124 Analytical Methods in Fine Particle Technology 3.5.4.3 Calibration ‘The integrated response of the TCD is calibrated to a reference quantity of adsorptive gas. The TCD output signal may be displayed, captured by a strip chart recorder, or monitored by computer software after conversion of the analog signal to digital form. The output ean be calibrated to gas quantity adsorbed or desorbed, or can be calibrated to read directly in surface area units. Calibration to gas quantity is straightforward; the gain of the output device simply is adjusted until the value shown on the output display (or other device) matches that of the known calibration quantity To calibrate for direct reading of surface area requires converting the injected calibration volume V; of adsorptive to the surface area occupied by that quantity of gas when adsorbed onto a solid surface. First, the injected volume is converted from ambient conditions to standard conditions (273.2 K at 760 tort) using the equation _(732) Pane : ve 2 (ax yy (3-80) Next, using Avogadro's number (Na), the number of moles of gas is calculated from the standard volume. The result is multiplied by the area of solid surface covered by a single molecule of gas (16.2 x 107° m? or 16.2 A for nitrogen). The total surface area coverage for the calibration injection then is (16.2 ANa(Vsrn) x 100 m2/A? 81) The output gain is adjusted to display this value, 3.5.4.4 Making a test Perhaps the most appropriate (and, being a manval instrument, the most con- venient) application of the dynamic flow technique is for quick, single point BET surface area determinations by aitrogen adsorption. A mixture of 30% by molar volume nitrogen gas and 70% by molar volume helium (the inert component) and a eryogenie liquid nitrogen bath at atmospheric pressure es- tablishes a relative pressure (P/Pq) for the adsorptive of approximately 0.30 which is near optimum for a BET analysis. However, neither the instrument system nor the analytical method is limited to the use of nitrogen as the Surface Area and Pore Structure by Gas Adsorption 125 adsorptive, For example, atmospheric pressure and ice water temperature ¢s- tablish appropriate conditions for an n-butane and helium mixture. Other gases 2 other conditions are usable also First the sample is conditioned by being heated and flushed five of con- taminating gas and vapor that may have been picked up from exposure to the atmosphere. Then itis cooled toLNz temperature, While maintained at cryogenic temperature, the 30% nitrogen-70% helium gas is flowed over the sample until the TCD signal returns to baseline, This indicates that adsorption on the sample has equilibrated at the current pressure and no more niteogen is being adsorbed from the mixture, Then the liquid nitrogen bath is removed and the sample is warmed to room temperature. This occurs rapidly and the adsorbed nitrogen is released from the sample surface producing a sharp peak in the ‘TED output signal. Integration of the peak and a comparison of the result to the calibration value reveals the surface area of the sample. For multipoint BET analyses the same processas described above is repented except with a range of nitrogen and helium mixtures, say 5, 10, 15, 20, and 30% by volume nitrogen, A multigas manifold may be used for convenient changeover from one gas mixture 0 another in the case of pre-mixed gases. Otherwise, an automatic flow-contraliing and mixing device can be used. For ‘maltipoint BET surface area determinations, the instrument is calibrated 10 provide volume data rather than to give direct indication of surface area as in the case of the single-point method, After the test the volume vs. relative pressure data are reduced using the method discussed in Sec. 2.2, Theory. ‘The above procedure also can be modified to achieve data points at higher relative pressures (higher concentration of nitrogen gas in the mixture). This allows sufficient data to be gathered to generate a desorption isotherm to which various data reduction techniques may be applied to extract pore structure information, The Gurvitsch (43) rule states that the volume of liquid condensed in the pores of a porous solid from a condensable gas near its saturation vapor pressure is equivalent to the volume of the pores. This rule is used conveniently 10 determine total pore volume by the flowing. gas technique. For this determination the instrument is calibrated to read directly the equivalent liquid volume of the injected calibration gas volume. The calibration gas volume fistis converted to standard conditions as described carlir, the ratio of the molar volume of the liquid state of the gas to the molar volume of the gaseous state providing the conversion factor. The molar volume of liquid nitrogen is 34.670 em? and of gaseous nitrogen at STP the molar volume is 22414 cm. Therefore the ratio (and conversion factor) for nitrogen is 0.015468,126 Analytical Methods in Fine Particle Technology First the sample is saturated with 100% nitrogen, then equilibrated with 95% nitrogen in helium at LNz temperature. When the sample temperature is then raised to ambient, the quantity of gas desorbed (which also is the quantity adsorbed) is measured, but the instrument, having been calibrated as described above, reports directly in total pore volume. 3.6 Recognition and avoidance of problems In tis section several ofthe more common difficulties associated with physical adsorption analyses are identified and techniques to avoid many of them are Giscussed. The order of discussion follows the sequence of checking instrument set-up, preparing the sample, performing the analysis, and evaluating the ex- perimenial data As has been stressed, the isotherm represents the fundamental analysis data. However, reported ordered pairs of quantity adsorbed vs. relative pressure data are not measured entirely by the instrument; they include other input terms. Quantity adsorbed is calculated using an expression such as eq. 3-61 which includes sample holder free space and instrument internal volume. Tem- peratures, particularly that of the hath, may be measured directly, entered by the user, or calculsted. Relative pressure is the ratio of absolute pressure to saturation pressure where absolute pressure may be a direct measurement or value calculated incorporating a thermal transpiration correction, and saturation pressure, may be measured directly, may be an interpolated value, or may be 8 user-entered value. Even if all these factors are absolutely correc, the isotherm may be distorted by nonequilibrated data ora poorly prepared sample. Therefore close inspection of the isotherm data and analysis conditions is warranted Extracting reliable information about the sample is possible only if the isotherm, is composed of high-quality dats. The steps below will help assure data reli- ability 3.6.1 Leaks ‘The integrity of the analytical system should be confirmed periodically. The instrument’s manual is the best guide to performing the tests. Detection and repair of leaks are of critical concern in any gas sorption system. Over time and for various reasons leaks may develop at any point in a system. Of course a leak can develop at a union of two paris such as a gas inlet connection, but an even more likely failure point is at a union where parts move relative to Surface Area and Pore Structure by Gas Adsorption 127 a seal. For example, sample holder seals receive considerable wear as a result of normal use and, if not routinely replaced, eventually 2 leak will develop. Pressure differences in a volumetric instrament are considerably higher than in @ flowing-gas system. In a volumetric system leaks may be to vacuum, to or from the atmosphere, or from the pressurized gas source. A clue to the nature of the leak likely will be superimposed upon the isotherm date and should be apparent om the plot of the data. An example