"Tried to resist using @ work, but ended up failing," the forensic chemist scribbled on a diary worksheet she kept as part of her substance abuse therapy. "Going to use phentermine," she wrote on another, "but when I went to take it, I saw how little (v. little) there is left = ended up not using."

The chemist, Sonja Farak, worked at the state drug lab in Amherst, Massachusetts, for more than eight years. Her notes record on-the-job drug use ranging from small nips of the lab's baseline standard stock of the stimulant phentermine to stealing crack not only from her own samples but from colleagues' as well. On another worksheet chronicling her struggle not to use, she described 12 of the next 13 samples assigned to her for testing as "urge-ful."

State police took these worksheets from Farak's car in January 2013, the same day they arrested her for tampering with evidence and for cocaine possession. The lead prosecutor on Farak's case knew about the diaries, as did supervisors at the state attorney general's office. And yet, despite explicit requests for this kind of evidence, state prosecutors withheld Farak's handwritten notes about her drug use, theft, and evidence tampering from defense attorneys and a judge for more than a year.

Farak's reports were central to thousands of cases, and the fact that she ran analyses while high and regularly dipped into "urge-ful" samples casts doubt on thousands of convictions. But without access to evidence showing how long Farak had been doing this, defendants with constitutional grounds for challenging their incarceration were held for months and even years longer than necessary.

Farak wasn't the first Massachusetts chemist to tamper with drug evidence. In fall 2012, just five months before her arrest, Annie Dookhan confessed to faking analyses and altering samples in the Boston testing facility where she worked.

The twin Massachusetts drug lab scandals are unprecedented in the sheer number of cases thrown out because of forensic misconduct. Between the two women, 47,000 drug convictions and guilty pleas have been dismissed in the last two years, many for misdemeanor possession. Many more are likely to follow, with the total expected to exceed 50,000.

Both scandals undercut confidence in the criminal justice system and the validity of forensic analysis. And both pose the obvious question about how chemists could behave so badly for years without detection.

This story is an effort to reconstruct what was known about Farak and Dookhan's crimes, and when, based on court filings, diaries, and interviews with the major players.

Dookhan's transgressions got more press attention: Her story broke first, she immediately confessed, and her misdeeds took place in big-city Boston rather than the western reaches of the state. But the Farak scandal is in many ways worse, since the chemist's crimes were compounded by drug abuse on the job and prosecutorial misconduct that the state's top court called "the deceptive withholding of exculpatory evidence by members of the Attorney General's office."

Thanks largely to the prosecutors' deception, the Massachusetts Supreme Judicial Court in October 2018 was forced to dismiss thousands of cases Farak may never have even touched, including every single conviction based on evidence processed at the Amherst lab from 2009 to the day of Farak's arrest in 2013. The court also dismissed all meth cases processed at the lab since Farak started in 2004. At least 11,000 cases have already been dismissed due to fallout from the scandal, with thousands more likely to come.

Because state prosecutors hid Farak's substance abuse diaries, it took far too long for the full timeline of her crimes to become public. In the aftermath, the court felt it necessary to make clear that "no prosecutor…has the authority to decline to disclose exculpatory information."

Looking back, it seems that Massachusetts law enforcement officials, reeling from the Dookhan case, simply felt they couldn't weather another full-fledged forensics scandal. Maybe fatigue made them sloppy, or perhaps they actively chose to look the other way as evidence piled up about the enormity of Farak's crimes. But why were a small handful of prosecutors allowed total control over evidence about one of the worst criminal justice failures in recent memory?

Annie Dookhan

"It was almost like Dookhan wanted to get caught," one of her former co-workers told state police in 2012.

Yet Dookhan's brazen crimes went undetected for ages. In the eight and a half years she worked at the Hinton State Laboratory in Boston, her supervisors apparently never noticed she certified samples as narcotics without actually testing them, a type of fraud called "dry-labbing." And when the tests she did run came back negative, Dookhan added controlled substances to the vials. Or she just lied about her results altogether: In one of the more ludicrous cases, she testified under oath that a chunk of cashew was crack cocaine. Even the master's degree on her résumé was fabricated.

The Hinton drug lab, operated by the Massachusetts Department of Public Health, appears to have been run largely on the honor system. It had no surveillance cameras, laughable security on evidence safes, and "laissez faire" management, which the state inspector general determined was the "most glaring factor that led to the Dookhan crisis."

In 2004, her first full year at the lab, Dookhan reported analyzing approximately 700 samples per month. On paper, these numbers made Dookhan the most productive chemist at Hinton; the next most productive averaged around 300 samples per month. One colleague called her the "super woman of the lab."

But another co-worker was suspicious, particularly since he "never saw Dookhan in front of a microscope."

Dookhan's output remained implausibly high even after the U.S. Supreme Court ruled in Melendez-Diaz v. Massachusetts (2009) that defendants were entitled to cross-examine forensic chemists about their analysis. Poetically, that landmark case originated from the Hinton lab, although Dookhan didn't conduct the analysis in question.

Martha Coakley, then attorney general for the state, argued in Melendez-Diaz that a chemist's certificate contains only "neutral, objective facts." If chemists had to testify in person, Coakley warned melodramatically, misdemeanor drug prosecutions "would essentially grind to a halt."

Prosecutors nationwide pretty uniformly backed this argument, which the Supreme Court rejected in a 5–4 opinion. "Forensic evidence is not uniquely immune from the risk of manipulation," Justice Antonin Scalia wrote for the majority. "A forensic analyst responding to a request from a law enforcement official may feel pressure—or have an incentive—to alter the evidence in a manner favorable to the prosecution."

Scalia may as well have been describing Dookhan.

In a letter filed with the Supreme Court, Julianne Nassif, a lab supervisor, wrote that Hinton had "appropriate quality control" measures. This was not true, as Nassif's department later conceded. "Dookhan's consistently high testing volumes should have been a clear indication that a more thorough analysis and review of her work was needed," an internal review found.

After the Supreme Court's decision, a skeptical colleague started tracking how many microscope slides Dookhan used to test samples for cocaine. Dookhan was now spending less time at her lab bench and more time testifying in court about her results. Finding that there did not appear to be enough slides in Dookhan's discard pile to match her numbers, the colleague brought his concerns to an outside attorney, who advised he should be careful making "accusations about a young woman's career," he later told state police.

It ultimately took a blatant violation to expose Dookhan, and even then her bosses twisted themselves in knots to hold on to their "super woman."

In June 2011, Dookhan secretly took 90 samples out of an evidence locker and then forged a co-worker's initials to check them back in, a clear chain-of-custody breach. Nassif considered it a lapse in judgment, but not a disqualifying one; Nassif's boss didn't think it necessary to alert the prosecutors whose cases relied on the samples, much less the defendants.

Nassif put Dookhan on desk duty but allowed her to finish testing cases already on her plate, including some of the samples she had taken from the locker. Her access to evidence was not restricted, and she continued testifying in court.

Democratic Gov. Deval Patrick's office didn't learn about the protocol breach until December 2011. Coakley's office finally launched a criminal investigation in July 2012, more than a year after the infraction was discovered by Dookhan's supervisors.

In her initial police interview, given at her dining room table, Dookhan said she "would never falsify" results "because it's someone's life on the line." Shown results suggesting otherwise, she copped to contaminating samples "a few times" during the previous "two to three years."

She couldn't be sure which cases these were, Dookhan told investigators. This threw every sample she had ever tested into question.

State officials rushed to condemn her loudly and publicly. "Annie Dookhan's alleged actions corrupted the integrity of the criminal justice system, and there are many victims as a result of this," Coakley said at a press conference. "First, of course, are the defendants, who when charged in the criminal justice system have the right to expect that they will be given due process and there will be fair and accurate information used in any prosecution against them." Coakley assigned the case against Dookhan to Assistant Attorney General Anne Kaczmarek and her supervisor, John Verner.

The criminal prosecution wasn't the only investigation of the Dookhan scandal. Even before her arrest, the Department of Public Health had launched an internal inquiry into how such misconduct had gone undetected for such a long time.

The governor also tapped a local attorney, David Meier, to count how many individuals' cases might be tainted. Meier put the number at 40,323 defendants, though some have called that an overestimate.

Soon after Dookhan's arrest, Coakley's office asked the governor to order a broader independent probe of the Hinton lab. The defense bar had raised concerns that prosecutors might be "perceived as having a stake" in such an investigation. "It is critical that all parties have unquestioned faith in that process from the beginning so that they will have full confidence in the conclusions drawn at the end," Coakley said.

Patrick appointed the state inspector general to look into it. His report deemed Dookhan the "sole bad actor" at the lab, a finding that remains disputed in some circles.

In November 2013, Dookhan pleaded guilty to obstruction of justice, tampering with evidence, and perjury.

"The gravity of the present case cannot be overstated," Kaczmarek wrote in her memo recommending a prison sentence of five to seven years. Dookhan had seeded public mistrust in the criminal justice system, which "now becomes an issue in every criminal trial for every defendant."

A judge sentenced Dookhan to three years in prison; she was granted parole in April 2016.

In a 6–1 ruling by the Supreme Judicial Court in 2017, the defense bar, led by public defenders and the Massachusetts branch of the American Civil Liberties Union (ACLU), won the dismissal of almost every conviction based on Dookhan's analysis—more than 36,000 cases in all. The lone dissenting justice called the decision "too little and too late" and argued that the severity of the scandal required tossing all the cases. But absent evidence of aggravating misconduct by prosecutors or cops, the majority ruled, Dookhan's tampering alone didn't justify a blanket dismissal of every case she had touched.

When the Farak scandal erupted, that misconduct came into view.

Sonja Farak

On a Friday afternoon in January 2013, a call came in to Coakley's office: "We have another Annie Dookhan out west."

Earlier that day, a chemist at the Amherst drug lab had tracked two samples that were missing from the evidence locker to Sonja Farak's bench. They were found with their packaging sliced open and their contents apparently altered. Farak was arrested the next day, and the attorney general's office assigned the case to Anne Kaczmarek.

Farak had started taking drugs on the job within months of joining the lab. She consumed meth, crack cocaine, amphetamines, and LSD at the bench where she tested samples, in a lab bathroom, and even at courthouses where she was testifying.

The Dookhan prosecution was barely underway, a grand jury having returned indictments a few weeks earlier. Kaczmarek was now juggling two scandals on opposite sides of the state. With the Dookhan case so fresh, reporters immediately labeled Farak "the second chemist."

Officials rushed to downplay the situation in Amherst. "These drugs…were tested fairly," Coakley claimed the day after Farak's arrest. Gov. Patrick said "the most important take-home" was that "no individual's due process rights were compromised."

Such strong claims were too hasty at best, since investigators had not yet finished basic searches; three days later, police executed a warrant for a duffel bag they found stuffed behind Farak's desk. It contained substances often used to make counterfeit cocaine, including soap, baking soda, candle wax, and modeling clay, plus lab dishes, wax paper, and fragments of a crack pipe. Another three days later, state police conducted a full search of Farak's workstation, finding a vial of powder that tested positive for oxycodone, plus 11.7 grams of cocaine in a desk drawer.

Exhausted from the ongoing scandal in Boston, state officials were desperate for damage control. It's not as bad as Dookhan, they asserted and implied over and over.

It took another three years for the truth to emerge. Farak had started taking drugs on the job within months of joining the Amherst lab in 2004. She consumed meth, crack cocaine, amphetamines, and LSD at the bench where she tested samples, in a lab bathroom, and even at courthouses where she was testifying.

But unlike with Dookhan, there were no independent investigations of Farak or the Amherst lab. The governor didn't appoint the inspector general or anyone else to determine how long Farak was altering samples or running analyses while high. Instead, Coakley's office served as gatekeeper to evidence that could have untangled the scandal and freed thousands of people from prison and jail years earlier, or at least wiped their improper convictions off the books.

As Kaczmarek herself later observed, Farak essentially had "a drugstore at her disposal" from her first day at the Amherst lab. But she proceeded on the hunch that Farak only became addicted in the months before her arrest, and her colleagues stonewalled people who were skeptical of that timeline. When defense lawyers asked to see evidence for themselves, state prosecutors smeared them as pursuing a "fishing expedition."

Most of the heat for this—including formal bar complaints—has fallen on Kaczmarek and another former prosecutor, Kris Foster, who was tasked with responding to subpoenas regarding the Farak evidence. Both have since left the attorney general's office for other government positions.

Meanwhile, other top prosecutors, including Coakley, largely escaped criticism for their collective failure to hand over evidence that they were bound by constitutional mandate to share with defendants. Coakley did not respond to multiple requests for comment for this story.

Thanks to Farak's testimony and those diary worksheets, we now know that, soon after joining the Amherst lab in 2004, Farak started skimming from the methamphetamine "standard," an undiluted oil used as a reference against which suspected meth samples are compared. Over the next four years, Farak consumed nearly all of it.

Like Hinton, the Amherst lab had no cameras. Each employee had a unique swipe card, but Farak simply used a physical key to get in after hours and on weekends. She had unrestricted access to the evidence room.

In 2009, Farak branched out to the lab's amphetamine, phentermine, and cocaine standards. She also starting dipping into police-submitted samples, a "whole other level of morality," as Farak called it during a fall 2015 special grand jury session.

"I remember actually sitting on the stand and looking at it," Farak said of her first time swiping from evidence in a trafficking case, "knowing that I had analyzed the sample and that I had then tampered with it."

She started seeing a substance abuse therapist around this time. During the next four years, she would periodically sober up and then relapse.

"That was one of the lines I had thought I would never cross: I wouldn't tamper with evidence, I wouldn't smoke crack, and then I wouldn't touch other people's work," Farak said.

She soon crossed all these lines. She started smoking crack cocaine in 2011 and was soon using it 10 to 12 times a day. She even made her own crack in the lab. In 2012, she began taking from co-workers' samples, forging intake forms and editing the lab database to cover her tracks.

A few months before her arrest, Farak's counselor recommended in-patient rehab.

"I was totally controlled by my addiction," Farak later testified. But she insisted the drugs didn't compromise her work—a belief that one judge would aptly declare "belies logic."

"Because on almost a daily basis Farak abused narcotics…there is no assurance that she was able to perform chemical analysis correctly," the judge found.

To multiple courts' amazement, her incessant drug use never caught the attention of her co-workers.

The crucial fact of her longstanding and frequent drug use also never made it into Farak's trial, much less to defendants appealing convictions predicated on her tainted analyses. As the state's top court put it, the criminal investigation into Farak was "cursory at best."

The Amherst lab had called state police when the two missing samples were noticed in 2013. Two detectives found Farak at a courthouse waiting to testify on an unrelated matter. They pulled her aside as she walked back to the courthouse from her car, where she had smoked "a fair amount of crack" during her lunch break.

At this point, Farak—unlike Dookhan—didn't admit anything. She stopped the interview when asked about crack pipes found at her bench, and state police towed her car back to barracks while they waited on a warrant.

Investigators either missed or declined opportunities to dig very deep. They never searched Farak's computer or her home. Even when she failed a post-arrest drug test—prompting the lead investigator to quip to Kaczmarek, "I hope she doesn't have a stash in her house!"—he didn't request a warrant. It was an astoundingly light touch for the second state chemist arrested in six months.

Still, the state was acquiring evidence. In Farak's car, police found a "works kit"—crack cocaine, a spatula, and copper mesh, often used as a pipe filter. There were also newspaper articles about other officials caught stealing drugs, including one with a scribbled note, "Thank god I'm not a law enforcement officer." Most important, they found seven worksheets from Farak's substance abuse therapy.

One was clearly dated November 16, 2011—a year and two months before her arrest. Another worksheet had the month and weekdays for December 2011, which police easily could have determined by cross-referencing holidays or looking up a New England Patriots game mentioned in one entry. (Conveniently, they also found a Patriots schedule from 2011 in the car.)

Officials recognized the worksheets for what they were: near-indisputable confessions. Sgt. Joseph Ballou, lead investigator for the state police, called them the most important documents from the car. He emailed them to Kaczmarek—subject: "FARAK Admissions." But when the relevant police reports were released to defense attorneys, there was no mention of the diary entries' existence, much less that they went back so far.

Kaczmarek quoted the worksheets in a memo to her supervisor, Verner, and others, summarizing that they revealed Farak's "struggle with substance abuse." But she worried they might be privileged as health information. Out of "an abundance of caution," Kaczmarek didn't present them to the grand jury that was convened to determine whether to indict Farak. When grand jury materials were eventually released to defense attorneys, then, they did not mention that these documents existed.

The attorney general's office—Kaczmarek or her supervisors—could have asked a judge to determine whether the worksheets were actually privileged, as Kaczmarek later acknowledged. If there's ever any uncertainty over "whether exculpatory information should be disclosed," the Supreme Judicial Court later wrote, "the prosecutor must file a motion for a protective order and must present the information for a judge to review."

The fact that she ran analyses while high and regularly dipped into samples casts doubt on thousands of convictions. Yet state prosecutors withheld Farak's handwritten notes about her drug use, theft, and evidence tampering from defense attorneys and a judge for more than a year.

Instead, Kaczmarek provided copies to Farak's own attorney and asked that all evidence from Farak's car, including the worksheets, be kept away from prying defense attorneys representing the thousands of people convicted of drug crimes based on Farak's work. Even as they filed numerous motions for information about how long Farak had been using drugs, the defense attorneys had no idea these worksheets existed. Local prosecutors also remained in the dark.

This might not have mattered as much if the investigators had followed the evidence that Farak had been using drugs for at least a year and almost certainly longer. If they'd kept digging, defendants might still have learned the crucial facts.

Instead, Kaczmarek proceeded as if the substance abuse was a recent development. Despite clear indications that Farak used a variety of narcotics—her worksheets mentioned phentermine, and that vial of powdered oxycodone-acetaminophen had been found at her bench—Kaczmarek also proceeded as if crack cocaine were Farak's sole drug.

"Please don't let this get more complicated than we thought," Kaczmarek replied when Ballou, the lead investigator, flagged irregularities in Farak's analysis in a case featuring pain pills. A local prosecutor also asked Ballou to look into a case Farak had tested as far back as 2005.

"If she were suffering from back injury—maybe she took some oxys?" Kaczmarek wrote back. She later called this dismissive exchange a "plea to God."

At some point, the attorney general's office stopped chasing leads entirely. It declined Farak's offer of a detailed confession in exchange for leniency, nixing the offer without even negotiating terms.

Having barely investigated her, prosecutors indicted Farak only for the samples in her possession the day she was caught. And so, when she pleaded guilty in January 2014, Farak got what one attorney called "de facto immunity." She was sentenced to 18 months in jail plus five years of probation.

Verner's "marching orders," he later testified, were to prosecute Farak with "what was in front of us, the car, things that were readily apparent. And then the bigger investigation was going to be someone else."

But unlike with Dookhan, no one launched a bigger investigation of Farak. And when defense attorneys tried to do it themselves, Coakley's office blocked their efforts.

In fall 2013, a Springfield, Massachusetts, judge convened hearings with the explicit aim of establishing "the timing and scope" of Farak's "alleged criminal conduct."

The attorney general's representative at these hearings was Assistant Attorney General Kris Foster, a recent hire. She had never quashed a subpoena before, but supervisors told her to fend off motions about Farak.

Foster's first step—per ethical obligations and office protocol—should have been to look through the evidence to see what had already been handed over. Instead, she submitted an intentionally vague letter to the judge claiming defense attorneys already had everything. Cleverly omitting pronouns, she wrote that "after reviewing" the file, "every document…has been disclosed." In court, she added that there was "no smoking gun" in the evidence.

In 2017, a different judge ruled that Foster's actions constituted a "fraud upon the court," calling the letter "deliberately misleading." He didn't buy her quibbling that there's a difference between an explicit lie and obfuscation by grammar.

Foster said that Kaczmarek told her all relevant evidence had been turned over and that her supervisor told her to write the letter, though both denied these claims. Asked for comment, Foster in January objected through an attorney that the judge never gave her an opportunity to defend herself and that his ruling left an "indelible stain on her reputation."

Without access to the diaries, the Springfield judge in 2013 found that Farak had starting stealing from samples in summer 2012. Thus, only defendants whose evidence she tested in the six-month window before her arrest could challenge their cases.

