Sigma Xi, The Scientific Research Society

Case Studies in Pathological Science Author(s): Denis L. Rousseau Reviewed work(s): Source: American Scientist, Vol. 80, No. 1 (January-February 1992), pp. 54-63 Published by: Sigma Xi, The Scientific Research Society Stable URL: http://www.jstor.org/stable/29774558 . Accessed: 21/10/2012 13:15 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

.

Sigma Xi, The Scientific Research Society is collaborating with JSTOR to digitize, preserve and extend access to American Scientist.

http://www.jstor.org

Case

in

Studies

Science

Pathological

How the loss ofobjectivityled tofalse conclusions in studies dilution and coldfusion ofpolywater,infinite

Denis L. Rousseau are often viewed as "com? mitted to truth, unbiased by emo? Scientists tion, open to new ideas, and profes? sionally and personally unselfish," J.Mahoney, an according toMichael American author and psychologist. Similar sentiments have given rise to a image of the archetypal widespread scientist?someone painstakingly ob? taining objective data, testing every side of a question and disregarding personal interests. Like other arche? types, however, this flawlessly compe? tentand dispassionate scientistdoes not exist. Even scientistsmay lose objectiv? ity in the pursuit of truth. John Locke, the 17th-century English empiricist, this possibility when he recognized wrote: "Error is not a fault of our knowledge, but a mistake of our judg? ment giving assent to thatwhich is not

The firstcharacteristic of pathological science is that the effectbeing studied is often at the limits of detectability or has a very low statistical significance. Thus itcan be difficult to do experiments that reliably test the effect. In some in? stances, subjective visual observations replace objective instrumentalmeasure? ments; in other cases, only sophisticated analyses can reveal a statistically signif? icant effect. If the effect is at the edge of detectability and ismeasured by visual observation, unconscious personal bias may affect the results. Because the effect is so weak or of such low statistical significance, there may be no consistent relationship be? tween themagnitude of the effect and the causative agent. Increasing the of the causative agent may not strength increase the size of the effect. This is un? true.... It is in man's to content power usually attributed to an incomplete himself with the proofs he has, if they derstanding of all of the variables that favor the opinion that suits with his in? control the effect.Once the investigator clinations or interest, and so stop from has become convinced that something new and important has been discov? further research." Errors in science created by a loss of ered, the fact that all of the parameters a exhibit simi? involved in itsdevelopment are not un? objectivity consistently der control is viewed as having little lar set of characteristics. Irving Lang muir, the late Nobel prize-winning consequence at the early stages of the chemist fromGeneral Electric, generat? "discovery." The second characteristic is a readi? ed a formal model of this syndrome ness to disregard prevailing ideas and and called itpathological science.He de? theories. Of course, if the effect thathas scribed six "symptoms" of this "dis? been discovered is not real, itmay not ease." I have condensed Langmuir's theoretical fit into the established six symptoms into two characteristics framework. Proponents of the effect and added a third,which I believe is themost important. might therefore concoct fantastic theo? for the new phe? ries to account nomenon. Some of these theories vio? Denis L. Rousseau is a Distinguished Member of the late a multitude of established physical Technical Staff at AT&T Bell Laboratories. He re? whereas others only mildly principles, ceived a Ph.D. in physical chemistry from Princeton distort fundamental ideas. When con? as a research associate in University. After two years frontedwith the dilemmas that the new thePhysics Department at theUniversity of South? theories create, theirproponents either ern is in he Bell Labs in 1969. He California, joined theBiological Computation Research Department, ignore the criticisms or offer ad hoc ex? cuses to dismiss the criticisms. By where his work is centered on biophysical studies of a new forth Bell 600 AT&T Address: Laboratories, proteins. theory, the investi? putting Mountain Ave., Murray Hill, N] 07974. gator becomes stillmore deeply com 54

American

Scientist, Volume

80

mitted to the new discovery because, with both a remarkable experimental observation and a revolutionary theory, major international prizes may be wait? ing over the horizon. To avoid these pitfalls, scientistsmust conceive and carry out a critical series of experiments. Ideally, the experiments the ef? give a definitive answer?either fect is real or it isnot. But the third iden? science is tifying trait of pathological that the investigator finds itnearly im? re? possible to do such experiments. The sults could be devastating. To avoid confronting the truth, the investigator selects experiments that do nothing, ex? cept perhaps add another significant figure to the result ormeasure a variant of the phenomenon. The investigator never finds the time to complete the that could bring critical measurement down thewhole house of cards. What happens ifsomeone else does a critical experiment that reveals a fatal flaw in the so-called discovery? The ex? periment is not accepted. Proponents of the effect claim that methodological mistakes, contamination or a missing key ingredient caused the negative re? sult.No matter how carefully the exper? iments are performed or how many at? tempts are made, there is always some excuse for rejecting a negative outcome. This description of science gone bad is not a portrait of deliberately fraudu? lent behavior. Pathological science aris? es

from

self-delusion?cases

where

sci?

entists believe they are acting in a manner but in? scientific methodical, stead have lost their objectivity. The practitioners of pathological science be? lieve that their findings simply cannot be wrong. But any idea can be wrong, any observation can be misinterpreted. There are many examples of non objective science. In contrast, deliberate? ly fraudulentwork is rare. It is self-delu? sion and the associated sloppiness that spawn

most

errors

in science.

Occasion

so impor? ally, the putative discovery is tant that itgets a great deal of attention and stimulates a large part of the scien? tificcommunity tomove in a new direc? tion.The three examples discussed here are such cases. I tell these stories not to ridicule those who turned out to be in error but rather towarn of a danger to which anyone in the scientific commu? nity could be vulnerable. Capillary Conjuring In the 1960s and early 1970s reports of a new form ofwater called polywater as? tounded the scientific community.N. N. Fedyakin of theKostrama Polytechnical Institute in the U.S.S.R. reported that water condensed in a capillary tube is different from normal water. Fedyakin joined with Boris V. Derjaguin of the In? stitute of Physical Chemistry of the of Sciences to de? U.S.S.R. Academy scribe the characteristics of thisnew wa? ter. They found that polywater froze into a glass-like material at -50 degrees Celsius and boiled at about 300 degrees Celsius. Polywater was more dense and viscous than normal water. Fedyakin that polywater had a new proposed and unknown structure. Polywater ismade by placing freshly drawn capillary tubes in an atmosphere that is nearly saturated with water. Through temperature control, thevapor pressure of thewater surrounding the capillary is held slightly below satura? tion to deter normal condensation of water in the tube. After a few days, a condensate forms inside the capillary tube. Normal water is removed from the condensate through evaporation, leaving only the thickpolywater. Polywater received little attention outside theU.S.S.R. until Deijaguin pre? sented his findings at international meetings in the late 1960s. His reports enticed Ellis R. Lippincott of theUniver? and Robert R. sity of Maryland National Bureau of the of Stromberg Standards to enter the polywater arena. Lippincott, Stromberg and their col? leagues applied infrared spectroscopy to thenew substance. This spectrum re? veals the geometry of a molecule and the energy of its bonds. Polywater yielded a surprising spectrum, entirely differentfrom thatof normal water. The spectroscopic results were interpreted as evidence for a polymeric structure (hence thename polywater) with water molecules arranged in a network of numer? hexagonal units. Subsequently, ous theories on the structure of polywa? ter filled the literature.

chamber generated more excitement than energy. The surprising reports Figure 1. Cold-fusion in 1989 can be interpreted as an instance of pathological cold-fusion of successful experiments the effect is nearly science can be defined by three general conditions: science. Pathological and crucial or statistically theories are disregarded; undetectable accepted irreproducible;

are and Steven Jones entered the experiments neglected. B. Stanley Pons, Martin Fleischmann that their science in the pursuit of cold fusion. They believed realm of pathological a at low a from amounts vast to extract of revealed energy simple apparatus way experiments cold fusion forced the and theoretical excitement surrounding cost. The economic potential

touted cold and news programs into the public media. While many publications as the greatest discovery since fire,many scientists remained skeptical. This is a as is copper, rather than platinum the positive electrode chamber inwhich demonstration of Utah.) device. used in the experimental courtesy of the University (Photograph debate

fusion

1992

January-February

55

a form of water, was the subject of a pathological polymeric Figure 2. Polywater, allegedly in the U.S.S.R., who in the 1960s. Polywater was discovered science episode by investigators it as a substance with the consistency of petroleum described jelly. They reported that it forms in capillary tubes, like the one shown here, and has properties different from those of normal

at -50 degrees Celsius and boils near 300 degrees more viscous and is denser than normal water. petroleum jelly, polywater 1970 by the AAAS.) and Porto 1970. Science 167:1715-1719. Copyright

water.

Polywater

freezes

And, like (From Rousseau

Celsius.

thermostatic jacket at highertemperature

Figure 3. Polywater reservoir connected

condenses

condenses days, material The remaining substance 56

American

inside

to the chamber

in the capillaries. is polywater.

Scientist, Volume

80

from a capillary tubes. Water evaporated the atmosphere with water vapor. In a few evaporates. Any normal water in the condensate

freshly drawn saturates

nearly

I began studying polywater as an as? sociate of Sergio Porto of theUniversity of Southern California. Porto reasoned that the environment holds many capil? lary-size pores; polywater could be formed naturally in these pores. This thought quickly swept us into spec? ulation. Could polywater alter biological ifpolywater processes? We wondered could extend longevity, possibly being the long-awaited "fountain of youth." The firstexperiments we planned on mea? polywater were Raman-scattering surements. This technique measures vi brational modes of themolecule, similar to an infrared spectrum. To obtain a Ra? man spectrum, a sample is irradiated by a laser, and the spectrum of the scat? tered light reveals the vibrational ener? gies. As soon as we directed our laser on polywater, it turned into a black char! This was no polymer ofwater but more likely a carbonaceous material. We quickly abandoned our grandiose im? plans for exploiting polywater's mortal qualities. ac? By that time,many scientists had even as without real, cepted polywater a thorough chemical analysis. In a pre hminary analysis of polywater, Porto and I found contamination by sodium. Years before, Derjaguin had given 25 samples of polywater toV. L. Talrose of the Institute of Physical Chemistry for mass spectroscopy. Talrose found sub? stantial organic contamination?lipids in quantities com? and phospholipids mass of the polywater. to the parable Still, Derjaguin argued that only those 25 samples were contaminated. The re? sults of the analysis appeared in an ob? scure journal. Aftermy initial discoveries with Por? to, I leftforBell Telephone Laboratories in the summer of 1969. At the time of my arrival, therewas great excitement about polywater. Some of themanagers of the laboratory invitedme to a meet? ing to discuss an interesting question. losses were increasing in Dielectric some of the transatlantic telephone ca? bles. Could itbe, theywondered, that polywater had seeped into the cables and changed their properties? William Slichter, director of chemical research, me to the analytical quickly introduced chemists and gave the analysis of my samples the highest priority. With the help of these chemists, I dis? covered many impurities in polywa? ter?the specimens were up to 60 per? cent sodium, 15 percent chlorine and 15 percent sulfate, none ofwhich appears

in normal water. The proponents of polywater accepted that my samples were contaminated. They claimed that their samples, however, were not. Itbe? came clear that a high concentration of contaminants would not be the silver stake in the heart of polywater. The proponents argued that the unique infrared spectrum of polywater proved that itwas a novel form ofwa? ter.This spectrum allegedly originated from vibrational modes involving oxy? gen and hydrogen atoms. Accordingly, polywater prepared from heavy water is replaced (D20)?in which hydrogen by deuterium (a heavier isotope of hy? drogen)?should produce different vi? brations because of the differentmass. I prepared polywater from heavy water and found that the infrared spectrum was identical to the spectrum of poly? water prepared fromnormal water. Determined to understand polywa ter's infrared spectrum, I turned tomy athletic passion, handball. After a lively game, I returned to the laboratorywith my sweaty T-shirt and wrung the per? spiration into a flask.When I placed the sweat in an infrared spectrometer, the spectrum looked strikingly similar to that of polywater. The implication was obvious: that the contamination of poly? water resulted from the condensation of bio-organic matter on the surface of the freshlydrawn capillary tubes.With the publication of this discovery, nearly all research on polywater stopped. Even without the evidence of chemi? cal contamination, the proponents of a polywater might have paused over more fundamental, thermodynamic, was too easy to dilemma?polywater make. At about the same time Fedyakin reported his firstobservations of poly? water, Kurt Vonnegut published his novel Cat's Cradle, inwhich a new form of water called ice-nine is discovered. Ice-nine has properties remarkably similar to those attributed to polywater. saw the in? however, Vonnegut, escapable thermodynamic conclusion. At the end of the novel, all of thewater in theworld becomes ice-nine. Ironical? ly,by some accounts the idea of ice-nine was originally suggested to Vonnegut by Irving Langmuir. The polywater episode illustrates the loss of objectivity that can accompany the quest forgreat new discoveries. The quantities of polywater available were so small thatmany useful experiments could not be done. Many theorieswere put forward to describe the structure of

6.66 wavelength

25.00

10.00

(micrometers)

4. Polywater (upper graph) and normal water (lower graph) produce different infrared dictates its infrared spectrum; materials with different spectra. The geometry of a molecule an structures have different spectra. Polywater's unusual infrared spectrum suggested the strongest evidence that polywater was a new kind of unusual structure, and this provided Figure

water.

(Data

from the author.)

2.50

3.33

wavelength Figure

10.00

5.00 (micrometers)

(upper graph) and human sweat (lower graph) have similar infrared shows significant levels of lipids and analysis of polywater common bio-organic the spectrum of The similarity between substances.

5. Polywater chemical

spectra. A

phospholipids, and

such as those in sweat, reveals that the spectra of other bio-organic molecules, in the capillary tubes. is not water at all but a product of organic contamination polywater (From Rousseau 1971a.) polywater

polywater without even considering the thermodynamic difficultyof accounting for itsvery existence. Finally definitive experiments showing high levels of contamination were done but not ac? cepted, until overwhelming evidence showed that a new polymer of water had not been discovered. In reflecting on the polywater saga, I am struck by the similarities between it and themore recent reports of infinite

dilution and cold fusion. These discov? eries, too, created great excitement in the scientific community. Finite Illusions and the Inquisition The physical principle of infinitedilu? tion is simple. A biologically active so? lution is diluted so many times thatno active molecules can be present, but the solution continues toproduce a biologi? cal effect.This curious notion is the ba 1992

January-February

57

Dinding of IgE bindingof IgE binding of IgE

a technique associated with homeopathic in the scientific literature in 1988. A had a brief appearance medicine, a type of white blood cell. In a standard experiment reported that infinitely diluted allergens affect basophils, investigators an allergen is added, it binds to the IgE and causes the E (IgE) binds to receptors on the surface of a basophil. When (left), immunoglobulin to release granules that lack granules. toluidine blue (a dye) fails to stain basophils basophil Jacques through exocytosis. After degranulation, Figure

group

6. Infinite dilution,

of French

of the University claimed that an infinitely diluted allergen of Paris and his collaborators (a solution containing essentially no of the allergen) also induced degranulation and his co-workers suggested of the basophils that the water in the (center). Benveniste of the allergen which attached to the IgE on the basophil's surface. As a control experiment infinitely diluted solution carried a "template" not present in the experimental (right), the allergen that binds to IgE was replaced by an allergen to IgG (a different immunoglobulin,

Benveniste molecules

The allergen to IgG fails to induce degranulation, preparation). in the control experiment that if there were more red basophils infinitely diluted allergen to IgE had induced degranulation. 58

American

Scientist, Volume

80

and so toluidine

blue

than in the experiment

stains

reasoned the basophils red. The French workers the infinitely diluted allergen, then the

that used

belief sis of homeopathic medicine?the -' ??-????i a that symptoms can be alleviated by ^fe^^, ; : v:;/;/^ ;:-v even when it is given in medication doses. small vanishingly In 1988 Jacques Benveniste of the University of Paris and his collaborators an in? a reported biological effect from controver? diluted The solution. finitely sy began when Nature published their paper after a two-year delay and fol? lowed itwith a note that expressed reservations about the validity of the phenomenon. During the two-year de? lay, the editors of Nature had insisted that Benveniste have his experiments repeated by independent laboratories. Benveniste's experiments involved human basophils, one type of white blood cell. Basophils hold many cyto plasmic granules that contain histamine ? "~? ??i Ii ?i-:?rr and other substances that induce aller? for reactions. 1 10* 1?40 10* 1010 1030 10* High-affinity receptors gic an? immunoglobulin E (IgE)?a class of dilutionofalleipen tibodies thatmediate some allergic reac? tions?cover themembrane of a baso? as the allergen is progressively diluted. After the varies periodically Figure 7. Degranulation an to degranulate. With first two dilutions, the allergen causes about 80 percent of the basophils phil. When the complex formed by at around 20 per? oscillates of degranulation of the allergen, the percentage further dilutions allergen and an IgE molecule binds to is consistent and replica this periodicity to Benveniste and his collaborators, cent. According one of these receptors, the basophil is et al. 1988.) variation. ble. Later experiments (From Davenas induced to release granules via exocyto reported considerable a The called sis, process degranulation. dye toluidine blue stains intact baso? The level of degranulation varied in ed more proof. Finally, an elaborate se? phils red; degranulated basophils do an ries of double-blind experiments was it is the odd manner with further dilutions not absorb the dye, because initiated. These tests ensured that none A the of the of There? stained. become that percent? plot allergen. granules ver? can in of the investigators doing the cell count? be of of the fore, age basophils degranulation degranulation degree sus the logarithm of the dilution was monitored by counting the number of ing knew which cells had received IgE to Ben? and allergen and which ones had not. red basophils after adding an allergen roughly periodic. According was consistent the and IgE and comparing the resultswith Indeed, Randi instituted an absurd pro? veniste, periodicity the identifying code a control sample inwhich degranula? and reproducible. cedure?wrapping in aluminum foil,placing it in a sealed entire story failed to Benveniste's tion has not occurred. the editor of convince JohnMaddox, Benveniste and his colleagues want? envelope and finally taping it to the ceil? no the measure Nature. to After the level of ed ing. These experiments produced original pa? publishing degranula? was an when the created Maddox tion as the allergen was serially diluted. allergen per, investigative degranulation committee composed of himself, James highlydiluted. They diluted the allergen tenfold and and Randi (a professional magician) then tested it on the basophils. Then, Maddox, Randi and Stewart conclud? that unintentional bias had influ? fraud ed dilut? Stewart and Walter the diluted took (an experienced they allergen the measurements. three enced After ed it tenfold again. After performing Analysis spending investigator). that duplicate readings of the the showed this process of progressive dilution as weeks in Benveniste's laboratory, committee discovered a few interesting same samples agreed more closely than many as 120 times, the experimenters facts about the experiments. The peri? stillobserved degranulation of thebaso? statistically expected, except in the dou? was es? not consis? ble-blind and his of Benveniste experiments. Therefore, the degranulation odicity colleagues phils. committee reported that the original ex? tent and varied from sample to sample; timated that this final dilution con? of the dye-marked basophils were difficult to periments were poorly controlled and tained only 10"107molecules was thatno efforthad been made to exclude count because basophils make up only The allergen infinitelydi? allergen. 1 in 100white blood cells; experiments luted?the probability was negligible systematic error or observer bias. There that even one allergen molecule was in simply failed towork foras long as sev? was no degranulation from the infinite? eral months in some cases; and one in? lydiluted allergen. the solution applied to the basophils. was the The infinite-dilutionexperiments had With no allergen present, what caused vestigator, Elizabeth Davenas, all at work. of the characteristics of pathological and his col? best Benveniste making experiments degranulation? The effectwas weak and inde? in time Ben? science. water the committee's acted that the EHiring leagues proposed as a template for the allergen and there? veniste's laboratory, they saw some de? pendent of the causative agent, the al? to even when the allergen was lergen. Itwas extremely difficult by carried the information even in the granulation the want count But the committee diluted. absence of the allergen. and, yet, experi basophils highly 1992

January-February

59

ments relied on visual measurements. measure of degranulation can be ob? committee, Ben? plying to theMaddox When more degranulation appeared in tained. Benveniste and his veniste wrote, "Itmay be that all of us colleagues a control are wrong in good faith. This is no sample than in a sample that ignored this line of study because, received the infinitelydiluted allergen, when the was crime but science as usual and only the diluted, allergen Wghly the investigators attributed this to error future knows." But, self-delusion is not they could not find histamine! Rather and recounted the basophils?building than accepting this negative result, the science as usual. in a bias for a positive result. a new Ran investigators sought theory. Benveniste and his colleagues also di expressed the need to do definitive Cold Confusion created a bizarre new theory?a persis? A torrid controversy experiments when reporting amazing began in 1989 tent structure ofwater mimicked the al? results by saying: "Look, if I told you when two research groups inUtah dis? that I kept a goat in the back yard ofmy covered that theywere independently lergen in its absence. No physical basis was offered to support such a in and house if Florida, you happened theory. working toward the same goal?cold to have a man nearby, you ask nuclear fusion. Electrochemists B. Stan? Indeed, the authors reported thatvigor? might ous agitation of the solution was neces? him to look over my garden fence, ley Ports of theUniversity of Utah and to observe man Certain? a when he'd That Martin Heischmann of theUniversity of sary degranulation. say, keeps goat.' But what would you do if I said, 1 keep ly,any putative structure imposed on Southampton in England directed one water would be destroyed by agitation. a unicorn inmy back yard'?" group, and physicist Steven Jones of Much like theproponents of polywa Benveniste and his colleagues were Brigham Young University directed the not doing fraudulent work. They ob? ter,Benveniste and his colleagues over? other group. Cold fusion offered thepo? looked the negative evidence. During served the effects that they reported. tential foran inexpensive, inexhaustible But they so believed in the phe? and clean source of energy. Some scien? degranulation, histamine is released from the basophils. By measuring the nomenon that they could ignore or rein? tistseven called it the greatest discovery amount of histamine in the solution, a since fire.And yet the experiments and terpretany questionable findings. In re apparatus appeared so ordinary that in your they could be duplicated kitchen sink. The principle of fusion is simple. It is the joining or fusing of two light nuclei (usually with a mass number below tomake one larger nucleus; ener? eight) D D helium3 neutron is released in theprocess. One specif? gy ic example of fusion is the joining of two deuterons. A deuteron is the nucleus of deuterium and consists of one proton and one neutron. If two deuterons are combined, they can produce an isotope of helium, namely helium 3 (with two protons and one neutron), plus an extra neutron. Less energy is required to hold together the two protons and one neu? tron of helium 3 than to hold together the two nuclei of thedeuterons. The ex? D D tritium proton tra energy is released, producing the power of fusion. But a substantial prob? lem, electrostatic repulsion, eliminates fusion as a present source of energy. Deuterons have a positive electric so and charge, they repel each other. To overcome this repulsion the deuterons must be heated to about 100million de? grees Celsius, about 10,000 times hotter man the surface of the sun. At this tem? perature deuterons collide violently II D Dhelium4 gamma rays enough to overcome the electrostatic re? pulsion. Cold fusion, however, requires to form larger nuclei is the process that Pons, Fleischmann and Figure 8. Fusion of deuterons only room temperature and no other ex? in their cold-fusion is a nucleus of Jones believed they had observed experiments. A deuteron or traordinary conditions. deuterium, (D) consists of one proton (p) and one neutron (n). hydrogen 2; each deuteron Under thermonuclear-fusion condi? a pair of deuterons When two are outcomes The constituent can fuse, likely. equally particles be rearranged to form a nucleus of helium 3,with the release of a neutron tions, two other fusion reactions can re? (top). Alternatively, the particles can form a nucleus of tritium, or sult from the combination of two 3,with the release of a proton hydrogen (middle).In a thirdreactionpath (bottom)theproduct ishelium 4,with onlygamma rays deuterons. Two deuterons can combine times less likely to produce tritium (another isotope of being emitted to carry off excess energy, but this reaction is about 10 million than the other two. hydrogen, with one proton and two 60

American

Scientist, Volume

80

neutrons) and a proton. Whether two deuterons combine to form helium 3 plus a neutron or tritiumplus a proton is simply a matter of chance; both reac? tions are equally likely.Two deuterons can also fuse to form a nucleus of heli? um 4, the common isotope (with two protons and two neutrons) plus gamma rays. This result is about 10 million times less likely than the production of helium 3 or tritium. Pons and Fleischmann collaborated on the study of complex processes in electrochemical cells formany years. They thought that,under the right con? ditions, the electrical forces acting at an electrode could squeeze atoms together and cause the nuclei to fuse. In 1985 an electrochemical they constructed chamber based on their ideas. During one experiment, Pons and Fleischmann claimed thatan electrode became so hot that itmelted right through the lab bench. Cold fusion had begun! Rather than risking premature release of their results, which would be necessary to get government funding, Pons and Fleischmann claimed to have spent $100,000 in personal funds for the next level of research. Jones had also wondered about cold fusion formany years, but from a dif? ferentperspective. He wanted to sim? ulate conditions proposed to exist in? side the earth. (One model suggests its heat that the earth generates through nuclear fusion.) He primarily monitored his experiments forneutron emissions, not heat. The basic experiments were similar for the two groups. A simple electro? chemical cell was constructed. The positive electrode (anode) was plat? inum and the negative electrode (cath? ode) was palladium, although Jones tried other metals. Palladium has long been known to absorb deuterium. Pons and Fleischmann filled their cell with a salt solution of lithium deuter oxide (LiOD) in heavy water. Jones used a solution, aMother Earth soup, consisting of a mysterious mixture of salts?some

concentrations

were

list?

ed simply as "a very small amount"? also inheavy water. Both groups relied on a similar theory.When a current is passed into the cell, the heavy water splits into deuteroxyl ions (OD") that move to the anode and deuterons that are absorbed into the cathode. Some of thedeuterons would be tightlypacked into the lattice of the palladium elec? trode,where theymight fuse. Regardless of the experimental simi

laxities, the two groups obtained radi? cally different results. Pons, Fleisch? mann and Marvin Hawkins, a graduate student in theDepartment of Chemistry at theUniversity ofUtah, ran the cell for extended periods of time to load the palladium electrode with deuterium be? foredetecting any production of heat, a possible indication of fusion. In the best case, they claimed that the thermal output exceeded the energy input by four and a half times. Jones detected no heat; but, he claimed that fusion began as soon as one hour into the experiment and that itdecreased after eight hours. Jones detected neutron emissions that were a million times smaller than those estimated by Pons, Fleischmann and Hawkins. However, the neutron emis? sion that Jones reported was about 40 orders of magnitude greater than the predicted background. In September 1988, two years after building his first fusion cell, Jones? unaware of the work being done by a re? Pons and Fleischmann?received search proposal to review for the De? partment of Energy. The proposal came from Pons and Fleischmann and concerned cold fusion via an electro? chemical cell. Joneswanted more infor? mation to assess the proposal. In Jan? uary 1989 the two groups met, and there was no indication that they would collaborate. Furthermore, Jones revealed that he had hastily submitted an abstract on cold fusion to theAmer? ican Physical Society to be presented at theirmeeting inMay. Although Pons and Fleischmann wanted more data before publishing, they also wanted the prestige of being the firstto publish results on cold fusion. The two groups failed to resolve the dilemma ofwhen and how the results should be published. Finally, the two universities attempted to arbitrate be? tween the research groups. At a meet? ing on March 6, both groups agreed to mail their papers simultaneously to Nature on March 24, even agreeing to meet at the Federal Express office. Pons, Reischmann and Hawkins, how? ever,welshed on the deal. They mailed a preliminary paper to the Journal of Electroanalytical Chemistry on March 11 and held a news conference on March 23 to announce their findings. Having heard about the press conference, Jones sent his paper by telefax toNature one

cathode (palladium)

anode

(platinum)

in a liquid-filled Figure 9. Two electrodes make a cold-fusion chamber. The positive electrode is a platinum wire that wraps

jar

around the negative ismade electrode, which of palladium. The solution in the chamber is a mixture of salts in heavy water (D20). across the electrodes provides energy Voltage to the chamber. The chamber ismonitored for products of nuclear neutrons or tritium.

fusion,

such as heat,

mann and Hawkins as quickly as possi? ble. And in this race to the press, accura? cy suffered.After publishing the origi? nal article, the journal published two pages of errata?19 changes in all. One change rectified "the inadvertent omis? sion" ofMarvin Hawkins as a coauthor. The significant heat reported by Pons, Fleischmann and Hawkins sug? gested a correspondingly significant number of fusion reactions, about 100 trillion per second. Although the level of neutron emissions should have been comparable, itwas nine orders ofmag? nitude lower. If the rate of neutron emission had been 100 trillion per sec? ond, everyone in the room would have been killed. Pons, Fleischmann and Hawkins recognized this paradox and day early. proposed an interesting solution: "the The Journal of Electroanalytical Chem? bulk of the energy release is due to a istrypushed their schedule ahead to hitherto unknown nuclear process." When other laboratories tried to repli publish the article by Pons, Fleisch 1992

January-February

61

cate the production of heat, however, it was never consistently found. Then other investigators wondered if any neutrons had been emitted. Pons, Reischmann and Hawkins based their claim of neutron emission on the gam? ma-ray emission spectrum. They pro? posed that this emission spectrum arose from neutrons ?feing captured by the surrounding water bath. Careful analy? sis, however, showed that the reported spectrum was a factorof twomore nar? row than the resolution of theirdetector. Furthermore, the spectral line associat? ed with neutron emission was not at the expected energy level.Many other labo? ratories repeated these experiments and found a definitive answer: No neutrons beyond the background level could be detected from electrochemical cells. These findings discredited thework of as and Hawkins Pons, Fleischmann well as that of Jones,whose only evi? dence for fusionwas neutron emission. Still, Pons, Fleischmann and Haw? kins reported finding tritium, a poten? tial product from fusion. Other labora? to have found tories also claimed tritium at relatively high levels. This was the strongest evidence for cold fu? sion. But by June 1990 tritium was shown tobe a contaminant in the palla? dium electrode. There was no evidence for the production of fusion products. Cold fusion was doomed from the startwhen a race to be firsttook prece? dence over the desire to be right.Most measurements reporting nuclear effects from cold fusionwere barely above the background noise, and extended peri? ods of failed experiments afflicted even Pons's laboratory. The proponents of cold fusion attributed the failure to sev? eral causes: differences in thematerials, the size of the electrodes, impurities in the electrodes, and low current density. The listgoes on. Nuclear reactions, however, are very well understood. Any theory offered to Figure 10. Close packing of deuterons within is the atomic matrix of metallic palladium the hypothetical mechanism of cold fusion.

the apparatus molecules of heavy wa? the platinum and palladium to is applied electrodes voltage (top). When

Within

ter surround

of heavy water split the electrodes, molecules to the positive into deuteroxyl ions that move to the thatmove electrode and deuterons

accu? electrode (middle). Deuterons negative in the spaces between the palladium mulate If two atoms of the negative electrode. are tightly packed in the palladi? deuterons um electrode, they may overcome their elec? and fuse (bottom). trostatic repulsion

62

American

Scientist, Volume

80

--

I

LABORIOUS RESEARCH, I RELENTLESS

?Jl^^?lliw

Hart Johnny B.C. By I IHAVE HEREBY

ANDINNUMERABLE

^ STUDY, CALCULATIONS,.

ASCERTAINED

THAT,../...

-? ^

NO TWO SNOWFLAKES....?

....AREALIKE....

J

-V.

_-HFigure destroys

_ ^^^^^-^

^^^^

science. Here, the scientist more by Johnny Hart portrays extremely pathological from Johnny Hart and Creator's Syndicate, it. (Reprinted with permission Incorporated.)

11. Cartoon

account for the reported observations deal of attention simply because they were scientifically and technologically must postulate new nuclear processes that only occur in the palladium elec? very important. In each of these exam? could have trodes. Indeed, Edward Teller has pro? ples, the investigators avoided the trap of nonobjectivity by posed that theremay be an undiscov? ered neutral particle involved! doing the definitive experiments? The investigators of cold fusion also those experiments that give a decisive answer. Definitive experiments existed ignored definitive experiments. Pons, Fleischmann and Hawkins only exam? for polywater, infinite dilution and ined theirwork in cells containing D20. cold fusion; but those experiments were An obvious control replaces either not done or not accepted when D20 with which would nuclear any prevent theywere done. The ability to define, H20, reactions. Likewise, no one showed that carry out and accept definitive experi? fusion products are formed at the same ments is the responsibility of every sci? time as the heat. Michael L. Salamon of entist, a responsibility thatmust be ful? thephysics department at theUniversi? filled at all costs. ty of Utah and an army of colleagues Bibliography spent more than fiveweeks in Pons's Cohen, James S., and John D. Da vies. 1989. The laboratory attempting these measure? cold fusion family. Nature 338:705-707. ments. No fusion products were detect? E., F. Beauvais, ed when heat was produced, orwhen it Davenas, J.Amara, M. Ober A. B. Robinzon, A. Miadonna, baum, was not. Salamon later said that they did

not

see a

"peep"

or an

"iota"

of con?

ventional fusion products. Pons, how? ever, rejected this negative result and said, "Maybe they should have been searching harder for nuclear particles instead of peeps and iotas." There aremany examples of scientif? ic projects in which objectivity was lost. Self-delusion in scientific research was recognized centuries ago, is evi? dent today and will no doubt continue into the future.Polywater, infinitedilu? tion and cold fusion received a great

B. Pomeranz, P. Fortner, P. Belon, J. B. Poitevin and Jacques Ben? Sanite-Laudy, 1988. Human veniste. basophil degranula? tion triggered by very dilute antiserum Nature 330:816-818. against IgE.

Tedeschi,

1973. Boris V., and N. V. Churaev. Derjaguin, Nature Nature of "anomalous water/' 244:430-431. and Stanley Pons. 1989. Martin, induced nuclear fusion Electrochemically of deuterium. Journal of Electroanalytical Chemistry 261:301-308.

Fleischmann,

Robert

N.

1989. Pathological Hall, Physics Today 42:36-48. Jones, Steven

E., E. Paul

Palmer,

science. J. Burt Czirr,

than overlooks

negative

evidence;

he

L. Decker, Gary L. Jensenjames M. Thorne, Stuart F. Taylor and Johann Rafelski. 1989. Observation of cold nuclear fusion in condensed matter. Nature 338:737-740.

Daniel

Lippincott, Ellis R., Robert R. Stromberg, War? ren H. Grant and Gerald L. Cessac. 1969. Vibrational indicate spectra Polywater. structure. Science stable polymeric unique 164:1482-1487. and Walter W. Maddox, John, James Randi Stewart. 1988. "High-dilution" experiments a delusion. Nature 334:287-290. Denis L. 1971a. An alternative expla? for polywater. Journal of Colloid and Interface Science 36(4):434-442.

Rousseau, nation

L. 1971b. "Polywater'' and Denis Rousseau, sweat: similarities between the infrared spec? tra. Science 171:170-172. Rousseau,

Denis

L., and Sergio P. S. Porto. 1970. or artifact? Science

Polywater: polymer 167:1715-1719.

Salamon, Michael H., M. E. Wrenn, H. E. Berge C. L. son, K. C. Crawford, W H. Delaney, Y. Q. Li, J.A. Rusho, G. M. Henderson, and S.M. Seltzer. Limits on the Sandquist electrons emission of neutrons, gamma-rays, elec? and protons from Pons/Fleischmann trolytic cells. Nature 344:401-405. A. Butler, D. S. Gin Schriber, James E., Michael I. Ewing. for 1989. Search ley and Ronald tita? cold fusion in high-pressure D2-loaded nium and palladium metal and deuteride. Fusion Technologu

16:397-400.

D. E., D. J. S. Findlay, D. H. Craston, Williams, M. R. Sene, M. Bailey, S. Croft, B. W. Hooton, C. P. Jones, A. R. J.Kucernak, J.A. Mason and R. I. Taylor. Upper bounds on 'cold fu? sion' in electrolytic cells. Nature 342:375-384.

1992

January-February

63