Modeling the optical coupling across the anterior chamber of the

eye towards polarimetric glucose sensing

Casey W. Pirnstill and Gerard L. Cot
Department of Biomedical Engineering, 3120 TAMU
Texas A&M University, College Station, TX 77843

ABSTRACT
Millions of people worldwide are affected by diabetes. While glucose sensing technology has come a long way
over the past several decades, the current commercially available techniques are still invasive, often leading to poor
patient compliance. To minimize invasiveness, focus has been placed on optical techniques to ascertain blood glucose
concentrations. Optical polarimetry has shown promise and progress as a viable technique for glucose sensing. Recent
developments in polarimetric glucose sensing have been focused on overcoming time varying corneal birefringence due
to motion artifacts. Beyond corneal birefringence, the next hurdle toward making this approach viable is the ability to
couple polarized light across the eyes anterior chamber. The eye is ideally suited to couple light to the retina. The index
mismatch between the air and cornea is partially responsible for the beam bending toward the retina and, while good for
vision, it complicates our ability to couple light across the anterior chamber without using an index matching device
when performing polarimetric glucose monitoring. In this report, we have designed and modeled a non-index matched
coupling scheme constructed with commercially available optics. The optical ray tracing model was performed using
CODE V to verify the feasibility of a reflective based non-index matched coupling scheme with respect to index of
refraction and anatomical restraints. The ray tracing model was developed for a dual-wavelength system and the effect of
refraction and reflection at each optical interface within the setup was evaluated. The modeling results indicate a
reflective based optical coupling design could be added to existing polarimetric glucose systems thus removing the need
for placing an index matched eye-coupling mechanism over the eye prior to data collection.

Keywords: Optical Modeling, Polarization Modeling, Optical Polarimetry, Glucose Sensing, Optical Biosensors.

1. INTRODUCTION
Millions of people are affected by diabetes mellitus on an annual basis [1]. Research has indicated that to achieve
optimal glycemic control of glucose levels in subjects with diabetes frequent glucose monitoring is an essential element
[2]. Currently, there are a large number of discrete and continuous based glucose monitoring devices commercially
available to aid in guided treatment for diabetes [3]. Although these devices are extremely important for physiologic
glucose monitoring and have been advanced significantly over the past few decades, current commercial devices often
have poor patient compliance due to the invasive nature the measurement is obtained. To overcome the invasive nature
of obtaining blood glucose concentration and improve compliance in patients suffering from diabetes to monitor glucose
levels, several noninvasive optical approaches have been explored both in academia and industry over the past three
decades such as optical near and mid infrared spectroscopy (NIR and MIR) [4-6], optical coherence tomography (OCT)
[7], Raman spectroscopy [8], photoacoustic spectroscopy [9], fluorescence spectroscopy [10], and polarimetry [11-12].
The majority of the proposed methods utilizing the techniques listed above measure the glucose concentration of blood
or interstitial fluid. However, optical polarimetry is unique in that it probes the anterior chamber of the eye [12]. Our
group has proposed an optical polarimetric method for noninvasive glucose monitoring across the anterior chamber of
the eye [12-18].
To date, research on the use of optical polarimetry towards noninvasively monitoring glucose levels in the aqueous
humor have focused on overcoming time varying corneal birefringence, one major limiting factor of the technique.
However, in these studies index coupling devices are utilized to overcome the index mismatch between the air-aqueous
humor barrier [12-18]. It is well known through the process of vision that the eye is ideally suited to couple light to the
back of the retina. While this approach is useful in removing the index mismatch limitation from experimental testing for
evaluating different polarimetric system configurations to monitor glucose in the presence of motion induced
birefringence, it is not a practical approach for a final device design. Previous success in measuring aqueous glucose
levels over physiologic range using polarimetry were all obtained either in a cuvette or utilizing index matched coupling
devices to overcome the index mismatch of the air-cornea interface. Thus, the next major limitation to overcome toward
Optical Diagnostics and Sensing XIV: Toward Point-of-Care Diagnostics, edited by
Gerard L. Cot, Proc. of SPIE Vol. 8951, 895107 2014 SPIE
CCC code: 1605-7422/14/$18 doi: 10.1117/12.2040021
Proc. of SPIE Vol. 8951 895107-1
Downloaded From: http://proceedings.spiedigitallibrary.org/ on 04/18/2014 Terms of Use: http://spiedl.org/terms
feasibility of polarimetric glucose monitoring is the ability to couple polarized light across the anterior chamber of the
eye without the use of a coupling device.
Several research groups have modeled the optical bending of light as it passes through the anterior chamber [19-
21]. Malik et al. produced optical simulations and calculations utilizing CODE V optical design software (Optical
Research Associates, Pasadena, CA) and MATLAB (The MathWorks, Natick, MA) [19]. The results produced in the
optical simulations indicated that light can be coupled through the anterior chamber over a narrow range in input angles
of the incident light relative to the entrance corneal surface. Previous models have indicated that light could be coupled
through the anterior chamber of the eye and out of the adjacent corneal surface with the input beam becoming a
divergent fanning beam shape at the exiting corneal surface.
In this report, we have further modeled light coupling through the anterior chamber of the eye in an unmatched
index environment and evaluated the addition of incoming and outgoing optics to provide a collimated beam both at the
entrance corneal surface and at the detection optics in a polarimetric setup. Utilizing the model previously published

