当前位置:首页 >> 工程科技 >>

Using quantum dots to tag subsurface damage in lapped and polished glass samples


Using quantum dots to tag subsurface damage in lapped and polished glass samples
Wesley B. Williams,1 Brigid A. Mullany,1,* Wesley C. Parker,2 Patrick J. Moyer,2 and Mark H. Randles3
1

Department of Mechanical Engineering and Engineering Science, University of North Carolina at Charlotte, 9201 University City Boulevard, Charlotte, North Carolina 28226, USA
2

Department of Physics and Optical Science, University of North Carolina at Charlotte, 9201 University City Boulevard, Charlotte, North Carolina 28226, USA

3

Northrop Grumman Synoptics, 1201 Continental Boulevard, Charlotte, North Carolina 28273, USA *Corresponding author: bamullan@uncc.edu Received 1 July 2009; accepted 13 August 2009; posted 25 August 2009 (Doc. ID 113634); published 11 September 2009

Grinding, lapping, and polishing are finishing processes used to achieve critical surface parameters in a variety of precision optical and electronic components. As these processes remove material from the surface through mechanical and chemical interactions, they may induce a damaged layer of cracks, voids, and stressed material below the surface. This subsurface damage (SSD) can degrade the performance of a final product by creating optical aberrations due to diffraction, premature failure in oscillating components, and a reduction in the laser induced damage threshold of high energy optics. As these defects lie beneath the surface, they are difficult to detect, and while many methods are available to detect SSD, they can have notable limitations regarding sample size and type, preparation time, or can be destructive in nature. The authors tested a nondestructive method for assessing SSD that consisted of tagging the abrasive slurries used in lapping and polishing with quantum dots (nano-sized fluorescent particles). Subsequent detection of fluorescence on the processed surface is hypothesized to indicate SSD. Quantum dots that were introduced to glass surfaces during the lapping process were retained through subsequent polishing and cleaning processes. The quantum dots were successfully imaged by both wide field and confocal fluorescence microscopy techniques. The detected fluorescence highlighted features that were not observable with optical or interferometric microscopy. Atomic force microscopy and additional confocal microscope analysis indicate that the dots are firmly embedded in the surface but do not appear to travel deep into fractures beneath the surface. Etching of the samples exhibiting fluorescence confirmed that SSD existed. SSD-free samples exposed to quantum dots did not retain the dots in their surfaces, even when polished in the presence of quantum dots. ? 2009 Optical Society of America OCIS codes: 220.5450, 160.2750.

1. Introduction

Subsurface damage (SSD) is a layer of defects and stressed material that exists beneath an apparently smooth surface. Investigations by Lawrence Livermore National Laboratory (LLNL) [1,2] proposed that SSD consists of a thin polished layer superim-

0003-6935/09/275155-09$15.00/0 ? 2009 Optical Society of America

posed on a fractured damaged layer followed by a highly stressed layer. These layers can account for a depth of up to 200 μm beneath the surface. SSD is typically induced by the more aggressive material removal processes such as grinding and lapping, i.e., a process that relies on brittle fracture to remove material. Once present, it can be removed by slower, less cost efficient polishing processes. The extent of the damage occurring in typical optical glasses is known to depend upon the aggressiveness of the previous
20 September 2009 / Vol. 48, No. 27 / APPLIED OPTICS 5155

grinding and lapping process, i.e., the size of abrasive being used, and the mechanical properties of the glass being polished. Preston and many researchers since have shown correlations between the resulting surface roughness and the extent of the depth of damage; for example, Lambropoulos et al. [3] found that the peak to valley surface roughness measured with a white light interferometer was a good upper bound for the average depth of subsurface damage in lapped BK7 and LHG8 glass. Theories supporting these relationships focus on microindentation models comprising the abrasive and workpiece surface. The abrasive under sufficiently high loading acts as an indenter and is capable of inducing both radial and lateral cracking in the material. Lateral cracking is considered essential for material removal, while the radial cracking is thought to form the damaged subsurface layer [4]. Cook [5] considered the abrasives as traveling indenters that induce compressive and tensile stress fields at their leading and trailing edges (Fig. 1). He thought the localized compressive stress fields induced brittle fractures and that these factures are further aggravated by the trailing tensile stress fields. Bowden [5] and Adams [6] hypothesized that these fractures may be covered over by mechanisms such as material flow [6] or redeposition [2,7]. Based on Griffith’s theory of brittle fracture occurring at dominant flaws in a material [8], it is possible that these fractures may never be open to the surface. Correct processing of optical and laser crystals should ensure that little to no damage remains after the final polishing step. Manufacturers need to be confident that their grinding, lapping, and polishing fulfill this requirement. To achieve this, they must know how much the damage is introduced into the samples, and which process step induces the damage. Failure to remove the SSD can negatively affect the mechanical integrity of the final surface. For example, a surface containing SSD can be highly degraded during common coating processing steps where the workpiece is subjected to high temperatures and pressures that promote defect propagation to the surface, resulting in “new” pits and cracks [9]. In laser applications, defects or highly stressed layers of material within the laser crystals will lower the laser

