*2.1. Sensor Concept*

The core of the presented concept is the development of a photoacoustic transducer chip that allows the spatial separation of the acoustic sensor from the photoacoustic cell. The transducer consists of a gas-tight and gas-filled cavity, acting as an optical absorption layer that transfers the irradiated infrared light energy to acoustical pulses and a thin, flexible membrane that can pass these pulses to a second cavity containing a microphone.

Figure 1 illustrates a conceptual sketch of the sensing mechanism. A MEMS hotplate emits infrared (IR) pulses of defined frequency and pulse length, which may be filtered by means of an optical filter with a transmission spectrum matching the characteristic absorption band of CO2 at *λ* = 4.26 μm. Although this filter is not necessarily needed, it can prevent wavelengths different from those of the CO2 absorption spectrum to get absorbed in Silicon or glass parts of the transducer and thus increasing the performance. The pulses of IR light pass an optical absorption path, where they get partly absorbed by potentially present CO2 molecules.

After the absorption path, the light pulses enter a hermetically encapsulated absorption cell through an optically transparent and rigid window. The absorption cell is filled with CO2, which acts as a gas-specific absorption medium and therefore absorbs the IR light pulses with the characteristic absorption wavelengths of CO2. The absorbed energy gets transformed into acoustic waves by means of the photoacoustic effect. The photoacoustic effect itself consists of three different steps: The radiative absorption by the gas molecules, the successive relaxation of the molecules while increasing the total thermal energy in the gas, and—by means of the ideal gas law in a closed system—the resulting emergence of a pressure signal [20]. As the IR light pulses are modulated with a defined frequency, the pressure signal also follows this frequency. The more energy is absorbed in the absorption path, the lower the resulting intensity of the acoustic waves generated in the cell is.

**Figure 1.** Conceptual sketch of the proposed sensing mechanism comprising: (A) pulsed IR-emitter (B) optical filter (optional) (C) optical absorption path (D) IR-transparent entrance window (E) hermetic cell with encapsulated CO2 (F) flexible membrane (G) acoustically tight cavity with microphone (H) sensor signal. (**a**) sensor without CO2 (**b**) sensor in presence of CO2.

One wall of the absorption cell consists of a thin and flexible, but gas-tight membrane, which slightly deforms with the pressure fluctuations and therefore allows the generated pressure pulses to be transmitted to the other side while maintaining the CO2 atmosphere inside the absorption cell. This way, the acoustic pulses can enter the second cavity, which needs to be only acoustically-tight, but not necessarily gas-tight. Inside this second cell, a MEMS microphone is placed, which detects the arriving acoustic pulses. The amplitude of the acoustic pulses is indirectly proportional to the number of CO2 molecules present in the absorption path.

The concept of using a slightly deforming membrane or diaphragm for gas sensing was first described by Golay as part of an infrared active pneumatic detector [21]. In contrast to the concept presented here, the membrane was utilized as a deforming mirror in an optical modulation circuit rather than as an acoustic transducer.