of a commonly observed symptom is presented in Figure 3.27. vw i een a a Pio Figure 3.27, Plot of reported volume adsorbed by an empty sample tbe, The plots ‘Mustrated are symptomatic of a system leak Micromeritcs’ static volumetric instruments make provisions for leak test ing which help confirm the presence of a leak and isolate the source. Leaks in instruments ofthe flowing gas design are often detected by plugging various flow paths to determine if a gas line branch mainiains pressure or allows pressure to bleed off. The sample holder seal is a frequent culprit and should be checked first regardless of instrument type. 3.6.2 Instrument calibration Accurate calibration of the system manifold volume obviously is critical 10 determining the quantity of gas dosed and subsequently to the quantity adsorbed by the sample. Equally important is the manifold temperature since this meas- urement, 100, is critical in accurate determination of quantity adsorbed. The other critical measuring devices are the pressure transducers which also require128 Analytical Methods in Fine Particle Technology periodic attention, primarily to the zero offset; this is an automatic process in some of Micromeritics’ instruments. A simple test of system function is to analyze an empty sample tube which should produce a value of essentially zero quantity adsorbed. A more complete test includes measuring a reference material; one week intervals between reference analyses is recommended for best results 3.6.3 Connecting an adsorptive gas supply to the instrument Whatever the gas or its source, it should be regulated to the pressure prescribed in the instrument manual. The regulator should be of a grade suitable for gas chromatography (GC) or mass spectrometry (MS) applications. It must not be ‘an “industrial-grade” regulator or one which employs internal lubricants which can migrate and contaminate downstream systems. Metal tubing should be used and the compatibility of the metal with the adsorptive ges should be considered, ‘On. slatic volumetric instrument itis important that all previous adsorptive gases be removed from the system as well as any atmospheric contaminants which aze likely to reside in the line or regulator of 2 new gas supply. The general procedure for scrupulous cleaning is first to connect any new gas line to the ixstrument, but without applying pressure, Assuming that a new gas line is being connected, atmospheric gases will be in the line and in the pressure regulator t© which the line is connected. ‘To clean all gas conduits of the system, the source tank valve should be turned off and the regulator valve ‘opened. The entire system should then be evacuated to ensure that all parts of the manifold as well as the gas inlet line and regulator of the new gas source are reached. Evacustion should be continued to 2 pressure of at least 0.010 torr. Next gas from the new source should be introduced to purge the system, This evacuate-and-purge cycle should be repeated a few times. Then the system should be evacuated for about 30 minutes. As a minimum the regulator and gas line should be purged prior to connecting to the instrument and a thorough purging of the manifotd and all lines in the instrument system should be performed. Removal of a previous gas from a flowing-gas analyzer is accomplished by purging alone since there is no vacuum system. Any flow path which can contain contaminants or previously used gas residues should be purged. A thermal conductivity detector requites calibration after changing gases. Surface Area and Pore Structure by Gas Adsorption 129 3.6.4 Sample preparation A sample not scrupulously cleaned of adsorbed contaminants will outgas during ‘an analysis. This gas becomes a rogue source of pressure within the system, Any such contribution by an outgassing sample will be interpreted by the sysiem logic as unadsorbed (or desorbed) analysis gas. Quite clearly, this can offset or can even overwhelm the true quantity adsorbed, producing erroneous data. Figure 3.28 illustrates the effect of an outgassing sample on an adsorption isotherm. ve Syrian a ask Figure 3.28. Syinptoms of an outgassing sample: suppression of reported quantity ad- sorbed and desorption curve falling below adsorption curve ‘The purpose of degassing is to remove previously adsorbed molecules from surfaces and from pores. By removing these the way is cleared for ad- sorptive molecules to imteract direcily with the sample and its pores. The adsorbate can then reveal its true properties, including the proper energy of adsorption. Practical determinations of volumes adsorbed and surface areas are seldom accurate or repeatable to better than 0.1%, Therefore if the surface is cleaned of previously adsorbed molecules so that less than 0.1% of a monolayer remains, the sample has been outgassed adequately for practical purposes Some Micromeritics instruments include built-in sample degassing capa- bility and others do not. For those that do not as well as for increasing the capacity of those instruments that do, Micromeritics supplies compact degassing units of two distinct types. One degasses samples by purging with an inert ‘gas while eating and the other by applying evacuation with heating. In general130 Analytical Methods in Fine Particle Technology cleaning samples by flowing an inert gas (purging) over them with heating is simpler, faster, and more efficient that applying a vacuum, but some users prefer evacuation, particularly with toxic or hazardous materials. For the latter materials, Micromeritics offers special valve units that ensure complet isolation both before and after degassing (see Section 3.6.6). Both degassing units have six stations and each can heat samples 0 400 °C. The temperature can be set and controlled to #5 °C so that any heating schedule up to 400 °C can be followed. ‘The identity of interfering or contaminating molecules must be considered when setting degassing criteria. Nitrogen, oxygen, and other “permanent gases” are removed easily from all but a few special materials by moderate vacuum applied at or above room temperature. Carbon dioxide adheres somewhat more tenaciously. Samples also may have picked up hydrocarbons, ammonia, alcohols, ‘and a wide variety of other vapors because of atmospheric exposure. Some of these, particularly ammonia, may chemisorb on the sample with such high energy that dislodging requires unusually high temperatures. Serious alteration of the semple may result from the attempt to clean it. Heating a sample too strongly can result in various types of damage, one being sintering. Sintering causes a deformation of the pore structure. The net result is a reduction in surface area and in pore volume. Sintering is avoided by keeping the deges temperature below a critical range, the critical range depending on the sample material. However, the most common molecule to be removed, and toward which most degassing strategies are directed, is water. Water uptake from the atmosphere is rapid for most moderate to high surface area materials so that only relatively short exposure to air leads to the need for degassing. Special caution should be exercised in removing water by heating certain microporous samples as some may be particalarly sensitive toa form of damage called steaming (hydrothermal alteration). Steaming occurs ‘when internal vapor pressure causes damage to the structure of the sample. Specifically in the case of zeolites a vaporizing and recondensing of water ‘within the sample causes recrystallization, The effect will be a reduction in surface area with little effect on total pore volume and a tendency to increase pore width. To avoid steaming, slow removal of the majority of the water shoul be done at less than 100 °C and then with stepwise increases in tem. perature CChemisorbed water, surface hydroxyl groups, or bulk chemically combined water often are desired