Learn to light a candle in the darkest moments of someone’s life. Be the light that helps others see; i

Uploaded by
Roy T. Bennett

103 downloads 5499 Views 2MB Size

Optical Coherence Tomography

I tried to make sense of the Four Books, until love arrived, and it all became a single syllable. Yunus

EZ-LIGHT K90TFDL Daylight Visible

In the end only three things matter: how much you loved, how gently you lived, and how gracefully you

Visible light for broadband communications

Open your mouth only if what you are going to say is more beautiful than the silience. BUDDHA

Optical spectroscopy of the Ni2

Love only grows by sharing. You can only have more for yourself by giving it away to others. Brian

Visible Light Communication, Networking and Sensing

Never wish them pain. That's not who you are. If they caused you pain, they must have pain inside. Wish

Soraa Optical Light Engine-Specsheet.pdf

If your life's work can be accomplished in your lifetime, you're not thinking big enough. Wes Jacks

Adaptive optics for fluorescence correlation spectroscopy

The beauty of a living thing is not the atoms that go into it, but the way those atoms are put together.

optical spectroscopy of bright fermi lat blazars

Be like the sun for grace and mercy. Be like the night to cover others' faults. Be like running water

Visible light optical coherence correlation spectroscopy Stephane Broillet,1* Daniel Szlag,1 Arno Bouwens,1 Lionel Maurizi,2 Heinrich Hofmann,2 Theo Lasser,1 and Marcel Leutenegger1 ´ d’Optique Biom´edicale LOB, Ecole Polytechnique F´ed´erale de Lausanne, CH-1015 Lausanne, Switzerland 2 Laboratory of Powder Technology LTP, Ecole ´ Polytechnique F´ed´erale de Lausanne, CH-1015 Lausanne, Switzerland

1 Laboratoire

* [email protected]

Abstract: Optical coherence correlation spectroscopy (OCCS) allows studying kinetic processes at the single particle level using the backscattered light of nanoparticles. We extend the possibilities of this technique by increasing its signal-to-noise ratio by a factor of more than 25 and by generalizing the method to solutions containing multiple nanoparticle species. We applied these improvements by measuring protein adsorption and formation of a protein monolayer on superparamagnetic iron oxide nanoparticles under physiological conditions. © 2014 Optical Society of America OCIS codes: (290.5850) Scattering, particles; (030.1670) Coherent optical effects; (170.0170) Medical optics and biotechnology; (170.4500) Optical coherence tomography.

References and links 1. P. Ghosh, G. Han, M. De, C. Kim, and V. Rotello, “Gold nanoparticles in delivery applications,” Adv. Drug Delivery Rev. 60, 1307–1315 (2008). 2. H. Jans and Q. Huo, “Gold nanoparticle-enabled biological and chemical detection and analysis,” Chem. Soc. Rev. 41, 2849–2866 (2012). 3. A. Gupta, R. Naregalkar, V. Vaidya, and M. Gupta, “Recent advances on surface engineering of magnetic iron oxide nanoparticles and their biomedical applications,” Nanomedicine 2, 23–39 (2007). 4. Q. Pankhurst, N. Thanh, S. Jones, and J. Dobson, “Progress in applications of magnetic nanoparticles in biomedicine,” J. Phys. D: Appl. Phys. 42, 1–15 (2009). 5. R. Singh and H. Nalwa, “Medical applications of nanoparticles in biological imaging, cell labeling, antimicrobial agents, and anticancer nanodrugs,” J. Biomed. Nanotechnol. 7, 489–503 (2011). 6. S. Broillet, A. Sato, S. Geissbuehler, C. Pache, A. Bouwens, T. Lasser, and M. Leutenegger, “Optical coherence correlation spectroscopy (occs),” Opt. Express 22, 782–802 (2014). 7. J. Robinson, T. Takizawa, D. Vandr´e, and R. Burry, “Ultrasmall immunogold particles: Important probes for immunocytochemistry,” Microsc. Res. Tech. 42, 13–23 (1998). 8. K. Huang, H. Ma, J. Liu, S. Huo, A. Kumar, T. Wei, X. Zhang, S. Jin, Y. Gan, P. Wang, S. He, X. Zhang, and X.-J. Liang, “Size-dependent localization and penetration of ultrasmall gold nanoparticles in cancer cells, multicellular spheroids, and tumors in vivo,” ACS Nano 6, 4483–4493 (2012). 9. T. Cedervall, I. Lynch, S. Lindman, T. Berggard, E. Thulin, H. Nilsson, K. Dawson, and S. Linse, “Understanding the nanoparticle-protein corona using methods to quntify exchange rates and affinities of proteins for nanoparticles,” Proc. Natl. Acad. Sci. U. S. A. 104, 2050–2055 (2007). 10. U. Sakulkhu, M. Mahmoudi, L. Maurizi, J. Salaklang, and H. Hofmann, “Protein corona composition of superparamagnetic iron oxide nanoparticles with various physico-chemical properties and coatings,” Sci. Rep. 4, 1–9 (2014). 11. P. Aggarwal, J. Hall, C. McLeland, M. Dobrovolskaia, and S. McNeil, “Nanoparticle interaction with plasma proteins as it relates to particle biodistribution, biocompatibility and therapeutic efficacy,” Adv. Drug Delivery Rev. 61, 428–437 (2009).

#214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21944

12. C. Walkey and W. Chan, “Understanding and controlling the interaction of nanomaterials with proteins in a physiological environment,” Chem. Soc. Rev. 41, 2780–2799 (2012). 13. C. Rocker, M. Potzl, F. Zhang, W. Parak, and G. Nienhaus, “A quantitative fluorescence study of protein monolayer formation on colloidal nanoparticles,” Nat. Nanotechnol. 4, 577–580 (2009). 14. S. Dominguez-Medina, S. McDonough, P. Swanglap, C. Landes, and S. Link, “In situ measurement of bovine serum albumin interaction with gold nanospheres,” Langmuir 28, 9131–9139 (2012). 15. I. Kohli, S. Alam, B. Patel, and A. Mukhopadhyay, “Interaction and diffusion of gold nanoparticles in bovine serum albumin solutions,” Appl. Phys. Lett. 102, 203705 (2013). 16. M. Mikhaylova, D. Kim, C. Berry, A. Zagorodni, M. Toprak, A. Curtis, and M. Muhammed, “Bsa immobilization on amine-functionalized superparamagnetic iron oxide nanoparticles,” Chem. Mater. 16, 2344–2354 (2004). 17. Z. Wang, T. Yue, Y. Yuan, R. Cai, C. Niu, and C. Guo, “Kinetics of adsorption of bovine serum albumin on magnetic carboxymethyl chitosan nanoparticles,” Int. J. Biol. Macromol. 58, 57–65 (2013). 18. M. Villiger and T. Lasser, “Image formation and tomogram reconstruction in optical coherence microscopy,” J. Opt. Soc. Am. A 27, 2216–2228 (2010). 19. R. A. Leitgeb, M. Villiger, A. H. Bachmann, L. Steinmann, and T. Lasser, “Extended focus depth for fourier domain optical coherence microscopy,” Opt. Lett. 31, 2450–2452 (2006). 20. C. Pache, N. Bocchio, A. Bouwens, M. Villiger, C. Berclaz, J. Goulley, M. Gibson, C. Santschi, and T. Lasser, “Fast three-dimensional imaging of gold nanoparticles in living cells with photothermal optical lock-in optical coherence microscopy,” Opt. Express 20, 21385–21399 (2012). 21. M. Villiger, C. Pache, and T. Lasser, “Dark-field optical coherence microscopy,” Opt. Lett. 35, 3489–3491 (2010). 22. J. Izatt and M. Choma, Optical Coherence Tomography: Technology and Applications (Springer Verlag, Berlin, 2008). 23. M. Wojtkowski, V. Srinivasan, T. Ko, J. Fujimoto, A. Kowalczyk, and J. Duker, “Ultrahigh-resolution, highspeed, fourier domain optical coherence tomography and methods for dispersion compensation,” Opt. Express 12, 2404–2422 (2004). 24. C. Dorrer, N. Belabas, J.-P. Likforman, and M. Joffre, “Spectral resolution and sampling issues in fouriertransform spectral interferometry,” J. Opt. Soc. Am. B 17, 1795–1802 (2000). 25. T. Wohland, R. Rigler, and H. Vogel, “The standard deviation in fluorescence correlation spectroscopy,” Biophys. J. 80, 2987–2999 (2001). 26. N. L. Thompson, Fluorescence correlation spectroscopy. In: Topics in Fluorescence Spectroscopy, vol. 1 (Plenum Press, 1991). 27. U. Meseth, T. Wohland, R. Rigler, and H. Vogel, “Resolution of fluorescence correlation measurements,” Biophys. J. 76, 1619–1631 (1999). 28. B. Berne and R. Pecora, Dynamic Light Scattering with Applications to Chemistry, Biology and Physics. Chapter 5: Model Systems of Spherical Molecules (John Wiley and Sons, 1976). 29. M. Leutenegger, R. Rao, R. Leitgeb, and T. Lasser, “Fast focus field calculations,” Opt. Express 14, 11277–11291 (2006). 30. R. Leitgeb, C. Hitzenberger, and A. Fercher, “Performance of fourier domain vs. time domain optical coherence tomography,” Opt. Express 11, 889–894 (2003). 31. M. Chastellain, A. Petri, and H. Hofmann, “Particle size investigations of a multistep synthesis of pva coated superparamagnetic nanoparticles,” J. Colloid Interface Sci. 278, 353–360 (2004). 32. L. Maurizi, U. Sakulkhu, L. Crowe, V. Dao, N. Leclaire, J.-P. Vallee, and H. Hofmann, “Syntheses of cross-linked polymeric superparamagnetic beads with tunable properties,” RSC Adv. 4, 11142–11146 (2014). 33. M. Ferrer, R. Duchowicz, B. Carrasco, J. De La Torre, and A. Acuna, “The conformation of serum albumin in solution: A combined phosphorescence depolarization-hydrodynamic modeling study,” Biophys. J. 80, 2422– 2430 (2001). 34. H. Kenneth Walker, W. Dallas Hall, and J. Willis Hurst, Clinical Methods, 3rd edition; The History, Physical, and Laboratory Examinations; Chapter 101: Serum Albumin and Globulin (Butterworths, 1990).

1.

Introduction

Nanoparticles (NPs) are studied thoroughly for applications in medicine and life sciences: of special interest are gold NPs due to their optical, chemical, and biocompatible properties [1, 2] and superparamagnetic iron oxide nanoparticles (SPIONs) [3–5]. Recently, we have shown that optical coherence correlation spectroscopy (OCCS) can measure the diffusion coefficient of gold NPs down to diameters of 30 nm with near-infrared light [6]. There are advantages to use even smaller gold NPs because of an enhanced labeling efficiency [7] and improved penetration into biological samples [8], showing the importance of measuring the dynamics of smaller gold NPs.

