After the industrial revolution1 of the 18th and 19th centuries, where manufacturing processes changed with machines and steam power, came the second industrial revolution with electrification of production lines, telephone, telegraph, expanding railroad networks, paving the way to globalisation.2 We are now in the third industrial revolution3, the information age or digital revolution, based on transformations such as the transistors, computers, internet, smartphones….
With a typical size of a few square millimetres and a selling price well below one euro for most of them, these MEMS and sensors make small and affordable what was before expensive and big.
Famous examples are accelerometers (e.g. for airbag control in cars), pressure sensors, gyroscopes, microphones, magnetometers, humidity sensors, micro-mirrors (e.g. for digital video projection), radio-frequency filter for communications, inkjet heads, micro-pumps.…
From the technical conference held by the MEMS & Sensors Industry Group in Munich this March4, we learn that future and emerging MEMS and Sensors are now believed to be gas sensors (CO2, CO, Ammonia, SOx, NOx, VOCs), particle sensors, automotive sensors like imaging or MEMS mirrors for LIDAR but also medical and healthcare sensors like non-invasive glucose sensors for diabetes, breath analysis, smart pills, sensors on skin, low power gyroscopes (< 1 µA), sensors for home monitoring and miniaturised spectroscopy.
Sensing devices generate data. A lot of data. The value is in the data, and its exploitation enables new businesses and new business models: this is digitalisation.5,6
This will affect productivity,7 healthcare, safety, smart homes…. everything. It should significantly increase the global economy.
The connected things (from smartphones to cars, production machines, and homes) and people are the internet of everything (IoE).
Silicon technology in VTT
Our silicon microfabrication cleanrooms are based in Espoo, Finland, in a national infrastructure called Micronova.8 Aalto University also uses the infrastructure. Companies use the facilities for their R&D and/or production. VTT Ltd has developed a few silicon competitive technologies.
Figure 2. Mapping of the cyanobacteria in the lake Lohjanjärvi. Project HSI-stereo in collaboration with University of Jyväskylä, FGI, Luode consulting, Lentokuva Vallas.
We are surrounded by an extraordinary amount of visual information. That is way more than our brain can process. And this is only for the visible part of the wavelength spectrum. Spectrometers analyse the light emitted, reflected or absorbed by objects and allow to determine their nature. They are expensive, big and fragile. But what if they were cheap, small and robust?
VTT invested in total more than 200,000 hours of research on the miniaturization of that technology, from ultraviolet to thermal infrared. Today we have a technology that allows to fabricate handheld microspectrometers solutions that are robust and can be mass manufactured. Moreover, we developed hyperspectral imaging devices, in which each pixel is a microspectrometer.
That generates a new family of imaging spectral sensors for a myriad of applications and new businesses in health, agriculture, pollution and environmental monitoring, defense, chemical analysis, industrial production monitoring, archeology – only the imagination is the limit. Everything has a spectral signature, and if we add the time dimension, no problem, hyperspectral video is also possible. If we add image analysis and machine learning, the imagination goes wild.
Let’s take a few examples of realisations:
In spring 2016, the Aalto 19 nanosatellite will be launched by SpaceX10 with one of our hyperspectral cameras on board (400 nm–1,100 nm, 5–10 nm resolution). That will be the first Finnish satellite and our first test in space. Much closer to the ground, unmanned air vehicles (UAVs) (Figure 5), drones, quadcopters, pseudo satellites, can carry hyperspectral imagers for many applications.
In the project HyperGlobal (Tekes) on offshore hyperspectral imaging, the remote sensing of sulphur oxides (SOx) pollution from ships is targeted. More stringent regulations on SOx emissions can only be forced though efficient monitoring technology as they increase the costs by billions of euros for the shipping industry.
As for the environmental monitoring let’s take two Finnish examples. The blue-green algae in lakes and pest insects detrimental to the forests. Blue-green algae or cyanobacteria releases cyanotoxins that are poisonous for animals and humans. In Finland, famous for its lakes, every summer, the presence of the blue-green algae is monitored by local sampling to inform people. Another method that has been tested is lake area mapping from a plane with a hyperspectral camera. In Figure 2 we see an example with the lake Lohjanjärvi taken from 2,000 m altitude. The high concentration of cyanobacteria is observed on the North-East part of the lake near the city center. The mapping instead of sampling is more powerful to understand the phenomenon and have a global view.
In the forests, there are sometimes pest insects, like the bark beetle, that affect trees, with environmental and economical impact. It is possible to map the trees affected by it with hyperspectral cameras. With the hyperspectral camera technology developed at VTT, this was realised on Norway spruce (Picea abies L. Karst.) by Näsi et al.11 in the MMEA research programme, coordinated by Cleen Ltd. in co-operation with FGI (Finnish Geospatial Research Institute) and the University of Helsinki. This is a very promising technology for preventing and monitoring bark beetle outbreaks.
