Skip to content


  • Open Access

Real-time gas identification on mobile platforms using a nanomechanical membrane-type surface stress sensor

EPJ Techniques and Instrumentation20141:9

  • Received: 23 June 2014
  • Accepted: 19 August 2014
  • Published:


Here we show real-time multiple gas identification on a mobile platform through the use of an array of nanomechanical membrane-type surface stress sensors (MSS). Commercially available hardware is used to integrate the MSS array into a portable unit with wireless capability. This unit transmits data to a consumer mobile tablet where data is displayed and processed in real-time. To achieve real-time processing with the limited computational power of commercial mobile hardware, a machine learning algorithm known as Random Forest is implemented. We demonstrate the real-time identification capability of the device by measuring the vapours of water, ethanol, isopropanol, and ambient air.


  • Random forest
  • MSS
  • Piezoresistive
  • Gas identification
  • Android
  • Mobile phone


1.1 Hardware implementation

The basis of our platform is a membrane-type surface stress sensor (MSS) that involves piezoresistors laid out in a full Wheatstone bridge configuration as described in previous papers [1],[2] (Figure 1a). The MSS has a theoretical detection limit of ~0.1 m N/m. By varying the properties of the coated polymer over each individual membrane such as its hydrophobicity and functional groups, differences in polymer-gas affinity can be utilized for identification of a wide range of VOC analytes.
Figure 1
Figure 1

System components and operation. a) Image of MSS in a four sensor array configuration. Each MSS is coated with a different polymer exhibiting different gas interaction properties. b) The enclosure containing the MSS sensor array chip and its electrical connections. c) Components in the device and the directions of information and gas flow between them. Information about the sample arrives at the tablet for analysis.

Two piezoelectric micropumps (Bartels Microtechnik mp6) flowed sample gases and ambient air over the MSS chip near their maximum rate of 0.3 m L/s[3]. A commercially available analog to digital converter with a resolution of 632 nV (ADS1258 EVM) measured the differential output voltage of the MSS with a bias voltage of -1.0 V. An Arduino Mega 2560 received this data via a Serial Peripheral Interface (SPI) to the analog-to-digital converter. The Arduino Mega 2560 also controlled the micropump switching. A custom breakout board mounted the MSS chip, which was then encased in a 3D-printed enclosure designed to maximize gas flow over the polymer receptor layers (Figure 1b). The Arduino Mega then sent the data to a consumer tablet (Google Nexus 7) over WiFi, using an Arduino WiFi shield.

1.2 Data processing

Random Forests [4] allow short characterization times of arbitrary input; characterization time is tunable through the size of the Forest. Each Forest can be tailored to complete its task on hardware of arbitrary speeds while maintaining a real-time analysis. Once generated offline, this machine learning algorithm can be moved to a target platform for quick, real-time analysis. Classification of data using a Random Forest simply involves traversal of many decision trees, which can be multithreaded easily for fast computation on multi-core processors. While this approach is sometimes coupled with Principle Component Analysis (PCA) to determine better candidates for predictors [5], the device is capable of identifying the chosen samples without requiring the full dataset in contrast to PCA. Voltage variations as a result of sample flowing through the device form unique curves when measured over time. These curves have several identifying characteristics, which can be extracted quickly by splitting the input into several windows, obtaining the difference of their averages, and using these as predictors for the Random Forest analysis.

The Random Forest was trained on sample data collected with the device using Scikit-learn [6] with Python 3.2. Converting the Forest into a custom, portable file format allows a consumer handheld tablet to predict outcomes with the CPU to be the only limiting factor in prediction speed.

1.3 Verification

The device was verified by identifying between ethanol, isopropanol, water, and ambient air. A training set consisting of 196 sample/purge cycles (48 cycles per sample type, and 52 for ambient air) was used to train a 100 tree Random Forest. Though the ADS1258 measured at a rate of 460 samples per second (SPS) per channel, data transmission was limited to 4800 bytes every 2 seconds (approximately 400 samples per second per channel) due to problems inherent within the Arduino WiFi module [7]. This caused the signal to appear discontinuous (Figure 2).
Figure 2
Figure 2

Sample ethanol data. An example of measured data during read cycles. Ethanol was pumped to the MSS on the first four segments (Sample point 0 to 1196) and ambient air was pumped to the MSS on the latter four segments to purge the absorbed ethanol vapour.

The training data was found to be easy to differentiate as a series of slopes in a voltage vs time series. Since the signal was divided evenly into eight segments per purge/sample cycle, the slopes between the averages of each segment indicated the general trend of the curve. Using these slopes as predictors for the Random Forest allowed the algorithm to identify sample gases with a high degree of accuracy (Additional files 1, 2).

The data classification technique traditionally performs well at avoiding overfitting of training data [4]. Using the out-of-bag estimator built-in to our Random Forest creation library, the Forest scored approximately 0.939, representing a 93.9% accuracy in predicting outcomes from inputs for which the Forest was not trained. The training set used to grow the Random Forest had a high amount of variability in its input (Figure 3), forcing the Forest to identify a corresponding sample from a wide variety of input data.
Figure 3
Figure 3

Training data channel response. An overlay of all input data used as the training dataset. Each graph (a-d) corresponds to the data for one channel (1-4) on the MSS chip, while the colours correspond to the analytes used.

1.4 Conclusion

We have demonstrated that the combination of an advanced algorithm (Random Forest) and the optimized nanomechanical sensor (MSS) can achieve real-time gas identification with commerical off-the-shelf hardware. Since the peripheral electronic components can be miniaturized by the introduction of application specific integrated circuits (ASIC) or field programmable gate arrays (FPGA), the present demonstration indicates the feasibility of integrating a real-time nanomechanical olfactory system into virtually any type of mobile platforms such as smartphones. Future developments towards real world applications will include a larger dataset with proper selection of parameters from output signals, effective receptor layers, and optimization of system components including the chamber and pumps.

Additional files



The authors express gratitude to Dr. Heinrich Rohrer, Dr. Terunobu Akiyama, Dr. Frederic Loizeau, Dr. Sebastian Gautsch, Dr. Peter Vettiger, Dr. Kota Shiba, Mr. Cory J. Y. Lee, Mr. Mayuran Saravanapavanantham, and Mr. Max Palumbo for their indispensable contributions to the development of the MSS platform and related devices, and Prof. Masakazu Aono, Dr. Tomonobu Nakayama, and Prof. Nico F. de Rooij for their help and support. This work was supported by WPI Research Center Initiative for Materials Nanoarchitectonics (MANA); the Grant-in-Aid for Young Scientist (A) 23685017 (2011), MEXT, Japan; Research Foundation for Opto-Science and Technology (REFOST); TEPCO Memorial Foundation; and Japan Science and Technology Agency (JST).

Authors’ Affiliations

World Premier International (WPI) Research Center Initiative, International Center for Materials Nanoarchitectonics (MANA), National Institute for Materials Science (NIMS), 1-1 Namiki, Tsukuba 305-0044, Japan
University of Waterloo, 200 University Ave W, Waterloo, N2L 3G1, Canada


  1. Yoshikawa G, Akiyama T, Gautsch S, Vettiger P, Rohrer H: Nanomechanical membrane-type surface stress sensor. Nano Lett 2011,11(3):1044–1048. doi:10.1021/nl103901a doi:10.1021/nl103901a 10.1021/nl103901aADSView ArticleGoogle Scholar
  2. Yoshikawa G, Akiyama T, Loizeau F, Shiba K, Gautsch S, Nakayama T, Vettiger P, Aono M: Two dimensional array of piezoresistive nanomechanical membrane-type surface stress sensor (MSS) with improved sensitivity. Sensors 2012,12(11):15873–15887. doi:10.3390/s121115873 doi:10.3390/s121115873 10.3390/s121115873View ArticleGoogle Scholar
  3. Operating manual micropump Mp6/mp6-pp and controller. , [].
  4. Breiman L: Random forests. Mach Learn 2001,45(1):5–32. doi:10.1023/A:1010933404324 doi:10.1023/A:1010933404324 10.1023/A:1010933404324MATHView ArticleGoogle Scholar
  5. Haindl M, Kittler J, Roli F (eds.): Multiple Classifier Systems. Lecture Notes in Computer Science, Vol. 4472, Berlin, Heidelberg: Springer; 2007. doi:10.1007/978–3-540–72523–7.Google Scholar
  6. Pedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, Blondel M, Prettenhofer P, Weiss R, Dubourg V, Vanderplas J, Passos A, Cournapeau D, Brucher M, Perrot M, Duchesnay E: Scikit-learn: Machine learning in Python. J Mach Learn Res 2011, 12: 2825–2830.MATHMathSciNetGoogle Scholar
  7. Notes regarding behaviour of arduino wifi shield as a TCP client. , [].


© Guerrero et al.; licensee Springer on behalf of EPJ. 2014

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited.