2557
Ind. Eng. Chem. Res. 1987,26, 2557-2558
that a kinetic model (their eq A l ) of over 20 parameters is required to fit basic data secured under ill-defined mixing conditions? Furthermore, the criteria for the presence or absence of external gradients such as those cited above cannot be employed unless CSTR conditions prevail. Given the small contact times (ca. 1 s) and exothermic nature of formaldehyde oxidation, one may well expect local external (interphase) gradients. But to detect the presence or absence of external (interphase) local gradients by the above-cited criteria (eq 1and 3) requires not only CSTR conditions but a value of k, for foils/ribbons packed in a tube. What value of k, was used if tests for external diffusional disguise were undertaken by the authors? Finally, Foster and Masel (1986) state that “the fact that the data is (sic) consistent with Conner and Bennetts’ mechanism does not necessarily imply that (their) mechanism is correct for this example”. I agree, as I’m sure would Conner and Bennett. Nomenclature a = external surfacefvolume of catalyst C = concentration C, = molar heat capacity 2l = intraphase diffusivity AH = reaction enthalpy change k , = interphase mass-transfer coefficient Le = Lewis number T = temperature, K
X = conversion Yf= reactant mole fraction in feed Greek Symbols
6, /3 = external, internal adiabatic AT, eq 3 and 4
A = difference 0 = contact time, volume of catalystfvolume of feed rate
Subscripts e = external f = feed i = internal o = bulk fluid, outlet value
Registry No. CH20, 50-00-0; nickel oxide, 11099-02-8. Literature Cited Carberry, J. J. Chemical and Catalytic Reaction Engineering; McGraw-Hill: New York, 1976. Foster, J. J.; Masel, R. I. Znd. Eng. Chem. Prod. Res. Deu. 1986,25, 563. Serrano, C.; Carberry, J. J. Appl. Catal. 1985, 19, 119.
* Visiting Fellow and Professor, Cambridge University, Cambridge, U.K. James J. Carberry* Chemical Engineering Department University of Notre Dame Notre Dame, Indiana 46556
Reply to Comments on “Formaldehyde Oxidation on Nickel Oxide” Sir: Professor Carberry points out that Foster and Masel (1986) have observed rather complicated behavior and wonders whether unusual reactor behavior or heat and mass transfer effects are contributing to these observations. This comment is surprising since Foster and Masel (1986) reported that they tested their reactor experimentally and found that Carberry’s (1976) criterion for negligible influence of heat and mass transfer and Anderson’s (1968) criterion for steady-state CSTR behavior were satisfied during the experiments. Carberry does not provide any data or analysis to contradict these conclusions, and he has never seen Foster and Masel’s reactor. It is curious, therefore, that such a comment would be made. The reply will concentrate on two points. (a) Complex behavior is the expected result for formaldehyde oxidation over nickel oxide. (b) Based on the criteria in Carberry (1976) and Anderson (1968),heat and mass transfer effecta or non-CSTR behavior should not be a problem in Foster and Masel’s work. The complexity issue will be considered first. As Carberry notes, rather complex behavior was reported by Foster and Masel (1986). To put the complexity in perspective, however, note that the oxidation of CO over nickel oxide catalysts has been studied in great detail and found to also show complex behavior. Foster and Masel (1986) found that they could fit their data for formaldehyde oxidation over nickel oxide via a rate equation derived from Connor and Bennett’s mechanism for CO oxidation over the same catalyst. Foster and Masel were careful to note that the good fit did not imply that the rate equation or the mechanism it was derived from was correct for formaldehyde oxidation over nickel oxide. However, Foster and Masel did conclude that formaldehyde oxidation over nickel oxide shows the same level of complexity as CO oxidation over the same catalyst. It is disappointing 0888-5885/87/2626-2557$01.50/0