induced damage threshold (LIDT) limit, resulting in shorter life spans or complete failure [2]. One of the main reasons why SSD is difficult to efficiently eliminate is because it is difficult to detect and quantify. Papers by Lucca et al. [10], Brinksmeier [11], and Shen et al. [9] provide details on SSD detection methods. They include X-ray techniques (diffraction, grazing, fluorescence), ultrasonic, micromagnetic, acoustic, laser modulated scattering, and Raman. An interested reader should refer to these papers for a comprehensive understanding of the different methods. In the optical fabrication industry SSD is often detected and quantified by dimpling or taper polishing techniques [1,12]. Both these techniques are destructive in nature, as they involve polishing away the top layers to reveal the undamaged bulk material. In the case of dimpling, a small sphere (diameter 25 mm) is used to polish a dimple into the workpiece surface. The resulting dimple is examined under the microscope, and the diameter of the circle containing no visible is damage is determined. Simple geometrical analysis combining knowledge of the polishing sphere geometry and the damage free circle size allows the depth of damage extending beneath the surface to be calculated. The smaller the damage free circle, the deeper the SSD. While simple in concept, this test is quite labor intensive [3], destructive in nature, and subject to the lateral resolution limitations of optical microscopy. Additional nondestructive, workshop floor friendly methods of detecting SSD in optical materials are still sought. The work presented here investigates if quantum dots (nanometer scale semiconductor crystals that fluoresce at a given wavelength [13]) can provide a means to quickly detect SSD. The method consists of tagging the abrasive slurries used at different stages of the manufacturing process with quantum dots. These dots will be present for all the dynamic events that may occur during polishing, e.g., cracks opening up to the surface. Their small size (diameters of 3:2–5:8 nm [13]) should allow them to travel into sample defects if they are open to the surface (Fig. 1). After polishing, the samples will be examined for fluorescence. A confocal microscope that scans areas on and beneath the surface will detect fluorescence from any remaining dots and provide information regarding their location. In addition to SSD detection, the technique has the potential to provide new insights into how material is removed during lapping and polishing processes. In this paper we report the findings of the proposed technique.
2. Quantum Dots and Their Detection

Fig. 1. Illustration of hypothesized interaction of quantum dots with fractures generated by lapping and polishing dynamics. 5156 APPLIED OPTICS / Vol. 48, No. 27 / 20 September 2009

The capability of quantum dots, nanosized (<10 nm) semiconducting crystals, to fluoresce at specified wavelengths was reported in the mid 1990s [14,15]. In this work, cadmium selenide quantum dots are used, but other material combinations such as lead sulfide, cadmium sulfide, indium arsenide, and indium phosphide can produce quantum dots. Regardless of the material type, they all have the following fluorescing

characteristics. When the electrons in their outer shells of the dots are excited, they can jump across the bandgap and reside in the higher conduction band. When the electron drops back to its original energy state, the energy difference between the two states is emitted as electromagnetic radiation. The emission wavelength depends on the actual bandgap distance, i.e., the excition Bohrs radius (EBR). Larger bandgap distances result in shorter wavelength emissions (higher energy). In the case of quantum dots, the actual sizes of the dots are comparable to the EBR; therefore any change in the size or structure of the quantum dots will affect the magnitude of the EBR and consequently the emission wavelength. The practical implication of this is, if the size and structure of the quantum dots can be tightly controlled, then the bandwidth of the emitted wavelength and the bandwidth of the incident wavelength required to excite electrons can be both finely tuned and distinct from each other. Quantum dots offer several advantages over cheaper organic fluorescent dyes, such as higher intensity fluorescence, narrower emission spectra, broad excitation spectra, and increased resistance to photobleaching (diminishing of the fluorescent response in successive excitations) [16]. The quantum dots utilized in all experiments detailed in this paper were “Hops Yellow” EviDots from Evident Technologies. These particles consist of a cadmium selenide core surrounded by a zinc sulfide shell, resulting in an estimated crystal diameter of 3:8 nm. This shell in turn is surrounded by a 2 nm thick layer of ligands giving a final particle diameter of 7:8 nm. The ligands allow the dots to remain in colloid suspension. The quantum dots have an emission peak at 553 nm ? 10 nm and are excited by wavelengths shorter than 540 nm. The quantum dots are shipped in a toluene solution.
A. Quantum Dots: Workpiece Theoretical Interactions

The equation governing the van der Waals force between a particle and surface is given by F VDW ? AD 12r2 ?N?; ?1?

where A is the Hamaker constant (Nm), D is the particle diameter (m), and r is the distance separating the particle and the surface (m). The Hamaker constant is a measure of the strength of the van der Waals forces in a system and depends on the geometry and composition of the interacting material pairs. Equation (1) does not take into account any deformation of the particle or the surface when the two bodies come in contact. As the deformation will increase the contact area, and thus the adhesive van der Waals forces, it should be factored into Eq. (1). The Derjaguin–Mueller–Toporov (DMT) theory [19], which was developed to calculate the elastic deformation of small hard particles, is used. Tabor based analysis [20,21] of DMT theory and the Johnson– Kendall–Roberts (JKR) theory [22], which focuses more on larger compliant diameter particles, confirms the suitability of DMT theory for this application. The contact diameter, a0 , of the particle on the surface under zero loading is given by a0 ? 1  πΔγD2 3 2K ?m?; ?2?

where Δγ?J=m2 ? is the work of adhesion (a measure of the force required to separate two materials [23]), and K is the composite Young’s modulus (Pa). The composite modulus, k, incorporates the elastic modulus, E, and Poisson’s ratio, ν, values of both the particle (E1 , ν1 ) and the surface (E2 , ν2 ): K?   4 1 ? ν2 1 ? ν2 1 2 ? E1 E2 3 ?Pa?: ?3?

In addition to quantum dots being trapped in defects, two other possible interactions with the workpiece surface are evaluated: (1) attractive forces between the quantum dots and the workpiece are so high that they are effectively absorbed into the workpiece and they remain on the surface even after cleaning, and (2) the hydrodynamic forces of the slurry flowing over the workpiece surface removes the quantum dots from the immediate surface. This would prevent the quantum dots from interacting with surface damage sites. To assess the possibility of these interactions the magnitude of the forces attracting and holding the quantum dots on the surface must be first estimated and then considered with respect to workpiece surface energies (absorption) and slurry hydrodynamic forces. In the case of small particles (diameters <50 μm) electrostatic van der Waals forces are the primary acting forces [17]. Other electrostatic forces (Casimir and electric double poles) are considered negligible, as are gravitational and interatomic forces [18].

As per Visser [24], Eq. (1) is modified to include the additional contact area, Eq. (4) [21]. The equilibrium spacing, ε, is the point of zero potential where the van der Waals forces shift from being attractive to repulsive due to the overlap of electron orbitals: F VDW ?   2a2 AD 1? 0 εD 12r2 ?N?: ?4?

By taking parameter values from literature best suited to the materials used in the experimental work (see Table 1) the values of a0 and F VDW are calculated to be of the order of 0:58 nm and 0:5 nN, respectively. According to work done by Zhang and Busnaina [25], in order for small particles to be removed from a surface by the action of a fluid flowing over the surface, the moment applied by the hydrodynamic force at 70% of the particle diameter must be greater than the moment created by the van der Waals force applied at a distance equal to the contact radius. The
20 September 2009 / Vol. 48, No. 27 / APPLIED OPTICS 5157

Table 1.

Values of Parameters Used in Eqs. (2) to (9) and Their Sources

Symbol A D Dabrasive Δγ E1 E2 ν1 ν2 E ρ Ω R

Name Hamaker constant Diameter of the quantum dot Diameter of the abrasive Work of adhesion Modulus of elasticity (ZnS) Modulus of elasticity (Glass) Poison ratio (ZnS) Poison ratio (glass) Equilibrium spacing Density of slurry Rotational speed of the platen Radius of the platen

Value 1 × 10?20 J 7:8 nm 450 nm 0:6 J=m2 74:5 GPa 75 GPa 0.28 0.2 0:4 nm 1:02 g=cc 20 rpm 300 mm

Ref. [33] [16] [34] [35] [31] [35] [31] [21]

hydrodynamic force, F D , can be described by[25] F D ? CD ρ v2 A 2 ?N?; ?5?

where CD is the coefficient of drag, ρ is the density of the slurry, v is the velocity, and A is the crosssectional area of the particle. For low Reynolds number flows [Eq. (6)], the coefficient of drag is given by Eq. (7): Re ? ρvD ; μ 24 ; Re ?6?

CD ?

?7?

where μ ?Pa:s? is the viscosity of the fluid, v ?m=s? is the mean velocity of the fluid, and D ?m? is still the particle diameter. The same approach as taken by Visser [24] and Busnaina et al. [26] to estimate the mean fluid velocity in chemical mechanical planarization (CMP) is taken here. The approach assumes that the peak fluid velocity occurs at a distance equal to the slurry abrasive particle diameter (not the quantum dot) from the workpiece surface. The velocity acting on the quantum dots will be a small fraction of the peak velocity, as the ratio between the quantum dot and the abrasive is 1∶56:25. As an extreme case, the velocity (vmax ) was taken to be the rotational speed of the platen (ω) multiplied by the radius of the platen (R): vmax ? ωR ?m=s?: ?8?

4:3 × 10?13 N. Table 1 details the values of the other parameters used in the calculations. The moment induced by the hydrodynamic force at 70% of the quantum dot diameter, M hydro , is 3:3× 10?12 N nm. The moment applied by F vdW at the contact radius, a0 , is 1:0 × 10?10 N nm. As M vdW > M hydro , it is not expected that the quantum dots will be “flushed” off the surface by the polishing slurry. Therefore the quantum dots can remain in the vicinity to interact with the workpiece surface. The other interaction to be considered is the absorption or diffusion of the quantum dots into the workpiece surface in the absence of any polishing action. When diffusion takes place, it is a result of diffusing atoms moving into the interstitial space between the existing atomic lattice [27] or the random movement of diffusing atoms jumping into voids in the existing lattice [27]. Both scenarios are aided by higher temperatures, where the lattice spacing is greater and there is more energy available to facilitate movement. In the case of interstitial movement, the diffusing atom needs to be significantly smaller than atoms in the matrix to be able to “squeeze in between” as is the case for carbon diffusing into iron. In the case of “Hops Yellow” quantum dots the overall dot diameter is approximately 7:8 nm. This makes it impossible for the quantum dots to move through the interstitial spaces, which are at least an order of magnitude smaller (angstroms instead of nanometers). By the same logic it would take much more than a single void in the crystal structure to accommodate a quantum dot. These considerations, in addition to the interactions taking place at room temperature, make it unlikely that there would be any significant diffusion of quantum dots into the glass sample.
B. Quantum Dot Detection

The mean velocity acting on the quantum dots is then determined by v ? vmax D Dabrasive ?m=s?: ?9?

1.

Confocal Fluorescence Microscopy

Values of 0.31 and 0:0055 m=s were determined for vmax and v, respectively. Substituting this value of v into Eq. (5) gives a hydrodynamic force of
5158 APPLIED OPTICS / Vol. 48, No. 27 / 20 September 2009

The confocal microscope used was developed at the University of North Carolina at Charlotte [28]. Confocal microscopy, pioneered by Minsky, consists of a setup where light from outside the focal plane is largely prevented from reaching the light detector

[29]. This rejection of light outside the focal plane, coupled with the exclusion of light from points adjacent to the focal point, reduces haze and increases the sharpness of the image [30]. The lateral resolution improvements over conventional microscopes and the capability to focus beneath a sample surface make it ideal for this body of work. The system used in this work excites at 470 nm. The excitation is provided by a PicoQuant PDL 800 diode laser driver pulsed at 10 MHz. The light illuminates the sample causing any fluorescent material to fluoresce. The light from the fluorescence is detected by an EG&G single photon counting module (SPCM) single photon avalanche diode. The sample stage provides motion in x, y, and z directions with ranges of 64.5, 49.7, and 31:5 μm, respectively [28]. For this work, measurement scan sizes of 40 μm × 40 μm, consisting of 256 × 256 data points, were collected with a scan rate of 4:2 Hz. The data from the confocal microscope is exported as a raw ASCII file. A MATLAB program was created to analyze the data and provide the relevant measurement statistics. Sample analysis involves first normalizing the fluorescence at each data point by the maximum level of detected fluorescence on a control sample. The control sample is a Corning 0215 soda lime glass slide (SiO2 73%, Na2 O 14%, CaO 7%, MgO 4%, and Al2 O3 2% [31]) that never had quantum dot exposure; thus any normalized values less than 1 recorded on samples exposed to quantum dots are automatically discarded. The remaining data points are then evaluated with respect to maximum, mean, and minimum levels of fluorescence, and the percentage of data points with normalized values greater than one. 2. Wide Field Fluorescence Microscopy

A.

Interaction between the Dots and the Workpiece

Testing was undertaken to confirm the following. First, glass slides exposed to quantum dots can be effectively cleaned. This ensures that any fluorescence detected on the samples is not due to poor cleaning. Second, the dots are not absorbed into the glass surface over time. As explained in Subsection 2.A, it is not expected that they are absorbed. To answer these questions, several Corning 0215 glass slides were each loaded with a quantum dot solution. The solution used consisted of the bottle of quantum dot solution further diluted with acetone to a concentration of 60 nmol=mL. The samples were then divided into three groups. The first was immediately cleaned with isopropyl alcohol (IPA) moistened tissues. The second group was cleaned in the same manner after 110 min, and the final group was not cleaned, i.e., the quantum dots were left to dry on the surface. The groups were examined under the confocal microscope. On each slide three measurement sites were examined. Measurements were taken on the surface and at depths of 2, 4, 6, 8, and 10 μm beneath the surface. Table 2 details the results obtained on the sample surface; the reported deviations are the standard deviation over the three measurement sites. The only samples to exhibit any significant level of fluorescence are the samples that were exposed to quantum dots but were not cleaned. This confirms that the cleaning method is sufficient to remove the dots from a surface and that dots are not absorbed into the surface with exposure time of 110 min. As expected, the level of fluorescence detected beneath the surface was lower than the surface readings; see Fig. 2.
B. Lapping and Polishing Process Characterization

