SPE 62525 Application of Semipermeable Membrane Technology in the Measurement of Hydrocarbon Gases in Drilling Fluids

A.O. Brumboiu*, SPE, D.P. Hawker*, SPE, D.A. Norquay*, *Datalog Technology Inc.; D.K. Wolcott, FIA Solutions

Copyright 2000, Society of Petroleum Engineers Inc. This paper was prepared for presentation at the 2000 SPE/AAPG Western Regional Meeting held in Long Beach, California, 1923 June 2000. This paper was selected for presentation by an SPE Program Committee following review of information contained in an abstract submitted by the author(s). Contents of the paper, as presented, have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material, as presented, does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members. Papers presented at SPE meetings are subject to publication review by Editorial Committees of the Society of Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper for commercial purposes without the written consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O. Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.

and inconsistencies previously associated with gas extraction from the drilling fluid. Housed within a small probe that can be positioned directly within the returning mud stream, the semipermeable membrane provides quantitative data that benefits both formation evaluation and drilling safety. Through various stages of testing and development, the paper will illustrate how the present semipermeable membrane wellsite system signifies the starting point for a whole new generation of gas analysis devices.

Abstract For three-quarters of a century, the detection and analysis of formation gas released into the drilling fluid and returned to surface, has been used to identify the presence of petroleum reservoirs. The value of this data source has been its immediacy. Reservoir zones can be evaluated while they are being penetrated for the first time and observations are not subject to post-drilling changes to the formation, such as mud invasion, that can limit the effectiveness of other evaluation techniques such as wireline logging. However, although the acquisition of this data is widespread in the petroleum industry, it has long had the reputation of being an unreliable source of data with inconsistent results. The inconsistencies of surface gas measurement results largely from the way that the gas is extracted from the drilling fluid. Virtually unchanged through the history of mud logging, dissolved gas is broken out of solution by applying some form of agitation to the mud. The released gas is held within a trap and transported to a remote gas analyzer by a flow of air. There are many variables and inconsistencies in this process that result in a purely qualitative gas measurement and leave important questions unanswered. Namely, how much gas is actually present in the drilling fluid and what exactly is the composition? This paper details the development of a new generation of gas measurement technology (Fig. 1) employing semipermeable membranes, providing a significant improvement in real-time wellsite gas-in-mud analysis and answers many of the previously unanswered questions. Laboratory and field results show how the semipermeable membrane functions and removes many of the inefficiencies

Fig. 1 The wellsite gas measurement system comprising gas extraction through the semipermeable membrane and analysis at the point of installation.

Introduction Analysis of formation gas returned to surface in the drilling fluid has long been an important first appraisal of a potential reservoir zone, providing important data to guide subsequent evaluation from wireline logging and testing. Furthermore, if formations become invaded or damaged after they are drilled, or if tools cannot reach the zone of interest, initial gas-in-mud analysis may provide the only reasonable data by which to evaluate a well.