#214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21945

Another very interesting feature would be to have the possibility to distinguish different species of scattering NPs, for instance based on their size. This would give the possibility to distinctively label several targets in measurements and to measure chemical reaction kinetics, for instance. The coating process of nanoparticles (NPs) with biomolecules generates a so-called protein corona [9, 10]. Understanding the properties of this coating is important for the application of NPs in living organisms [11]. Most strategies involve the removal of the NP–protein complex from their original physiological environment [12]. A complementary approach investigates the adsorption of proteins under physiological conditions. So far, the adsorption of either bovine serum albumin (BSA) or human serum albumin (HSA) has been studied with different types of NPs using fluorescence correlation spectroscopy (FCS), as for example the coating process of FePt NPs and quantum dots [13] or gold NPs [14, 15]. Adsorption of BSA on SPIONs has already been studied [16, 17], however, a quantitative analysis of their diffusion properties in physiological conditions has not been performed yet. OCCS exploits the backscattered light of NPs illuminated by a broadband light source, giving simultaneous access to several sampling volumes along the optical axis. Over a large concentration range, this technique allows the extraction of the diffusion coefficient and concentration of NPs using an auto-correlation analysis. In this work we greatly extend the application potential of OCCS by shifting the OCCS instrument from the near-infrared to the visible wavelength range. The signal-to-noise ratio (SNR) has been increased by a factor of more than 25 due to a much increased NA allowing the extraction of the diffusion coefficient of gold NPs down to a size of 10 nm diameter. In addition, the auto-correlation model has been generalized to solutions hosting multiple NP species and this generalized model is used to retrieve the particle fractions in a mixture of two NP species. Using our novel OCCS technique, we show its application to the study of BSA adsorption on SPIONs by measuring their hydrodynamic radius under physiological conditions. This combines the advantages of not having to use fluorescent labels (thus preventing the risk that these labels change the properties of the NPs) and of avoiding additional separation steps to remove unbound protein from the solution. 2. 2.1.

Materials and methods Low coherence interferometer

Figure 1 depicts the OCCS instrument which is based on a Mach-Zehnder interferometer [6, 18–21]. This visible light OCCS instrument is based on low coherence interferometric principles as described in [6]. The light from a super-continuum source (Koheras SuperK Extreme, NKT Photonics) is spectrally filtered to obtain a broad visible spectrum centered at 590 nm wavelength with a bandwidth of 160 nm full width at half-maximum (FWHM). The light is linearly polarized by a polarizing beam splitter and delivered to the interferometer by a singlemode fiber (mode field diameter of 3.4 µm, Fibercore). After collimation, the light is split into a reference beam and an illumination beam by a non-polarizing 50:50 beam splitter. The reference beam passes through glass prisms with variable thicknesses for compensating the dispersion difference in the illumination and sample arms. The illumination beam passes through an axicon lens (176◦ cone angle, Asphericon) generating a Bessel beam, which is further imaged by a 300 mm achromatic lens and an apochromatic water immersion objective (40×, NA 1.15, Olympus) into the sample at the focal plane of the objective. The power impinging on the sample is adjusted between 2–10 mW, depending on the experiment. The annular mask Fill blocks residual stray light from the tip of the axicon. The back-scattered sample field is collected by the objective, superimposed with the reference field and focused into a singlemode fiber (mode field diameter of 3.4 µm, Thorlabs). The fiber guides the collected field to the customized spectrometer (transmission grating; 600 lines/mm, Wasatch) for the registration #214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21946

L2

L3

Detection

Axicon

Fdet

Illumination

Reference

Illumination fiber

L1

Beam splitter

40x/1.15NA

Sample

Detection fiber

Fill

Dispersion compensation

Fig. 1. Visible light OCCS interferometer with a Bessel-Gauss configuration [18]. The axicon generates a Bessel beam illumination, whereas the detection mode is Gaussian. The apertures Fill and Fdet provide a dark field contrast [21].

of the spectral interferogram from 420 nm to 760 nm wavelength (linear CMOS array Sprint spL2048-140km with 2048 pixels, Basler). As illustrated in Fig. 1, the illumination and detection beams do not overlap in the aperture plane of the objective: any specular reflected light from the glass slide is blocked by the diaphragm Fdet . The interferometer was conceived with a minimum number of optical elements in order to increase the detection efficiency and to simplify the alignment of components. Because the glass slide is always closer to the objective than the sample, a dark field is obtained even though the filters Fill and Fdet are not placed in a pupil plane but just as close to the objective as possible. The dark field contrast is essential for a high SNR while detecting the weak back-scattered field from the NPs. This interferometer implements a Bessel illumination and Gaussian detection configuration [18]. The illumination field corresponds to a radial zero-order Bessel distribution in the focal plane with the first minimum located at 0.22 µm lateral radius, whereas the detection mode is Gaussian with an effective numerical aperture (NA) close to 0.9. In typical experiments the detector was driven at a line rate of 20 kHz with an exposure time of 48 µs. The reference arm intensity was adjusted to about 75% of the available dynamic range of the camera. By taking the Fourier transform of the spectrum (resampled at equidistant wave numbers, λ to k mapping), we obtain the time-dependent signal traces extending over several sampling volumes [22]. Here, the background corrected depth profile containing a sequence of sampling volumes (center to center distance of 0.35 µm in water) was then extracted by using an algorithm based on the discrete Fourier transform (DFT). The signal processing leading to the time-dependent signal traces is indicated in Fig. 2. An integration kernel to get the tomogram from the sample spectrum by a matrix product was calculated such that (i) it resampled the spectrum at equidistant wave numbers (λ to k mapping); (ii) it compensated the residual dispersion φ (k) by multiplying the spectrum with calibrated phase factors exp(i φ (k)) calculated with an automatic numerical iterative procedure developed by Wojtkowski et al. [23]; and (iii) Fourier transformed the corrected spectrum to yield the tomogram in (iv) the region of interest.

#214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21947

t

I(z) tim et

Background subtraction

λ k mapping

Dispersion compensation

FFT

Selection of the z indices

depth z

wavevelength λ

I(λ) tim e

DFT

Modulus

Fig. 2. Signal processing leading to the time-dependent signal traces. The interference signal is recorded via a spectrometer. The average spectrum is then subtracted from the recorded signal (background subtraction). The new DFT integration kernel replaces the following steps: (i) λ to k mapping; (ii) dispersion compensation; (iii) fast Fourier transform of the spectrum; and (iv) selection of the sampling volumes of interest. The modulus of the signal is taken to yield the time-dependent signal traces.

Thereby, the matrix multiplication of this kernel with the measured spectra directly yielded the depth profiles. This algorithm is only performed on the sampling volumes in the optical focal area in order to minimize the calculation time. In all the calculations that follow, 150 sampling volumes are calculated. Auto-correlations of the absolute values of these time-dependent signals were then calculated. In this application, the new DFT algorithm is about six times faster than the previous FFT algorithm [6]. In detail, the DFT kernel F is an M × N matrix, where M is the number of tomogram samples m ∈ [0, N) and N is the number of interferogram samples. The tomogram is obtained as matrix product I(z) = F · I(λ ). In other words, I(zm ) = ∑N−1 n=0 Fmn I(λn ), where zm = m∆z = m·0.35 µm. The fast Fourier transform is performed by Fmn = exp(−2i πmn/N) for m, n ∈ {0, N − 1}. In the DFT algorithm, the dispersion is compensated by multiplying the interferogram samples with a factor fn = exp (i φ (λn )). An extra amplitude component would allow reshaping the spectral amplitudes. Therefore, the kernel implementing steps (ii–iv) writes as Fmn = fn exp(−2i πmn/N). The λ to k mapping (i) is performed by multiplying this kernel with an N × N identity matrix resampled at wavelengths corresponding to equidistant wave number samples. A high-quality resampling is achieved by Fourier-upsampling the columns of the identity matrix to yield L & 3N samples, which are then linearly interpolated at the desired wavelengths [24]. 2.2.

Sample preparation

Sample solutions of monodisperse polystyrene microspheres (PS MSs) with a diameter of 109 nm (POLYBEAD Microspheres 109 nm, Polysciences) and gold colloids (gold NPs) with diameters of 10, 15, 20, 30, 40 and 50 nm (EM.GC10, EM.GC15, EM.GC20, EM.GC30, EM.GC40, EM.GC50, British Biocell International) were used. To obtain the nominal concentrations, the solutions were prepared by diluting an adequate volume of stock solution in ultrapure water or in glycerol. All OCCS measurements were performed in plastic wells (µSlide 8 well, uncoated, sterile, Ibidi) with a well volume of 300 µl.

#214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21948

2.3.

Correlation analysis of multiple components mixtures

The signal acquisition in OCCS and its auto-correlation analysis for single species was described in previous work [6], where the interested reader finds the detailed derivation of Eq. (1). The auto-correlation analysis of the OCCS signal Im (t) in a sampling volume m was subdivided into three distinct regimes depending on the number of particles in the sampling volume of interest N: the single particle (N 1), the few particles (N ∼ 1) and the many particles (N 1) regimes. The general case is the few particles regime, which takes into account the diffusion of particles in and out of the sampling volume and a dynamic light scattering (DLS) contribution depending on the number of particles N in the sampling volume. In the few particles regime, the OCCS autocorrelation function can be written as r −1 τ τ τ τ γ 1+ 1+ 1 + Ab exp − 1 + N exp − , (1) GF,m (τ) = N τxy τz τb τc where γ is the volume contrast [25] and N the average number of particles in the sampling volume. τxy = r02 (4D)−1 and τz = z20 (4D)−1 are the lateral and axial diffusion times, where D is the diffusion coefficient, and r0 and z0 are the lateral and axial e−2 radii of an equivalent threedimensional Gaussian volume. The exponential decay τb = rb2 (4D)−1 with the characteristic length rb and amplitude parameter Ab account for the Bessel profile. The last factor is the coherent contribution caused by speckle fluctuations of the back-scattered field of the NPs in one sampling volume. This coherent contribution is characterized by the decorrelation time τc = (8n2 k02 D)−1 known from dynamic light scattering (DLS), where n is the refractive index of the medium and k0 is the central wave number of the source spectrum. Here we extend this initial model to the study of multiple independently diffusing species i with different diffusion coefficients. Without the coherent contribution, Eq. (1) describes a correlation function for a signal similar to the intensity fluctuations in FCS. Therefore, we can apply the multi-species model derived for FCS, see for instance Thompson [26] and Meseth et al. [27]. For the coherent contribution, Berne and Pecora [28] describe a multi-species model in the context of DLS. In both cases, the multi-species models are linear combinations of the single-species correlation functions. Therefore, we can write the multi-species auto-correlation function as hBi 2 ∑ Ni Q2i DF,m,i (τ) mix GF,m (τ) = γ 1 − , (2) hIm i (∑ Ni Qi )2 where DF,m,i (τ) =

1+

τ τxy,i

r 1+

τ τz,i

−1 τ τ 1 + Ab exp − 1 + Ni exp − . (3) τb,i τc,i

hIm i = hBi + ∑ Ni Qi is the average signal in the sampling volume m and hBi is the average background. Ni is the average number and Qi is the brightness of particles of species i. Qi is proportional to the absolute value of the particle’s electric field reflectivity. The squared prefactor stands for the background influence on the correlation amplitude. The total number of particles is N = ∑ Ni and Pi = Ni /N are the fractions of the species. Each species contributes to the correlation amplitude by its concentration times the squared brightness which limits the discrimination of species with significant brightness differences. 2.4.

