Benchtop SS-OCT – layout and performance evaluation


Benchtop SS-OCT – layout and performance evaluation  
José P. Dominguesa,b, Susana F. Silvab, Rui Bernardesb and A.M. Morgadoa,b 

a Department of Physics, University of Coimbra, Coimbra, Portugal 
bIBILI, Faculty of Medicine, University of Coimbra, Coimbra, Portugal 

jpd@uc.pt   susana.f.s.89@gmail.com   rmbernardes@fmed.uc.pt   miguel@fis.uc.pt 

 

Abstract — Optical coherence tomography (OCT) is a non-

invasive biomedical imaging technique that provides high speed 

and high resolution three dimensional and cross sectional images 

of biological samples, in vivo and in situ. OCT applications 

targeting small animals is believed to bring developments in 

medical techniques, instruments, diagnosis and therapies for a 

number of human diseases as always have been the case of animal 

experimentation. 

With the swept source OCT (SS-OCT) system presented in this 

work, we were able to achieve performance parameters that meet 

the requirements to image the retina of small animals. 

Performance characteristics include 105 dB for system 

sensitivity, a roll-off below 1 dB/mm over 3 mm depth and an 

axial resolution of 8 µm. We describe the layout and 

acquisition/processing solutions towards fast imaging of in vivo 

samples. 

 
Keywords— Swept Source OCT, Imaging, Instrumentation. 

I. INTRODUCTION 

Since its first appearance in the early 1990’s, Optical 

Coherence Tomography (OCT) was recognized as a very 

helpful tool for ophthalmology. Based on the optical 

interference phenomenon, the OCT is capable of producing 

high-resolution cross-sectional images of non-homogeneous 

samples, among which the biological tissue, being particularly 

tailored to image ocular structures such as the retina. Over the 

last twenty years, OCT has experienced a fast and steady 

growth. It now offers a wide range of applications, is clinically 

very well accepted and is the subject of active research 

worldwide. The technique and theory behind it – in all possible 

configurations, being it Time-Domain OCT (TD-OCT), 

Spectral-Domain OCT (SD-OCT) or Swept-Source OCT (SS-

OCT) – is extensively described in the literature [1][2][3][4]. 

SS-OCT has been appointed as the most promising 

technology for OCT imaging by offering higher scanning 

speed, reduced sensitivity roll-off and better overall 

performance, when associated with balanced detection,[3] in 

comparison to time and Fourier/spectral domain approaches. 

OCT systems specifically developed to image small animals, 

used as physiological models of disease, are fundamental to 

test and develop new medical therapies. Our group is engaged 

in the development of a dedicated SS-OCT platform for small 

animals based on the most recent technological advances. It 

should allow testing different configurations and explore new 

concepts. The basic layout has already been presented [5] 

 

 

along with simulations and preliminary performance 

evaluation. In this paper, we report recent developments 

regarding in vitro performance and discuss key details with 

respect to acquisition/scanning synchronism.  

The motivation for this development is definitely the 

possibility of building a valuable tool for research using animal 

models of disease. 

In the field of biomedical research, small animals are very 

often used to develop, validate and test new techniques and 

therapies. OCT imaging can provide researchers with the 

means to understand physiology, pathology and phenotypes in 

health and disease.  

Furthermore, this benchtop OCT system can be used as a 

platform for the continuing development and test of new OCT 

components, instrumentation and methods. 

II. MATERIALS AND METHODS 

In its simpler formulation, the OCT can be considered as an 

open-air Michelson interferometer (Fig 1) with a beam-

splitter, reference and sample paths and a Gaussian-shaped 

broadband source. 

 

Fig. 1 Michelson interferometer schematics based OCT. 

Light coming from the source can be represented by its 

electric field wave component Ein expressed as a complex 

exponential [1][2]: 

𝐸𝑖𝑛 = 𝑠(𝑤)𝑒
𝑖(𝑤𝑡−𝑘𝑧) (1) 

 

i-ETC: ISEL Academic Journal of Electronics, 
Telecommunications and Computers 
CETC2016 Issue, Vol. 3, n. 1 (2017) ID-8 

http://journals.isel.pt 

mailto:jpd@uc.pt
mailto:susana.f.s.89@gmail.com
file:///C:/Users/JPD/Documents/rmbernardes@fmed.uc.pt
file:///C:/Users/JPD/Documents/miguel@fis.uc.pt


where s(w) is the source amplitude spectrum, w the angular 

frequency, k the wavenumber and z the distance along the 

propagation direction. Phase appears due to different 

interactions throughout the interferometer both for the 

reference and the sample arms, respectively: 

𝐸𝑅 = (𝑇𝑟 𝑇𝑠)
1

2⁄ 𝐸𝑖𝑛 𝑒
−𝑖2𝑘𝑛𝑧𝑅  (2) 

𝐸𝑆 = (𝑇𝑟 𝑇𝑠 )
1

2⁄ 𝐸𝑖𝑛 𝑒
−𝑖2𝑘𝑛𝑧𝑆  (3) 

where n is the refraction index, zR and zS are the path lengths 

in the reference and sample arms and the factor of 2 arises from 

the double pass of light. Tr and Ts are the beam-splitter 

reference and sample arm intensity transmission coefficients. 

The sample is not a simple reflecting surface and has a 

multilayered structure with each layer interface having a 

reflectance, r, given by the Fresnel’s equations. 

The internal sample refraction index structure is revealed by 

the Eout field that results from superposition (interference) of 

ES and ER. 

Optical detectors measure Irradiance, which is proportional 

to the square of the amplitude of the E field. The detector 

output photocurrent, I(w,k), is proportional to the time average 

of the Irradiance and can be written as: 

𝐼(𝑤, 𝑘) = 𝑙𝑖𝑚
𝑇→∞

1

2𝑇
𝜌 ∫ 𝐸𝑜𝑢𝑡 𝐸𝑜𝑢𝑡

∗𝑇

−𝑇
𝑑𝑡 (4) 

where ρ represents detector responsivity. Photocurrent appears 

as a sum of a self-interference term (arising from source self- 

interference and autocorrelation from the multiple reflectors of 

the sample) and a cross-interference term. 

𝐼(𝑘) = 𝐼0 + 𝜌𝑆(𝑘)√𝑅𝑟 𝑅𝑆𝑛 𝑐𝑜𝑠2𝑘(𝑧𝑟 − 𝑧𝑆𝑛 ) (5) 

I0, also called the dc term, accounts for all self-interference 

contributions. The second term, obtained by using the Euler’s 

formula, is the cross-interference term, which contains the 

information on the sample structure. R is the power reflectance 

and subscripts Sn and r refer to sample layer n and reference 

mirror respectively.  

Following Time-domain OCT, the technique evolved to the 

Fourier-domain OCT where the interference information is 

spatially encoded (SD-OCT), requiring a detector array, or 

time encoded (SS-OCT). In this case, a single-element detector 

is enough but the broadband source must be wavelength-swept 

and the data acquisition system must handle high acquisition 

rates. In the Fourier domain, the reference plan is fixed, an 

important advantage over Time-domain OCT, which requires 

mechanical sweeping of the reference mirror position. The 

layout presented here corresponds to fiber optic based SS-

OCT. 

III. EXPERIMENTAL APPARATUS 

A. OCT layout 

The basic SS-OCT layout, composed of interferometer, 

scanning and acquisition system, is schematically depicted in 

Fig 2. The fundamental blocks of the system are: an Axsun 

Laser Swept-Source AXP50125-3 1060 nm (Axsun 

Technologies, Billerica, MA, USA), with central 1060 nm 

wavelength, 110 nm bandwidth and sweep frequency of 100 

kHz; a 400 MSPS multi I/O acquisition board (X5-400M – 

Innovative Integration, California, USA) with two A/D and 

two D/A channels; an InGaAs balanced amplified 

photodetector (PDB471C – Thorlabs) and a X-Y 

positioning/scanning system (GSV002 Galvo Scanning 

System, Thorlabs GmbH, Germany). Optical components such 

as fiber cables, attenuators, couplers, polarization controllers, 

collimators and focusing objectives complete the system. 

 

 
Fig. 2 – Overall SS-OCT Layout 

 

The scanning, for cross-sectional (B-scan) and volume scans, 

is ensured by a X and Y galvanometer/mirror system which is 

controlled, respectively, through a precision waveform 

generator (AFG3101 Arbitrary Function Generator, 

Tektronix) and a DAC board (NI 6010, National Instruments, 

Austin, Texas). Digitized fringe data undergoes Fourier 

analysis to produce individual A-scans (along the beam 

direction), which can be composed in final B-scan images 

(XZ) and/or volumes (XYZ). These tasks are carried out by 

dedicated acquisition/control and processing software. 