by
our group, a computer aided drawing (CAD) rendering in Fig. 1 below shows that by using glancing angles with a round
mirror at 45 degrees (a small mirror at the nose and a larger off-axis parabolic collecting mirror at the outside of the
skull) we can potentially overcome the index mismatch and propagate polarized light across the anterior chamber of the
eye. All parameters for the eye were obtained from the literature [19-23].

2. MATERIALS AND METHODS
To evaluate a non-index matched optical coupling scheme, models were designed and verified using Code V, an
optical modeling software package, evaluating various coupling schemes and optical component variations. For the eye
model, the corneal parameters and dimensions utilized in this report were the same as described in detail previously by
Malik et al. [19]. In the final design, shown below, we use a 90 degree off-axis parabolic mirror to collect and re-
collimate the light beam after exiting the cornea. The parabolic mirror is needed because the beam shapes, after traveling
through the aqueous humor, will fan out as shown in Fig. 1; thus, the beams need to be re-collimated in order for it to
travel through the receiving optics and arrive onto the detectors. As depicted in Figs. 2-4, we have developed a CAD
rendering of the model similar to that shown in Fig. 1 in which all the optical simulations performed in CODE V optical
design software (Optical Research Associates, Pasadena, CA). The initial variables adjusted were the input mirror angle
and entry point of the beam as well as the output parabolic mirror angle and distance. Optimization of the model was
then achieved by adjusting the coupling optics position and angle in three dimensions to ensure the two optical beams
undergo roughly the same optical path as they pass through the polarimetric system. An additional 90 degree off-axis
parabolic mirror was added after the first parabolic mirror at the output of the eye to help ensure collimation of the beam
before entering the detection optics. The posterior and anterior corneal surfaces were modeled as spheres centered on the
optic axis of the eye with radii of 7.7 mm and 6.8 mm, respectively. The central thickness of the cornea was set to be 0.5
mm and the thickness increased monotonically (as a conicoidal function) to 1.97 mm at 6 mm away from the center [21].
The index of refraction at the outside of the cornea was assumed to be that of air (n=1.00). The refractive indices of
cornea and aqueous humor were set to values of 1.376 and 1.336, respectively (i.e. |n|=1.6 x 10
-3
). The smoothness of
cornea is defined by the formation of a liquid tear film over the posterior surface which makes the cornea behave similar
to a polished optical lens [22]. As indicated in the optical simulation produced by Malik et al. the effect of motion
artifact can be minimized when the beam is positioned in the region of the entrance near the midpoint of the corneal apex
and limbus (i.e. 1.5 mm below the apex) [19]. For optical ray tracing of each wavelength, the beam diameter was chosen
to be fixed at a diameter of 1 mm to be consistent with the model utilized previously by our group [19]. The CODE V
software package allows for a polarization ray tracing method to be employed, which solves for the path/surface
interactions through the optical system for each wavelength, further details of which have been described elsewhere [23].
CODE V has the ability to divide the optical surface into a rectangular or circular grid like pattern. The user then has the
ability to define the birefringence parameters such as the direction of fast axis and the birefringence (neno, where ne and
no are the extraordinary and ordinary refractive indices, respectively) for each individual grid section [24].
Proc. of SPIE Vol. 8951 895107-2
Downloaded From: http://proceedings.spiedigitallibrary.org/ on 04/18/2014 Terms of Use: http://spiedl.org/terms

