148th ASA Meeting, San Diego, CA

[ Lay Language Paper Index | Press Room ]

Chemical Detection With a MEMS Microphone

Dr. Michael Pedersen- mpedersen@comcast.net
Novusonic Corp.
17942 Pond Road
Ashton, MD 20861

Popular version of paper 2aEA7
Presented Tuesday morning, November 16, 2004
148th ASA Meeting, San Diego, CA

In this paper we discuss a project currently under way to develop a micro mechanical MEMS microphone specifically for use in photoacoustic instruments. Photoacoustic spectroscopy (PAS), which measures the molecular absorption of light energy to provide identification of molecules with high resolution, is a simple and promising technology for use in biochemical sensors and detectors.

What is photoacoustics?

The photoacoustic effect of matter was first discovered by Alexander Graham Bell in 1880. He found that if he aimed a strong light source at a surface, and pulsed the light by turning it on and off, an acoustic signal with similar frequency was emitted from the surface. His discoveries led to the spectrophone shown in figure 1. The discovery was considered an oddity of nature and it took 50 years to develop microphones sensitive enough to actually measure a photoacoustic signal. It was not until 1973, when a detailed theoretical model (the RG theory) was developed by Rosenzweig and Gersho, that the phenomenon was well understood. The basic principle of photoacoustic conversion is shown in figure 2.

spectrophone  principle
Figure 1 (upper) and 2 (lower): The original spectrophone designed by Bell and basic principle of photoacoustic conversion.

If the incoming light has energy (a wavelength), that matches the difference between the steady state and an allowed excited state in the molecule, there is a good chance the light will be absorbed putting the molecule into an excited state. The molecule will only remain in this state for a short period and eventually decay back to the steady state. During the decay, the excess energy can be dispersed in the form of fluorescence, a photochemical reaction, or simply as heat. It is the heat that is exploited, since the local heat generation leads to a pressure rise in the surrounding gas, which can be detected with a pressure sensitive microphone. In figure 3, a basic photoacoustic detector is shown. The light source is pulsed either by turning it on and off, or with a screen wheel as shown. When the wavelength of the light is chosen to coincide with an absorption line in the gas, a photoacoustic signal is generated at the microphone with a frequency similar to the light modulation frequency. It is important to understand that all molecules have unique sets of allowed states, as predicted by quantum theory, which gives each molecule a unique fingerprint. Therefore, with this method it is possible to detect specific molecules with very high sensitivity. Sensitivities to the ppt (parts per trillion) level have been demonstrated. In figure 4, a sample absorption fingerprint is shown for ethane, which illustrates the strong absorption peaks.

PAS setup         
Figure 3 (upper) and 4 (lower): A basic photoacoustic gas detector and an example absorption spectrum for ethane.

Why use MEMS microphones for this?

The microphones presently used in state-of-the-art photoacoustic instrumentation are very high quality measurement microphones, that were originally designed for a general purpose. The most important problems with measurement microphones in this application are:

            Mismatch of frequency response (Bandwidth too large).
            High vibration sensitivity.
High cost.

The frequency range of interest in microphones for photoacoustics is from about 100Hz to a few kHz. Since the microphones currently used have bandwidths sometimes in excess of 10 kHz, a lot of potential microphone sensitivity is lost. As a result, as shown in table 1 below, it is possible to design a much smaller MEMS microphone, which still has higher sensitivity and lower thermal noise level than the measurement microphone. The vibration sensitivity of microphones is directly related to the thickness of the pressure sensing membrane inside the microphone. Conventional microphones have much thicker membranes than what can be realized in MEMS microphones. As a result, a reduction in vibration sensitivity of more than 10 times can easily be realized. The vibration sensitivity is one of the most important issues, which has largely limited photoacoustic spectroscopy to the laboratory environment where structure-born noises can be minimized. MEMS microphones will help in the realization of rugged photoacoustic field equipment, and hence help the technology compete against other detection methods in the field. The performance of a photoacoustic MEMS microphone is summarized in table 1 with the specifications of two other microphones currently used in instrumentation.

Table 1: Performance comparison of conventional microphones used for PAS and proposed MEMS microphone.


Brüel & Kjær 4189
Capacitive Microphone
Knowles EM-3446
Capacitive Microphone

Photoacoustic MEMS microphone

Diaphragm size

~12.5mm Ø



Diaphragm thickness


~25 micron

0.5 micron

Diaphragm resonance frequency

14 kHz

5.2 kHz

4.3 kHz

Open-circuit sensitivity

50 mV/Pa

17.8 mV/Pa

432 mV/Pa

Frequency response

6.3 Hz to 20 kHz ±2 dB

300 Hz to 9 kHz ±3 dB

10 Hz to 4 kHz ±2 dB

DC bias voltage

0 V (Internal bias)

0 V (Internal bias)

1.2 V

Microphone capacitance

13 pF

~3 pF

1.5 pF

Input referred noise level

14.6 dB(A) SPL

25 dB(A) SPL

0 dB(A) SPL

Maximum sound pressure

158 dB SPL


94 dB SPL

Vibration sensitivity @ 1 g

82.3 dB SPL

63 dB SPL

55 dB SPL

Operating temperature range

-30 to 150 ºC

-17 to 63 ºC

-40 to 150 ºC


>$200 w/o amplifier

~$10 w/ amplifier

~$10 w/ amplifier

[ Lay Language Paper Index | Press Room ]