Configuration parameters in the software can be tuned 

according to the needs: the scanning angles, the number of A-

scans per B-scan and number of B-scans per volume. After 

acquisition, the whole volume can be stored for posterior 

analysis or discarded. 

Overall timing parameters that determine the A-scan rate, 

A/D conversion sampling time and scanning positioning are 

obtained from two output signals provided by the laser source: 

SWEEP TRIGGER and K-CLOCK. Fig 3 shows the 

characteristics of the trigger and clock signals as well as the 

optical power at the useful sweep range. 

B. Software 

A customizable software was developed using object-

oriented programming (Microsoft Visual C++/IDE) for the 64-

bit Microsoft Windows 7 operating system resorting to 

libraries from Innovative Technologies to deal with data 

acquisition and hardware control. 

J.P. Domingues  et al. | i-ETC -  CETC2016 Issue, Vol. 3, n. 1 (2017) ID-8

i-ETC: ISEL Academic Journal of Electronics, Telecommunications and Computers 



 
Fig. 3 – Typical spectral power, trigger and clock data for 

the Axsun source (from Axsun OCT source manual Rev. 09) 

 

With the current hardware implementation, 1510 data points 

are digitalized in less than 5 µs, which correspond to the 

number of sample wavelengths outputted by the laser source. 

A full frequency sweep corresponds to the entire A-scan that 

can be computed owing to the Fourier transform. The number 

of useful data points is, nevertheless, restricted to 1376, with 

the remaining ones produced by the dummy clock outputted 

by the laser source. 

TABLE 1 

TYPICAL SCAN PARAMETERS (FROM AXSUN OCT 

SOURCE MANUAL REV. 09) 

Parameter Value 

Wavelength Range 985.0-1095.0 nm 

Sweep Frequency 100 kHz 

Maximum Samples 1510 

Selected Number of Samples 1376 

Duty Cycle 45% 

 

Synchronization between the acquisition and laser emission 

is paramount for SS-OCT. As such, the laser source has an 

embedded Mach-Zehnder interferometer (MZI) to provide an 

uniform clock signal in the wavenumber space (k-space). This 

allows direct A/D sampling of the optical detector signal event 

though the wavelength sweep of the laser output is non-linear. 

The period of the k-clock signal is defined by the free spectral 

range of MZI. Linearized fringe signals with equal k-spacing 

can be achieved by clocking the high-speed A/D channel of the 

acquisition board with the clock provided by the source. The 

Fourier transform analysis can then be directly applied to the 

acquired data. Moreover, the laser source also provides a 

trigger signal which is connected to the SYNC port of the 

acquisition board. This signal is responsible for starting the I/O 

module. Fig 4 shows a typical result of an A-Scan obtained 

over a 1 mm thickness glass. 

 

Fig. 4 - Interferogram and FFT spectrum for 1mm thick 

coverglass. Peaks represent air-glass and glass-air interfaces. 

IV. PERFORMANCE ANALYSIS 

The performance of the system was established based on 

commonly parameters found on the literature [6][7]: axial 

resolution (AR), sensitivity (S), dynamic range (DR) and depth 

sensitivity roll-off. Preliminary results have been obtained 

with current setup which compares and surpass the required 

values for biomedical applications which are listed in table 2.  

TABLE 2 

MINIMUM VALUES FOR THE MAIN OCT PARAMETERS IN 

BIOMEDICAL APPLICATIONS.[6] [7] 

 

 

Parameter Value 

Axial Resolution < 10 𝜇𝑚 

Sensitivity Fall-Off 20 𝑑𝐵 over 2 𝑚𝑚 depth 

Dynamic Range 40 − 50 𝑑𝐵 

Sensitivity > 95𝑑𝐵 

The Point Spread Function (PSF), the system’s response to a 

point object (or impulse), corresponds, for the described OCT 

system, to the Fourier transform of the interference signal due 

to a single, perfect reflector. This is accomplished using a gold 

mirror as sample. The PSF of the OCT system is shown in 

Fig 5. 

