LLNL-CONF-686901

Overview and Comparison of

Spatial Light Modulator
Calibration Methods

A. E. M. Browar, M. Shusteff, R. M. Panas, J. D.

Ellis, C. M. Spadaccini

March 24, 2016

31st ASPE Annual Meeting

Portland, OR, United States
October 23, 2016 through October 28, 2016
Disclaimer

This document was prepared as an account of work sponsored by an agency of the United States
government. Neither the United States government nor Lawrence Livermore National Security, LLC,
nor any of their employees makes any warranty, expressed or implied, or assumes any legal liability or
responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or
process disclosed, or represents that its use would not infringe privately owned rights. Reference herein
to any specific commercial product, process, or service by trade name, trademark, manufacturer, or
otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the
United States government or Lawrence Livermore National Security, LLC. The views and opinions of
authors expressed herein do not necessarily state or reflect those of the United States government or
Lawrence Livermore National Security, LLC, and shall not be used for advertising or product
endorsement purposes.
 OVERVIEW AND COMPARISON OF SPATIAL LIGHT MODULATOR
CALIBRATION METHODS

Allison E.M. Browar1,3 , Maxim Shusteff3,4 , Brett E. Kelly3,5 , Robert M. Panas3 ,

Jonathan D. Ellis1,2 , and Christopher M. Spadaccini3
1
Department of Mechanical Engineering
2
The Institute of Optics, University of Rochester, Rochester, NY, USA
3
Materials Engineering Division
Lawrence Livermore National Laboratory, Livermore, CA, USA
4
Department of Electrical Engineering and Computer Science
Massachusetts Institute of Technology, Cambridge, MA, USA
5
Department of Mechanical Engineering
University of California, Berkeley
Berkeley, CA, USA

INTRODUCTION sectional layers of the SLM consist of a glass

Wavefront control is critical for advancing a wide substrate, a transparent electrode, liquid crystals,
array of optical systems, yet it is difficult to pixelated electrodes and a reflective silicon
integrate a closed loop diagnostic subsystem. backplane. As light passes through the liquid
In this work, we explore different metrology crystals, the phase is modulated proportionally to
approaches to identify fundamental electrical and the effective refractive index of the liquid crystal,
physical errors in a wavefront-controlling spatial which can be altered by applying a voltage.
light modulator (SLM) and compensate for them. Pixelated electrodes allow independent control
Three drop in diagnostic systems were created of potential difference, providing the capability of
to analyze phase modulation by an SLM. A sys- spatially varying phase. However, fundamental
tematic approach to calibration is demonstrated. electrical and physical hardware errors mandate
Our aim is to develop an optimal phase calibration calibration of the SLM.
apparatus which will enable significant advances
in precision, repeatability, and predictability in The voltage applied to each pixel in the SLM
optical systems such as digital holography and is assigned through a look up table (LUT)
holographic optical tweezers [1, 2]. programmed in the SLM control software. Eight-
bit grey values displayed on the SLM control the
voltages applied at each pixel. After observing
a nonlinear relationship between the measured
overall phase and applied grey value, it was
determined that the LUT needed to be calibrated
to recover a linear response. The gamma table
and voltages that were encoded in the SLM are
optimized by the manufacturer to yield a linear
FIGURE 1. Framework of a parallel aligned phase shift with grey value of range 2.1π.
nematic liquid crystal on silicon spatial light
modulator. Layers include a glass substrate, Additionally, due to fabrication limitations, the
continuous transparent electrode, liquid crystals, silicon backplane and glass substrate of the SLM
pixelated electrodes, and a silicon backplane. may not be flat nor parallel [4]. This contributes
Thickness of SLM at A and B may vary due to the nonlinearity of the response, since the
to fabrication imperfections, such as a non-flat induced phase delay is proportional to thickness,
silicon backplane. All pixels bear zero potential illustrated in Figure 1. For this calibration, more
difference except the pixel at A. precise measurements of spatially varying phase
In this study we use a parallel aligned nematic over the entire SLM surface needed to be made.
liquid crystal on silicon spatial light modulator
(LCoS SLM) as illustrated in Figure 1 [3]. Cross MEASUREMENT METHODS
Previous studies suggest three diverse tech-
niques including interferometry, linear polarizers,
and diffraction based measurements of phase
delay to quantify variable phase error as a result
of a non-flat backplane of the SLM, shown in
Figure 1. Phase correction is dependent on both
the physical backplane error and the nonlinear
electrically induced birefringence inherent to liq-
uid crystals. Interferometry based measurements
lead to an analysis of both relative phase delay
from active liquid crystals with respect to the
off state (grey value of 0) and any curvature
within the physical layers of the SLM. The linear
polarizer technique described below measures
the absolute phase change imposed by the
FIGURE 2. Michelson Interferometer with the
liquid crystal, which can be measured for each
SLM as the device under test. Reference mirror,
pixel. Diffraction based measurements quantify
M1, is on a nano-positioning actuator, A. A
an overall phase of the entire illuminated SLM
rectangle with variable 8-bit values is displayed on
area relative to an SLM with no applied voltage.
the SLM with a constant 0 grey value perimeter.
Our aim is to perform a rigorous comparison of
An interferogram is captured using the camera
the techniques and quantify the reliability and
and is converted into a phase map after using a
repeatability of the processes aiming to create an
5-step phase shifting technique.
in situ calibration method.
value, a rectangle with constant grey values was
The SLM used to evaluate all of these techniques displayed on the SLM, surrounded by a constant
is a Holoeye-Pluto SLM (HES 6010 VIS) set to a 0 grey value. The phase, φ, at each grey value
5:6 bitplane driving scheme provided by Holoeye was then calculated using
support team to minimize flicker often seen at
higher driving schemes [5]. The source for the
 
