To provide a good level of indoor air quality and thermal comfort, mechanically-based ventilation systems in high-rise buildings play a vital role in delivering adequate fresh air exchanges, because densely populated and insufficiently ventilated buildings can greatly affect human health. Generally, low-grade filters inside their mechanical ventilation ductworks have a minimum efficiency rating value (MERV) of less than 1-4 (ASHRAE Standard 52.2.-1999). However, they are not effective for capturing fine or ultrafine particles at a sub-micrometre scale. Instead, the use of high-efficiency particulate air (HEPA) filters is a common alternative to achieving a higher capturing efficiency, yet it may induce a greater pressure drop with strong reliance on HVAC fan power to overcome the resistance, resulting in additional significant power consumption. Acoustic agglomeration is known as an efficient pre-treatment technique for promoting the formation of particle clusters to enhance the capturing efficiency of low-grade filters in mechanical ventilation systems. However, a key limitation in implementing electromagnetic acoustic transducers is that, they fail to be sufficiently sensitive to generate swift and precise responses which can cause the variation of unsteady airflows inside ventilation ducts. In addition, piezoelectric- or electrostatic-based acoustic devices are typically very large in size, thereby limiting their applicability in commercial engineering solutions. For the first time, we propose to develop a smart acoustic agglomeration technique enabled by thermo-acoustic (TA) waves, which can be directly generated from carbon nanotube (CNT) thin-film materials, and also 2D materials (e.g., graphene and hexagonal boron nitride (hBN)). Advances in nanotechnology over the last decade have realized the thermophone concept, which was discovered a century ago, and fundamentally differs from the mechanism of conventional acoustic devices to produce sound by mechanical vibration. The operating mechanism of this technique can be described by providing an alternating current to CNT thin-film materials, causing the surrounding medium (air) to be heated periodically. Therefore, an oscillating temperature field can be induced in the medium, resulting in thermal expansion and contraction to generate acoustic waves. This innovative control technique can be scalable to deliver a flat-band frequency range as well as high-intensity sound pressure outputs. In this work, we demonstrate TA effect, in the form of a standing wave, which can be modulated by an adaptive system to promote the coagulation of suspended particles in air.