A.O. BRUMBOIU, D.P. HAWKER, D.A. NORQUAY, D.K. WOLCOTT

SPE 62525

Despite this, the evaluation provided by gas analysis is often over-looked and mis-understood. This results from the qualitative and inconsistent nature of the data stemming from the way that the gas sample is extracted for analysis. Problems With Agitation. Agitator traps have remained largely unchanged for many years and have been used, virtually exclusively in the industry, to extract the gas sample. Yet, those who have used such devices have recognized the many problems associated with the technique, as detailed below. The extraction efficiency of different hydrocarbon gases varies with a number of different parameters. Gas density and solubility the heavier and more soluble the gas, the less efficient the extraction. Mud type oil base muds and mineral oils hold a greater amount of methane in solution, therefore less is likely to be extracted compared to a water base system. In addition, mud density and viscosity reduce the rotational speed of agitators and reduce extraction efficiency. While the gas volume in the drilling fluid remains in solution, the relative extraction efficiency is the only real issue, since air will be re-introduced into the trap at approximately the same rate as the sample is drawn out. However, when the drilling fluid is saturated to the point that gas also appears as bubbles, air is displaced from the trap vessel so that the apparent gas-air concentration is increased. The non-linearity resulting from such trap saturation leads to erroneously high measurements that hinder both formation evaluation and well safety. Fig. 2 illustrates a similar phenomenon where the air volume within a trap can be reduced by increased mud volume, reduced airflow or by a change in the velocity of wind entering a trap port. At approximately 19:00 hrs, a gas response is equally detected by both trap and membrane gas measurements, reflecting a true increase in gas volume in the drilling fluid. However, with the second group of responses, the CO2 level drops, resulting from the reduced air volume in the trap vessel as the mud level increases. The corresponding chromatographic and total gas values show an erroneous response as the gas-air concentration is falsely increased. Over the same interval, note that there is no such increase in the gas measurement sampled through the semipermeable membrane. Variations in mud flow and/or mud level have a dramatic effect on gas trap extraction. With a reduction in mud flow, and corresponding drop in the mud level at the trap location, less mud volume is being sampled while, at the same time, the air volume within the trap increases, lowering the gas-air concentration and causing a drop in gas measurement. Fig. 3 shows a period of 27 minutes where the pump rate was reduced from 118 to 85 SPM. Note how, over the corresponding period, chromatographic gas measurements obtained from a gas trap are lower than the levels seen with the higher pump rate. This reduction is a false reduction, caused simply by a slight drop in the mud level at the shaker box, and does not reflect a change in actual gas concentration in the mud. Observe how the gas sampled by the gas membrane, being unaffected by mud level, actually shows a slight increase in volume over the same period.

The relative proportion of air can also vary due to wind variations. Work by Texaco/GRI1 demonstrated that a 50% drop in gas measurement resulted from an 8-mph wind being blown directly into the trap port and increasing air dilution.

Fig. 2 Illustration of conventional gas trap saturation, or reduction in air volume leading to erroneous gas responses.

In addition, the air dilution and transit of gas-air mixture from the trap to the remote analyzer results in a delay in analysis time as well as the integration and poor definition of gas responses (peak spreading). Colder air temperatures may also cause the heavier hydrocarbons (certainly pentanes, often butanes, and, sometimes, even propane) to condense in the gas sample line and subsequently be trapped by filters. Thus, they are prevented from reaching the analyzer or, worse, released at a later time. The trap, being a mechanical device, is prone to changes in operational efficiency or breakdown. It is limited in location to the shaker box at the end of the flowline (by which time methane will have escaped), and is limited in underbalanced drilling applications that use gas or foam type drilling fluids.

SPE 62525APPLICATION OF SEMIPERMEABLE MEMBRANE TECHNOLOGY IN THE MEASUREMENT OF HYDROCARBON GASES IN DRILLING FLUIDS 3

Fig. 3 Comparison of gas measurements made with a conventional gas trap and a gas membrane, showing the effect of changes in flowrate and mud level.

What results from these limitations and effects is an extraction process that only provides a qualitative gas measurement with exact quantity and composition remaining unknown. For the oil company, this means that it is very hard or, in some cases, impossible, to reliably correlate results between wells, either from the same field or from the same formation in different fields. Improvements to Gas Analysis. Several techniques have been introduced to improve the quality and repeatability of surface gas-in-mud measurements. The basic evaluation technique2 of evaluating gas responses over and above a

recycled and contamination3 background gas level removes the influence of many of the variables, but still doesnt yield accurate quantitative data in terms of volume and composition. Many techniques have been developed to compare chromatographic components and use quantitative gas ratios4,5,6,7 to evaluate zones of interest. This has certainly improved the quality and usefulness of the data, allowing for reservoir fluid characterization during and post- reservoir penetration, but the basic limitation of variable extraction efficiency remains. The exact gas volume in the mud, together with the relative proportion of individual components that are extracted, remain an unknown.

