Content uploaded by Anuj Baskota
Author content
All content in this area was uploaded by Anuj Baskota on Feb 15, 2022
Content may be subject to copyright.
Gigahertz Ultrasonic Imaging of Nematodes in
Liquids, Soil, and Air
Justin Kuo1, Anuj Baskota1, Scott Zimmerman1, Frank Hay2, Sarah Pethybridge2, and Amit Lal1,3
1 Geegah, Inc., Ithaca, NY, USA
2 School of Integrative Plant Science, Cornell University, Ithaca, NY
3 SonicMEMS Laboratory, School of Electrical and Computer Engineering, Cornell University, Ithaca, NY, USA
Abstract— Plant parasitic nematodes are soil-borne
microscopic worms that parasitize plant roots, reducing crop
yields for economically important crops such as soybeans, sugar
beets, and strawberries, to name a few. This paper presents the
concept of using a GHz ultrasonic imaging array to image
nematodes, as a way to quantify nematode populations within soil.
A 128-pixel by 128-pixel, 1.85 GHz monolithic CMOS integrated
ultrasonic imager was successfully used to image nematodes in
moist soil, air, and thin layers of water.
Keywords—GHz ultrasonic imaging, aluminum nitride, MEMS,
nematodes
I. INTRODUCTION
Nematodes are worm-like animals that comprise one of the
most diverse phyla of organisms, with over 1 million species
estimated to exist [1]. Out of all these types of nematodes, those
of particular interest to humans are nematodes that live within
soil, which includes beneficial nematodes and plant parasitic
nematodes. Beneficial nematodes are nematodes that parasitize
insects and therefore can be used for pest control. Plant parasitic
nematodes are nematodes that are of great economic
importance because they can parasitize crops, thereby reducing
crop yields.
For example, soybean cyst nematodes are known to be the
greatest contributor to soybean crop losses [2], causing annual
crop losses of more than 9% in the United States [3]. These
nematodes were found to be responsible for mean loss per year
of more than $1 billion in damages in the U.S. between 2006-
2009 from soybean crop loss [4].
Due to high toxicity and cost of nematicides for controlling
nematode populations, it is desirable to restrict the usage of
nematicides to only where they are needed. In order to diagnose
whether or not low crop yields are due to nematodes, to
determine whether or not nematicides should be applied, it is
necessary to perform soil sampling of nematode populations.
This is done by taking several soil and core samples per acre.
These samples are sent to a lab where nematode, cyst, and egg
counting are performed by filtering the soil samples and
manually counting the filtered nematodes, cysts, and eggs
through an optical microscope.
Not only is this process time-consuming, but soil core
samples are generally taken at specific times during the year,
potentially backlogging labs and resulting in delays in testing.
In order to expedite this process and get results faster to the
farmer, the use of a micro-imager that is capable of operating
and sensing beneath soil is proposed as a way to quantify local
nematode populations either in the field or in diagnostic
equipment to detect whether or not a given plot of plant has
nematodes and to determine whether or not a chemical product
should be applied to that plot in the field.
The proposed solution utilizes a GHz ultrasonic imager, as
shown in Fig. 1. Thin film AlN (aluminum nitride) transducers
fabricated on a silicon substrate transmit GHz frequency
ultrasound into the silicon substrate. The ultrasound travels
through the thickness of the silicon substrate until it hits the
back surface of the silicon, where the ultrasound is reflected
back into the silicon, where it is received by the AlN
transducers and the echo amplitude per pixel is measured and
formed into an image. Depending on the acoustic impedance of
the material present on the back surface of the silicon, the
reflection coefficient of the reflected wave, and therefore the
measured echo amplitude, is different.
Monolithic CMOS integration of the AlN transducers with
CMOS transmit and receive circuits is required to realize a fine
pitch 2D imager array due to the number of interconnects
required to access all transducers and the potential fanout issues
that are incurred for heterogenous integration of MEMS
transducers and CMOS circuits on separate chips. Furthermore,
CMOS integration allows for a massive signal level
improvement simply due to the elimination of low input
impedances, typically from parasitic load capacitances or 50
ohm load resistances, that typically attenuate GHz transducer
echo amplitudes to very low voltages in cases of heterogeneous
integration.
This work was supported by ARPA-E under award number DE-AR0001049
and NSF under award number 1746710.