Sampling volume characterization

The determination of r0 and z0 is essential for an appropriate fit model requiring an accurate characterization of the OCCS sampling volumes. In analogy to the near-infrared light #214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21949

OCCS procedures described in [6], we characterized the sampling volume by imaging individual 50 nm gold NPs. The scattering particles were dispersed in an agarose gel with 0.3% weight/volume ratio. We scanned individual NPs using a three-axis piezoscanning stage in the x-y plane (repeatability 0.01 µm, Physik Instrumente) while illuminating with a power of 5 mW. We compared the brightness profile measurements with ab initio calculations at the cen-

(a)

y

Signal along x,y [a.u.]

(b)

1

0.4 0.2 0

x

(e)

Signal [a.u.]

1

Signal along z [a.u.]

(d)

z

Au 50nm

0.6

x (c)

Calculation

0.8

0

0.1

1

0.6

0.7

Coh. gate

0.8

Au 50nm

0.6 0.4 0.2 0 −2

−20

0.2 0.3 0.4 0.5 Radial distance ρ [µm]

−10

−1.5

−1

−0.5 0 0.5 Depth z [ µm]

Sampling volumes Vm 0 10 20 30

40

1

1.5

2

50

Measurement FWHM=7.1 µm Calculation FWHM=3.3 µm

0.8 0.6 0.4 0.2 0

−5

0

5 10 Depth z [ µm]

15

Fig. 3. Brightness profiles characterization. (a,c) Brightness profile cross-sections xy and xz p measured with a 50 nm single gold NP. Scalebars: 1 µm. (b,d) Average radial ρ = x2 + y2 and axial (z) brightness profiles measured on ten 50 nm gold NPs in comparison with calculations. (e) Depth of field characterization with 109 nm PS MSs freely diffusing in water (concentration: 330 pM; illumination power: 5 mW; average on 10 measurements of 50 seconds). The useful DOF is indicated by the dashed lines showing the measured FWHM. V0 corresponds to the focal sampling volume. The calculations in (b) and (e) were performed using the focus field calculation framework by Leutenegger et al. [29]. The coherence gate in (d) is calculated from the measured source spectrum S(k) using the Wiener-Khintchine relation.

#214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21950

tral wavelength using the focus field calculation framework by Leutenegger et al. [29]. The brightness profile is rotationally symmetric around the optical axis z. Figures 3(a)–3(d) show the measured brightness profile W (r). Figure 3(b) shows the measured radial brightness profiles W (r) compared with the calculation performed at the central wavelength of 590 nm. We attribute the non-zero minimum (10% of the maximal signal) and larger profile in the measurement to the large illumination bandwidth and to the finite size of the NPs used. Figure 3(d) shows the measured axial brightness profiles W (r) compared with the coherence gate determined by the measured source spectrum S(k) and calculated using the Wiener-Khintchine relation. The e−2 half width z0 of this calculated axial coherence gate is 0.95 µm. There is a good agreement between the calculation and the measurement in the profiles’ FWHM. The difference in the tail is < 15% of the maximal signal and is due to residual dispersion difference between the reference path and the illumination plus detection paths. The depth of field (DOF) assessment is important for the evaluation of the number of useful sampling volumes along the optical axis. For assessing this parameter, we used a solution of 109 nm PS MSs (concentration of 330 pM) freely diffusing in water, an illumination power of 5 mW and measured the time-averaged signal amplitude hIm i for the sampling volumes Vm along the optical axis. Figure 3(e) compares the DOF (average over 10 measurements during 50 seconds) with our ab initio calculations. The measurements indicate a DOF of approximately 7 µm FWHM corresponding to 20 sampling volumes, which can be observed simultaneously in axial separation steps of ∆z = 0.35 µm. The significant increase in DOF as compared to calculations is most likely caused by longitudinal chromatic aberrations in the illumination and detection paths. 2.5.

Data Analysis

Care should be taken to carefully calibrate several parameters summarized in Table 1. Figure 3(d) confirms the good agreement between the calculated and the measured axial brightness profile. Therefore, the axial extent z0 was extracted from the coherence length given by the spectrum of the light source in Fig. 3(d). The volume contrast factor γ was calculated from the measured PSF [25]. The lengths r0 , rb and the parameter Ab were calibrated for each sampling volume with an auto-correlation measurement of 109 nm PS MSs matched to Eq. (1) by using the theoretical value for the diffusion coefficient D (concentration of 330 pM; illumination power of 5 mW; 10 measurements of 50 seconds). This calibration was repeated for every measurement session. The characteristic lengths r0 , rb , z0 as well as the parameters Ab and γ were then kept fixed and Eq. (1) was used to fit the auto-correlations shown in Fig. 5(a) and Fig. 7(a), where only N and D have been retained as free fit parameters. The hydrodynamic radii RH of particles in Fig. 7(b) and Fig. 7(c) were calculated using the Stokes-Einstein relation D = kB T /(3πη2RH ), where kB is the Boltzmann constant, T the absolute temperature (295 K), and η the fluid viscosity.

Table 1. Typical values of the calibrated fit parameters for the focal sampling V0 .

Parameter Value Standard deviation

z0 0.95 µm 0.04 µm

r0 0.49 µm 0.06 µm

rb 0.17 µm 0.03 µm

Ab 0.37 0.09

γ 0.25 0.04

At the single particle level, scaling γ promotes linearly to N. With increasing concentration the coherent contribution becomes significant and the influence of γ on N diminishes. Scaling all three characteristic lengths by ε scales D by ε 2 . Independent errors in the characteristics lengths and/or the parameter Ab may lead to errors in D and/or N, but certainly worsen the fit

#214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21951

quality because the shape of the model correlation function differs from the measurement. For the two-species mixture measurements, the same procedure is first applied to solutions of single species to extract also the diffusion coefficients D1 and D2 . In order to calculate the relative brightness of the two species, the average signal hI0 i in the focal volume (V0 in Fig. 3(e)) averaged on 10 measurements of 20 seconds is calculated (same conditions for the two species: concentration of 150 pM; illumination power of 5 mW). The brightness was then estimated by subtracting the background: Q = hI0 i − hBi. The extracted parameters are then used for fitting the auto-correlations shown in Fig. 6(a) using Eq. (2) with only N1 and N2 as free fit parameters. Errors scaling the absolute concentrations N of all species equally do not affect the species fractions P. Errors scaling the diffusion constants D of all species equally also have little effect on P. However, erroneous shapes of the single-species correlation functions affect P because the fit will find a better match of the measured correlation function at an erroneous P. Hence, all errors in characteristic lengths and/or the parameters Ab and γ have little effect on P unless the shape of the model function changes. Obviously, an error in the relative brightness Q is critical as it affects the weights Q2 P and QP in Eq. (2). 3. 3.1.

Experimental results Improvement of the signal-to-noise ratio

The OCCS interferometer contains a high NA objective to enhance the collection efficiency and a variable aperture Fdet to adjust the NA of the detection mode. In order to demonstrate the improvement in signal detection, we characterized the SNR in function of the detection NA. Following Leitgeb et al. [30], the SNR in an optical coherence microscope is given as

2 I (4) SNR = m2 , σN where hIm i is the signal amplitude in sampling volume Vm and σN2 is the noise variance. The noise variance is evaluated by calculating the root mean square of a signal-free region within the tomogram at higher depths to avoid cover slip artifacts. This signal-free region may contain some background from multiply scattered light. Numerical Aperture

SNR

0.11 0.14 0.17

500 200 100 50 20 10 5 2 1

1

1.25 1.5

0.28

0.44

2 2.5 3 4 5 Diaphragm Opening [mm]

0.89

6

7 8

Fig. 4. SNR measurements with different diaphragm Fdet settings and corresponding detection NA. The solid line starts at the SNR measured with 1mm opening and follows the increase of the solid detection angle.

We measured 109 nm PS MSs diffusing freely in water with different diaphragm Fdet openings (concentration of 330 pM; illumination power of 5 mW; 10 measurements during 50 seconds for every opening value; line rate of 20 kHz with an exposure time of 48 µs). The SNR #214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21952

was calculated with respect to the averaged signal hI0 i in the sampling volume V0 . The noise variance was calculated at V50 (17.5 µm from the center of V0 ) well outside the DOF but close enough to the optical focus. Figure 4 shows the results of the measurements with respect to the Fdet opening. The NA of the detection mode is calculated from this opening diameter. The diaphragm opening of 1.25 mm diameter corresponds to the near-infrared light OCCS described in [6]. The solid line represents the calculated solid angle seen by the detection fiber Ω = 2π [1 − cos (arcsin (NA/n))] normalized by the SNR at the smallest opening. The measured SNR improved proportionally with the solid detection angle Ω. At intermediate NA, the measured SNR was about twice the expected value, most likely due to a higher back-scattered signal at intermediate detection angles. In contrast, the SNR improved only slightly at high NA, which could be caused by residual spherical aberration becoming particularly marked at high NA. Due to the diaphragm opening increase to 8 mm, the SNR improved by a factor of up to 28× as compared to the 1.25 mm diameter Fdet used in [6]. 3.2.