constituents of the sample, without which the sample is no longer representative of the product. Only informed judgment can be used in the degassing of such samples. Surface Area and Pore Structure by Gas Adsorption 131 The vapor pressure of a liquid rises exponentially with temperature and the same is truc for an adsorbed film of the liquid. Vacuum cannot remove adsorbed gases until the bonds holding them to the surface are broken, Thus, applying heat is far more effective than decreasing the pressure, and for every ‘vapor there is a temperature which will clean the sample leaving only negligible amounts of contaminant, Flowing inert gas over a sample is a very efficient means for removing contaminants from the vicinity. ‘Many hydrocarbon vapors have such low saturation pressures that high heat must be used to remove them. In this situation, cracking of the molecules may occur and carbon may be left behind on the sample. With continuous pumping the indicated system pressure may be much lower than the actual pressure just at the sample, What the actual pressure is at the sample can be deduced by isolating the sample from the system vacuum and allowing the indicated pressure (o rise until relatively steady. An isolated sample should exhibit a pressure rise of less than 10° torr per minute if the sample is adequately degassed. 3.6.5 Helium entrapment Some microporous materials, particularly carbons with micropores below 8 A diameter (or width) and zeolites, adsorb a considerable quantity of helium, both during the course of sample preparation if helium is used to backfill the sample holder and during the measurement of free space. This helium is sub. sequently released during adsorption of nitrogen or argon at cryogenic tem: perature. The result is a higher equilibrium pressure in the ultza low pressure ange than would be the case if the sample had mot been exposed to helium. Figure 3.29 illustrates the effect of retained helium on an adsorption isotherm, ‘Helium entrapment is unlikely to be noticed on a linear plot ofthe isotherm but, if present, is readily apparent on a logarithmic plot. When the pressure range in which the data are being collected is extremely Iow, the pressure of the outgassing helium is most significant and may even dominate. At higher pressure the increasing proportion of pure adsorptive in the sample tube reduces the net effect of the helium and the isotherm resumes its normal shape. ‘The remedy is never to expose a sample to helium after final degassing. After exposure of the sample to helium during free-space measurement, 2 degas procedure with heat should be performed since evacuation alone ai room temperature has little effect over the short term. ‘When using nitrogen or argon as the analysis gas, one of these gases can be used as the backfill gas when moving the sample from a prep station to182 Analytical Methods in Fine Particle Technology the analysis port. A judicious choice of backfill gas will have to be made considering the affinity of the sample for the gas at room temperature if using any other gas, It usually is advisable to allow extra time for outgassing on the sample port to ensure that all of the backfill gas has been removed. Whether cor not heat is required depends on the affinity of the sample for the backfill gas. Ye se Be Einar we 10? 407 10" 109 PP Figure 3.29. The effect of helium release from a microporous carbon sample. Avoiding exposure of the sample to helium after outgassing and prior to analysis, means that the free-space values must be manually entered rather than being measured. This raises the problem of how to determine an accurate free-space value. The preferred method is to pre-calibrate a sample tube as previously explained in Sec. 3.5.1.2.1, It also is important to note the position of the Isothermal Jacket on the calibrated sample holder and to be certain thet it ig im the same position during subsequent analyses. Select the “entered free-space option when making the test. The entered warm free space should bbe the empty sample tube volume less the absolute, or skeletal, volume of the sample, The entered cold free space should be the empty sample tube volume less the sample skeletal volume converted to standard units. This is explained in more detail in previous sections of this chapter. 3.6.6 Use of SealFrits™ and PrepSeals™ SealFrits are stopper-like devices that fit into the sample holder neck. They contain a valve that opens automatically when a semple bolder is installed on either a degas or an analysis port. ‘The valve closes automatically when the Surlace Area and Pore Structure by Gas Adsorption 133 holder is removed. A special fixture in both degas and analysis ports opens the valve and allows gas to enter or leave each sample holder, The benefit of using such a device is that after degassing a sample and backfilling, the sample holder can be removed from the degas port without exposing the sample 0 almospheric contaminants, Therefore the use of SealFrits is recommended with instruments that have built-in degassing capability. PrepSeals, on the other hand, are designed for use with Micromeritics’ auxiliary degassing units. They are inline, zero dead volume, mechanical valves that also prevent sample ex- posure to atmospheric contamination, 3.6.7 Selecting the appropriate adsorptive ‘There are choices to be made concerning which adsorptive gas to use for a particular application, Much depends on the objective of the analysis, ic. whether to characterize the surface or to study adsorption phenomenon, The adsorptive is just a tool to probe the surface of the sample and, as with any choice of toois, some are better suited than others to achieve the end purpose. ‘There arc guidelines for using or not using particular adsorptives. The bases of these guidelines may involve the gas-solid interaction, the instrument system, or the data redvetion method 3.6.7.1 Krypton and argon vs. nitrogen for low surface area determinations ‘The measured quantity of adsorbed gas has an uncertainty (an error band) which Micromerities atiempts to minimize through system design. In the case of surface area determinations, there are samples such as organic materials, povidered metals, and certain geological samples which have surface areas well below 1 m°/g and which adsorb a very small quantity of gas. Depending on the instrument, there is a lower limit of surface area at which the size of the relative error becomes unacceptable. “The objective is 10 increase the difference hetween the quantity of gas dosed and the quantity of gas remaining after equilibrium when working with samples having very low surface area. An obvious approach is to maximize the gas uptake by increasing the quantity of sample. However, this hes an upper limit usually imposed by the sample holder and, at most, may only produce an increase in uptake by about one order of magnitude.134 Analytical Methods in Fine Particle Technology ‘A betier approach is to change analysis gas. Krypton has a saturation pressure approximately 1/300 that of nitrogen at liquid nitrogen temperature (krypton: 2.5 torr; nitrogen: 760 torr). Therefore, compared to nitrogen, there is in the free space above the sample about 1/300 the number of krypton ‘molecules present at the same relative pressure. Since about the same number of krypton and nitrogen molecules are required to form a monolayer, this number represents 2 far greater proportion of the quantity dosed than in the case of ritrogen, ‘Argon at liquid nitragen temperature (P9= 195 tort) also may be employed for low surface area materials. However, the benefit gained (about four-fold) vill not be as great as achieved by using krypton, ‘The primary source of uncertainty in the quantity of gas adsorbed in static volumetric analyzers of the types depicted in Figures 3.20 and 3.23 is in the prerequisite determination of free space. A primary means