Panel Says Bell Labs Scientist Faked Discoveries

A series of extraordinary advances in physics claimed by scientists at
Bell Labs relied on fraudulent data, a committee investigating the
matter reported yesterday. 

The findings, in effect, dismiss as fiction results from 17 papers
that had been promoted as major breakthroughs in physics, including
claims last fall that Bell Labs had created molecular-scale
transistors. 

The committee concluded that data in the disputed research, published
between 1998 and 2001, had been improperly manipulated, even
fabricated, confirming suspicions raised by outside scientists in
May. The committee placed the blame for the deceit on one Bell Labs
scientist, Dr. J. Hendrik Schön. 

"He committed scientific misconduct," said Dr. Malcolm R. Beasley, a
professor of applied physics at Stanford University who headed the
committee. "Nobody else did." 

Bell Labs immediately fired Dr. Schön, 32, a scientist who a year ago
had been thought to be on a fast path to a Nobel Prize. 

Dr. Schön did not return a phone call asking for comment. In written
comments in the report, he admitted he had made mistakes and
apologized, but he insisted that his findings were all based on
experimental observations. "I am convinced that they are real," he
wrote, "although I could not prove this to the investigation
committee." 

Despite the panel's finding that no other scientists were guilty of
misconduct, the scandal tarnishes surrounding participants, including
the co-authors who noticed nothing amiss, the scientific journals that
critics say moved too quickly to publish the sensational findings, and
Bell Labs' parent company, Lucent Technologies, which has been
buffeted by a collapse of the market for its telecommunications
equipment and tens of thousands of layoffs. 

The case also raises questions about the core of the scientific
process, in which scientists critique each other's work for errors but
rely on trust that the data is honest. If the panel is correct,
Dr. Schön pursued his fabrications in one of the hottest areas of
research, working with a revolving cast of co-researchers, and managed
to continue the charade for several years.  

Dr. Schön's molecular-scale transistors were seen as particularly
exciting because they worked the same way as current silicon
transistors. That suggested this technology could be used for computer
chips when the shrinking of silicon circuits hits fundamental physical
limits in about a decade. 

Scientists in the nascent field of molecular electronics ^× building
circuits out of individual molecules ^× worry that the negative
publicity could diminish the reputation and financing of the field. 

Bell Labs officials portrayed the scandal as the transgressions of one
scientist, not a wider problem at the laboratory. "This is an
individual case performed by an individual," said William T. O'Shea,
president of Bell Labs. "In this case, we had an individual who didn't
live up to the scientific requirement for integrity." 

The committee examined 24 accusations of scientific misconduct and
found Dr. Schön guilty in 16. The committee did not directly judge
whether the findings in the papers were invalid, but it is clear that
scientists in the field no longer believe them. 

The committee exonerated all 20 of Dr. Schön's collaborators of
complicity or knowledge in the fraud. But it also suggested that
perhaps Dr. Bertram Batlogg, the former director of solid state
physics research at Bell Labs who hired Dr. Schön in 1998 as a
post-doctoral researcher, should have taken a more critical look at
data Dr. Schön was producing. Dr. Batlogg, now a professor at the
Swiss Federal Institute of Technology in Zurich, was the senior author
of several of the papers. 

The responsibility of scientists for the work of their co-authors is,
the report said, "an extremely difficult issue, which the scientific
community has not considered carefully."  

Dr. Batlogg could not be reached for comment. 

The report does not address if higher-level managers should have
picked up signs of the problems earlier. But some outside scientists,
including former Bell researchers, said they thought such a scandal
would not have occurred at a Bell Labs of an earlier era, because
scientists scrutinized each other's work more closely. 

"I honestly don't believe it would have," said Dr. Robert C. Haddon, a
professor of chemistry at the University of California at Riverside,
who worked at Bell Labs until 1997. He cited an experiment of his at
Bell Labs in which soccer-ball molecules of carbon known as buckyballs
unexpectedly lit up when a current passed through them. 

"As soon as we tried to release this for publication, we had a
director and two department heads coming down and demanding to see
this experiment work," Dr. Haddon said. 

Most of Dr. Schön's disputed experiments, it turned out, were not even
performed at Bell Labs in Murray Hill, N.J., but at the University of
Konstanz in Germany, where Dr. Schön received his doctorate in
1997. With one exception, none of his collaborators ever witnessed any
of the experiments described in the papers, the report
said. Typically, organic crystals were grown by Dr. Schön's
collaborators, and he then assembled them into electronic devices. 

The committee also could not find any evidence to support the veracity
of the reports. Dr. Schön told the committee he had deleted almost all
of the original data files because his computer lacked hard disk space
to store the files. He said he had no laboratory notebooks. Dr. Schön
also could not reproduce any of the findings for the committee. 

Dr. Schön, a native of Germany, first worked as an intern at Bell Labs
in the spring of 1997. As a post-doctoral researcher, Dr. Schön
started publishing a dazzling series of papers. Based on ideas of
Dr. Batlogg, Dr. Schön assembled transistors on top of crystals of
organic materials. The transistors applied an electric field to add or
remove electrons, allowing scientists to study the materials'
electronic properties in an unusually systematic way. 

Dr. Haddon saw a talk of Dr. Schön's at the Materials Research Society
on Dec. 2, 1999. Impressed with the findings, Dr. Haddon suggested to
Dr. Schön that he redo an experiment that he had unsuccessfully tried
a few years before, trying to turn buckyballs into a superconductor by
applying an electric field. 

Six weeks later, Dr. Schön sent an e-mail message reporting
success. "The e-mails look genuine to me," said Dr. Haddon, a
co-author of the resulting paper. "It's just what I would have
written. He was probably more conservative than I would have been." In
the more than two years since then, no other scientist has been able
to reproduce the findings, nor subsequent experiments where Dr. Schön
claimed to raise the buckyball superconducting temperature to
minus-249 degrees Fahrenheit, which is surprisingly warm for a
superconductor. 

When scientists who were frustrated in trying to reproduce the work
approached Dr. Schön, he said the technique was difficult to
master. Later, he said only a few of the devices actually worked. 

Further doubts arose because it seemed impossible to do that much
work. In 2001, Dr. Schön averaged one scientific paper every eight
days. For most scientists, a few papers a year is productive.  

The molecular transistor papers led to his downfall.

In a paper in Nature last October, Dr. Schön and two colleagues said
they had constructed a transistor where the main switching component
was a layer one molecule thick. Two months later they had an article
in Science reporting they had now made a transistor where the switch
was a single molecule. 

But other scientists noticed that the two papers included an identical
graph. Dr. Schön said he had accidentally included the wrong graph in
the Science paper and submitted a correction, which was published
later. 

Then, in May, Dr. Paul L. McEuen, a professor of physics at Cornell
University, noticed more identical graphs, which supposedly
represented data from different experiments. Dr. McEuen notified Bell
Labs officials, who quickly assembled Dr. Beasley's committee to
investigate. 

The committee found more identical graphs. Other graphs appeared to be
spliced together from different sets of data, often with identical
curves appearing multiple times in the same graph. 

The data in some other graphs were too perfect. Dr. Schön admitted
that in some cases, he used curves of mathematical functions to
represent experimental data, the report said. 

The journals that published the research must now figure out what to
do with the discredited articles. At Science, for example, a paper can
be withdrawn only at the request of all of the authors. 

"Obviously, the authors are going to have to come together and do
something with this," said Dr. Donald Kennedy, editor in chief of
Science. "If for some reason that does not happen, then we will have
to make some announcement of the journal's position." 

Dr. Kennedy said the peer review system that underlies scientific
publication is not designed to catch fraud. "I don't think it's ever
been expected to detect fraud wherever fraud occurs," he said. 

It is also not clear what will happen to the honors that Dr. Schön
garnered in recent years, like the Outstanding Young Investigator
award and $3,000 from the Materials Research Society. 

At Bell Labs, existing policies about documenting research and keeping
computer data are being reinforced, said Dr. Cherry Murray, vice
president for physical sciences research at Bell Labs. But she said
the incident would not greatly change the laboratory.  

"We still will continue doing world-class research," she said. "Lucent
is committed to basic research." 

In a way, said Dr. Beasley, the head of the investigative panel, the
scandal proves that the scientific process succeeded in battling
fraud. "It got understood and exposed," he said. But he said the case
of Dr. Schön also showed the need for scientists to consider how much
responsibility they needed to take for their collaborators'
work. "Organizations that represent the profession need to examine
these issues," he said.