Instrument validation with gold NPs

We validated this novel visible light OCCS instrument by measuring freely diffusing gold NPs with different NP diameters. Figure 5 summarizes the results of these experiments. The normalized auto-correlations in V0 (see Fig. 3(e)) of differently sized gold NPs are shown in Fig. 5(a) (NPs concentrations: 150 pM except 35 pM for 50 nm; 60 pM for 40 nm. Illumination powers: 2 mW for 50nm; 5 mW for 40 nm, 30 nm; 10 mW for 20 nm, 15 nm, 10 nm). These NPs were freely diffusing in glycerol/water solutions with 60% w/w of glycerol. The diffusion coefficients were extracted from the signal in the sampling volumes in the focal area from V−2 to V2 independently and then averaged. These sampling volumes offered the best SNR and their small size allowed measurements at the few particles regime at the applied NPs concentrations. Extracting the diffusion coefficients was based on the fitting procedure described in the data analysis section. The fit residuals were small (< 2%) for NPs diameters ≥ 20 nm but increasing for smaller NPs due to low SNR. The extracted diffusion coefficients in Fig. 5(b) are related by the Stokes-Einstein relation. The viscosity of the glycerol/water solutions was obtained by measuring 109 nm PS MSs in this mixture. The deviation of the diffusion coefficient from the absolute theoretical values varies between 3.5% and 12% (smallest NPs). The improved SNR allows characterizing gold NPs of 10 nm which is 3× smaller than with the previous OCCS instrument [6]. Because the back-scattered field scales with the NP volume, this finding confirms an ∼27× higher sensitivity. 3.3.

Measurements of gold NPs/PS MSs mixtures

OCCS allowed measuring the fractions P1 of gold NPs and P2 of PS MSs in a two component solution. Figure 6 shows the results of freely diffusing 40 nm gold NPs mixed with 109 nm PS MSs. The normalized auto-correlations Gmix F,0 (τ) with varying fractions P1 from 0% to 100% of 40 nm gold NPs are shown in Fig. 6(b). The total particle concentration was kept constant at 150 pM and the host medium was pure water. The diffusion coefficient of each species was first extracted separately from the single species measurements using Eq. (1). For the fraction analysis, the 40 nm gold NPs brightness was defined as Q1 = 1 and a relative brightness Q2 = 0.86 for the 109 nm PS MSs was measured. A fit per sampling volume to the extended auto-correlation function Gmix F,m (τ) (Eq. (2)) for the five sampling volumes V−2 to V2 was used for extracting the fractions (see data analysis). The residuals of <4% underline the quality of these fits. Figure 6(b) shows that the measured average fractions P1 increase linearly with the prepared fractions P1 . The red line represents the theoretical values, while the dotted black line is the linear regression of the measured P1 . For the solutions with only one type of particles, fitting Eq. (2) for two species never totally removed the contribution for the other particle. Noise

#214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21953

and residual mismatches between the fit model and the measured auto-correlations caused the fit to find up to 8% of the other type of particles. The measured fractions P1 had an uncertainty of ≈ 6% and an absolute error < ±8%. Varying the characteristic lengths and the parameter Ab within the mean values ± the standard deviation given in Table 1 worsened neither the uncertainty nor the linearity significantly. However, the absolute errors at P1 = 0% and 100% increased to ±20% in the worst case. In all cases, the two NPs species could be clearly distinguished. 3.4.

Protein adsorption on SPIONs

We measured the adsorption of BSA onto SPIONs with OCCS. Naked SPIONs (maghemite phase of 8 nm±2 nm crystallite’s diameter) were preliminary coated with a mixture of polyvinyl alcohol (PVA) as described in previous works [31, 32] (these PVA-coated SPIONs will simply be called SPIONs for the remainder of this document). Their hydrodynamic radius RH in a PBS buffer (phosphate buffered saline, Sigma-Aldrich Chemie) as measured by OCCS yielded a mean diameter of 42.0 nm with a standard deviation of 4.7 nm (20 measurements lasting 20 seconds each; line rate of 50 kHz with an exposure time of 18 µs; illumination power of 5 mW). These SPIONs were also measured with DLS (Malvern Zetasizer, using DTS Nano 7.02 software) and a mean diameter of 43.1 nm with a standard deviation of 13.5 nm was obtained (measured polydispersity index = 0.24). The mean values are in close agreement but the standard deviation is greater with DLS. The OCCS standard deviation was estimated with the different measurements, each based on approximately 1000 particle transits. In DLS the standard deviation is calculated on the NPs population directly. The BSA at a neutral pH can be approximated as an equilateral triangular prism with sides of 8.4 nm and a height of 3.15 nm [33]. The hydrodynamic radius RH (N) of the SPIONs versus

(a)

(b)

1

50nm

60

40nm

0.8

50

20nm 15nm

0.4

D [ µm2 /s]

G( τ)

30nm 0.6

10nm

0.2

Residuals

0

40 30 20 10

0.02 0 −0.02

10 0.1ms

1ms

10ms 0.1s Lag time τ

1s

10s

20 30 40 NPs diameter [nm]

50

Fig. 5. (a) Normalized OCCS curves from gold NPs in V0 of different diameters in glycerol/water solutions with 60% w/w of glycerol. Thin lines with markers show the average auto-correlations in V0 of 10 measurements lasting 50 seconds each (line rate of 20 kHz with an exposure time of 48 µs). Thick lines show the fits using Eq. (1) and the residuals. (b) The extracted diffusion coefficients from V−2 to V2 matched well with the theoretical values.

#214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21954

(a)

(b) Au 40 0%

1

Au 40 40% Au 40 60%

0.6

Measured fraction P1

G( τ)

1

Au 40 20%

0.8

Au 40 80%

0.4

Au 40 100%

0.2

Residuals

0 0.04 0.02 0 −0.02 −0.04

0.8 0.6 0.4 0.2 0

0.1ms

1ms

10ms 0.1s Lag time τ

1s

10s

0

0.2

0.4 0.6 0.8 Au NPs 40nm prepared fraction P1

1

Fig. 6. OCCS of mixtures with two particle species. (a) Normalized OCCS auto-correlation curves from mixtures with different fractions of 40 nm gold NPs and 109 nm PS MSs in V0 . Thin lines with markers show the average auto-correlations in V0 of 10 measurements lasting 20 seconds each (line rate of 50 kHz with an exposure time of 18 µs; illumination power of 5 mW). Thick lines show the fits using Eq. (2) and the residuals. (b) The measured fraction of the species P1 versus the mixed fraction P1 . The red line represents the prepared values. The dotted black line is the linear regression of the measured P1 .

the number N of adsorbed proteins can be described by √ RH (N) = RH (0) 3 1 + cN,

(5)

where c = Vp /Vb is the ratio of the volume Vp of the protein and the volume Vb of the bare NP [13]. The SPIONs RH (0) = 21.0 nm yields Vb = 38 800 nm3 . Estimating the molecular volume of BSA from the size of the triangular prism we obtain Vp = 96 nm3 and hence c = 2.5 · 10−3 . The measurements were performed for 330 pM SPIONs in PBS buffer solutions containing BSA concentrations ranging from 330 nM to 33 µM (20 measurements lasting 20 seconds each; line rate of 50 kHz with an exposure time of 18 µs; illumination power of 5 mW). Figure 7(a) shows the average OCCS auto-correlation curves in V0 for measurements of solutions with 3.3 µM BSA and without BSA. Figure 7(c) shows the distribution of the hydrodynamic radius that has been calculated for every sampling volumes ranging from V−2 to V2 on the measurements with 3.3 µM BSA and without BSA along with a fitted normal distribution. The difference between the extracted mean values of the distributions is ∆R = 3.1 nm, suggesting that BSA molecules form a monolayer with their larger face on the SPIONs surface. Using Eq. (5), the maximum number of adsorbed proteins Nmax is about 194 for the observed increase of RH by 3.1 nm. For comparison, the coverage of the surface of a sphere with radius RH (0) would require about 181 BSA molecules which is in good agreement with Nmax and further indicates the formation of a monolayer. Figure 7(b) shows the hydrodynamic radii of the NPs as a function of BSA concentration. The standard deviations are mainly due to the polydispersity of the SPIONs. Following the approach by Rocker et al. [13], the experimental

#214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21955

(a)

SPIONs in PBS SPIONs in PBS, 3.3µM BSA

1

G(τ)

0.8 0.6 0.4 0.2

(c)

0.04 0.02 0 −0.02 −0.04

(b) 35

0.1ms

1ms

10ms Lag time τ

0.1s

Probability of occurence

Residuals

0

1s

SPIONs radius [nm]

30

0.4

0.3

SPIONs in PBS SPIONs in PBS, 3.3 µM BSA

0.2

0.1

0 25

0 5 10 15 20 25 30 35 40 45 50 Radius [nm]

20

15 10nM

100nM 1µM BSA concentration

10µM

Fig. 7. Adsorption of BSA on SPIONs. (a) Normalized OCCS auto-correlation curves in V0 of SPIONs 330 pM with 3.3 µM BSA and without BSA. Thin lines with markers show the average auto-correlations in V0 of 20 measurements lasting 20 seconds each (line rate of 50 kHz with an exposure time of 18 µs; illumination power of 5 mW). Thick lines show the fitting results using Eq. (1) and the residuals. The extracted hydrodynamic radii were RH = 21.4 nm without BSA and 22.5 nm with BSA. (b) The hydrodynamic radii of the NPs versus BSA concentration. The thick red line corresponds to a fit with Eq. (5). (c) Distribution of hydrodynamic radii extracted from V−2 to V2 of the auto-correlations.

data is fitted to a modified Langmuir model: N = Nmax

1

h , 1 + KD / hCBSA i 0

(6)

where Nmax is the maximum number of proteins binding to the nanoparticle, hCBSA i is the 0 average concentration of BSA, h is the Hill coefficient, and KD is the dissociation coefficient, which quantifies the strength of the protein–NP interaction. The fit of the data gives a trend line, however the standard deviation of the Hill coefficient is rather elevated, which does not allow a definite statement on the cooperativeness of the BSA adsorption.

#214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21956

We observed a full coverage above approximately 3.3 µM BSA concentration. Using this result, we can estimate the maximal concentration of SPIONs in serum for achieving full coverage by serum albumin. The reference range for albumin concentrations in serum is approximately 35–50 mg/ml [34]. By taking the lower bound of 35 mg/ml (corresponding to 526 µM), the maximal concentration of SPIONs in serum for a full coverage is estimated at 53 nM. 4.

Conclusions

In comparison to the previous OCCS instrument working in the near-infrared wavelength range, we present an increase of the SNR by a factor of more than 25 with this visible light OCCS instrument. This improved SNR allowed the measurement of the diffusion coefficient of gold NPs down to a size of 10 nm diameter. The generalization of the fitting model to multiple component solutions has been applied to the measurement of the fractions of a mixture containing two different NPs species. For two NPs species with large difference in size (2.7:1) but similar brightness (1.2:1), we obtained excellent agreement of the measured fractions versus the prepared fractions. The measurement uncertainty and absolute error were small enough to qualify for an unambiguous distinction of the two NPs species. We also measured the protein adsorption and formation of a protein monolayer on SPIONs, i.e. a study of the protein corona on NPs. These adsorption measurements were performed under physiological conditions using an intrinsic signal from the SPIONs, thus preventing potential modifications of these NPs that can occur when using fluorescent labels or separation steps. This achievement opens the door to the study of adsorption of other proteins of interest on SPIONs, in particular to the study of protein adsorption in serum. In summary, the OCCS technique has been substantially enhanced and first applications highlight its potential for nanoscience and life sciences. Acknowledgments We acknowledge funding by the Swiss National Foundation (grant 205321L 135353) and support by the COST action MP1302 ”NanoSpectroscopy”.

