, thus providing a direct comparison of the measurement structure factor to those computed by taking the Fourier transform of the real-space images. In order to predict the color that a given nanostructure might produce, we will use the 2-D Fourier tool developed by Prof. Richard Prum at Yale University [67-69]. Such predictions can then be compared with the measurements that we would make of the color of the individual wing scale as well as the entire wing of a given butterfly.
5.5.4 Photonic properties of the wings based on Photonic crystals: Photonic crystals are materials that

possess periodic dielectric structures with different refractive indices that forbid propagation of 1-D 2-D 3-D electromagnetic waves in any particular directions and for any particular polarization, in a certain frequency Figure 15. Schematic illustration of onerange, the so-called photonic band gap (PBG). Such a dimensional (1-D), two-dimensional (2-D), material is schematically shown in Fig. 15 where two and three-dimensional (3-D) photonic materials denoted by A and B are stacked alternatively. crystals. Such photonic crystals can be classified as onedimensional (1-D), two-dimensional (2-D) and three-dimensional (3-D) based on how the materials are stacked (Fig. 15). The optical characteristics of such photonic band gap structures depend on the dielectric contrast between the structures, the symmetry of the structure, and the filling factor, i.e. the ratio of the volumes occupied by each dielectric with respect to the total volume of the photonic crystal structure. So far, complete PBGs are found for relatively high dielectric constant materials. Should the modulation of the contrast be weak, as in the case of the wings of butterflies, pseudogaps can occur. The propagation of light inside the structure can be strongly affected by the characteristics of the band structure regardless of the completeness of the band gaps [7-11]. Hence, these periodic structures can then exhibit unusual optical phenomena, like the phenomenon of negative refraction [70-74].

13
Many butterflies that are characterized as exhibiting structural colors have structures that are quite periodic in nature, thus mimicking or behaving as photonic crystals. Some examples of such wonderfully periodic structures in butterfly wings are shown in Fig. 16 [75]. Hence, knowing the structure and the refractive indices of the periodic material, one can calculate the band structure of such a system. The photonic band diagram indicates all possible electromagnetic scattering interactions in a periodic system. Hence, we will calculate such photonic band diagrams for the structures that we encounter in the range of butterflies that we study. We will also look for instances where the band structure influences the fluorescent emission of pigments found on the wing scales of many species of butterflies [76]. It has been recently demonstrated that the blue-green coloration found on the wings of Swallowtail (Papilio) butterflies has directional emission of fluorescence whose direction has been modified by a photonic crystal slab (PCS) into which the fluorescent pigments have been infused [76]. This is quite analogous to what has been demonstrated for organic light emitting diodes (OLEDs) [77, 78]. In order to increase the efficacy of coupling the light out of an organic light emitting diode (OLED), photonic crystals in conjunction with distributed Bragg Reflectors (DBRs, which are 1-D photonic crystals - see Fig. 15) have been used. It should come as no surprise that many of the design elements that the butterflies use are technologically Figure 16. SEM of a fractured green scale from Parides sesostris (left) and cracked bristle of Tomares ballus (right). relevant [78]. Scale bar is 1m. From ref. 68. On the one hand, we will study the photophysics of the natural pigments that have already been infused into the wing scales of butterflies. On the other hand, we will also perform experiments where we infuse organic molecules (including polymeric ones) that are widely studied for their optical, and semiconducting properties, which have found use in a variety of devices ranging from light emitting diodes to solar cells [79-82]. These materials will be dissolved in a suitable solvent and infused into the individual wing scales of a variety of butterflies that exhibit color due to the nano-structures on the wing scales. We will study the emission characteristics which include the fluorescence spectrum, lifetime and fluorescence polarization, of such molecules infused into the nanostructures. Such a systematic study should provide us with valuable data to understand how the photophysical properties (fluorescence emission characteristics and lifetimes) change due to the fact that the molecules are trapped in a nano-structured environment.

6. Work plan
6.1. Monteiro Lab: Year 1: Start artificial selection experiment on two replicate lines of B. anynana butterflies. After first generation, determine heritability for structural scale colors and estimate the number of generations necessary to produce a significant change in scale color. Continually select butterflies until end of year, possibly middle of year 2. In between selection bouts, collect high-resolution SEM and TEM data for all the different wing colors in all the collected Bicyclus species. Send samples of all the Bicyclus species to Srinivasarao lab for light reflectivity analysis. Year 2: Continue collecting SEM and TEM data for different Bicyclus species as well as for B. anynana material from different stages of the selection experiments. Send samples from selection material to Srinivasarao lab for light measurements and modeling.

14
Year 3: Analyze the correlation between structure and color data between Bicyclus species, and between the selection lines. Understand which nano-morphologies are primitive and which are derived, and the extent to which particular morphologies have evolved independently. Help analyze data for all the parameters collected by the Srinivasarao lab. Write papers. 6.2. Srinivasarao Lab: Year 1: Build a dedicated instrument that will measure the angle dependent reflectivity of the wing scales. Take microspectrophotometric data on wing scale specimens sent by Monteiro, in both reflective and transmission mode, as a function of angle and wavelength. Obtain optical images of the wings and scales using polarized optical microscopy and confocal laser scanning microscopy. Perform these with different incoming polarization to determine the polarization dependent characteristics of the butterfly wing scales. Year 2: Continue the optical and photophysical measurements for all the available wing scales. Perform optical modeling calculation and construct chromaticity charts for all the wing scales for which the optical properties were measured in year 1. Study how the nanomorphologies described in Year 1 correlate with the wing scale colors. Year 3: Simulate interference/diffraction patterns of the wing scales. Help correlate the morphological/structural characteristics with the color/optical/photophysical properties in a phylogenetic context.

