Figure 6. Comparison of P-T instrument to Mid IR ATR withinfluence of temperature / pressure spikes

Figure 6 above shows the comparison of ATR CO2 measurement and an in-line P-T measurement method. The P-T instrument is subject to erroneousreadings due to pressure spikes as shown in the graph. This data is notuseful for process control and must be filtered out of the data stream.

Figure 7. CIP cycle measurement

Figure 7 above shows the measurement of CO2 using ATR before, during,and after a CIP cycle. In-line P-T measurement methods are unable tomeasure, and are in fact unusable, during this period and typically for 10-15 minutes after CIP termination, while the ATR method shows theprocess engineer a clear picture of the cleaning cycle.

CONCLUSION

There is a preponderance of evidence to show that the Mid IR ATRmeasurement of CO2 concentration method offers many advantages overtemperature-pressure methods. CO2 specificity and elimination of errorsdue to alcohol and the specific gravity of the beer, (which affect thesolubility of CO2 in beer), are key attributes. When used in a processstream for automated production management, further benefits arerealized when compared to manual sampling including 24 x 7 operation,minimization of human error and the ability to see related qualityparameters such as CIP cycles. All of the mentioned factors areinterrelated and contribute to the cost and quality of the beer as impacted byrework costs, product waste, utility costs, engineering time, and throughput.

The selectivity of the Mid-IR ATR measurement method produces a “true CO2”measurement of dissolved CO2. In most cases, the true CO2 measurementswill differ from traditional temperature-pressure derived measurements.The advantages of the true CO2 measurement include accuracy,repeatability and the subsequent relationship of these parameters to theproduction, shelf-life, drinkability and quality of the beer.

REFERENCES

1. The In-Line Determination of Carbon Dioxide in Beer by Infrared Analysis,P.A.Wilkes, MBAA Technical Quarterly Vol 25, No.4, pp.113-116

2. Attenuated Total Reflection Spectroscopy method for measuring dissolvedCO2 concentration in Beer, R. O’Leary Thermo Electron Corporation

3. The Brewers Handbook, T. Goldhammer, KVP Publishers, 1999pp.312-314, 318

NOVEL ONLINE SENSOR FOR MEASURING DISSOLVED CO2

USING ATTENUATED TOTAL REFLECTANCE (ATR) TECHNOLOGYF.M. Cash,1 G.P. Broski,1 P.J. Slier1

1Thermo Fisher Scientific

2007 ASBC Annual Meeting June 16 – 20, 2007 Fairmont Empress

Victoria, British Columbia, Canada

ABSTRACT

Traditional methods for measuring carbon dioxide both in the laboratoryand online are based on Henry’s Law and use pressure and temperature tocalculate dissolved CO2 concentration. These methods are not specific tocarbon dioxide and measure all gases including oxygen and nitrogen. A novel sensor has been developed which measures only dissolved CO2

and is suitable for online applications. Based on infrared spectroscopy, thesensor uses the Attenuated Total Reflectance sampling technique tocontinuously analyze the beverage in flow and no flow applications. Thesensor is compact and is designed to fit directly into the process. Theoperation and design of the sensor are discussed. The performance of theATR-based sensor will be compared with traditional temperature-pressurebased devices. The ATR-based sensor provides improved measurementaccuracy which is not dependent on beer type, density or alcohol content,is not affected by pressure surges, and can operate during CIP cycles.

INTRODUCTION

There are many references to the topic of CO2 control in beer. CO2 levelsare universally considered critical to final product quality from thesupplier’s, as well as the consumer’s, perspective. Taste, presentation,drinkability and packaging of the final product depend on reliable controlof this ingredient.

During processing of ‘final beer’ dissolved carbon dioxide levels aremeasured along with other key parameters, including specific gravity, realextract, alcohol %, temperature, etc. Manufacturers require CO2 levels tobe within the quality control standards set by their brewing chemists andprocess engineers. These levels in the U.S. may vary from lower levels of1.5-2.0 Volume/Volume on (Ales) to higher levels of 2.5-2.8 Volume/Volume on (Lagers). To account for shelf loss in plastic packaging, theselevels may be increased ~10%. The allowable ranges for the CO2 readingsfrom a tank may be ±0.1 V/V , or ~4% to 5 %.

Traditional methods for measuring carbon dioxide both in the laboratoryand online are based on Henry’s Law which uses pressure andtemperature measurements to calculate the dissolved CO2 concentration.These methods are not specific to carbon dioxide and measure all gasesincluding oxygen (O2) and nitrogen (N2).

There is widespread industry understanding and acceptance of thePressure - Temperature (“P-T”) measurement method and its limitations.Pressure / Temperature tables have been developed that obtain a goodcorrelation to dissolved CO2 in “standard beer.” These relationships have further been integrated into “on-line” systems that determinedissolved CO2 from pressure and temperature sensors in a slipstream of the brewing process.

It is not uncommon for a large brewer to integrate in-line CO2 instrumentsbased on these P-T relationships into their plant-wide Distributed ControlSystem, (DCS). Through the DCS, real-time monitoring of critical CO2

levels may be made available to the operations personnel, so decisions onproduct release to packaging, or blending ratios, can be made withconfidence and traceability. In some cases, the level of CO2 isautonomously controlled through the process-control software.

The days of a single plant or brewer making one brand of product arevirtually gone. Process engineers are challenged to accommodate multiplebrands, and brand change-over’s in the course of a week or even in thesame day.

These multiple brands may vary fairly widely by alcohol % and density suchthat a standard beer P-T dissolved CO2 measurement will no longer servethe process as well as desired. Density directly affects the solubility of CO2

in beer. An improved method, using ATR, or Attenuated Total ReflectanceMid-Infrared Spectroscopy, has been developed to address these limitations.

MATERIALS AND METHODS INCLUDING DESIGN

Attenuated Total Reflectance Mid-IR Spectroscopy is well suited forapplication in the beverage industry (Figure 1) for the following reasons.First, the CO2 molecule has a strong absorption mode in the mid-infraredregion around 4.3 microns. Second, the properties of ATR, which are well-known in laboratory science as a method for increasing the signal

of interest without corresponding increases in system noise, are perfectlysuited to in-line process measurement.

In this case, evanescent waves from the IR source couple into the beverageon each of multiple ‘bounces’, being preferentially absorbed by CO2

molecules at each instance.

Figure 1. ATR Reflection

The lack of other absorbents at the 4.3 micron region ensures theabsorption is based solely on CO2. A reference sensor factory calibrated tothe thermal properties of the sensor ambient maintains stableperformance over the process range temperature of the finished beer.

Because the ATR CO2 system measures CO2 absorption selectively, plants using these types of systems need not be concerned with thefactors that contribute to CO2 measurement errors:

• Dissolved gases such as O2 and N2

• Alcohol %• Specific Gravity• Headspace errors (from bottle/can shaker testing)• Manual sampling errors

The ATR CO2 measurement method produces stable and accuratemeasurements for brewers to control their processes with confidence,avoiding unnecessary rework of fundamentally good product, (but thoughtto be out of specification when measured as a ‘standard beer’), or worse,passing product either too high or too low in “True CO2”.

Figure 2. ATR Measurement Fundamentals

The basic system employed for our discussion is shown above in Figure 2.It incorporates an infrared light source, a sapphire crystal for theattenuated total reflectance sampling, and two detector channels, areference and an active channel.

Figure 3. CO2 Infrared Spectrum

Figure 3 above illustrates the infrared absorption spectrum of CO2. Thereference filter obtains a signal from a region far below the active filtercentered on the area around 4.3 um.

It is plainly visible from this figure that signal from gases such as N2O(shown) are in areas that cannot be passed by either reference of activefilters, and therefore can have no influence on the sensor reading.

An ATR CO2 sensing system can be placed without modification into astandard VariVent process port. There are only three materials in contactwith the product: Stainless Steel, Sapphire, and a PMMA gasket forsealing. The sensor has a sanitary design for the product interface and isalso environmentally protected for survival, external to the product stream.

The lack of pistons, slipstreams, membranes or diaphragms, or any other moving parts, reduces maintenance issues significantly. There are no consumables associated with the solid-state ATR sensor.

Additionally, pressure variations that ‘blind’ on-line P-T measurementinstruments do not affect ATR CO2 systems. This keeps measurements and available closed-loop controls operational during critical start-up after CIP cycles.

RESULTS AND DISCUSSION

Results from ATR CO2 measurements and P-T measurements have beensuccessfully compared over the last few years at various brewery sites.Selected results are shown below.

Figure 4 shows ATR CO2 measurements made over a period of 19 hours. During this period, five production cycles and three brandchanges were recorded. In addition, Alcohol and specific gravity wererecorded at each sample point noted with “X”.

Figure 4. CO2 in beer results. Five production cycles, three brandchanges with no corrections

Figure 5. CO2 in beer results. Five production cycles, three brandchanges, corrected for alcohol and density

Figure 5 above shows the same data from the offline measurementmethod (figure 4), corrected for alcohol and specific gravity.

BEER Attenuation

ATR Crystal

IR Emitter IR Receiver

b ca

Carbonation Sensor Host System

Host Processer

SerialCommunication

InfraredLight

CO2

Unit Status

Beer’s Law

[∑ ( )]T_filter =gwi

e-α, xxL

-I0 + Lxx

Sum of all Frequencies

CO2

Digital Data010111001010111001110011100100111001

CO2 Filter

Ref Filter

DataAcquisition