Fig. 1: Modeling of the optical path through the anterior chamber of the eye using the new optical coupling method. Note that light
still has to be incident at a relatively glancing angle with respect to the posterior corneal surface and there is no visible difference (on
the current scale) between the optical paths taken by the two beams at different wavelengths. However, the output beam is now
collimated, which is critical for transmission through the detection optics.

In the dual-wavelength polarimetric system design, two optical coupling schemes were explored utilizing reflective
based systems. In the reflective based optical coupling setup that offered the best results, a 90 degree off-axis parabolic
mirror was placed at the output of the cornea to collect the fanning output beams of each of the wavelengths in the
system. The off-axis parabolic mirror reflected the beam back into the system towards the collection optics while
collimating the fanned beam. Additionally, a second off-axis parabolic mirror was placed after the first to further
collimate the overlapping wavelengths in the system prior to directing the beams through the analyzer onto the detector.
Commercial optics were evaluated from multiple vendors and an optical coupling scheme was constructed and
evaluated in modeling software utilizing reflective based optical components. In the final design a round mirror was
employed for guiding the entrance beam into the eye at a glancing angle and two 90 degree off-axis silver coated
parabolic mirrors were utilized to couple the light for both wavelengths out of the cornea and then providing a relatively
collimated beam onto the detection optics. For the optical ray trace simulation, CAD models were downloaded from the
THORLABS (Newton, NJ) website for each optical component and the listed specifications for the products were input
into CODE V to simulate the use of commercial optical mirrors. The 90 degree off-axis parabolic mirror item #:
MPD127075-90-P01 (THORLABS, Newton, NJ) was chosen for collimating the output beam in the coupling model
because it offered the most optimal parameters for the specified design. Utilizing two off-axis parabolic mirrors it was
expected that the design would re-collimate the output beam accounting for the expected beam divergence at the output
of the cornea as described by Malik et al. [19]. The off-axis parabolic mirror listed above was chosen because it provided
the shortest reflected focal length (RFL), diameter, and parent focal length of any 90 degree parabolic mirror
commercially available from several optical device vendors. The diameter of the parabolic mirror is 12.7 mm, the RFL is
15 mm, and the parent focal length was listed as 7.5 mm for the parabolic mirror listed above. Additionally, a round
protected silver mirror was determined to be optimal for providing the input beam into the cornea at a glancing angle. In
the modeling software the specifications for the round mirror were obtained from THORLABS for item #: PF03-03-P01.
The specified mirror had a diameter of 7 mm and a thickness of 2 mm.

3. RESULTS AND DISCUSSION
Utilizing CODE V, an optical model was constructed and ray trace simulations were evaluated for both 635 nm and
532 nm wavelengths, as illustrated Fig 2. In the model a glancing angle of 16 degrees was utilized for the entrance beam.
After the most suitable optical elements were determined to include in the reflective based optical coupling scheme, an
optical ray trace model was constructed in Code V to evaluate the re-collimating capability of the proposed design.
Initially the mirrors were placed arbitrarily in position with respect to an eye model. Then to optimize the performance of
the coupling system the beam position for the individual wavelengths at the detector and the appearance of the ray traces
for each wavelength were analyzed. To achieve suitable collimation of the beams at the detector the x, y, and z position
and angles of the two 90 degree off-axis parabolic mirrors were varied until optimal results were obtained maximizing
the overlap between the wavelengths and minimizing the change in the optical signal observed on the detector over
multiple simulations. Additionally, after the optimal position parameters were determined and entered for each optical
component the position of the optical elements were further modified until the results indicated an optimal ray trace of
Proc. of SPIE Vol. 8951 895107-3
Downloaded From: http://proceedings.spiedigitallibrary.org/ on 04/18/2014 Terms of Use: http://spiedl.org/terms
the overlapping wavelengths providing the most visually collimated output product, as shown in Fig. 2 below. After the
ray trace model was generated, as shown in Fig. 2, the output beam profile and spot profile were evaluated over multiple
runs. The sensor size was set to 3 mm x 3 mm for a 5x5 circular pattern of rays for both wavelengths. Then the beam
overlay on the detector was plotted to determine if the beam shape will remain similar over multiple runs in order to fit
on the surface area of commercially available optical detectors. As shown in Fig. 3A and 3B the spot diagram analysis
indicated the beam size remained within the 3 mm diameter detector area. Following the examination of the beam profile
on the detector the optimal position was determined for each element and a CAD rending of the ray trace starting with
the entrance beam up to the rays onto the detection optics were output in STL format. Utilizing SolidWorks a 3D model
was generated of the respective ray traces for the individual wavelengths with the exported file as a reference for
constructing the ray trace CAD models. Using the CODE V output ray traces and CAD models for the optical elements,
a human eye model [20], and a rending of a human face (obtained via scanning a face into a 3D representation) a model
was re-constructed similar to the Code V model evaluating anthropometric constraints. Using this model it was possible
to evaluate the placement of the optical elements and beam path as shown by the models in Fig. 4 below. Using the
exported ray traces generated in CODE V the model was then evaluated with respect to anatomical features on the
human face such as the nose and corneal/sclera surface to determine if the optical model would be feasible with
anatomical constraints. Figs. 4A and 4B illustrate the compiled 3D model of the optical simulation provided from CODE
V via multiple different viewing angles. As illustrated in Figs. 4A and 4B it appears the proposed non-index matched
optical coupling scheme is feasible from an anatomical and optical standpoint.
The proposed non-index coupling scheme including the input beam entering the eye at a glancing angle and 90
degree off-axis parabolic mirrors placed at the exit point of the eye were modeled in the CODE V and the ray trace for
635 nm and 532 nm wavelengths of the respective models are shown in Figs. 2 through 4 below. For each optical model
the beam diameter was chosen to be circular with a diameter of approximately 1mm. This is representative of the beam
diameter utilized in the system designed by our group. Additionally, a glancing angle of approximately 16 degrees was
introduced between the input beam and the entrance corneal surface of the eye.


Fig. 2: CODE V simulation of proposed non-index matched eye-coupling scheme (A) utilizing two 90 degree off-axis parabolic
mirrors re-collimating the beam exiting the cornea in addition to a round mirror guiding the entrance beam into the cornea at a
specified glancing angle and (B) close-up image of the model illustrating the ray trace change within the cornea and aqueous humor
surfaces.

M1-90P
D

M2-90P
A.)
B.)
Proc. of SPIE Vol. 8951 895107-4
Downloaded From: http://proceedings.spiedigitallibrary.org/ on 04/18/2014 Terms of Use: http://spiedl.org/terms
A . )
S U R F A C E 1 5
B . )
R M S = 1 . 4 7 3 3 4
0 e t 1 0 0 % = 2 . 5 6 4 3 2 3
I y 0 A 0 P ,
P
9
. 9 6 9 M M

Fig. 3: CODE V simulation spot ray trace simulation output at the detector for model shown in Fig. 2 (A) profile showing repeatable
beam shape under 3 mm x 3 mm which is the current size of the detector surface and (B) spot diagram using rays in circular pattern
5x5 rays of red and green beams with the total surface area set to 3 mm x 3 mm the detector area size.