#214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21957

1 Laboratoire

* [email protected]

Abstract: Optical coherence correlation spectroscopy (OCCS) allows studying kinetic processes at the single particle level using the backscattered light of nanoparticles. We extend the possibilities of this technique by increasing its signal-to-noise ratio by a factor of more than 25 and by generalizing the method to solutions containing multiple nanoparticle species. We applied these improvements by measuring protein adsorption and formation of a protein monolayer on superparamagnetic iron oxide nanoparticles under physiological conditions. © 2014 Optical Society of America OCIS codes: (290.5850) Scattering, particles; (030.1670) Coherent optical effects; (170.0170) Medical optics and biotechnology; (170.4500) Optical coherence tomography.

References and links 1. P. Ghosh, G. Han, M. De, C. Kim, and V. Rotello, “Gold nanoparticles in delivery applications,” Adv. Drug Delivery Rev. 60, 1307–1315 (2008). 2. H. Jans and Q. Huo, “Gold nanoparticle-enabled biological and chemical detection and analysis,” Chem. Soc. Rev. 41, 2849–2866 (2012). 3. A. Gupta, R. Naregalkar, V. Vaidya, and M. Gupta, “Recent advances on surface engineering of magnetic iron oxide nanoparticles and their biomedical applications,” Nanomedicine 2, 23–39 (2007). 4. Q. Pankhurst, N. Thanh, S. Jones, and J. Dobson, “Progress in applications of magnetic nanoparticles in biomedicine,” J. Phys. D: Appl. Phys. 42, 1–15 (2009). 5. R. Singh and H. Nalwa, “Medical applications of nanoparticles in biological imaging, cell labeling, antimicrobial agents, and anticancer nanodrugs,” J. Biomed. Nanotechnol. 7, 489–503 (2011). 6. S. Broillet, A. Sato, S. Geissbuehler, C. Pache, A. Bouwens, T. Lasser, and M. Leutenegger, “Optical coherence correlation spectroscopy (occs),” Opt. Express 22, 782–802 (2014). 7. J. Robinson, T. Takizawa, D. Vandr´e, and R. Burry, “Ultrasmall immunogold particles: Important probes for immunocytochemistry,” Microsc. Res. Tech. 42, 13–23 (1998). 8. K. Huang, H. Ma, J. Liu, S. Huo, A. Kumar, T. Wei, X. Zhang, S. Jin, Y. Gan, P. Wang, S. He, X. Zhang, and X.-J. Liang, “Size-dependent localization and penetration of ultrasmall gold nanoparticles in cancer cells, multicellular spheroids, and tumors in vivo,” ACS Nano 6, 4483–4493 (2012). 9. T. Cedervall, I. Lynch, S. Lindman, T. Berggard, E. Thulin, H. Nilsson, K. Dawson, and S. Linse, “Understanding the nanoparticle-protein corona using methods to quntify exchange rates and affinities of proteins for nanoparticles,” Proc. Natl. Acad. Sci. U. S. A. 104, 2050–2055 (2007). 10. U. Sakulkhu, M. Mahmoudi, L. Maurizi, J. Salaklang, and H. Hofmann, “Protein corona composition of superparamagnetic iron oxide nanoparticles with various physico-chemical properties and coatings,” Sci. Rep. 4, 1–9 (2014). 11. P. Aggarwal, J. Hall, C. McLeland, M. Dobrovolskaia, and S. McNeil, “Nanoparticle interaction with plasma proteins as it relates to particle biodistribution, biocompatibility and therapeutic efficacy,” Adv. Drug Delivery Rev. 61, 428–437 (2009).

#214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21944

12. C. Walkey and W. Chan, “Understanding and controlling the interaction of nanomaterials with proteins in a physiological environment,” Chem. Soc. Rev. 41, 2780–2799 (2012). 13. C. Rocker, M. Potzl, F. Zhang, W. Parak, and G. Nienhaus, “A quantitative fluorescence study of protein monolayer formation on colloidal nanoparticles,” Nat. Nanotechnol. 4, 577–580 (2009). 14. S. Dominguez-Medina, S. McDonough, P. Swanglap, C. Landes, and S. Link, “In situ measurement of bovine serum albumin interaction with gold nanospheres,” Langmuir 28, 9131–9139 (2012). 15. I. Kohli, S. Alam, B. Patel, and A. Mukhopadhyay, “Interaction and diffusion of gold nanoparticles in bovine serum albumin solutions,” Appl. Phys. Lett. 102, 203705 (2013). 16. M. Mikhaylova, D. Kim, C. Berry, A. Zagorodni, M. Toprak, A. Curtis, and M. Muhammed, “Bsa immobilization on amine-functionalized superparamagnetic iron oxide nanoparticles,” Chem. Mater. 16, 2344–2354 (2004). 17. Z. Wang, T. Yue, Y. Yuan, R. Cai, C. Niu, and C. Guo, “Kinetics of adsorption of bovine serum albumin on magnetic carboxymethyl chitosan nanoparticles,” Int. J. Biol. Macromol. 58, 57–65 (2013). 18. M. Villiger and T. Lasser, “Image formation and tomogram reconstruction in optical coherence microscopy,” J. Opt. Soc. Am. A 27, 2216–2228 (2010). 19. R. A. Leitgeb, M. Villiger, A. H. Bachmann, L. Steinmann, and T. Lasser, “Extended focus depth for fourier domain optical coherence microscopy,” Opt. Lett. 31, 2450–2452 (2006). 20. C. Pache, N. Bocchio, A. Bouwens, M. Villiger, C. Berclaz, J. Goulley, M. Gibson, C. Santschi, and T. Lasser, “Fast three-dimensional imaging of gold nanoparticles in living cells with photothermal optical lock-in optical coherence microscopy,” Opt. Express 20, 21385–21399 (2012). 21. M. Villiger, C. Pache, and T. Lasser, “Dark-field optical coherence microscopy,” Opt. Lett. 35, 3489–3491 (2010). 22. J. Izatt and M. Choma, Optical Coherence Tomography: Technology and Applications (Springer Verlag, Berlin, 2008). 23. M. Wojtkowski, V. Srinivasan, T. Ko, J. Fujimoto, A. Kowalczyk, and J. Duker, “Ultrahigh-resolution, highspeed, fourier domain optical coherence tomography and methods for dispersion compensation,” Opt. Express 12, 2404–2422 (2004). 24. C. Dorrer, N. Belabas, J.-P. Likforman, and M. Joffre, “Spectral resolution and sampling issues in fouriertransform spectral interferometry,” J. Opt. Soc. Am. B 17, 1795–1802 (2000). 25. T. Wohland, R. Rigler, and H. Vogel, “The standard deviation in fluorescence correlation spectroscopy,” Biophys. J. 80, 2987–2999 (2001). 26. N. L. Thompson, Fluorescence correlation spectroscopy. In: Topics in Fluorescence Spectroscopy, vol. 1 (Plenum Press, 1991). 27. U. Meseth, T. Wohland, R. Rigler, and H. Vogel, “Resolution of fluorescence correlation measurements,” Biophys. J. 76, 1619–1631 (1999). 28. B. Berne and R. Pecora, Dynamic Light Scattering with Applications to Chemistry, Biology and Physics. Chapter 5: Model Systems of Spherical Molecules (John Wiley and Sons, 1976). 29. M. Leutenegger, R. Rao, R. Leitgeb, and T. Lasser, “Fast focus field calculations,” Opt. Express 14, 11277–11291 (2006). 30. R. Leitgeb, C. Hitzenberger, and A. Fercher, “Performance of fourier domain vs. time domain optical coherence tomography,” Opt. Express 11, 889–894 (2003). 31. M. Chastellain, A. Petri, and H. Hofmann, “Particle size investigations of a multistep synthesis of pva coated superparamagnetic nanoparticles,” J. Colloid Interface Sci. 278, 353–360 (2004). 32. L. Maurizi, U. Sakulkhu, L. Crowe, V. Dao, N. Leclaire, J.-P. Vallee, and H. Hofmann, “Syntheses of cross-linked polymeric superparamagnetic beads with tunable properties,” RSC Adv. 4, 11142–11146 (2014). 33. M. Ferrer, R. Duchowicz, B. Carrasco, J. De La Torre, and A. Acuna, “The conformation of serum albumin in solution: A combined phosphorescence depolarization-hydrodynamic modeling study,” Biophys. J. 80, 2422– 2430 (2001). 34. H. Kenneth Walker, W. Dallas Hall, and J. Willis Hurst, Clinical Methods, 3rd edition; The History, Physical, and Laboratory Examinations; Chapter 101: Serum Albumin and Globulin (Butterworths, 1990).

1.

Introduction

Nanoparticles (NPs) are studied thoroughly for applications in medicine and life sciences: of special interest are gold NPs due to their optical, chemical, and biocompatible properties [1, 2] and superparamagnetic iron oxide nanoparticles (SPIONs) [3–5]. Recently, we have shown that optical coherence correlation spectroscopy (OCCS) can measure the diffusion coefficient of gold NPs down to diameters of 30 nm with near-infrared light [6]. There are advantages to use even smaller gold NPs because of an enhanced labeling efficiency [7] and improved penetration into biological samples [8], showing the importance of measuring the dynamics of smaller gold NPs.