 

J.P. Domingues  et al. | i-ETC -  CETC2016 Issue, Vol. 3, n. 1 (2017) ID-8

i-ETC: ISEL Academic Journal of Electronics, Telecommunications and Computers 



 

Fig. 5 – Point Spread Function using a gold mirror reflector 

A. Axial Resolution 

Axial or depth resolution is one of the key parameters to 

evaluate an OCT system performance. It is defined as the 

shortest interface separation that can be measured along the 

light beam´s propagation direction. The theoretical value of the 

axial resolution, usually specified in the air and assuming a 

Gaussian laser beam power spectrum, can be calculated as 

follows. Being  

𝑆(𝑘) =
1

𝜎𝑘 √2𝜋
exp [−

(𝑘−𝑘0)
2

2𝜎𝑘
2 ] (6) 

the Gaussian power spectrum and 

𝛥𝑘 = (2√2 ln(2))𝜎𝑘 (7) 

its full width at half maximum (FWHM), its Fourier transform 

is given by 

𝛾(𝑧) =
1

√𝜋
exp [−

𝛥𝑘2(𝑧−𝑧0)
2

16ln(2)
] (8) 

This is the coherence function whose FWHM corresponds to 

the coherence length of the laser beam. If we used the well 

established criterium of defining the axial resolution as half of 

the source coherence length, we obtain for our system 

(λ0=1060 nm and Δλ=100 nm). 

𝑧 − 𝑧0 = 𝛥𝑧 =
2𝑙𝑛 (2)𝜆0

2

𝜋𝛥𝜆
= 4.51 µ𝑚 (9) 

Actually, the power spectrum of our laser source does not 

show a Gaussian profile. As such, the PSF is experimentally 

determined by fitting a Lorentzian function to the measured 

PSF profile (A-scan) when using a gold mirror sample (Fig 6). 

For our system we obtained an axial resolution of 8.1 µm, in 

the air, which corresponds to 6.2 µm in tissue. 

B. Sensitivity 

Sensitivity is related with the smallest sample reflectivity, 

Rs,min, that can be detected, defined as the signal level when 

SNR=1. In dB units: 

𝑆𝑑𝐵 = 10𝑙𝑜𝑔10
𝑅𝑠

𝑅𝑠,𝑚𝑖𝑛
 (10) 

 

 

Fig. 6 – Experimental axial resolution measurement using 

curve fitting to the gold mirror PSF 

One way of experimentally measure S is to use a gold mirror 

(Rs=1) on the sample side. Thus, (6) becomes: 

𝑆𝑑𝐵 = 10𝑙𝑜𝑔10
1

𝑅𝑠,𝑚𝑖𝑛
 (11) 

In terms of the measured quantities, sensitivity of the system 

can be estimated from the ratio between the output from the 

optimal reflector (the system point spread function, PSF) and 

that from the noise (no sample), the latter being defined as the 

standard deviation of the readings, leading to:  

𝑆𝑑𝐵 = 20𝑙𝑜𝑔10
𝑃𝑆𝐹𝑝𝑒𝑎𝑘

𝜎𝑛𝑜𝑖𝑠𝑒
 (12) 

Notice that the factor 2 arises from the square correlation 

between the photodetector current and the reflectance, R, that 

is: 𝑃𝑆𝐹 ∝ 𝐼𝐷 (𝑧) ∝ √𝑅 . 

One way to experimentally determine the sensitivity is to 

place a neutral density filter of a known optical density (OD) 

before the gold mirror. Light will then be doubly attenuated by 

the filter, resulting in a total attenuation of 20×OD (dB). 

Added to (12) results in:  

𝑆𝑑𝐵 = 20𝑙𝑜𝑔10
𝑃𝑆𝐹′𝑝𝑒𝑎𝑘

𝜎𝑛𝑜𝑖𝑠𝑒
+ 20×𝑂𝐷 (13) 

In these conditions, a sensitivity of 105 dB was 

experimentally determined, allowing to calculate Rs,min from 

(11). 