A.O. BRUMBOIU, D.P. HAWKER, D.A. NORQUAY, D.K. WOLCOTT

SPE 62525

The most significant improvement, to date, has been the development of a quantitative gas measurement (QGM) system by Texaco and the Gas Research Institute1. Here, the trap design8 successfully eliminates the influences of mud level and wind variation to provide a constant mud volume sample. However, the variation in extraction efficiency of individual components remains a problem. To alleviate this, periodic gas measurements are made from a mud sample using a steam still9. The gas-in-mud measurement obtained is then used to provide a correction factor to the gas-in-air measurement obtained from the trap. However, the dissolvedfree gas transition can still lead to non-linearity in the gas measurement.

Gas measurements can also be normalized by using a reference gas10,11 with known dispersion, extraction and detection properties in relation to the formation gases. The resultant measurement, of a known volume of reference gas, after a circulation through the wellbore, can be used to calibrate the formation gas measurement. Unfortunately, one reference gas cannot adequately compensate for all hydrocarbon gases being analyzed. The semipermeable membrane system12 improves the quality of data by removing the problem at the source, employing the new semipermeable membrane technology, positioned directly in the returning mudstream, by way of a probe. Using the semipermeable membrane technology13,14, a more accurate determination, by volume, of gas in liquid can be derived15. The quantitative measurement is derived regardless of whether the gas is dissolved or present as bubbles (free gas). Without the traditional trap vessel and associated dead volume of air diluting gas samples, gas responses are much better defined and can be more accurately correlated to depth. Analysis is at the point of extraction, providing very rapid resolution compared to conventional systems (Fig. 4) and, in addition, the paper will demonstrate the improved linearity and better ratiometric extraction capability of the membrane technology. The Semipermeable Membrane Basic Principle A semipermeable membrane is generally considered to be impermeable to liquids, while permeable to gases. Gas permeation through the membrane wall is driven by the difference in the partial pressures, the pressure outside the membrane wall and the pressure inside, of that particular gas. There are a variety of membrane type permeation devices made from natural or synthetic polymers, and the particular type would be selected according to the gas, or gases, that are to be extracted and, also, to the carrier fluid and its particular chemical and mechanical resistance. The geometry of the membrane device may also vary according to the specific application although the most common used is a disc shaped device. For the application of hydrocarbon gas detection in drilling fluids, a specially designed device is used where a capillary type membrane is wound, in a double helix configuration13, around a support groove (Fig. 5). This design has several advantages over the basic disc type, including a very low dead volume inside the membrane tube and a much greater surface area in contact with the fluid. In addition, the resulting cylindrical probe (membrane plus support) only requires an inch diameter space, allowing for installation in hard to reach locations such as flow pipes and tank walls. In order to keep, on a continual basis, the maximum difference between the outside and inside partial pressures, the concentration of the specific gas has to be kept close to zero inside the membrane. This is achieved by flowing a neutral carrier gas, such as nitrogen, helium or air, through the inside of the membrane.

Fig. 4 Log illustrating the high resolution and fast response time obtained from the semipermeable membrane.

SPE 62525APPLICATION OF SEMIPERMEABLE MEMBRANE TECHNOLOGY IN THE MEASUREMENT OF HYDROCARBON GASES IN DRILLING FLUIDS 5

When measuring gas-in-air mixtures, Daltons Law states that the partial pressure (Pi) of the gas is equal to the quantitative concentration (Qi) within the mixture.

In other words, a mixture with almost twice as much methane as propane would result.

Qi = Pi....................................................Eq.1
In this case, the permeating flow of hydrocarbon gas through the membrane wall and, therefore, the concentration of gas measured by the analyzer attached to the membrane, are directly proportional to the concentration (Qi) of the gas in the mixture.

P1

P3

= (1 / 0.033)

(1 / 0.065)

= 1.97

Therefore, when extracting hydrocarbon gases from drilling fluids using mechanical agitation, boiling, or membrane permeability, the quantitative approach must take into consideration the distinctive solubility of the individual hydrocarbon gas. The heavier the gas (the more carbon molecules) the higher the solubility and the lower the extraction efficiency. Inverse Measuring Procedure. Comparison with the extraction principles of agitation and boiling was simulated with an inverse measurement experiment (possible with the gas membrane technology due to its high sensitivity). Rather than the normal measurement of gases released by agitation (gas trap principle) or boiling (the basic principle of the steam or microwave still), the experiment was conducted in order to determine the volume of dissolved hydrocarbon gases still remaining in different liquids after incremental periods of time. Agitation Of Water. During a period of agitation, a mixture of methane, ethane and propane, dissolved in water, is released selectively (methane better than ethane, better than propane) as shown in Fig. 6. After 15 minutes of continual agitation, relative to the original amounts, 10% of methane remained, 22% of ethane and 32% of propane. Boiling Of Synthetic Oil. A similar experiment was conducted through boiling, to determine the amount of hydrocarbon gases remaining in solution, in synthetic oil base mud, after prolonged boiling together with agitation supplied by an impeller. Naturally, the solubility of hydrocarbon gas is much greater in an oil base mud than in water. A total liquid volume of 1 litre was used for the experiment, holding the following initial gas amounts: 27,000ppm methane, 66ppm ethane, 64ppm propane, 34ppm i-butane, 29ppm n-butane, 28ppm i-pentane and 21ppm n-pentane. Fig. 7 illustrates the relative efficiencies in hydrocarbons released during the period of boiling. After 15 minutes, methane was no longer detected by the chromatograph (operating at 1ppm resolution), almost half of the initial propane remained and close to 80% of the initial pentanes were still present. After a further hour had passed, the propane was now completely released whereas over 20% of the pentanes still remained in solution. These experiments show the difficulty in quantitatively extracting hydrocarbon gases from liquids, especially from oil base or invert type muds (in this case synthetic oil), and illustrates the necessity to apply correction factors in order to provide truly quantitative analysis. In order to meet this requirement, a new volumetric calibration method was developed (described later in this paper).

Fig. 5 Schematic of the probe and semipermeable membrane

This leads to the straightforward application of the device to measure the hydrocarbon concentration, in air or nitrogen for example, used during underbalanced drilling applications, an area where traditional gas detection has been poor or impossible. Measuring Dissolved Gases in (Drilling) Fluids The same basic principle of gas permeation through the membrane wall is used to determine the volume of dissolved gas in a given liquid. According to Henrys Law of Solubility, the amount of gas dissolved in a liquid is directly proportional to the partial pressure of that gas above the liquid surface. This relationship may be expressed as:

Qi = KiPi......or.....Pi =

1 Qi..........................Eq.2 Ki

Usually, hydrocarbon gases are kept dissolved in liquids by Van der Waals forces and the solubility increases with the number of C-H bonds in the hydrocarbon molecule. The same applies to the liquid, where water has a lower solubility for methane than diesel. As shown in Table. 1, column 1, the solubility of methane through butane increases significantly with the number of carbon atoms in the molecule (Ki constant in Eq. 2). For example, if equal amounts of Q1=1% methane, and Q2=1% propane, are dissolved in water at the same time, the gas mixture above the water surface, according to Eq. 2, is:

P1

P3 =

(Q1 / K 1) ( Q 3 / K 3)

.......... .......... ...... Eq .3

A.O. BRUMBOIU, D.P. HAWKER, D.A. NORQUAY, D.K. WOLCOTT

SPE 62525

The Quantitative Measurement of Methane using the Semipermeable Membrane The first analyzer, using the semipermeable membrane technology, was designed to measure the methane since it is the most abundant hydrocarbon gas and associated with both gas and oil reservoirs. For this application, a methane selective semipermeable membrane was required and the analyzer employed was based on a catalytic combustible detector. The configuration of the final, complete device, is shown in Fig. 1. The membrane is wound around the probe located at the end of a flexible steel hose that carries the air and gas mixture and includes filtering elements. At the other end of the hose, an explosion proof box houses many features including the detector, air pump and processor. Volumetric Calibration. The device was initially calibrated for the partial pressure of methane in air, using known concentrations from readily obtainable cylinders. For measuring the methane volume in drilling fluids, a new calibration method was devised that required a volumetric injection of gas, into a known volume of liquid, within a sealed, closed loop calibrator. The calibrator possesses an impeller type circulating pump that provides a continuous flow of fluid over the probe and membrane and, simultaneously, homogenizes the gas-liquid mixture. With the device calibrated, initially, to determine the methane concentration in air, the calibration procedure in liquid was carried out using three varieties of drilling mud, namely, a water-base mud (WBM), a diesel invert mud (DIM) and a synthetic (mineral) oil-base mud (SOBM). The solubility of methane in the three mud types was 3.1% (WBM), 11.8% (DIM) and 15.3% (SOBM). The response of the probe is shown in Fig. 8 and, as can be seen, the response from all three mud types is characterized by good linearity. The linear approximation modeling (as illustrated in Fig. 9) of the three slopes shows a higher response for the water-base mud, with a slope of 6.251. The response from the diesel invert is much lower, with a slope of 1.604, whereas the response from the synthetic oil is lower still, with a slope of 1.319. These differences in response correlate nicely with the solubility of methane in the three mud types (the higher the solubility, the lower the slope) and according to the theoretical relationship of solubility (Eq. 2). For example, compare the relationship between the WBM and SOBM slopes when derived experimentally and from the relative solubilities. Experimentally, the ratio WBM:SOBM equals 4.74 (6.251/1.319). From the solubility values, the slopes are equal to 1/K (as defined by Eq. 2). For WBM, this is equal to 1/0.031 and, for SOBM, it is equal to 1/0.153. This gives a ratio of WBM:SOBM equal to 4.93, comparable to the 4.74 derived experimentally. Following the same procedure to compare the slopes for DIM and SOBM, the experimental ratio is 1.216 and that derived from the solubilities 1.296, again comparable.

When the volume of gas entrained in the drilling fluid is above the saturation limit, gas bubbles are released at the mud surface, so that we also have the presence of free gas. The free gas, even when measured with the same type of units (% gas by volume mud) provides a much lower membrane response than the same concentration of dissolved gas in liquid. The free gas behaves in a similar fashion to a free gas mixture as described by Daltons Law (where partial pressure equals the percent by volume). This foaming type of mud is very difficult to reproduce for calibration purposes, but a theoretical linear model is used to describe it. Fig. 9 shows the membrane response to varying amounts of methane in a diesel invert drilling mud, at concentrations above (free and dissolved gas) and below (dissolved gas) the saturation limit of approximately 12%. Experimentation displays a slope change between the two states of gas, from an initial dissolved gas response slope to a second, free gas response slope. Linear Modeling Of Gas Behaviour. A linear model (Fig. 10) is illustrated to describe the response. The graph shows that the concentration of the gas measured in the liquid consists of two distinct phases around the saturation limit. The portion of the measured curve representing the gas in solution increases rapidly as the overall concentration is increased up to the saturation limit (S). The dissolved behaviour linear equation is determined from experimental volumetric injection values below the saturation limit. Beyond this point, the amount of dissolved gas slowly declines towards zero, as the percentage of gas bubbles increases and displaces liquid. In theory, the gas cannot begin to exist as foam bubbles until the saturation point has been exceeded. However, in practice, this is not a sharply defined inflexion point. There will always be a region in which there exists both bubbles and dissolved gas around the saturation point of the curve. Ideally, the percentage gas as foam in the liquid rises from zero, beginning at the saturation point. As the total concentration is increased beyond the saturation point, increasingly more of the gas is in bubble form, displacing more of the liquid, until a concentration of 100% gas is reached, at which point (Y100) no more liquid is present. As more gas is introduced, an equal amount of liquid is necessarily displaced and thus, the amount of gas in solution decreases to zero. The overall representation of the gas, by percent volume in the liquid, above the saturation point, is the sum of the bubble curve and the solution curve. The characteristic curves for all gases in fluids (including air) will have two common end points at 0% and 100% concentrations. At 0%, there is no gas to measure, and the measurement can be adjusted to zero, regardless of the fluid type. At 100%, there is only gas present (assuming a constant volume) and again the measured concentration is easily reproducible. Between these two points, a mathematical model, as described above, exists. Below the saturation point of the gas in the liquid, the measured concentration is determined by only the amount of gas in solution, while above the saturation point it is determined by the maximum soluble gas plus the free gas present as entrained bubbles.