#214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21945

Another very interesting feature would be to have the possibility to distinguish different species of scattering NPs, for instance based on their size. This would give the possibility to distinctively label several targets in measurements and to measure chemical reaction kinetics, for instance. The coating process of nanoparticles (NPs) with biomolecules generates a so-called protein corona [9, 10]. Understanding the properties of this coating is important for the application of NPs in living organisms [11]. Most strategies involve the removal of the NP–protein complex from their original physiological environment [12]. A complementary approach investigates the adsorption of proteins under physiological conditions. So far, the adsorption of either bovine serum albumin (BSA) or human serum albumin (HSA) has been studied with different types of NPs using fluorescence correlation spectroscopy (FCS), as for example the coating process of FePt NPs and quantum dots [13] or gold NPs [14, 15]. Adsorption of BSA on SPIONs has already been studied [16, 17], however, a quantitative analysis of their diffusion properties in physiological conditions has not been performed yet. OCCS exploits the backscattered light of NPs illuminated by a broadband light source, giving simultaneous access to several sampling volumes along the optical axis. Over a large concentration range, this technique allows the extraction of the diffusion coefficient and concentration of NPs using an auto-correlation analysis. In this work we greatly extend the application potential of OCCS by shifting the OCCS instrument from the near-infrared to the visible wavelength range. The signal-to-noise ratio (SNR) has been increased by a factor of more than 25 due to a much increased NA allowing the extraction of the diffusion coefficient of gold NPs down to a size of 10 nm diameter. In addition, the auto-correlation model has been generalized to solutions hosting multiple NP species and this generalized model is used to retrieve the particle fractions in a mixture of two NP species. Using our novel OCCS technique, we show its application to the study of BSA adsorption on SPIONs by measuring their hydrodynamic radius under physiological conditions. This combines the advantages of not having to use fluorescent labels (thus preventing the risk that these labels change the properties of the NPs) and of avoiding additional separation steps to remove unbound protein from the solution. 2. 2.1.

Materials and methods Low coherence interferometer

Figure 1 depicts the OCCS instrument which is based on a Mach-Zehnder interferometer [6, 18–21]. This visible light OCCS instrument is based on low coherence interferometric principles as described in [6]. The light from a super-continuum source (Koheras SuperK Extreme, NKT Photonics) is spectrally filtered to obtain a broad visible spectrum centered at 590 nm wavelength with a bandwidth of 160 nm full width at half-maximum (FWHM). The light is linearly polarized by a polarizing beam splitter and delivered to the interferometer by a singlemode fiber (mode field diameter of 3.4 µm, Fibercore). After collimation, the light is split into a reference beam and an illumination beam by a non-polarizing 50:50 beam splitter. The reference beam passes through glass prisms with variable thicknesses for compensating the dispersion difference in the illumination and sample arms. The illumination beam passes through an axicon lens (176◦ cone angle, Asphericon) generating a Bessel beam, which is further imaged by a 300 mm achromatic lens and an apochromatic water immersion objective (40×, NA 1.15, Olympus) into the sample at the focal plane of the objective. The power impinging on the sample is adjusted between 2–10 mW, depending on the experiment. The annular mask Fill blocks residual stray light from the tip of the axicon. The back-scattered sample field is collected by the objective, superimposed with the reference field and focused into a singlemode fiber (mode field diameter of 3.4 µm, Thorlabs). The fiber guides the collected field to the customized spectrometer (transmission grating; 600 lines/mm, Wasatch) for the registration #214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21946

L2

L3

Detection

Axicon

Fdet

Illumination

Reference

Illumination fiber

L1

Beam splitter

40x/1.15NA

Sample

Detection fiber

Fill

Dispersion compensation

Fig. 1. Visible light OCCS interferometer with a Bessel-Gauss configuration [18]. The axicon generates a Bessel beam illumination, whereas the detection mode is Gaussian. The apertures Fill and Fdet provide a dark field contrast [21].

of the spectral interferogram from 420 nm to 760 nm wavelength (linear CMOS array Sprint spL2048-140km with 2048 pixels, Basler). As illustrated in Fig. 1, the illumination and detection beams do not overlap in the aperture plane of the objective: any specular reflected light from the glass slide is blocked by the diaphragm Fdet . The interferometer was conceived with a minimum number of optical elements in order to increase the detection efficiency and to simplify the alignment of components. Because the glass slide is always closer to the objective than the sample, a dark field is obtained even though the filters Fill and Fdet are not placed in a pupil plane but just as close to the objective as possible. The dark field contrast is essential for a high SNR while detecting the weak back-scattered field from the NPs. This interferometer implements a Bessel illumination and Gaussian detection configuration [18]. The illumination field corresponds to a radial zero-order Bessel distribution in the focal plane with the first minimum located at 0.22 µm lateral radius, whereas the detection mode is Gaussian with an effective numerical aperture (NA) close to 0.9. In typical experiments the detector was driven at a line rate of 20 kHz with an exposure time of 48 µs. The reference arm intensity was adjusted to about 75% of the available dynamic range of the camera. By taking the Fourier transform of the spectrum (resampled at equidistant wave numbers, λ to k mapping), we obtain the time-dependent signal traces extending over several sampling volumes [22]. Here, the background corrected depth profile containing a sequence of sampling volumes (center to center distance of 0.35 µm in water) was then extracted by using an algorithm based on the discrete Fourier transform (DFT). The signal processing leading to the time-dependent signal traces is indicated in Fig. 2. An integration kernel to get the tomogram from the sample spectrum by a matrix product was calculated such that (i) it resampled the spectrum at equidistant wave numbers (λ to k mapping); (ii) it compensated the residual dispersion φ (k) by multiplying the spectrum with calibrated phase factors exp(i φ (k)) calculated with an automatic numerical iterative procedure developed by Wojtkowski et al. [23]; and (iii) Fourier transformed the corrected spectrum to yield the tomogram in (iv) the region of interest.

#214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21947

t

I(z) tim et

Background subtraction

λ k mapping

Dispersion compensation

FFT

Selection of the z indices

depth z

wavevelength λ

I(λ) tim e

DFT

Modulus

Fig. 2. Signal processing leading to the time-dependent signal traces. The interference signal is recorded via a spectrometer. The average spectrum is then subtracted from the recorded signal (background subtraction). The new DFT integration kernel replaces the following steps: (i) λ to k mapping; (ii) dispersion compensation; (iii) fast Fourier transform of the spectrum; and (iv) selection of the sampling volumes of interest. The modulus of the signal is taken to yield the time-dependent signal traces.

Thereby, the matrix multiplication of this kernel with the measured spectra directly yielded the depth profiles. This algorithm is only performed on the sampling volumes in the optical focal area in order to minimize the calculation time. In all the calculations that follow, 150 sampling volumes are calculated. Auto-correlations of the absolute values of these time-dependent signals were then calculated. In this application, the new DFT algorithm is about six times faster than the previous FFT algorithm [6]. In detail, the DFT kernel F is an M × N matrix, where M is the number of tomogram samples m ∈ [0, N) and N is the number of interferogram samples. The tomogram is obtained as matrix product I(z) = F · I(λ ). In other words, I(zm ) = ∑N−1 n=0 Fmn I(λn ), where zm = m∆z = m·0.35 µm. The fast Fourier transform is performed by Fmn = exp(−2i πmn/N) for m, n ∈ {0, N − 1}. In the DFT algorithm, the dispersion is compensated by multiplying the interferogram samples with a factor fn = exp (i φ (λn )). An extra amplitude component would allow reshaping the spectral amplitudes. Therefore, the kernel implementing steps (ii–iv) writes as Fmn = fn exp(−2i πmn/N). The λ to k mapping (i) is performed by multiplying this kernel with an N × N identity matrix resampled at wavelengths corresponding to equidistant wave number samples. A high-quality resampling is achieved by Fourier-upsampling the columns of the identity matrix to yield L & 3N samples, which are then linearly interpolated at the desired wavelengths [24]. 2.2.

Sample preparation

Sample solutions of monodisperse polystyrene microspheres (PS MSs) with a diameter of 109 nm (POLYBEAD Microspheres 109 nm, Polysciences) and gold colloids (gold NPs) with diameters of 10, 15, 20, 30, 40 and 50 nm (EM.GC10, EM.GC15, EM.GC20, EM.GC30, EM.GC40, EM.GC50, British Biocell International) were used. To obtain the nominal concentrations, the solutions were prepared by diluting an adequate volume of stock solution in ultrapure water or in glycerol. All OCCS measurements were performed in plastic wells (µSlide 8 well, uncoated, sterile, Ibidi) with a well volume of 300 µl.

#214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21948

2.3.

Correlation analysis of multiple components mixtures

The signal acquisition in OCCS and its auto-correlation analysis for single species was described in previous work [6], where the interested reader finds the detailed derivation of Eq. (1). The auto-correlation analysis of the OCCS signal Im (t) in a sampling volume m was subdivided into three distinct regimes depending on the number of particles in the sampling volume of interest N: the single particle (N 1), the few particles (N ∼ 1) and the many particles (N 1) regimes. The general case is the few particles regime, which takes into account the diffusion of particles in and out of the sampling volume and a dynamic light scattering (DLS) contribution depending on the number of particles N in the sampling volume. In the few particles regime, the OCCS autocorrelation function can be written as r −1 τ τ τ τ γ 1+ 1+ 1 + Ab exp − 1 + N exp − , (1) GF,m (τ) = N τxy τz τb τc where γ is the volume contrast [25] and N the average number of particles in the sampling volume. τxy = r02 (4D)−1 and τz = z20 (4D)−1 are the lateral and axial diffusion times, where D is the diffusion coefficient, and r0 and z0 are the lateral and axial e−2 radii of an equivalent threedimensional Gaussian volume. The exponential decay τb = rb2 (4D)−1 with the characteristic length rb and amplitude parameter Ab account for the Bessel profile. The last factor is the coherent contribution caused by speckle fluctuations of the back-scattered field of the NPs in one sampling volume. This coherent contribution is characterized by the decorrelation time τc = (8n2 k02 D)−1 known from dynamic light scattering (DLS), where n is the refractive index of the medium and k0 is the central wave number of the source spectrum. Here we extend this initial model to the study of multiple independently diffusing species i with different diffusion coefficients. Without the coherent contribution, Eq. (1) describes a correlation function for a signal similar to the intensity fluctuations in FCS. Therefore, we can apply the multi-species model derived for FCS, see for instance Thompson [26] and Meseth et al. [27]. For the coherent contribution, Berne and Pecora [28] describe a multi-species model in the context of DLS. In both cases, the multi-species models are linear combinations of the single-species correlation functions. Therefore, we can write the multi-species auto-correlation function as hBi 2 ∑ Ni Q2i DF,m,i (τ) mix GF,m (τ) = γ 1 − , (2) hIm i (∑ Ni Qi )2 where DF,m,i (τ) =

1+

τ τxy,i

r 1+

τ τz,i

−1 τ τ 1 + Ab exp − 1 + Ni exp − . (3) τb,i τc,i

hIm i = hBi + ∑ Ni Qi is the average signal in the sampling volume m and hBi is the average background. Ni is the average number and Qi is the brightness of particles of species i. Qi is proportional to the absolute value of the particle’s electric field reflectivity. The squared prefactor stands for the background influence on the correlation amplitude. The total number of particles is N = ∑ Ni and Pi = Ni /N are the fractions of the species. Each species contributes to the correlation amplitude by its concentration times the squared brightness which limits the discrimination of species with significant brightness differences. 2.4.

Sampling volume characterization

The determination of r0 and z0 is essential for an appropriate fit model requiring an accurate characterization of the OCCS sampling volumes. In analogy to the near-infrared light #214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21949

OCCS procedures described in [6], we characterized the sampling volume by imaging individual 50 nm gold NPs. The scattering particles were dispersed in an agarose gel with 0.3% weight/volume ratio. We scanned individual NPs using a three-axis piezoscanning stage in the x-y plane (repeatability 0.01 µm, Physik Instrumente) while illuminating with a power of 5 mW. We compared the brightness profile measurements with ab initio calculations at the cen-

(a)

y

Signal along x,y [a.u.]

(b)

1

0.4 0.2 0

x

(e)

Signal [a.u.]

1

Signal along z [a.u.]

(d)

z

Au 50nm

0.6

x (c)

Calculation

0.8

0

0.1

1

0.6

0.7

Coh. gate

0.8

Au 50nm

0.6 0.4 0.2 0 −2

−20

0.2 0.3 0.4 0.5 Radial distance ρ [µm]

−10

−1.5

−1

−0.5 0 0.5 Depth z [ µm]

Sampling volumes Vm 0 10 20 30

40

1

1.5

2

50

Measurement FWHM=7.1 µm Calculation FWHM=3.3 µm

0.8 0.6 0.4 0.2 0

−5

0

5 10 Depth z [ µm]

15

Fig. 3. Brightness profiles characterization. (a,c) Brightness profile cross-sections xy and xz p measured with a 50 nm single gold NP. Scalebars: 1 µm. (b,d) Average radial ρ = x2 + y2 and axial (z) brightness profiles measured on ten 50 nm gold NPs in comparison with calculations. (e) Depth of field characterization with 109 nm PS MSs freely diffusing in water (concentration: 330 pM; illumination power: 5 mW; average on 10 measurements of 50 seconds). The useful DOF is indicated by the dashed lines showing the measured FWHM. V0 corresponds to the focal sampling volume. The calculations in (b) and (e) were performed using the focus field calculation framework by Leutenegger et al. [29]. The coherence gate in (d) is calculated from the measured source spectrum S(k) using the Wiener-Khintchine relation.

#214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21950

tral wavelength using the focus field calculation framework by Leutenegger et al. [29]. The brightness profile is rotationally symmetric around the optical axis z. Figures 3(a)–3(d) show the measured brightness profile W (r). Figure 3(b) shows the measured radial brightness profiles W (r) compared with the calculation performed at the central wavelength of 590 nm. We attribute the non-zero minimum (10% of the maximal signal) and larger profile in the measurement to the large illumination bandwidth and to the finite size of the NPs used. Figure 3(d) shows the measured axial brightness profiles W (r) compared with the coherence gate determined by the measured source spectrum S(k) and calculated using the Wiener-Khintchine relation. The e−2 half width z0 of this calculated axial coherence gate is 0.95 µm. There is a good agreement between the calculation and the measurement in the profiles’ FWHM. The difference in the tail is < 15% of the maximal signal and is due to residual dispersion difference between the reference path and the illumination plus detection paths. The depth of field (DOF) assessment is important for the evaluation of the number of useful sampling volumes along the optical axis. For assessing this parameter, we used a solution of 109 nm PS MSs (concentration of 330 pM) freely diffusing in water, an illumination power of 5 mW and measured the time-averaged signal amplitude hIm i for the sampling volumes Vm along the optical axis. Figure 3(e) compares the DOF (average over 10 measurements during 50 seconds) with our ab initio calculations. The measurements indicate a DOF of approximately 7 µm FWHM corresponding to 20 sampling volumes, which can be observed simultaneously in axial separation steps of ∆z = 0.35 µm. The significant increase in DOF as compared to calculations is most likely caused by longitudinal chromatic aberrations in the illumination and detection paths. 2.5.

Data Analysis

Care should be taken to carefully calibrate several parameters summarized in Table 1. Figure 3(d) confirms the good agreement between the calculated and the measured axial brightness profile. Therefore, the axial extent z0 was extracted from the coherence length given by the spectrum of the light source in Fig. 3(d). The volume contrast factor γ was calculated from the measured PSF [25]. The lengths r0 , rb and the parameter Ab were calibrated for each sampling volume with an auto-correlation measurement of 109 nm PS MSs matched to Eq. (1) by using the theoretical value for the diffusion coefficient D (concentration of 330 pM; illumination power of 5 mW; 10 measurements of 50 seconds). This calibration was repeated for every measurement session. The characteristic lengths r0 , rb , z0 as well as the parameters Ab and γ were then kept fixed and Eq. (1) was used to fit the auto-correlations shown in Fig. 5(a) and Fig. 7(a), where only N and D have been retained as free fit parameters. The hydrodynamic radii RH of particles in Fig. 7(b) and Fig. 7(c) were calculated using the Stokes-Einstein relation D = kB T /(3πη2RH ), where kB is the Boltzmann constant, T the absolute temperature (295 K), and η the fluid viscosity.

Table 1. Typical values of the calibrated fit parameters for the focal sampling V0 .

Parameter Value Standard deviation

z0 0.95 µm 0.04 µm

r0 0.49 µm 0.06 µm

rb 0.17 µm 0.03 µm

Ab 0.37 0.09

γ 0.25 0.04

At the single particle level, scaling γ promotes linearly to N. With increasing concentration the coherent contribution becomes significant and the influence of γ on N diminishes. Scaling all three characteristic lengths by ε scales D by ε 2 . Independent errors in the characteristics lengths and/or the parameter Ab may lead to errors in D and/or N, but certainly worsen the fit

#214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21951

quality because the shape of the model correlation function differs from the measurement. For the two-species mixture measurements, the same procedure is first applied to solutions of single species to extract also the diffusion coefficients D1 and D2 . In order to calculate the relative brightness of the two species, the average signal hI0 i in the focal volume (V0 in Fig. 3(e)) averaged on 10 measurements of 20 seconds is calculated (same conditions for the two species: concentration of 150 pM; illumination power of 5 mW). The brightness was then estimated by subtracting the background: Q = hI0 i − hBi. The extracted parameters are then used for fitting the auto-correlations shown in Fig. 6(a) using Eq. (2) with only N1 and N2 as free fit parameters. Errors scaling the absolute concentrations N of all species equally do not affect the species fractions P. Errors scaling the diffusion constants D of all species equally also have little effect on P. However, erroneous shapes of the single-species correlation functions affect P because the fit will find a better match of the measured correlation function at an erroneous P. Hence, all errors in characteristic lengths and/or the parameters Ab and γ have little effect on P unless the shape of the model function changes. Obviously, an error in the relative brightness Q is critical as it affects the weights Q2 P and QP in Eq. (2). 3. 3.1.

Experimental results Improvement of the signal-to-noise ratio

The OCCS interferometer contains a high NA objective to enhance the collection efficiency and a variable aperture Fdet to adjust the NA of the detection mode. In order to demonstrate the improvement in signal detection, we characterized the SNR in function of the detection NA. Following Leitgeb et al. [30], the SNR in an optical coherence microscope is given as

2 I (4) SNR = m2 , σN where hIm i is the signal amplitude in sampling volume Vm and σN2 is the noise variance. The noise variance is evaluated by calculating the root mean square of a signal-free region within the tomogram at higher depths to avoid cover slip artifacts. This signal-free region may contain some background from multiply scattered light. Numerical Aperture

SNR

0.11 0.14 0.17

500 200 100 50 20 10 5 2 1

1

1.25 1.5

0.28

0.44

2 2.5 3 4 5 Diaphragm Opening [mm]

0.89

6

7 8

Fig. 4. SNR measurements with different diaphragm Fdet settings and corresponding detection NA. The solid line starts at the SNR measured with 1mm opening and follows the increase of the solid detection angle.

We measured 109 nm PS MSs diffusing freely in water with different diaphragm Fdet openings (concentration of 330 pM; illumination power of 5 mW; 10 measurements during 50 seconds for every opening value; line rate of 20 kHz with an exposure time of 48 µs). The SNR #214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21952

was calculated with respect to the averaged signal hI0 i in the sampling volume V0 . The noise variance was calculated at V50 (17.5 µm from the center of V0 ) well outside the DOF but close enough to the optical focus. Figure 4 shows the results of the measurements with respect to the Fdet opening. The NA of the detection mode is calculated from this opening diameter. The diaphragm opening of 1.25 mm diameter corresponds to the near-infrared light OCCS described in [6]. The solid line represents the calculated solid angle seen by the detection fiber Ω = 2π [1 − cos (arcsin (NA/n))] normalized by the SNR at the smallest opening. The measured SNR improved proportionally with the solid detection angle Ω. At intermediate NA, the measured SNR was about twice the expected value, most likely due to a higher back-scattered signal at intermediate detection angles. In contrast, the SNR improved only slightly at high NA, which could be caused by residual spherical aberration becoming particularly marked at high NA. Due to the diaphragm opening increase to 8 mm, the SNR improved by a factor of up to 28× as compared to the 1.25 mm diameter Fdet used in [6]. 3.2.

Instrument validation with gold NPs

We validated this novel visible light OCCS instrument by measuring freely diffusing gold NPs with different NP diameters. Figure 5 summarizes the results of these experiments. The normalized auto-correlations in V0 (see Fig. 3(e)) of differently sized gold NPs are shown in Fig. 5(a) (NPs concentrations: 150 pM except 35 pM for 50 nm; 60 pM for 40 nm. Illumination powers: 2 mW for 50nm; 5 mW for 40 nm, 30 nm; 10 mW for 20 nm, 15 nm, 10 nm). These NPs were freely diffusing in glycerol/water solutions with 60% w/w of glycerol. The diffusion coefficients were extracted from the signal in the sampling volumes in the focal area from V−2 to V2 independently and then averaged. These sampling volumes offered the best SNR and their small size allowed measurements at the few particles regime at the applied NPs concentrations. Extracting the diffusion coefficients was based on the fitting procedure described in the data analysis section. The fit residuals were small (< 2%) for NPs diameters ≥ 20 nm but increasing for smaller NPs due to low SNR. The extracted diffusion coefficients in Fig. 5(b) are related by the Stokes-Einstein relation. The viscosity of the glycerol/water solutions was obtained by measuring 109 nm PS MSs in this mixture. The deviation of the diffusion coefficient from the absolute theoretical values varies between 3.5% and 12% (smallest NPs). The improved SNR allows characterizing gold NPs of 10 nm which is 3× smaller than with the previous OCCS instrument [6]. Because the back-scattered field scales with the NP volume, this finding confirms an ∼27× higher sensitivity. 3.3.