 

C. Dynamic Range 

Dynamic range (DR) regards the ratio between the maximum 

and the minimum reflectivity signal that can be measured 

within the same A-scan. In OCT, the smallest reflectivity is 

taken as the standard deviation of the noise floor when we 

measure the PSF using a gold coated mirror as sample. This 

way the expression is similar to (8) but all measurements refer 

to the same A-Scan:  

𝑆𝑑𝐵 = 20𝑙𝑜𝑔10
𝑃𝑆𝐹𝑝𝑒𝑎𝑘

𝜎𝑛𝑜𝑖𝑠𝑒
 (14) 

J.P. Domingues  et al. | i-ETC -  CETC2016 Issue, Vol. 3, n. 1 (2017) ID-8

i-ETC: ISEL Academic Journal of Electronics, Telecommunications and Computers 



Experimentally we measured the Dynamic Range always 

above 55 dB, a value better than the minimum acceptable for 

biomedical tissues (40-50). Better results were obtained with 

the inclusion of an electronic 100 MHz low-pass filter at the 

output of the balanced amplifier to remove high frequency 

noise. Although the presence of the filter limits the maximum 

measurement depth, this is not a problem for the experimental 

setup as the mirror position is far from the filter’s influence.  

In the case depicted in Fig 7 for the current setup we 

obtained:  

𝐷 = 20 log (
1627.9

1.1820
) = 62.78𝑑𝐵 (15) 

 
Fig. 7 - Ratio of maximum to the minimum reflectivity that 

can be measured simultaneously (same A-scan). 

D. Sensitivity Roll-off 

In OCT, a depth dependent sensitivity fall-off occurs 

independently of absorption or scattering from the sample. 

This fall-off, which places a limit to the maximum tissue depth 

that can be assessed, is primarily due to the washout of spectral 

fringes caused by the limited spectral sampling resolution of 

the spectrometer (in case of SD-OCT) or by the limited 

sweeping resolution and detector bandwidth (in case of SS-

OCT). The sensitivity roll-off of our system was evaluated by 

doing the sensitivity measurement procedure for different 

axial positions of the mirror, used as sample. The sensitivity 

roll-off with depth was assessed over an axial range of 4 mm 

to find it to be less than 1 dB/mm in the first 2.8 mm and about 

3 dB/mm thereafter as shown in Fig  8. 

 

V. CONCLUSIONS 

The presented layout and setup allowed to achieve the 

required performance for biomedical field and opens the 

possibility to future developments in the instrumentation and 

application sides. 

 

 
Fig. 8 – Sensitivity measurement over 4 mm depth range. 

The corner for a depth near 2.8 mm is due to the 100 MHz LP 

electronic filter. Higher electronic frequencies on the detector 

signal correspond to higher penetration depths. 

 

ACKNOWLEDGMENT 

This study was supported by The Portuguese Foundation for 

Science and Technology (PEst-UID/NEU/04539/2013) and by 

FEDER-COMPETE (POCI-01-0145- FEDER-007440). 

REFERENCES 

[1] Tomlins P H and Wang R K (2005) Theory, 
developments and applications of optical coherence 

tomography. Journal of Physics: Appl. Phys. 38, pp. 

2519-2535 DOI 10.1088/0022-3727/38/15/002 

[2] Targowski, P Wojtkowski M et al (2004) Complex 
spectral OCT in human eye imaging in vivo. Optics 

Communications 229 pp. 79-84. 

[3] Marschall S, Sander B et al (2011) Optical coherence 
tomography – current technology and applications in 

clinical and biomedical research. Anal Bioanal Chem 400 

pp. 2699-2720 DOI 10.1007/s00216-011-5008-1 

[4] Huang D, Swanson C et al (1991) Optical Coherence 
Tomography. Science, 254(5035) pp. 1178-1181 

[5] Silva, S., Domingues, J., Agnelo J, Morgado A, 
Bernardes R (2014), Development of an Optical 

Coherence Tomograph for Small Animal Retinal 

Imaging, IFMBE Proceedings Volume 41, , pp 419-422. 

DOI: 10.1007/978-3-319-00846-2_104 

[6] W. Wieser, B. R. Biedermann, T. Klein, C. M. 
Eigenwillig, and R. Huber, “Multi-megahertz OCT: High 

quality 3D imaging at 20 million A-scans and 4.5 

GVoxels per second.,” Opt. Express, vol. 18, no. 14, pp. 

14685–14704, 2010. 

[7] M. D Bayleyegn, H. Makhlouf et al (2012) Ultrahigh 
Resolution spectral-domain optical coherence 

tomography at 1.3 µm using a broadband 

superluminescent diode light source, Opt Commun., vol 

285, nº 24, pp 5564-5569  

J.P. Domingues  et al. | i-ETC -  CETC2016 Issue, Vol. 3, n. 1 (2017) ID-8

i-ETC: ISEL Academic Journal of Electronics, Telecommunications and Computers