A wide field fluorescence microscope pairs a conventional optical microscope with a lamp emitting at a specified wavelength to excite any fluorescence present. Additional filters can be used to permit only the expected fluorescent wavelength to reach the imaging camera. In this work a mercury lamp provided the excitation and a “Lucifer Yellow” filter was used to filter unwanted wavelengths. As in conventional microscopy, the entire scan area is illuminated and imaged at the same time; this obviously makes it faster than the confocal imaging setup.
3. Experimental Procedures

Prior to performing the lapping and polishing tests with tagged slurries, it was necessary (1) to determine how the quantum dots interact with the workpiece surface and (2) to ensure that the lapping and polishing processes used to generate a smooth polished surface leaves SSD beneath the surface. Details of these two sets of prerequisite tests along with subsequent lapping and polishing of the Corning 0215 glass slides with tagged slurries are given in this section.

As already stated, it is critical to have a lapping and polishing procedure that produces samples with both a good surface finish (low roughness) and SSD. The lapping process will induce damage by brittle fracture mechanisms, while the polishing will improve the surface quality sufficiently for imaging on the confocal microscope. The following processes were used to achieve these objectives. The samples were first lapped on a cast iron plate with alumina slurry. The slurry was composed of 40 g of 20 μm particles (UnAlum 600) mixed in distilled water. The samples were hand lapped for 20 min with the platen rotating at 20 rpm. This resulted in matte surfaces with average Ra values of 0:37 μm (σ ? 0:03 μm). They were measured with a Mitutoyo SJ-400 surface roughness tester, using a 0:8 mm scan length averaged over five sample lengths. Based on sample weight loss, measured on an Ohaus Adventurer Pro scale, a depth of approximately 17:5 μm was removed by lapping. Polishing was done on a Strasbaugh overarm polisher using a 300 mm Dacron woven cloth pad (Allied High Tech Limited) on a metal platen. A 0:45 μm ceria based slurry (Hastilite PO) diluted with deionized water (ratio of 1∶8) was supplied to the pad at a rate
20 September 2009 / Vol. 48, No. 27 / APPLIED OPTICS 5159

Table 2.

Normalized Fluorescent Response of Glass Samples Exposed to Quantum Dots Under Varied Conditions

Test Details Baseline Testing Short exposure and cleaned 110 min exposure and cleaned Exposed and not cleaned Lapping and Polishing Tests Lapped and pad polished with quantum dots Lapped and pad polished with quantum dots and a 5 min pitch polish Pad polished with quantum dots

Maximum 1.59 (σ ? 0:19) 1.75 (σ ? 0:32) 43.2 (σ ? 4:1) 64.8 (σ ? 11:2) 13.49 (σ ? 4:2) 1.13 (σ ? 0:04)

Mean 1.22 (σ ? 0:06) 1.05 (σ ? 0:22) 28.9 (σ ? 5:1) 2.04 (σ ? 1:05) 2.86 (σ ? 1:14) 1.05 (σ ? 0:03)

% Points Fluorescing 0.0007% (σ ? 0:00003) 0.014% (σ ? 0:022) 100% (σ ? 0%) 1.3% (σ ? 2:2) 1.3% (σ ? 0:89) 0.0008% (σ ? 0:0016)

of 120 mL= min. A peristaltic pump recirculated the slurry. The samples were polished for 30 min under a load of approximately 20 kPa with a platen speed of 15 rpm and an arm sweep of 6 rpm. An average of 1:8 μm of material was removed from each of the samples (based on sample weight loss). The polished surface had a roughness value of approximately 1 nm Ra as measured on the Zygo NewView white light interferometer (110 μm × 80 μm). Atomic force microscope (AFM) measurements (40 μm × 40 μm) gave Ra values of the order of 1:5 nm. To confirm the existence of SSD, the samples were etched in a 2% solution of hydrofluoric acid for 30 s. This etching procedure removes approximately 130 nm. Optical images of the surface were acquired on a Mitutoyo Finescope microscope at magnifications ranging from 20× to 100×. The etched samples showed definite cracks and pitting that were not present in the pre-etch images; see Fig. 3. An additional pitch polishing stage was also developed to clean, further improve the surface quality, and remove the generated SSD. When using this pitch polishing process, no quantum dots were added to the process. This polishing step is only used selectively to help further understand and analyze the da-

mage formation mechanisms. While its necessity is explained in Section 4, the process parameters are given here. In the process a 300 mm diameter pitch tool made of Acculap Standard (synthetic pitch, equivalent to Gugolz 64) was used. X–Y grooves with approximately 10 mm spacings were scored onto the surface. The tool was broken in and charged with the same Hastilite PO slurry used for polishing. The samples were polished under a load of 7:5 kPa, with a platen speed and arm sweep of 15 and 3:5 rpm, respectively. Slurry was continuously supplied to the platen at a rate of 60 mL= min.
C. Lapping and Polishing with Tagged Slurries