by which these types of instruments minimize free-space error is by controlling the levet of EN} with a device such as the Isothermal Jacket discussed in Sec. 3.5.1.4. ‘The Gemini design significantly reduces the band of uncertainty in the ‘measurement of quantity adsorbed, and as a result permits very low surface acca determinations with nitrogen. As already explained, this insirumentemploys a method of common mode rejection which significantly reduces free space error, This enables the Gemini design to extend the minimum surface area specification by about two orders of magnitude 3.6.7.2. Nitrogen vs. argon for micropore analyses Argon is inert, and its molecule is spherical, monatomic, and nonpolar. Nitrogen is diatomic, nonspherical, and has a quadrupole moment which may lead to localized adsorption, particularly on adsorbents with polar sites. Argon and nitrogen molecules are similar in size and, other than for nitrogen adsorption fn polar sites, adsorb with similar heats of adsorption. The boiling point of argon at standard pressure is 87.29 K and of nitrogen is 77.35 K. ‘Argon al liquid argon temperature is preferable to nitrogen or argon at LN temperature as @ probe molecule in studies of microporosity. Being inert is an advantageous attribute in regard to the solid-surface interaction because there are fewer specific interactions, Being at a higher temperature provides the experimental advantage of shorter equilibrium times. In addition, the argon pressure at which a given pore width fills is higher than with nitrogen and therefore is easier 0 measure accurately. Surface Area and Pore Structure by Gas Adsorption 135 3.6.8 Cryogen and adsorptive pairing A convenient combination of adsorptive gas and eryogen, although not nec- essarly required, is to pair the adsorptive with its liquid phase. This mey not always be possible or practical. In some areas of the worl, liquid argon is not available and analyses with argon as the adsorptive sre performed at LN temperature. Since the bath temperature of LNa (~77 K) is less than that of LAr (-87 K), argon gas will condense to a solid at approximately 195 torr, ‘Argon adsorption phenomena are not well understood at 77 K since this is below the tiple point of argon Argon and krypton ate used most often as analysis gases in liquid nitrogen baths. A common limitation is imposed by these pairings. All three will solidify at TT K al pressures Jess than one atmosphere. BJH pore size distributions are not meaningful since there are no surface tension values for these gases a 77 K, nor are there commonly accepted thickness equations and contact angles which are required in some data reduction methods 3.6.9 Sample temperature ‘The sample temperature ideally should be known exactly and should remain constant (isothermal) for the duration of the analysis. The bath temperature is obtainable from handbooks if the purity of the bath liquid is sufficiently high and the barometric pressure is known. Otherwise the temperature must be measured ‘The liquid phase of some adsorptives, perticularly liquid nitrogen, although ppure when drawn from the storage vessel, may dissolve atmospheric gases, primarily oxygen, over time while in an open dewar. This increases slightly the temperature of the bath. There is a flow outward of vaporized coolant from within a dewar which essentially eliminates migration of atmospheric contaminants into the dewar with a reasonably tight-fitting cover on the dewar during analysis. 3.6.10 Sample quantity ‘A sample may be of small quantity or have low surface area. Too little total surface area means that the uptake of adsorbed gas by the sample at each step produces a very small pressure difference signal in volumetric systems or, in the case of a flowing-gas analyzer, a very small TCD signal to be integrated, Not taking advantage of the dynamic range of the measuring device leeds to186 Analytical Methods in Fine Particle Technology low resolution and unreliable measurements. A sample is obviously of insuf- ficient quantity if its weight does not permit accurate measurement on an analytical balance. ‘Another consideration is the buoyancy effect which becomes significant ‘when using small quantities of sample and a ges other than the adsorptive 5 (particularly helium) for backfill. Sample weight typically is determined by subirecting the weight of the empty sample holder from the combined weight of the sample holder and degassed sample. The sample holder without sample may be weighed with air inside, but the sample tube is sometimes backfilled with another gas after the degassing process. The density difference between the other gas and nitrogen may be significant in cases when small samples are involved. Nitogen, when used for backfilling, contributes an in- significant difference from weighing in air. ‘Too much sample surface or mass can increase significantly both degas and analysis time, Although itis logical that the analysis time increase is due to increasing dosing requirements, this is not the only cause of lengthened analyses, Additional time also is consumed due to having to dissipate more hheat from a large sample quantity. The heat comes from the gas being dosed ‘al ambjeat temperature to 2 sample at cryogenic temperature and from the liberated heat of adsorption. Heat transport under ultra low pressure conditions is primarily by conduction through the sample material. Only afier the sample achioves thermal equilibrium will pressure equilibrium be attained. 3.6.11 Free-space values Free-space errors propagate through all calculations of quantity adsorbed since warm and cold free-space volumes are fundamental parameters in determining the quantity of gas adsorbed by a sample at each and every point. The relationship of free-space Factors to quantity adsorbed is developed in See. 3.5.1. Figure 3.30 shows an example of an isotherm altered by a free-space error, The dizoction of the deviation at the upper end of the isotherm depends on the sign of the free-space error. Filler rods, which as the name implies, are simply inert glass or polymer rods that slip into a sample holder and fill some of the otherwise empty volume, reduce the variability of the warm and cold free-space determinations. When the free space is reduced, so is the potential error, Surface Area and Pore Structure by Gas Adsorption 137 Berar synao o Log PP Figure 330. Effecs of free-space error on isotherm, 3.6.12 Thermal transpiration correction ‘Thermal transpiration correction, the theory of which is explained in Sec 3.5.15, may be applied only to a system which has an unobstructed path across the diameter of the sample holder. For this reason use of a filler rod precludes the use of thermal transpiration correction. Otherwise, the thermal ‘wanspiration correction should be applied anytime measurements involving equilibration pressure below I torr are expected to be obtained. 3.6.13 Equilibrium time Not allowing suificient time for equilibration results in the quantity adsorbed or desorbed at a given pressure being less than would have been the case had equilibrium been achieved. Since a sequence of data points is collected over a range of pressures, there is a systematic effect of incomplete equilibrium. On the adsorption branch of the isotherm, for example, some of the gas which id not adsorb at the previous nonequilibrated pressure will adsorb at the next higher pressure (higher energy) and will be attributed to adsorption at the higher pressure. This tends to shift all features of the adsorption isotherm to higher pressures. On the desorption branch, the same effect shifts features to lower pressures, opening wider the hysteresis Joop or producing hysteresis where none may have existed hat equilibrium been established. Examples are shown in Figure 3.31 for nitrogen on silica-alumina at 77 K.138. Analytical Methods in Fine Particle Technology ve Nye / | ‘Squats chore PP Figure 3.31. The offect of nonequilibrated daia on the isotherm plot Unless the adsorption behavior of the sample material is well known, it is difficul if not impossible to estimate the optimum equilibration time without veloping at least a partial igotherm using an estimated time. A very general ‘guideline for mesoporous samples is to begin with about a 100-second equili- bration time (which means an equilibrium imterval of about 10 seconds based on Mieromerities’ equilibrium detection method as described in Sec.3.5.1.6.2.5) For microporous samples, about a 500-sevond equilibration time (about .45-second equilibrium interval) would be a reasonable estimate when using argon as the adsomptive: more time is required for nitrogen. In the ultra low. pressure region, however, pressure equilibrium may require extended periods, ‘A minimum equilibration time of 20 to 30 minutes is appropriate. The isotherm should be allowed to develop to the end of monolayer for- mation (the knee). Then other isotherms with increased equilibration times should be gencrated to determine the appropriateness of equilibration times. If the data shift, then further adjustment is required. Ifthe data do not shift, and if minimizing the duration of the analysis is important, then it may be beneticial to shorten the equilibrium time to determine the minimum which oes not significantly affect the data ‘Unfortunately, these steps are necessary to Team the equilibration charac- teristics of an unknown sample to arrive at optimum analysis conditions. There aze several factors, such as pore size, tormosity, or intercomnectivity of the sample pores which affect equilibrium for a specific sample. These generally are unknown prior to probing the sample material with the adsorptive molecule, Surlace Area and Pore Structure by Gas Adsorption 139 3.6.14 Measured vs. entered saturation pressure Micromeritics’ volumetric instruments allow the choice of emtering the setu- ration pressure (Pp) value manually or having the instrament measure it when such measurements are within the pressure range of the system. Entered values may not be as accurate as measured ones but, for BET surface area analyses, significant inaccuracy occurs only when the Po error is large, However, the accuracy of Po is critical when working at pressures near saturation such as with total pore volume determinations. The pressure on the sample must be very close to saturation, say 0.995 P/Po, in order to fill pores near the upper end of the pore size range. In this case entering an underestimated value of Po results in a larger pore size range being reported than actually exists. Entering a value of Pp larger than the tue Po results in gas condensing in the sample tube accompanied by the erroneous reporting of an excess of ‘gas adsorbed. Therefore Po should be measured by the instrument when data are to be collected near the saturation pressure. Periodic measurement of Po during an analysis is recommended for the most accurate measurement isotherms associated with materials having long equilibration times since atmospheric pressure, purity of bath cryogen and, therefore, Po is subject to change over time. 3.7 Evaluation of results Analysis reports provide a record of test conditions, experimental data, and information extracted from the experimental data by application of various reduction methods. The sections which follow discuss the elements of various reports presented by Micromeritics’ static volumetric physical adsorption ana- lyzers and suggest ways by which the merit of the reported information may be evaluated, Regardless of the precautions exercised before the analysis, problems still ‘may occur during the analysis, or as a result of using inappropriate parameters or even inappropriate methods. The isotherm data should be inspected for evidence of experimental error. The traditional method of confirming the quality of the experiment is to repeat the analysis. Toward that end Micromeritics’ analyzers log and report the exact conditions of each analysis,140 Analytical Methods in Fine Particle Technology 3.7.1 Inspection of the isotherm plot ‘The first evaluation of data should be a visual inspection of the isotherm plot. ‘The plot should be composed of data which have not been subjected to mathe matical smoothing (the report will indicate if the data are or are not smoothed). Ifthe data describe a Type I isotherm, then the plot is best shown on a logarithmic pressure axis so that details of the low pressure region are revealed. Date in this region are important particularly for micropore studies, Examine the plot o determine if any points are outliers or if a region of the isotherm exhibits characteristics (spikes, steps, ete.) which are inconsistent with the physical process being monitored. The philosophical question of whother or not these suspected extraneous data points should be removed from the raw data is not ‘considered here, but it may be appropriate to exclude an outlier from reduced data, Too many outliers can cause the integrity of the total data set to come under suspicion, Examine specific reported values to confirm that the isotherm data were collected under reasonable conditions and using reasonable parameters. For ‘example, confirm that the warm and cold free-space values reported are typical for the sample holder and bath in use. A problem with either warm or cold free space values may indicate a free-space measurement error and affect all calculations of quantity adsorbed. Other parameters influencing the accuracy of the isotherm and requiring examination are the nonideality correction factor, which appears in all calcu: Jations-of quantity adsorbed, and the thermal transpiration molecular diameter and sample holder inside diameters which appear in the calculations when @ thermal transpiration correction is applied. If a thermal transpiration correction ‘was applied, then it should be confirmed that no filler rod was used. 3.7.2 Evaluating the isotherm tabular data set Another place te look for reasonableness of the data is the adsorptive uptake by the sample in the BET range (P/Po = 0.05 to 0.30). Total uptake is the specific quantity adsorbed (em®/g) times the sample weight (g). As an example, the level of uncertainty in this range typically is less than 0.1 cm’ STD for a high performance system. Total uptake quantities should be some multiple of this level of uncertainty. Otherwise an unfavorable signal-to-noise ratio and unreliable data result. The solution is to use greater quantity of sample 10 increase adsorptive upteke, Surface Area and Pore Structure by Gas Adsorption 141 Another valuable bit of information resides in the tebulated saturation Pressure, This pressure is expected to change somewhat over the duration of fan analysis, but it is not expected to do so with large or abrupt transitions. Unreasonable seturation pressures or unusual changes may indicate thal a gas different from the adsorptive was used in determining Po, that the level of the eryogen fell too far, or that the cryogen is impure or inappropriate. With experience obvious signs of problems can be detected by a quick inspection of the tabular and graphical data, If the date appear satisfactory, the next step is to evaluate the reduced data 3.7.3 Reduced data Isotherm data may be analyzed by any one of several reduction methods de- pending on instrument model and pressure range employed, The quality of the results depends on the quality of the isotherm, the congruity of the data reduction parameters with experimental conditions, the agreement of the theo- zetical model with the physical gas-solid system, and compliance tothe pressure range over which the method is valid. Typically, results can be appraised by examining a few salient areas of the report; the sections which follow give details 3.7.3.1 Physical parameters ‘The value of physical parameters which are used only in data reduction routines should be reviewed to assure that they agree with experimental conditions. These parameters can be changed and the experimental data recalculated if ‘an error is discovered or if exploring an alternate value is desired. Run condition ‘values used in the calculation of quantity adsorbed can be changed also. These ‘re the manually entered cold and warm free spaces, nonideality correction, factor, bath temperature, and the thermal transpiration constant, ‘The area occupied by a single adsorbed molecule is a required parameter in the calculation of surface area by the BET and Langmuir methods. The software provides a default value, but other values are found in the literature, McClellan and Hamsberger (44) provide a comprehensive review of such values. The volume of pores of a specific size range is calculated from the gas ‘quantity adsorbed in them by converting the quantity to its liquid equivalent volume. This is achieved through use of @ density conversion factor calculated from the ratio of molar densities of the condensed adsorbate at bath temperature142, Analytical Methods in Fine Particle Technology to the gaseous phase at STP, The necessary information is found in handbooks. ‘The software contains default values for common adsorptives: values for other adsorptives must be calculated. ‘The terms for liquid surface tension 7, contact angle between solid and Tiquid phase 8, molar volume of the adsorbate v, gas constant R, and sample temperature T are treated as one constant, the adsorbate property factor A expressed by ngs @-82) RF a using which, the Kelvin equaation (eq. 3-33) reduces to Px_A (3-83) Either surface tension, contact angle, or molar volume can be revised individually to give a new value for the factor A, or A can simply be altered axbitrarily for exploratory purposes, ‘The thermal transpiration correction requires two parameters which may be adjusted from those of the default values. The first is the inside stem (neck) diameter of the sample holder, and the second is the hard-sphere diameter of the adsorptive molecule. The sample holder inside diameter is available from the documentation provided with it or is measurable. Information on hard- sphere diameters of molecules may be obtained from handbooks. For terms such as the interaction parameter found in the Horvath-Kawazoe calculation (eq. 3-26), the Dubinin affinity coefficient (eq. 3-21) or Astakhov ‘exponent (eq. 3-23), the default value(s) as provided by the software generally are adequate. A search of the technical literature is required if the analysis involves a gas-solid system other than that covered by the default values. .7.3.2 Guidelines and evaluation of data reduction methods ‘The BET, Langmuir, t-Plot, Dubinin-Radushkevich and Dubinin-Astakhov data reduction methods require a linear regression analysis on transformed isotherm data. For the BET wransform plot, the y-axis is 1/(Val(Po!P)-1)} and is plotted ‘agains: relative pressure (P!Po) and for the Langmuir transform plot, I/LVq(Po/P)] is ploued against relative pressure. The t-Plot method plots quantity Surlace Area and Pore Structure by Gas Adsorption 143 adsorbed (Vg) against thickness (f) derived from a thickness equation, and the Dubinin transform plots quantity adsorbed against log(P/PoY". All of these data reduction methods were first proposed for specific applications. The user must make a judgment as to the applicability of the method wo a ges-solid system. If applied appropriately, all transform plots will exhibit a linear range and the regression analysis must be applied only over the linear range and within the range of application. Fitting a regression line to surface area transformation plots should yield 2 correlation coefficient of 0.999 or betier and for Lplots and Dubinin plots the correlation coefficient should be 0.99 or better. If the data reduction model does not apply to the gas-solid system under examination, then it may be that either no linear range exists within the pressure range of validity, or that solutions derived from the regression line of the Hinear range arc intuitively incorrect, that is, they have no relevance 10 the physical situation, such as @ negative C-value from a BET wansform. 1.3 BET C-value BET theory assumes uniform surface coverage with no favored adsorption sites and it also assumes that the gas is more strongly attracted to the surface than t0 other gas molecules. The typical range of BET C-values is from about 5 to well over 100. Values much less than 5 imply that the gas-to-gas affinity is competing with the gas-to-solid affinity which conflicts with the basic as- sumptions of BET theory. C-values much greater than 100 indicate very strong autraction for the surface or preferential adsorption. Provided the isotherm was determined with negligible error and the re- axession line to the BET transformation data was fit properly, then an out-of range C-value probably indicates that the gas-solid interaction for the particular sample material does notconfort tothe BET model. An inappropriate adsorption ‘model may be indicated also by the coefficient of correlation of the regression line, 0.999 being about the minimum value expected with five more ot less equally spaced points, In the case of indications of poor conformance to the BET model, the Langmuir data reduction method should be examined 3.7.3.4 Data analyses by the BJH method In general, this method visualizes the incremental decomposition of an ex- perimental isotherm, starting at the highest relative pressure or pore size. At144 Analytical Methods in Fine Particle Technology ccach step the quantity of adsorptive involved is divided between pore-emptying and film-thinning processes and is accounted for totally. This computational algorithm frequently leads to inconsistencies when carried to small mesopore sizes. If the thickness curve used is too steep, ultimately it will predict a larger increment of adsorptive for given pressure increment than is actually observed, The algorithm must stop since negative pore volume is nonphysical ‘Accumulated error results in the calculation of a too large volume of (possibly nonexistent) small pores ifthe thickness curve used underestimates film thinning, 3.8 A systematic approach to sample analyses by physical adsorption Guidelines follow for a systematic approach to physical adsorption analyses. Details of the analysis and data reduction methods are located in the sections above aad in the instrument operating manual. The objective of this section is to indicate, for a given analysis type, which categories of information need. to be explored for details, This section assumes that a properly calibrated instrument is being used and that all data points are equilibrated suitably. 3.8.1 Sample preparation Contaminants on the surface of every sample that has been exposed to the atmosphere must be removed, This is done effectively with heat and flowing, ‘gas, lypically nitrogen or helium, or by heat and vacuum application. The flowing gas or vacuum are (0 carry away the desorbed gases. Some samples require care when heat is applied due to the possibility of thermal damage. In the case of zeolites, for example, it is necessary (o hold the temperature under 100 °C for an extended period to allow all water to be removed. The sample mass should be determined after degassing since adsorbed contaminants can contribute significant mass to the sample. Of course the sample must be protected from exposure to the atmosphere after degassing, 3.8.2 Analysis temperature When nitrogen is to be the adsorptive, the bath typically is liquid nitrogen: when argon is the analysis gas, then liquid argon is preferred but liquid nitrogen way be used. Carbon dioxide analyses usually are performed in an ice slush bath, Caroful attention should he paid to the limitations imposed on the range Surface Area and Pore Structure by Gas Adsorption 145 of analysis pressure if a bath other than the liquid phase of the analysis gas is used. Be certain that any data reduction parameters associated with the analysis gas or bath temperature are adjusted properly to match analysis con ditions. 