Measurements of gold NPs/PS MSs mixtures

OCCS allowed measuring the fractions P1 of gold NPs and P2 of PS MSs in a two component solution. Figure 6 shows the results of freely diffusing 40 nm gold NPs mixed with 109 nm PS MSs. The normalized auto-correlations Gmix F,0 (τ) with varying fractions P1 from 0% to 100% of 40 nm gold NPs are shown in Fig. 6(b). The total particle concentration was kept constant at 150 pM and the host medium was pure water. The diffusion coefficient of each species was first extracted separately from the single species measurements using Eq. (1). For the fraction analysis, the 40 nm gold NPs brightness was defined as Q1 = 1 and a relative brightness Q2 = 0.86 for the 109 nm PS MSs was measured. A fit per sampling volume to the extended auto-correlation function Gmix F,m (τ) (Eq. (2)) for the five sampling volumes V−2 to V2 was used for extracting the fractions (see data analysis). The residuals of <4% underline the quality of these fits. Figure 6(b) shows that the measured average fractions P1 increase linearly with the prepared fractions P1 . The red line represents the theoretical values, while the dotted black line is the linear regression of the measured P1 . For the solutions with only one type of particles, fitting Eq. (2) for two species never totally removed the contribution for the other particle. Noise

#214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21953

and residual mismatches between the fit model and the measured auto-correlations caused the fit to find up to 8% of the other type of particles. The measured fractions P1 had an uncertainty of ≈ 6% and an absolute error < ±8%. Varying the characteristic lengths and the parameter Ab within the mean values ± the standard deviation given in Table 1 worsened neither the uncertainty nor the linearity significantly. However, the absolute errors at P1 = 0% and 100% increased to ±20% in the worst case. In all cases, the two NPs species could be clearly distinguished. 3.4.

Protein adsorption on SPIONs

We measured the adsorption of BSA onto SPIONs with OCCS. Naked SPIONs (maghemite phase of 8 nm±2 nm crystallite’s diameter) were preliminary coated with a mixture of polyvinyl alcohol (PVA) as described in previous works [31, 32] (these PVA-coated SPIONs will simply be called SPIONs for the remainder of this document). Their hydrodynamic radius RH in a PBS buffer (phosphate buffered saline, Sigma-Aldrich Chemie) as measured by OCCS yielded a mean diameter of 42.0 nm with a standard deviation of 4.7 nm (20 measurements lasting 20 seconds each; line rate of 50 kHz with an exposure time of 18 µs; illumination power of 5 mW). These SPIONs were also measured with DLS (Malvern Zetasizer, using DTS Nano 7.02 software) and a mean diameter of 43.1 nm with a standard deviation of 13.5 nm was obtained (measured polydispersity index = 0.24). The mean values are in close agreement but the standard deviation is greater with DLS. The OCCS standard deviation was estimated with the different measurements, each based on approximately 1000 particle transits. In DLS the standard deviation is calculated on the NPs population directly. The BSA at a neutral pH can be approximated as an equilateral triangular prism with sides of 8.4 nm and a height of 3.15 nm [33]. The hydrodynamic radius RH (N) of the SPIONs versus

(a)

(b)

1

50nm

60

40nm

0.8

50

20nm 15nm

0.4

D [ µm2 /s]

G( τ)

30nm 0.6

10nm

0.2

Residuals

0

40 30 20 10

0.02 0 −0.02

10 0.1ms

1ms

10ms 0.1s Lag time τ

1s

10s

20 30 40 NPs diameter [nm]

50

Fig. 5. (a) Normalized OCCS curves from gold NPs in V0 of different diameters in glycerol/water solutions with 60% w/w of glycerol. Thin lines with markers show the average auto-correlations in V0 of 10 measurements lasting 50 seconds each (line rate of 20 kHz with an exposure time of 48 µs). Thick lines show the fits using Eq. (1) and the residuals. (b) The extracted diffusion coefficients from V−2 to V2 matched well with the theoretical values.

#214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21954

(a)

(b) Au 40 0%

1

Au 40 40% Au 40 60%

0.6

Measured fraction P1

G( τ)

1

Au 40 20%

0.8

Au 40 80%

0.4

Au 40 100%

0.2

Residuals

0 0.04 0.02 0 −0.02 −0.04

0.8 0.6 0.4 0.2 0

0.1ms

1ms

10ms 0.1s Lag time τ

1s

10s

0

0.2

0.4 0.6 0.8 Au NPs 40nm prepared fraction P1

1

Fig. 6. OCCS of mixtures with two particle species. (a) Normalized OCCS auto-correlation curves from mixtures with different fractions of 40 nm gold NPs and 109 nm PS MSs in V0 . Thin lines with markers show the average auto-correlations in V0 of 10 measurements lasting 20 seconds each (line rate of 50 kHz with an exposure time of 18 µs; illumination power of 5 mW). Thick lines show the fits using Eq. (2) and the residuals. (b) The measured fraction of the species P1 versus the mixed fraction P1 . The red line represents the prepared values. The dotted black line is the linear regression of the measured P1 .

the number N of adsorbed proteins can be described by √ RH (N) = RH (0) 3 1 + cN,

(5)

where c = Vp /Vb is the ratio of the volume Vp of the protein and the volume Vb of the bare NP [13]. The SPIONs RH (0) = 21.0 nm yields Vb = 38 800 nm3 . Estimating the molecular volume of BSA from the size of the triangular prism we obtain Vp = 96 nm3 and hence c = 2.5 · 10−3 . The measurements were performed for 330 pM SPIONs in PBS buffer solutions containing BSA concentrations ranging from 330 nM to 33 µM (20 measurements lasting 20 seconds each; line rate of 50 kHz with an exposure time of 18 µs; illumination power of 5 mW). Figure 7(a) shows the average OCCS auto-correlation curves in V0 for measurements of solutions with 3.3 µM BSA and without BSA. Figure 7(c) shows the distribution of the hydrodynamic radius that has been calculated for every sampling volumes ranging from V−2 to V2 on the measurements with 3.3 µM BSA and without BSA along with a fitted normal distribution. The difference between the extracted mean values of the distributions is ∆R = 3.1 nm, suggesting that BSA molecules form a monolayer with their larger face on the SPIONs surface. Using Eq. (5), the maximum number of adsorbed proteins Nmax is about 194 for the observed increase of RH by 3.1 nm. For comparison, the coverage of the surface of a sphere with radius RH (0) would require about 181 BSA molecules which is in good agreement with Nmax and further indicates the formation of a monolayer. Figure 7(b) shows the hydrodynamic radii of the NPs as a function of BSA concentration. The standard deviations are mainly due to the polydispersity of the SPIONs. Following the approach by Rocker et al. [13], the experimental

#214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21955

(a)

SPIONs in PBS SPIONs in PBS, 3.3µM BSA

1

G(τ)

0.8 0.6 0.4 0.2

(c)

0.04 0.02 0 −0.02 −0.04

(b) 35

0.1ms

1ms

10ms Lag time τ

0.1s

Probability of occurence

Residuals

0

1s

SPIONs radius [nm]

30

0.4

0.3

SPIONs in PBS SPIONs in PBS, 3.3 µM BSA

0.2

0.1

0 25

0 5 10 15 20 25 30 35 40 45 50 Radius [nm]

20

15 10nM

100nM 1µM BSA concentration

10µM

Fig. 7. Adsorption of BSA on SPIONs. (a) Normalized OCCS auto-correlation curves in V0 of SPIONs 330 pM with 3.3 µM BSA and without BSA. Thin lines with markers show the average auto-correlations in V0 of 20 measurements lasting 20 seconds each (line rate of 50 kHz with an exposure time of 18 µs; illumination power of 5 mW). Thick lines show the fitting results using Eq. (1) and the residuals. The extracted hydrodynamic radii were RH = 21.4 nm without BSA and 22.5 nm with BSA. (b) The hydrodynamic radii of the NPs versus BSA concentration. The thick red line corresponds to a fit with Eq. (5). (c) Distribution of hydrodynamic radii extracted from V−2 to V2 of the auto-correlations.

data is fitted to a modified Langmuir model: N = Nmax

1

h , 1 + KD / hCBSA i 0

(6)

where Nmax is the maximum number of proteins binding to the nanoparticle, hCBSA i is the 0 average concentration of BSA, h is the Hill coefficient, and KD is the dissociation coefficient, which quantifies the strength of the protein–NP interaction. The fit of the data gives a trend line, however the standard deviation of the Hill coefficient is rather elevated, which does not allow a definite statement on the cooperativeness of the BSA adsorption.

#214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21956

We observed a full coverage above approximately 3.3 µM BSA concentration. Using this result, we can estimate the maximal concentration of SPIONs in serum for achieving full coverage by serum albumin. The reference range for albumin concentrations in serum is approximately 35–50 mg/ml [34]. By taking the lower bound of 35 mg/ml (corresponding to 526 µM), the maximal concentration of SPIONs in serum for a full coverage is estimated at 53 nM. 4.

Conclusions

In comparison to the previous OCCS instrument working in the near-infrared wavelength range, we present an increase of the SNR by a factor of more than 25 with this visible light OCCS instrument. This improved SNR allowed the measurement of the diffusion coefficient of gold NPs down to a size of 10 nm diameter. The generalization of the fitting model to multiple component solutions has been applied to the measurement of the fractions of a mixture containing two different NPs species. For two NPs species with large difference in size (2.7:1) but similar brightness (1.2:1), we obtained excellent agreement of the measured fractions versus the prepared fractions. The measurement uncertainty and absolute error were small enough to qualify for an unambiguous distinction of the two NPs species. We also measured the protein adsorption and formation of a protein monolayer on SPIONs, i.e. a study of the protein corona on NPs. These adsorption measurements were performed under physiological conditions using an intrinsic signal from the SPIONs, thus preventing potential modifications of these NPs that can occur when using fluorescent labels or separation steps. This achievement opens the door to the study of adsorption of other proteins of interest on SPIONs, in particular to the study of protein adsorption in serum. In summary, the OCCS technique has been substantially enhanced and first applications highlight its potential for nanoscience and life sciences. Acknowledgments We acknowledge funding by the Swiss National Foundation (grant 205321L 135353) and support by the COST action MP1302 ”NanoSpectroscopy”.

#214988 - $15.00 USD Received 1 Jul 2014; revised 21 Aug 2014; accepted 21 Aug 2014; published 3 Sep 2014 (C) 2014 OSA 8 September 2014 | Vol. 22, No. 18 | DOI:10.1364/OE.22.021944 | OPTICS EXPRESS 21957

*When life gives you a hundred reasons to cry, show life that you have a thousand reasons to smile*

© Copyright 2015 - 2021 PDFFOX.COM - All rights reserved.