7. Anticipated Results
7.1. Artificial selection: We expect the selection experiment to result in a visible change in the hue of the colors of the wings of Bicyclus anynana within 7-9 generations of selection (1 year). Unpredictable results will be the extent to which all the scales in the same wing surface, or other wings (hindwings) or surfaces (ventral surface) will show correlated changes. It is possible that the alterations in hue will be localized just to the region that was placed under selection as was found for vein pattern selection in Drosophila [83] and eyespot shape in Bicyclus [84]. It is difficult to predict whether the replicate lines will evolve similar colors via the same morphology, or alternative morphologies, as this type of artificial selection has never been done before. 7.2. Evolution of color with the genus Bicyclus: Through our survey of scale color and TEM, SEM measurements for the Bicyclus butterflies, we will determine to what extent the similarly brown, white, blue, and purple colors shared by several members of the genus have similar nano-morphologies. We predict that there will be several different nano-morphological solutions to the same “color” but at this stage we cannot predict whether some morphologies will more readily evolve into a diverse array or colors relative to other morphologies. This will be a novel finding for this study. We will also learn which components of a scale’s ultra-structure are actually likely responsible for the structural color observed, and which components are likely redundant (relative to color). 7.3. Modeling of optical properties and target-oriented study for potential applications: For the first time we will obtain systematic optical data on an individual wing scale that will allow us to understand the relationship between the subwavelength structures and the observed coloration/pattern on the wings of butterflies. Such data will allow for a comprehensive modeling of the optical properties of the wings from the point of view of “Color Science” and “Photonic Crystals”.

8. Broader impacts of the proposed project
Education and Outreach Plan.

15
Monteiro and Srinivasarao are currently building a mutually enriching collaboration between a Materials and a Biology department, capitalizing on their shared interest in butterfly scales and providing students with a cross-disciplinary training environment As one of only three women faculty in the Department of Ecology and Evolutionary Biology at Yale University, A. Monteiro plays an important role in serving to attract, retain, and mentor graduate and undergraduate women in science. Since 2003 A. Monteiro has recruited three graduate women, and ten undergraduate women. She also has a strong tradition of including undergraduate students on small, self-contained but valuable projects that will contribute to the understanding of Bicyclus anynana biology, and thus leverage progress among the larger projects in the lab. She has systematically pursued REU supplements to her grants, and has been awarded 3 of these fellowships in the last three years that have helped recruit over nineteen undergraduates to work in her lab. Several of these students have gone on to pursue research careers in graduate school. As an outreach component, the work in the Monteiro lab will be featured as part of a public exhibit about butterflies at the Natural History Museum of Lisbon, Portugal, that will later travel through Europe (http://www.borboletasatravesdotempo.com). The work will also be featured in an upcoming television series about a “new generation of scientists”, produced by the Portuguese television channel RTP2. In addition to providing education and training for graduate students and undergraduate students, Srinivasarao has been quite active in outreach activities some of which are listed below. In the past four years we have organized 10 high school science teacher workshops and all have been quite successful. We also go to local schools to lecture at least twice during a semester. High school Science Teachers Day Workshop dealing with Polymer Science and Optics At least four workshops have been organized in various cities in conjunction with National ACS meetings during the period 1997-2000. The workshops were aimed at addressing some of the missing links based on "Project 2061: Bench Marks in Science Literacy", published by The American Association of Advancement in Science. Project 2061 describes, in great detail, what students should know at every level from K-12. It is rather interesting and quite puzzling to find that the words polymer, macromolecule, biopolymer and plastic do not exist in the index of Project 2061. The teachers have always liked the workshops and wanted to participate in future workshops. This activity was funded by Lucent Technologies, Polaroid Corporation, NSF and Exxon, among others. In 2001, we switched to having workshops at Georgia Tech. There have been four workshops in the last year at Georgia Tech for local teachers with support from Exxon, NSF and Georgia Tech. We will continue this effort during the years to come. Also, in conjunction with NASA, CEISMC has a high school science teachers’ day in which we actively participate every year and where we discuss issue related to solar power. We plan to have more workshops aimed at addressing the need for education of high school science teachers in the area of liquid crystals, optics, color science, polymer science, and energy conversion from the sun. The workshops will be conducted so that the teachers can use what is done in the workshop in their classes. At a later stage, this effort will be carried over to other professional societies that we are members of. We will strive to have at least one lecture for the general public on energy – this would be done with the blessing of the upper administration at Georgia Tech. Collaboration with CEISMC: Research Experiences for Teachers The Georgia Industrial Fellowships for Teachers (GIFT) program places approximately 85 science, mathematics, and social studies teachers per summer in 8-week business, industry or laboratory research experiences to provide them with a greater understanding of the use of science and mathematics in the real world. Approximately 20-30 of these teachers per year spend their summer participating in research internships at Georgia Tech and Emory. Some GIFT teachers are also hired to assist faculty in developing K-12 educational materials. CEISMC recruits the teachers, and the participating faculty member provides a stipend plus $1,250 program fee (which we have written into this proposal).


								
To top