Figure 1: Concept of using a GHz ultrasonic imager to image nematodes
below soil
978-0-7381-1209-1/21/$31.00 ©2021 IEEE
2021 IEEE International Ultrasonics Symposium (IUS) | 978-1-6654-0355-9/21/$31.00 ©2021 IEEE | DOI: 10.1109/IUS52206.2021.9593762
Authorized licensed use limited to: Cornell University Library. Downloaded on February 15,2022 at 17:43:53 UTC from IEEE Xplore. Restrictions apply.
A GHz ultrasonic imager is proposed for several reasons: 1)
Ultrasound does not require a light source and therefore can be
deployed under the soil. While GHz ultrasound does not
penetrate into the soil due to the high attenuation of GHz
ultrasound in non-crystalline materials, objects, such as
nematodes, that come into contact with the imager surface will
be imaged. 2) GHz ultrasound potentially allows for high
imaging resolution due to the smaller wavelength of ultrasound
at higher frequencies – around 4.5 µm at 1.85 GHz in <100>
silicon for bulk acoustic waves. 3) The imaging occurs using
the unpatterned back side of the silicon, allowing for increased
sensor robustness and lifetime in soil environments. 4) CMOS-
MEMS integration allows for additional functionality to be
added into the imager chip, such as GPS or RF
communications, allowing for a highly compact and integrated
multi-modal agricultural sensor to be realized at large volumes
and low cost, increasing sensor density per acre and thereby
increasing the usefulness of digital agriculture techniques.
The GHz transducers used in this work utilize a similar
device layer stack to previous work from the SonicMEMS
Laboratory at Cornell University [5-8]. While traditionally the
use of GHz ultrasound has been confined to the realm of
acoustic microscopy [9], in recent developments from Cornell
University, GHz ultrasound has been employed in a wide array
of applications, such as on-chip ultrasonic communication,
reconfigurable phased array on-chip wireless communication
links, fingerprint imaging, temperature sensing, analog
computation, as well as single pixel transducers integrated with
CMOS transmit and receive circuits [5-8].
This technology has been further developed at Geegah to
achieve a monolithic single-chip 128 by 128 pixel imager, with
50 um pitch pixels, where the RF transmit and receive
electronics are situated within the area of a pixel, directly
underneath the pixel transducer. Geegah has developed an
evaluation system (Fig. 2c) such that images can be acquired
from the sensor with frame rates of 6 to 12 frames per second,
depending on operation mode, over a USB interface. The
ultrasound carrier frequency is typically selected to be 1.85
GHz, although operation between 1.6 GHz to 1.9 GHz is
possible due to the wide bandwidth of the GHz transducers.
II. EXPERIMENT
In order to demonstrate that a GHz ultrasonic imager is
capable of imaging nematodes, a series of experiments were
performed using the 128 by 128 pixel imager system. As shown
in Fig. 3a, nematodes are placed in contact with the sensor on
the exposed unpatterned silicon surface of the sensor chip.
The type of nematodes used in the experiments are
Steinernema carpocapsae nematodes (BioBest Sustainable
Crop Management). These nematodes are not plant parasitic
nematodes, but are instead beneficial nematodes. The reason
these nematodes were used for the experiments is because they
are easily obtainable and disposable, whereas plant parasitic
nematodes are typically only available after annual soil
sampling of fields. As these nematodes are packaged as a paste
comprised of a mix of nematodes and nematode food, it is
necessary to dissolve some of the paste in water first to activate
Figure 2: (A) Micrograph of 128 x 128 imager chip, (B) Micrograph
of pixel with pixel transducer shown, (C) Imager system with USB
interface for data acquisition.
(A) (B)
(C)
8 mm
7.5 mm
50 µm
Figure 3: (A) Experiment setup showing how nematodes are places on
top of the sensor for the water and air experiments. (B) Sensor placement
in soil for the nematodes in wet soil experiment. The imager surface is
fully immersed in the soil. (C) Optical micrograph of the Steinernema
nematodes used in the experiment. The nematodes in this image have an
average length of 597.5 ± 22.3 µm and an average width of 28.7 ± 2.6
µm.
(A)
(B) (C)
Authorized licensed use limited to: Cornell University Library. Downloaded on February 15,2022 at 17:43:53 UTC from IEEE Xplore. Restrictions apply.
the nematodes. As shown in Fig. 3, an optical microscope
(Hayear HY-2307) was mounted on top of the imager to
provide optical images to prove that the acoustic images are
indeed showing nematodes. Using this optical microscope, the
average length of the nematodes was determined to be around
600 µm in length and around 30 µm in width, noting that the
width is slightly smaller than the size of a pixel.
The first type of experiment performed was to take a droplet
of water (~10 µL) with nematodes in it and deposit it on the
sensor. The droplet of water is allowed to dry, such that only
nematodes are left. The purpose of this experiment is to
demonstrate that it is possible to distinguish individual
nematodes on the sensor in air, without any water or soil.