The lapping and polishing steps were as outlined above. The quantum dots were added to the process as follows. In the lapping process, 10 mL of quantum dots, first diluted in 5 mL of acetone and 5 mL of distilled water, were introduced directly into the lapping mixture and added to the platen. Following lapping, the samples were immersed in a solution of 10 mL of quantum dots in toluene and 30 mL of acetone (final concentration was approximately 60 nmol=mL for 60 min). During polishing, the dots were added directly to the pad at the start of the polishing and every 5 min thereafter (up to 25 min). The added solution consisted of 10 mL of quantum dots diluted in acetone; again the final concentration was approximately 60 nmol=mL. After polishing, the samples were thoroughly cleaned with the isopropyl soaked tissues. The polished samples were visually examined by an optical microscope. Their surface roughness values and morphology were evaluated by both a Zygo NewView interferometer and a Digital Instruments AFM. Fluorescence was evaluated by the confocal and wide field microscopes. The results are presented in Section 4.
4. Results and Discussion

Fig. 2. Maximum and mean relative fluorescence detected on and beneath the surface of a quantum dot contaminated (not cleaned) glass sample, with error bars denoting the standard deviation. 5160 APPLIED OPTICS / Vol. 48, No. 27 / 20 September 2009

Figures 4(a)–4(d) illustrate the interferometric, confocal, wide field, and AFM measurements taken after the samples were lapped and pad polished with the tagged slurries. The interferometric image [Fig. 4(a)] demonstrates that the surfaces are clean and smooth

Fig. 3. Optical images of the sample surface after (a) lapping and polishing and (b) with ?1 μm etched away to reveal SSD.

(Ra ? 1:2 nm, σ ? 0:3 nm). Both the confocal scan [Fig. 4(b)] and the wide field image [Fig. 4(c)] show a number of fluorescent spots indicating that dots are embedded in, or under, the surface. Additional confocal scans taken to a depth of 10 μm below the surface show that the maximum level of fluorescence typically exists between the surface and 2 μm below the surface. Occasional features have been observed where the maximum fluorescence persists up to 10 μm beneath the surface (the deepest focal plane measured). The lower portion of Table 2 details the detected levels of fluorescence on the surface of

the sample, while Fig. 5 illustrates how the values drop off with increasing depth beneath the surface. Etching techniques were employed to estimate the fracture depth beneath the surface. Very conservative estimates put the fracture depth at between 3 and 7 μm. This indicates that the dots do not travel deep into the fracture sites, but are embedded in the surface at damage sites. The AFM scan supports this hypothesis, whereby raised features (black spots) are seen on the surface. The frequency of their occurrence is similar to the frequency of hot spots in the confocal image. It is worth noticing that the AFM

Fig. 4. (a) White light interferometer, (b) confocal fluorescence, (c) wide field fluorescence, and (d) atomic force microscopy images of the sample surface after lapping and polishing with tagged slurries. 20 September 2009 / Vol. 48, No. 27 / APPLIED OPTICS 5161

scan clearly shows that the surface is covered with light scratches, yet the incidence of either the black spots in the AFM or the fluorescing data points on the confocal scan do not typically exhibit a high degree of collinearity, which would be expected if they were aligned in the scratches. The mechanism of light scratch formation does not impart sufficient energy into the surface to enable the retention of quantum dots on the surface. The light scratches are most likely a by product of plastic deformation during the polishing process. To confirm that the quantum dots are strongly embedded in the surface, one sample was lightly etched in 2% hydrofluoric acid for 10 s. This would remove anything that is lightly attached to the surface. AFM scans of the etched surface reveal the same surface morphology as before etching. A confocal scan of the surface did not detect any fluorescence; however, this is not unexpected as the HF is thought to react strongly with the zinc sulfide shells, thus severely modifying the physical structure of the quantum dots and consequently their capability to fluoresce [32]. Samples were then subjected to a final 5 min pitch polishing process to further confirm that the quantum dots are not lightly adhered to the surface. This length of polishing time is not considered sufficiently long to remove all SSD. The depth of material removed was estimated to be approximately 250 nm (based on weight loss methods). After polishing, the Ra values of the sample measured by the Zygo had not altered significantly. The confocal microscope did detect fluorescence; however, the level dropped significantly from that present before the 5 min polish (see Table 2 and Fig. 5). The high occurrence of low intensity spots that are characteristic of the images captured after lapping and pad polishing [Fig. 6(a)] are notably absent from the confocal images of samples subject to the pitch polishing [Fig. 6(b)], where only the higher intensity sites remain. To investigate if the polishing process is only responsible for removing quantum dots from the sur-

