A Mach-Zender digital holographic microscope with sub-micrometer resolution for imaging and tracking of marine micro-organisms Jonas Kühn, Bimochan Niraula, Kurt Liewer, J. Kent Wallace, Eugene Serabyn, Emilio Graff, Christian Lindensmith, and Jay L. Nadeau Citation: Review of Scientific Instruments 85, 123113 (2014); doi: 10.1063/1.4904449 View online: http://dx.doi.org/10.1063/1.4904449 View Table of Contents: http://scitation.aip.org/content/aip/journal/rsi/85/12?ver=pdfcov Published by the AIP Publishing Articles you may be interested in A simple backscattering microscope for fast tracking of biological molecules Rev. Sci. Instrum. 81, 113704 (2010); 10.1063/1.3495960 Submersible digital in-line holographic microscope Rev. Sci. Instrum. 77, 043706 (2006); 10.1063/1.2193827 In situ imaging of micro-organisms in intense magnetic fields Rev. Sci. Instrum. 76, 103706 (2005); 10.1063/1.2103427 Two-dimensional tracking of a motile micro-organism allowing high-resolution observation with various imaging techniques Rev. Sci. Instrum. 76, 034301 (2005); 10.1063/1.1857632 Compact device for assessment of microorganism motility Rev. Sci. Instrum. 75, 4727 (2004); 10.1063/1.1809266

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP: 198.252.15.206 On: Thu, 08 Jan 2015 01:05:41

REVIEW OF SCIENTIFIC INSTRUMENTS 85, 123113 (2014)

A Mach-Zender digital holographic microscope with sub-micrometer resolution for imaging and tracking of marine micro-organisms Jonas Kühn,1 Bimochan Niraula,2 Kurt Liewer,1 J. Kent Wallace,1 Eugene Serabyn,1 Emilio Graff,3 Christian Lindensmith,1 and Jay L. Nadeau2,a) 1 Jet Propulsion Laboratory, California Institute of Technology, 4800 Oak Grove Dr., Pasadena, California 91009, USA 2 Department of Biomedical Engineering, McGill University, 3775 University St., Montreal, Quebec H3A 2B4, Canada 3 Division of Aerospace Engineering, California Institute of Technology, 1200 E. California Blvd., Pasadena, California 91125, USA

(Received 13 October 2014; accepted 5 December 2014; published online 19 December 2014) Digital holographic microscopy is an ideal tool for investigation of microbial motility. However, most designs do not exhibit sufficient spatial resolution for imaging bacteria. In this study we present an off-axis Mach-Zehnder design of a holographic microscope with spatial resolution of better than 800 nm and the ability to resolve bacterial samples at varying densities over a 380 μm × 380 μm × 600 μm three-dimensional field of view. Larger organisms, such as protozoa, can be resolved in detail, including cilia and flagella. The instrument design and performance are presented, including images and tracks of bacterial and protozoal mixed samples and pure cultures of six selected species. Organisms as small as 1 μm (bacterial spores) and as large as 60 μm (Paramecium bursaria) may be resolved and tracked without changes in the instrument configuration. Finally, we present a dilution series investigating the maximum cell density that can be imaged, a type of analysis that has not been presented in previous holographic microscopy studies. © 2014 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4904449] I. INTRODUCTION

Digital Holography was only made practical in the mid1990s, more than 40 years after the discovery of holography by the Nobel Prize recipient Denis Gabor.1 Thanks to breakthroughs in pixelated sensing technologies and computation, it is possible to reconstruct digitally recorded holograms on a computer in almost real time.2–4 Most digital holography implementations use coherent interference between a clear “reference” beam and a so-called “object” beam transmitted through, or reflected by, a sample of interest. This encodes the complex wavefront (amplitude and phase) as intensity modulation (fringes) in the detector plane (see Fig. S1 in the supplementary material for a discussion of the principles of off-axis holography5 ). Digital Holographic Microscopy (DHM)6–9 adds a lens or microscope objective in the object path to obtain a magnified image that is interfered with the reference wave; the object wavefront of interest is the magnified replica of the sample. Following the recording process, usually requiring the use of temporally coherent light (laser), reconstruction can be undertaken on a computer as a pure post-processing step. First, the encoded object wavefront of interest is retrieved from the digital hologram, typically using phase-shifting10–12 or spatial-filtering in the Fourier domain.13 Second, this complex quantity can be numerically propagated to any plane of interest using classical Fresnel6, 14 or Angular Spectrum15–17 classes of algorithms. Finally, the amplitude and phase components of the complex wavefront in the plane of destination are computed and saved for later analysis. a) Author to whom correspondence should be addressed. Electronic mail:

[email protected] 0034-6748/2014/85(12)/123113/6/$30.00

DHM possesses several key advantages over light microscopy. First is its ability to image a 3D volume by generating stacks of successive in-focus images corresponding to different layers in the sample, providing a so-called extended depth-of-field over conventional microscopy.18–20 This “digital focusing” capability can also be employed to re-focus blurred images in post-processing, thus alleviating mechanical stability requirements. Second, DHM is fast and realtime compatible. A modern computer using GPU-accelerated processing can reconstruct 4-Mpx digital holograms in realtime at up to 15 frames per second (1-Mpx ∼ 30+ frames per second). Recording just holograms, with reconstruction in post-processing, is limited only by camera speed. In addition, as the laser source usually exhibits several mW of optical power, camera shutter times in the few μs-range can be set to “freeze” temporal fluctuation frequencies up to 10– 100 kHz. Combining this with defocus insensitivity enables DHM to be inherently robust regarding environmental disturbance, despite its interferometric nature. This makes it suitable for use in extreme environments.21–24 Finally, DHM can deliver quantitative phase-contrast images corresponding to optical path difference (OPD) information with an interferometric accuracy down to a few nanometers,25 while the lateral resolution remains diffraction-limited, usually set by the microscope objective numerical aperture (NA). In a reflection configuration OPD directly corresponds to twice the sample topography (for a homogeneous material), while in transmission mode OPD is the depth-integrated by-product of thickness and refractive index. This provides label-free enhanced contrast for transparent or semi-transparent samples, such biological cells, without the need for fluorescent dyes. Such a

85, 123113-1

© 2014 AIP Publishing LLC

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP: 198.252.15.206 On: Thu, 08 Jan 2015 01:05:41

123113-2

Kühn et al.

Rev. Sci. Instrum. 85, 123113 (2014)

label-free imaging capability explains the growing interest in DHM for biological applications ranging from pharmacology26, 27 to cellular biology28–31 and marine biology.32–35

II. INSTRUMENT DESCRIPTION: AN OFF-AXIS MACH-ZEHNDER CONFIGURATION

Previous DHM implementations used for biology relied on a single-beam, lensless pinhole configuration,36, 37 which has the advantage of compactness, low cost, and design simplicity. However, the sub-micrometer resolution needed to image most bacteria cannot be realistically reached with this configuration. Either geometrical scales of the same order as the camera chip are required, or the sample must be immersed in an invasive high-refractive index solution.38 Furthermore, this approach (and other alternate in-line DHM geometries39, 40 ) inherently introduces ghost artefacts caused by the “halo” of the out-of-focus complex conjugate image. This introduces noise which is detrimental to imaging small structures. In addition, in-line implementations cannot rely on spatial frequency modulation to isolate the interference term of interest in the Fourier domain,13 and typically require phase-shifting10 to get rid of the conjugate image, hence prohibiting real-time reconstruction. An off-axis dual-beam implementation is free of these drawbacks, with the sole disadvantage being loss of a factor of ∼2 in field of view compared to in-line arrangements, as twice the magnification is required to achieve the same pixel sampling. This led us to retain a dual-beam off-axis Mach-Zehnder (MZ) DHM configuration, using a magnifying lens to reach the targeted sub-micrometer lateral resolution required for bacterial imaging (Fig. 1). The MZ optical design also takes into account the need to minimize on-axis optical interfaces in order to reduce coherent parasitic reflections. Typically, antireflection (A/R)-coated plate beamsplitters (BS) are used in place of cube BSs, and we refrained from using a lens-assembly commercial microscope objective (MO). Therefore, in order to meet the goal of submicrometer lateral resolution without using a bulky MO, we R asretained an A/R-coated C240-TM-A Thorlabs Geltech pheric lens with a focal length f = 8 mm and NA of 0.5 as a magnification M = 20× single-lens objective (image plane at ∼160 mm from the lens back focal length). Combined with the use of a blue (λ = 405) nm laser diode (Thorlabs S1FC405), this sets the theoretical diffraction-limited resolution at ∼λ/NA ∼ 0.8 μm (0.5 μm under the Rayleigh criterion). The reference beam goes through the same lens in its own path before being combined with the object beam at a slight angle, as required by the off-axis scheme. The camera is a 2448 × 2050 Baumer TXG-50 digital CCD camera with 3.45 μm pixel size, delivering up to 22 fps in 1024 × 1024 (1 Mpx) mode and 14 fps with 2048 × 2048 pixels (4 Mpx). For high-speed hologram reconstruction, we employed the R off-axis DHM reconstruction software41 LynceeTec Koala that can deliver camera-limited live reconstruction from holograms up to 4 Mpx, and wavefront phase retrieval using polynomial fit42, 43 and reference conjugated hologram44 algorithms.

FIG. 1. Instrument design. (a) General schematic showing a Typical MachZehnder off-axis DHM scheme. O, Object beam; R, reference beam; BS, beamsplitter; M, mirrors; L, lens; MO, microscope objective. (b) Photo of the actual breadboard instrument (arrow indicates the sample).

The MZ approach to holographic microscopy has been previously reported.45–48 The novelty of the instrument reported here lies in its optimization for real-time submicrometer imaging of bacterial cultures at realistic densities. Several of the elements are critical to achieving this performance: the 405 nm laser, the A/R coated 20× objective lens, and the choice of camera. These elements combined led to spatial resolution that has not been previously reported in holographic microscopy, including imaging of low-contrast microorganisms in liquids, as detailed in Sec. III. Section IV demonstrates the ability of this instrument to image single microorganisms in crowded samples, a unique application of the off-axis geometry.

III. INSTRUMENT PERFORMANCE A. 3-D field-of-view and resolution

We used a NIST-compliant high-resolution USAF test target in order to evaluate three-dimensional FOV and optical resolution (Fig. 2). The lateral FOV was 0.385 × 0.385 mm2 over a slightly cropped area of 1980 × 1980 pixels due to apodization processing,49 indicative of an effective magnification of about 18×: this is in accordance with the CCD camera location along the optical axis, about 2–3 cm upstream of the 160-mm image plane of the single-asphere objective. The DHM was able to resolve the finest USAF elements, with

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP: 198.252.15.206 On: Thu, 08 Jan 2015 01:05:41

123113-3

Kühn et al.

Rev. Sci. Instrum. 85, 123113 (2014)

FIG. 2. MZ-DHM optical performance assessment with USAF test target sample. (a) Intensity reconstruction image, with inset showing the highest resolution elements: the white circle indicates 0.78 μm wide bars. (b) and (c) Zoomed images on the smallest elements (groups 8 and 9) of (a) but reconstructed for the target moved along the optical axis (focus) by −1 mm (b) or +1 mm (c) away from the nominal working distance used in (a). Note that the 3rd element of group 8 (1.55 μm wide) is still well resolved in both cases as shown in the cross-profiles, indicating less than a factor two loss in resolution vs. the nominal 0.8 μm.

a line thickness of 0.78 μm (Fig. 2(a)). In depth, we moved the sample away from the design working distance by ±1 mm along the optical axis and the resolution degraded by less than a factor 2 (Fig. 2(b)). This depth corresponded to ±300 μm in image space, which is the limit imposed by the KoalaTM software. We also confirmed a lateral resolution better than 0.8 μm over a range of ±0.3 mm in object space from the initial working distance, which would correspond to the nominal-resolution volume of investigation. Overall, this gives a volume of investigation in excess of 0.38 × 0.38 × 2 mm3 , with nominal

A Mach-Zender digital holographic microscope with sub-micrometer resolution for imaging and tracking of marine micro-organisms.

Digital holographic microscopy is an ideal tool for investigation of microbial motility. However, most designs do not exhibit sufficient spatial resol...
2MB Sizes 0 Downloads 7 Views