Figure 3. Visualization of classification results using FPI color-infrared and RGB images on background. Image courtesy Roope Näsi from National Land Survey of Finland.
Figure 4. Resulting forest classification in the Kerinkallio test site. Image courtesy Roope Näsi.
In the health field, we developed a camera able to assess the skin field cancerization in seconds.12 We also developed a camera for the hyperspectral imaging of the eye retina with a potential in helping diagnosis in oxygen saturation, retina structure, glaucoma and diabetes.13 Other devices were made like a hyperspectral microscope, chemical imagers, a wireless microspectrometer in the cover of a phone for CO2 monitoring, and recently an hyperspectral system on an iPhone camera. You can read more on our spectral sensing technology.14
Others use hyperspectral imaging for Archeology,15 biodiversity quantification and nature management.16 If you have insights on the possible exploitation of spectral information I invite you to contact us. We probably can tailor a camera for your specific business.
Figure 5. Test hyperspectral imagers on-board of a drone used in the project HSI-Stereo (Tekes).
Infrared room temperature nanobolometers: we developed silicon based infrared nanobolometers that, from our measurements on test structures, should eventually be one or two orders of magnitude more sensitive than the state-of-the-art (resistive and thermoelectric). 17
The great sensitivity gain represents a key advantage for small aperture lower-cost optical systems and furthermore can be a lower cost alternative to other sensors above 5 µm wavelength.
Figure 6. Estimation of detectivity of VTT’s nanobolometers for two different pixel sizes.
The internet of things or the wearable technologies rely much on the fact that the MEMS and sensors and their systems have very little energy needs or even are energy autonomous and still wirelessly connected. In many cases using and changing batteries is too costly. For these reasons, energy harvesting (from heat, vibrations, WiFi, sunlight….) and energy storage are important.
One way to store energy is to use supercapacitors. They can release energy very fast (e.g. for communication) and they can store a lot of energy per mass. A classic way to rank supercapacitors is to plot their performance in a Ragone plot18 where the power density is plotted versus the energy density. This illustrates how fast and how much energy is delivered per volume of supercapacitor. VTT’s supercapacitors and current state-of-the-art devices are based on porous silicon and a titanium nitride coating grown by atomic layer deposition.19
As supercapacitors are extremely thin, it makes sense to talk about energy per surface. In our case we reach power densities of ~50 mW/mm2 or energy densities of ~2 mWs/mm2.
Figure 7. Ragone plot. Areal energy and power densities of Porous Silicon-TiN supercapacitors with aqueous (blue rectangles) and organic (red diamonds) electrolyte and comparison with literature. Maximum values for 100 µm PDMS case with organic electrolyte are also shown (open red diamonds). The available data of several relevant devices from the literature are presented: laser graphene oxide20, graphene coated porous silicon21 and direct laser writing22 (for those three cases the areal values were calculated from volumetric values using given device thicknesses); graphitization of silicon carbide23 and silicon carbide nanowires24 (volumetric values were calculated from areal values using device thickness); doped silicon nanowires25 (device or layer thickness not given in the article).
Figure 8. Detailed structure of the porous silicon supercapacitors.
MEMS based CO2 and relative humidity sensor
There is an increasing demand for sensors measuring the quality of our environment and typically the air we breathe. Some sensors are already on the market, the latest being the BME680 from Bosch.20 In VTT Ltd we are working on a MEMS based solution that measures CO2 in air21 very fast and from small concentrations to much higher ones, like e.g. in exhaled breath. We now work on its stability versus temperature. If we succeed we believe this 1 mm2 active area MEMS sensor will give CO2 readings in less than one second and only consume a few mW during the measurement. Given its size, this could become a very low cost CO2 sensor for consumer electronics. We also work on simultaneous measurements of relative humidity and CO2 using the same principles.
We developed over the years a competitive silicon-on-insulator (SOI) micron-scale waveguide technology with unique properties like losses below 0.1 dB/cm, ultra-wide bandwidth (from 1.2 to 8 µm), small polarisation dependence and a bending radius of less than 10 µm with less than 0.1 dB/90° loss.
The know-how in this field allows us to offer high integration density on chips. Our traditional activities on passive components are now completed by active components and their integration towards cheaper, faster and more energy-efficient data communication. Other applications in biological or gas sensing and in the medical field are considered. You can read all the details in a former issue of this magazine.22
Figure 9. Illustration (left), microscope image and a scanning electron microscope image of waveguide curves and mirrors developed by VTT, with bending radius down to one micrometre, which is around a thousand times smaller than traditional curves.
Besides the contract R&D for customers, process development, IP licensing and participation in jointly funded research projects, VTT Ltd is offering silicon manufacturing services via its daughter company VTT MEMSFAB.23 Low volumes in the order of several tens of thousands of 150 mm wafer starts per year can be offered on given processes. After a process development is complete, production of our specialty MEMS, sensors and components or pilot production can be performed for the customer. The key components can then be integrated in high added value instruments for target applications by partners or customers. Technology transfers are also possible. Typical examples for manufacturing can be radiation detectors, MEMS Fabry-Perot Interferometers, etc.
If we consider only the silicon activities, a few companies emerged from VTT. Among them we have VTT MEMSFAB23, Aivon,24 Advaplan (CMP services), Advacam25 (Large area photon counting radiation sensors and cameras), Spectral Engines (Spectral sensors)26 and Asqella (THz based security screening solutions).27 As VTT can vertically integrate its technology from the hardware to the cloud based service, more are expected in the future.
From silicon to cloud
We have this slogan in our business area led by the executive Vice President Petri Kalliokoski. Knowledge intensive products and services are our core business. Our asset in VTT Ltd is the ability to build vertical value beyond each special competence within a same institution. For example, our unique roll-to-roll printed electronics lines28,29 and our silicon microsystems technologies are meant to explore lower cost hybrid flexible electronics. Our sensor systems experts further integrate components into solutions, devices and prototypes. Our communication systems take the lead in the future networks and related hardware, e.g. 5G. Finally health applications, accredited metrology30 (VTT Mikes) and digital systems and services complete the skillset to offer solutions for business and generate new economical value.
A recent example among many is how we combined our environmental metrology know how and hyperspectral technology. Initially we demonstrated the identification of materials from the record breaking distance of 1.5 km.31 The method called active hyperspectral detection (AHS) in the infrared uses a supercontinuum light source. Latest development of the technology under the TransSmart programme32 included significant miniaturisation of the system using MEMS Fabry-Perot technology and implementation of the latest supercontinuous laser. The AHS instrument was successfully tested in traffic conditions, spectrally resolving objects ahead of the driving vehicle.
Figure 10. Normalised NIR spectra of different targets (left)37, and AHS instrument positioned on top of the VTT vehicle for field tests (right).
Philippe Monnoyer is
the head of VTT’s Microsystems research area. He got his DSc (Physical
Chemistry) degree from the University of Namur in Belgium in 1998.
has worked at the Liège Space Centre, Imec and Motorola (later
Freescale Semiconductors) in a research alliance with ST
Microelectronics and Philips (later NXP) in Crolles, near Grenoble in
France. Monnoyer moved to Finland in 2007 and joined VTT, where his work
focuses on MEMS technology. His research team’s most important research
areas include microspectrometers, MEMS, radiation sensors,
supercapacitors, infrared nanobolometers, and piezoelectric materials.
research team is part of a VTT’s business area that provides technology
from silicon to cloud, including printed electronics, sensor systems,
metrology and health care applications, communication systems, and
 Street view https://email@example.com,24.8188365,3a,46.3y,296.53h,90.53t/data=!3m7!1e1!3m5!1saRxfdVeUG6nespjQnUeqjw!2e0!6s%2F%2Fgeo2.ggpht.com%2Fcbk%3Fpanoid%3DaRxfdVeUG6nespjQnUeqjw%26output%3Dthumbnail%26cb_client%3Dmaps_sv.tactile.gps%26thumb%3D2%26w%3D203%26h%3D100%26yaw%3D49.058609%26pitch%3D0!7i13312!8i6656?hl=en
Summary Aalto-1 Spectral Imager has been built in an ESA-Strin project “MEMS Fabry-Perot Interferometer Technology for Miniaturized Hyperspectral Imagers and Microspectrometers”, ESA contract number No. 4000106267/12/NL/CP.
 Näsi, R., Honkavaara, E., Lyytikäinen-Saarenmaa, P., Blomqvist, M., Litkey, P., Hakala, T., & Holopainen, M. (2015). Using UAV-Based Photogrammetry and Hyperspectral Imaging for Mapping Bark Beetle Damage at Tree-Level. Remote Sensing, 7(11), 15467-15493.
 Kaare et al., Development of tunable Fabry-Perot spectral camera and light source for medical applications, AIP Conf. Proc. 1537, 231 (2013); http://dx.doi.org/10.1063/1.4809717
 U. Dillner et al., J. Sens. Sens. Syst. 2, 85 (2013).
 Koppinen et al., A novel MEMS gas sensor based on ultrasonic resonance cavity, Ultrasonics Symposium (IUS), 2014 IEEE International, pp 655–658.
 A. Manninen et al. Opt. Express 22, 7172-7177 (2014).