Fig. 4: 3D anatomical analysis for proposed non-index matched design configuration utilizing CODE V ray trace exported in CAD
format. (A) 3D rending top view of proposed design coupled to eye with facial anatomy included and (B) additional side view of
proposed setup illustrating the collimated beam profile for input and output beams.

Based on the results illustrated in Figs. 4A and 4B it appears that the proposed non-index matched coupling scheme
could be achieved without being limited anatomically by the facial structure. It also appears significant room is available
to modify the position of the input and output mirrors for varying anthropometric requirements that exist between people.

4. CONCLUSION
In summary, we have developed and modeled an index unmatched optical coupling scheme using a reflective based
approach capable of evaluating the potential for coupling light into and out of the cornea without the use of index
matching between the air-cornea interface. Models were generated in CODE V to evaluate the feasibility of a non-index
matched coupling approach consisting of only commercially available optical elements. A reflective optical
configuration for each wavelength was modeled. The results produced from the optical modeling indicate the use of a
reflective based non-index matched eye coupling design, such as the one described in this report has the potential to
provide polarimetric glucose detection without the use of index matching. The model indicated that with the right
commercial optics it is possible to maintain a collimated beam through the optical coupling into and out of the eye if a
A.) B.)
Proc. of SPIE Vol. 8951 895107-5
Downloaded From: http://proceedings.spiedigitallibrary.org/ on 04/18/2014 Terms of Use: http://spiedl.org/terms
glancing angle is used to input the beam into the eye and off-axis parabolic mirrors are used to re-collimate the fanned
beam exiting the cornea. The CODE V software was shown to work great for exporting the beam profile of the ray trace
into 3D CAD software for anatomical consideration in the coupling design. However, in CODE V importing existing
commercial models available is not possible in their existing formats. Thus, a modeling program such as Zemax would
potentially be better at incorporating the optical elements from THORLABS because the models are provided for Zemax
by THORLABS including the optical properties and CAD dimensions of the elements.

REFERENCES
[1] Centers for Disease Control and Prevention, "National diabetes fact sheet: national estimates and general
information on diabetes and prediabetes in the United States," U.S. Department of Health and Human Services
Center for Disease Control and Prevention, (2011).
[2] T. D. Control, and Complications Trial Research Group, "The Effect of Intensive Treatment of Diabetes on the
Development and Progression of Long-Term Complications in Insulin-Dependent Diabetes Mellitus," N. Engl J.
Med 329(14), 977-986 (1993).
[3] Kuo, C. Y., Hsu, C. T. , Ho, C. S., et al., "Accuracy and precision evaluation of seven self-monitoring blood
glucose systems," Diabetes Technol Ther. 13(5), 596-600 (2011).
[4] Suzuki, Y., et al., "Preliminary evaluation of optical glucose sensing in red cell concentrations using near-infrared
diffuse-reflectance spectroscopy," J. Biomed Opt. 17(1), 017004-8 (2012).
[5] Huang, Z.-H., et al., "Noninvasive blood glucose sensing on human body with near-infrared reflection
spectroscopy," Proc. of SPIE 8193, 10 (2011).
[6] Burmeister, J.J., Arnold, M.A. and Small, G.W., "Noninvasive blood glucose measurements by near-infrared
transmission spectroscopy across human tongues," Diabetes Technol Ther. 2(1), 5-16 (2000).
[7] Esenaliev, R. O. , Larin, K. V. , Larina, I. V., et al., "Noninvasive monitoring of glucose concentration with optical
coherence tomography," Opt. Lett. 26(13), 992-994 (2001).
[8] Dingari, N., et al., "Investigation of the specificity of Raman spectroscopy in non-invasive blood glucose
measurements," Analytical and Bioanalytical Chemistry 400(9), 2871-2880 (2011).
[9] Weiss, R., et al., "Noninvasive continuous glucose monitoring using photoacoustic technology - Results from the
first 62 subjects," Diabetes Technol Ther. 9(1), 68-74 (2007).
[10] Pickup, J.C., et al., "Fluorescence-based glucose sensors," Biosensors and Bioelectronics 20(12), 2555 (2005)
[11] Cameron, B. D. and Cot, G. L., "Noninvasive glucose sensing utilizing a digital closed-loop polarimetric
approach," IEEE Trans. Biomed. Eng. 44(12), 1221-1227 (1997).
[12] Pirnstill, C. W., Malik, B. H., Gresham, V. C. and Cot, G. L., "In vivo glucose monitoring using dual-wavelength
polarimetry to overcome corneal birefringence in the presence of motion," Diabetes Technol Ther. 14(9), 819
(2012).
[13] Malik, B. H. and Cot, G. L., "Real-time, closed-loop dual-wavelength optical polarimetry for glucose monitoring,"
J. Biomed Opt. 15(1), 017002-6 (2010).
[14] Malik, B. H. and Cot, G. L., "Characterizing dual wavelength polarimetry through the eye for monitoring glucose,"
Biomed. Opt. Express 1(5), 1247-1258 (2010).
[15] Malik, B.H., Pirnstill, C.W. and Cot, G.L., "Polarimetric Glucose Sensing in an Artificial Eye Anterior Chamber,"
Proc. of SPIE 8229, 6 (2012).
[16] Cot, G.L., Pirnstill, C.W. and Malik, B.H., "In Vivo optical glucose monitoring using polarized light," Gordon
Research Conference Lasers in Medicine & Biology, (2012).
[17] Malik, B.H., Pirnstill, C.W., and Cot, G.L., "Dual Wavelength Polarimetric Glucose Sensing in the Presence of
Birefringence and Motion Artifact Using Anterior Chamber of the Eye Phantoms," J. Biomed Opt. 18(1), 017007-
017007 (2013).
[18] Pirnstill, C. W., Grunden, D. T. and Cot, G. L., "Polarimetric glucose sensing in vitro: a high frequency approach,"
Proc. of SPIE 8591, 859101-859107 (2013).
[19] Malik, B. H. and Cote, G. L., "Modeling the corneal birefringence of the eye toward the development of a
polarimetric glucose sensor," J. Biomed Opt. 15(3), 037012 (2010).
Proc. of SPIE Vol. 8951 895107-6
Downloaded From: http://proceedings.spiedigitallibrary.org/ on 04/18/2014 Terms of Use: http://spiedl.org/terms
[20] Winkler, A. M., Bonnema, G. T. and Barton, J. K., "Optical polarimetry for noninvasive glucose sensing enabled by
Sagnac interferometry," Appl. Opt. 50(17), 27192731 (2011).
[21] Liou, H.-L. and Brennan, N. A., "Anatomically accurate, finite model eye for optical modeling," J. Opt. Soc. Am. A
14(8), 16841695 (1997).
[22] Maurice, D., "The Charles Prentice award lecture 1989: the physiology of tears," Optom. Vis. Sci. 67(6), 391399
(1990).
[23] Sokolova, E., Kruizinga, B. and Golubenko, I., "Recording of concave diffraction gratings in a two-step process
using spatially incoherent light," Opt. Eng. 43(11), 26132622 (2004).
[24] Doyle, K. B., Hoffman, J. M., Genberg, V. L. and Michels, G. J., "Stress birefringence modeling for lens design and
photonics," Proc. of SPIE 4832, 436447 (2002).



Proc. of SPIE Vol. 8951 895107-7
Downloaded From: http://proceedings.spiedigitallibrary.org/ on 04/18/2014 Terms of Use: http://spiedl.org/terms