3.8.3 Quantity of sample A sample quamtity totsling at least $ to 10 square meters of surface area should be used for good precision. However, approximately one-half square meter of surface is acceptable in the case of krypton analyses. The question should be asked: is the sample of high specific surface area (contains micro- pores), medium surface area (few or no micropores, but with mesoporosity), cor low surface area (essentially nonporous, or containing extremely large pores)? Krypton or argon should be chosen if the sample is estimated to be of low specific surface area (m2/g). The exception is the Gemini method in which case nitrogen will suffice. Krypton analysis by the static volumetric technique requires thatthe instrument be equipped with an optional high-vacuum system, High surface area samples such as activated carbon and most zeolites exhibit high volumes of gas uptake per unit mass of material. Although perhaps 1 mg of sample would provide sufficient total surface for a good analysis, such a small quantity is difficult to weigh with accuracy. Excessive sample quantity, although easily weighed with accuracy, can significantly increase analysis time. As a general rule, more than a few hundred square meters of total surface is excessive 93.8.4 Surface area determination Pressures above about 0.4 P/Pp are not required for surface area analyses. For single-point BET analyses by either the dynamic or static volumetric technique, the final data point is taken typically at 0.30 relative pressure, Inert helium and active nitrogen are blended in a ratio which results in a partial pressure ‘of nitrogen of this magnitude when using s dynamic analyzer for these analyses. In the case of multipoint BET analyses, a series of mixed gases is required when using the Nowing-gas technique. The static volumetric technique requires conly that a pressure table be available to guide the instrument as it steps from pressures typically between 0.05 and 0.30 P/Pg. Such default pressure tables are built into Micromeritics’ instrument software but custom tables can be programmed if desired.146 Analytical Methods in Fine Particle Technology 3.85 Porosity ‘As with surface area, a general assessment of the porosity of the sample should bbe mad, but this ean be limited to whether or not the sample is microporous. If so, the question is then whether a micropore volume distribution by size is important, or is only the total mieropore volume of interest? Points in the ultra low-pressure region are necessary to obtain a complete rmicropore volume distribution. This requites a static volumetric analyzer quipped for ultra low-pressure operation. A limited range of micropores may be analyzed using carbon dioxide as the adsomptive without the high-vacuum ‘option. Experimental data are reduced by the Horvath-Kawazoe (20) and density fanctional theory methods (32,33,36) ‘Total micropore volume is obtained by the Dubinin and t-plot methods. A thickness curve appropriate for the material under analysis must be selected when using the tplot method. The t-plot method does not require the collection of data in the high-vacuum region. Pore volume distribution as a function of pore size in the mesopore and low macropore size range is attained using the BJH method by reducing isotherm (adsorption or desorption) dats between about 0.05 P/Po and saturation. The BIH method is not valid in the micropore range; thus a bigh-vacuum option is of no benefit. Complete pore distributions are almost exclusively achieved with state volumetric instruments since developing an isotherm with a lowing ‘488 analyzer is labor intensive. As with the tplot method, a thickness curve must be employed and the curve type must be appropriate for the sample ‘material, The DFT Plus dara eduction package slso produces complete mesopore distributions and requires no thickness models although its parameters should be selected to match the pore shape (sit or cylindrical), adsorptive, and bath temperature. ‘Surface Area and Pore Structure by Gas Adsorption 147 Symbols used in this chapter ‘Number in parentheses indicates equation in which symbol first appears. A slope of the Harkins and Jura tplot (3-11) A adsorption potential in D-R equation (3-20) A adsorbate property factor (3-82) An constant in Lennard-Jones potential for the adsorbate (3-26) Aa constant in Lennard-Jones potential for the adsorbent (3-26) B intercept of the Harkins and Jura t-plot (3-11) c BET constant (3-5) Fo characteristic energy of adsorption for a reference vapor G21) F Helmholtz free energy (3-38) Fee free space factor for the sample holder when immersed in the cold bath (3-70) Eu free space factor for sample holder when not immersed in the cold bath (3-64) Fal) hard sphere repulsive forces (3-39) Fie fiee-space factor for the portion of the sample holder at bath temperature (3-67) Fn free space factor for the gas manifold (3-61) FD difference between chemical potential of bulk gas and equilibrium potential of adsorbed layer at thickness ¢ (3-14) Fip] Helmholtz free energy expressed as functional of molecular density distribution (3-39) K a composite constant converting gas volume (STP) to liguid volume and A units to centimeters (3-19) K Avogadro's number as used in the Horvath-Kavwazoe equa: tion (3-26) K € constant in the Everett-Pow! equation (3-24) L nucleus-to-nucleus slit width in the Everett-Powl equation G24) Lb radius of a spherical pore in the Cheng and Yang spherical pore model (3-28) M molecular weight (3-8) Na the Avogadro constant (3-3) Na number of molecules per unit area of adsorbate as used in the Horvath-Kawazoe equation (3-26)148 Anaiytical Methods in Fine Particle Technology Pi, Po Op) number of atoms per unit area of adsorbent as used in the Horvath-Kawanoe equation (3-26) number of oxygen atoms on the wall surface of the eavity 8-28) rmumber of adsorbate atoms inside the cavity (3-28) absolute pressure of bulk gas above sample (3-1) aan implicit function of the adsorbate-adsorbate-edsorbent interactions in the development of the Horvath-Kawazoe equation (3-25) # clement of a set of absolute pressures (3-50) pressure of the manifold (3-61) saturation pressure of the adsorptive (3-5) standard pressure (760 torr) (3-50) critical condensation pressure (3-33) partial pressure of gas 1 and gas 2, respectively (3-54) total quantity adsorbed per unit weight at pressure p by DFT theory (3-44) universal gas constant (3-6) 4 constant in the Dubinin and Astakhov equation (3-23) surface areas obtained from the slopes of tangents 10 thick- ness curve (3-19) absolute temperature (3-6) bath temperature (3-70) <éimensionless terms in the Cheng and Yang spherical pore ‘model, where i ranges from 1 to 4 (3-28) # clement of a set of temperatures (3-50) temperature of manifold (3-53) temperature of reference volume (3-53) standard temperature (273.15 K) (3-50) expression of the adsorbent-adsorbate interactions in the development of the Horvath-Kawazoe equation (3-25) mean-field calculation of the attractive energy (3-39) pairwise interaction energy between atoms (3-37) physical volume of gas container (3-47) volume of gas adsorbed by sample (3-1) volume of calibration chamber (3-55) warm free-space volume (3-63) volume of calibration gas injected (3-80) # clement of a set of volumes (3-50) Wipe) y abe. b a dy flasb,o..) fi) fe) ul) k 1 Nay My Surface Area and Pore Structure by Gas Adsorption 149 the volume of the sample tube which is below the level of the bath liguid (3-66) volume of gas adsorbed when the entire surface is covered with a monomolecular layer (3-1) manifold volume @-53) molar volume of the gas (3-3) volume of pores Vp between any (wo tangent points on MP plot (3-19) potential acting on molecule at r (3-38) standard volume containing gas at standard pressure and standard temperature (3-50) reference volume (3-53) the volume of the sample tube which is above the level of the bath liquid (3-66) quantity adsorbed (Dubinin-Radushkevich method) (3-21) micropore capacity (Dubinin-Radushkevich method) (3-21) ‘grand potential energy (3-38) constant containing physical parameters of the system as used in the equation for thermal transpiration (3-77) set of distributed properties (3-44) an empirical constant in the Langmuir equation (3-1) iameter of the adsorbate molecule (3-26) the arithmetic mean of the diameters of the adsorbate and adsorbent atoms (3-28) distribution function of the properties a,b,c (3-44) total area of pores of size H in the sample (3-45) total area of surface of energy € in the sample (3-46) Helmholtz free energy density of @ uniform hard-sphere fuid @-40) indexing integer (3-27) distance between nuclei of two layers (slit widkh) (3-26) mass of the adsorbing sample (3-3) chemical potential (3-38) aan exponent in the D-A equation (3-23) ‘molar quantity of gas in a system (3-47) quantities of gas in manifold during manifold volume Calibration procedure (3-55 through 3-57) moles of gas contained in the manifold (3-78) molar quantity of gases 1 and 2 in a 2-gas mixture (3-53) pressure (3-441)150 Analytical Methods in Fine Particle Technology ae gP.Ab.c...) pH) ape) differential heat liberated on adsorption; from Ross and € Olivier (3-30) Kemel function describing the adsorption isotherm on unit, 7 surface of material with fixed properties a,b,c. (3-44) quantity adsorbed per unit area at pressure p in an ideal powe of size H (3-45) quantity adsorbed per unit area at pressure p in an ideal five surface of energy © (3-6) etn2 hheat of adsorption of the first layer (3-6) heat of liquefaction of the adsorptive (3-6) e2 radius of the pore (3-16) ‘general 3-space coordinate (3-38) 7 mean radius of curvature (3-33) : radius of cylindrical micropore (3-27) specific surface arca (3-3) separation distance between molecules (3-37) . statistical thickness of the adsorbed layer (3-10) ° equilibration thickness of the adsorbed layer (3-15) ptr) thickness of adsorbed layer at tangent point » oa MP plot 3 G19) Lennard-Jones 6-12 pairwise potential (3-41) 6 molar volume of the condensed adsorptive (3-33) ° molar volume of the adsorbate (3-16) distance of the adsorbate molecule from a surface atom in v the Everett-Powl equation (3-24) distance along the z-axis normal to the solid surface (3-43) two-dimensional spreading pressure (3-30) nonideality correction factor unique to a specific gas at a specific temperature (3-52) Weber's coefficient in calculation of thermal transpiration correction (3-77) calculated by gamma functions of k following the deriva- tion by Everett and Pow! (3-27) affinity coefficient (Dubinin-Radushkevich method) (3-21) ‘Weber's coefficient in calculation of thermal transpiration correction (3-77) ‘two-dimensional van der Waals constant (3-30) change in the Gibbs free energy (3-20) ‘molar change of enthalpy upon adsorption (3-30) Surface Area and Pore Structure by Gas Adsorption 181 potential energy of interaction in the Everett-Powl equation G24) depth of the intermolecular potential well from calculations of the Lennard-Jones potential (3-41) characteristic energy of the adsorptive (3-37) potential energy minimum in the Everett-Pow! equation G24) ‘minimum adsorbent-adsorbate potential energy of interac tion in the Cheng and Yang spherical pore model (3-28) minimum adsorbate-adsorbate potential energy of interac- tion in the C-Y spherical pore model (3-28) surface tension of the adsorbed film (3-16) fraction of surface that is occupied by adsorbed molecules G30) contact angle between the solid and condensed phase (3-33) density of the liquid adsorbate (3-8) molecular equilibrium density distribution (3-38) atea of surface occupied by a single adsorbed gas molecule G3) diameter of the adsorptive molecule (3-37) distance from a surface atom at zer0 interaction energy in the Evereit-Powl equation (3-24) che152 Analytical Methods in Fine Particle Technology References 1 11, 18 19. 20, 21 23 24. 25 Gregg, S.J. and Sing, K.S.W., Adsorption, Surface Area and Porosity, 2nd Ed, New York (1982), Ross, S. and Olivier, LP, On Physical Adsorption, Interscience Publishers, New York (1964), Brunauer, S., The Adsorption of Gases and Vapors. Vol. 1, Physical Ad- sorption, Princeton University Press, Princeton, NJ. (1943). Barrer, RM., McKenzie, N., and Reay, 1.8.S., J. Coll. Sei, 1, 479 (1956). Langmuir, L, J. Am. Chem. Soc., 40, 1631 (1918). Brunaver, S., Emmett, PH, and Teller, E., J. Am. Chem. Soc, 60, 309 (1938). Harris, M.R. and Sing, K.S.W., Chem. Ind. 487 (1959) Lippens, B.C, Linsen, B.G., and de Boee, JH. J. Catalysis, 3, 32 (1964) Shull, C. G.. J. Am. Chem. Soc., 70; 1405 (1948). Harkins, W.D. and Jura, G., J. Am. Chem. Soc., 66, 1366 (1944), Halsey, G.D., J. Chem. Phys., 16, 931 (1948). Brockhoff, J.C.P. and de Boer, IH, “The Surface Area in Intermediate Pores,” Proceedings of the International Symposium on Surface Area De- termination, D.H. Everett, RH. Ottwill, eds., UK. (1969). de Boet, JH, Lippens, B.C., Linsen, B.G., Brockhoff, J.C.P, van den Heuvel, A., and Osinga, Th. J., J. Coll. and Interface Sci., 21, 405 (1966). Magee, R.W., Presented Paper, Rubber Div, J. Am. Chem. Soc. Meeting, Pittsburgh, Pa, Oct. 11-14 (1994). Mikhail, R. Sh., Brunauer, S., and Bodar, E.E., J. Coll. andi Interface Sci, 26, 45 (1968). Dubinin, M.M. and Radushkevich, LLV., Proc. Acad. Sci. USSR, $8, 331 1947), Polanyi, M., Verb. Deutsch. Phyrik. Ges., 16, 1012 (1914). Marsh, H., Carbon, 25, 49 (1987), Dubinia, MM. and Astakhov, V.A., Adv: Chem. Soc., No. 102, 69 (1971). Horvath, G. and Kawazoe, K. J. Chem. Eng. Japan, 16, 470 (1983) Everett, DH. and Powl, .C., J. Chem. Soc., Faraday Trans. 1, 72, 619 976) Saito, A. and Foley, H.C., AIChEJournal, 37, 429 (1991), Cheng, LS. and Yang, RT, Chemical Engineering Science, 49, 2599 4994), ‘Thomson, W., Phil. Mag. S., 42, 448 (1871). On, C., Dechema - Monographien NR, 1589-1615, 79(B), p. 39 (1976) 2, 28. 29. 20. 31 22, 34, 35, 36. 37. 38, 39, 40. 4. 2. 3. ‘Suriace Area and Pore Structure by Gas Adsorption 183 Barrett, EP, Joyner, L.G., and Halenda, PP, J. Am. Chem, Soe. 73, 373 (950). Pierce, C., J. Phys. Chem, 87, 149 (1953). Tarazone, P, Physical Review, 31, 2672 (1985). Tarazona, P, Marconi, UM.B., and Evans, R., Molecular Physics, 60, 543 (1987). Tarazona, P, Molecular Physies, 52, 847 (1984), Seaton, N.A., Walton, JLPR.B., and Quirke, N., Carbon, 27, 853 (1989). Peterson, B.K, Walton, I.PR.B.. and Gubbins, KE., J Chem Soc, 82, 1789 (1986), Olivier, J. P and Conklin, W.B., Presented Paper, International Symposium ‘on Effects of Surface Heterogeneity in Adsorption and Catalysis on Solids, Kazimierz Dolay, Poland (1992), Olivier. J.P, Conklin, W.B, and von Szombathely, M., Presented Paper, “Studies in Surface Science and Catalysis (Characterization of Porous Solids MD.” Elsevier, Amsterdam, (1994) de Boes, ILH., The Dynamical Character of Adsorption, Oxford at the Clarendon Press (1953). Fowler, R. and Guggenheim, E.A., Staristical Thermodynamics, Cambridge at the University Press (1956). Olivier, IP, J. Porous Materials, 2, 9 (1995). Maddox, MW., Olivier, IP, and Gubbins, KE. in press Langmuir. Loebenstein, W.V., Journal of Colloid and Interface Science, 36,391 (1971). Partington, LR., An Advanced Treatise on Physical Chemistry, Vol. 1, Longman, Green, London (1949). Millet, G.A.. J. Pigs. Chem. 67, 1359 (1963). Nelson, FM. and Eggertsen, FT, Anal. Chem 30, 1387 (1958) Gurvitsch, L, J. Phys. Chem. Soc. Russ. 47, 805 (1915). McClellan, A.L. and Hamnsberger, HLE, Journal of Colloid and Interface Science, 23, $77 (1967).