Fig. 6. Confocal fluorescence image of a glass sample (a) lapped and polished with quantum dots, then (b) pitch polished at a focal plane coincident with the surface (contrast enhanced for printing).

face, a new undamaged glass slide was pad polished with a quantum dot tagged slurry. After polishing, this sample did not show any significant levels of fluorescence. Fewer than 0.006% of the data points in any one scan exceeded the background noise. This confirms that the polishing material removal mechanism does not create damage sites of sufficiently high surface energy to enable quantum dots to be embedded into the surface. For completion, a sample was polished on the pitch tool for 90 min. Again, no quantum dots were in the slurry. After polishing, the surface roughness was <0:7 nm Ra (Zygo, 65×). Little to no fluorescence was detected by the confocal microscope, and etching revealed that very little SSD was present in the sample. Successful removal of quantum dots from the surface strongly suggests that the SSD has also been eliminated.
5. Conclusions

Fig. 5. Maximum and mean fluorescence values for glass samples lapped and polished with quantum dots, then pitch polished at various focal planes, with error bars denoting the standard deviation. 5162 APPLIED OPTICS / Vol. 48, No. 27 / 20 September 2009

Quantum dots were added to the lapping and polishing slurries. Their detection after the polishing process indicates the presence of surface damage sites, which strongly suggests the presence of SSD. The same samples showed no indication of surface defects under optical or white light interferometric examinations. AFM scans showed raised features embedded in the sample surface that matches the occurrence of the fluorescing data points in the confocal scans. The raised features are embedded quantum dots. As quantum dots were not found on samples that were simply exposed to quantum dots or only polished with the tagged slurries, it suggests that the dots are embedded into the surface at lapping induced damage sites. The dots do not appear to travel deep into any cracks that may exist orthogonal to the surface. Additional testing is underway to confirm this hypothesis. This includes closer examination of fluorescent sites in the existing confocal images as well as examination of lapped samples that have not undergone any polishing steps. Quantum dots were removed from the surface by a long pitch polishing step; subsequent etching and examination of the same surface revealed that very little SSD was present. Again this strongly indicates

that the absence of fluorescence correlates to the absence of SSD. While the wide field imaging did not provide high resolution detail of the damage morphology, it does allow for very quick detection of quantum dots and thus offers the potential for a quick, nondestructive SSD detection test. Further work is also continuing here. This material is based upon the work supported by the National Science Foundation (NSF) under grant 0620783. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation. Northrop Grumman Synoptics has been an invaluable industrial partner, with Bryan Stanley, Kevin Stevens, and Adam Dittli making notable contributions to this work. Dr. Gloria Elliott in the UNC Charlotte Department of Mechanical Engineering and Engineering Science must be thanked for offering both her equipment and expertise in wide field fluorescence imaging. Dr. Jimmie Miller in the UNC Charlotte Center for Precision Metrology provided significant assistance with the atomic force microscopy.
References
1. P. P. Hed, D. F. Edwards, and J. B. Davis, “Subsurface damage in optical materials: origin, measurement and removal,” paper presented at the ASPE Spring Conference on Sub-Surface Damage in Glass, Tucson, Arizona, 25–27 April 1989. 2. D. W. Camp, M. R. Kozlowski, L. M. Sheehan, M. A. Nichols, M. Dovik, R. Raether, and I. Thomas, “Subsurface damage and polish compound affects the 355 nm laser damage threshold of fused silica surfaces,” Proc. SPIE 3244, 356–364 (1998). 3. J. C. Lambropoulos, Y. Li, P. Fukenbusch, and J. Ruckman, “Non-contact estimate of grinding-induced subsurface damage,” Proc. SPIE 3782, 41–50 (1999). 4. B. R. Lawn and T. R. Wilshaw, “Review indentation fracture: principles and applications,” J. Mater. Sci. Technol. (Sofia) 10, 1049–1081 (1975). 5. L. M. Cook, “Chemical processes in glass polishing,” J. Non-Cryst. Solids 120, 152–171 (1990). 6. F. P. Bowden and T. P. Hughes, “Physical properties of surfaces. IV. Polishing, surface flow and the formation of the Beilby layer,” Proc. R. Soc. London Ser. A 160, 575–587 (1937). 7. N. K. Adam, “The polishing of surfaces,” Nature 119, 162–163 (1927). 8. A. A. Griffith, “The phenomona of rupture and flow in solids,” Phil. Trans. R. Soc. London Ser. A 221, 163–198 (1921). 9. J. Shen, S. Liu, K. YI, H. He, J. Shao, and Z. Fan, “Subsurface damage in optical substrates,” Optik (Jena) 116, 288–294 (2005). 10. D. A. Lucca, E. Brinksmeier, and G. Goch, “Progress in assessing surface and subsurface integrity,” CIRP Ann. Manuf. Technol. 47, 669–693 (1998). 11. E. Brinksmeier, “State of the art of nondestructive measurement of subsurface material properties and damages,” Precis. Eng. 11, 211–224 (1989).

12. A. Lindquist, S. D. Jacobs, and A. Feltz, “Surface preparations for rapid measurement of subsurface damage depth,” presented at the OSA Science of Optical Finishing Topical Meeting, Monterey, California (1990). 13. “Quantum dots features,” (Evident Technologies, 2008), http:// www.evidenttech.com/quantum?dots?explained/ quantum?dot?features.html, retrieved 14 April 2008. 14. B. O. Dabbousi, M. G. Bawendi, O. Onitsuka, and M. F. Rubner, “Electroluminescence from CdSe quantum-dot/polymer composites,” Appl. Phys. Lett. 66, 1316–1318 (1995). 15. B. O. Dabbousi, O. Onitsuka, M. F. Rubner, and M. G. Bawendi, “Size dependent electroluminescence from CdSe nanocrystallites (quantum dots),” Mater. Res. Soc. Symp. Proc. 358, 707– 712 (1995). 16. “Nanomaterials catalog,” (Evident Technologies, 2005), Vol. 7. 17. R. A. Bowling, “A theoretical review of particle adhesion," in Particle on Surface. I. Detection, Adhesion, and Removal, K. L. Mittal, ed. (Plenum, 1988), pp. 129–142. 18. M. Wautelet, “Scaling laws in the macro-, micro- and nanoworlds,” Eur. J. Phys. 22, 601–611 (2001). 19. B. V. Derjaguin, V. M. Muller, and Y. P. Toporov, “Effect of contact deformations on the adhesion of particles,” J. Colloid Interface Sci. 53, 314 (1975). 20. D. Tabor, “Surface forces and surface interactions,” J. Colloidal Interface Sci. 58, 2–13 (1977). 21. G. M. Burdick, N. S. Berman, and S. P. Beaudoin, “Describing hydrodynamic particle removal from surfaces using the particle Reynolds number,” J. Nanopart. Res. 3, 453–465 (2001). 22. K. L. Johnson, K. Kendall, and A. D. Roberts, “Surface energy and the contact of elastic solids,” Proc. R. Soc. London Ser. A 324, 301–311 (1971). 23. Y.-P. Zhao, X. Shi, and W. J. Li, “Effect of work of adhesion on nanoindentation,” Rev. Adv. Mater. Sci. 5, 348–353 (2003). 24. J. Visser, “Particle adhesion and removal: a review,” Part. Sci. Technol. 13, 169–196 (1995). 25. F. Zhang and A. A. Busnaina, “Submicron particle removal in post-oxide chemical–mechanical planarization (CMP) cleaning,” Appl. Phys. A 69, 437–440 (1999). 26. A. A. Busnaina, H. Lin, N. Moumen, J.-W. Feng, and J. Taylor, “Particle adhesion and removal mechanisms in post-CMP cleaning processes,” IEEE Trans. Semicond. Manuf. 15, 374–382 (2002). 27. E. L. Cusser, Diffusion: Mass Transfer in Fluid Systems (Cambridge University, 1984). 28. K. Elliott, “Development of a versatile scanning system for multiprobe biomedical measurements,” Ph.D. dissertation (University of North Carolina at Charlotte, 2008). 29. M. Minsky, “Microscopy apparatus,” U.S. patent 3013467 (19 December 1961). 30. D. Semwogerere and E. R. Weeks, “Confocal microscopy,” in Encyclopedia of Biomaterials and Biomedical Engineering (Taylor & Francis, 2005), pp. 1–10. 31. “Soda lime glass 0215 Corning glass slides,” (Ted Pella Inc.), www.tedpella.com/technote_html/0215%corning%20glass.pdf. 32. M. Hines, Evident Technologies, Troy, New York (personal communication, 2008). 33. L. Bergstrom, “Hamaker constants of inorganic materials,” Adv. Colloid Interface Sci. 70, 125–169 (1997). 34. Z. Wang, L. L. Daemen, Y. Zhao, C. S. Zha, R. T. Down, X. Wang, Z. L. Wang, and R. J. Hemley, “Morphology-tuned wurtzite-type ZnS nanobelts,” Nat. Mater. 4, 922–927 (2005). 35. “Zinc sulfide specifications,” (ISP Optics), www.ispoptics.com/ OpticalMaterialsSpecs.htm.

20 September 2009 / Vol. 48, No. 27 / APPLIED OPTICS

5163


相关文章:
更多相关标签: