Measurement of the glucose concentration in human urine with optical

refractometer

Rui-Yang Wu, Cheng-Chih Hsu*, Ching-Tang Meng, Chih-Ching Cheng, and Yu-Ching Liao
Department of Photonics Engineering, Yuan Ze University, 135, Yuan-Tung Road, Chung-Li,
32003 Taiwan
* E-mail address: cchsu@saturn.yzu.edu.tw, Tel: 886 3 4638800; Fax: 886 3 451428

ABSTRACT

In this paper, a new type of human urine glucose measurement system is proposed. We measured the phase variation of
human urine with/without glucose-urine mixture (to simulate diabetes mellitus). We were able to achieve high
resolution with the proposed method. The relation curve between the phase difference and glucose concentration can be
estimated, and the glucose concentration of a urine sample can be determined by using this relation curve. The proposed
method showed that theoretical resolution is approximated of 1.47 mg/dl.

Keyword: human urine, glucose measurement, optical refractometer

1. INTRODUCTION

Diabetes mellitus currently afflicts 200 million people worldwide [1], will double to 300 million by 2025 [1, 2], and
causes many complications such as heart disease, kidney failure, and ophthalmic complications. These long-term
complications become more evident as the duration of diabetes increases and various organ-specific complications co-
exist. To prevent or at least slow the progression of these complications, self-monitoring of blood glucose for diabetic
patients is the current standard of care.

There are two categories of human blood glucose monitoring: invasive [3] and non-invasive [1, 3-10]. The relatively
pain-free non-invasive technologies have become the most attractive approach for diabetic patients at present. Many
methods [1, 3-10] have been reported, such as near infrared (NIR) spectroscopy [3], far infrared (FIR) spectroscopy [3],
dielectric spectroscopy [3] (DS), and optical polarimeter [3, 6, 10]. But many effects influence the result, such as the
difficulty of measuring the tiny test signal coming from the human body, the operation at specific frequency, and the
scattering effect of the test tissue. Hence, these methods have so far had limited applications.

To avoid these drawbacks, the refractive index measurement [7-8] of body fluids with an optical refractometer is another
approach to blood glucose concentration monitoring. There are many body fluids, such as urine [3, 5, 7], tears [9], and
interstitial fluid [3] that have strong relevance to blood glucose. And the fluids’ refractive index change as the glucose
concentration varies, implying the phase variation of the investigated laser beam. To demonstrate the validity of glucose
testing by refractive index variation monitoring, an alternative measurement method with optical refractometer is
proposed. Human urine was chosen as the test sample in manifold body fluids and glucose intermixed into it to simulate
diabetes mellitus. Then we obtained the relation curve between the phase difference and glucose concentration variation.

International Conference on Optical and Photonic Engineering (icOPEN 2015), edited by Anand K. Asundi, Yu Fu,
Proc. of SPIE Vol. 9524, 95240W · © 2015 SPIE · CCC code: 0277-786X/15/$18 · doi: 10.1117/12.2188975

Proc. of SPIE Vol. 9524 95240W-1

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 08/03/2015 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx

 In our method, because the light beam is normally reflected from one point on the surface of the test medium, the
scattering effect of the test medium is not strictly required and only a little quantity of the test medium is necessary. In
addition, there is no longer the need for an additional sensor [8] to be kept in contact with the test medium, eliminating
the need to clean after measuring a liquid. Based on the theoretical prediction, the resolution of the proposed method is
approximated of 1.47 mg/dl under the optimum experimental conditions. In brief, this method features easy, real-time
operation, better sensitivity, and simple optical configuration. Although our sample was very small, the measurement
results led us to infer that our method might be suitable for monitoring of glucose concentration in human urine,
especially in the sample of the first urine of the day.

2. PRINCIPLES

Figure 1 shows a schematic diagram of the proposed method. For simplicity, the +z axis is chosen to be along the light
propagation direction and the y-axis is along the direction perpendicular to the paper plane. The heterodyne light source
[11] consists of a linearly polarized light source, an electro-optic modulator, and a linear voltage amplifier. The angular
frequency difference between x- and y- polarizations is ω. The light beam coming from this heterodyne light source
passes through a quarter-wave plate Q1 with the fast axis at 0° to the x-axis, and the Jones vector of the light can be
written as [11]

⎛ ωt ⎞
⎜ cos( ) ⎟
2 ⎟ 1 ⎡1⎤ ωt 1 ⎡1 ⎤ ωt
Ei = Q(0°) ⋅ EO(ωt ) ⋅ E 0 = ⎜ = ⎢ ⎥ exp(i ) + ⎢ ⎥ exp(−i ) . (1)
⎜ ωt ⎟ 2 ⎣i ⎦ 2 2 ⎣− i ⎦ 2
⎜ − sin( ) ⎟
⎝ 2 ⎠
From Eq. (1), it is obvious that left- and right-circular polarizations also have an angular frequency difference ω. Then
the light beam enters a modified Twyman-Green interferometer, which consists of a beam-splitter BS, a test medium T
with refractive index n, a reference mirror M, a quarter wave plate Q2 with the fast axis at 45° to the x-axis, an analyzer
AN with the transmission axis at α degree to the x-axis, and a photodetector D. In this interferometer, the paths of two
beams are (1) BS → T → BS → AN(α) → D, and (2) BS → Q2(45°) → M → Q2(-45°) → BS → AN(α) → D. The light
beam in path 2 passes through Q2 twice, so its x- (or y-) polarization interchanges with its orthogonal counterpart, y- (or
x-) polarization. Thus, the interference signal measured by D can be written as
I t = I 0 [1 + γ cos(ωt + φ )] , (2)

where I0, γ, and φ are the mean intensity, the visibility, and the phase of the interference signal, respectively. They are
represented by the following equations:
1
I 0 = [ A 2 + 2rm2 + 2 Arm sin 2α sin(φd 1 − φd 2 )] , (3)
2
1
1 1
{ ( A 2 − 4rm2 ) 2 cos 2 (2α ) + [2 Arm − ( A 2 − 4rm2 ) sin(2α )]2 } 2
γ= 4 2 , (4)
1 2
[ A + 2rm + 2 Arm sin 2α sin(φ d 1 − φ d 2 )]
2

( A 2 − 4rm2 ) tan(2α ) − 4 Arm sec(2α )

φ = tan −1 [ ]. (5)
A 2 − 4rm2

Proc. of SPIE Vol. 9524 95240W-2

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 08/03/2015 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx

 In Eqs. (3) ~ (5), φd1 and φd2 are the phase variations associated with the optical path lengths of path 1 and path 2,
n −1
respectively; rm is the normal reflection coefficient of the mirror and A = .
n +1

He -Ne Laser EO

Heterodyne
Light
r
Source z
-VI
-r. 1 ....
z
AN BS Q2 M
rn
R H H H -"---+ n N I

T ,...L ,,.

Modified Twyman -Green Interferometer

Fig.1 The schematic diagram of the proposed method.

On the other hand, the electrical signal generated by the function generator FG is filtered and becomes the reference
signal that can be described as
1
I r = [1 + cos(ωt )] . (6)
2
Both of these two sinusoidal signals It and Ir are sent into the lock-in amplifier, and φ can be obtained. Theoretically,
various glucose concentration in solution lead to different phase difference which resulted from the refractive index of
the solution changed. Therefore, the glucose concentration in human urine and corresponding refractive index n can be
determined by measuring φ under the optimum measurement conditions of rm and α.

3. EXPERIMENT AND RESULTS

To show the validity of the proposed method, we measured the phase variations of human urine (first urine of the day)
intermixed with glucose, obtained from a normal people with no food for at least 8 hours prior. A He-Ne laser with
632.8nm wavelength and the EO (New Focus, Inc., Model: 4002) with a half-wave voltage 125 V were used to set up a
heterodyne light source. Its frequency difference was 1k Hz. A high-resolution lock-in amplifier (Stanford Research
System, Model: SR850) with an angular resolution of 0.001° was used to measure the phase difference. The human urine

Proc. of SPIE Vol. 9524 95240W-3

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 08/03/2015 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx

 was collected from one non-diabetic male of age about 31 years and intermixed with glucose to simulate different
glucose concentrations in urine. The glucose weighting was performed by an accuracy weight meter (Sartorius, Model:
RC210P). Then we measured these samples (pure urine and glucose-urine mixture) at room temperature. The values rm =
0.99 and α = 40° were chosen in our experiment. The measurement results are shown in Fig. 2. In this figure, the symbol
… and the solid line represent the average value of 50 measured data points and the linear fitting curve, respectively. The
equation of the linear fitting curve can be represented p (c) = 59.7675 + 0.00679c , in which p(c) and c are the
corresponding phase difference and the glucose concentration, respectively. Most of the measured data were within the
upper and lower 95% confidence limit (dash line). We also determined the refractive index of the test sample by
substituting phase difference into Eqs. (4) and (5). Then the results plotted together in Fig. 2. The relationship between
the refractive index and glucose concentration can be represented as n(c) = 1.3317 + 0.0002c , where the n(c) is the
corresponding refractive index. We also measured the phase variation of human urine samples (first urine of the day) and
monitored the variation for a long period (about 1 month); results are shown in Fig. 3. In this figure, the peak of the
normal distribution curve is 59.787°

Fig. 2 Measurement results and associated fitting curve of the phase difference versus glucose concentration at room
temperature.

to -:Normal distribution
1 Mean value: 59.787
J SD: 0.13813

8j 7 8
$

\
6] / \
I
5
\
4/

1 2 / N
211 /

59.5 59.6 59.7 59.8 59.9 60.0 60.

Phase difference (deg)

Fig.3 Phase variation of human urine from first morning urine samples- statistical analysis of measurement data.

Proc. of SPIE Vol. 9524 95240W-4

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 08/03/2015 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx

 4. DISCUSSION
To consider the angular resolution of the lock-in amplifier, the polarization-mixing errors, and the second-harmonic
error, the total phase difference errors Δφ could be decreased to 0.03° in our experiments [12]. The relation of the phase
difference error and resolution of the proposed method can be written as:

Δφ
Δc = . (7)
s

In which, symbols s, Δφ, and Δc are the slope of the fitting curve in Fig. 2, total phase difference errors and resolution of
the proposed method, respectively. Substituting measured data into the Eq. (7), the resolution of the proposed method
(Δc) can be obtained and approximated of 4.42 mg/dl.

From Eq. (5), it is obvious that two independent factors, which are the azimuth angle of the analyzer and the reflectance
of the reference mirror, will affect the phase difference. Hence, the relationship between them can be described as

Δφα = (∂ α φ ) ⋅ Δα , (8)

Δφm = (∂ rm φ ) ⋅ Δrm , (9)

where Δφα and Δφm are the errors in phase difference coming from azimuth angle error and reflectance error; Δα and Δrm
are the errors in azimuth angle and reflectance.

To understand the relation between them, the curves of Δφα versus α and Δφm versus rm for different test media have
been simulated and shown in Fig. 4 and Fig. 5, respectively. Obviously, the fluctuated behaviors of the phase difference
are acute for values of α in the region of 60° to 75° for different test media. This means that small variations of the
azimuth angle, which comes from the inaccurate alignment of the azimuth angle, will induce large errors in phase
difference. Hence, to reduce this effect in our method we can choose the azimuth angle of the analyzer from 20° to 50°.
In this region, Δφα will vanish when the azimuth angle of the analyzer equals 45°; these results are plotted in Fig. 4(b).
Unfortunately, we can measure the phase difference as α = 45°, because the divergence will occur at this condition and
obviously in Eq. (5).
s ,,,,,,,,,,,,,, _
,,,,, , . X 0-3
n=1.52
7. n=1.52.,,,,,,
6

4
= n=1
cy)
a) 3.

d 1., n=1.33
o

-1 .,

-20
10 20 30 40 50 60 70 80 90 100 41 42 43 44 45 46

a (deg) a (deg)

(a) (b)

Fig. 4 The relation curve of Δφα versus α. (a) Full range of α, (b) near 45 degree.

Proc. of SPIE Vol. 9524 95240W-5

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 08/03/2015 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx

 Moreover, the light beam reflected from the reference mirror M is used to encode and convert the reflectivity into the
phase term of the interference signal. Hence, its normal reflection coefficient rm should be related to the measurement
resolution. The curves of Δφm versus rm for different test media have been considered and shown in Fig. 5. From this
figure, it is obvious that Δφm decreases as rm increases. This means that the high reflectance mirror can achieve better
experimental results. Based on the theoretical prediction, the theoretical phase difference error can be estimated by the
summation of Δφα and Δφm which is approximated of 0.01° under the experimental conditions. Therefore, the theoretical
resolution of the proposed method can be reached to 1.47 mg/dl.

92x10-3

8.8

8.6

ú) 8.4
CD n=1.52
8.2
E
-9
a 8

7.8

7.6

7.4 - -

09 0.92 0.94 0.96 0.98

rm

Fig. 5 The relation curve of Δφm versus rm with different refractive index of test sample.

5. CONCLUSION

In this paper, we present an alternative method to determine the glucose concentration of human urine. According to the
experimental results, high resolution can be achieved because of the heterodyne interferometry. According to the
theoretical prediction, the resolution of this method can be approximated of 1.47 mg/dl. To realize the monitoring of
glucose in human urine, more experiments will be carried out, such as 24 hour urine sample testing, urine testing of
diabetics and non-diabetics, and urine testing of people with different dietetic-culture.

6. ACKNOWLEDGEMENTS

The authors thank the National Science Council of the Republic of China for financially supporting this research under Contract No.
NSC 102-2221-E-155-076-MY3. The authors acknowledge the contributions of Prof. J. Y. Lee and M. D. W. S. Yang in
discussing experimental issues in this study.

Proc. of SPIE Vol. 9524 95240W-6

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 08/03/2015 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx

 REFERENCES
[1] R. Rawer, Q. Li, W. Stork, and K. D. Muller-Glaser, “Implantable osmotic-pressure based glucose sensor with
non-invasive optical readout,” in BioMEMS and Nanotechnology, D. V. Nicolau and U. R. Muller, eds., Proc.
SPIE 5275, 247-256 (2004).
[2] A. Dove, “Seeking sweet relief for diabetes,” Nat. Biotechnology 20, 977-981 (2002).
[3] R. C. Bonney, “ Diabetes monitoring and devices”, Chapter 3, (PJB Publications Ltd., UK:Surrey, 1998).
[4] A. Caduff, E. Hirt, Y. Feldman, Z. Ali, and L. Heinemann, “First human experiments with a novel non-invasive,
non-optical continuous glucose monitoring system,” Biosens Bioelectron 19, 209-217 (2003).
[5] K. Dubayova, J. Kusnir, and L. Podracka, “Diagnostic monitoring of urine by means of synchronous fluorescence
spectrum,” J. Biochem. Biophys. Methods 55, 111-119 (2003).
[6] C. Chou, C. Y. Han, W. C. Kuo, Y. C. Huang, C. M. Feng, and J. C. Shyu, “Noninvasive glucose monitoring in
vivo with an optical heterodyne polarimeter,” App. Opt. 37, 3553-3557 (1998).
[7] S. Y. El-Zaiat, “Interferometric determination of refraction and dispersion of human blood-serum, saliva, sweat
and urine,” Opt. Laser Eng. 35, 55-60 (2003).
[8] K. H. Chen, C. C. Hsu, and D. C. Su, “Interferometric optical sensor for measuring glucose concentration,” Appl.
Opt. 42, 5774-5776 (2003).
[9] K. Mitsubayashi, Y. Wakabayashi, S. Tanimoto, D. Murotomi, and T. Endo, “Optical-transparent and flexible
glucose sensor with ITO electrode,” Biosens Bioelectron 19, 67-71 (2003).
[10] R. J. McNichols and G. L. Cote, “Optical glucose sensing in biological fluids: an overview,” J. Biomed. Opt. 5, 5-
16 (2000).
[11] D. C. Su, and C. C. Hsu, “Method for determining the optical axis and (ne, no) of a birefringence crystal,” Appl.
Opt. 41, 3936-3940 (2002).
[12] M. H. Chiu, J. Y. Lee, and D. C. Su, “Complex refractive-index measurement based on Fresnel’s equations and the
uses of heterodyne interferometry,” Appl. Opt. 38, 4047-4052 (1999).

Proc. of SPIE Vol. 9524 95240W-7

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 08/03/2015 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx