Three post-doctoral fellows doing an inventory of laboratory chemicals Friday morning set off an explosion in the Converse Chemical Laboratories that sent all three to the hospital and their Chemistry Department colleagues scrambling for an explanation.

The three researchers, whose names were not released, were in an attic in the Converse Laboratories, doing an annual inventory of chemicals kept for the laboratory of Lawrence Professor of Chemistry David A. Evans.

In the course of the inventory, a little before 10:30 a.m., the researchers handled a glass two-liter bottle of the caustic chemical isopropyl chloroformate, an eye and skin irritant, and then replaced it on a top shelf.

Moments later, the bottle exploded, spilling the liquid chemical into the face and eyes of one researcher and spraying some on the clothes of the other two.

All three were helped to safety showers and eye wash stations one floor below. The first Harvard police officers on the scene found the researcher hit in the face still having difficulty breathing, according to Lt. John Rooney.

All three were taken to Massachusetts General Hospital, where the two less injured researchers were treated and released within an hour. The more seriously injured man went home after a few hours of treatment and observation.

Percy H. Carter, a Chemistry graduate student working one floor below the three men when the explosion happened, said none of those injured would suffer permanent effects.

The event went out over police band radio as a "chemical explosion," and that designation sent trucks and three rescue units from the Cambridge Fire Department to the building, located on the corner of Oxford and Cambridge streets near Memorial Hall.

As a television news helicopter cir- cled that area, police, fire fighters,Chemistry Department officials and HarvardEnvironmental Health and Safety workers struggledto explain the explosion.

Isopropyl chloroformate, while irritating tothe skin, eyes and lungs, is not explosive on itsown, and so those on the scene were firsthypothesizing a neighboring bottle of explosiveether had touched off the blast.

A day later, Carter confirmed that a bottle ofisopropyl chloroformate had exploded on its own.He said the compound had likely been opened atsome earlier date and then reacted with water inthe air to produce gases including carbon dioxide.

These gases then built up inside the compound'sglass container, Carter said.

When the researchers moved the bottle in orderto include it in the inventory, he hypothesized,the pressurized gases were stirred and soon thepressure was too great to be contained.

Carter said incidents like this one are notcommon, but almost impossible to prevent.

He said the only solution could be simplybuying chemicals only in the amounts needed, sothat no bottles like this one could sit on theshelves untouched for long periods of time.

The explosion occurred in a relatively largespace, and soon after the blast the chemical haddissipated to less dangerous levels. A privatecompany arrived in the early afternoon toeliminate irritating fumes, and the building wasreopened later that day.

Two Harvard police officers exposed to fumes asthey assisted one of the injured researchers wentto Mt. Auburn Hospital for observation, and werereleased soon after. Another officer injured hisankle on the building's stairs, and was alsotreated at Mt. Auburn