that Carberry finds such a conclusion surprising. However, it would have been much more surprising if formaldehyde (H,CO) oxidation over nickel oxide was found to be less complex than CO oxidation over the same catalyst. Carberry notes that complex behavior can also be caused by heat- and mass-transfer effects or insufficient recycle. However, the criteria of Carberry (1976) and Anderson (1968) suggest that this was not a problem in Foster and Masel’s experiments. Carberry’s comment notes that internal heat and mass transfer is not an issue with a foil catalyst. One does have to consider external heat- and mass-transfer effects. However, Carberry (1976) shows that interphase heat- and mass-transfer effects “do not effect our observed catalytic rate” when AT,/Tf < 0.01 and AC,/C, = AT,/p < 0.01 (1) where the constants are defined in Carberry’s comment. One can show that for Foster and Masel’s reactor, AT, < TmB,- Tmin,where Tmaxis the temperature of the foil at the top of the catalyst bed and T- is the wall temperature (which equals the feed temperature). Foster and Masel measured Tminand TmBIduring their experiments and found that AT, was small enough that eq 1 was satisfied. Thus, the data in Foster and Masel (1986) should be “undisguised by heat and mass transfer effects” within the catalyst bed. Carberry also raises the question whether the steadystate performance of Foster and Masel’s reactor approximates that of a CSTR. According to Anderson (1968), steady-state CSTR behavior is obtained when X / ( l + R) < 5 % (2) Foster and Masel (1986) measured the internal circulation rate of their reactor, when running with pure helium, and 0 1987 American Chemical Society
2558 Ind. Eng. Chem. Res., Vol. 26, No. 12, 1987
used that measurement to estimate a recycle ratio and t,hereby verify that eq 2 is satisfied (actually X/(1 + R ) < 1% 1. Foster and Masel also varied the recycle ratio, by varying the agitation rate, and detected no change in the performance of their reactor. Estimated recycle ratios were in the range 23-47. Thus, according to Anderson’s (1968) criterion, the steady-state behavior of Foster and Masel’s reactor should approximate that of a CSTR.’ Further, it is useful to note that the complex behavior reported by Foster and Masel occurred under conditions where the conversion was between 3% and 12%. At 3% conversion, Anderson’s (1968) criterion are satisfied even in the absence of recycle. As a result, we conclude that based on the criteria in Carberry (1976) and Anderson (1968), Carberry’s concerns are unwarranted. Finally, it is also useful to note that Carberry seems to claim that it is difficult to circulate gas past a foil with a fan. In reality, however, this is not a problem provided the foil is oriented parallel to the flow and the fan is designed appropriately. A recycle ratio of 20 corresponds to an internal flow rate of 2 SCFM. By comparison, the 3-in. fans used to cool electronics are rated at 68 SCFM. One does have to be careful with fan design; Foster and Masel’s original fan did not work properly, so it was modified (the modification was inadvertently not mentioned in the paper). However, if one is careful, it is not difficult to circulate gas past a foil with a fan. Packed beds, of course, require much more elaborate designs because of pressure drop considerations.
In summary, then, Carberry does not provide any data or analysis to justify his comments; he has never seen Foster and Masel’s reactor to know how it will perform. The performance of Foster and Masel’s reactor was tested experimentally and found to obey the criteria in Carberry (1976) and Anderson (1968). Thus, it seems that Carberry’s concerns are unjustified. Registry No. CH20, 50-00-0;nickel oxide, 11099-02-8. Literature Cited Anderson, R. B.Experimental Methods in Catalytic Research, Academic: New York, 1968; Vol. I, p 10. Carberry, J. J. Chemical and Catalytic Reaction Engineering; McGraw-Hill: New York, 1976; p 407. Foster, J. J.; Masel, R. I. Znd. Eng. Chem. Prod. Res. Deu. 1986,25, 563. t More extensive testing would have been needed if Foster and Masel had wanted to use their reactor for non-steady-state experiments. However, one of the advantages of a recycle reactor is that details of the recycle loop (even dead time) do not affect steady-state performance.
Richard I. Masel* Department of Chemical Engineering University of Illinois, Urbana Urbana, Illinois 61801 James J. Foster Westvaco Corporation Couington, Virginia 24426