SPE 62525APPLICATION OF SEMIPERMEABLE MEMBRANE TECHNOLOGY IN THE MEASUREMENT OF HYDROCARBON GASES IN DRILLING FLUIDS 7

The model allows the free gas response slope to be determined by equation 4:

Slope = (Y 100 Ys )

(100 S )

.........................Eq.4

This predicts the response behaviour without the need to experimentally reproduce the difficult foaming mixtures or sample foam from the mudstream. The Chromatographic Gas Measurement For detecting the hydrocarbon gases of interest, usually methane through pentane, another membrane is used that permeates all of the required gases. In order to optimize the heavier hydrocarbon gas extraction (butane and pentane), which as seen in Table 1, column 2, are more readily held in solution, a membrane that permeates the heavier hydrocarbons more efficiently is used (see column 3). The permeation characteristic of the membrane, when exposed to a mixture of air and hydrocarbon test gas (column 4) can be seen in columns 5 and 6. Note that the membrane permeates pentane almost twenty times more efficiently than methane (column 6 slope relative to C1). When the same test gas mixture is applied to the volumetric injection calibrator, the percentage gas values obtained, by volume, in water, are shown in column 7. The response of the membrane, shown in column 8, is close to a true concentration, illustrating a near to true ratiometric extraction even before the individual component calibration from the chromatograph. The rising response slopes, obtained from each component, are compared to the methane slope in column 9, although the accuracy is subject to the 1ppm resolution capability of the chromatograph. A more detailed family of response curves for methane through pentane mixtures, in water, can be seen in Fig. 11. The response slopes of the gases are fairly close, ranging between 0.88 for n-pentane and 1.57 for propane. The situation becomes more dramatic when dealing with fluids with higher solubilities for hydrocarbon gases. Fig. 12 shows the responses obtained from a methane through pentane mixture in a synthetic (mineral) oil base drilling fluid. The responses show good linearity, but the extraction efficiency of the membrane cannot compensate for the huge increase in solubility characteristic of the synthetic oil. Slope ratios between the high:low responses are now as high as 10 to 1, much larger than the slope ratio of 1.8 to 1 obtained from the lower solubility water base mud. Conclusions The physical attributes and operation of the semipermeable membrane device provide many significant improvements to gas-in-mud measurement at the wellsite. Variations in parameters such as mud flow, mud level and wind velocity have no effect on the gas extraction through the membrane and resulting gas-air concentration.

With the analyzer located at the point of installation, faster response times with improved definition and depth correlation result, leading to enhanced formation evaluation in terms of accurate determination of reservoir fluid contacts, fractures etc. It also ensures that none of the heavier hydrocarbons go undetected (through condensing in a gas sample line). The size of the probe allows for many more location possibilities for gas measurement, including the mud flow return line and suction line to provide accurate differential gas measurement. The ability of the membrane technology to operate in any fluid type makes the system equally applicable in conventional and underbalanced drilling operations. It even allows for possible applications in ambient gas detection, such as the presence of gas in enclosed tanks or spaces, degasser efficiency, etc. Conventional gas analysis has always suffered from variations in gas extraction, and non-linearity, caused by different drilling fluid types, gas density and solubility, and the transition between the two gas states (dissolved or free). The new volumetric calibration method and the linear modeling technique, described in this paper, show that the behaviour of different hydrocarbon gases, in various drilling fluids, can be more accurately predicted. Taking into consideration the varying permeation rates of gases through the membrane allows for a more accurate quantitative gas measurement to be obtained. At present, the semipermeable membrane devices are being used by the petroleum industry for the methane application described in the paper. The chromatograph device is still at a prototype stage undergoing laboratory and field tests with different drilling fluid systems. Acknowledgements We would like to express our sincere appreciation to Larry Krusel and Sam Quantz of PetroCanada for allowing us to conduct wellsite tests during their drilling operations. Our thanks also go to the wellsite geologists Bill Sparrow and Bob Redford for helping to manage two simultaneous gas detection systems. We would like to thank the management at Datalog for supporting the project and for granting permission for this paper to be published. Additional thanks go to Dave Occleston and Mark Pickell for their assistance in the field testing program. Nomenclature a given gas; where i=1, the gas is methane i= for example solubility coefficient of a gas in liquid, as Ki = shown in table 1 partial pressure of a gas above a liquid Pi = surface , % volumetric concentration of a gas in liquid, Qi = % Saturation limit in a liquid S=

A.O. BRUMBOIU, D.P. HAWKER, D.A. NORQUAY, D.K. WOLCOTT

SPE 62525

Ys = Y100 = References

the response of the probe at the saturation limit of the gas in liquid the response of the probe at 100% gas

1. Wright, A.C., Hanson, S.W. and DeLaune, P.L.: A New Quantitative Technique for Surface Gas Measurements, SPWLA 34th Annual Logging Symposium, June 13-16,1993. 2. Whittaker, M.: Mud Logging Handbook, Chap 6 The Formation Gas Show, 186-192, Prentice- Hall Inc, 1991. 3. Mercer R.F.; Liberated, Produced, Recycled or Contamination, SPWLA 15th Annual Logging Symposium, June 2-5, 1974. 4. Pixler, B.O.: Formation Evaluation by Analysis of Hydrocarbon Ratios, SPE 2254 presented at the 43rd Annual Fall Meeting, 1968, Sept 29-Oct 2. 5. Ferrie, G.H., Pixler, B.O. and Allen, S.: Well-Site Formation Evaluation by Analysis of Hydrocarbon Ratios, Petroleum Society of CIM, 81-32-20, presented at the 32nd Annual Technical Meeting, Calgary, May 3-6, 1981. 6. Haworth, J.H., Sellens, M.P. and Gurvis, R.L.: Reservoir Characterization by Analysis of Light Hydrocarbon Shows, SPE 12914 presented at the 1984 Rocky Mountain Regional Meeting, May 21-23. 7. Haworth, J.H., Sellens, M. and Whittaker, A.: Interpretation of Hydrocarbon Shows Using Light (C1-C5) Hydrocarbon Gases from Mud-Log Data, AAPG Bulletin, V69 No8, Aug 1985, 1305. 8. Wright, A.C. et al: Gas Trap, U.S. Patent # 5,199,509, 1993 9. Moore, P.J.: Steam Still, U.S. Patent # 3,050,449, 1959 10. Amen, R.M.: Quantifying Hydrocarbon Shows Using On-Line Gas Referencing, SPWLA 35th Annual Logging Symposium, June 19-22, 1994. 11. Amen, R.M.: Gas Referencing, U.S. Patent # 5,277,263, 1994 12. Hawker, D.P.: Direct Gas in Mud Measurement at the Well Site, Petroleum Engineer International, V72 No9, Sept. 1999, 31-33. 13. Westlake and Wolcott, D.K.: Semipermeable Membrane Technology, U.S. Patent # 5, 317, 932, 1994 14. Wolcott, D.K.: Use of Capillary-Membrane Sampling Device to Monitor Oil Drilling Muds, U.S Patent # 5, 469, 917, 1995. 15. Brumboiu, A.O. and Norquay D.A.: Method for Determining the Concentration of Gas in a Liquid, US Patent Applied for, 2000. 16. Dean J.A.: Langes Handbook of Chemistry, 14th Edition, McGraw-Hill, Inc. 1992.

SPE 62525APPLICATION OF SEMIPERMEABLE MEMBRANE TECHNOLOGY IN THE MEASUREMENT OF HYDROCARBON GASES IN DRILLING FLUIDS 9

100 90
Hydrocarbons remaining (% of initial amount)

%C1 %C2 %C3

80 70 60 50 40 30 20 10 0
0 2 4 6 8 10 12

14

16

Elapsed time (minutes)

Fig 6

Hydrocarbon fraction remaining dissolved in water during agitation

100

%C1
90 80 70 60 50 40 30 20 10 0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80

%C2 %C3 %iC4 %nC4 %iC5 %nC5

Hydrocarbons remaining (% of initial amount)

Elapsed time (minutes)

Fig 7 Hydrocarbon fraction remaining dissolved in synthetic oil after prolonged boiling

10

A.O. BRUMBOIU, D.P. HAWKER, D.A. NORQUAY, D.K. WOLCOTT

SPE 62525

14 12 10 8 6 4 2 0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 C1 dissolved, % by volume

WBM natural WBM linear Diesel Invert Diesel Inv. linear Synthetic Oil Synth. linear

Membrane response

Fig 8

Methane dissolved in various drilling fluids

35 30

Membrane response

25 20 15 10 5 0
natural linear approximation of dissolved gas response linear approximation of free gas response

10 15 20 % C1 by volume

25

Fig 9

Linear modeling of dissolved and free gas responses

SPE 62525APPLICATION OF SEMIPERMEABLE MEMBRANE TECHNOLOGY IN THE MEASUREMENT OF HYDROCARBON GASES IN DRILLING FLUIDS 11

Fig. 10 Theoretical modeling of free and dissolved gas in a liquid

Solubility coefficient in water at STP (Ki) Methane (C1) Ethane (C2) Propane (C3) i-Butane (iC4) n-Butane (nC4) i-Pentane (iC5) n-Pentane (nC5)

Solubility relative to C1, in water

Permeability in membrane

Test gas in air (% vol)

Membrane Response (ppm)

Slope relative to C1

Dissolved Gas in Water (ppm by vol)

Membrane Response (ppm)

Slope relative to C1

0.033 0.047 0.065 0.15 0.15

1.0 1.424 1.97 4.5 4.5

1.0 2.5 6.154 7.69 7.69

1.5 0.1 0.1 0.1 0.1 0.1 0.1

700 173 347 469 694 846 931

1.0 3.7 7.4 10.03 14.86 18.10 19.98

59 3.94 3.94 3.94 3.94 3.94 3.94

61 5 6 4 5 5 4

1.0 1.227 1.43 0.981 1.227 1.227 0.981

Table 1 Chromatographic permeation analysis

12

A.O. BRUMBOIU, D.P. HAWKER, D.A. NORQUAY, D.K. WOLCOTT

SPE 62525

80

70

60

Membrane response

50

C1 ppm C2 ppm C3 ppm

40

iC4 ppm nC4 ppm iC5 ppm

30

nC5 ppm

20

10

0 0 5 10 15 20 25 30 35 40 45 50 Dissolved Hydrocarbons (ppm by volume)

Figure 11

Methane Pentane response in water-base mud

140

120

100 Membrane response

C1 ppm

80

C2 ppm C3 ppm iC4 ppm nC4 ppm iC5 ppm nC5 ppm

60

40

20

0 0 200 400 600 800 1000 1200 Dissolved Hydrocarbons (ppm by volume)

Figure 12

Methane Pentane response in synthetic oil-base mud