2(I2 − I4 )
interferometric and diffraction based methods is φ = tan−1 (1)
2I3 − I5 − I1
a 532 nm solid state laser (Coherent Verdi 6)
which is spatially filtered using a pinhole and where I1 , I2 , I3 , I4 and I5 are image of intensity
collimated. A ThorLabs 530 nm Light Emitting taken at δ = 0, π/2, π, 3π/2, and 2π, respectively.
Diode (LED) is used for the cross polarizing
diagnostic to eliminate undesirable interference To recover the phase measurement at each gray
phenomena that arise from coherent illumination. value, first, any high-frequency image content
was removed by Fourier low-pass filtering. Then
Interferometry a MATLAB routine was used to unwrap the phase
In this study we use a Michelson interferometer profile and remove any tilt of the SLM. The phase
as shown in Figure 2. The SLM takes the place offset was removed by subtracting the average
of the test mirror and a reference mirror with phase on the outer perimeter of the rectangle.
a flatness of λ/20 or better is mounted on a The resulting average phase value within the
nano-positioning stage with closed loop feedback rectangle thereby yields the actual phase change,
control (ThorLabs NFL5DP20S). A reduction ∆φ, for each grey value.
telescope with magnification 0.66X images the
SLM onto the CMOS camera. One complication for determining the phase
change between the zero valued perimeter and
Of the variety of shifting methods available, the inner rectangle is that the results must be
we chose the 5-step, Hariharan, method as a unwrapped from the −π to π domain. Figure 3
suitable compromise between measurement time shows the phase data with grey value 172
and accuracy [6]. The nano-positioning stage taken using the 5-step method (top) compared
was moved in δ = π/2 increments, collecting 5 to the phase map after unwrapping using the
intensity measurements from the CMOS camera. MATLAB built in unwrap function, removing tilt,
and subtracting the phase profile for a gray value
To measure the phase change for each grey of 0 (bottom). The standard MATLAB unwrapping
 FIGURE 4. Change of phase relative to zero
FIGURE 3. Wrapped (top) and unwrapped grey value displayed on the SLM. The calibrated
(bottom) phase of the SLM with 172 grey gamma table we used was linear 2.1π phase
value rectangle in the center and 0 grey value range.
outside. The unwrapped phase was calculated
by unwrapping the phase discontinuities and
removing the tilt, and subtracting phase variance
due to the non-flat surface of the SLM.

algorithm has trouble handling measured phase

values around π and 2π, so this aspect of the
analysis will need to be made more robust in the
future.

Figure 4 shows the phase change as a function

of grey value displayed on the SLM. The phase is
displayed between −π and π and values around
π and 2π needed to be determined manually.
Our measurement implies that the experimental FIGURE 5. Phase measurement setup using
values for phase, as measured by the Michelson linear polarizers. The first linear polarizer P1 is
Interferometer, do not match with the encoded oriented at 45◦ . The SLM delays the horizontally
calibration table. polarized light, the light is reflected off the back
plane and is delayed again by the SLM. The
Polarization Approach light is reflected off the beam splitter and travels
In a similar configuration, we implemented two through a second polarizer, P2 oriented at ±45◦ .
linear polarizers to measure the relationship The image of the resulting intensity is displayed
between phase and grey value as in [7, 8]. on the CMOS camera.
Figure 5 shows the experimental apparatus. The
first linear polarizer, LP1 is oriented at 45◦ to intensity values of the recorded image. The Jones
the horizontal. The SLM delays the phase of Matrix of a reflective SLM, derived from [9], is
only one polarization, the one oriented along the
 −i2φ
horizontal; therefore the phase reflected from the

e 0 −i2kno d
SLM is elliptical. The second polarizer, LP2 is e (2)
0 1
oriented at ±45◦ to the horizontal to correct for
the light that was unchanged by the SLM. The where d is the thickness of the liquid crystal layer
phase change, ∆φ, induced by the liquid crystals and no is the refractive index of the SLM with no
can be calculated using Jones matrices from the induced phase delay.
Following through the system using Jones matri- single SLM pixel, compared with the analytically-
ces and an incoherent light source with electric predicted functional form.
field, E,
Ex eiφx
 
These measurements match well to the analytical
E= (3)
Ey eiφx expression, in which the intensity of a modulated
phase through 45◦ linear polarizers is propor-
where Ex , Ey , φx and, φy are the independent tional to cos2 (φ/2+δ). Work is ongoing to map the
amplitudes and complex phases, respectively, full phase response of the SLM area on a pixel-
for the components of light perpendicular to the by-pixel basis.
propagation direction, the resulting electric field,
Esystem , is

Ex eiφx + Ey eiφy
 
1
Esystem = (eiφ ± 1) (4)
4 ±Ex eiφx ± Ey eiφy

where the positive equation arises from polarizers

rotated parallel to each other and the negative
equation is a result of perpendicular polarizers.
The effect is a phase shift of π which is addressed
when the intensity is simplified below. The
intensity for each pixel on the SLM is given by
†
I(∆φ) = Esystem Esystem (5)
1 FIGURE 6. Experimentally validated normalized
= (Ex eiφx + E iφy )2 (ei∆φ ± 1)2 (6)
8 intensity versus grey values for a single pixel. The
theoretical intensity is proportional to cos2 (φ/2+δ)
which can be reduced to where δ is a phase shift.
 
φ Diffraction
2
I(φ) = I0 cos or (7) Several authors display a binary grating on an
2
  SLM because the phase delay from addressed
φ
I(φ) = I0 sin2 (8) values can be analytically extracted from the
2 intensity of the diffracted orders using the
Fourier transform [4, 10, 11]. When a perfect
where I0 is a constant with change of φ and only
binary phase grating is displayed on the SLM
depends on incident amplitude and phase. This
and is passed through a Fourier lens, the
result shows the phase response of each pixel
normalized power of the zeroth and first orders
when imaged through the linear polarizers can
are proportional to
be calculated by normalizing the intensity and
solving for φ. 1 + cos(φ)
P0 (φ) = and (9)
2
This approach can be easily adapted to mea-  2
2 1 − cos(φ)
suring the spatially-varying phase response over P±1 = (10)
π 2
the area of the SLM. Accordingly, as a first
step, we measured the phase of one portion respectively, where P0 and P±1 are the nor-
of the SLM at high magnification using a Nikon malized power and φ is the phase modulation
40X Plan Fluor microscope objective and a from the SLM [10]. Ronzitti suggests that
75 mm tube lens resulting in a field of view the output phase from the SLM has inter-pixel
that images 104x84 pixels on the SLM. The phase effects and results in a sinusoidal rather
use of high NA optics allows recording of sub- than crenel-shaped phase response. With that
pixel information. Similarly to the interferometric modification, the resulting power for the zeroth
approach, a 100x80 pixel rectangle at varying and first diffraction orders has the forms
gray values was displayed in the center of the   2
0 φ
field of view, and the measured intensity of the P0 (φ) = J0 and (11)
resulting image was extracted, using a custom 2
  2
MATLAB routine. Figure 6 shows the measured 0 φ
intensities as a function of grey value for a P1 (φ) = J1 (12)
2
respectively, where J0 and J1 are the zero and
first order Bessel functions, respectively.

Figure 7 shows the setup used to measure the

intensity of the diffraction orders from the SLM
when a linear phase grating is applied. A binary
grating with a period of two pixels was displayed
on the SLM. A Fourier lens with focal length
250 mm was used one focal length from the SLM.
The power at the zeroth and first order were taken
at various grey values using a Thorlabs digital
power meter (PM121D). Figure 8 shows the
normalized power measurements for the zeroth
and first diffraction orders. Theoretical values
for a crenel-shaped and sinusoidal response are
shown for comparison.

FIGURE 8. Measurements of normalized power

versus grey values of the zeroth and first
diffraction order from a linear grating. The theo-
retical values for a crenel-shaped and sinusoidal
FIGURE 7. Setup to measure the intensity of response are shown with fitting parameters for
the diffractive modes from the SLM. A linear phase offset and are normalized.
diffraction grating is displayed on the SLM. A
Fourier lens, FL, is placed one focal length away
from the SLM. Each mode is measured using the
power meter, PM.
The phase measured using the diffraction method
corresponds to the overall phase change of the
entire illuminated area of the diffraction grating (in
this case the whole SLM), including any inherent
physical defects. Preliminary results suggest that
more refined calibration and analytical models are
necessary due to physical defects in the SLM
and approximation of phase response shape.
Engström suggests splitting the SLM into regions
to measure the phase delay at each particular
region of the SLM [4]. Yu suggests that the
response of the SLM should account for the
distance between the pixels [11]. Further analysis
will be carried out to approximate the phase
response using images taken with the linear
polarizer setup mentioned above.

RESULTS AND DISCUSSION

Figure 9 compares the phase response mea-
sured from each method with their associated
FIGURE 9. Graph comparing phase measured
uncertainties. The phase for the polarizer method
by each method
was calculated using equation 7 and solving for
φ. Phase for the zeroth and first order diffraction
TABLE 1. Method Comparison

Phase Shifting Linear

Parameter Interferometry Polarizer Diffraction
Measurement Continuous Discrete Overall
Type (over SLM
face)
Light Source coherent incoherent coherent
Angle of Inci- 0 any any
dence
Advantages can measure flatness of can calibrate each pixel, can be measured in
SLM, can calibrate each one measurement per Fourier domain, can
pixel grey value, absolute create subsection of
measurement, low SLM and calculate phase
spacial footprint delay of region
Disadvantages No discretization of pixels, Highly data intensive, measures overall phase,
must unwrap phase, only must have at least π cannot resolve single
relative measurement range to extract phase, pixel phase variation
compared to zero value, must have incoherent
low confidence at π and source
2π phase values

was calculated numerically using Equations 11 Distributed Renewable Energy: From Science
and 12, respectively, and solving for φ. The and Technology to Entrepreneurship and Policy
uncertainties for the polarization and diffraction (NSF Award #0966089)
method were calculated by taking the derivative
of equations associated with each method and REFERENCES
solving for dφ. [1] Lutkenhaus J, George D, Moazzezi M, et al.
Digitally Tunable Holographic Lithography
We have demonstrated three concepts to mea- Using a Spatial Light Modulator as a Pro-
sure the fundamental errors affecting the phase grammable Phase Mask. Optics Express.
in a spatial light modulator. Table 1 some 2013 October;21(22):26227–26235.
considerations for comparing the performance
of interferometry, polarization, and diffraction [2] Curtis JE, Koss BA, Grier DG. Dynamic
methods to measure the phase variance on a Holographic Optical Tweezers. Optics
spatial light modulator. Ongoing work is being Communications. 2002;.
performed to rigorously quantify and systemat-
[3] Takiguchi Y, Otsu T, Inoue T, et al.
ically correct for errors associated with non-flat
Self-distortion Compensation of Spatial
back plane and nonlinear electrically induced
Light Modulator Under Temperature-
birefringence. With this information we will be
Varying Conditions. Optics Express. 2014
able to construct an optimal in situ calibration
June;22(13):16087.
apparatus for our system to use with optical
tweezers and holographic lithography. [4] Engström D, Persson M, Bengtsson J,
et al. Calibration of Spatial Light Mod-
ACKNOWLEDGMENTS ulators Suffering from Spatially Varying
This work was performed in part under the Phase Response. Optics Express. 2013
auspices of the U.S. Department of Energy by June;21(13):16086–16103.
Lawrence Livermore National Laboratory under
contract DE-AC52-07NA27344. The authors [5] Martı̀nez FJ, Màrquez A, Gallego S, et al.
acknowledge support from the LLNL LDRD Electrical Dependencies of Optical Mod-
program (14-SI-004). LLNL-CONF-686901. We ulation Capabilities in Digitally Addressed
would also like to acknowledge funding support Parallel Aligned Liquid Crystal on Sili-
from the NSF Integrative Graduate Education con Devices. Optical Engineering. 2014
and Research Traineeship (IGERT) Program: June;53(6):067104.
 [6] Malacara D, editor. Optical Shop Testing. 3rd
ed. John Wiley & Sons, Inc.; 2007.
[7] Reichelt S. Spatially Resolved Phase-
Response Calibration of Liquid-Crystal-
Based Spatial Light Modulators. Applied
Optics. 2013 April;52(12):2610–2618.

[8] Stigwall J. Optimization of a Spatial Light

Modulator for Beam Steering and Tracking
Applications. Swedish Defence Research
Agency; 2002.
[9] Yu FH, Kwok HS. Comparison of Extended
Jones Matrices for Twisted Nematic Liquid-
Crystal displays at oblique angles of inci-
dence. Journal of the Optical Society of
America. 1999;.
[10] Ronzitti E, Guillon M, de Sars V, Emiliani
V. LCoS Nematic SLM Characterization and
Modeling for Diffraction Efficiency Optimiza-
tion Zero and Ghost Orders Suppression.
Optics Express. 2012 July;20(16).
[11] Yu CJ, Park JH, Kim J, et al. Design of
Binary Diffraction Gratings of Liquid Crystals
in a Linearly Graded Phase Model. Applied
Optics. 2004 March;43(9):1783–1788.