The second type of experiment is to place a droplet of water
(~100 µL) with numerous nematodes in it on the sensor while
imaging. The purpose of this experiment is to determine
whether or not the nematodes can be imaged in water.
The third and last type of experiment is to insert the sensor
into soil and apply water (~1000 µL) containing nematodes to
the top of the soil. The image data is collected over time and
analyzed to detect the presence of nematodes.
III. R
ESULTS AND
D
ISCUSSION
The imager system outputs quadrature measurement data for
each pixel – for each measurement, the system generates two
quadrature 128 x 128, 12-bit matrices –
and
.
The value of each pixel in the matrix is the quadrature
component of the first acoustic echo from the silicon backside
interface. For best image contrast, a baseline measurement is
taken with air backing, at the beginning of each experiment,
generating matrices
and
. While amplitude and
phase can be extracted from measurement data, for simplicity,
the image matrices
or
shown in this work are
simply the difference between the two air-backed and
measurement matrices:
(1)
(2)
If the shape of nematodes can be seen in the GHz ultrasonic
images, then the image indicates the presence of nematodes.
Nematodes are also known to move at speeds of ~100 to 600
µm/s [10] – movements in measurement data corresponding to
these speeds are therefore likely to be indicative of nematode
motion. Analyzing nematode movement across adjacent imager
frames, nematode movement velocity has been measured to be
approximately 120 µm/s, which corresponds to literature
values.
The results from the first air-backed nematode imaging
experiment are shown in Fig. 4. In this experiment, a droplet of
water containing nematodes and dissolved nematode food was
deposited on the sensor surface and allowed to dry. From the
optical image in Fig. 4a, nematodes can be seen within a
circular ring that is likely due to deposits of dried nematode
food formed by the coffee-ring effect. The exact same
Figure 5: Optical and GHz ultrasonic images taken during nematodes in
water experiment. Note that the optical and ultrasonic images are not
taken at the same time. (A) Optical image showing nematodes dispensed
onto the surface of the sensor in a water surface, (B) 1.85 GHz ultrasonic
image showing nematodes crawling out of the water droplet onto the
sensor. (C) Nematodes under PDMS block on sensor surface. (D) 1.85
GHz ultrasonic image showing nematodes in thin water layer under
PDMS block. A sensor board with a recessed cavity is used, hence less
than 64 x 64 pixels are exposed to the droplet of water.
Voltage (mV)
Column
Row
(A)
Voltage (mV)
50
0
-50
(B)
(C) (D)
Nematodes
Nematodes
Nematodes
Sensor surface
PDMS block
Sensor surface
exposed to water
Figure 4: (A) Optical image of nematodes in air on the sensor surface.
(B) Corresponding 1820.2 MHz ultrasonic image of the nematodes on
the surface. Note that the optical and ultrasonic images are flipped
because the optical camera looks down on the sample from the top, and
the ultrasonic imager looks up at the sample from the bottom.
(A)
(B)
Authorized licensed use limited to: Cornell University Library. Downloaded on February 15,2022 at 17:43:53 UTC from IEEE Xplore. Restrictions apply.
nematodes, as well as the coffee ring from the dried nematode
food, can be seen in the ultrasonic image, Fig. 4b, which was
measured at an 1820.2 MHz ultrasonic carrier frequency.
The next experiment involved placing a droplet of water
containing nematodes onto the sensor. As time passes,
nematodes can be observed crawling out of the water droplet,
as shown in Fig. 5a. However, nematodes cannot be seen in the
center of the droplet. There are two possible causes for this: 1)
There are so many nematodes in that region that they look like
one single continuous mass from the imager. 2) Nematodes, due
to their light weight, tend to float on top of a water layer if the
water layer is too thick. This second possibility is not a cause
for concern for sensing in the field, as it is known that
nematodes travel through small pores in moist soil [11].
To further prove that nematodes are visible in thin films of
water, a block of PDMS was placed on top of a droplet of water
containing nematodes. Pressing down on the PDMS block,
nematode movement could be detected using the GHz
ultrasonic imager, as in Fig. 5d, although the exact shape of the
nematodes cannot easily be distinguished due to varying
thicknesses of water and the large number of nematodes.