This very well could have been the end of the investigative trail but for a few stubborn defense lawyers, who appealed the ruling. A year later, in October 2014, prosecutors relented, granting access to the full evidence in Farak's case to attorney Luke Ryan. He was floored when he found the worksheets.

"It would be difficult to overstate the significance of these documents," Ryan wrote to the attorney general's office. "Whether law enforcement officials overlooked these papers or intentionally suppressed them is a question for another day."

Fallout

The four years since Ryan discovered Farak's diaries have been a bitter fight over this question of culpability—whether Kaczmarek, Foster, and their colleagues were merely careless or whether they deliberately hid crucial evidence.

The Massachusetts Supreme Judicial Court ruled in 2015—by which time the current state attorney general, Maura Healey, had been elected—that it was "imperative" for the government to "thoroughly investigate the timing and scope of Farak's misconduct." State prosecutors gave Farak the immunity they had declined to grant two years earlier, then asked when she started analyzing samples while high.

Her answer: more than eight years before her arrest.

Defense lawyers doubled down on challenges to every case she might have tainted—not just her own, which district attorneys ultimately agreed to dismiss, but also her co-workers', based on Farak's admission that she stole from other chemists' samples. The defense bar also demanded answers on how such crucial evidence stayed buried for so long.

"I suspect that if another entity was in the mix"—perhaps the inspector general or an independent investigator—"the Attorney General's Office would have treated the Farak case much more seriously and would have been much more reluctant to hide the ball," Ryan writes in an email. One reason that didn't happen, he says: "the determination Coakley and her team made the morning after Farak's arrest that her misconduct did not affect the due process rights of any Farak defendants." Because the attorney general had "portrayed Farak as a dedicated public servant who was apprehended immediately after crossing the line, there was also no reason…to waste resources engaging in any additional introspection."

In the only quasi-independent probe of the Farak scandal ever ordered, Attorney General Healey and a district attorney appointed two retired judges to investigate in summer 2015. Without even interviewing Foster, they determined there was "no evidence" of obstruction of justice by her, by Kaczmarek, or by any state prosecutor.

Fortunately, the courts largely ignored this shallow investigation. In June 2017, following hearings in which Kaczmarek, Foster, Verner, and others took the stand, a judge found that Kaczmarek and Foster together "piled misrepresentation upon misrepresentation to shield the mental health worksheets from disclosure."

Verner, who testified that he didn't "micromanage" Kaczmarek, escaped criticism.

The state's top court took an even harsher view, ruling in October 2018 that the attorney general's office as an institution was responsible for the prosecutorial misconduct of its former employees. The justices ordered Healey's department to cover all costs of notifying all defendants whose cases were dismissed. In a separate opinion in October 2018, the Supreme Judicial Court also ordered the state to return most court fines and probation fees to people whose cases were dismissed; one estimate puts that price tag at $10 million.

Despite such unequivocal findings of misconduct, the court removed language about Kaczmarek and Foster from notification letters to those whose cases have been dismissed, which will be sent out in early 2019.

"Thousands of defendants were kept in the dark for far too long about the government misconduct in their cases," the ACLU and the Committee for Public Counsel Services, the state's public defense agency, wrote in a motion. "The need to inform defendants of government misconduct does not disappear when that misconduct was committed by a government lawyer as opposed to a government chemist."

That motion was denied, and the notice letters will explain Farak's tampering without any mention of prosecutorial misconduct.

In addition to ordering the dismissal of many thousands of cases, the Supreme Judicial Court directed a committee to draft a "checklist" for prosecutors, clarifying their obligation to turn over evidence to defendants.

But a crucial issue was not before the court. Lost in the high drama of determining which individual prosecutors hid evidence was a more basic question: In scandals like these, why are decisions about evidence left to prosecutors at all?

"We shouldn't be in the position of having to be saying, 'Don't close your eyes to the duration and scope of misconduct that may affect a whole lot of cases,'" the exasperated Massachusetts chief justice told prosecutors during oral arguments.

Maybe it's not a matter of checklists or reminders that prosecutors have to keep their eyes open for improprieties. Perhaps, as criminal justice scandals inevitably emerge, we need to get more independent eyes on the evidence from the start.

Reporting for this story was supported by the Fund for Investigative Journalism.