VOL. 10 January ISSUE YEAR 2009

Science Update

in Vol. 10 - January Issue - Year 2009
CO2 Snow Blasting
Fig. 1: Mechanism of blasting with solid carbon dioxide, left: thermal effect, right: mechanical effect

Fig. 1: Mechanism of blasting with solid carbon dioxide, left: thermal effect, right: mechanical effect

Fig. 2: Half section of the developed blasting device

Fig. 2: Half section of the developed blasting device

Fig. 3: Velocity of the air stream at a blasting pressure of 7 bar

Fig. 3: Velocity of the air stream at a blasting pressure of 7 bar

Fig. 4: Path lines at a pressure of 7 bar

Fig. 4: Path lines at a pressure of 7 bar

A new Blasting Device – Development aided by CFD-Simulation


In recent years, blasting with solid carbon dioxide has been established in industry as a cleaning technology. In a variety of applications blasting with solid carbon dioxide can substitute chemical cleaning technologies. Therefore it contributes to environment protection and improves employee safety levels. The used CO2 in industry has no additionally effect on the global warming, because it is a by-product of the chemical industry and released into atmosphere anyway. The usage of this CO2 as a blasting media provides an economical and ecological advantage.

CO2 is a colourless, odourless, non-flammable, electrically non-conductive and chemically inert gas [1]. It is stored in the liquid phase either in low pressure tanks ( 20 °C, 20 bar) or in high pressure tanks (20 °C, 57 bar). When the pressurized liquid CO2 is expanded to ambient pressure solid carbon dioxide is generated. During the expansion the gas cools down as a result of the Joule-Thomson-Effect. A certain fraction of the gas goes into the solid state, which is called CO2 snow. Blasting with solid carbon dioxide can be divided into two variants, blasting from the solid phase and blasting from the liquid phase. Blasting from the solid phase is called dry ice blasting. Thereby the CO2 snow is forced trough moulds and dry ice pellets are generated. These pellets are feed in blasting devices to a compressed air stream and accelerated within nozzles on the surface to be cleaned. When blasting from the liquid phase, the expansion of the liquid CO2 takes place only in the blasting device. The generated CO2 snow is then used immediately as the blasting agent and also accelerated by a compressed air stream on the surface to be machined. The advantage over dry ice blasting is the reduced complexity in terms of automation. The dry ice pellets sublime continuously due to the contact with air under ambient conditions. Furthermore they stick together as a result of freezing out of air humidity. To avoid this, an additionally technical effort is needed.
Dry ice blasting as well as CO2 snow blasting possess two important advantages compared to conventional blasting processes. The hardness of the dry ice pellets is about 1,5 Mohs, the hardness of the CO2 snow is even lower. Due to the low hardness it is possible to machine sensitive surfaces without causing damage [2]. Furthermore CO2 exists at ambient conditions either in the solid or in the gaseous phase. During the blasting process it changes directly from the solid into the gaseous phase, it sublimes. That’s why only the removed coating has to be disposed. The gaseous CO2 is released to atmosphere. Blasting with solid CO2 uses the following two effects (figure 1):

Thermal Effect
• Embrittlement as a result of the low temperature
• Separation of the coating because of different thermal expansion coefficients between the coating and the base material

Mechanical Effect
• Mechanical impact of the CO2 particles
• Force due to the compressed air stream

Another effect, known as the sublimation effect, does not contribute to the cleaning process, but is an essential advantage of blasting with solid carbon dioxide [2].


Blasting with CO2 snow is less abrasive compared to conventional blasting processes. Nevertheless, available blasting devices for CO2 snow blasting exhibit also differences in abrasiveness due to their working principle.

Blasting devices with a high abrasiveness use one or several agglomeration chambers for generating the CO2 snow. In these chambers the CO2 snow particles can join together and strengthen due to built-in components. The compressed air, used for the acceleration of the particles, is generated by compressors with a volume flow rate of 15 m³/min at a maximum pressure of 15 bar. This high volume flow rate is necessary to accelerate the CO2 particle to the speed of sound and thus a high cleaning capacity can be achieved.

Blasting devices with a low abrasiveness use a concentric nozzle. The liquid carbon dioxide is expanded at the outlet of this nozzle and the generated CO2 snow is accelerated by a compressed jacket air stream. These snow particles are smaller than the particles of the devices mentioned before. Therefore they have a lower abrasiveness and also a lower cleaning capacity. The advantage of this variant is the low consumption of compressed air and CO2.

The combination of the advantages, low air consumption at high abrasiveness, was the goal of developing a new blasting device. This new blasting device is a development of the Institute for Machine Tools and Factory Management (IWF) of the Technical University Berlin. The provision of compressed air should be achieved by usual mobile compressors or by an existing workshop air supply. This represents the focus of applications in industrial and practical use. Beside this the blasting device has multiple setting options to match the blasting power to the cleaning task. For this purpose the CO2 mass flow rate, the size and the position of the agglomeration chamber, the blasting pressure and the geometry of the blasting nozzle can be adjusted. Furthermore the blasting device is a modular design, so it can be upgraded if necessary.

Figure 2 shows the modular design of the blasting device. It is apparent, that the blasting nozzles can be changed easily. Furthermore the adjustable agglomeration chamber can be seen. The CO2 mass flow rate is adjustable by a flow restrictor inside the inlet of the agglomeration chamber. The casing was designed to be able for setting up an array to increase the cleaning capacity.
The blasting nozzles have a laval geometry. The manufacturing of this geometry is very complex, so the nozzles were made by selective laser sintering. This manufacturing technology allows a quick production with a sufficient accuracy. The nozzles to be tested were designed and optimised, aided by the computational fluid dynamics (CFD) simulation.


For the development of dry ice blasting nozzles CFD is extensively used at the IWF. With the aid of the simulation, it is possible to analyse different nozzle geometries. So only the most promising nozzles have to be tested under real conditions. The effort for the real testing can be reduced to a minimum by simultaneously increasing the parameter variety. The simulation has additionally the effect that all conditions remain constant in contrast to real testing. For example occurring random errors due to the operator can be avoided. During the simulation all relevant global values of the flow field are calculated so they can be displayed and analysed. During tests under real conditions that is impossible. The simulation of the CO2 snow blasting process in this paper is based on the results of the dry ice blasting simulations. The results of the dry ice blasting simulation correspond sufficiently to the results of the conducted blasting tests [3]. So it can be assumed that the results of the simulation of the CO2 snow blasting also achieve a good correlation to the real flow field. The goal of this simulation was to show its feasibility and the possibilities for the further development of CO2 snow blasting. For  round and flat nozzle, simulations of the air stream both in and outside the IWF blasting device were conducted. The blasting pressure was set to 7 bar, according to the concept of this nozzle.

Figure 3 shows the flow velocities for the round and flat nozzle. It can be seen that the simulated velocities are almost the same. Tests for abrasiveness have shown that the flat nozzle has a significant lower abrasiveness than the round nozzle. All parameters are the same, so that the difference in abrasiveness can be found in the CO2 snow generation. In Figure 4a the pathlines of the air flow for the round nozzle and in Figure 4b for the flat nozzle are displayed. It can be clearly seen that for the flat nozzle a higher turbulence occurs. This turbulence disturbs the snow generation and therefore the abrasiveness is lower for the flat nozzle.

With the simulation the observed differences in abrasiveness were identified as differences in the CO2 snow formation. The nozzle types play an important role in the CO2 snow formation. It is assumed that the position and the size of the agglomeration chamber have a similar effect to the flow field inside the nozzle and therefore, also to the snow formation. The results of the simulation demonstrate the advantage of the CFD by the development of new blasting devices and nozzles. Due to the simulation, the time for development can be shortened and therefore, the costs will be reduced. The number of prototypical devices and nozzles, which have to be tested under real conditions, are reduced to a minimum.

Nevertheless it has to be taken into account that the generation of a realistic flow simulation needs a wide knowledge in the field of fluid dynamics and also in the mathematical principles of the simulation software. The analyses of the simulation and the post processing require also a deep understanding of the physical background as mentioned above.


The results presented in this paper are the basis for further investigations of the presented blasting device. Thereto further simulations of the flow field are necessary and will be conducted. In a submitted public funded research project the mechanism of the CO2 snow formation will be investigated. Also structural measures for influencing the properties of the CO2 snow particles are in the focus of this project. Beyond this, additional simulation models should be created. The simulation of a multiphase flow, such as compressed air with solid and gaseous carbon dioxide, and a sublimation model are main parts of further simulation activities.


[1] N.N.: Eigenschaften der Kohlensäure. Fachverband Kohlensäure-Industrie e.V. 2004
[2] Krieg, M.: Analyse der Effekte beim Trockeneisstrahlen. Dissertation 2008, Fraunhofer IRB Verlag
[3] Kretzschmar, M.; Uhlmann, E.: CFD-Simulation in the development of dry ice blasting nozzles. In: Proceedings of the 18th International DAAAM Symposium, Vienna, Austria, 2007, S. 403-404

Technical University Berlin
Pascalstraße 8-9, 10587 Berlin, Germany
Tel- +49.30.314 - 24 96 0, Fax 314 - 25 89 5
E-Mail: kretzschmar@iwf.tu-berlin.de