The last experiment involved seeing if the sensor can detect
nematodes in soil. The sensor board is placed in soil, with the
entire sensor surface immersed in soil. Water containing
nematodes was added to the soil. As shown in Fig. 6A, as the
water seeps through the soil, the nematodes within it move into
contact with the sensors. As opposed to the nematodes in water
experiment, the nematodes can be very easily distinguished in
the ultrasonic image. This is likely because the small gaps
within the soil that the water flows through brings the
nematodes directly in contact with the sensor surface. As the
water dries up, the nematodes that do not manage to move away
in time are left on the imager surface and remain visible in the
ultrasonic image.
IV. C
ONCLUSIONS AND
F
UTURE
W
ORK
This work has shown, for the first time, the first ever
published images from a monolithic CMOS-integrated GHz
ultrasonic imaging array. Using this new type of imager,
imaging of nematodes has been achieved in air, water, and in
soil. Most notably, imaging within soil provided clear images of
nematodes as the small pores in soil allowed nematodes to get
very close to the sensor.
Although the nematodes being sensed in the current work are
beneficial nematodes, plant parasitic nematodes are the
nematodes of interest in the intended application for the sensor.
Therefore, efforts are ongoing to improve imager resolution and
image processing in order to allow the imager to determine if a
nematode is a plant parasitic nematode, which can typically be
distinguished through their mouthparts. In addition, further
sensor testing in fields and greenhouses is required to investigate
how nematodes interact with the sensor in a less controlled, and
therefore more realistic, environment and to determine the
appropriate packaging for the sensor for field deployment.
The authors believe that the nematode imaging demonstrated
in the current work is only one application out of the many
potential imaging and sensing applications that monolithic
CMOS-integrated GHz ultrasonic imager arrays can be applied
to. In particular, this new sensing technology will make possible
space-constrained and cost-constrained applications that were
not possible before, and allow them to be achieved with high
imaging resolution.
A
CKNOWLEDGMENT
This work was supported by ARPA-E under award number
DE-AR0001049 and NSF under award number 1746710. This
work was performed in part at the Cornell NanoScale Facility, a
member of the National Nanotechnology Coordinated
Infrastructure (NNCI), which is supported by the National
Science Foundation (Grant NNCI-2025233).
R
EFERENCES
[1] J. Hugot, P. Baujard, and S. Morand, “Biodiversity in helminths and
nematodes as a field of study: an overview,” Nematology, 2001, vol. 3(3),
199-208.
[2] A. Bandara, et al., “Dissecting the economic impact of soybean diseases
in the United States over two decades,” PLOS ONE, Apr. 2020.
[3] S. Savary, et al., “The global burden of pathogens and pests on major food
crops,” Nature Ecology & Evolution, vol. 3, Mar. 2019, 430-439.
[4] S. Koenning and K. Wrather, “Suppression of Soybean Yield Potential in
the Continental United States by Plant Diseases from 2006 to 2009,” Plant
Health Progress, Nov. 2010.
[5] J. Kuo, et al., “64-Pixel Solid State CMOS Compatible Ultrasonic
Fingerprint Reader”, IEEE International Conference on Micro Electro
Mechanical Systems (IEEE MEMS) 2017.
[6] J. Hoople, “Integrated Piezoelectric Transducer Arrays for On Chip
Communication and Sensing,” Ph.D. Thesis, Cornell Univ., Ithaca, NY,
USA, 2016.
[7] M. Abdelmejeed, “AlN Piezoelectric CMOS Compatible Ultrasound
Transducers: Applications on Sensing and Computation,” Ph.D. Thesis,
Cornell Univ., Ithaca, NY, USA, 2020.
[8] M. Abdelmejeed, et al., “Monolithic 180nm CMOS Controlled GHz
Ultrasonic Impedance Sensing and Imaging,” 2019 IEEE International
Electron Devices Meeting (IEEE IEDM), 2019.
[9] A. Briggs and O. Kolosov, Acoustic Microscopy: Second Edition. Oxford
University Press. 2009.
[10] S. Jung, “Caenorhabditis elegans swimming in a saturated particulate
system,” Physics of Fluids, vol. 22, no. 3, 2010.
[11] H. Wallace, “The Dynamics of Nematode Movement,” Annu. Rev.
Phytopathol., 1968.
Figure 6: 1.85 GHz ultrasonic images from nematodes in wet soil
experiment: (A) Water is completely covering the sensor. (B) As the
water dries up, nematodes can still be observed in the dried regions.
(A)
Row Row
Column
Voltage (mV)
(B)
Epoxy covered regions
Wet soil
Nematodes
Dried soil
Nematodes
Authorized licensed use limited to: Cornell University Library. Downloaded on February 15,2022 at 17:43:53 UTC from IEEE Xplore. Restrictions apply.